Advances in Experimental Social Psychology, Volume 26

ADVANCES IN

Experimental Social Psychology

VOLUME 26

This Page Intentionally Left Blank

ADVANCES IN

Experimental Social Psychology

EDITED BY

Mark P. Zanna DEPARTMENT OF PSYCHOLOGY UNIVERSITY OF WATERLOO WATERLOO, ONTARIO, CANADA

VOLUME 26

ACADEMIC PRESS, INC. A Division of Harcourt Brace & Company

San Diego New York Boston London Sydney Tokyo Toronto

This book is printed on acid-free paper. @

Copyright 0 1994 by ACADEMIC PRESS, INC. All Rights Reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher.

Academic Press, Inc. 525 B Street, Suite 1900, San Diego, California 92101-4495 United Kingdom Edition published by ACADEMIC PRESS LIMITED 24-28 Oval Road. London NW 1 7DX

International Standard Serial Number: 0065-2601 International Standard Book Number: 0-12-015226-6

PRINTED IN THE UNITED STATES OF AMERICA 9 3 9 4 9 5 9 6 9 1 9 8 BB 9 8 7 6 5 4 3 2 1

CONTENTS

Contributors

.....................................................

ix

Attitudes toward High Achievers and Reactions to Their Fall: Theory and Research Concerning Tall Poppies N. T. Feather 1. 11.

111. IV. V. VI. VII. VIII.

1x.

...

................................ Relevant Theories and Their Implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Research Program ............................ ............ Studies That Varied Status ........................................... General Attitudes toward Tall Poppies: Theoretical Perspectives . ..... Correlational Studies of General Attitudes toward Tall Poppies . . . . . . . . . . . . . Studies of Public Figures Summary and Discussion References . . . . . . . . . . . . .......................

I

3 4 22 23 40 46

69

Evolutionary Social Psychology: From Sexual Selection to Social Cognition Douglas T. Kenrick I. 11.

111.

IV. V. VI.

Introduction . . . . . . . . . . . ............. __._. General Principles of Evolutionary ogy . . . . . . . . . . . . . . . . . . . . . . . . . . Integrating Social Rychology and Evolutionary Biology: Two Programs of Research on Human Mate Selection ........... ... ..... Evolutionary Social Cognition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Combining Traditional Experimental and Evolutionary Paradigms to Examine Mediators of Gender Differences ...................................... Conclusion: Toward a Unified Science of Behavior . . . . . . . . . . . . . . . . . . . . . . . _ .. , References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

15 16 85 101 I10

I13 115

vi

CONTENTS

Judgment in a Social Context: Biases, Shortcomings, and the Logic of Conversation Norbert Schwarz I. 11.

111. IV. V.

Introduction ....................................................... The Logic of Conversation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Communicated Information Comes with a Guarantee of Relevance . . . . . . . . . . Making One's Contribution Informative . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . General Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . , . . . References ........................................................

.

i23 124

128 144

154 156

A Phase Model of Transitions: Cognitive and Motivational Consequences Diane N. Ruble I. 11. 111. IV.

Introduction . , . . . . , . , . . . , . . . , . . . , . , . , . , . , Illustration from Data on Transition to First Mother .................. A Phase Model of Transitions . . . . . . . . . . . . . . . . . . . . . , . . . . . . Theoretical Background .......................

V. VI. VII. VIII.

IX. X.

The Gender Transition . . . . . . The Transition in the Meaning o The Transition in Perceptions of Characteristics . . . . . . . . . . . . . . . . . . Implications of the Phase Model . . . . . . . . . . . . . . . . . . . . . . , . , . . . . , . , . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . .......................................

163 164 166

170 180 181

187 193 197 202 206

Multiple-Audience Problems, Tactical Communication, and Social Interaction: A Relational-Regulation Perspective John H. Fleming I. 11. 111.

IV. V.

VI.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conceptual Overview of the Relational-Regulation Perspective . . . . . . . . . . . . A Tactical Relational-Regulation Analysis of Role Conflict . . . . . . . . . . . . . . . . Distancing and Embracing: Properties and Consequences of RelationalRegulation Behaviors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Communicating Hidden or Mixed Messages in Multiple-Audience Situations . . Conclusions and Future Directions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

.

.

.

.

I

215 216 228 237 255 270 279

CONTENTS

vii

From Social Inequality to Personal Entitlement: The Role of Social Comparisons, Legitimacy Appraisals, and Group Membership Brenda Major I. 11. 111.

IV. V. VI. v11. VIII.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Social Groups and Reactions to Social Inequality: The Cases of Gender and Race . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Toward a Rychology of Entitlement The Impact of Social Inequ Becomes Deserved . . . . . From Social Disadvantage When Inequality Becomes Undeserved ................................. The Consequences of Entitlement . . . . . . . . . . . . . . . . ....... Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . ....

293 295

329 338 344 347

Mental Representations of Social Groups: Advances in Understanding Stereotypes and Stereotyping Charles Stangor and James E. Lange 1. 11.

111. IV. V.

VI. VII.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ....... Stereotypes as Mental Representations . . . . . . . . . . . . A New Way of Defining Stereotypes . . . . . . . . . . . . . . Stereotype Measurement .................... Stereotype Accessibility: What “Comes to M i n d ......... Stereotype Maintenance and Change . . . . . . . . . . . Future Issues: On the Sufficiency of the Abstractionist Approach . . . . . . . . . . . . References ................................... .

........................................................... ........................................

Index Contents of Other Volumes

357

378 392 406

417 427

This Page Intentionally Left Blank

CONTRIBUTORS

Numbers in parentheses indicate rhe pages on which the authors’ contributions begin.

N. T. FEATHER (l), School of Psychology, Faculty of Social Sciences, The Flinders University of South Australia, Adelaide 500 1, Australia JOHNH. FLEMING(2 15), Department of Psychology, University of Minnesota, Minneapolis, Minnesota 55455

T. KENRICK( 7 3 , Department of Psychology, Arizona State UniverDOUGLAS sity, Tempe, Arizona 85287 JAMESE. LANCE (357), Department of Psychology, University of Maryland, College Park, Maryland 20742 BRENDAMAJOR (293), Department of Psychology, State University of New York at Buffalo, Buffalo, New York 14260 DIANEN. RUBLE (163), Department of Psychology, New York University, New York, New York 10003 NORBERT SCHWARZ(123), Institute for Social Research, University of Michigan, Ann Arbor, Michigan 48106 CHARLES STANGOR(357), Department of Psychology, University of Maryland, College Park, Maryland 20742

ix

This Page Intentionally Left Blank

ATTITUDES TOWARD HIGH ACHIEVERS AND REACTIONS TO THEIR FALL: THEORY AND RESEARCH CONCERNING TALL POPPIES N. T. Feather

I. Introduction What kinds of attitudes do people hold toward the person who occupies a very high position in some field of endeavor? How do they react when that person suffers a fall from the top of the ladder? These are very general questions and it could reasonably be argued that they permit no general answer. It is surely the case that attitudes toward a person who has achieved high status will depend on a number of factors that relate to the high-status person, the context, and to whoever is making the judgment, and that these variables will also affect reactions when the high-status person falls from grace. For example, how we feel about a particular high achiever and how we react to a high achiever’s fall will depend upon our perceptions of how the person attained the high position in the first place, whether he or she deserves to be there or deserves to fall, the personality characteristics of the high achiever, how the high achiever behaves while occupying the position of status, and what our own needs and values are. Yet it may also be the case that people hold generalized attitudes toward high achievers in addition to their specific attitudes toward particular high achievers and that one can identify the variables that influence these generalized attitudes. My interest in this possibility was stimulated by the observation commonly made within my culture that Australians like to see those in high positions experience a fall, as when an important politician is caught in a foolish act, when a wellknown business leader loses a lot of money in a stock market crash, or when a much-publicized personality in the TV industry suffers a sudden decline in popularity. These people are called tall poppies, and it is claimed that Australians feel a certain amount of satisfaction when tall poppies are cut down to size and suffer a major reverse in status. The belief that Australians like to see tall poppies fall refers to a generalized attitude toward those who occupy positions of high status. A I l V A h C t S IN tXPFRIMFh’TAl S O C I A L PSY(’HO1I)GY

VOL ?h

1

Copyright 0 1994 by Acadcniic Press. Inc. All rights of reproduction in any form rccerved.

2

N. T. FEATHER

It is a belief that is assiduously promoted by the mass media, and it would be foolish to accept its validity without checking to see whether it is a myth or whether there is some factual basis for it. The Oxford English Dictionary (Simpson & Weiner, 1989) defines a tall poppy in Australia as “an especially well-paid, privileged, or distinguished person.” The 1988 edition of the Australian National Dictionary defines a tall poppy as “a person who is conspicuously successful” and (frequently) as, “one whose distinction, rank, or wealth attracts envious notice or hostility” (Ramson, 1988). The term has a long history, and its origins go back to ancient times. The Roman historian Livy refers to the symbolic decapitation of the heads of the tallest poppies by the elder Tarquinius when the Roman ruler was walking in his garden. This message was conveyed to his son, Sextus Tarquinius, who then rid himself of the chief men of the state of Gabii and thus delivered the state unresisting to the Roman king. The destruction of the tall poppies of Gabii was an exercise in the pursuit of power and self-interest in a particular context. The message conveyed to Sextus Tarquinius was clear: Destroy the leaders if you want to attain power. That message continues throughout history. Those who compete for important stakes actively try to surpass their rivals. Politicians seek to undermine their competitors for political office; business leaders strive for advantage in a competitive market; golfers, tennis champions, and others who have a high profile in sport work to defeat the tall poppies who are in competition with them. These are situations in which a person’s success depends upon the defeat or failure of others who are competing for a valued outcome. Those who are engaged in these competitive conflicts are personally involved in trying to improve their own positions at the expense of others. There are other situations, however, in which the tall poppies are not competitors in some interpersonal struggle but are viewed at a distance. They may be presidents or prime ministers, kings or queens, high-profile entertainers, business leaders, individuals who have risen to the top of the ladder in the field of sport, or high-status people in literature, sciences, and the arts. We do not know them personally but we know about their high status and achievement, usually through the mass media. We may become involved in their fortunes but only from afar. Their actions and outcomes do not impact on us directly in a personal sense, although in time they may affect our lives and those of other people. They are tall poppies in a more distant field. The research to be described in this article was concerned with tall poppies who are viewed impersonally from a distance rather than with tall poppies who are seen as rivals in situations where there is a struggle for achievement or power. The research was conducted in order to identify some of the variables that affect our attitudes toward tall poppies and how we react to their fall. It investigated not only attitudes toward specific tall poppies who were high achievers but also attitudes toward tall poppies in general.

ATTITUDES TOWARD HIGH ACHIEVERS

3

Although the research program was stimulated by my interest in tall-poppy attitudes in Australia, the significance of the topic goes far beyond national concerns. There is a large literature on the achievement motive and the personality characteristics of people who have done well at their chosen pursuits (e.g., Atkinson, 1983; Atkinson & Feather, 1966; McClelland, 1985; Spence, 1983), but by comparison the literature is relatively sparse on how others regard the person in a high position, whether this status be achieved or ascribed. The topic is an important one because it confronts us with the question of how we react to high status and, by extension, to the loss of high status. Systematic study of the topic should provide us with a deeper understanding of the variables that influence our attitudes toward leaders, public figures, and others who have been conspicuously successful, and how we react when they fall or are cut down to size.

11. Theoretical Orientation The research on tall poppies was not guided by a single formalized theory. In that regard it contrasts with those research programs in psychology that test a particular theoretical model set up in advance. 1 was exploring a new area and trying to discover some of the major variables that affect our attitudes toward high achievers and how we react when they fall. The formalization of variables into a theoretical model can sometimes be premature, closing off the possibility of developing a deeper and more systematic account of the fundamental variables that are involved in the phenomenon under investigation. 1 wanted to examine some of the major variables that might affect tall-poppy attitudes, so I cast the net widely, drawing on a number of theoretical ideas. This is not an easy strategy to pursue because one runs the risk of being overwhelmed by complexity, but it is a strategy that is basic to any scientific enquiry as one moves through induction to formalized models that can then be systematically tested. I expected that new theoretical ideas would emerge as the research progressed. That did occur. However, the research program has not led to one elegantly formulated theory about tall-poppy attitudes and reactions to a tall poppy’s fall. We will find that the topic is a complex one and that it can be viewed from different theoretical perspectives depending upon the context and the questions that are asked. Indeed it may turn out in the end that more than one theoretical model will be necessary to account for all of the diverse findings. In succeeding sections I will describe some of the main sources for the theoretical ideas that guided the research program. These main sources were as follows: 1. Heider’s (1958) discussion of how a person reacts to the lot of another person, 2. Attribution theory as related to motivation and emotion,

4

N. T. FEATHER

3. The conceptual analysis of personal and cultural values, 4. Theoretical approaches to the analysis of justice and deservingness, and 5. Social comparison approaches to the analysis of envy. Other theoretical ideas will be presented in the context of particular studies (e.g., social identity theory; the role of authoritarianism). What follows is a distillation of the main ideas from these various sources that were used in the research program. The focus will be on those aspects of each approach that turned out to be useful in the research that was conducted. As we will see, the variables suggested by each approach do not form unrelated sets. Some come together, especially in a theoretical analysis of deservingness that I will subsequently describe.

111. Relevant Theories and Their Implications A. REACTIONS TO ANOTHER’S LOT Heider’s (1958) influential book, The Psychology of Interpersonal Relations, contains a number of theoretical ideas that are relevant to understanding how individuals react to tall poppies who occupy positions of high status. His seminal contributions to attribution theory and balance theory are well known and will be referred to subsequently. Less well known is his discussion of how individuals react to another’s lot. His analysis is somewhat cryptic in style but it is a rich source of ideas. The variables that Heider considers involve liking or sentiment relations; unit or “closeness” relations; pressures that relate to “oughtness,” values, and justice; status and power differences and their interactions in situations that involve one person ( p ) and another person (0).Overall, there is the assumption, central to Heider’s analysis of interpersonal relations in general, that perceived relations tend to be structured so that they fit together in a consistent or balanced way. Some of the ideas from Heider’s analysis of reactions to another’s lot emerge again in more formalized form in recent developments in social comparison theory, especially in the conceptual analysis of envy. Some of them also foreshadow the formal analysis of deservingness that I will refer to later. The degree to which a tall poppy is perceived to deserve the high position that he or she holds and the degree to which he or she is seen to deserve any fall that might occur from that high position are key variables that affect attitudes to a tall poppy’s lot. Heider (1958) describes four possible reactions of a person ( p ) to the lot of another person (0).First, o might have a positive experience that is positive forp. Second, o might have a negative experience that is negative for p. These two

ATTITUDES TOWARD HIGH ACHIEVERS

5

types of reaction are concordant, and they presuppose conditions of sympathetic identification. The first case involves sympathetic enjoyment, the second involves sympathy or compassion. The two remaining reactions are those in which there is a discordance or antagonism between p and o. First, o might have a positive experience that is negative for p . Second, o might have a negative experience that is positive for p . The former reaction may be called envy. The latter reaction involves malicious joy or Schadenfreude. In the present context it should be evident that these four reactions are all possible responses to a tall poppy’s rise or fall. Under what conditions do we experience sympathetic enjoyment about a high achiever’s success as opposed to feelings of envy? Under what conditions do we experience sympathy and compassion about a high achiever’s fall as opposed to feelings of Schadenfreude or malicious joy? Answers to these questions are suggested in Heider’s subsequent discussion. He follows his formal statement of possible reactions with a detailed discussion of how these reactions are related to other psychological variables that involve cognitions, wishes, and sentiments. This discussion also refers to balance tendencies in p’s reactions to 0’s lot. For example, whether p likes or dislikes o will influence his or her reactions to o’s lot. Person ( p ) may be unhappy about the success of an enemy and happy when the enemy fails. Thus, in a context involving competition or a struggle for power, one would be unhappy when a tall poppy who is a rival succeeds and happy when the tall poppy falls. Reactions to another’s lot will also be affected by the relation of the lot to the perceived worth of o. As Heider (1958) states, “p may feel great satisfaction that o got what he deserved, whether in a positive or negative sense; and he may feel annoyed or disgusted when 0’s lot does not agree with 0’s worth as he sees it” (p. 283). These reactions in turn tie in with the requirements of justice. Heider (1958) illustrates this connection as follows: “If a , through his own efforts, obtains a good to which he has no right, then he violates the ought force and p will think that o deserves to be punished. Similarly, if o sacrifices a good beyond the call of duty, he will be seen as deserving of reward . . . the situation is balanced if the experiences of another are in accord with what he deserves” (pp. 283-284). Again these statements are relevant to tall poppy attitudes. They imply that one would be more satisfied about the success of a tall poppy if the success was seen to be deserved rather than undeserved. Similarly, one would be less upset about the fall of a tall poppy if the fall was seen to be deserved rather than undeserved. Note, however, that although these statements can be related to how individuals react to high achievers or tall poppies and to their actual or hypothetical fall, they are not restricted to tall poppies but apply more generally. They are not unique to those who hold high-status positions.

6

N. T. FEATHER

Heider (1958) considers the status variable in his discussion of the interaction between the lots of p and o, especially in relation to envy. He notes that comparing the lots of p and o may have the effect of enhancing or undermining the selfconcept. Hence a person’s reaction to the lot of another may be highly influenced by personal needs relating to the maintenance or enhancement of the selfconcept. Heider (1958) discusses the envy that may occur when ”. . . p feels bound up with o in some way, similar to o, in the same class with o, in the same group” (p. 287). When lots are unequal and when this “closeness” is present, there may be a wish as well as action to eliminate the inequality. These reactions would occur in a way that promotes consistency or balance in the overall system of relations. If the relative lots or outcomes cannot be altered, then other changes may occur in the direction of achieving a balanced or harmonious set of relations. Thus, the sentiment or liking relations between p and o may change, p’s evaluation of 0’s outcome may alter, or there may even be a change in the status relations between p and o. Again, Heider (1958) provides us with examples of these various possibilities (p. 287). He also notes that envy does not necessarily follow inequality in the fortunes of people who are close. The personal relevance of the outcome for p is an important variable. If the comparison of lots has little importance for p , then there may be no envy. Thus the “. . . mother may not be envious of her son’s military distinctions because she has no aspirations in this direction” (Heider, 1958, p. 287). Furthermore, the reaction of envy may be controlled by strong sentiment or liking relations between p and o, sympathetic pleasure being evoked at 0’s positive outcome rather than envy. The tendency to equalize lots when p and o are close or form some sort of unit may become bound up with considerations of justice and partake of the character of an ought force, that is, outcomes should be equal because p and o are close. Heider (1958) notes that in such cases the force toward equality may act against p’s personal interests. He also indicates that not all envy is related to the requirements of justice or the tendency toward equalization of lots. Nor is every case where p desires or attempts to obtain what another person possesses an expression of envy. Other emotions and motives may also be involved. The goal in wanting x may be “. . . either to harm o or to have x, whereas in envy, p desires x because another person has it” (Heider, 1958, p. 288). Heider recognizes, however, that p might not be close to o and may regard o as belonging to another class. The other person may be seen as more valuable or more worthy, and the value difference between p and o may then justify or offset the difference in their lots. This discussion of envy is clearly relevant to an analysis of attitudes toward tall poppies. Heider’s discussion implies that whether or not a person envies a high achiever would depend upon the closeness of the high achiever to the person, the relevance of the high achiever’s success to the person, and the

ATTITUDES TOWARD HIGH ACHIEVERS

7

person’s needs relating to self-maintenance and self-enhancement. Note also that there is a clear line between Heider’s analysis and subsequent more formalized discussions of social comparison envy (e.g., Salovey, 1991 ; Salovey & Rodin, 1984; Tesser, 1988, 1991). A discrepancy in lots may also determine an adjustment in the status of the person concerned. As Heider (1958) puts it, . . . should p have more than o, he will easily think that after all he is a better person and deserves more. This relieves him of the pressure to share his fortunes and of possible guilt feelings. Likewise, p will not be disturbed should he have less than o if he feels that he deserves less. (p. 291)

Finally, Heider considers the case in which a person is secure in his or her power and value. In such cases p may feel pleasure when o also has value or even has a superior position. However, where p has doubts about his or her value and where this value is based on comparison with others, p may try to improve his or her position by relevant actions if he or she has the power to do so. Where this power is lacking, p may develop new values that in some way justify the weakness or inferior position. The fox may see the grapes as sour. In the present context, the person with low self-esteem and blocked opportunities may disparage the tall poppy or high achiever in various ways. It is clear that Heider’s analysis suggests variables that are likely to be important when considering attitudes toward tall poppies and reactions to their fall. For example, we can ask whether a tall poppy is liked, has value or worth, is envied, or deserves the high position that he or she holds and how these variables affect reactions to the tall poppy’s fall. We can consider attitudes toward tall poppies and their possible fall in relation to a person’s own position and power and a person’s level of self-esteem. We can also examine interrelations between variables in order to discover whether they form a system and follow the balance principle. Subsequently we will find some clear links between Heider’s analysis of reactions to another’s lot and the kinds of variables that were included in the research program on tall poppies.

B. ATTRIBUTION THEORY Heider’s (1958) discussion of how a naive observer interprets the causes of actions and outcomes is also relevant to understanding attitudes toward tall poppies and reactions to the fall of tall poppies. The distinction that Heider makes between the effective personal force and the effective environmental force as determinants of an action outcome, his separation of the personal force into a power factor and a motivational factor, and the contrast he makes between

8

N. T. FEATHER

dispositional and variable factors, are all seminal ideas that have influenced subsequent developments in attribution theory (e.g., Weiner, 1986). Heider’s further analysis of personal causality and the role of intention enables him to clarify the nature of purposive action and personal control and, by extension, the meaning of personal responsibility. Thus he argues that . , , intention is the central factor in personal causality . . . that brings order into a wide variety of possible action sequences by coordinating them to a final outcome . . . [If] we account for an act by a person’s stupidity or clumsiness, that is by ability factors, we tend to hold him less responsible than if we take the act as an indication of his motives. . . . People are held responsible for their intentions and exertions but not so strictly for their abilities. (Heider, 1958, p. 112)

When attribution for an action outcome also involves some appeal to environmental factors such as luck or task difficulty, personal responsibility “. . . then varies with the relative contribution of environmental factors to the action outcome; in general, the more they are felt to influence the action, the less the person is held responsible” (p. 113). One would expect that these variables that concern the causal interpretation of an action outcome will influence how we react to the success of tall poppies and to their fall. For example, people may rejoice in a tall poppy’s achievement if the success followed years of hard work and reflected high ability. There may be little sympathy for a person who attained a high position because of good luck and who then failed to maintain that position because of lack of effort or carelessness. People are judged differently according to whether or not they are deemed to be responsible for their successes or failures. Note that this form of analysis that refers to causal attributions for action outcomes applies generally and not only to those who occupy high positions. There is now an extensive literature that is concerned with the role of causal attributions in motivation and emotion, a literature containing theoretical developments and empirical studies that were directly stimulated by Heider’s (1958) pioneering contributions. Thus, Weiner (1986, 1992) has described reactions that may follow success or failure in relation to an attributional theory of motivation and emotion that gives special emphasis to the locus of a cause (internal or external), the stability of a cause (stable or unstable), and the controllability of a cause (controllable or uncontrollable). The latter causal dimension has also been discussed in relation to an attributional analysis of stigmas (Weiner, Perry, & Magnusson, 1988). Weiner et a]. (1988) showed that others are more likely to react with pity, liking, no anger and with help-giving when the onset of the stigma was seen to be outside of a person’s control than when the stigma was seen to be an outcome of a person’s own volitional and controllable behavior. Others will feel more sympathy, for example, for a blind person whose blindness was the result of an accident caused by a co-worker than for a blind person whose blindness was due to repeated carelessness.

ATTITUDES TOWARD HIGH ACHIEVERS

9

I have argued that emotional reactions involve a motivational base as well as a complexly determined attributional base (Feather, 1992a). That is, the feelings that we experience are related not only to the way we construe situations and events; they also depend upon our needs and values. These links between affective reactions and underlying needs and values have largely been ignored by attributional theorists. Heider (1958) discussed the role of ought and value in the context of interpersonal relations and the ways in which considerations of justice may influence reactions to the lot of another person. More recently, however, the emphasis has been on relations between affect and particular kinds of causal attribution rather than on relations between affect and a person's needs and values (e.g., Weiner, 1986). Three examples will suffice to indicate how affect may relate to values and needs as well as to causal attributions. In the first place, in Western societies high achievement that can be attributed to individual effort is likely to be praised when compared with high achievement that can be attributed to external causes that are beyond the control of the person (e.g., Covington & Omelich, 1979; Weiner & Kukla, 1970). Indeed Weiner (1986) considers that ". . . the pattern of data reported by Weiner and Kukla (1970), especially the documented reward for trying and punishment for not trying, has been replicated many times and can be considered an established truth (p. 146). He relates affective reactions in this case to an attribution to an internal, controllable cause (effort or lack of effort). However, it should also be clear that the affect that is experienced depends not only on this type of attribution but also on the fact that hard work is valued in individualistic, achievement societies whereas laziness is condemned. If this were not the case, the affective reactions of praise or disapproval would not occur. As another example, guilt depends not only on how individuals construe situations in terms of causal influences but also relates to the violation of ethical norms that involve values that are central to the self-concept (Weiner, 1986). People feel guilty when they lie to others or behave dishonestly (Weiner, Graham, & Chandler, 1982). They condemn others who have succeeded by dishonest means, and they feel satisfied when the unethical behavior of others is punished. As a final example, people tend to experience positive affect when their needs are satisfied and to experience negative affect when their needs are frustrated or unfulfilled. It has been assumed, for example, that those with strong achievement needs will be more likely to be pleased about their successes than those with weaker achievement needs, and more displeased when their achievement needs are frustrated (Atkinson & Feather, 1966). An important implication of this analysis is that the affective reactions that are associated with the success or failure of high achievers or tall poppies will be related to a complex network that involves not only causal attributions but also underlying needs and values. In the next section I consider the relevance of value theory to the analysis of attitudes toward tall poppies and reactions to their fall.

10

N.T. FEATHER

C. VALUE THEORY in a wide-ranging discussion Heider (1958) conceptualizes oughts and values as “. . . impersonal objective requirements which are also dispositional in character and intersubjective in validity” (p. 242). Both concepts involve an appeal to some suprapersonal objective order, some system of rights and preferences that extend beyond the individual and that have the validity of objective existence. However, whereas values and likes are assumed to belong to the dimension of potentiality, oughts and wants are assumed to belong to the dimension of force or force field. Thus, Heider (1958) asserts that . . . “ p values x” can give rise to many different actions, or to none at all, whereas “p ought to do x” means there is an actual objective force present. The relation between “ p values x” and “ p ought to do x” is analogous to the relation between “p likes X” and “p wants to do x.” (p. 224)

Lewin’s (195 1) earlier discussion of the conceptual properties of values also recognized that although values influence behavior, they do not have the characteristics of a goal (or force field). He saw values as having an important relationship to power fields where “. . . the concept of power refers to the ‘possibility of inducing forces’ of a certain magnitude on another person” (Lewin, 195I, p. 40). Lewin (1951) considered that it was probably correct to say “. . . that values determine which types of activity have a positive and which a negative valence for an individual in a given situation. In other words, values are not force fields but they ‘induce’ force fields” (p. 41). The legacy of Lewin is apparent in some of my own contributions to value theory, especially in the way 1 conceptualize relations between values and actions using expectancy-value theory (Feather, 1990a, 1992b). However, I would argue that although values are influenced by a suprapersonal order that involves one’s culture and its institutions, they also take on a personal meaning and identity as they become incorporated into the self-concept. Following Rokeach (1973) 1regard values as properties of persons that involve general beliefs about desirable or undesirable ways of behaving and about desirable or undesirable goals or end-states. They differ from wants and needs in that they involve a dimension of goodness and badness and they have an oughtness quality about them. Although wants and oughts may coincide, they may also be in opposition. For example, a person may feel that he or she ought to behave in a certain way (e.g., honestly) to achieve a certain end, and the person may also want to behave in that way. In other cases wanting something that is attractive (e.g., eating a lot of cake) may violate norms and values that relate to a desirable condition or potential outcome (e.g., maintaining a healthy body). The desired is not always the desirable, although as Heider (1958, p. 233) points out, balance tendencies may lead to some conjunction between the two.

ATTITUDES TOWARD HIGH ACHIEVERS

11

Values are also assumed to transcend more specific attitudes toward objects and situations, but they influence the form these attitudes take. Values provide standards or criteria that people use to evaluate actions and outcomes, to justify opinions and conduct, to plan and guide behavior, to decide between different alternatives, to engage in social influence, and to present self to others. Values are also assumed to be core aspects of the self-concept and to be closely linked to the affective system. They are certainly not affectively neutral. We feel happy when our values are fulfilled, sad and angry when they are frustrated. Moreover our values vary in their importance for self. They become organized into hierarchies of importance or value systems with some values highly important to a person and others relegated to positions lower down in the hierarchy. Values are considered to be relatively stable aspects of the person but not unchangeable. The relative importance of different values may alter over the course of the life-span as people take on new roles and responsibilities (e.g., getting a job and raising a family). Value systems vary across individuals, groups, and cultures and their antecedents are complex, involving culture, society and its institutions, and personality. My treatment of values allows for these different antecedents and functional properties, but it also conceptualizes values as motives (Feather, 1990a, 1992b). There would probably be little debate about the assumption that values can function as motives and influence a person’s actions. Like needs, the values that people hold affect their initiation of new goal-directed activities, the amount of effort they put into an activity, how long they persist at an activity in a given direction in the face of alternative activities, the choices that they make between different activities, the way they construe and appraise situations, and how they feel when the goal-directed activity that they undertake is successful or unsuccessful according to the standards that are set. That is not to say, however, that values are the same as needs. As indicated previously, needs are not necessarily connected to an evaluative framework. We talk about the strength of a need but not whether a need is good or bad. However, values are tied to a normative base that involves a dimension of desirability or goodness-badness. Indeed, philosophers in earlier times sometimes provided lists of human virtues, using the term virtue in a way that is similar to a value that a person might hold to be important for self. Values are also usually held to more readily verbalized and closer to conscious awareness than are many underlying needs (Weinberger & McClelland, 1990). Depending on definition, needs can be conceived as either momentary states of the person (e.g., an immediate need for food) or as relatively stable dispositional properties of persons (e.g., a general need to achieve). However, as I noted previously, values are conceived to be relatively stable in nature when compared with the more transient states of the person, although the relative importance of values for a person can alter as a result of socialization and other life experiences.

12

N . T. FEATHER

How do these various discussions relate to research on tall poppies? I have examined the role of values in three different ways in the research program to be described subsequently. In the first place, I have related respondents’ attitudes toward tall poppies in general to the value priorities that the respondents hold. It can be argued that people who assign high priority to values concerned with achievement and power will be more disposed to want to reward tall poppies for their success and high status when compared with those who value achievement and power less, The former individuals would also be less likely to want to see tall poppies fall. In contrast, people who hold strong equalitarian values would be expected to be less likely to want to reward tall poppies and more likely to want to see them fall when compared with those who place less value on equality. Thus, our attitudes toward those who achieve considerable success within a society and our reactions to their fall would be influenced in different ways by values concerning achievement and power on the one hand and equality on the other. Second, we can extend the analysis of the role of values to consider cultural differences in attitudes toward high achievers and reactions to their fall. Cultures differ in their characteristic values, and we would expect these value differences to be associated with differences at the attitudinal level. For example, some cultures (e.g., Japan) have a collectivist orientation and emphasize values concerned with harmony and interpersonal relations, whereas other cultures (e.g., the United States) place more value on individualism (e.g., Markus & Kitayama, 1991; Triandis, 1990). A culture and its institutions may be identified as a basic component of the supraindividual order referred to by Heider (1958) in his discussion of ought and value. The literature on values across cultures is now extensive and refers to different values and value orientations that characterize cultures in their totality (e.g., Feather, 1975, 1986a; Fiske, 1991; Schwartz, 1992; Triandis, 1990). Some research in our program has examined differences between Australian and Japanese students in regard to their tall-poppy attitudes (Feather & McKee, 1992, 1993). Third, values are important variables in determining how individuals construe events and outcomes. Following Lewin (1951), I have argued that values like needs can induce valences on events and outcomes. They influence a person’s subjective definition of a situation so that some objects, activities, and potential outcomes within the immediate situation become invested with goal properties and are seen as having positive valence (they become attractive) or negative valence (they become aversive), to use Lewin’s (1936) terminology. Thus, needs and values when activated are assumed to influence a person’s cognitiveaffective appraisal of a situation in relation to both means and ends. As I have stated previously (Feather, 1992b): Just as a hungry person sees food as attractive and an insecure person sees the environment as threatening, so a person who values freedom sees the restoration of threatened

ATTITUDES TOWARD HIGH ACHIEVERS

13

freedoms as worthwhile and desirable, and a person who values honesty sees dishonest ways of behaving as reprehensible and not to be undertaken. In each of these cases, actions and their possible outcomes become linked to the cognitive-affective system via a person’s dominant needs and values. The immediate situation therefore takes on affective meaning, with some potential actions and outcomes having positively valent characteristics, and other potential actions and outcomes having negatively valent characteristics. (p. 1121

How general needs and values become linked to positive and negative goal structures within specific situations requires detailed conceptual analysis. I have considered this question in another context, drawing on cognitive models that deal with information processing, affective networks, and attitude accessibility (Feather, 1990a). The assumption that values induce valences on objects, events, and outcomes is involved in a conceptual model of deservingness to be described in the next section. In this model the primary interest is in how a person comes to judge whether another’s outcome is deserved or undeserved. The conceptual analysis is applied to specific tall poppies who hold a high position, and it acknowledges that judgments of deservingness are closely linked to the value priorities that a person holds.

D. AN ANALYSIS OF DESERVINGNESS The model of deservingness that I have applied to the analysis of attitudes toward specific tall poppies (e.g., to well-known public figures) combines ideas from attribution theory, balance theory, and value theory. The model was developed as an attempt to address the question of what variables influence a person’s judgment that another person deserves or does not deserve an outcome. The question is a complex one, and the philosophical, psychological, and legal literature contains various attempts to answer it, with discussions concerned with such issues as justice, rights, responsibility, and blame. The model to be described is a relatively simple one and can be extended in the future as new variables are added to the formalization. However, it does incorporate variables that are assumed to be important in regard to how people decide whether an outcome is deserved or not deserved. I assume that one variable that influences whether or not a person is seen to deserve an outcome is the degree to which he or she is perceived to be personally responsible for it, Personal responsibility for an outcome is assumed to be attributed to a person when the outcome is seen to be produced by the person and related to the person’s intentions. This usage is consistent with discussions in the psychological and legal literature. For example, both Piaget’s (1932) analysis of moral judgment and Heider’s (1958) definition of different levels of responsibility attribution included intentionality as an important determinant of respon-

14

N. T. FEATHER

sibility. It is also an important component in Shaver’s (1985) analysis of the attribution of blame for a negative event and in legal definitions of responsibility (Fincham & Jaspars, 1980; Hart, 1968; Hart & Honori, 1959). Mikula (1993) also includes perceived intention as one of a number of variables that are assumed to influence perceptions of injustice. I have linked responsibility to volition in the structural balance model of communication effects (Feather, 1964, 1967), and, as we have seen, Weiner (1986) includes controllability as a dimension of causes in his attributional theory of motivation and emotion. These various approaches, as well as commonsense usage, suggest that outcomes and their consequences are more likely to be perceived as undeserved when a person is seen not to be responsible for them, that is, when they are unintended or due to uncontrollable causes that have an internal or external locus. The tall poppy who rose to high status because of a lucky event (e.g., being born into a wealthy family) may be seen to deserve the high status less than a person whose high status followed years of hard work. Luck would be perceived as an uncontrollable cause, whereas effort is a cause that is within the control of a person. Does this mean that judgments of deservingness always involve some perception that the person being judged caused the event and is accountable for the outcome and its consequences? 1 have argued that although responsibility is a key element, the analysis of deservingness also has to consider other variables, especially those related to values and justice (Feather, 1992a). In particular, one can relate deservingness to valued behaviors and valued outcomes. Specifically, the conjunctions of a positively valued behavior with a positively valued outcome and a negatively valued behavior with a negatively valued outcome are both assumed to determine judgments of deservingness. The two structures in Fig. la and Fig. l b illustrate these cases. In these structures a positive relation is indicated by a solid line and a negative relation by a dashed line. Action and outcome are bound together by a positive unit relation represented by a bracket; that is, the outcome is associated with or belongs to the action. Both structures are balanced in Heider’s (1958) sense (i.e., the relations fit together in a consistent manner). In contrast, outcomes are assumed to be seen as less deserved when positively valued behavior leads to negatively valued outcomes and when negatively valued behavior leads to positively valued outcomes. These two situations are represented in Fig. lc and Fig. Id. Both structures are unbalanced in Heider’s (1958) sense. This analysis disentangles controllability and value. Although outcomes (whether good or bad) are assumed to be perceived as not deserved when they occur outside of a person’s control and intentions (i.e., when the person is deemed to be not responsible for them), there may be outcomes that follow controllable, instrumental behavior that are seen to be undeserved. The model

ATTITUDES TOWARD HIGH ACHIEVERS

15

DESERVED OUTCOMES Person

Person

a

b

UNDESERVEDOUTCOMES

Person

Person

C

Fig. I . Balanced and unbalanced structures representing deserved outcomes and undeserved outcomes. Reprinted from Feather (1992a) with permission of the author and the British Rychological Society.

implies that negative outcomes that follow a person’s controllable and positively valued behavior would be seen to be undeserved. Similarly, positive outcomes that follow a person’s controllable and negatively valued behavior would be seen to be undeserved. For example, failure at an exam (a negatively valued outcome) that followed positively valued behavior (e.g., preparation and hard work) would be perceived as undeserved by most people; success at an exam (a positively valued outcome) that followed negatively valued behavior (e.g., cheating) would also be perceived as undeserved. In each of these examples the person would be seen as responsible for the instrumental action that led to the outcome, but the outcome would be seen to be undeserved. It might be argued that in some of these cases one could still infer that the person was not responsible for the outcome because of the operation of other causes beyond the person’s control. For example, when a student fails an exam despite a lot of preparation and hard work, it might be inferred by an observer that the exam was extremely difficult for everyone and that the person was therefore not responsible for the failure. Even when such alternative interpretations are possible, however, it can still be argued that the particular form taken by the conjunction of a positively or negatively valued instrumental behavior and a positively or negatively valued outcome is an important variable in the analysis of deservingness. This conjunction sets the analysis of deservingness within the

16

N. T. FEATHER

means-end context of purposive behavior, where action takes place within a defined goal structure and is influenced by needs and values. As applied to tall poppies or high achievers, this analysis implies that when high status is seen as due to controllable and positively valued behavior (e.g., hard work), the positively valued high status is seen as more deserved than when it can be attributed to negatively valued behavior (e.g., dishonest practices). Similarly, when a tall poppy falls from the top of the pedestal, the negatively valued fall is seen as more deserved when it can be attributed to negatively valued behavior (e.g., laziness or lack of effort) than when it follows positively valued behavior (e.g., diligence and hard work). In these examples, the subjective values assigned to the instrumental behaviors and the outcomes may be related to the strength of values concerning honesty and achievement within each person’s value system. As indicated previously, underlying general values are assumed to induce subjective values (or valences) on specific actions and their associated outcomes (Feather, 1990a, 1992b). Other theorists have recognized the important role that norms and values play in judgments of deservingness and entitlement. The extensive literature on distributive justice and equity describes general principles (e.g., proportionality, equality, need) that people may use to judge the fairness of allocations in different contexts, providing a means of determining whether or not a particular level of reward is deserved or appropriate (e.g., Deutsch, 1985; Feather, 1990b, 1991c; Mikula, 1980). Relative deprivation theory also refers to feelings of deservingness and their relation to general principles of distributive justice and rules, requirements, or standards (e.g., Crosby & Gonzales-Intal, 1984; Prentice & Crosby, 1987). Lerner (1980, 1991) has argued that people develop a belief in a just world where people get what they deserve and this more primitive belief may override socially determined normative prescriptions when the situation is emotionally involving. Shaver’s ( 1985) analysis of moral responsibility also acknowledged that judgments of moral responsibility presuppose some comparison of behavior with ethical standards. One would expect that attitudes toward tall poppies and affective reactions associated with their success or failure would be related to the degree to which the high achiever is seen to be responsible for his or her position and deserves to be there. The deserved success should be accompanied by more positive affect than the undeserved success for both an actor and an observer; the undeserved failure should be associated with more negative affect than the deserved failure for both an actor and an observer (Feather, 1992a). Note, however, that this analysis is not restricted to tall poppies but applies to all people who experience deserved or undeserved outcomes, whatever the status that they occupy. The analysis that has been developed so far has been related to the relatively simple structures that were presented in Fig. 1. But the analysis can be taken further by adding as an element the other person (0)and his or her relation to

17

AITITUDES TOWARD HIGH ACHIEVERS DESERVED OUTCOMES INGROUP OTHER

Olher[#.

OUTGROUP OTHER

Person

Action

Outcome

Person

/In\

P R \\

Other(/;

,

*\

/

/d

d

b

Outcome

*\

/

\

Action

a

Person

, . ” ; ( r eh tO

d

/

Person

Outcome

Acllon

b

\

Action d

b

Outcome

d

C

UNDESERVED OUTCOMES INGROUP OTHER Person

OUTGROUP OTHER Person

__

__

e

f

Person

9

Person

_ I

h

Fig. 2. Balanced and unbalanced structures representing deserved outcomes and undeserved outcomes and including ingroup other and outgroup other.

person ( p ) . For example, is p assumed to like, dislike, or feel indifferent toward o? Is o assumed to belong to p’s group, to be a member of some outgroup, or to be unconnected to p? Figure 2 presents eight structures that include both person ( p ) and other (0) as separate entities. In these structures a solid bracket denotes a positive unit relation and a dashed bracket denotes a negative unit relation. In Fig. 2 the positive bracket linking p and o is used to represent the situation where both p and o belong to the same ingroup; the dashed bracket linking p and o is used to represent the situation where o is a member of an outgroup from p’s point of view. The positive bracket linking other and action denotes that o is bound to the action, i.e., is seen to own the action or to be responsible for it. As in Figure 1, solid lines denote that actions or outcomes are positively valued by person ( p ) ; negative lines denote that actions or outcomes are negatively valued, again from p’s point of view. The structures in Fig. 2 have interesting implications as far as ingroup and outgroup effects are concerned. The only structures that are balanced in Heider’s (1958) sense are the structures in Fig. 2a and Fig. 2d. These two structures

18

N. T. FEATHER

respectively represent a situation where an ingroup member is seen to deserve a positively valued outcome and a situation where an outgroup member is seen to deserve a negatively valued outcome. These relations fit together in a consistent way. However, the other two structures that are classified in Fig. 2 as implying deserved outcomes are unbalanced in Heider’s (1958) sense, i.e., they involve inconsistent sets of relations. These structures respectively represent a situation where an ingroup member is seen to deserve a negatively valued outcome (Fig. 2b), and a situation where an outgroup member is seen to deserve a positively valued outcome (Fig. 2c). In both of these cases p may not be as willing to judge a as deserving the outcome when compared with the former two situations where the sets of relations are balanced. There may also be a tendency in the two unbalanced situations for p to judge o as less responsible for the action that o took (e.g., o may be seen to have been coerced by external pressures to act in that way or to be the victim of difficult circumstances). In these cases the unit relation linking other and action may be represented as negative rather than positive and the two structures would then be balanced. In a sense, the other would be seen as not owning the action. The action is outside of the person’s control. This perception of reduced responsibility for action would moderate p’s judgement that o deserved the positive or negative outcome because o is seen as less responsible for it. Similar comparisons can be made for those situations where outcomes are perceived by p to be undeserved. Here it may be the case that a negatively valued outcome is seen as more undeserved and a positively valued outcome as less undeserved when o is a member of p ’ s ingroup (Fig. 2e and Fig. 2f) than when o is a member of an outgroup (Fig. 2g and Fig. 2h). All of the structures in Fig. 2 that relate to undeserved outcomes are unbalanced and the imbalance in these structures would be difficult to resolve. The structures in Fig. 2 extend the analysis of deservingness. Further extensions to the model would involve allowing for positive, negative, or null liking relations between p and o. With these modifications the theoretical approach could be applied to the analysis of the effects of both social identity (e.g., ingroup versus outgroup) and interpersonal relations on perceptions of deservingness.

E. SOCIAL COMPARISON ENVY Situations in which one person is aware of the high achievement and high status of a tall poppy may involve social comparison between self and other. The literature on the social comparison of opinions and abilities is extensive (e.g., Suls & Wills, 1991). Wheeler (1991) has recently provided a brief history,

ATTITUDES TOWARD HIGH ACHIEVERS

19

beginning with the seminal article by Festinger (1954). The goals of social comparison include self-evaluation, self-improvement, and self-enhancement (Wood & Taylor, 1991), and the choice of a target with whom a person compares herself or himself may involve an upward comparison or a downward comparison. The literature on social comparison also encompasses research on relative deprivation and justice (e.g., Folger, 1984; Masters & Smith, 1987). The comparison of self with a high achiever or tall poppy would in most cases involve an upward comparison along the dimension that defines the status difference. Recent developments in social comparison theory enable one to specify some of the variables that are important in accounting for how a person deals with these upward comparisons. Of particular interest are theoretical analyses that consider feelings of envy. It can be argued that individuals may sometimes rejoice at a tall poppy’s fall because they are envious of the tall poppy’s status and attainments. Because of this possibility I included envy as a variable in some of the studies in my research program. Two examples of conceptual approaches to the analysis of envy will be provided: Tesser’s (1988, 1991) analysis of comparison and reflection processes, and Salovey’s (1991) discussion of social comparison processes in envy and jealousy. As noted previously, Tesser’s (1988, 1991) model of self-evaluation maintenance is clearly related to Heider’s (1958) discussion of reactions to another’s lot. Tesser’s model considers the effects of social comparison when it occurs along relevant or nonrelevant attributes but does so in the context in which the other is a close associate of the person or in some form of unit relation, to use Heider’s (1958) term. The “pain of comparison” occurs when unfavorable comparisons are made with others on self-relevant dimensions and where positive selfevaluation is under threat. When comparison with a higher performing other is made on a dimension that has low personal relevance, the person may bask in the reflected glory of the other (Cialdini, Borden, Thorne, Walker, Freeman, & Sloan, 1976). Tesser (1988, 1991) assumes that performance, closeness, and relevance interact in a systemic way. For example, when relevance is high, a person might interfere with or derogate a close other’s superior performance. When relevance is low, a person might try to facilitate the superior performance of a close other. Tesser (1 988, I99 1) reports laboratory studies involving behavioral variables that provide some validation for the hypothesized reflection and comparison processes. He also describes studies that use emotion as a marker on the assumption that when the comparison involves a threat to self-evaluation, a negative emotion should be experienced; when reflection results in a boost to selfevaluation, a positive emotion should be experienced. For example, a study by Tesser and Collins (1988) found that jealousy and envy were manifestations of the comparison process when self-evaluation was under threat; pride was higher when self outperformed the other. Both of these effects were stronger when the

20

N. T. FEATHER

comparison involved a high-relevance dimension. However, closeness of the other had no significant effect in either case. Tesser (1991) argues that emotional arousal plays a role in the unfolding of self-evaluation maintenance behaviors together with attributions concerning the source of the emotion. Salovey and Rodin (1984) provided an analysis of envy or social comparison jealousy that was influenced by Silver and Sabini’s (1978) assertion that social comparison envy will result when the possessions, attributes, and attainments of one person have the effect of diminishing the status of another person (see also Sabini & Silver, 1982). They assumed that envy would be reported in situations where there was negative feedback or negatively valenced information about oneself relative to another person, when the information had high self-relevance, and when the other person was similar or close to self in various respects. The results of their study showed that subjects reported significantly greater envy or jealousy toward another person when they were told that their standing on a selfrelevant characteristic was worse than that of the successful other on the same characteristic. Subjects also disparaged the successful other under these conditions, and they were less likely to desire his or her friendship. They also tended to feel more depressed and anxious about interacting with the comparison person. The Salovey and Rodin (1984) analysis is clearly similar to Tesser’s (1988, 1991) model of self-evaluation maintenance (see also Salovey & Rothman, 1991). In a recent discussion, Salovey (1991) notes that what seems especially important in envy are “. . . threats to self-evaluation that result from loss of status relative to comparison others” (p. 268). Salovey (1991) discuses the emotions of envy and jealousy and their interrelationship using a schematic framework that is linked to an analysis by Bryson (1977) and, more distantly, to Heider (1958). Like Heider (1958) and Foster (1972), Salovey (1991) recognizes that individuals may try to conceal their envy in various ways, because it can be a destructive force that interferes with the fabric of social relationships (see also Schoeck, 1969). It will be evident that the two examples of recent developments in social comparison theory that I have chosen to describe show considerable overlap in the variables that they consider. Both models involve a comparison other who is close to the person and both also emphasize the degree to which the attribute along which comparison occurs is relevant or irrelevant, having implications for the self. Similar variables are also contained in a recent discussion of upward and downward social comparison in relation to esteem-relevance and perceived control (Major, Testa, & Bylsma, 1991). The Tesser (1988, 1991) and Salovey (1991) models imply that tall poppies or high achievers who are close to the person will be envied and denigrated in various ways if their superior performance is on a relevant dimension. The Major et al. (1991) analysis adds the variable of personal control over attempts to change the relative standing of self with comparison other as a further influence on the type of response that is made to the difference in relative standing.

ATTITUDES TOWARD HIGH ACHIEVERS

21

It is important to note, however, that not all tall poppies or high achievers are close to the person who may make a comparison. The conspicuously successful person may be a public figure who is known only through media reports and not through personal contact. Such people may also be envied. As noted previously, a tall poppy or high achiever may also be a competitor or rival rather than someone who is psychologically close to the person. There may be strong motivation to defeat the rival and claim the prize for oneself. The fall of the tall poppy then implies the person’s success. The general point to be made is that the recent models of social comparison that I have described may be limiting because they focus on comparison others who are close or similar to the person. There are other kinds of personiother relations that should be considered, including null relations and negative relations. Moreover, the closeness variable has a certain looseness about it and requires more precise definition. Not all kinds of closeness are the same or have the same implications. Two people may be physically close (e.g., two adjacent workers in a factory, students who share a room in a hall of residence) yet psychologically distant. The failure to find an effect of closeness in some of the recent studies from social comparison research may reflect this looseness. The studies to be described later in this article did not include closeness as a variable but were concerned with tall poppies considered at a distance. It is evident that tall-poppy attitudes can be considered generally, relating to any person who is conspicuously successful, whether the person is close or otherwise. The social comparison theories that I described are pertinent in that they deal with the emotion of envy and include relevance as a variable. One implication of these theoretical analyses is that tall poppies will be envied more when their high status and attainments are along a dimension that is relevant to the person making the social comparison than when they are not relevant. Another implication is that feelings of envy may be associated with other negative affective reactions to a tall poppy’s high status and achievement and with positive affective reactions (e.g., Schadenfreude) when a tall poppy falls. There are other analyses of envy that involve a phenomenological point of view, drawing distinctions between envy and jealousy and envy and resentment (Parrott, 1991; Smith, 1991). For example, Smith (1991) argues that the experience of envy typically involves both discontent and hostility. The hostility may be a product of the frustration involved when someone else does better than self, thereby highlighting one’s inability to obtain a desired attribute at a higher level. Smith (1991) believes, however, that the hostility is usually related to a sense of injustice that is associated with a person’s disadvantaged position, a claim that is consistent with research on relative deprivation, equity theory, and justice-related motives (e.g., Folger, 1987; Masters & Smith, 1987; Walster, Walster, & Berscheid, 1978). Smith (1991) sees hostile envy as involving a sense of injustice that has personal rather than universal validity (see also Heider, 1958; Scheler, 1915/ 1961). That the hostility has a subjective, unsanctioned character may

22

N . T. FEATHER

account for the fact that envy is often concealed from others (see also Foster, 1972). When the hostile feelings are legitimized and shared by others, they may be expressed more freely. In this case one is dealing with resentment proper rather than hostile envy. Note that Smith (1991) also recognizes that there may be qualitatively different forms of envy (see also Parrott, 1991). These discussions of the subjective nature of envy reinforce the conclusion that the conceptual analysis of attitudes toward tall poppies should include a justice component that relates to deservingness and entitlement. As noted in the previous section, affective reactions to a person’s positive and negative outcomes will partly depend on whether or not these outcomes are seen to be deserved.

F. SUMMARY The theoretical approaches and ideas that I have discussed are those that involve variables and forms of analysis that were relevant to the studies in the research program that will be described in the remainder of this article. We will find that these studies included variables that were concerned with causal attributions, values, deservingness, and affective reactions to a tall poppy’s lot (including envy). It should be apparent that attitudes toward tall poppies may involve a mixture of emotions and that affective reactions following the fall of a tall poppy may also reflect mixed feelings. The tall poppy may be admired but also envied, respected but also distrusted, be a popular figure but one who is not without blemish. When a tall poppy falls, there may be a certain amount of sympathy but also some degree of satisfaction in observing the discomfiture of a person who has fallen from a high position. This review of theoretical approaches is not meant to exhaust the range of theoretical ideas that can be applied to the analysis of tall poppy attitudes and reactions to their fall. I have concentrated on those theoretical approaches that I used to frame my research program. However, from time to time I will refer to other theoretical ideas in the context of specific studies as these ideas become relevant.

IV. The Research Program The research program on tall poppies has involved both experimental and correlational studies. In some studies I used hypothetical scenarios in which the status variable was manipulated. Respondents’ attitudes to high achievers and average achievers were compared, and their reactions were investigated when the high or average achiever suffered a fall. These studies were specifically designed

ATTITUDES TOWARD HIGH ACHIEVERS

23

to investigate whether in fact there is a tall poppy effect, that is, whether respondents would report more pleasure (or less unhappiness) when a high achiever fell than when an average achiever fell. In other studies 1 have been more concerned with investigating specific variables in addition to status that may influence respondents’ attitudes toward high or average achievers and their fall. For example, I have examined the effects of personal responsibility, deservingness, causal attributions, the personality characteristics of the stimulus person, and the domain in which the achievement occurred on how respondents react to the high or average achiever. Still other studies have involved the use of a specially constructed tall poppy scale to measure generalized attitudes toward tall poppies and to investigate the correlates of these attitudes. The main correlates that have been investigated are global self-esteem, measures of the respondents’ own relative standing, the importance of different values relating to such domains as achievement, power, equality, and right-wing authoritarianism. Finally, I have begun to investigate tall poppy attitudes at the cross-cultural level, using students from Australia and Japan. The cross-cultural research enables one to address the question of whether there are features of tall poppy attitudes that are distinctive to Australia. The use of hypothetical scenarios in questionnaires runs the risk of not engaging respondents sufficiently. The judgments made by respondents may occur under conditions of low personal involvement without much interest or commitment on their part. Under these conditions the responses that are provided may reflect “cold” cognition, and they may differ from the sorts of responses that would be made in realistic situations where the respondents are personally involved and where the affective system comes into play. In an attempt to overcome this problem I have also investigated attitudes toward prominent public figures who are real-life tall poppies, although this procedure removes the experimental controls that one can exert when presenting hypothetical scenarios. Although I have sampled beyond student populations in some of the studies, the bulk of the research has involved respondents who either were attending university courses or were in the senior grades of high school.

V. Studies That Varied Status 1 begin by describing studies that used vignettes or scenarios in which the initial status of the stimulus person was experimentally manipulated and where this status declined as a result of performance or other factors. A major interest in these studies was to examine the effects of status on attitudes toward the stimulus person and reactions to the stimulus person’s fall in different contexts. In addi-

24

N. T. FEATHER

tion, later studies investigated deservingness in more detail and cross-cultural differences. A. STUDY 1 : A FALL IN ACADEMIC PERFORMANCE The first study was concerned with tall poppies who had a consistently high level of performance at school when compared with students whose level of performance was average (Feather, 1989). I manipulated the extent to which the stimulus person with high or average status fell from his or her original position. Would subjects report some degree of satisfaction or Schadenfreude when a stimulus person fell? If this “malicious joy” occurred, would it be more pronounced for the tall poppy or high-achieving student than for the average student? Would these affective reactions depend upon the extent of the fall? It might be the case that subjects report sympathy or unhappiness about any type of fall rather than malicious joy, and that these feelings are related to the nature of the fall (e.g., whether the fall in performance takes the person to near the bottom of the class or to near the average performance level of the class). I expected that the high and average achievers would be perceived differently both in relation to their personalities and to the causes of their performance. I also expected that how subjects reacted to the fall of the high and average stimulus person would be related to their initial attitudes toward the stimulus person. Stronger initial attraction to a stimulus person should be accompanied by more regret and disappointment when a fall occurs. Negative attraction or dislike for a stimulus person may be accompanied by feelings of pleasure when that person falls. The subjects who participated in this study were 531 male and female high school students in metropolitan Adelaide who were all enrolled in Year 11 classes, the penultimate grade before completing high school. These subjects completed a questionnaire that contained a scenario describing a stimulus person who was either a high or average achiever and either male or female. Subsequently the high or average achiever was described as suffering a fall in academic performance in the very important final public examination at high school that qualifies students for tertiary studies. The fall could take one of three different forms: from high performance to middle performance, from high performance to low performance, and from middle performance to low performance. ’A completely crossed design would involve two levels of status (high achiever, average achiever) and two levels of fall (latge fall, medium fall). Such a design could not be realized in the present study because the average achiever cannot fall further than to the bottom of the scale and that fall (from middle to bottom) is classified as a medium fall rather than as a large fall. Note also that the fop, middle, and bottom of the scale refers to a person’s position relative to others and that the middle position on the performance scale corresponds to average performance, representative of the group as a whole.

25

ATTITUDES TOWARD HIGH ACHIEVERS

Subjects used a variety of scales to rate the high and average achiever in regard to causal attributions for their performance, personality characteristics, and attraction, before information about the fall was provided. The causal attribution items referred to ability, effort, an easy task, good luck, and friendly teachers. The personality characteristics were based upon a factor analysis of the ratings of the high and average achievers along 26 semantic differential, bipolar adjective scales. Three scales were derived from this analysis: Positive qualities, introversion, and achievement orientation or assertiveness. Attraction was assessed by using seven items that referred to close social interaction with the stimulus person, to perceived similarity, and to liking. A factor analysis justified the use of a composite scale of attraction based on all seven items. Table I shows that subjects attributed the high achiever's performance more to ability, effort, and to friendly teachers and less to good luck in comparison with the average achiever's performance. The high achiever was also seen as having more positive qualities, as being more introverted, and as much more achievement oriented and assertive when compared with the average achiever. Note, however, that although the total positive qualities score was significantly higher for the high achiever, scores on two components of this scale (sociable, good mixer) were significantly lower than for the average achiever. So the picture that emerged was of a high achiever who was perceived to be ambitious and with positive qualities but who was also seen as more reserved and less sociable than TABLE I MEANSCORESFOR CAUSAL ATTRIBUTIONS, SEMANTIC DIFFERENTIAL SCALES,AND ATTRACTION FOR HIGHACHIEVER AND AVERAGE ACHIEVER BEFORE FAILURE Mean scoreso Variable Causal attributions Ability Effort Easy task Good luck Friendly teachers Semantic differential Positive qualities Introversion Achievement orientation Attraction

Midpoint of scale

High achiever

Average achiever

df

F

4 4 4 4

5.02 6.12 3.17 2.07 3.57

4.19 5.02 3.34 2.57 3.28

1,514 ,517 ,514 ,518 ,516

49.34** 96.34** 1.14 15.44**

32 24 32 28

39.19 28.97 39.02 28.69

37.54 26.53 34.25 28.63

,494 ,492 1,480 1,509

6.50* 30.92** 70.93** .I0

4

5.05*

"N = 348 for high achiever; N = 183 for average achiever. Minor variations from these Ns occurred because of missing cases. Reprinted from Feather (1989) with permission of the author and the Australian Psychological Society. * p < .05. * * p < ,001.

26

N . T. FEATHER

the average achiever. Table I also shows that the mean attraction scores were almost identical for the high and average achievers. Thus, there was no evidence that the high achiever was liked less than was the average achiever. After subjects were presented with information that the stimulus person suffered a fall in performance, they provided ratings for a further set of items. These items concerned whether the stimulus person deserved the mark that he or she obtained in the final examination, causal attributions for the stimulus person’s fall (lack of ability, lack of effort, difficult exam, bad luck, unfriendly teachers), whether they would feel privately pleased about the stimulus person’s fall, how they thought the stimulus person would feel about the fall, and whether they would feel more friendly toward the stimulus person, about the same as before, or less friendly. Table I1 shows that there were no statistically significant effects for type of fall when the causal attribution ratings and deserve-to-fall scores were analyzed. Lack of effort was seen as the most important cause of the fall followed by the difficulty of the examination. The stimulus person was seen to be very displeased with the fall in all conditions but especially when the fall was from top to bottom. Of greater interest are the results that concern respondent satisfaction and more friendly attitude. Table I1 shows that subjects in general provided ratings that

TABLE I1 MEANSCORES FOR CAUSAL ATTRIBUTIONS A N D OTHERPOST-FALL VARIABLES IN RELATION TO TYPEOF FALL^ Mean scoresh Variable Causal attributions Lack of ability Lack of effort Difficult task Bad luck Unfriendly teacher Deserve to fall Respondent satisfaction Stimulus person’s satisfaction More friendly attitude

Midpoint of scale

4 2

High to middle

High to bottom

Middle to bottom

df

3.48 5.10 4.57 2.62 3.02 4.34 3.57

3.49 5.04 4.77 2.73 2.88 4.05 3.16

3.72 5.01 4.42 2.85 3.18 4.31 2.78

2,512 2,514 2,514 2,513 2,513 2,509 231 1

.06 1.87 .86 .94 1.91 11.38***

1.72 2.16

I .33 2.12

I .66 2.04

2,513 2,506

6.49** 3.40*

F

I .40

aReprinted from Feather (1989) with permission of the author and the Australian Psychological Society. b N = 175 for high to middle fall; N = 173 for high to bottom fall; N = 183 for middle to bottom fall. Minor variations from these N s occurred because of missing cases. *p

< .05. * * p < .01. * * * p < ,001.

ATTITUDES TOWARD HIGH ACHIEVERS

27

were below the neutral point of the scale for respondent satisfaction about the fall; that is, generally they were unhappy about the fall. Thus they did not report experiencing Schadenfreude or malicious joy. However, they expressed least displeasure about the fall when the stimulus person fell from high to middle and most displeasure when the fall occurred from middle to bottom. Similarly, the most friendly attitudes occurred when the fall was from high to middle and the least friendly attitudes when it was from middle to bottom. The “painvise” comparison of the means for respondent satisfaction showed that all of the differences between means were statistically significant for the three possible comparisons. However, in the case of friendly attitude, the difference between the means for the high to middle fall, when compared with the middle to bottom fall, was the only significant difference. These results could be interpreted as indicating some preference for the average position in the classroom, a preference that may reflect the influence of equalitarian values. When a tall poppy in the classroom falls to the average position she or he becomes more similar or equal to others. In contrast, when an average achiever falls away from the middle of the class, he or she is removed from the collectivity that is defined by the middle or normative range of accomplishment where most others are located. Falling to the average position in the middle of the scale was the least aversive outcome and falling away from the middle of the scale was the most aversive outcome in terms of how pleased the subjects were about the fall and how friendly they felt they would be toward the stimulus person after the fall. The middle or average position may be taken as the group norm, the position that represents the collectivity and that defines some degree of equality. Perhaps a fall to the average position also made the high achiever appear more human in the eyes of the subjects. Helmreich, Aronson, and Le Fan (1970) found that a competent person was liked more among subjects with average self-esteem after that person suffered a pratfall or blunder (clumsily spilling a cup of coffee), although subjects who were either high or low in self-esteem were more attracted to the stimulus person when he or she did not blunder. Deaux (1972) also found that a highly competent individual who made a blunder was preferred to one who did not, but only when males were judging males and not in the other possible combinations. Helmreich et al. (1970) explained the increased liking for the competent stimulus person in their study as probably due to the humanizing of the competent individual. The foolish mistake made him closer and more similar to the person who was average in self-esteem and therefore more approachable. According to Helmreich et al. (1970), subjects who were high in self-esteem may have seen the competent stimulus person as similar in status to themselves and they would therefore be more contemptuous of this person when he made a foolish mistake. Those subjects who were low in self-esteem may have liked the competent person less when he made the blunder because they needed a perfect hero to respect and could not tolerate any imperfections. Note that this interpreta-

28

N. T.FEATHER

tion draws attention to the effects of similarity and the possible influences of values concerned with equality and collectivism. In regard to the results reported in Table I , the Helmreich et al. findings suggest that subjects may prefer a situation where a fall makes a tall poppy more like themselves and in that sense more part of the human collectivity. It is important to note that in the present study a fall below the average position would denote a failure on the scale of performance. It is not surprising that subjects, who were students themselves, would report feeling especially displeased about this outcome (see Table 11). Note also that although there was evidence that the high achiever and the average achiever were initially perceived differently in regard to their personalities and the causes of their performance (Table I), there was no evidence that subjects were initially more negative toward the high achiever. The results did show that there were statistically significant negative correlations between initial attraction toward the stimulus person and subjects’ reports about how pleased they felt about the stimulus person’s fall for all three fall conditions; that is, the less they liked the stimulus person, the more pleased they were about the fall.*

B. STUDY 2: A FALL RESULTING FROM A MISDEMEANOR

The main focus of the first study was on the effects of a fall in relative academic performance for a high or average achiever in the classroom. In this first study no information was provided about the cause of the decline in performance. Subjects were left to make their own interpretations. The second study also concerned a high achiever and an average achiever who suffered a fall, but in this case the cause of the fall was specified (Feather, 1989). Subjects were informed that the fall was the result of a misdemeanor (cheating at an examination). Specifying the cause of the fall in performance made it possible to relate the results to some of the theoretical ideas presented in preceding sections of this article, especially in relation to attribution theory and the analysis of deservingness. I expected that subjects would report being more punitive in their actions toward a transgressor when the misdemeanor was committed by a high achiever than by an average achiever and that they would be more pleased about the high achiever’s fall. zVery few gender effects occurred in this study or in Study 2. Gender effects were either very limited or somewhat inconsistent throughout the studies in this research program. There was some evidence for gender differences in values and self-esteem, with females having more of a prosocial, communal orientation and reporting lower self-esteem. The reader is referred to the published papers for details.

A’ITITUDES TOWARD HIGH ACHIEVERS

29

Why might this be so? Information that a high achiever has cheated may change one’s perceptions of a high achiever and the way one attributes causality for the high achiever’s success. The high achiever’s success in the past may be seen as due to cheating, and the benefits and privileges of the high achievement may be seen as undeserved. Considerations of justice and equity would demand that the high achiever “pay the piper” and be punished more than the average achiever who cheated. Both would be perceived as deserving some form of sanction or punishment, but the high achiever would be seen to deserve a stiffer penalty. It could also be argued that high status confers on a person extra responsibility for setting a good example to others, though some idiosyncrasies might be allowed (Hollander, 1964). A high achiever who commits a misdemeanor would transgress normative expectations about appropriate behavior and would be seen as deserving a greater penalty than would a person who had not achieved high status. It may also be the case that a transgression such as cheating provides an opportunity for observers to cut the high achiever down to size, making the tall poppy more like others. This latter interpretation corresponds to the analysis presented in relation to the first study, but it generalizes it to a fall produced by a transgression that violates moral standards. There were 361 male and female subjects who were involved in this study. They were enrolled in an introductory psychology class at Flinders University. The procedure was similar to that used in the first study. Subjects completed a questionnaire that presented subjects with a hypothetical scenario describing a male or female person who was either a high or average achiever on university examinations. They then rated the stimulus person on the same 26 semantic differential scales that were used in the first study and on the seven items that were designed to measure attraction. However, measures of causal attribution for the performance were not included. Factor analyses led to derived scales of positive qualities, introversion, achievement orientation and assertiveness, and attraction that were essentially the same as those that were used in the first study. Subsequently, information was presented that the high or average achiever had been discovered cheating on a major examination and subjects answered questions that concerned the actions they would take about the cheating and how they would privately feel. The results showed that, prior to subjects receiving the information about cheating, the high achiever was perceived to be significantly more achievement oriented and assertive than the average achiever and significantly less sociable and less of a good mixer when compared with the average achiever. These results were consistent with those of the first study. The average achiever had significantly higher scores on the positive qualities variable when compared with the high achiever, a difference that was in the reverse direction to that found in Study 1. This difference in results may reflect the effects of sampling from different populations in different contexts (high school students in Study 1; university

30

N . T. FEATHER

TABLE 111 MEANSCORESON POST-CHEATING DEPENDENT VARIABLES FOR HIGHACHIEVER AND AVERAGE ACHIEVER" Mean scores Variable Report to authorities Discuss with others Pleased about discovery Penalty for cheating Deserve penalty Pleased about expulsion

Midpoint of scale

High achiever

Average achiever

4 4 4 3.5 4

2.86 4.64 4.81 3.32

2.09 4.42 4.33 3.04

4.17

3.46

4

3.52

2.93

df

F

1,351 1,353 1,353 1,352 1,352 1,351

16.06**

1.20 7.96* 9.41* 15.98** 19.40**

aN = 361. Minor variations from this N occurred because of missing cases. Reprinted from Feather (1989) with permission of the author and the AustraIian Psychological Society. * p < .Ol. * * p < ,001.

students in Study 2). Introversion scores and attraction scores were very similar for the high and average achiever in Study 2 (see Feather, 1989, Table 4). After subjects were presented with information that the stimulus person had been discovered cheating, they provided ratings on a further set of items. These items concerned whether they would report the stimulus person to the authorities, whether they would discuss the cheating with their fellow students, how they would privately feel if the stimulus person was caught cheating by someone in authority, what level of penalty should be exacted by a disciplinary committee, to what extent the stimulus person deserved to be expelled from the university, and how pleased they would be if the stimulus person were expelled. Table 111 shows that subjects reported feeling significantly more pleased when the high achiever was caught cheating than when the average achiever was caught and also more pleased when the high achiever was expelled from the university than when the average achiever was expelled. They also indicated that they would be more likely to report the high achiever to the authorities, that they would award a stiffer penalty to the high achiever for cheating, and that they considered that the high achiever deserved the penalty more when compared with the average achiever. As in the first study, significant negative correlations were obtained between initial attraction to the stimulus person and reported feelings of pleasure relating to the fall. These results support the hypothesis that subjects will report being more punitive when a misdemeanor is committed by a high achiever or tall poppy than by an average achiever and that they will also report feeling more pleased about the high achiever's fall. Note that in this second study subjects reported that the high achiever deserved to be punished more than the average achiever for cheat-

ATTITUDES TOWARD HIGH ACHIEVERS

31

ing whereas there were no differences in deservingness relating to the status of the stimulus person in the first study. The two studies differed, however, in that the cause of the fall (cheating) was made explicit in the second study whereas no information was provided in the first study about the causes of the fall in academic performance. A transgression can trigger feelings that concern justice, equity, and retribution. When a high achiever commits a misdemeanor, subjects’ perceptions of the high achiever and how he or she obtained high status can change. Subjects may perceive that the mistake made by the high achiever is more serious and has wider consequences because the high achiever has acted irresponsibly and has set a bad example to others. If we focus on the other side of the coin, the results indicated that subjects reported feeling less pleased when the average achiever was punished, a finding that is consistent with the low ratings of respondent satisfaction in Study 1 when the average achiever suffered a fall. Thus, in both studies, there was evidence that the fall of the average achiever was viewed by subjects in a more sympathetic manner.

C. STUDY 3: FURTHER RESEARCH ON DESERVINGNESS As we have seen, the second study extended the investigation of the fall of a tall poppy to a new context, one that involved a moral transgression (cheating). Cheating would generally be perceived as a negatively valued behavior. Consistent with the conceptual analysis of deservingness that I presented in an earlier section, the results showed that subjects perceived a penalty (a negatively valued outcome) to be deserved when cheating occurred, though the high achiever was seen to deserve the penalty more than did the average achiever. The main aim of the third study (Feather, 1992a) was to conduct a detailed investigation to test the attributional and value analysis of deservingness. It will be recalled that in this model deservingness was assumed to be related to a person’s responsibility for an outcome and to whether positively (or negatively) valued actions lead to positively (or negatively) valued outcomes. I also included the status variable in this third study along with conditions in which initial performance was either subsequently maintained or fell. The study was a complex one and not all findings from it will be reported. A key feature of the design was the provision of information about the causes of success and failure. The scenarios described a student (gender unspecified) who was either a high achiever or an average achiever at school. Information about the causes of the high or average achievement was vaned in three ways. In one condition no information was provided; in a second condition the stimulus person was described as a student who doesn’t possess a lot of natural ability but who

32

N. T. FEATHER TABLE IV VARIABLES AND F VALUES FOR MAINEFFECTS FROM UNIVARIATE MEANSOF DEPENDENT ANOVAs (SUCCESSOUTCOME)^ Information about stimulus person Variable

No information

High effort/ average ability

Low effort/ high ability

df

F for information

5.75, 5.80, 4.94, 3.65 4.80,

6.04, 5.96, 5.05, 3.43 4.96,

4.69, 5.02, 4.06, 3.44 3.72,

2,163 2,163 2,162 2,163 2,162

13.01** 6.47* 8.72** 0.27 10.89**

5.43

5.32

4.89

2.163

1.94

~~

Responsible Deserve Pleased Envy Praise Stimulus person’s pleasure

ON = 169; minor variations in Ns occurred across variables due to missing cases. Means in each row not sharing a common subscript differ significantly; p < .05 (Thkey). Reprinted from Feather (1992a) with permission of the author and the British Rychological Society. * p < .01 **p < ,001.

puts in a lot of effort (high effort/average ability); in a third scenario the stimulus person was described as a student with a lot of natural ability who doesn’t study very hard (low effort/high ability). The subjects in the study were 689 male and female high school students in their final two years of study. In one condition 169 subjects responded to a scenario in which the stimulus person maintained the high or average performance level on the final examinations at the end of high school. They rated the high or average achiever on items that asked how pleased they privately felt that the stimulus person obtained the final score, whether the stimulus person deserved that final score, how responsible the stimulus person was for obtaining the final score, how much envy they privately felt toward the stimulus person because of the final score that was obtained, how much they thought the stimulus person should be praised for the final score that was obtained, and how pleased they thought the stimulus person would feel about the final score that was obtained. Table IV presents the main results. These results show that the stimulus person whose past performance was attributed mainly to effort (a controllable cause) was rated as significantly more responsible for the successful outcome than the stimulus person whose past performance was attributed mainly to natural ability (an uncontrollable cause). The mean ratings of deservingness, positive affect, and praise were also higher in the high effort/average ability condition where the stimulus person was perceived to be more responsible for the success and where success (assumed to be positively valued) followed high effort (also assumed to be positively valued). The corresponding ratings were lower in the low ef-

33

ATTITUDES TOWARD HIGH ACHIEVERS

Achievement status

F for status

Average achiever

High achiever

df

5.29 5.35 4.20 2.63 3.84

5.61 5.76 5.07 4.28 5.03

1,163 1,163 1,162 1,163 1.162

2.40 3.42 16.04** 36.04** 24.25 * *

4.55

5.78

1,163

22.57**

fort/high ability condition where the stimulus person was seen to be less responsible for the success and where success (assumed to be positively valued) followed low effort (assumed to be negatively valued). These results are consistent with the conceptual analysis of deservingness that was presented previously. Table IV also shows that the mean ratings for pleased, envy, praise, and the stimulus person’s pleasure were significantly higher for the high achiever than for the average achiever. Two statistically significant interaction effects were obtained involving status and information. The mean ratings showed that the high achiever was judged to be significantly more responsible and significantly more deserving of the success in the no-information and high-effortlaverage-ability conditions when compared with the average achiever. However, this difference was reversed in the loweffort/high-ability condition. In this condition, the average achiever was seen as significantly more responsible for the success than the high achiever and significantly more deserving of it. A second part of the study presented 420 male and female subjects with a scenario in which the stimulus person obtained a low score in the final examinations, a score that was a lot worse than expected. I presented three variations of this scenario. In the first variation no further information was provided about the fall. In the second variation subjects were told that the stimulus person had contracted a serious illness just before the examination that affected the stimulus person’s performance. It was assumed that subjects would perceive the stimulus person not to be responsible for the low grade. In the third variation the stimulus person was described as slacking off and not putting in the extra effort required. Thus, the cause of the low performance was represented as related to controllable and intended behavior. It was assumed that the reduction in effort would be

34

N . T. FEATHER

negatively valued behavior and that subjects would perceive the stimulus person to be responsible for the low grade. After reading their scenario, subjects rated the stimulus person on items concerning deservingness and responsibility for the outcome. They also rated how sorry they privately felt for the stimulus person because of the low score, how unhappy they privately felt that the stimulus person obtained the low score, how pleased they thought the stimulus person would feel about the low score, and whether they would be in favor of letting the stimulus person “resit” a second examination with the opportunity of improving his or her score. The results are presented in Table V. The most dramatic effects are those that relate to type of failure situation (see the bottom part of Table V). The stimulus person was seen to be less responsible for the low grade, less deserving of the low grade, and less pleased about the low grade when the cause could be attributed to sickness than when the cause could be attributed to reduced effort or when no cause was provided. Subjects also reported feeling more unhappy about the stimulus person’s failure, more sorry for the stimulus person, and more willing to allow the stimulus person to redeem his or her grade when the low grade could be attributed to sickness (an uncontrollable cause), than to reduced effort (a controllable cause), or to other unknown factors (the no-information condition). Note that again these results are consistent with the conceptual analysis described previously. The low grade (assumed to be a negatively valued outcome) was seen to be deserved more when it followed reduced effort (also assumed to be negatively valued behavior), and there was less evidence of sympathy and willingness to let the stimulus person redeem the low grade in this situation. However, when the stimulus person failed because of sickness, he or she was more likely to be seen as not responsible for the failure. In this situation the failure was seen to be undeserved and subjects showed more sympathy and willingness to let the stimulus person redeem the low grade. Table V shows that information about the stimulus person also affected the ratings provided by subjects when the outcome was a low grade (see the top part of Table V). The mean ratings for responsibility and deservingness were significantly lower in the high-effort/average-abilitycondition when compared with the means in the other two conditions. Subjects also reported feeling significantly more unhappy about the stimulus person‘s low grade and significantly more sorry about the failure when the stimulus person was described as high effort/average ability than when the information referred to low effort/high ability. Thus, once more there was evidence that perceptions of deservingness and affective reactions relating to the stimulus person’s negative outcome depended on the relation between the subjective values of both the action and the outcome itself. In this case, however, the conjunction involved positively valued behavior (high effort) that was followed by a low grade (a negatively valued outcome), a conjunction that, according to the model, would lead to perceptions of undeservingness. Note

35

ATTITUDES TOWARD HIGH ACHIEVERS TABLE V MEANSOF DEPENDENT VARIABLES AND F VALUES FOR MAINEFFECTSFROM UNlVARlATE ANOVAs (FAILURE OUTCOME)^ Information about stimulus person

No information

High effort/ average ability

Low effort/ high ability

df

F for information

sow

4.14, 3.16, 3.76, 4.28,

4.30, 3.40, 3.87, 4.47,

4.99, 4.51, 3.06, 3.59,

2,502 2,501 2,501 2,502

1.51** * 19.84*** 10.18*** 12.63***

Stimulus person’s pleasure Redeem grade

2.04 5.61

2.04 5.86

2.13 5.54

2.498 2.501

Variable Responsible Deserve Unhappy

0.41 1.12

Achievement status Average achiever

High achieve1

df

F for status

sow

4.86 4.15 3.52 3.82

4.51 3.65 3.60 4.39

1,502 1,501 1,501 1,502

2.63 1.56** 0.08 10.40***

Stimulus person’s pleasure Redeem grade

2.06 5.60

2.09 5.13

1,498

Responsible Deserve Unhappy

1,501

0.21 0.40

Failure situation

No information

Reduced effort

Sickness ~

~

df

F for situation

~

sow

5.40, 4.13, 3.33, 4.04,

3.30, 2.58, 3.85, 4.99,

5.31, 5.00, 3.51, 3.21,

2,502 2,501 2,501 2,502

18.75*** 85.06* * * 3.96* 39.11 ***

Stimulus person’s pleasure Redeem grade

2.38, 5.47,

1.65, 6.09,

2.20, 5.44,

2,498 2,501

14.34*** 1.61* * *

Responsible Deserve Unhappy

ON = 520; minor variations in N s occurred across variables due to missing cases. Means in each row not sharing a common subscript differ significantly; p < .05 (Tukey). Reprinted from Feather (1992a) with permission of the author and the British Psychological Society. * p < .05 * * p < .01 * * * p < ,001.

that in this condition subjects may also have assumed that the low grade was obtained because of other causes. For example, they may have assumed that the final exam was very difficult and that the student was therefore not really responsible for the low grade. The results are also consistent with this interpretation. There were two statistically significant effects of achievement status. The high

36

N. T. FEATHER

achiever was perceived to deserve the low grade less and subjects reported feeling more sorry for the high achiever when compared with the average achiever. Note that subjects did not report feeling less unhappy about the high achiever’s failure when compared with the average achiever’s failure, contrary to the results of Study 1 (Feather, 1989). Nor were there any statistically significant interaction effects that involved the status variable in combination with the other variables. There were two statistically significant interaction effects involving information about the stimulus person and failure situation, but discussion of these results is outside the scope of the present report (see Feather, 1992a). In summary, the results of this study supported the conceptual analysis of deservingness that was presented in a previous section. There was clear evidence that a stimulus person’s deservingness for achievement outcomes could be related both to attributions of responsibility and to the relation between the subjective values of actions and outcomes. Affective reactions were also related to the same set of variables. The interpretation of the results involved going beyond attribution theory by including the concept of deservingness (a justice-related concept) and by allowing for the effects of values. The analysis may therefore be seen as building bridges between areas that broadly encompass attributions, justice, values, and affect. There were some effects of achievement status in regard to such variables as envy, positive affect, and sympathy but no evidence that subjects were happier (or less unhappy) when a high achiever or tall poppy fell than when an average achiever obtained a low grade. Hence, the results of the previous two studies were not replicated in this more complex study that presented subjects with more detailed information about the conditions of performance.

D. STUDY 4:A FALL IN DIFFERENT CULTURES

I indicated at the beginning of this article that one reason for embarking on the program of research on high achievers or tall poppies was to examine whether or not there was any validity in the statement that Australians enjoy seeing tall poppies fall. The results of the studies that I have described so far indicate that the statement is far too simple and that a number of different variables influence our attitudes toward tall poppies and our reactions to their fall. However, the statement does have a cultural dimension in its implication that Australians are especially prone to want to see tall poppies fall and to react to their fall with a certain amount of malicious joy or Schadenfreude. We have conducted one cross-cultural investigation in an attempt to discover whether tallpoppy attitudes occur in another culture (Feather & McKee, 1992). This study involved a comparison of how Australian and Japanese students reacted to hypothetical scenarios that again involved high and average achievers who either maintained their achievement status or suffered a subsequent fall. Another part of

ATTITUDES TOWARD HIGH ACHIEVERS

37

the study was concerned with attitudes toward tall poppies in general (Feather & McKee, 1993). Results relating to these more general attitudes will be described in a later section. What kinds of differences might one expect to find between the Australian and Japanese students in their tall poppy attitudes? There is a well-known Japanese proverb that “the nail that sticks up gets pounded down.” This proverb could be interpreted as cautioning people not to rise above the group, and it is consistent with a lot of evidence that suggests that the Japanese culture emphasizes group harmony and cohesion and is collectivist rather than individualistic in its orientation (Feather & McKee, 1992, 1993). Indeed, Markus and Kitayama (1991) describe the Japanese construal of self as one that is likely to emphasize interdependence with the social context rather than independence from others. We would argue that the Australian culture is probably more individualistic in its orientation. Various authors have commented on the mixture of individualism and collectivism in Australia and the tension between an ideology that values mateship, group solidarity, and equality and a system that rewards individual achievement and personal advancement (e.g., Encel, 1970; Feather, 1975, 1980, 1986a, 1986b). Because the scenarios we used in our study emphasized individualistic achievement and did not refer to achievement that was set within a group context, we predicted that the Japanese students would view the high achiever more negatively when compared with the Australian students. Our study involved 127 male and female Australian students who were enrolled in undergraduate courses in psychology at the Flinders University of South Australia in Adelaide in 1989. The Japanese sample consisted of 112 male and female students who were also enrolled in undergraduate psychology courses at Keio University in Tokyo. The students in both samples completed a questionnaire that presented them with a hypothetical scenario describing a male stimulus person who was either a student or a golfer. Within each of these activity domains, the stimulus person was described as either a high achiever or an average achiever. Having read the description of the stimulus person, subjects then provided a personality description by rating the stimulus person on 12 semantic differential bipolar adjective scales. A factor analysis led to two derived scales: sociability and assertiveness. Subjects also responded to seven items that provided measures of attraction or liking for the stimulus person, how pleased they were about the stimulus person’s performance, and how much they envied the stimulus person. After completing these items, subjects were provided with another scenario (matched for domain with the first) that described how the student or golfer performed after a lapse of 10 years. In this scenario, the stimulus person was described as maintaining his previous level of performance or failing to maintain it (suffering a fall). Subjects then again completed the personality description and attitude items in terms of how they would now feel about the stimulus person. There is only space to present a selection of the results. We would expect that

38

N . T. FEATHER TABLE VI MEANSCORESFOR “PLEASED ABOUTPERFORMANCE” IN RELATION TO NATION, INITIAL PERFORMANCE, SUBSEQUENT PERFORMANCE, AND SCENARIOo Australian students

Initial performance High Average

Japanese students

Subsequent performance

N

Initial scenario

Subsequent scenario

N

Initial scenario

Subsequent scenario

Maintain Fall Maintain Fall

25 29 32 30

5.40 5.34 4.47 4.43

5.44 3.72 5. I3 3.53

28 27 28 25

4.21 4.85 3.54 4.52

4.21 2.93 4.39 3.40

“Reprinted from Feather and McKee (1992) with permission of the authors and the Australian Psychological Society.

subjects would be generally more pleased if a stimulus person maintained status rather than suffered a fall. A tall poppy effect would be obtained if subjects reported that they were less unhappy (or relatively more pleased) when a high achiever suffered a loss of status than when an average achiever suffered a fall and if this difference did not occur or was in the reverse direction when a high- or average-status position was maintained by the stimulus person. The means in Table VI show that this effect did not occur for either the Australian or Japanese subjects. There was no evidence for either a statistically significant initial status by subsequent performance by scenario interaction effect or for a statistically significant nation by initial status by subsequent performance by scenario interaction effect. There were significant main effects of nation, subsequent performance, and scenario, but these main effects were modified by higher order interaction effects. One of these was a statistically significant initial status by scenario interaction effect. Subjects were more pleased about the performance of the high achiever than the average achiever following the initial scenario, but there was little difference in the means following the subsequent scenario when the stimulus person either maintained his original position or fell from grace. There was also a statistically significant subsequent performance by scenario interaction effect. For both the Australian and Japanese students there was a sharp decrease in the ratings of pleased about performance when the stimulus person suffered a fall but a small increase in these ratings when the stimulus person maintained his original position. The results did show that the high achiever was perceived to be more assertive than the average achiever and that initially the high achiever was envied more. Subsequently, subjects were more positive in various ways toward the stimulus person who maintained performance when compared with the stimulus person

39

ATTITUDES TOWARD HIGH ACHIEVERS TABLE VII MEANSCORESFOR “ENVY”IN RELATIONTO INITIALSTATUS,DOMAIN, AND SCENARIO^^ High initial performance

Average initial performance

Domain

N

Initial scenario

Subsequent scenario

N

Initial scenario

Subsequent scenario

Student Golfer

56 53

9.13 8.26

6.64 6.70

58 57

5.33 7.11

5.83 6.86

~~

~

~

~~

~


whose performance declined. The results also showed that the golfer was especially favored by the Japanese students. The Japanese students made a stronger distinction between the high achiever and the average achiever in regard to assertiveness than the Australian students, and they were also more negative in various ways than the Australians when the stimulus person suffered a fall rather than maintaining status (see Feather & McKee, 1992). These differences are consistent with other analyses of the Japanese culture that indicate a high level of achievement concern and a fear of losing face (e.g., Christopher, 1982; De Vos, 1973; Lebra, 1976). Consistent wiih the recent extensions of social comparison theory described in earlier sections of this article (e.g., Salovey, 1991; Salovey & Rodin, 1984; Tesser, 1988, 1991), the results also showed that the relevance of the domain influenced subjects’ reported envy. Table VII indicates that in both cultures the high-achieving stimulus person was envied more than the average achiever following the first scenario, but the difference in envy was much greater for the student than for the golfer. The performance of the student was clearly more relevant for the student subjects used in the investigation than was the performance of the golfer. The effects of relevance were specific, however, and did not extend to other variables.

E. CONCLUDING COMMENTS It should be evident that the more recent experimental studies that I have described were more complex in form than the earlier ones, with new variables being added. There was some evidence that subjects were less unhappy about a high achiever’s fall than about an average achiever’s fall in the first two studies where the vignettes or scenarios were relatively simple. As new variables were added, however, the effects of status differences were diminished and the tallpoppy effects that were observed in the earlier studies were not obtained. Even in

40

N. T. FEATHER

the first two studies the effects of status on subjects’ ratings cannot be considered to be substantial. The power of the statistical tests was enhanced by the fact that I used large samples. Given these large samples, the differences in achievement status would account for only a relatively small proportion of the variance in subjects’ responses to the rating scales. The message that comes from the results of these experimental studies is that attitudes toward stimulus persons who are tall poppies or high achievers and reactions to their fall depend upon a range of variables that include the perceived causes of their success and failure, judgments about the extent of their deservingness, the domain in which success and failure occurs, and whether the fall is an extreme one or results in the tall poppy being more like others. Many of these variables exert their influence independent of status. Thus, attitudes toward the high achiever when compared with the average achiever are not necessarily more favorable or less favorable but depend upon the information given and the context of achievement. So do reactions to the fall of a high achiever. There is no justification for any general statement that people are negative toward high achievers and like to see them fall. That is probably the case for “bad” tall poppies who are seen not to deserve their high position, but it is not likely to be the case for “good” tall poppies who are perceived to deserve to be where they are.

VI. General Attitudes toward Tall Poppies: Theoretical Perspectives A. THE TALL POPPY SCALE A second group of studies examined attitudes toward tall poppies in general and the correlates of these attitudes. These studies did not involve vignettes in which the variables of interest were experimentally manipulated. Instead they examined generalized attitudes toward high achievers, using a specially constructed scale called the Tall Poppy Scale. The Tall Poppy Scale is presented in Table VIII. It consists of 20 items, half of which express positive attitudes toward tall poppies and half of which express negative attitudes. Subjects answer each item by using a 6-point scale labeled as follows: + 1, I agree a little; +2, I agree on the whole; +3, I agree very much; - 1, I disagree a little; -2, I disagree on the whole; -3, I disagree very much. A constant of 4 is subsequently added to the item responses so as to obtain a positive scale. The results of factor analyses have consistently indicated that the 20 items can be separated into two subscales corresponding to two different factors. The first

ATTITUDES TOWARD HIGH ACHIEVERS

41

TABLE VIII THETALLPOPPYSCALE^

1. People who are very successful deserve all the rewards they get for their achievements. 2. It’s good to see very successful people fail occasionally. 3. Very successful people often get too big for their boots. 4. People who are very successful in what they do are usually friendly and helpful to others. 5 . At school it’s probably better for students to be near the middle of the class than the very top student. 6 . People shouldn’t criticise or knock the very successful. 7. Very successful people who fall from the top usually deserve their fall from grace. 8. Those who are very successful ought to come down off their pedestals and be like other people. 9. The very successful person should receive public recognition for hislher accomplishments. 10. People who are “tall poppies” should be cut down to size. 11. One should always respect the person at the top. 12. One ought to be sympathetic to very successful people when they experience failure and fall from their very high positions. 13. Very successful people sometimes need to be brought back a peg or two, even if they have done nothing wrong. 14. Society needs a lot of very high achievers. 15. People who always do a lot better than others need to learn about what it’s like to fail. 16. People who are right at the top usually deserve their high position. 17. It’s very important for society to support and encourage people who are very successful. 18. People who are very successful get too full of their own importance. 19. Very successful people usually succeed at the expense of other people. 20. Very successful people who are at the top of their field are usually fun to be with. ~

YItems1 , 4 , 6 , 9 , 11, 12, 14, 16, 17, and 20 comprise the Favor-Reward subscale; items 2 , 3 , 5 , 7 , 8, 10, 13, 15, 18, and 19 compnse the Favor-Fall subscale. Reprinted from Feather (1989) with permission of the author and the Australian Psychological Society.

subscale contains the 10 items that express negative attitudes toward tall poppies and that favor the tall poppy’s fall (e.g., “People who are very successful get too full of their own importance”; “Those who are very successful ought to come down off their pedestals and be like other people”). I call this variable Favor Full. The second subscale contains the remaining 10 items that express positive attitudes toward tall poppies and that favor rewarding tall poppies (e.g., “It’s important for society to support and encourage people who are very successful”; “People who are right at the top usually deserve their high position”). I call this variable Favor Reward. Total scores for each subscale can range from 10 to 70. These two variables, favor fall and favor reward, were the tall poppy attitudes that were investigated in the studies to be presented. These studies used samples from different populations. All of the studies included the Rosenberg measure of global self-esteem (Rosenberg, 1969, usu-

42

N. T. FEATHER

ally in a modified 10-item version (Bachman, O’Malley, & Johnston, 1978). Other variables that were assessed varied from study to study. These variables included value priorities, perceived competence in the classroom, and right-wing authoritarianism. One of the studies involved a cross-cultural comparison of findings from Australian and Japanese students. The studies to be described drew upon some of the theoretical ideas that were presented earlier in this article (e.g., value theory), but they were also related to other theoretical approaches. The main theoretical perspectives are reviewed in the following sections.

B. TALL POPPY ATTITUDES AND VALUES I expected that attitudes toward tall poppies would be related to underlying value priorities, consistent with the discussion of the role of values that I presented previously. In particular, I predicted that tall poppy attitudes would be related to the importance that individuals assigned to values concerned with achievement, power, and equality. There is a tension between these values in Australian society, and similar tensions may exist in other societies because the resolution of conflicts between achievement, status, and equality is a general problem confronting all societies in the search for an acceptable balance between opposing interests. 1 hypothesized that individuals who assigned high priorities to achievement and power values would be more likely to favor rewarding tall poppies and less likely to want to see them fall when compared with those individuals who give these values much less weight in their value systems. An emphasis on achievement and power is consistent with support for high achievers who occupy highstatus positions and also consistent with not favoring their fall. In contrast, I expected that individuals who placed a high value on equality would be less likely to favor rewarding tall poppies and more likely to want to see them fall when compared with those individuals who valued equality less. An emphasis on equality implies a preference for societal arrangements in which status differences are reduced and in which opportunities are equal for all. Equalitarian values are generally opposed to values concerned with achievement and power and, as hypothesized, implications for tall poppy attitudes should reflect this conflict. C. TALL POPPY ATTITUDES AND GLOBAL SELF-ESTEEM Hypotheses that concerned relations between tall poppy attitudes and global self-esteem were developed on the basis of a number of theoretical considerations. I expected that individuals who favored the fall of tall poppies would tend

ATTITUDES TOWARD HIGH ACHIEVERS

43

to be lower in global self-esteem when compared with those who were less in favor of their fall. In contrast, I expected that individuals who favored rewarding tall poppies would tend to be higher in global self-esteem when compared with those who were less in favor of rewarding tall poppies. Similar hypotheses were tested in regard to relations between tall poppy attitudes and perceived competence. It was expected that attitudes favoring the fall of tall poppies would be negatively related to perceived level of competence and that attitudes favoring rewarding tall poppies would be positively related to perceived level of competence. These hypotheses can be related to theories concerned with similarity and attraction, with social comparison and social identity, with the mediating role of values, and with the general effects of frustration and failure in a person’s life. In regard to the latter, it can be argued that attitudes toward tall poppies will reflect the amount of failure and frustration that individuals have encountered in their lives and the amount of envy and other negative feelings that they experience. Disappointment and anger, along with feelings of envy toward those who have moved up the ladder of success, would follow a failure to achieve goals that are important to the individual. Feeling pleased about a tall poppy’s fall may then be linked to an individual’s frustrated ambitions and to the anger and envy that may occur. If one assumes that consistent failure to achieve important goals determines lower global self-esteem and lower perceived competence, then it would also be expected that individuals with low self-esteem and low perceived competence would be more envious of a tall poppy and happier to see a tall poppy fall than would individuals with high self-esteem and higher perceived competence. The latter individuals may be more likely to identify with the tall poppy because the tall poppy is seen as closer to self in relation to achievement status. From the perspective of the similarity/attraction hypothesis (Byrne, 197 l), it can be argued that individuals with higher levels of global self-esteem and higher levels of perceived competence may see tall poppies as more similar to themselves than do individuals with lower levels of self-esteem and lower levels of perceived competence. One would then expect the former individuals to be more attracted to tall poppies than are the latter individuals and perhaps even to identify with them. The same kinds of predictions can be generated from social identity and selfcategorization theory (e.g., Hogg & Abrams, 1988; Tajfel, 1978; Tajfel & Turner, 1985; Turner, Hogg, Oakes, Reicher, & Wetherell, 1987). In this case it can be argued that individuals with high global self-esteem and high perceived competence would be more likely to categorize themselves as high achievers, seeing themselves as belonging to the tall poppy group. They should therefore display respect and admiration toward high achievers and want to see tall poppies rewarded for their achievements. In contrast, individuals with low levels of global self-esteem and low perceived competence would be more likely to have a record of failures and to categorize themselves as low achievers and as not belonging to

44

N. T. FEATHER

the tall poppy group. They may show some admiration and respect toward tall poppies because of their achievements, but they may also display negative attitudes such as envy because high achievers are seen to belong to a different, highstatus group. These various forms of analysis can also be applied to a tall poppy who falls. In this case one can hypothesize that individuals with lower levels of global selfesteem and lower levels of perceived competence would be more likely to hold positive attitudes toward a fallen tall poppy, and that individuals with higher levels of global self-esteem and higher levels of perceived competence would be more likely to hold negative attitudes toward a fallen tall poppy. One can assume that the low-self-esteem/low-competenceindividuals would now see the fallen tall poppy as more similar to themselves, and he or she may be liked and supported more because of this similarity or because the tall poppy has now joined the category of low achievers to which these individuals belong. On the other hand, one can assume that the high-self-esteem/high-competence individuals would now see the fallen tall poppy as dissimilar to themselves, and he or she may be disliked because of this dissimilarity or because the tall poppy is now categorized as belonging to a different group and is no longer the recipient of ingroup favoritism. One can also derive predictions about relations between global self-esteem and tall poppy attitudes that are based on their joint association with underlying values. The results of previous research showed that global self-esteem was positively related to both school performance and to the importance assigned to achievement values by Australian high school and university students (Feather, 1991b). If global self-esteem expresses achievement values within a particular culture, then one would expect global self-esteem to relate positively to favoring the reward of tall poppies and negatively to favoring the fall of tall poppies, because these two variables are assumed to be associated with high and low valuations of achievement respectively.

D. TALL POPPY ATTITUDES AND RIGHT-WING AUTHORITARIANISM The hypotheses concerning tall-poppy attitudes and right-wing authoritarianism were based on Altemeyer’s (198 1 , 1988) discussion. Altemeyer (198 1) reviewed the literature on the authoritarian personality beginning with the pioneer research of Adorno, Frenkel-Bnmswick, Levinson, and Sanford (1950). He conceives of authoritarianism as a personality variable that involves authoritarian submission, authoritarian aggression, and conventionalism. Authoritarians are assumed to be more disposed to submit to established authority, to control the behavior of others through punishment, and to accept and be committed to the traditional social norms of their society when compared with nonauthoritarians.

ATTITUDES TOWARD HIGH ACHIEVERS

45

Altemeyer’s analysis has some interesting implications in regard to tall poppy attitudes. If a society provides authoritative support for high achievement, then authoritarians would be expected to regard those who have achieved conspicuous success and the high status that often comes with it more favorably than nonauthoritarians. If, on the other hand, there is a tendency for accepted authorities to downgrade conspicuous success for whatever reason, then authoritarians may regard those who have achieved notable success and high status less favorably when compared with nonauthoritarians. I indicated previously that although Australians are commonly held to value achievement within a context of individualism, there are also collectivist concerns for equality and group solidarity within the society. There appear to be ambivalent views concerning achievement and the high status and public prominence that may come with it. On the one hand high achievers may be praised and seen as important for the nation’s progress; on the other hand high achievers may be seen as tall poppies who should be lopped and brought back to a common level, depending upon whether they are perceived to be “good” tall poppies or “bad” tall poppies and on other variables that I discussed earlier in this article. If we assume that the Australian culture sanctions rewarding those who have enjoyed conspicuous success, then we would expect to find that authoritarians should be more in favor of rewarding tall poppies when compared with nonauthoritarians. Paradoxically, however, if the culture is also seen to sanction punishing high achievers by undermining their status, then we would expect to find that authoritarians should be more in favor of punishing tall poppies when compared with nonauthoritarians. These two hypotheses are consistent with an analysis that assumes that ambivalent attitudes toward tall poppies within a society may lead to complex and possibly contradictory relations between tall poppy attitudes and authoritarianism.

E. CROSS-CULTURAL DIFFERENCES IN TALL POPPY ATTITUDES

Finally, as I indicated previously, one would expect to find differences in tallpoppy attitudes depending on the cultural context. For example, in Japan there may be a stronger tendency to want to see tall poppies fall than in more individualistic cultures such as Australia because tall poppies violate the normative expectation that achievement should occur with proper regard for interdependence, group harmony, and humility. The cultural prescription for selfeffacement in Japan may also be associated with some resistance to endorsing items from Westem-based self-esteem scales that present self within an independent and individualistic framework. In the next section I will report the results of correlational studies in which some of the hypotheses that I have discussed were tested.

46

N. T. FEATHER

VII. Correlational Studies of General Attitudes toward Tall Poppies A. STUDY 1: TALL POPPY ATTITUDES, VALUES, AND GLOBAL SELF-ESTEEM This study involved 205 male and female subjects who were sampled from the general population of metropolitan Adelaide (Feather, 1989). They completed a questionnaire that included the Tall Poppy Scale, the Schwartz Value Survey (Schwartz, 1992), and the Rosenberg Self-Esteem Scale (Rosenberg, 1965), as modified by Bachman et al. (1978). The Tall Poppy Scale has already been described (Table VIII). The Schwartz Value Survey involves 30 terminal values and 26 instrumental values that subjects rase for importance on a scale from - 1 to 7 in regard to their importance for self. Eleven subscales were constructed that involved different sets of values and that represented different value types. These value types were consistent with Schwartz and Bilsky’s (1987, 1990) discussion of the universal content and structure of human values, and they emerged from smallest space analysis (Guttman, 1968) of the ratings of the 56 values from samples in different cultures (Feather, 1989; Schwartz, 1992).3 The Rosenberg Self-Esteem Scale (Bachman et al., 1978; Rosenberg, 1965) was included to provide a measure of global self-esteem. Subjects rated items such as “I feel that I’ma person of worth, at least on an equal plane with others” and “I am a useful person to have around” on a 5-point scale ranging from “Never true” to “Almost always true.” I also included a measure of political preference that enabled me to compare responses from subjects who favored the Liberal Party in Australia with the responses of subjects who favored the Labor Party. I assumed that subjects who preferred the Liberal Party would be more conservative and respectful of authority and would put more stress on individual enterprise and achievement in contrast to those who preferred the Labor Party. Supporters of the Labor Party were assumed to be more concerned with social welfare and with reducing inequalities between people. This assumption is consistent with the policy platforms of the two parties. Full details about these measures are contained in the original report (Featner, 1989). Table IX presents a sample of the results. As predicted, these results show that tall poppy attitudes were related to values concerned with achievement and power. In the classification that I used, the achievement values comprised being ambitious, influential, capable, intelligent, and successful; the social power val3The value types were an earlier version of those reported by Schwartz (1992).

47

ATTITUDES TOWARD HIGH ACHIEVERS

TABLE IX AND CORRELATIONS FOR VALUE SUBSCALES, MEANSCORES, STANDARD DEVIATIONS, FAVORFALL,FAVORREWARD, GLOBALSELF-ESTEEM, GENDER, A N D LEET-WING POLITICAL PREFERENCEO Correlations

Variable Value subscale Hedonism Achievement Social Power Self-Direction Stimulation Maturity Prosocial Security Restrictive conformity Tradition maintenance Spiritual Tall-poppy variables Favor fall Favor reward Gender Political preference Global self-esteem

Mean

SD

4.36 4.22 2.38 4.55 3.55 4.33 4.83 4.12 4.07 2.52 4.08

I .55 1.10 1.38 1.06 1.51 1.14 .94 1.03 I .25 1.31 I .64

38.02 44.82 1.49 1.54 41.19

11.95 10.52

so .50 5.26

Favor fall

-16% -.19*' .02 - .07 .I2 - .05 - -05 .06 .I0 .19** -.I1

Favor reward

.09 .40*** ,33*** .03 .IS* .04

.oo .II .22**

.os .02

Gender

.03 -.04 -.08 .03 - .04

.IS* .21** .09 .16*

.I0 .21**

-

-.36***

-.36*** -.06 .21* -.31***

-

-.06 .04

.04 -.32*** .06

.04 -.16*

-

Left-wing political preference

Global selfesteem

.04

- .02

-.21* -.32***

- .07

.I5 -.08 .20* .25** - .05 -.15 .01 .22*

.2\ * -.32*** .04

-

-.01

.14*

.II - .03

.oo - .02 -.I0 -.17* -.25*** .07 - .3 1

***

.06 -.16* -.Ol -

aN = 205 except for correlations involving political preference, where N = 122. There were minor variations from these Ns due to missing cases. Means for value subscales have been converted to the - 1 to 7 scale. Political preference was coded as follows: Liberal Party = 1, Labor Party = 2. Gender was coded as follows: Male = 1, Female = 2. Tests of significance are two-tailed. Reprinted from Feather (1989) with permission from the author and the Australian Psychological Society. * p < .05. * * p < .01. * * * p < ,001.

ues comprised social power, wealth, social recognition, authority, and preserving my public image. Subjects who were more in favor of the fall of tall poppies assigned less importance to achievement values when compared with subjects who were less in favor of the fall of tall poppies. Those who were more in favor of rewarding tall poppies assigned more importance to achievement and power values when compared with subjects who were less in favor of rewarding tall poppies. Other value types were also involved in statistically significant relations with the favor-fall and favor-reward variables (see Table IX). For example, those subjects who saw tradition-maintenance values as more important were more in favor of the fall of tall poppies. The tradition maintenance values comprised respect for tradition, being humble, accepting my portion in life, and devout. Favoring the reward of tall poppies was positively related to restrictive conformi-

48

N.

T.FEATHER

ty values. These values comprised politeness, self-discipline, honoring of parents and elders, being obedient, and being clean, all of which involve some form of conformity to authority. The statistically significant correlations involving political preference are also consistent with the assumption that value priorities influenced attitudes toward tall poppies in general. Table IX shows that subjects who expressed a preference for the more to the left Labor Party assigned higher importance to maturity values, prosocial values, and spiritual values when compared with those who preferred the more to the right Liberal Party. Labor Party supporters saw achievement values and social power values as less important for themselves when compared with Liberal Party supporters. Thus, subjects’ political preferences reflected different sets of value priorities. These value differences probably account for the fact that Labor Party supporters were more likely to favor the tall poppy’s fall and were less likely to support rewarding the tall poppy when compared with Liberal Party supporters. Finally, as predicted, support for the fall of tall poppies was linked with lower global self-esteem. However, subjects with higher global self-esteem were not more likely to favor rewarding tall poppies when compared with those lower in self-esteem. Thus, the hypothesis of a positive relation between favor reward and global self-esteem was not supported for this sample.

B. TALL POPPY ATTITUDES, PERCEIVED COMPETENCE, AND GLOBAL SELF-ESTEEM This study involved 105 male and female students in the penultimate year of a South Australian high school (Feather, 1991a). These subjects completed the Tall Poppy Scale (Table VIII) and the modified version of the Rosenberg Self-Esteem Scale (Rosenberg, 1965; Rosenberg et al., 1978). In addition I obtained from the students their own ratings of competence level relative to others, based on their perceived performance in their best five school subjects. I also obtained a rating of their relative competence level from the teacher who administered the questionnaire. The former measure is called student performance and the latter measure is called teacher rating. The main results of this study are presented in Table X. The statistically significant findings indicate that global self-esteem was positively related to favor reward and negatively related to favor fall. The results involving relative academic performance assessed either by students’ own ratings or by teacher’s ratings present a similar picture. Subjects who were low in perceived competence were less likely to agree that tall poppies should be rewarded and more likely to agree that tall poppies should fall from their high positions when compared with subjects who were high in perceived competence. Note that global self-esteem

49

AITITUDES TOWARD HIGH ACHIEVERS

TABLE X MEANS,STANDARD DEVIATIONS, A N D INTERCORRELATIONSFOR MAJOR VARIABLES~ ~~

~~

Correlations Competence measures

Variable Competence measures Global self-esteem Student performance Teacher rating Attitude measures Favor reward Favor fall Mean SD

Selfesteem

-

Student performance

.34***

-

Attitude measures Teacher rating

.32*** .68*** -

Favor reward

Favor fall

.29** .27** .22*

- .20*

-

-.36***

-.32*** - .24*

36.32 5.38

3.56 .65

2.23 .77

46.89 8.17

44.54 10.63

ON = 105. There were minor variations from this N due to missing cases. Tests of significance are two-tailed. Reprinted from Feather (1991a) with permission of the author and the Australian Rychological Society. * p < .05. * * p < .01. * * * p < ,001.

and the two measures of competence were positively related, as one would expect (Feather, 1991b). The self-esteem of the high school students was bound up with their relative standing as far as their school performance was concerned.

C. STUDY 3: TALL POPPY ATTITUDES, PROTESTANT ETHIC VALUES, COMMUNAL ORIENTATION, AND EQUALITARIANISM

In one recent study that involved different samples of students as well as parents and their children, subjects completed measures of Protestant ethic values, communal orientation, and equalitarianism in addition to the Tall Poppy Scale, the Schwartz Value Survey, and the Rosenberg Self-Esteem Scale (Feather, 1993a). The detailed results involving the Schwartz value types will not be presented here. The value types were defined using the most recent classification presented by Schwartz (1992). These value types were power, achievement, hedonism, stimulation, self-direction, universalism, benevolence, tradition, conformity, and security. The specific values making up these value types were very similar to those that defined the corresponding earlier value types that were used as variables in the first study that I described (Table IX). The results of the

50

N . T. FEATHER

present study showed that, although the correlations between favor fall, favor reward, and the value types varied across samples, there were some regularities in regard to the correlations involving the favor-reward variable. Consistent with the findings reported for the first study, favoring the reward of tall poppies was positively related to the importance assigned by subjects to achievement values, power values, and conformity values. Favoring the fall of tall poppies was also positively related to the importance that subjects assigned to conformity values. Subjects in two of the samples completed the Protestant Ethic Scale devised by Mirels and Garrett (1971), the Communal Orientation Scale developed by Clark, Ouellette, Powell, and Milberg (1987), and a measure of equalitarianism designed by the author. The latter two scales are rather more specific in their orientation than is the Protestant Ethic Scale, which was constructed to assess the moral outlook described in Weber’s (1904- 190511976) classic analysis of the spirit of modem capitalism, a moral outlook that stressed the virtues of hard work, self-discipline, the denial of pleasure for its own sake, and individual activism as a person attempted to fulfill his or her duty in a calling or vocation. Sample items are “Any person who is able and willing to work hard has a good chance of succeeding,” “Most people spend too much time in unprofitable amusements,” and “The self-made person is likely to be more ethical than the person born to wealth.” The Communal Orientation Scale (Clark et al., 1987) contains items designed to assess whether the subject behaves in a typically communal way toward other people (e.g., “I believe people should go out of their way to be helpful”), as well as whether the subject expects others to behave in a communal way toward him or her (e.g., “It bothers me when other people neglect my needs”). The Equalitarianism Scale consists of the six items presented in Table XI. All of these items involve positive statements that concern equality of opportunity and equality of condition and state. The results of a principalcomponents factor analysis of the inter-item correlation matrix for each sample provided a one-factor solution in each case.

TABLE XI THE EQUALITARIANISM SCALE ~~

~

~

~

Item I . There is too large a gap between the rich and the poor in our society. 2. Society functions better when everyone has an equal share of the profits. 3. It is wrong for top executives to be paid very high salaries when other workers receive a much lower wage. 4. Everyone should have an equal chance and an equal say in most things. 5. We should try to level out the differences that exist in our society so that people can become more equal. 6. People should be treated equally because we are all human beings.

_

AnITUDES TOWARD HIGH ACHIEVERS

51

Subjects responded to items in both the Protestant Ethic Scale and the Equalitarianism Scale by using the same 6-point scale as was used for the Tall Poppy Scale, with a constant of 4 subsequently added to the item responses. They answered the Communal Orientation Scale by using the 1 to 5 scale described by Clark et al. (1987). Results are presented in Table XI1 for two samples. The first sample consisted of 281 subjects (96 males, 185 females) who were enrolled in the introductory psychology course at Flinders University in 1991. The second sample comprised 339 subjects, consisting of parents and their children from 113 two-parent samples in metropolitan Adelaide. These subjects were also tested in 1991. Because of time constraints it was necessary to use different versions of the questionnaire for Sample 1, each containing a different mixture of scales. Approximately fourfifths of subjects answered the Schwartz Value Scale in Sample 1, and approximately two-fifths of subjects answered the other scales. Subjects in Sample 2 did not answer the Schwartz Value Survey, but they did complete the other measures. The correlations in Table XI1 involving the tall poppy attitudes reveal a consistent pattern of relations between these attitudes and the measures of Protestant ethic and equalitarianism. Statistically significant positive correlations were obtained between favor fall and favor reward and the Protestant ethic variable in both samples. As predicted, the favor-fall variable was positively related to equalitarianism, and this correlation was also statistically significant in both samples. However, there was no consistent evidence for a negative correlation between equalitarianism and favoring the reward of tall poppies, although there was a statistically significant negative correlation between those variables in Sample 2. The communal orientation variable was generally unrelated to taIl poppy attitudes except for the statistically significant positive correlation between this variable and favor reward in Sample 2. Global self-esteem was positively related to favor reward and negatively related to favor fall in Sample 2, and both correlations were statistically significant. However, this pattern of correlations was not obtained for Sample 1. Most of these results are consistent with expectation. One would expect that individuals who support the Protestant work ethic would also favor rewarding tall poppies and that those who hold equalitarian values would favor bringing tall poppies down to a level where status differences are reduced. The positive correlation obtained in both samples between the Protestant ethic variable and favoring the fall of tall poppies is more difficult to explain and may reflect the fact that the Protestant Ethic Scale is multidimensional and not simply a measure of achievement values. For example, the evidence from the current study and from a previous one (Feather, 1984) indicated that scores on the Protestant Ethic Scale were also positively related to the importance assigned by subjects to conformity values (e.g., being obedient). Thus, the positive correlation between the Protestant ethic variable and the favor-fall variable may reflect their joint dependence

MEANSCORES,STANDARD DEVIATIONS, AND’ CORRELATIONS

TABLE XI1 FAVORFALL,FAVORREWARD,VALUES MEASURES, AND GLOBAL SELF-ESTEEW

BETWEEN

Sample 2

Sample 1 Correlations

Variable

Theoretical range

Mean

SD

N

Favor fall Favor reward Global self-esteem Protestant ethic Communal orientation Equalitarianism

10-70 10-70 10-50 19-133 14-70 6-42

35.80 48.10 39.60 83.58 53.34 30 25

10.13 8.98 6.65 1245 6.44 6.78

272 276 111

Correlations

Favor fall

Favor reward

-

-.21*** .04 .26** .05 - .02

108 118

SD

35.93 50.02 40.81 87.46 51.85 29.60

10.89 8.79 5.49 14.51 6.54 7.40

-.21*** 14 .23* .06 .24**

N

Favor fall

Favor reward ____

~

~

~~

~~

110

Mean

336 336 337 332 330 336

-

- .26***

-.26*** -.30*** .14** -.07 .25***

.12* .26*** .12* -.11*

-

UTests of significance are two-tailed. * p < .05. * * p < .01. ***p < ,001 TABLE XI11 STANDARDIZED BETACOEFFICIENTS AND MULTIPLECORRELATIONS FROM MULTIPLEREGRESSION ANALYSES WITH FAVORFALL AND FAVOR REWARDAS DEPENDENT VARIABLES“ Independent variables _

~

Dependent variables

Punitive aggression

Respectful submission

Favor fall Favor reward

.38** -.I5

.I1 .I1

_

_

_

~

~

~

~~

Conventionality

Rule-following

Global self-esteem

Right-wing political preference

.19*

-.31** .21*

-.17* .18*

- .02 .14

.14

Gender

Age

R

RZ

.07

.oo

-.I4 -.03

.52 .41

.27 .I6

aFrom “Authoritarianism and Attitudes toward High Achievers,” by N. T. Feather, 1993b, Journal ofpersonaliry and Social Psychology, 65, 152-164. Copyright 1993 by the American Psychological Association. Reprinted by permission. * p < .05. * * p < .001.

ATTITUDES TOWARD HIGH ACHIEVERS

53

on other variables such as conformity values. In any case, this positive correlation requires replication.

D. STUDY 4: TALL POPPY ATTITUDES AND RIGHT-WING AUTHORITARIANISM Another recent study investigated relations between tall-poppy attitudes, rightwing authoritarianism, and global self-esteem, as well as affective reactions to the rise and fall (actual or hypothetical) of three Australian political leaders (Feather, 1993b). There were 204 subjects (104 males, 100 females) in this study who were recruited from the general population of metropolitan Adelaide in March 1992. These subjects completed a questionnaire that contained the Tall Poppy Scale, the 24-item version of Altemeyer’s (1981, pp. 305-306) RightWing Authoritarianism Scale (RWA), the Rosenberg Self-Esteem Scale (Bachman et al., 1978), and items relating to the three Australian political leaders. They also provided information about their voting preference in regard to the current political parties in Australia. The responses were coded in the direction of right-wing political preference (1 for the Labor Party, 2 for the Liberal Party) after deleting subjects who nominated other political parties. The items in the Right-Wing Authoritarianism Scale cover authoritarian submission, authoritarian aggression, and conventionalism, and they are balanced to control for acquiescence response set. Subjects responded to each item by using the same 6-point scale as they used for the Tall Poppy Scale, with a constant of 4 subsequently added to the item responses. A principal-components factor analysis of the inter-item correlation matrix yielded four interpretable factors that I called punitive aggression, respectful submission, conventionality, and rulefollowing. Four subscales were constructed involving items with high loadings on each respective factor. There is space to present only a selection of the results. As predicted, there was a statistically significant positive correlation between favoring the reward of tall poppies and total scores on the Right-Wing Authoritarianism Scale, r( 189) = .23, p < ,001. The correlation between favoring the fall of tall poppies and total scores on the Right-Wing Authoritarianism Scale was also positive, but the correlation was not statistically significant, r( 188) = .1 1. Multiple regression analyses were conducted with favor fall and favor reward as the respective dependent variables and with scores on the four authoritarian subscales, right-wing political preference, global self-esteem, gender, and age as the independent variables. Table XI11 presents the results of these regression analyses. The statistically significant standardized regression coefficients in Table XI11 show that when the effects of other variables were controlled for, punitive ag-

54

N . T. FEATHER

gression and conventionality were positive predictors of favor fall, and rulefollowing and global self-esteem were negative predictors. When favor reward was the dependent variable, rule-following and global self-esteem were positive predictors. The pattern of coefficients indicates that punitive aggression, rulefollowing, and global self-esteem predicted in opposite directions to favor fall and favor reward respectively when the effects of other variables were controlled for in the regression analysis. Thus, it was the rule-follower with relatively high global self-esteem who was more likely to favor rewarding tall poppies, and it was the more aggressive rebel with relatively low global self-esteem who was more likely to favor the fall of high achievers. These results indicate the importance of unpacking the authoritarianism construct when considering relations with tall poppy attitudes. Correlations that involve the different facets of authoritarianism are more informative than those that involve the total scores.

E. STUDY 5: TALL POPPY ATTITUDES AND GLOBAL SELF-ESTEEM IN AUSTRALIA AND JAPAN

The Tall Poppy Scale and the Rosenberg Self-Esteem Scale (Rosenberg, 1965) were also included in the study with Australian and Japanese students that was described in a previous section (V,D). We hypothesized that the Japanese subjects would be more likely to agree that tall poppies should fall and that the protruding nail should be leveled when compared with the Australian subjects (Feather & McKee, 1993). We based this hypothesis on Markus and Kitayama’s (1991) analysis of cultural differences in the construal of self. They report a lot of evidence that indicates that the Japanese culture involves a concern with interdependence, harmony, ingroup solidarity, and humility (see also Cousins, 1989; Fiske, 1991; Lebra, 1976; Triandis, 1990; Weisz, Rothbaum, & Blackburn, 1984). This concern would be expected to lead to a rejection of those who express independence from others and who stand above the group. As noted previously, we assumed that the Australian culture involves a mixture of collectivism and individualism. However, compared with Japan, the Australian culture would appear to be more independent and less interdependent insofar as the construal of self is concerned (Feather & McKee, 1993). In Japan the nail that sticks up metaphorically implies independent assertion and autonomy and conflicts with norms for modest and self-effacing behavior so that the nail has to be leveled. In Australia the phrase “bringing down the tall poppy” is directed at those who stand out by virtue of their eminent status and, while collectivist concerns may be involved, they are not the only basis for the attitude. Note that the hypothesis that the Japanese subjects would be more likely to

55

ATTITUDES TOWARD HIGH ACHIEVERS

favor the fall of tall poppies is not meant to imply that achievement is not valued in cultures such as Japan that are interdependent and collectivist in their orientation. The meaning of achievement varies across cultures (e.g., Fyans, Salili, Maehr, & Desai, 1983; Maehr & Nicholls, 1980). There is plenty of evidence that achievement is highly valued in Japan (e.g., De Vos, 1973; Hamilton, Blumenfeld, Akoh, & Miura, 1990; Markus & Kitayama, 1991), but the valued achievements tend to be those that are the result of effort directed toward social or group goals rather than the pursuit of personal ambition. We also expected that the Japanese subjects would be less likely to assert the independent construal of self that is implied in Westem-based measures of global self-esteem such as Rosenberg’s (1965) scale. In Japan outward expressions of personal worth would tend to be constrained because they conflict with a collectivist orientation and an interdependent view of self that emphasizes the interpersonal context, ingroup cohesion and harmony, and self-effacement (Markus & Kitayama, 1991). In Australia, however, there should be more readiness to endorse items that assert independence and autonomy because such endorsement is consistent with a more individualistic orientation within the culture and with an independent construal of self. Hence, we expected that the Japanese subjects would report lower levels of global self-esteem on the Rosenberg Self-Esteem Scale (Rosenberg, 1965) when compared with the Australian subjects. The main results are presented in Table XIV. The Japanese subjects had a significantly higher mean score than the Australian subjects for the favor-fall variable (means of 45.37 and 34.17 respectively, t(233) = 10.60, p < .001), but TABLE XIV BETWEEN GLOBAL SELF-ESTEEM A N D TALLPOPPYATTITUDES MEANSCORESA N D CORRELATIONS FOR JAPANESE A N D AUSTRALIAN SAMPLESO Correlations Japanese sample Variable

Mean

SD

Global self-esteem

Global self-esteem Favor reward Favor fall

31.19 48.41 45.37

5.88 8.50 6.73

.09 -.24** 39.32 6.36

47.58 8.55

-

Favor reward

Favor fall

.11 -

-.I5 -.I2

-.33***

~~

Australian sample

Mean

SD

34.17 9.09

aNs varied from 108 to 109 for the Japanese sample and from 122 to 126 for the Australian sample due to missing cases. Correlations for the Japanese sample are above the diagonal; correlations for the Australian sample are below the diagonal. Tests of significance are two-tailed. Reprinted from Feather and McKee (1993) with permission of the authors and the American Sociological Association. * * p < .01. * * * p < ,001.

56

N. T. FEATHER

the difference in the mean favor-reward scores was not statistically significant (means of 48.41 and 47.58 respectively; t(230) = .75, ns). However, the mean global self-esteem score was significantly lower for the Japanese sample than for the Australian sample (means of 31.19 and 39.32 respectively; t(233) = - 10.1 1, p < .001). The differences that were obtained for the global self-esteem and favor-fall variables were consistent with our predictions. Table XIV also shows that statistically significant negative correlations were obtained between global self-esteem and the favor-fall variable, and between the favor-fall variable and the favor-reward variable, replicating the results that have been described in previous sections (Feather, 1989, 1991b). However, none of the correlations were statistically significant for the Japanese sample, although the correlations were in the same direction as the respective correlations for the Australian sample. The negative correlation between global self-esteem and the favor-fall variable became stronger and was statistically significant after trimming items from the favor-fall scale so as to improve its internal reliability (see Feather & McKee, 1993, for further discussion of these results).

F. CONCLUDING COMMENTS The studies reviewed in this part of the chapter were concerned with attitudes toward tall poppies in general, and they were correlational in nature. The results of these studies showed that these attitudes were related to value prioritiesespecially to the importance that subjects assigned to values concerned with achievement, power, and equality-to global self-esteem and perceived competence, and to different facets of right-wing authoritarianism. These results can be interpreted in terms of a range of theories that deal with value-attitude relations, with social comparison and related ideas concerned with social identity and self-categorization, with similarity and attraction, with the nature of right-wing authoritarianism, and with cultural differences.

VIII. Studies of Public figures A third group of studies in my research program has examined attitudes toward tall poppies who are highly visible public figures. These studies resemble the experimental studies that were described previously, except that achievement status was not varied and real persons rather than hypothetical persons were the subject of interest. These studies drew upon the theoretical approaches that were described earlier in this article, especially attribution theory and the conceptual analysis of deservingness.

AlTITUDES TOWARD HIGH ACHIEVERS

57

A. STUDY 1: ATTITUDES TOWARD HIGH ACHIEVERS IN PUBLIC LIFE The hypotheses that were tested in this first study (Feather, Volkmer, & McKee, 1991) were based upon a theoretical analysis of relations between causal attributions and different kinds of affect, and relations between the subjective values of actions and outcomes and beliefs about deservingness, as discussed in a previous section of this article. Thus, we hypothesized that subjects would report feeling more pleased about a high achiever’s success and would see the high achiever as deserving the success more when the success could be attributed to internal causes that involved controllable effort and ability as distinct from external causes such as outside assistance and luck. In the former case the tall poppy would be perceived as responsible for the success, especially if the success was seen to follow a lot of preparation and hard work, that is, behavior that is under personal control. The success would also be seen as more deserved because the positively valued outcome (success) followed behavior that was positively valued within our culture, namely behavior that reflected effort and skill (Feather, 1992a). When a person’s success can be attributed to external causes, however, the person does not gain credit for the success and the responsibility for the success is seen to lie elsewhere. When this external attribution occurs, subjects would report feeling less pleased about a high achiever’s success and they would perceive the success to be less deserved. A similar analysis can be applied to a high achiever’s fall from a high position. In this case we hypothesized that subjects would report feeling less pleased, more sorry, and more disturbed about a high achiever’s fall the more the tall poppy was seen to be responsible for and deserving of the high position that was held prior to the fall. Subjects were also expected to report feeling less pleased, more sorry, and more disturbed when the tall poppy was seen to be not responsible for the fall than when responsibility for the fall could be attributed to the tall poppy. Thus, we expected that those high achievers who were seen to have gained their high position because of hard work and ability would be viewed more sympathetically after a fall when compared with those high achievers whose high position was seen to be due more to external and less to internal causes. Both kinds of high achievers, however, should receive more expressions of sympathy when they were seen as not responsible for the fall than when the fall was seen to be their fault. These hypotheses were concerned with variables that related to how the high achiever or tall poppy attained his or her high position and how the high achiever came to fall. We also expected that reactions to the high achiever would depend upon how the high achiever was seen to use the high position and what kind of personality characteristics he or she was perceived to possess. The results of two of my studies described in previous sections (VA and VB) showed that the extent

58

N. T. FEATHER

to which subjects reported feeling pleased about the fall of a hypothetical person was negatively related to a measure of initial attraction toward the stimulus person (Feather, 1989). In the current study we hypothesized that subjects who evaluated a public figure more positively in terms of his or her personality characteristics would be more likely to report feeling pleased about that person’s success, more likely to report that the success was deserved, and more likely to report feeling sorry and disturbed when that person fell, when compared with subjects who viewed a tall poppy less favorably or even negatively. A secondary aim of our study was to investigate relations between the specific measures of deservingness and affect that we included in the investigation and general attitudes toward tall poppies as assessed by the Tall Poppy Scale (Feather et al., 1991). We also included the modified Rosenberg Self-Esteem Scale (Bachman et al., 1978; Rosenberg, 1965) in the study and a measure of Labor versus Liberal Party preference in order to test whether the relations between these measures and the favor-fall and favor-reward variables from the Tall Poppy Scale that were described previously would be replicated. The subjects in this study were 377 male and female students who were enrolled in an introductory course in psychology at Flinders University in March 1990. Subjects completed a questionnaire that presented three well-known Australian tall poppies who came from either politics, sport, or entertainment. There were three versions of the questionnaire corresponding to each of these three domains. The politicians were Robert Hawke (Labor Party), Andrew Peacock (Liberal Party), and Janine Haines (Democrat Party), all of whom were prominent in Australian politics at the time. The high-profile Australian sports people were Pat Cash (tennis), Alan Border (cricket), and Lisa Curry-Kenny (swimming). The well-known Australian entertainers were Kylie Minogue (actress and singer), Paul Hogan (actor), and John Farnham (singer). Subjects first rated each of three tall poppies in the version of the questionnaire that they received in regard to the causes of that person’s high achievement. The causal attribution items referred to ability or talent, hard work or effort, outside or external assistance (e.g., publicity through the mass media, help from others), and opportunity or luck. They then rated the degree to which they thought each tall poppy deserved to be in his or her high position and whether the tall poppy deserved to maintain his or her high position in the future. They also rated how pleased they were that the tall poppy held his or her current high position and how pleased they would be if the tall poppy were to rise even further or to fall from his or her current high position. The scales developed to measure personality characteristics were based on a factor analysis of the ratings of the tall poppies along 16 semantic differential, bipolar adjective scales (see Feather et al., 1991, for details). Two scales were derived from this analysis: self-centered personality and good mixer. A final set of items asked subjects to suppose that for some reason the tall

59

ATTITUDES TOWARD HIGH ACHIEVERS

poppy lost his or her current high position. They then rated 10 adjectives (e.g., pleased, sorry, angry) to indicate how they would feel about the hypothetical fall of each tall poppy. These ratings were made under two conditions, a responsible condition where the tall poppy was described as responsible for his or her fall, and a not responsible condition where the tall poppy was described as not responsible for his or her fall. A factor analysis of the ratings led to the development of three affect scales that were designated sorry, pleased, and disturbed. Finally, subjects completed the Tall Poppy Scale (Feather, 1989), the modified Rosenberg Self-Esteem Scale (Bachman et al., 1978; Rosenberg, 1965), and items about age, gender, political preference, and family background. Table XV presents a selection of the results, pooling across the three domains. Table XV shows that, as predicted, tall poppies who were perceived to have achieved their success by ability and effort were also seen to deserve their success more. Subjects also reported feeling more pleased about the high achievement and the possibility of even greater success, and less pleased should a fall occur, when the achievement of tall poppies could be attributed more to ability and effort and when it was seen to be deserved. These relations were in the reverse direction when the achievement was attributed more to external assistance. The achievement was then seen to be less deserved, and subjects reported feeling less pleased about the success but more pleased about a hypothetical fall. TABLE XV CORRELATIONS BETWEEN DESERVE A N D PLEASED MEASURES A N D CAUSAL ATTRIBUTIONS AND PERSONALITY CHARACTERISTICS“ ~~

Deserve Variable Causal attributions Ability Effort Assistance Luck Deserve Present position Maintain position Personality Self-centered Good mixer

Pleased

Present position

Maintain position

Present position

Rise further

.61*

.49* .52* -.23*

.01

.54* .53* - .23* - .03

.49* .50* - .22* .03

.64* -.21*

-

.83*

.83*

-

-.56*

-.55*

.55*

.51*

-.01

.69* .67*

.69* .70*

-.59* .57*

-.58* .56*

Fall - .40* - .42*

.30* .I0 -.56* -.61*

.53* - .50*

“Ns ranged from 345 to 367 due to some missing cases. Tests of significance are two-tailed. Reprinted from Feather, Volkrner, and McKee (1991) with permission of the authors and the Australian Psychological Society. * p < ,001.

60

N. T.FEATHER

Table XV also indicates that judgments of deservingness and ratings of positive affect were related to the personality characteristics of the tall poppies. The more the tall poppies were seen to be self-centered, the less they were seen to deserve their success, the less pleased subjects reported feeling about the tall poppies’ present high position and a hypothetical future rise, and the more pleased subjects reported feeling about a hypothetical future fall. These relations were in the reverse direction for tall poppies who scored higher on the goodmixer variable. These tall poppies were seen to deserve their high position more, and subjects reported feeling more pleased about their success and less pleased about a hypothetical fall. These results are consistent with predictions. The results for the three derived affect variables (pleased, sorry, angry) were consistent with those that were just reported (see Feather et al., 1991, for details). In both the responsible and not-responsible conditions subjects who reported feeling more sorry, less pleased, and in most cases more disturbed about the hypothetical fall of each tall poppy were more likely to attribute the tall poppy’s current position to ability and effort, less likely to attribute the high position to external assistance, more likely to report that the tall poppy deserved his or her present high position, less likely to see the tall poppy as self-centered, and more likely to see the tall poppy as a good mixer. An analysis of differences in the mean scores for each of the three affect variables showed that subjects reported feeling more sorry, less pleased, and more disturbed about the hypothetical fall of a tall poppy when the high achiever was described as not responsible for the fall when compared with the responsible condition. These results are also consistent with predictions. The results also showed that reactions to the public figures also depended on the domain in which they were eminent. Subjects were generally more favorable in various ways to tall poppies in the sports domain when compared with tall poppies from the politics and entertainment domains (Feather et al., 1991). This finding needs to be checked in future studies with larger samples of public figures taken from a variety of domains that vary in their relevance for the subjects who are involved in the research. Finally, the results in Table XVI provide information about the relations between the favor-fall and favor-reward subscales of the Tall Poppy Scale (Feather, 1989) and the other variables that were included in the study. Table XVI shows that, in general, scores from the subscales of the Tall Poppy Scale correlated in the predicted direction with variables concerned with how subjects reacted to actual tall poppies, thereby providing validation evidence. For example, the statistically significant correlations indicate that subjects who were more in favor of the fall of tall poppies tended to report feeling more pleased if a tall poppy were to fall, and they were less likely to report that the tall poppy deserved the high position, less likely to attribute the tall poppy’s success to internal causes, less likely to report feeling pleased about the tall poppy’s high position, and less

61

ATTITUDES TOWARD HIGH ACHIEVERS TABLE XVI AND CORRELATIONS MEANSCORES, STANDARD DEVIATIONS, BETWEEN FAVOR FALL,FAVORREWARD,A N D OTHERVARIABLESO Correlations

Variable Causal attributions Ability Effort Assistance Luck Deserve Present position Maintain position Pleased Present position Rise further Fall Personality Self-centred Good mixer Feelings about fall Responsible sorry

Pleased Disturbed Not responsible sorry

PIea sed Disturbed Tall Poppy Scale Favor fall Favor reward Global self-esteem Political preference

Favor fall

Favor reward

Mean

SD

4.99 5.44 5.16 4.59

1.08 1.01 1.12

- .09

1.11

- .08

5.02 4.63

1.08 1.09

-.17***

4.49 4.35 3.53

1.01 1.14 .92

-.11* .15**

.29*** .28*** -.24***

31.97 22.51

6.10 3.25

.I0 -.21***

-.27*** .28***

3.52 2.36 2.16

2.26 2.29 2.93

-.12* .23*** .05

.16** - .07 .11*

5.57 1.42 3.85

2.72 I .89 3.51

- .08 .19*** .09

- .08

36.82 48.12 40.26 1.60

9.78 8.73 5.54 .49

-.33*** -.25*** .06

-.13* .05

-.16**

-.16**

.21*** .17** - .08 .01 .22*** .30***

.19*** .05 -.33*** -

.13* -.19**

aNs ranged from 333 to 365 due to missing cases, except for political preference, where Ns ranged from 201 to 203. Political preference was coded as follows: Liberal Party = 1, Labor Party = 2. Tests of significance are twotailed. Reprinted from Feather, Volkmer, and McKee (1991) with permissicn of the authors and the Australian Psychological Society. * p < .05. * * p < .01. * * * p < ,001.

62

N. T. FEATHER

likely to report feeling sorry when the tall poppy fell. These relations were in the opposite direction for the favor-reward variable .4 Table XVI also indicates that support for the fall of tall poppies was negatively related to global self-esteem, as found in most of the previous studies that I have described. Favoring the reward of tall poppies was positively correlated with global self-esteem, also replicating some of the previous results. In addition, subjects who reported a voting preference for the more to the left Labor Party were less likely to favor rewarding tall poppies when compared with the more to the right Liberal Party supporters, again replicating previous results. However, we did not obtain a statistically significant correlation between support for the Labor Party and the favor-fall variable, in contrast to the results of one of the studies described previously (Feather, 1989). The results of this study with conspicuous public figures supplement the findings from the experimental and correlational studies that I reported in previous sections. They are consistent with the theoretical approaches that I described previously in which affective reactions and judgments of deservingness were related to causal attributions and the subjective values assigned to actions and their outcomes, and with the results of the experimental study of deservingness that was based on this analysis (Feather, 1992a). The results also indicate that the personality characteristics of the high achievers affected how subjects reacted to them. Subjects not only gave credit to tall poppies whose success could be attributed to ability and effort, they were also more positive toward tall poppies who occupied their high positions with integrity and without displays of egotism and self-interest. Thus, again the results indicated that tall poppies varied in the extent to which they were vulnerable to being lopped or cut down to size. In the present context, the quiet achievers who were seen to have earned their success in positively valued ways were more immune from attack when compared with those tall poppies who were selfcentered, quick-tempered, and uncaring in their attitudes, and whose integrity and care for others was suspect. Tall poppies in the sports domain also had an advantage when compared with tall poppies in the other two domains. B. OTHER STUDIES: REACTIONS TO FALLEN PUBLIC FIGURES

I have also investigated reactions to tall poppies in public life who suffered a dramatic fall from their high positions. These tall poppies were Ben Johnson, the 4The correlations in Table XVI were based on data pooled across the three domains (i.e., the average ratings for the three tall poppies). Relations involving favor fall and favor reward with reported affect would depend on context and domain and would be influenced by other variables. For example, subjects with a right-wing political preference may be more in favor of rewarding tall poppies than subjects with a left-wing political preference, but they may report feeling less pleased about a left-wing political leader holding office. In this case the relation between favor reward and reported pleasure would be influenced by political values.

63

ATTITUDES TOWARD HIGH ACHIEVERS

sprinter who lost his gold medal for the 100 meters in the Seoul Olympics following a test that indicated that he had taken anabolic steroids, and Robert Hawke, the Australian Labor Party leader who was displaced as Prime Minister of Australia and Labor Party leader by Paul Keating in December 1991. I was interested in how subjects would react to each of these tall poppies after their respective falls. In both cases we were not examining hypothetical falls, as in the previous study, but real and important falls that actually occurred. The study that involved Hawke also included Paul Keating, the Labor Party politician who replaced him, and John Hewson, the leader of the opposition Liberal Party (Feather, 1993b, 1993~).The questions used were very similar to those that were employed in the study described in the previous section (Feather et al,, 1991) except that, in the case of Hawke, the questions referred to a fallen tall poppy. Furthermore, in this study the analysis of subjects’ responses to the bipolar adjective scales relating to each of the three political leaders led to the development of two scales, one that was called integrity, the other that was called arrogance. In addition, in this study subjects also rated each leader’s responsibility for the hypothetical or actual loss of his high position. The results of this study were generally consistent with those of the previous one (Feather et al., 1991), and they can be interpreted in the same way. Table XVIl presents a sample of these results for Hawke, the fallen Prime Minister. These results show that subjects who judged Hawke as more responsible for his fall and who reported feeling less sorry, more pleased, and less disturbed about TABLE XVII CORRELATIONS BETWEEN RESPONSIBILITY, FEELINGS ABOUT ACTUAL FALL,CAUSAL ATTRIBUTIONS, DESERVE, A N D PERSONALITY CHARACTERISTICS FOR HAWKEO Hawke Feelings about fall Variable Causal attributions Lack of ability Lack of effort External forces Bad luck Deserve Present position Maintain position Personality Integrity Arrogance

Responsible .30*** .33*** -.16*

-.26*** .57*** .36***

-.34*** .40***

sow

Pleased

Disturbed

-.33*** -.23*** .28*** .19**

.la** .22** -.22** -.I1

-.22** - .07 .21** .I2

- a*** -.36*** .47*** - .48***

.52*** .29***

-.43*** .39***

-.39*** -.20** .15* -.I3

aNs ranged from 194 to 202 because of some missing cases. Tests of significance are two-tailed. * p < .05. **p < .01. * * * p < ,001.

64

N. T. FEATHER

his fall, were more likely to attribute Hawke’s fall to lack of ability and lack of effort, less likely to attribute his fall to external forces and bad luck, more likely to report that he deserved to lose the Prime Ministership and to stay where he is, less likely to see Hawke as possessing integrity, and more likely to see him as being arrogant. Other results showed that subjects who perceived Hawke’s fall as due more to lack of ability and lack of effort were also more likely to see Hawke’s fall as deserved. These relations were in the reverse direction when the fall of Hawke was attributed more to external forces or bad luck. Finally, there was also strong evidence that subjects’ responses for all three political leaders were influenced by their own voting preference (Labor versus Liberal). Indeed, subjects’ voting preference moderated relations between right-wing authoritarianism and their affective reactions to each political leader (Feather, 1993b). The questions about Ben Johnson, the Olympic sprinter, were included as part of the study that investigated relations between tall poppy attitudes, perceived competence, and global self-esteem (Feather, 199la). This study was described in Section VII, B. I included a set of seven questions to gauge how subjects reacted to Ben Johnson’s fall from grace. The results of a factor analysis showed that these items could be combined to form a scale that measured the extent to which subjects supported Johnson after the fall (e.g., by feeling less pleased about his fall, by feeling sorry for him, by believing that he should be allowed to compete in the next Olympics). The results showed that subjects were generally negative toward Ben Johnson (the mean for the support-Johnson variable was below the midpoint of the scale), presumably because his failure was a result of cheating-a negatively valued and controllable behavior. The results also showed that subjects with lower global self-esteem were more supportive of Ben Johnson after his loss of the gold medal. The negative correlation between global self-esteem and the support variable (r = - . 2 5 ) was statistically significant ( p < .05). A similar result was obtained in the Hawke study. Subjects with lower global self-esteem reported feeling less pleased and more sorry about Hawke’s fall (i.e., more supportive of Hawke) when compared with subjects whose self-esteem was higher. The positive correlation between global self-esteem and feeling pleased was statistically significant r( 196) = .16, p < .05. The negative correlation between global selfesteem and feeling sorry was also statistically significant using a one-tailed test, r(197) = -.12, p < .05. These latter results again suggest that the generalized belief that a person holds about his or her self-worth colors attitudes toward successful or unsuccessful others. The winner is more likely to identify with and to favor other winners and the loser tends to identify with and support other losers. Thus subjects’ reactions partly relate to where the other who is judged stands in relation to one’s own position in regard to a general dimension of competence and self-worth.

ATTITUDES TOWARD HIGH ACHIEVERS

65

C. CONCLUDING COMMENTS These studies of public figures who hold or have held positions of high status and achievement extend the investigation of attitudes toward high achievers to tall poppies in realistic settings. Thus, they provide evidence relating to the ecological validity of the results described in previous sections. The results of these studies of public figures indicate that subjects’ attitudes toward tall poppies and their actual or hypothetical fall can again be understood in terms of conceptual analyses that involve a consideration of causal attributions, values, beliefs about deservingness, personality characteristics, and social comparison processes that relate to a person’s own relative standing.

IX. Summary and Discussion This research program on tall poppies has covered a wide territory, both in relation to the nature of the studies that have been conducted and to the theories to which I have referred. What have we learned about attitudes toward tall poppies and reactions to their fall? Here are some conclusions that are based upon the results that were reported. 1. When status was made salient and when other information was reduced, there was some tendency for subjects to report feeling less unhappy about the fall of a high achiever when compared with the fall of an average achiever. However, the effects of status were diminished when more information was provided about the person and the conditions that led to success or failure, enabling judgments to be made about such variables as perceived causality, personality attributes, likability, responsibility, and deservingness. 2 . Reactions to the fall of tall poppies depended upon the extent of the fall. In a classroom situation, a fall that brought the high achiever to the average position in the group was viewed less negatively than a fall that took the average achiever to a lower position. 3. It was necessary to distinguish between “good” tall poppies and “bad” tall poppies in a particular domain in terms of the degree to which they were seen to deserve or not deserve their high status, how attractive or unattractive they were seen to be, their positive or negative personality characteristics, and whether they used their high position in desirable or undesirable ways. The evidence showed that subjects were less pleased and more sorry when “good” tall poppies fell from their high position than when “bad” tall poppies fell. These differential reactions would not be restricted to tall poppies but would apply to “good” versus “bad” people in general.

66

N. T. FEATHER

4. Reported feelings of envy were related to the level of another person’s achievement and to the relevance of the high achievement for the individual. Students reported relatively more envy for a high versus an average achiever, and the difference in envy was more pronounced when the achiever was a student rather than a golfer. 5 . Attitudes toward tall poppies varied depending upon the domain in which the achievement occurred (e.g., politics, sport, entertainment), and they also varied across groups and cultures. Tall poppies in sport were viewed more favorably than tall poppies in politics and entertainment. Japanese subjects were more in favor of the fall of tall poppies when compared with Australian subjects. 6 . A person’s attitudes toward tall poppies were related to that person’s values, especially values that concerned achievement, power, equality, and conformity. There was some evidence that subjects who placed high value on achievement and power were more in favor of rewarding tall poppies and that those who placed high value on equality were more in favor of seeing tall poppies fall. 7. A person’s attitudes toward tall poppies were related to that person’s level of global self-esteem and level of competence. The evidence indicated that those subjects with low global self-esteem and low perceived personal competence were more likely to favor the fall of tall poppies; those with high global selfesteem and high perceived personal competence were more likely to favor their reward. 8. Support and sympathy for a fallen tall poppy tended to be stronger among those subjects who were lower in global self-esteem. 9. Right-wing authoritarians tended to favor rewarding tall poppies. They also favored the fall of tall poppies but this relation was attributable in part to the higher standing of authoritarians on punitive aggression. These conclusions are relative to the samples and procedures used in the studies that I have described. It should also be noted that in some instances the effect sizes were small. There is plenty of scope for further research in this area. Future research could add to our knowledge about the effects of the domain in which the tall poppy is eminent, the effects of the relevance of the tall poppy’s achievement to those who judge the tall poppy, the effects of personality needs and general attitudes (e.g., authoritarianism, conservatism) on subjects’ reactions to tall poppies, similarities and differences in tall poppy attitudes both across cultures and across groups within the same culture, possible differences in reactions depending on whether the tall poppy’s high status is achieved or ascribed, the effects of the nature and magnitude of a tall poppy’s fall, and the effects of whether or not the fall is seen as temporary or permanent, involving the possibility of redemption or no redemption at all. It may be the case that the perceived personality characteristics of tall poppies

ATTITUDES TOWARD HIGH ACHIEVERS

67

exert their influence via the deservingness variable (Feather, 1992a). My analysis of deservingness focused on responsibility attribution and the relation of instrumental behaviors and outcomes to needs and values. However, a more complete analysis might also include the valued attributes of persons. These attributes might derive their influence from their association with positively or negatively valued behaviors. For example, a politician would be seen to have high integrity if he or she is perceived to behave honestly and fairly and with a respect for human rights; a politician would be called arrogant if he or she is seen to behave arrogantly or rudely. Thus, the perceived personality characteristics of the politician are linked to positively or negatively valued behaviors. Following the balance model of deservingness (Fig. l ) a politician of high integrity who fails to be reelected may then be seen not to deserve that negative outcome, whereas an arrogant politician who is not reelected may be seen to have received his or her just deserts. As noted previously, it would also be possible to include positive and negative liking relations or attitudinal relations between person (p) and other (0) in a set of structures that extend those presented in Fig. 1 by includingp and o as separate entities (e.g., as in Fig. 2). We also need further investigation of the conditions under which Schadenfreude or malicious joy might occur when a tall poppy falls. In general, I would expect that the malicious joy that may be experienced by person (p) when other (0) suffers a negative outcome (Heider, 1958) would depend upon the nature of the relation betweenp and o and on the importance or seriousness of the negative outcome and its relevance to self. The results of my studies that used hypothetical stimulus persons showed that in general subjects were not pleased when a high status person fell from a high position; that is, there was no evidence of malicious joy (e.g., Feather, 1989). However, it is likely that subjects would report feeling some degree of pleasure when an envied rival or disliked tall poppy falls from a high position, as when a high-status competitor for office is not appointed or when an unpopular political leader fails to be re-elected. They might also be entertained and find it mildly amusing when a friend or stranger makes a foolish mistake, providing it is not too important or damaging to the person (e.g., Helmreich et al., 1970). Note, however, that it may be difficult to gauge the extent of envy and malicious joy in any particular case. These emotions are often censored or concealed because expressing them may lead to disapproval from others. They may be expressed in situations where there is a social consensus that they are appropriate reactions. A strong test of the hypothesis that status influences reported feelings of malicious joy would involve the experimental manipulation of status and other variables so as to gauge the extent of their effects both singly and in combination. The experimental studies described earlier in this chapter were designed with this purpose in mind, but more research is needed. It may still be the case that some individuals experience a degree of Schadenfreude even when a “good” tall poppy

68

N. T. FEATHER

falls, either because their own status is low and they are envious of the tall poppy or because they are generally reactive against authority and have strong equalitarian values. Some people may also hold the view that tall poppies should not be exempt from the negative events that other “lesser” mortals experience in their lives and that tall poppies may even become better people and in a sense more human from learning how to cope with adversity. There may also be a reaction against their distinctiveness coupled with a wish to bring them back into the human collectivity (cf. Brewer, 1991). Future research might also explore the effects of shifts in status that involve success rather than failure. Again it would be possible to use some of the theoretical approaches described in this chapter to generate predictions about the effects of these shifts. For example, social identity theory would imply that high achievers may feel more supportive toward a low achiever who experiences a conspicuous success when compared with how low achievers react to this shift. Those in the latter category may feel somewhat ambivalent about someone who has left their ranks to join the class of high achievers. Such mixed attitudes are not uncommon in families where upwardly mobile children move to a higher status when compared with their parents. Again, however, I would expect that reactions to the rise would depend upon value priorities, causal attributions for the rise, the magnitude of the rise, justice considerations that relate to deservingness, the personality characteristics of the person who succeeds, the relevance of the achievement, the nature of the culture, and the other variables that I considered in relation to the fall of a high achiever. These topics are important ones for social psychology. They bring us face to face with questions that concern status, values, and justice. They are relevant to many areas of our everyday life, including how we react to the success and failure of our leaders. Studies of status, values, and justice should move more to the center in social psychological research. We have tended to neglect these areas of inquiry and to overlook the fact that social relations occur not only within a context of close relationships but also within a social field that is formally structured along vertical lines. As I noted in a previous section (111, E), even the recent theory and research on social comparison (e.g., Major et al., 1991; Tesser, 1988, 1991) is mainly restricted to situations in which the comparison other is close to the person who makes the comparison. However, research on the status variable should extend far beyond these limiting conditions to include situations in which person and other are not necessarily close to each other in the sense of forming a unit (Heider, 1958). We also need to move beyond explanations that relate comparison processes to self-evaluation, self-improvement, and self-enhancement (e.g., Suls & Wills, 1991). We need to recognize that a person’s reactions to the status of others and to changes in their status are bound up with that person’s construal of how that status came about, with value priorities, with considerations of deservingness and social justice, with the person’s own relative position, and with the other

ATTITUDES TOWARD HIGH ACHIEVERS

69

variables that I have discussed. I have found Heider’s (1958) discussion of reactions to another’s lot to be a useful source of ideas for my research, but, as I indicated, other theoretical approaches are also relevant. The research that I have described extends to the investigation of emotional reactions beyond those commonly dealt with by attribution theory (e.g., Weiner, 1986) and considers some of the emotions that relate to comparison with others. Thus, I have included envy as a variable in some of the studies (e.g., Feather, 1992a; Feather & McKee, 1992). Ortony, Clore, and Collins (1988) classify envy and jealousy as resentment emotions. They also consider other emotional states that accompany social comparison, namely appreciation emotions (e.g., admiration, awe, respect) and reproach emotions (e.g., contempt, disdain, indignation). These emotions also deserve investigation in future research where status is an important variable. It is also important to distinguish between emotions that relate to wanting to see a tall poppy fall and striving to bring a tall poppy down (e.g., envy) and those emotions that may occur once a tall poppy has fallen (e.g., malicious joy, unhappiness, compassion). Finally, the present research program developed from my curiosity about the validity of a commonly held belief that Australians like to see tall poppies fall. As I noted previously, the results of my studies indicated that this is an oversimplified belief. The real situation is much more complex, and one has to pay close attention to the context of achievement and must distinguish between good tall poppies and bad tall poppies. I believe that it may be heuristic for social psychologists in other cultures to examine beliefs about attitudes and behaviors that are held to be distinctive of their own culture and to generate research programs that test the validity of these beliefs. Such research has the power to separate myth from reality. It may also extend social psychology beyond Westem-based models and, in the search for general principles, may focus attention on areas that have been somewhat neglected.

Acknowledgments The research reported in this article was funded by the Australian Research Council.

References Adomo, T. W., Frenkel-Brunswick, E., Levinson, D. J., & Sanford, R. N. (1950). The authoritarian personality. New York: Harper. Altemeyer, B. (1981). Right-wing authoritarianism. Winnipeg: University of Manitoba Press. Altemeyer, B. ( 1988). Enemies of freedom: Understanding right-wing authoritarianism. San Francisco: Jossey-Bass. Atkinson, J. W. (1983). Personulity, motivation, and action. New York: Praeger.

70

N. T. FEATHER

Atkinson, J. W., & Feather, N. T. (Eds.). (1966). A theory of achievement motivation. New York: Wiley. Bachman, J. G., O’Malley, P. M., &Johnston, J. (1978). Adolescence to adulthood. Ann Arbor, MI: Institute for Social Research. Brewer, M. (1991). The social self: On being the same and different at the same time. Personality and Social Psychology Butletin. 17, 475-482. Bryson, J. B. (1977, September). Situational determinants of the expression of jealousy. Paper presented at the annual meeting of the American Psychological Association, San Francisco, CA. Byrne, D. (1971). The attraction paradigm. New York: Academic Press. Christopher, R. C. (1983). The Japanese mind: The Goliath explained. New York: Linden Press. Cialdini, R. B., Borden, R. J., Thorne, A,, Walker, M. R., Freeman, S., & Sloan, L. R. (1976). Basking in reflected glory: Three (football) field studies. Journal of Personaliry and Social Psychology, 34, 366-375, Clark, M. S., Ouellette, R., Powell, M. C., & Milberg, S. (1987). Recipient’s mood, relationship type, and helping. Journal of Personality and Social Psychology, 53, 94- 103. Cousins, S. D. (1989). Culture and self-perception in Japan and the United States. Journal of Personality and Social Psychology, 56, 124- 13I . Covington, M. V., & Omelich, C. L. (1979). It’s best to be able and virtuous too: Student and teacher evaluative responses to successful effort. Journal of Educational Psychology, 71, 688-700. Crosby, F., & Gonzales-Intal, A. M. (1984). Relative deprivation and equity theories: Felt injustice and the undeserved benefits of others. In R. Folger (Ed.), The sense of injustice: Socialpsychological perspectives (pp. 141- 166). New York: Plenum. Deaux, K. (1972). To err is humanizing: But sex makes a difference. Representative Research in Social Psychology, 3, 20-28. Deutsch, M. (1985). Distributive justice: A social-psychological perspective. New Haven, CT: Yale University Press. De Vos, G. A. (1973). Socialization for achievement: Essays on the cultural psychology of the Japanese. Berkeley, CA: University of California Press. Encel, S . (1970). Equality and authority. London: Tavistock. Feather, N. T. (1964). A structural balance model of communication effects. Psychological Review, 71, 291-313. Feather, N. T. (1967). A structural balance approach to the analysis of communication effects. Advances in Experimental Social Psychology, 3, 99- 164. Feather, N. T. (1975). Values in education and society. New York: Free Press. Feather, N. T. (1980). Values in adolescence. In J. Adelson (Ed.), Handbook of adolescentpsycholog y (pp. 247-294). New York: Wiley. Feather, N. T. (1984). Protestant ethic, conservatism, and values. Journal of Personality and Social Psychology, 46, 1132-1141. Feather, N. T. (1986a). Cross-Cultural studies with the Rokeach Value Survey: The Flinders Program of Research on values. Australian Journal of Psychology, 38, 269-283. Feather, N. T. (1986b). Value systems across cultures: Australia and China. International Journal of Psychology, 21. 697-715. Feather, N. T. (1989). Attitudes towards the high achiever: The fall of the tall poppy. Australian Journal of Psychology, 41, 239-267. Feather, N. T. (1990a). Bridging the gap between values and actions. In E. T. Higgins & R. M. Sorrentino (Eds.), Handbook of motivation and cognition: Foundations of social behavior (Vol. 2, pp. 151-192). New York: Guilford Press. Feather, N. T. (1990b). Reactions to equal reward allocations: Effects of situation, gender and values. British Journal of Social Psychology, 29, 315-329. Feather, N. T. (1991a). Attitudes towards the high achiever: Effects of perceiver’s own level of competence. Australian Journal of Psychology, 43, 121-124.

ATTITUDES TOWARD HIGH ACHIEVERS

71

Feather, N. T. (1991b). Human values, global self-esteem, and belief in a just world. Journal of Per.Tonality, 59, 83- 107. Feather, N. T. (1991~).Variables relating to the allocation of pocket money to children: Parental reasons and values. British Journal of Social Psychology, 30, 221-234. Feather, N. T. (1992a). An attributional and value analysis of deservingness in success and failure situations. British Journal of Social Psychology, 31, 125- 145. Feather, N. T. (1992b). Values, valences, expectations, and actions. Journal of Social Issues, 48, 109- I 24. Feather, N. T. (1993a). Attitudes toward the high achiever: Value correlates and family resemblances. Unpublished manuscript. Feather, N. T. (1993b). Authoritarianism and attitudes toward high achievers. Journal of Personality and Social Psychology. 65, 152-164. Feather, N. T. (1993~).The rise and fall of political leaders: Attributions, deservingness, personality, and affect. Australian Journal of Psychology, 45, 61-68. Feather, N. T., & McKee, 1. R. (1992). Australian and Japanese attitudes towards the fall of high achievers Australian Journal of Psychology, 44, 87-93. Feather, N. T., & McKee, I. R. (1993). Global self-esteem and attitudes toward the high achiever for Australian and Japanese students. Social Psychology Quarrerly, 56, 65-76. Feather, N. T., Volkmer, R. E., & McKee, I. R. (1991). Attitudes towards high achievers in public life: Attributions, deservingness, personality, and affect. Australian Journal of Psychology, 43, 85-91, Festinger, L. (1954). A theory of social comparison processes. Human Relations, 7, 117-140. Fincham, F., & Jaspars, J. (1980). Attribution of responsibility: From man the scientist to man as lawyer. Advances in Experimental Social Psychology, 13, 81- 138. Fiske, A. P. (1991). Structures of social life: The four elementary forms of human relations. New York: Free Press. Folger, R. (1984). The sense of injustice: Social psychological perspectives. New York: Plenum. Folger, R. (1987). Reformulating the conditions of resentment: A referent conditions model. In J. C. Masters & W. P. Smith (Eds.), Social comparison, social justice, and relative deprivation: Theoretical, empirical, and policy perspectives (pp. 183-215). Hillsdale, NJ: Erlbaum. Foster, G. M. (1972). The anatomy of envy: A study in symbolic behavior. Current Anthropology, 13, 165-186. Fyans, L. J., Salili, F., Maehr, M. L., & Desai, K. A. (1983). A cross-cultural exploration into the meaning of achievement. Journal of Personality and Social Psychology, 44, 1000-1013. Guttman, L. (1968). A general nonmetric technique for finding the smallest coordinate space for a configuration of points. Psychometrika, 33, 469-506. Hamilton, V. L., Blumenfeld, P. C., Akoh, H., & Miura, K. (1990). Credit and blame among American and Japanese children: Normative, cultural, and individual differences. Journal of Personality and Social Psychology, 59, 442-45 1. Hart, H. L. A. (1968). Punishment and responsibility. New York: Oxford University Press. Hart, H. L. A., & Honore, A. M. (1959). Causation in the law. Oxford: Oxford University Press. Heider, F. (1958). The psychology of interpersonal relations. New York: Wiley. Helmreich, R., Aronson, E., & Le Fan, J. (1970). To err is humanising-sometimes: Effects of selfesteem, competence, and a pratfall on interpersonal attraction. Journal of Personality and Social Psychology, 16, 259-264. Hogg, M. A , , & Abrarns, D. (1988). Social identifications: A social psychology of intergroup relations and group processes. London: Routledge. Hollander, E. P. (1964). Leaders, groups, and injuence. New York: Oxford University Press. Lebra, T. S. (1976). Japanese patterns of behavior. Honolulu: University of Hawaii Press. Lerner, M. J. (1980). The beiief in a just world: A fundamental delusion. New York: Plenum Press.

72

N. T. FEATHER

Lerner, M. 1. (1991). Integrating societal and psychological rules of entitlement: The basic task of each social actor and fundamental problem for the social sciences. In R. Vermunt & H. Steensma (Eds.), Social justice in human relations (Vol. I , pp. 13-32). New York: Plenum Press. Lewin, K . (1936). Principles of topological psychology. New York: McGraw-Hill. Lewin, K. (1951). Field theory in social science: Selected theoretical papers (D. Cartwright, Ed.). New York: Harper & Row. McClelland, D. C. (1985). Human motivation. Glenview, 1L: Scott, Foreman & Company. Maehr, M. L., & Nicholls, J. G. (1980). Culture and achievement motivation: A second look. In N. Warren (Ed.), Studies in cross-cultural psychology (Vol. 2, pp. 221-267). London: Academic Press. Major, B., Testa, M., & Bylsma, W. H. (1991). Responses to upward and downward social comparisons: The impact of esteem-relevance and perceived control. In J. Suls and T. A. Wills (Eds.), Social comparison: Contemporary theory and research (pp. 237-260). Hillsdale, NJ: Erlbaum. Markus, H. R., & Kitayama, S. (1991). Culture and the self: Implications for cognition, emotion, and motivation. Psychological Review, 98, 224-253. Masters, J . C., & Smith, W. P. (Eds.) (1987). Social comparison, social justice, and relarive deprivation: Theoretical, empirical. and policy perspectives. Hillsdale, NJ: Erlbaum. Mikula, G. (Ed.). (1980). Justice and social interaction. Bern: Huber. Mikula, G. (in press). On the experience of injustice. European Review of Social Psychology, 4 . Mirels, H. L., & Garrett, I. B. (1971). The Protestant ethic as a personality variable. Journal of Consulting and Clinical Psychology, 36, 40-44. Ortony, A,, Clore, G. L., & Collins, A. (1988). The cognitive structure of emotions. New York: Cambridge University Press. Parrott, W. G. (1991). The emotional experiences of envy and jealousy. In P. Salovey (Ed.), The psychology of jealousy and envy (pp. 3-30). New York: Guilford. Piaget, I. (1932). The moral judgment of the child. New York: Harcourt Brace. Prentice, D. A., & Crosby, F. (1987). The importance of context for assessing deservingness. In J. C. Masters & W. P. Smith (Eds.), Social comparison. social justice, and relative deprivation: Theoretical, empirical, and policy perspectives. Hillsdale, NJ: Erlbaum. Ramson, W. S. (1988). Australian national dictionary. Melbourne: Oxford University Press. Rokeach, M.(1973). The nature of human values. New York: Free Press. Rosenberg, M. ( 1965). Sociery and the adolescent self-image. Princeton, NJ: Princeton University Press. Sabini, J., & Silver, M. (1982). Moralities of everyday life. Oxford: Oxford University Press. Salovey, P. (1991). Social comparison processes in envy and jealousy. In J. Suls & T. A. Wills (Eds.), Social comparison: Contemporary theory and research (pp. 261-286). Hillsdale, NJ: Erlbaum . Salovey, P., & Rodin, J. (1984). Some antecedents and consequences of social-comparison jealousy. Journal of Personality and Social Psychology, 47, 780-792. Salovey, P., & Rothman, A. J. (1991). Envy and jealousy: Self and society. In P. Salovey (Ed.), The psychology of jealousy and envy (pp. 271-286). New York: Guilford. Scheler, M. (1961). Ressentiment (L. A. Coser, Ed., W. W. Holdheim, Trans.). Glencoe, IL Free Press. (Original work published 1915) Schoeck, H. (1969). Envy: A theory of social behaviour. London: Secker & Warburg. Schwartz, S . H. (1992). Universals in the content and structure of values: Theoretical advances and empirical tests in 20 countries. Advances in Experimental Social Psychology, 25, 1-65. Schwartz, S. H., & Bilsky, W. (1987). Toward a universal psychological structure of human values. Journal of Personality and Social Psychology, 53, 550-562. Schwartz, S. H., & Bilsky, W. (1990). Toward a theory of the universal content and structure of values: Extensions and cross-cultural replications. Journal of Personality and Social Psychology, 58, 878-891.

ATTITUDES TOWARD HIGH ACHIEVERS

73

Shaver, K. G. (1985). The attribution of blame: Causality, responsibility, and blameworthiness. New York: Springer-Verlag. Simpson, J.A., & Weiner, E.S.C. (1989). The Oxford English Dictionary (2nd ed.). Oxford: Clarendon Press. Silver, M., & Sabini, J. (1978). The perception of envy. Social Psychology, 41, 105-1 17. Smith, R. H. (1991). Envy and the sense of injustice. In P. Salovey (Ed.), The psychology ofjealousy and envy. New York: Guilford. Spence, J. T. (Ed.). (1983). Achievement and achievement motives: Psychological and sociological approaches. San Francisco: Freeman. Suls, J . , & Wills, T. A. (Eds.). (1991). Social comparison: Contemporary theory and research. Hillsdale, NJ: Erlbaum. Tajfel, H. (1978). Dixerentiation between social groups: Studies in the social psychology of intergroup relarions. London: Academic Press. Tajfel, H . , & Turner, J. C. (1985). The social identity theory of intergroup relations. In S. Worchel & W. G . Austin (Eds.), Psychology of intergroup relations (pp, 7-24). Chicago: Nelson-Hall. Tesser, A. (1988). Toward a self-evaluation maintenance model of social behavior. Advances in Experimental Social Psychology, 21, 181-227. Tesser, A. (1991). Emotion in social comparison and reflection processes. In J. Suls & T. A. Wills (Eds.), Social comparison: Contemporary theory and research (pp. 117- 145). Hillsdale, NJ: Erlbaum. Tesser, A,, & Collins, J. (1988). Emotion in social reflection and comparison situations: Intuitive, systematic, and exploratory approaches. Journal of Personality and Social Psychology, 55, 695709. Triandis, H. C. (1990). Cross-cultural studies of individualism and collectivism. In J. J. Berman (Ed.), Nebraska symposium on motivation, 1989: Cross-cultural perspectives (pp. 41 - 133). Lincoln, NE: University of Nebraska Press. lbmer, J. C., Hogg, M. A , , Oakes, P. J . , Reicher, S. D., & Wetherell, M. S . (1987). Rediscovering the social group: A self-categorization theory. Oxford: Basic Blackwell. Walster, E., Walster, G . W., & Berscheid, E. (1978). Equify: Theory and research. Boston: Allyn & Bacon. Weber, M. (1976). The Protestant ethic and the spirit of capitalism (T. Parsons, Trans.). London: George Allen & Unwin. (Original work published 1904-1905) Weinberger, J., & McClelland, D. C. (1990). Cognitive versus traditional motivational models: Irreconcilable or complementary? In E. T. Higgins & R. M . Sorrentino (Eds.), Handbook of motivation and cognition: Foundations of social behavior (Vol. 2, pp. 562-597). New York: Guilford. Weiner, B. (1986). An attributional theory of motivation and emotion. New York: Springer-Verlag. Weiner, B. (1992). Human motivation: Metaphors, theories. and research. Newbury Park: Sage. Weiner, B., Graham, S . , & Chandler, C. (1982). Causal antecedents of pity, anger, and guilt. Personality and Social Psychology Bulletin, 8 , 226-232. Weiner, B., & Kukla, A. (1970). An attributional analysis of achievement motivation. Journal of Personality and Social Psychology, I S , 1-20. Weiner, B., Perry, R. B., & Magnusson, J. (1988). An attributional analysis of reactions to stigmas. Journal of Personaliiy and Social Psychology, 55, 738-748. Weisz, J. R., Rothbaum, F. M., & Blackburn, T. C. (1984). Standing out and standing in: The psychology of control in America and Japan. American Psychologist, 39, 955-969. Wheeler, L. (1991). A brief history of social comparison theory. In J. Suls & T.A. Wills (Eds.), Social comparison: Contemporary theory and research (pp. 3-21). Hillsdale, NJ: Erlbaum. Wood, J. V., & Taylor, K . L. (1991). Serving self-relevant goals through social comparison. In J. Suls & T. A. Wills (Eds.), Social comparison: Contemporary theory and research (pp. 2349). Hillsdale, NJ: Erlbaum.

This Page Intentionally Left Blank

EVOLUTIONARY SOCIAL PSYCHOLOGY: FROM SEXUAL SELECTION TO SOCIAL COGNITION

Douglas T. Kenrick

I. Introduction Two revolutionary interdisciplinary syntheses have been generating increasing energy among social psychologists and other social scientists over the last decade or two. Evolutionary psychology incorporates new findings and theoretical insights in behavior genetics, ethology, anthropology, and psychology (Barkow, Cosmides, & Tooby, 1992; Lurnsden & Wilson, 1981). Cognitive science encompasses developments in cognitive psychology, computer science, linguistics, anthropology, and neuroscience (Gardner, 1985; Lakoff, 1987). Thus far, these two developments have been somewhat independent. However, evolutionary psychology and cognitive science are both centrally concerned with the structure and function of the human brain, and it seems inevitable that there will be a higher level fusion. Because the evolution of the human brain has been shaped at least as much by the challenges of social group living as it has by the challenges of letter recognition, social psychology should necessarily be at the epicenter of the eventual meta-fusion. An incidental byproduct should be an incineration of the artificial barriers between researchers working on social cognition, interpersonal processes, and personality traits, and beyond that, between researchers in social/personality and researchers in experimental, developmental, and clinical psychology. Unfortunately, some of the long-standing epistemological biases of social psychologists raise the question of whether we will ourselves make the heat or simply stand by with wide eyes and singed eyebrows. Those biases include a sometimes dogmatic preference for small-range mini-theories, a very narrow understanding of the physiological underpinnings of social behavior, and a sometimes schizophrenic avoidance of any discussion of (1) individual differences between humans, (2) a species-specific human nature, or (3) the similarities and ADVANCES IN EXPERIMENTAL SOCIAL PSYCH0U)GY. VOL. 26

15

Copyright 0 1994 by Academic Press, Inc All nghts of reproduction In any form reserved

76

DOUGLAS T.KENRICK

differences between humans and related species. Of course, it is doubtful that many social psychologists believe that humans did not evolve, or that the research conducted by physiologists, geneticists, animal ethologists or anthropologists is not of interest in its own right. Many simply seem to feel that research in those areas is just not directly relevant to studies of ongoing social information processing or interrelationships between people living in modern society. The main goal of this paper is to argue that a consideration of the evolutionary literature is not only a potentially useful tool in generating hypotheses about social psychological phenomena but that it is also an essential framework for a full understanding of the phenomena we have already uncovered. A secondary goal is to show that social psychological methods and findings can help provide essential pieces of the puzzle connecting evolutionary psychology and cognitive science. Evolutionary biologists and anthropologists have been at least as ill informed about social psychology as we have been about developments in their disciplines, and the mutual ignorance has worked to their detriment as well. This article will proceed in three general sections. In the first section, I will summarize the general principles of evolutionary social psychology. In the second section, I will consider a number of studies my colleagues and I have conducted in the area of gender differences in attraction and mate selection that wed the traditional evolutionary and social psychological approaches. Perhaps because of the importance of individual differences and differential reproduction to evolutionary theory, mate selection is one of the research areas that has resulted in a great deal of fruitful cross-fertilization (e.g., Buss, 1989; Feingold, 1992; Kenrick & Keefe, 1992). In the final section, I will argue that the implications of these and related findings should concern all social psychologists, and not only those who study heterosexual relationships. Several researchers working in the areas of aggression, altruism, and group processes have considered the ultimate implications of their findings (e.g., Daly & Wilson, 1988a; 1988b; Rushton, 1989; Sidanius, Cling, & Pratto, 1991). However, I will argue that an evolutionary perspective also raises questions that should be of central concern to researchers in all areas, even those working on very “proximate” problems of ongoing social cognition. Researchers studying topics such as social schemata, person memory, self-concepts, or the influence of affect on social cognition must begin to consider the ultimate functional significance of these processes if they wish to gain a full understanding of the significance of their findings and of how those findings relate to findings in other areas of social psychology, other subdisciplines of psychology, and other scientific disciplines beyond our own field.

11. General Principles of Evolutionary Psychology

Evolutionary theory consists, at base, of three simple assumptions, outlined by Charles Darwin (1859/ 1958) in his book Origin ofSpecies. The first assumption

EVOLUTIONARY SOCIAL PSYCHOLOGY

77

is that animals are engaged in a struggle for existence. Even slowly reproducing animals, such as elephants, reproduce so rapidly that any species would overrun the earth in a few centuries. This does not happen, because resources are limited: When the elephant population becomes too large, the pachyderms begin to exhaust their supply of vegetation and some of them starve. The resources necessary for survival are limited, and other members of the same species are the main competitors for those resources. Therefore, animals of a given species must struggle against one another to survive. Elephants must beat the other elephants to the limited greenery on the trees. The second crucial assumption of evolutionary theory presumes heritable variation within a species. This assumption, quite controversial when Darwin first advanced it, is that animals within a species are not all exactly the same; they differ in many ways and can pass some of those differences along to their offspring. Darwin (1859/ 1958) showed how such variations were exploited by pigeon domesticators: The diversity of breeds is something astonishing. The carrier, more especially the male bird, is . . . remarkable from the wonderful development of the carunculated skin about the head; and this is accompanied by greatly elongated eyelids, very large external orifices to the nostrils, and a wide gape of mouth . . . the common tumbler has the singular inherited habit of flying at a great height in a compact flock, and tumbling in the air head over heels. . . . The pouter has a much elongated body, wings, and legs; and its enormously developed crop, which it glories in inflating, may well excite astonishment and even laughter. . . . Great as are the differences between the breeds . . . all are descended from the rock pigeon. (pp. 41-42)

One can readily observe the results of such selective breeding in domesticated dogs. Although all dogs are descended from the same species, Chihuahuas, Great Danes, and Saint Bemards have been selectively bred to pass on obviously different sets of characteristics to their offspring. The third principle of evolution is that of natural selection, and it follows from the other two. If animals must compete with one another to survive, and if animals vary in ways that can be inherited, then animals whose variations assist in the struggle for survival will have more offspring. Those offspring will, in turn, have relatively more offspring than less well adapted strains, and so on. Over generations, the genotypes that are most closely suited to their particular environments will replace those that are less well adapted.

A. NATURAL SELECTION AND BEHAVIOR Most people understand how Darwin’s theory applies to the physical characteristics of animals, but often misunderstand the implications of the theory for behavior. As Darwin made clearer in his later book, The Expression of Emotions

78

DOUGLAS T. KENRICK

in Man and Animals (1872/1965) behaviors evolve in the same way as do physical features. A little thought reveals why this should be the case. Seals are closely related to wolves, but if a seal inherited a brain programmed to run a wolf’s body and tried to run down large-hoofed mammals on dry land, it would not survive well; neither would a wolf that attempted to swim out to sea and dive for fish 50 feet below the surface. Along with their bodies, seals, wolves, bats, giraffes, and wallabies inherit brains programmed to do certain things with those bodies. Thus, evolution applies to survival-related behaviors in much the same way that it applies to physical characteristics. Those animals with behavioral variations most suited to their environments (and to their bodily equipment) will survive and out-reproduce those animals with less well adapted behavioral variations. Behavior genetics research provides evidence that behavioral characteristics can be selectively bred in the same way as are morphological characteristics. For instance, by interbreeding the most aggressive or least aggressive rats from a litter for several generations, it is possible to produce strains of especially gentle and especially violent rodents (Lagerspetz, 1979). The same process has occurred in domesticated dogs: Some strains are notoriously aggressive, such as pit bulls; some are notoriously good-natured, such as golden retrievers; and some are notoriously jittery, such as Irish setters.

B. UNIVERSALS AND HUMAN BEHAVIOR Evolutionary psychologists since Charles Darwin have been especially interested in universals of human behavior. In his Expression of Emotions, Darwin noted the importance of cross-cultural research to his theory, and in fact collected some preliminary data on the universality of emotional expression, which has since been corroborated in the work of Ekman and his colleagues (Ekman & Friesen, 1971; Ekman et al., 1987). This work has suggested that people in different societies around the world attribute the same meaning to the same emotional expressions (smiles always indicate happiness, and the thrusted tongue face that Americans would take to mean disgust means disgust in Japan or in Brazil). This agreement cannot be attributed to similar cultural influences because it is found even in comparisons involving members of isolated Stone Age cultures with little or no exposure to outsiders (Ekman & Freisen, 1971). One of the most exciting developments in the area of evolutionary social psychology has been the discovery of a number of such universals of social behavior that are difficult to explain using traditional sociocultural or learning approaches. For instance, Eibl-Eibesfeldt filmed natural instances of flirtation in numerous cultures and noted that micro-analyses of the movement patterns led to the discovery of universal flirtation gestures. Eibl-Eibesfeldt (1 975) reported “agreement in the

EVOLUTIONARY SOCIAL PSYCHOLOGY

79

smallest detail in the flirting behavior of girls from Samoa, Papua, France, Japan, Africa (Turcana and other Nilotohamite tribes, Himba, Bushmen) and South American Indians (Waika, Orinoko)” (p. 465). A number of other universal patterns of human social behavior have been found (Kenrick & Trost, in press). For instance, males commit over 85% of all homicides in every society that has ever been investigated (Daly & Wilson, 1988b). There are also a number of similarities in mate preferences that are consistently found across cultures (Buss, 1989; Daly &Wilson, 1983; Kenrick & Keefe, 1992). We will discuss these in some detail below.

C. DIFFERENT STRATEGIES WITHIN AND BETWEEN SPECIES Because all the members of a particular species must meet certain environmental demands, natural selection often leads to a reduction of variance on certain characteristics-most humans have two eyes and walk upright, for instance. It is for this reason that evolutionary theorists often search for species-typical characteristics, such as the flirtation gestures described above. However, it is a misconception to assume that natural selection removes all the important differences among the members of a species. If that were the case, neither artificial selection nor natural selection would have raw materials with which to work. Behavior genetic studies of twins and adoptees have supported the assumption that humans, like other animals, have many heritable variations that affect behavior (Plomin, DeFries & McClearn, 1990). Variations in personality characteristics ranging from friendliness and intelligence through depression and schizophrenia appear to involve heritable components. Some of the differences within a species are due to genetic variations within that species (as is the case with Irish setters and Doberman pinschers); others are due to interactions between features of the environment and species-wide genetic proclivities keyed to those variations in the environment. There are several reasons why evolutionary pressures will not lead all animals sharing a particular environment to be identical. For instance, if the resources in the environment vary in distribution from time to time or from place to place, or if there are advantages to heterozygosity (as in the case of sickle-cell anemia), different forms of the same animal may persist. Perhaps most interesting to social psychology is a process that evolutionary theorists call frequency-dependent strategiescertain behavior patterns are adaptive only when they are not universal. This notion of a frequency-dependent strategy should be familiar to psychologists who have studied game theory, which has also influenced work on laboratory conflict and cooperation. The “defection” strategy in a prisoner’s dilemma game is successful in winning points only if a player can assume that others are unlikely to

80

DOUGLAS T. KENRICK

use it, and players will vary their strategies based on their estimates of what other players will do. It appears that other animals are programmed to vary their life strategies based on a combination of their own phylogenetically programmed predispositions and their ontogenetic experiences with their social and physical environments. Within the area of behavioral ecology, the metaphor of “hawks and doves” is commonly used to explicate this principle (e.g., Dawkins, 1976). If nonaggressive doves predominate in an environment, it pays to behave like an aggressive hawk and to prey on other birds. As the number of hawks increases, however, it becomes increasingly dangerous to attack other birds (who may viciously counterattack). Under these circumstances, a pacifist dove (who runs from any conflict) will fare better. Evolutionary theorists assume that different species of animals who share the same environment often maintain an equilibrium. If the hawk population increases, they begin to destroy the dove population upon which they prey. Hawks begin to starve and die, and the dove population increases. Of course, when the dove population increases dramatically, there are more prey for hawks, and the hawks increase again. With this type of interdependence, the populations of hawks and doves will tend to stabilize at some mutually limiting equilibrium. Such a reciprocal equilibrium can occur across species, as in the case of hawks and doves, but it may also occur within a species. For instance, there are two types of adult male blue-gilled sunfish (Gross, 1984). One type is a large territorial male whose colorful body is highly attractive to females. A second type of male is smaller and drab in appearance, resembling the less resplendent female. These smaller males are known as “sneak-copulators.” Rather than investing nutritional energy in developing a large, flashy physique, they develop enormous sperm-producing organs. When a large territorial male is mating with a female, the smaller male will thwart his larger opponent by darting in and releasing his sperm. Obviously, the success of the smaller males’ strategy depends partly on the existence of the larger males in the vicinity to attract females, and is decreased by too many other smaller sneak-copulators in the neighborhood. As another example, some animals can actually change from one physical form into another depending upon “environmental openings.” In the cleaner wrasse, for instance, females are most numerous and congregate in harems around a large, territoried male. When a large male dies, the largest female in his harem goes through a series of rapid physiological changes during which she grows larger and transforms into a male (Warner, 1984). Thus, the success of a particular combination of body type and behavior is linked to variations in the environment, and some species have evolved to change body types as the environment changes. Do humans play more than one mating strategy? Several authors have speculated about how notions such as frequency-dependent selection might apply to humans (Kenrick & Trost, 1987; Rushton, 1985; Simpson & Gangestad, 1992).

EVOLUTIONARY SOCIAL PSYCHOLOGY

81

Such models have been applied to differences in relative promiscuity, for instance (Simpson & Gangestad, 1992). As has been found with animals, environmental variations have been found to lead to changes in hormones related to bodily characteristics and sexual behavior. For instance, successful competition has been found to be linked to increased testosterone flow (Mazur, Booth, & Dabbs, 1992). The connections between these findings and long-term differences in mating strategies remain speculative at this point. However, the clearest support for an evolutionary model of within-species differences in human mating strategies comes from research on gender differences, to which I now turn. The most prevalent morphological and behavioral divisions of labor within species are based on gender. Males and females of a given species tend to differ in size and in behavior. Many of these differences are unique to particular species: The nature of the differences between a peacock and a peahen are not the same as those between a bull and a cow. However, some wide generalizations in sex differences can be found across a wide range of vertebrate species. Darwin noted in his Origin of Species (185911958) that males tend to be relatively larger and more showy. If one member of a fish species has more decorative fins (as in the Siamese fighting fish), if one member of a bird species has more colorful plumage (as in the peacock), or if one member of a mammalian species has larger antlers (as in the elk), it tends to be the male. There are fairly general behavioral differences between the sexes as well. Males tend to be more aggressive and more inclined toward dominance competitions. These sex differences are associated with some common differences in mating arrangements across species. In particular, polygyny (one male, many females) is more common than polyandry (one female, many males) among vertebrates. This mammalian generalization holds for humans as well. Of 849 societies examined in Murdock’s Ethnographic Atlas, 708 were polygynous, whereas only 4 were polyandrous (Daly & Wilson, 1983). In addition, each of the 4 polyandrous societies were also polygynous, whereas the reverse was not true. Two general principles are often used to explain these differences: differential parental investment and sexual selection.

D. DIFFERENTIAL PARENTAL INVESTMENT AND SEXUAL SELECTION Dijferential parental investment refers to the fact that males and females are initially different in the minimum amount of resources they must invest in order to reproduce (Trivers, 1972). For all animal species, eggs are generally more costly to produce than sperm. In species that utilize internal fertilization, as do most mammals, this difference is enhanced considerably. To produce a single offspring, a mammalian female must carry a fetus that requires a large amniotic sac and that absorbs considerable nutrition from her body. It is generally believed

82

DOUGLAS T. KENRICK

that the higher ratio of body fat to muscle in the human female originally stemmed from the need to ensure survival of the costly fetus in nutritionally uncertain times (Frisch, 1988). Following birth, the female mammal nurses the newborn, again sacrificing her own nutritional intake to provide nutrition for her progeny. Female elephant seals lose two kg for every kilogram gained by their pups, and the chance of a female red deer surviving from one season to the next goes down if she bears offspring (Clutton-Brock, 1984; Trivers 1985). In some species, such as humans, the offspring must be fed and cared for even after they are weaned. Therefore, the minimum female parental investment is quite large. Males, on the other hand, could father a child with a very low investmentthe amount of energy required for one act of intercourse. The record number of legitimate children recorded for one man is 899 (Daly & Wilson, 1983), and the actual number could conceivably be higher than that. It is biologically difficult for a woman to have more than 24 children, regardless of the number of husbands she has. For example, among the Xavante, a hunter-gatherer group, the average number of offspring for males and females is 3.6. However, the variance for women is 3.9, whereas for men it is 12.1. In other words, some Xavante men have many offspring, some have very few. Only 1 of 195 Xavante women are childless at age 20; whereas 6% of men are still childless by age 40. One man in the group had 23 children, whereas the highest number of children for a woman was 8 (Daly & Wilson, 1983, Salzano, Neel, & Maybury-Lewis, 1967). Some of the physical differences between males and females are due simply to natural selection based on this differential parental investment. Females need a different body to produce eggs, and in the case of mammals, other specialized physical features are required to nurture the fetus and the newborn baby. However, those different parenting requirements are not sufficient to explain the vast differences between the sexes. One might expect that a slightly larger body would be of more use to a female mammal (Ralls, 1976). (The female must contribute her physical resources directly to the offspring). However, males tend to be larger in most vertebrate species. Males are also more likely to have decorative features like antlers and peacock’s feathers, and to use some of those features (such as antlers) to compete with one another. Darwin used the concept of sexual selection to explain such differences. Sexual selection consists of two separable processes. Zntrasexual selection refers to the selection pressure that the members of one sex exert on one another via competition. In a species in which males compete for access to females by butting their heads, those individuals with the boniest heads, the strongest shoulder muscles, and the largest antlers will be more likely to win dominance competitions and survive. Epigamic selection is the other part of sexual selection. If one sex selects sexual partners on the basis of a certain feature, such as the possession of large antlers or bright displays of feathers, those features will be more characteristic of one sex than the other. Indeed, experimental studies have verified that sexual selection does operate in the manner that Darwin suspected. For example, Anderson (1982) manipulated

EVOLUTIONARY SOCIAL PSYCHOLOGY

83

the lengths of tail-feather displays in widowbirds. Because males have longer tails, and these do not appear to serve any obvious survival function, Anderson suspected that longer male tails were selected via female choice. Consistent with this reasoning, females were more likely to mate with males whose tails were experimentally elongated, and to reject males whose tails were experimentally shortened. Darwin suspected that epigamic sexual selection applies more to female choice of males, and female choice could explain why male vertebrates tend to be larger, showier, and more dominance-oriented. The reason that females are more likely to be choosy relates back to the concept of differential parental investment. Because females have an initially higher investment in their offspring, they are better served to be selective about their partners. An ill-chosen mating partner is, on average, less likely to be costly for a male. This inherent difference is most pronounced in “arena” species, such as the fallow deer (Clutton-Brock, 1991). In such species, males compete for one of a small number of territories, and only males who have a territory are attractive to females. The territoried males mate with many females in the course of a single day. These males appear to be completely nonselective, whereas females will mate only with the most dominant males. A number of human sex differences seem to fit a pattern in which females are relatively more selective about mating partners and males are relatively less selective (Daly & Wilson, 1983; Hinde, 1984; Townsend & Levy, 1990). In choosing mates, human females, compared to males, also appear to be relatively more attentive to signs of dominance and status (Feingold, 1992; Sadalla, Kenrick, & Vershure, 1987; Kenrick, Sadalla, Groth, & Trost, 1990). In sum, the inherent difference in parental investment has led to a general tendency for female vertebrates to be more selective about their mating partners. Females are likely to mate selectively with males who have demonstrated their ability to compete successfully with other males. In partial response to female selection, males compete with one another for positions in dominance hierarchies. This general pattern of sex differences has been used to explain certain features of human behavior. It is thus possible to make some broad generalizations about behavior that span a wide range of species. However, some of the most interesting evolutionary analyses go beyond these broad generalizations and consider the unique adaptations of particular species to particular physical and social ecologies. In the next section, we consider some hypotheses that follow from a consideration of the unique pattern of parental investment in humans.

E. QUALIFYING THE PARENTAL INVESTMENT MODEL IN APPLICATION TO HUMANS This model of differential parental investment and sexual selection provides a reasonable explanation for the behaviors of species like the fallow deer and the peacock, but it cannot be applied to humans without some important qualifica-

84

DOUGLAS T. KENRICK

tions. Human males invest heavily in their offspring. In a monogamous relationship, the man may provide resources for one wife and her offspring for all of his adult life. Because human males invest heavily in their offspring, they should, like females, be selective about choosing a mating partner. In fact, sex differences tend to be diminished in species in which the males invest in their offspring (Lancaster, 1985). Male and female humans are relatively similar in size and decoration in contrast to peacocks and peahens, or even to our more sexually unrestricted primate cousins such as baboons. This suggests that our male and female ancestors were more nearly equal in exerting epigamic sexual selection pressures on one another. Human males are therefore expected, unlike fallow deer, to be selective in choosing mates. However, males would not be expected to choose females along the same dimensions that females use for choosing males. Men and women invest different resources, so one would expect the two sexes to value different characteristics in a mate. Males invest indirect resources (such as food, money, protection, and security). Thus, women would be expected to value men who show the ability to provide those resources. On the other hand, women directly invest their bodily resources in the offspring. Thus, males would be expected to value women who demonstrate signs of fertility and health. From an evolutionary perspective, human males and females are also expected to differ in the age of mates they find most attractive. Aging limits a woman’s reproductive potential, and around age 50, ends it through menopause. Menopause is an unusual adaptation, and it is not found in our primate relatives (Lancaster & King, 1985). One explanation of why human females do not continue breeding until they die is related to the relative advantages of grandparental effort. As the mother’s age increases, so do maternal and infant mortality as well as birth defects. At some point, therefore, it becomes more genetically costly for a female to attempt to produce more offspring than to invest in her existing offspring and grandchildren (Alexander, 1987). Menopause would have been adaptive in a species that was long-lived and in which offspring were likely to have continued living in proximity to their mothers, as has been the case for most of human evolutionary history. Probably because of the low physical demand that reproduction places on the male body, males do not undergo menopause, and older males are still capable of fathering offspring (Nieschlag & Michel, 1986). Male health does decrease with age, but the features that females find attractive, such as economic resources and social status, continue to accrue well past a male’s physical prime. These differential changes, decreasing female fertility and increasing male resources, would lead males and females to place different value on signs of youth and physical health. In fact, survey data across a number of cultures have suggested that males tend to value relatively younger partners, whereas females value relatively older partners (Buss, 1989). This is not to say that females disregard physical attractiveness. However, indications of attractiveness are different for the two sexes. In males, a mature large chin and a hint

EVOLUTIONARY SOCIAL PSYCHOLOGY

85

of beard are attractive, whereas in females, a youthful small chin and smooth skin are seen as relatively more attractive (Cunningham, Barbee, & Pike, 1990). Physiognomic features that are regarded as attractive in males seem to'be indications of dominance (Keating, Mazur, & Segall, 1971). Several authors have suggested that males' judgments of female physical attractiveness, on the other hand, seem closely related to signs of youth and physical health-which would historically have translated into reproductive capacity (Alley, 1992; Cunningham, 1986; Cunningham et al. 1990; Symons, 1979). Thus, human males and females are expected to be similar in that both may invest heavily in offspring and both are often selective about mating partners. However, males and females are expected to be selective along different dimensions. Females are expected to place more emphasis on physical and behavioral characteristics that signal maturity, resources, and social status; males are expected to place more emphasis on characteristics related to youth and physical attractiveness (as cues to health and reproductive capacity). Social psychologists who hear evolutionary theorists describe these sex differences may be tempted to note that we knew them all along. Do we need evolutionary theory to tell us that males are generally more promiscuous than women, but that when they are being selective, they are more likely to value physical attractiveness than dominance and social status? One response to this question is that an evolutionary perspective helps us understand why these sex differences exist. It also allows us to integrate our own research findings with a wide range of findings from a diverse range of disciplines including ethology, behavior genetics, endocrinology, and anthropology. A theory that helps us understand how the behaviors of hermaphroditic fish, of human twins raised apart, of monkeys exposed to testosterone, and of flirting Bushmen are all connected to the behavior of undergraduates in a laboratory in Michigan is simply more intellectually satisfying than an insular and microscopic proliferation of unconnected mini-theories. Parsimony is a somewhat aesthetic consideration, and there are circumstances when it can be ignored. However, there are other advantages of considering an evolutionary perspective as well. One advantage is that the perspective generates new research questions and has already led to several interesting findings. Another advantage is that the approach can better explain a number of existing findings that have been misinterpreted within existing social psychological models.

111. Integrating Social Psychology and Evolutionary Biology: ' h o Programs of Research on Human Mate Selection In this section, I will describe two programs of research in which my colleagues and I have been involved. These lines of research attempt to illustrate the heuristic and explanatory value of an evolutionary perspective. The first series of

86

DOUGLAS T.KENRICK

studies, conducted with Richard Keefe (1992), examines the phenomenon of sex differences in age preference. It is fairly common knowledge that females, in general, prefer to date and marry men slightly older than themselves, and that men, in general, show a complementary tendency to prefer females slightly younger than themselves. Social scientists have tended to explain this violation of the similarity-attraction principle in terms of cultural norms. However, we have collected data suggesting that the normative theories cannot explain the phenomenon. An evolutionary perspective provided a more parsimonious account of existing data, and generated new data that further undermined the adequacy of the normative account. The second series of studies, conducted with Gary Groth, Melanie Trost, and Edward Sadalla (1993) examined mate preferences for different levels of commitment in relationships. In that research, we used insights from social psychological studies of relationships to explain some inconsistencies generated by earlier evolution-based studies. By combining both approaches, it has been possible to understand the circumstances under which humans will act like “typical” polygamous mammals, with relatively high female selectivity and low male selectivity, and when humans will act like monogamous animals, with high selectivity by both sexes.

A. AGE PREFERENCES IN MATES: CULTURAL

NORMS OR BIOLOGICAL CONSTRAINTS? One of the most robust findings in the attraction/relationships literature is the matching principle-people tend to mate with others who are similar to them on most dimensions that have been examined (Antill, 1983; Byme, 1971; Critelli & Waid, 1980; Hendrick, 1981). These include physical attractiveness, wealth, religion, personality, political attitudes, degree of psychopathology, and even personal habits. Social psychological theorists invoke economic exchange to explain the similarity findings. Presumably, partners attempt to match themselves with others who have similar social value. The exceptions to the principle that “like prefers like” are few and worthy of careful examination. One consistent exception to the principle is the finding that females are attracted to older males, while males are attracted to younger females (Bolig, Stein, & McKenry, 1984; Cameron, Oskamp, & Sparks, 1977; Harrison & Saedd, 1977). Economic models attribute this age-preference complimentariness to gender discrepancies in social value. Females presumably exchange youth and physical attractiveness for economic security (Brehm, 1985). Why the gender difference in the valuation of age? Economic models generally explain this and other features of mate selection in terms of cultural norms. For example, Presser (1975) alludes to the age difference and offers a tentative explanation in terms of the “norm” that

EVOLUTIONARY SOCIAL PSYCHOLOGY

87

. . . a husband should be, or at least appear to be, mentally and physically superior to his wife. Not only should he be taller than she (for the appearance of superiority) but also older (which gives him the advantage of more time to become better educated and more experienced). (p. 202)

As another author puts it, “Traditionally, in our society, males have been valued for their economic success, and females for their physical attractiveness” (Brehm, 1985, p. 76, emphasis added). In a similar vein, Cameron et al. (1977) explain their findings that females prefer older, taller, high-status males as due to “traditional sex-role specifications . . . frequently valued as sex appropriate in American Society,” which specify that women should “look up to” their male partners (p. 29, emphasis added). Along similar lines, Deutsch, Zalenski, and Clark (1986) speculated that such sex differences are due to a “double standard of aging” in our society. Although social psychologists do not usually examine cross-cultural data, the above quotes reveal a tacit assumption that there are other societies in which gender differences in mate preference are fundamentally different from those found in our society (see also Rosenblatt, 1974). Evolutionary theorists have noted the same gender difference in age preference, but explained it in terms of inherent sex differences in the types of resources invested in the offspring (e.g., Symons, 1979). As indicated earlier, males invest indirect resources (like food, money, protection, and security) that don’t necessarily diminish as they get older. As noted earlier, women’s reproductive potential decreases with age, and ends with menopause around age 50. Males do not undergo the same change. Different predictions follow from the evolutionary and normative explanations of age differences in mate preference. Those who have adopted the normative perspective have assumed that the rule is a fairly simple one-analogous to the rule about differential height: Men should be a few years older (as they should be a few inches taller; e.g., Cameron et a]., 1977; Presser, 1975). None of the normative accounts I came across indicated that the size of the difference should vary with the age of the individuals involved. Because there is evidence that younger people (particularly teenage males) are particularly sensitive to sex-role norms (Deutsch et al., 1986), it might be expected that the difference would be slightly more pronounced among younger males. On the other hand, a consideration of the issue of the differential reproductive value of males and females leads to a different set of predictions. Because a female’s fertility reaches a peak in her mid-20s and then declines more rapidly than does a male’s, a male should change the preferred age difference between him and his partner as he ages. For a teenage male, the choice of a similar partner results in choosing someone with maximum remaining reproductive years, but low fertility. A woman slightly older than him, in her early 20s, on the other hand, still has many reproductive years, and higher fertility. So a

88

DOUGLAS T. KENRICK

teenage male should make little or no discrimination against women a few years older than him. For a man in his 40s, however; a woman’s remaining reproductive years should be a more important consideration, which should act against his using similarity as his only criterion. Because females mature earlier, it makes sense for teenage females to be attracted to older males. Older males also lose physical resources (like health and sexual “arousability”), but at the same time, they may gain in the indirect resources and social status that females generally look for in a mate. As we indicated earlier, an older man is fully capable of fathering offspring. Those factors may simply cancel one another out. Janet Leonard (1989) argued that, from an evolutionary perspective, a woman would optimize her reproductive potential by mating with a man ten years older. Such a male will have more resources and status than someone her own age. However, he is not so old that he will be likely to die while their children are still young. Thus, an evolutionary perspective makes a similar prediction to the normative perspective regarding female age preferences, but differs from the normative perspective in predicting that males will change their preferences as they grow older. Very young males will have no preference for younger women, and no bias against older women. As males grow older, they will progressively prefer females who are dissimilar in age, and increasingly younger than themselves. The other major difference between the two perspectives is that, from an evolutionary perspective, one would be led to expect that the gender difference in age preference will be relatively cross-culturally robust. Male and female differences in lifetime fertility no doubt vary across cultures, and are influenced by nutrition, age of parenting onset, and number of offspring. However, females in all societies bear the young and undergo menopause. 1 . Age Preferences in Mate Advertisements

As a first test of our evolutionary model, Kenrick and Keefe (1992) borrowed a technique that has been used in a number of studies of the “economic exchange” process. We analyzed a sample of 2 18 personal advertisements from three issues of a “singles” newspaper from the western United States. The research included only advertisements that specified the age of the advertiser and a minimum and/or maximum desired age for a partner. When the advertisements were broken down by advertiser’s gender and decade of age, they yielded the pattern shown in Fig. 1. As can be seen in the figure, female specifications remained fairly constant throughout the age range. Males, on the other hand, changed their preferences in a systematic fashion as their own age increased. Males in their 20s were equally attracted to women above and below their own age, specifying partners ranging, on average, from 5 years younger to 5 years older. As males got older, however, their preferences increasingly diverged from those of females in the same age

EVOLUTIONARY SOCIAL PSYCHOLOGY 0 Oldest preferred 0 Youngest preferred W

2

zr,

k-

w

*OI 10

2 5

0 a LL

0

W

0

z

W

a W

-10

LL

lL c3

-20

204

-

1

-20

10s 20s 30s 40s 50s 60s MALE’S AGE

10s 20s 30s 40s 50s 60s FEMALES AGE

Fig. I . Age differences preferred in mate advertisements, plotted as minimum and maximum differences from advertiser’s age (based on data from Kenrick & Keefe, 1992).

group. Among males in their 50s and 60s, the maximum acceptable age was several years below their own age, and the minimum specified was almost a generation discrepant from their own age. The same sex-differentiated pattern was found among a sample of East Coast men and women whose selfdescriptions (in singles advertisements in the Washingtoaim magazine) indicated that they were relatively wealthy (Kenrick & Keefe, 1992). When age preferences from mating advertisements are examined more carefully, then, they yield results that do not fit well with normative models. If there is a “norm” for men to prefer younger women, it should show up most markedly in younger men, who tend to be more concerned with behaving in a stereotypically sex-typed manner. Yet the interesting feature of younger men’s preferences is that they extended equally above and below their own ages. Thus, these data provide no evidence that a preference for younger women is a consistent feature of the normatively defined role for males. The results do fit well with an evolutionary model. However, the data from this study are inherently limited. For one thing, preferences expressed in singles advertisements may not reff ect preferences in the larger population. For another, such preferences may not transfer into actual reproductive behaviors because they may be limited by a number of constraints. Although older males might be interested in females who are several years younger than themselves, for instance, they must still deal with female preferences.

90

DOUGLAS T. KENRICK

2 . Age Direrences in American Marriage Statistics To determine whether the age differences expressed in singles advertisements would be reflected in actual mate choice, Kenrick and Keefe (1992) examined marriage statistics for all marriages (n = 753) for the month of January 1986 in Seattle, Washington. A random sample of marriages for the same year from Phoenix, Arizona, was also collected. The data were first analyzed using the male’s age as the predictor, and his partner’s age as the dependent variable. Next, the same data were analyzed using the female’s age as predictor. The results of both samples followed the same pattern found in the singles preferences. Figure 2 compares the Phoenix marriages with the Phoenix singles advertisements. The results for actual marriages were thus consistent with an evolutionary model, and they fell cleanly within the preference ranges found in the first study. This suggests that the pattern is not limited to singles’ advertisers, but that it applies to marriages, and thus has a high likelihood of translating into actual reproductive behavior. Consistent with this reasoning, Rajecki and his colleagues found that relatively younger females in their sample of singles advertisers received more responses to their advertisements (Rajecki, Bledsoe, & Rasmussen, 1991). From an evolutionary perspective, human mate preferences reflect selection

Phoenix marriages Phoenix advertisements

I0 0 Oldest preferred 0 Youngest preferred

a v, lo

i

0

-1 0

-20

--

10s 20s 30s 40s 50s 60s MALE’S AGE

-20

10s 20s 30s 40s 50s 60s FEMALE’S AGE

Fig. 2. A comparison of marriage ages and age preferences stated in personal advertisements in Phoenix during the 1980s (based on data from Kenrick & Keefe, 1992).

EVOLUTIONARY SOCIAL PSYCHOLOGY

91

pressures that have been operating over several million years. If the sex differences in age preference reflect past evolutionary constraints, then they should be relatively impervious to changing norms, like the ones that have affected sex roles in the United States during the latter half of this century. A sample of marriages recorded in Phoenix, Arizona, for 1923 (the earliest year for which organized records were available) yielded the same sex-differentiated pattern obtained in the 1980s studies (Kenrick & Keefe, 1992). 3. Cross-Cultural Mating Advertisements

As we indicated above, an evolutionary model would also lead us to expect that these sex differences in preferred age of a partner would be consistent across cultures. Guus van Heck of the University of Tilburg provided data from singles advertisements in Holland; and Ute Hoffman and Kirstin Schaefer of the University of Bielefeldt provided some similar German data. The patterns were, in both instances, the same as those obtained in American singles advertisements (Kenrick & Keefe, 1992). Of course, there are important similarities between American and European cultures. A more interesting comparison came from an examination of Indian marital advertisements. In contrast to American singles advertisements, Indian marital advertisements indicate very different cultural criteria. Most ads include strict limitations by caste and religion, and they commonly ask for horoscope information. An example follows:

Wanted: A non-Bharadwaj smart good-looking preferably employed Kerala lyer girl below 25 for a Kerala lyer boy 29. Chemical engineer. Contact with horoscope. (Times of India, Bombay, Sunday, January 29, 1989) Kenrick and Keefe (1992) obtained a sample of such Indian advertisements and again used those advertisements that included the advertiser’s age, and that stipulated a minimum and/or a maximum age for preferred partners. There were no advertisements for women above age 40, and an Indian informant explained that women who do not marry by their 30s are not considered marriageable. Nevertheless, gender differences found in the pattern of the Indian data are similar to those found in the American sample (see Fig. 3). A die-hard cultural determinist could still argue that modern residents of India and Europe share some cultural influences with residents of Phoenix in 1923. Kenrick and Keefe (1992) were also able to obtain data from the village of Poro-a small fishing community on a remote Philippine island, far removed from any urban area, and isolated from any European contact to this day. The

DOUGLAS T. KENRICK

92

0 Oldest preferred 0 Youngest preferred

2ol

“ 4 4

10

-20

20s 30s 40s 50s

MALES AGE

FEMALES AGE

Fig. 3. Age preferences indicated in marital advertisements from India (based on data from Kenrick & Keefe, 1992).

data in Fig. 4 describe all the marriages recorded on Poro during the years 19131939. In commentary published along with the original findings, two anthropologists examined data from several traditional African cultures, and found the same sexdifferentiated pattern to hold there (Broude, 1992; Harpending, 1992). Other w

9

2oj

-10

-20

1

-I u 10s 20s 30s 40s 50s 60s

MALE’SAGE

FEMALES AGE

Fig. 4. Age differences at marriage, based on all marriages on the island of Poro between 1913 and 1939 (based on data from Kenrick & Keefe, 1992).

93

EVOLUTIONARY SOCIAL PSYCHOLOGY

anthropological commentators argued that the pattern is more sex-differentiated in non-American cultures (Symons, 1992; Thornhill & Thornhill, 1992; Townsend, 1992; Van den Berghe, 1992). It thus seems that the explanation of this particular phenomenon in terms of American cultural norms was rather off the mark.

I

4 . Teenage Preferences

In addition to its problems with the cross-cultural data, the normative explanation also failed to explain why preferences of young males included older as well as younger females. Most of the data reported above includes males in their 20s as the youngest group. It is worth noting that the average age of marriage partner for the small number of teenage males in the Phoenix sample was actually older than their own age. We have recently obtained preference data from a sample of teenagers (Kenrick, Engstrom, Cornelius, & Keefe, 1992). Subjects ranging from 12 to 18 years of age were asked to indicate the youngest and oldest partners they would consider dating, and to indicate the most desirable age of a partner. Teenaged females were similar to older females in specifying partners ranging from around their own age to several years older. However, when the preferences of teenaged males are placed in the context of the adult singles advertisements, they indicate a particularly interesting lifetime trend (see Fig. 5).

Adult advertisements

w

20

8

1 2o

I

o Oldest preferred 0 Youngest preferred

1'

0

Teenagers

'I

- 1-

-20

10s 20s 30s 40s 50s 60s

-lo -20

10s 20s 30s 40s 50s 60s

MALES AGE FEMALES AGE Fig. 5. Age preferences stated by teenagers in the context of adult preferences stated in singles advertisements (based on data from Kenrick, Engstrom, Cornelius, & Keefe, 1992, and Kenrick & Keefe, 1992).

94

DOUGLAS T.KENRICK

Teenaged males were very different from older males in specifying ages ranging from a few months younger to several years older than themselves. Most interesting was the fact that the young males’ preferences extended much farther above their own age than below. Further, teenaged males indicated that the age of an ideal partner would be older than their own age. This preference not only goes against the likely reward structure determined by female preferences, but it also goes against the supposed social norm (which should operate most strongly on males during the highly sex-typed teenage years; Deutsch, et al., 1986). On the other hand, it is quite consistent with an evolutionary model-for a teenaged male, the most fertile females are older, not younger. B . PARENTAL INVESTMENT AT DIFFERENT RELATIONSHIP LEVELS

The studies discussed so far suggest some of the ways in which social psychological researchers might profit from a consideration of simple evolutionary principles. Given that the evolutionary biologists have Darwin on their side, it is not surprising that most of the cross-fertilization between social psychology and evolutionary biology has been in one direction-with the biologists giving, and the social psychologists taking. This state of affairs has been rather less like cross-fertilization than like artificial insemination. However, the theories and methods of social psychology offer tools that can be used to clarify issues for evolutionary theorists. As one example, my colleagues and I were able to use an insight from the social psychological literature to resolve a discrepancy between two lines of research based in evolutionary theory (Kenrick, Groth, Trost, & Sadalla, 1993; Kenrick, Sadalla, Groth, & Trost, 1990). In Section II,D, I discussed the concepts of differential parental investment and sexual selection. To recapitulate briefly, females are assumed to have an initially higher investment in potential offspring, and to therefore be more selective about the partners with whom they will mate. Males, on the other hand, provide less in the way of direct resources for any given offspring, and have less to lose from mating with any given available partner. As part of the process of sexual selection, Darwin suggested that females often may mate preferentially with males who have either won the struggle for dominance over others of their sex or who have “special weapons” suggesting their “general vigour” (Darwin, 185911958). Over generations, a preference for males who display obvious signs of superiority over their competitors will lead to a steep dominance hierarchy, and at same time, select for greater size and weaponry among males. This would explain the sort of dramatic sex differences one sees among baboons, where males are perhaps twice the size of females, and also more dominant. Male baboons are that way, at least in part, because of past mating choices made by female baboons.

EVOLUTIONARY SOCIAL PSYCHOLOGY

95

Consistent with this model, there is evidence that human females, like other mammalian females, are relatively more attentive to a male’s position in the dominance hierarchy (Hill, 1984; Mealey, 1985; Sadalla, Kenrick, & Vershure, 1987; Townsend & Levy, 1990). In one series of studies, Sadalla and his colleagues exposed college men and women to target people who varied in dominance. The subjects’ ostensible task was to attempt to guess what a target person’s personality was like from small amounts of information. In one study, dominance was manipulated nonverbally: The target person entered an office, sat down, and conversed with a same-sexed other in a manner that had previously been demonstrated to indicate either social dominance or submissiveness. In another study, subjects read an account of a tennis player who won most of his or her matches, but who was described by a coach as playing in a manner that tended to dominate (or be dominated by) his or her opponents. In a final study, subjects read bogus personality descriptors that included adjectives related to leadership and social dominance versus submissiveness. In the course of rating the target’s personality, subjects were asked to estimate how sexually attractive this person would be to members of the opposite sex. None of the various dominance manipulations had an effect on males’ judgments of a female target’s attractiveness, whereas female subjects in each experiment rated the dominant male as more sexually attractive. This gender/dominance effect was quite robust. Sadalla and his colleagues found it with each of the different operationalizations of dominance, in two distinct regions of country, with college students and with middle-aged factory workers (Sadalla & Fausel, 1980; Sadalla et al., 1987). Cross-cultural studies also indicate that high-status males are more successful at obtaining desirable wives, and in polygynous societies, at attracting multiple wives (e.g., Mealey, 1985; Turke & Betzig, 1985). On the other hand, Buss and Barnes (1986) also applied the differential parental investment model to human mate preferences, but found less prominent gender differences. They used a survey design and asked students to rank the characteristics they would prefer in a mate. Consistent with Buss and Barnes’ evolutionary perspective, females ranked traits related to resources as more important, and males ranked physical attractiveness as more important. However, the most striking feature of Buss and Barnes’ data was the similarity of male and female preferences-7 out of 10 of the highest rated preferences were identical for the two sexes. Buss and Barnes explained the lack of strong sex differences as due to the fact that differential parental investment and sexual selection are diminished in species like homo supiens, which are monogamous and in which most mating-age individuals pair off. If one explains sex differences in mateselection strategies in terms of evolutionary theory, and also argues that a lack of sex differences is consistent with an evolutionary perspective, a critic might justifiably complain that evolutionary hypotheses are untestable tautologies. To avoid this criticism, it is necessary to delimit our predictions. In this case, it is necessary to answer the question: When do humans act like polygynous species,

96

DOUGLAS T. KENRICK

with females being more selective (as in the Sadalla et al. 1987 data), and when do we act like monogamous species, with both sexes showing high selectivity (as in the Buss and Barnes 1986 data)? To answer this question, we turned to social psychological models of mate selection. Social psychologists have long argued for a consideration of longitudinal factors in relationships, noting that what is considered desirable at one phase of a relationship may be less important at another phase (e.g., Levinger, 1974; Murstein, 1970). Although there has been little support for earlier conceptualizations of rigid relationship “stages,” those approaches have been replaced with a notion of more flexible “phases” (Brehm, 1992; Levinger, 1983). Abundant evidence suggests that different patterns of exchange characterize different levels of involvement in relationships (cf. Clark 8z Mills, 1979; Clark & Reis, 1988). The longitudinal perspective helped clarify differences between the findings of Sadalla et al. (1987) and those of Buss and Barnes (1986). Buss and Barnes had asked their subjects about desirable characteristics in a marriage partner, whereas Sadalla et al. (1987) had asked people to rate the sexual attractiveness of a stranger. Although evolutionary theorists had not previously considered the longitudinal perspective, it is consistent with an evolutionary perspective that sex differences in selectivity will interact with the amount of time and resources partners have invested in one another. Monogamous species typically have a lengthy courtship that precedes mating; this presumably allows both members to evaluate one another’s values as a potential parent. Polygynous species, on the other hand, have brief courtships. If the above distinction is worthwhile, one would expect that sex differences would be most prominent in casual mating opportunities, for which males stand to invest less time and resources in any resulting offspring than females do. In choosing a marriage partner, on the other hand, a male may invest resources for a lifetime; thus males and females should both be selective. 1 . Two Studies of Minimum Standards for Partners at Different Relationship Levels

To examine the distinction suggested above, we asked college men and women about their minimum standards in a partner at four levels of involvement. Figure 6 depicts the results for the minimum intelligence that subjects demanded in a date, a sexual partner, a steady dating partner, and a marriage partner. In another study, we asked subjects about a partner for a “one-night stand,” a sexual partner whom the person would never see again. It was at this level of investment that females and males showed the largest differences in criteria. Regardless of the order in which these relationship levels were presented to the subjects, and regardless of the criterion trait in question, females demanded more than males did in a partner for a “one-night stand.” Figure 7 depicts the results for the intelligence variable.

97

EVOLUTIONARY SOCIAL PSYCHOLOGY Male

-._ ...

Female

Fig. 6 . Minimum intelligence criteria (expressed as percentile) for males and females at different levels of involvement (based on data from Kenrick, Sadalla, Croth, & Trost, 1990).

Several features of Figs. 6 and 7 are of interest. First, both sexes indicate increasing demands in moving from a single date to steady dating to mamage. A date represents a minimal investment of resources for both sexes, whereas a marriage represents a substantial investment. On the other hand, the interactions (significant in both cases) suggest that males seem to regard sexual relationships as (at least potentially) low investment, whereas females regard them as high investment. This is consistent with the fact that females could become pregnant, and potentially invest heavily from even a single sexual encounter. Males, on the other hand, could potentially reproduce with minimal investment (especially in the explicitly defined “one-night stand). One might wonder why, in an age of contraception, females and males might not have the same criteria. It is important to note that from an evolutionary perspective, the cognitive and affective mechanisms that underlie these differences in preference are the products of evolutionary pressures that operated over thousands of years, and not the products of “rational” reflection. Although the

98

DOUGLAS T.KENRICK

Fig. 7. Minimum intelligence criteria (expressed as percentile) for males and females at different levels of involvement, including “one-night stand” (based on data from Kenrick, Groth, Trost, & Sadalla, 1993).

very recent advances in birth control ought to influence a female’s decisions if the process were a “thoughtful” one, there is presumed to be a strong evolved “heuristic” operating. 2 . A Study of Selj-Appraisals, Equity, and Relationship Investment In addition to connecting with the research on relationship phases, this research can be connected with another aspect of research on relationships. Social psychological economic models have emphasized the importance of one’s selfevaluation as a mediator of one’s trading value in the heterosexual marketplace. As indicated earlier, social psychological models of relationship formation often focus on processes such as “social exchange,” “equity,” or “investment” (e.g.,

EVOLUTIONARY SOCIAL PSYCHOLOGY

99

Hatfield, Traupmann, Sprecher, Utne, & Hay, 1985; Rusbult, 1980). These economically based models adopt a view of potential partners as searching for the best bargain in exchange for their own social assets. According to this approach, social amenities such as beauty, intelligence, charm, wealth, and social status are traded for desirable characteristics in the partner. This approach is well expressed by the authors of a study of singles advertisements. Cameron et al. (1977) described singles ads as depicting a “heterosexual stock market,” and noted that “the ads in this paper read a little bit like the ask-bid columns of the New York Stock Exchange. Potential partners seek to strike bargains which maximize their rewards in the exchange of assets” (Cameron, et al., 1977, p. 28). Consistent with these economic models, people with more social assets have been found to demand more of them in others (Cameron et al., 1977; Harrison & Saedd, 1977; Walster, Aronson, Abrahams & Rottman, 1966). The fact that dating and married couples tend to be closely matched on such characteristics is interpreted as a result of such exchange processes. Exchanging like assets (e.g., good looks for good looks) is, presumably, one easy rule of thumb for obtaining a fair exchange. Evolutionary models have tended to ignore an individual’s self-evaluation, although such models actually share some of the basic assumptions of economic theories (Cooper, 1987; Frank, 1991; Kenrick & Keefe, 1992; Kenrick & Trost, 1987). It is quite consistent with evolutionary models that self-evaluation should be important. Given the assumption that individuals compete for positions in dominance hierarchies to gain access to more desirable mates (e.g., Sadalla et al., 1987), individuals should be very aware of their position in those hierarchies when considering what they can expect in a mate. This would apply to females as well as to males because evolutionary theorists have argued that it makes adaptive sense for people of both sexes to attempt to find partners who are at or above their own level of social status and attractiveness (Sloman & Sloman, 1988). In a related line of reasoning, some personality theorists believe that the major dimensions of personality judgement are cognitive schemas that evolved in large part to assist us in appraising our position in our social group (Goldberg, 1981; Hogan, 1982). Interestingly, these same dimensions appear across widely divergent cultures, with “dominance versus submissiveness” always emerging as one of the two major dimensions (White, 1980). Based upon this reasoning, it could be predicted that a person’s self-evaluation would be highly predictive of the minimum criteria he or she would find acceptable in a partner. These considerations illustrate the congruence between social exchange models and evolutionary models of mate selection. Both suggest a system in which humans make social comparisons between self and same-sexed others, and form a concept of their own relative value or attractiveness. This self-concept in turn plays an active role in courtship decisions. Evolutionary models can be used to supplement social exchange models by suggesting a consideration of gender

100

DOUGLAS T. KENRICK

differences in the use of such self-evaluations at different levels of involvement. In general, women’s self-evaluations should be highly predictive of the criteria desired in a mate for any type of sexual relationship. Males, on the other hand, should show differential selectivity depending upon the level of investment in a sexual relationship. In choosing a partner for a more committed relationship, men’s self-appraisals should match their criteria for a partner fairly closely.

One-night Stand

Marriage

Males

Females SUBJECTS’ SEX

Fig. 8 . Correlation between self-rated status and minimum criteria for partner’s status (based on data from combined results for Studies 1 and 2; Kenrick, Groth, Trost, & Sadalla, 1993).

EVOLUTIONARY SOCIAL PSYCHOLOGY

101

Given men’s comparatively lower potential investment in casual sexual liaisons, however, a man’s self-appraisals should be relatively less predictive of his minimum criteria for such relationships. In the most recent series of studies in the program of research on relationship development, we accordingly asked subjects to rate themselves. When those self-ratings were correlated with criteria for a partner, they indicated an interesting pattern in line with the notion of differential parental investment. Figure 8 depicts correlations between subjects’ self-rated social status and their minimum criteria for partner’s status. Both males’ and females’ criteria for status in a marriage partner correlate significantly with self-ratings-the higher a subject’s self-perceived status, the higher his or her criteria for a spouse. The same holds true for females’ ratings of partners for a one-night stand-the higher a woman’s self-perceived status, the higher her criteria for a one-night stand. However, males’ self-rated status is not significantly correlated with criteria for a one-night stand. These findings strongly support the importance of “equity,” but they suggest that the application of equity principles is related to evolved sex differences. When shopping for a onenight stand, males do not appear to drive much of a bargain. Note that this integrated perspective does not pit the social psychological and evolutionary approaches against one another, or argue that one is wrong. Without the social psychological perspectives on self-appraisal and changes over relationship course, several findings would remain either invisible or anomalous to those adopting an evolutionary perspective. Likewise, without the evolutionary approach’s ultimate perspective on gender differences in reproductive strategies, a number of findings on gender differences would remain either invisible or anomolous to someone who focused only on proximate models of social exchange. By integrating the two perspectives, we gain a more articulated understanding of proximate processes and of their ultimate significance.

IV. Evolutionary Social Cognition Some of the studies I have described thus far in this article, including observational research done across different species and archival research done across different cultures, might seem to be a bit outside the mainstream of “experimental social psychology.” In fact, the accepted wisdom has been that evolutionary theory concerns itself with “ultimate” explanations of behavior, explanations that stress historically distant causes. Psychologists, as I have been informed by more than one experimental social psychologist, are more interested in “proximate” explanations that stress contemporaneous events. However, psychologists conducting half-hour experiments in the laboratory sometimes forget that proximate

102

DOUGLAS T. KENRICK

and ultimate levels of explanation cannot be separated. The general lesson is one we all encountered in our high school history courses when we learned the distinction between the immediate cause of World War I (the assassination of Archduke Francis Ferdinand) and the background causes (like militarism, nationalism, and multinational alliances). Imagine studying politics with the limitation that we could not consider anything other than this week’s news. Without a consideration of the ultimate context of Francis Ferdinand’s assassination, it would have been no more important than any of the other millions of murders committed during this century. Political scientists would of course never consider limiting themselves to the current week’s news. However, psychologists who study cognitive processes often limit themselves in a similar way. By failing to consider the evolutionary context of human behavior, such psychologists miss the significance of proximate process and structure. For instance, regardless of the task at hand, subjects in laboratory experiments on social perception spontaneously attend to the categories of gender, age, race, and attractiveness. Experts on social cognition frequently note this without any mention of the potential evolutionary significance of such categorizations (e.g., Brewer & Lui, 1989; Hastie & Park, 1986; Markus & Zajonc, 1985). That seems a bit like mentioning that an archduke named Francis Ferdinand was assassinated early in the century, without mentioning that the event may have been connected to the start of World War I. What is the point of studying social cognition without asking why people attend to certain things more than others, why people encode some things more easily than others, and why people remember some social events more easily than they remember others? One answer to the question about why social cognition researchers have paid relatively little attention to the content of cognition is that it has been assumed that the same general cognitive processes are used to think about different types of content (Markus & Zajonc, 1985). In adopting this assumption, social cognition researchers follow cognitive psychologists in the traditional experimental areas (e.g., Glass & Holyoak, 1986). Traditional information-processing approaches to the study of cognition have focused predominantly on “general processes” rather than specific content. Within cognitive psychology proper, the processes involved in reading printed words have received perhaps the most attention and are often considered prototypical. For instance, Glass and Holyoak (1986) note in the opening chapter of their popular textbook on cognition that reading printed words “calls into play virtually every aspect of the cognitive processes that we will be exploring in this book” (p. 15). A perusal of the numerous articles published over the years in the “Attitudes and Social Cognition” section of the Journal of Personality Q Social Psychology would indicate that social cognition researchers have taken this viewpoint quite literally: The “social” stimuli presented to subjects in social cognition experiments have fre-

EVOLUTIONARY SOCIAL PSYCHOLOGY

103

quently consisted entirely of printed words, and have involved paradigms imported directly from the cognitive psychology of word recognition, such as the work on lexical priming. Researchers who adopt an evolutionary perspective, on the other hand, have tended to argue against the existence of general all-purpose cognitive processes. Instead, they have argued that cognitive processes differ depending upon the specific content of the information being processed (e.g., Hansen & Hansen, 1988; Sherry & Schacter, 1987; Tooby & Cosmides, 1989). Sherry and Schacter (1987) reviewed extensive evidence suggesting that animals have qualitatively different memory systems to deal with different types of information. For example, a bird learns to sing the songs of its species during an early sensitive period, often before it is capable of performing the song adequately, and later practice may serve only to improve the fidelity of performance. By contrast, memory for food stores is frequently erased within a few days and is “re-programmed” repeatedly throughout life. Memory for sickening foods shows a completely different pattern, in which the animal may learn an aversion on one trial and store its memory permanently with little or no input from experience. Support for content-specific cognitive processing also comes from psychophysiological research with humans, which has indicated that specific patterns of brain damage may lead to very specific and limited cognitive deficits. For instance, patients with one pattern of bilateral damage to the occipital-temporal areas suffer from a disorder called prosopagnosia (Damasio, Damasio, & Van Hoesen, 1982). Individuals with this disorder are unable to recognize specific faces, even their own, although they can recognize friends and relatives from their voices, they can find emotional expressions in photographs, and they can even pick out features such as noses, lips, and eyes. They can also distinguish the age and sex of the people they see, but not their identity. More recent research with these same patients indicates that although they are unable to accurately provide a verbal identification of which of a series of faces are familiar and which are unfamiliar, they do demonstrate marked changes in skin conductance in response to familiar faces (Tranel & Damasio, 1985). These findings are illustrative of a large body of literature that questions the assumption of a monolithic process of “memory” or “cognition” that applies uniformly to all types of content. The alternative to the view of a monolithic cognitive processor is the view of the brain as organized into a confederation of relatively specialized modules that operate in parallel (e.g., Martindale, 1991). We are already accustomed to thinking of the brain as modular to the extent that we differentiate between auditory and visual areas of the cortex, and at a finer level, between areas such as Wernicke’s area (related to understanding speech) and Broca’s area (related to production of speech). Instead of considering the brain as a single organ, we can think of it as a group of interrelated organs, each somewhat specialized for particular tasks. From an evolutionary perspective, it makes sense to ask about

104

DOUGLAS T. KENRICK

the kinds of tasks that a human brain would have evolved to perform. Many of these are nonsocial-such as the control of heartbeat and breathing, or the judgment of depth. However, evolutionary theorists argue that the rapid development of the cortex went hand in hand with the solution of problems that were specifically social in nature: Our ancestors needed to survive within social groups, they needed to be aware of their position in the social hierarchies of those groups, and they needed to attract and keep mates, for instance (Lancaster, 1976). From this perspective, it is to be expected that, in addition to modules designed to process information about the direction of sounds, the color patterns on potentially dangerous insects, and the sweetness of fruits, the human brain also contains modules designed to solve problems in social living (Lumsden & Wilson, 1981; Tooby & Cosmides, 1989). Social psychologists are not necessarily interested in, or equipped to study, the neurophysiology underlying the operation of different brain areas. However, it is important to be aware of the functional consequences for social cognition. In particular, there are marked qualitative differences in the processing of the same sorts of problems when the stimuli do and do not involve people (DeSoto, Hamilton, & Taylor, 1985; Tooby & Cosmides, 1989). In the past we have tended to import paradigms from cognitive psychology, asking whether the same sorts of processes that apply to word recognition (such as priming) might also apply to the processing of social stimuli. From an evolutionary perspective, a more fruitful approach would be to ask which types of problem social situations our ancestor’s brains would have had to solve, and to look for evidence of qualitative and quantitative differences in the way stimuli related to those situations are processed. For instance, Tooby and Cosmides found that a normally difficult logical problem (called the Wason task) was solved quite easily by subjects when it was cast in terms of a problem of “looking for a cheater.” They reasoned that our ancestors would have better survived with brains programmed to dedicate space to the detection of social cheaters. As another example, Hansen and Hansen (1988) used a tachistoscope to present subjects with rapidly flashed arrays of facial photographs. In some cases, one of the faces included a discrepant expression. They found that, in comparison with other emotional expressions, angry faces seemed to “jump out” of the stimulus array; subjects noticed those angry expressions with shorter presentation times, and their detection was less influenced by the number of distractor faces. Hansen and Hansen argue that the very short latencies needed to detect angry faces suggest that they are detected “preperceptually.” They further suggest that such rapid detection would have served an obvious evolutionary purpose-rapidly detecting someone in a dangerous state of mind. The findings of Tooby and Cosmides (1989) and Hansen and Hansen (1988) are quite preliminary, and one can argue with their interpretations. Nevertheless, the point of these types of studies is to indicate that content is important, and

EVOLUTIONARY SOCIAL PSYCHOLOGY

105

that, in formulating our hypotheses about social cognition, it might be worthwhile considering which sorts of social situations would have been most crucial for our ancestors’ survival. Given that differential reproduction is the central task of natural selection, it might be especially fruitful to consider social cognition in light of human reproduction. As discussed above, a consideration of the gender differences in parental investment and sexual selection has already led to a number of hypotheses about human mate choice. My colleagues and I have begun to consider how those same considerations might lead to hypotheses about social cognition. I consider two preliminary studies, and then suggest a systematic program of cognitive studies that follows from this line of reasoning.

A. A STUDY OF DEFAULT SCHEMATA The concept of a schema has been an influential one in social cognition research (Fiske & Taylor, 1991). The term schema is used in several ways, but it can be usefully thought of as a cognitive structure that acts as a filter guiding the selection of information for attention, encoding, and retrieval from memory. Cognition researchers note that it would be impossible to attend to every aspect of every situation in which we find ourselves; schemata guide the selection of information in ways that are consistent with ongoing motives (Fiske & Neuberg, 1990). Most of the studies on schema activation have tried to manipulate artificially people’s information processing. Research on priming suggests that recent exposure to words related to a particular concept (e.g., “reckless” as opposed to “adventurous”) can alter the interpretation of new information about a person (e.g., a description of him as a sky diver) (Srull & Wyer, 1979; 1980). There has been little investigation of default schemata. What do people pay attention to and remember from social situations when they are operating on automatic and have not been specifically primed by the experimenter? An evolutionary theorist would assume that we automatically process for information that is relevant to survival and reproduction. Before even processing the cut of someone’s clothes or his or her regional accent, we should notice the person’s gender, his or her sexual attractiveness, and whether this person poses a potential threat or a mating opportunity. Given these considerations, it makes eminent sense that gender, attractiveness, age, and race are used as spontaneous categorizations in social cognition experiments (e.g., Brewer & Lui, 1989; Hastie & Park, 1986; Markus & Zajonc, 1985). Once again, it is possible to use an evolutionary perspective to make predictions about gender differences in such “default schemata.” In line with the differential parental investment hypothesis, which sees males as actively competing for the attention of choosier and more reticent females, males should be more likely to spontaneously process information about potential mates. Dengelegi and I obtained some preliminary data

106

DOUGLAS T.KENRICK

Same sex remembered Opposite sex remembered

Males

Females SUBJECTS’ SEX

Fig. 9. Judge’s rating of the physical attractiveness of photographs recalled by subjects who had just looked at a yearbook (based on data from Kenrick & Dengelegi, 1991).

regarding this type of “default processing” in a free-recall study (Kenrick & Dengelegi, 1991). We asked subjects to look through a yearbook under the pretense that they were to develop an idea of the different types of students who attended different schools. After they had examined the yearbook, we unexpectedly asked them if they could bring any particular face to mind, and then we asked them to locate that face. We also collected information on subjects’ relational status: single, dating steadily, or married. Uninformed judges later rated the attractiveness of the faces that were and were not spontaneously chosen by members of each sex. As depicted in Fig. 9, when a subject recalled someone of

EVOLUTIONARY SOCIAL PSYCHOLOGY

107

Males 8(

x

W

Females

0

7(

v)

W

k

0

6C 0 (3

5

J J

a

2

50

U

cn

+

2

m’ 3

v)

40

&

+ z

W

$

30

W

a

20

10

Unattached

Attached

SUBJECT’S STATUS Fig. 10. Percentage of subjects recalling a photograph of the opposite sex, as a function of subject’s sex and relationship status (based on data from Kenrick & Dengelegi, 1991).

the opposite sex, that person tended to be more physically attractive than when a subject remembered someone of the same sex. That finding itself is not inconsistent with the idea that people “automatically” attend to mating relevant people (or at least spontaneously recollect them). However, another aspect of the study is more interesting and fits with the differential parental investment hypothesis discussed above. As shown in Fig. 10, females

108

DOUGLAS T.KENRICK

were about equally likely to bring a male as a female to mind, whether they were attached or not. Males, on the other hand, were more likely to recall a female than a male. This tendency appears to be reduced when the males were attached, but even attached males were more likely to selectively recall an attractive member of the opposite sex. This finding suggests that the reproductive constraints on males and females, discussed earlier, may have implications at the level of “on-line” processing. When males and females are asked to spontaneously recall a face, males’ recollections indicate that they are more likely to have their schematic antennae tuned for a mating opportunity. In itself, this type of finding does not prove an evolutionary model over alternative perspectives. However, it does fit nicely into the nomological network of findings on mate selection discussed earlier. As I will discuss in more detail below, an evolutionary model raises a number of different questions about social information processing.

B. A STUDY OF CONTRAST EFFECTS AND ATTRACTIVENESS CRITERIA Another line of research also suggests possibly interesting differences in spontaneous schema use by males and females. This research began with a series of studies unconnected to an evolutionary model. However, an evolutionary perspective provided a possible explanation of some unexpected sex differences found in the data, and led to empirically testable hypotheses. In the original work, we found that subjects exposed to physically attractive people later rated average looking targets as less attractive (Kenrick & Gutierres, 1980). A similar effect was found when males were asked to rate their mates after exposure to attractive centerfold photographs (Kenrick, Gutierres, & Goldberg, 1989). A parallel effect was not found for women exposed to male centerfolds-these women did not show reliably lower attraction for their male partners (Kenrick, et al., 1989). At first this finding made no sense to us, and we suspected (incorrectly) that perhaps the male stimuli were not highly attractive to the women. However, as noted in the earlier discussion of differences in mate selection criteria, evolution-based studies of mate criteria have found that women’s judgments of partners may be, compared with men’s, based relatively less upon physical attractiveness and more upon social status and dominance. A consideration of this literature led us to conduct a study in which subjects were asked to rate their commitment to their current relationships after exposure to targets who varied not only in attractiveness but also in descriptions of their dominance (Kenrick, Neuberg, Zierk, & Krones, in press). Subjects were led to believe that they were evaluating a series of profiles for a campus dating service. They were exposed to a series of profiles that contained

109

EVOLUTIONARY SOCIAL PSYCHOLOGY

photos of opposite-sex individuals who were either physically attractive or average looking. The profiles also included personality information, including descriptions of the targets as either high or low in dominance, leadership, and influence over others. A given subject saw profiles that were all relatively low or relatively high in social dominance, and that were either all attractive looking or all average looking. The results again indicated that exposure to attractive males had little effect on females’ levels of commitment to their current mates, but exposure to attractive females did significantly affect males’ commitment. On the other hand, female’s judgments were influenced by the dominance of the men to whom they were exposed. Compared to females exposed to nondominant men, females exposed to more dominant men rated themselves as relatively less committed to the current relationship. The data also indicated an unexpected, but interesting, interaction for the males-the attractiveness of the targets they saw was more important when the women were also non-dominant. The effect of exposure to attractive women held only for nondominant women: Men reported relatively lowest commitment to their current relationships after exposure to a series of non-dominant beautiful women (see Fig. 11). The different patterns suggest that the information of an abundance of avail-

0Dominant

DOMINANCE OF PROFILES

Nondominant

FEMALE SUBJECTS

6.4i

MALE SUBJECTS 6.4

6.2

6.2

6.0 5.8

6.0

5

=i= 8 4

5.6

5.6

5.4

5.4

fi 2

5.2

5.2

t-

$g

‘6

5

5.8

5.0

5.0

4.8

4.8 Unattractive

Attractive

r

Unattractive

Attractive

ATTRACTIVENESS OF PHOTOS IN PROFILES Fig. 11. Commitment to current relationship as a function of subject sex and the dominance and physical attractiveness of opposite-sexed targets to which subjects were previously exposed (based on data from Kenrick, Neuberg, Zierk, & Krones, in press).

110

DOUGLAS T. KENRICK

able women who are good-looking and submissive may lead a man to be unsatisfied with his current relationship, whereas the information of an abundance of available men who are highly dominant has the same effect on a woman. Although we have not previously found that men are “turned off” by social dominance in women, we had not previously investigated the interaction between attractiveness and dominance. It may be that a woman who is both highly physically attractive and highly socially dominant is regarded as too potentially powerful for the average male. However, further research is needed to explore this interaction. In any case, these findings again suggest that men and women use different schemata to process members of the opposite sex and to establish a comparison level with which to compare their current partners. For women, dominance rather than physical attractiveness appears more relevant. By using an evolutionary heuristic, therefore, we were able to predict an aspect of ongoing cognitive response to the social environment, and to consider how such ongoing processes might fit with a larger body of literature on the social behavior of humans and other animals.

V. Combining Traditional Experimental and Evolutionary Paradigms to Examine Mediators of Gender Differences

The two preliminary lines of research described in the previous section suggest that males and females might use somewhat different cognitive schemas in approaching social situations, but those studies only begin to examine the extensive possibilities. In combination with the results of the research discussed earlier in the paper, those two studies suggest the potential fruitfulness of a systematic program of studies on evolutionary social cognition. In collaboration with my colleagues Richard Keefe and Edward Sadalla, we have begun to develop a program that would integrate the central features of the cognitive and evolutionary paradigms. A systematic integration of these two approaches could be used not only to develop hypotheses about which types of content will be processed differently by males and females but also to examine the underlying mechanisms, which have hitherto been specified in rather vague terms. At what levels of cognitive processing do males and females differ? We do not yet know the answer to the question, and it has a number of empirically testable possibilities. Several of these are listed below. 1. Diferential attention: We have found that men are more likely to spontaneously recall attractive members of the opposite sex. One possible explanation for this is that they pay more attention to them in the first place. To examine this

EVOLUTIONARY SOCIAL PSYCHOLOGY

111

possibility, it would be possible to modify traditional methods to examine possible evolutionary constraints. For instance, Rock and Gutman (198 1) exposed subjects to a series of slides, each of which had two oddly shaped geometric figures on it, one red and one green. Half the subjects were asked to rate the aesthetic qualities of the red forms, the other half were asked to judge the aesthetic qualities of the green forms. In a later recognition test, subjects did no better than chance in recognizing the unattended color. This sort of method could be used to test the attention-grabbing qualities of any number of social stimuli, some of which, unlike geometric forms, ought to impose themselves on attention even when instructions focused attention on different aspects of the stimulus. In line with the above discussion, we would expect that the two sexes might differ in the extent to which their attention was grabbed by distractors varying in gender, age, attractiveness, and dominance. Given our earlier discussion of differential polygamy thresholds for the two sexes, for instance, women in relationships might actually suppress attention to potentially attractive or dominant members of the opposite sex. 2. Habituation: A line of research beginning with the early Russian psychologists has established a clear picture of the pattern of habituation to a novel stimulus (Martindale, 1991).At first exposure there is what Pavlov (1927) called an orienting response. This consists of a momentary freezing of motor activity accompanied by dilation of pupils and a general change in EEG, along with a decrease in respiration and heart rate. After 10 or 15 repetitions, only a localized change in EEG is evident; and after 25 to 30 additional repetitions, the EEG response drops out as well. There might be interesting sex differences in the speed of habituation to various stimuli, and those would be expected to fit with the theory and research described earlier. It would also be possible to examine how different arousal states interact with these responses. From an evolutionary perspective, emotional and motivational states can be seen as meta-cognitive programs designed to facilitate survival and reproductive goals (Kenrick & Hogan, 1991; Plutchik, 1980; Scott, 1980). Again, since the reproductive goals of males and females differ somewhat, one might expect that the state of sexual arousal would have different effects on males and females. For females, acting on sexual attraction for a stranger has less potential genetic rewards than the same response has for a male. In addition, a female who is attached stands a much higher risk of physical harm from a jealous partner than does a male (Daly & Wilson, 1988b). Given these constraints, an attached female who is sexually aroused may be less likely to attend to, and more likely to quickly habituate to, an attractive stranger. On the other hand, her attention may be more drawn to her partner, in line with findings that females are more attracted to familiar partners (Dewsbury, 1981). For males, sexual arousal should have an equally facilitating effect for partners and for strangers. Thus, sexual arousal might lead males to be more distracted by, and to less quickly habituate to, attractive strangers of the opposite sex, whereas the reverse might be true for females.

112

DOUGLAS T. KENRICK

3. Priming: Several studies have suggested that even stimuli presented below thresholds for accurate recognition may influence the interpretation of later stimuli (e.g., Bargh & Pietromonaco, 1982). The same approach could be varied using facial stimuli as primes, and such variations might yield some interesting interactions involving subject sex. For instance, if an attractive face of an opposite-sex stranger is presented as a subliminal prime to a subject, males might more quickly recognize words associated with approach, interest, or sexual arousal, or they might judge a personality profile as more appealing. For females, on the other hand, exposure to an attractive opposite-sex stranger might prime semantic associations to discomfort or avoidance, or lead them to rate a person described in a personality profile as more threatening (analogous to the effects Devine, 1989, found for subliminally presented race-related stimuli). It is also possible that these sex differences may be moderated by commitment to an ongoing relationship (Johnson & Rusbult, 1989; Simpson, Gangestad, & Lerma, 1990). These tentative hypotheses regarding attention, habituation, and priming are simply examples of the kinds of variations in social cognition that an evolutionary perspective might suggest. It would be worth searching for analogous sex differences at every phase of processing, from sensation through perception, memory, and complex cognitive processing. Some of the existent findings in social psychology suggest that such a model would apply to more complex cognitive processes. Consider Abbey’s (1987) study of sex differences in attributional patterns. She found that males were more likely to interpret a stranger’s friendly behavior as flirtation than were females. Such differences in attribution are consistent with the differential reproductive goals of males and females. It is more genetically advantageous for a male to avoid “false negatives” than “false positives” if a stranger might be flirting, but the same is not true for a female. Because a female can carry only one offspring at a time, she has little to gain from additional partners beyond one. In particular, a female gains no particular advantage from pursuing an ambiguously committed partner. Because a male could potentially inseminate a number of females during the same time period, it is more advantageous for him to pursue any potential partner, even one who may not have a strong long-term interest in him. Given the importance of reproduction and the long evolutionary history of sex differences in reproductive strategies, it could be argued that there will be much more action in the social domain than in the areas where psychologists have traditionally looked for sex differences (e.g., general “intelligence,” mathematical abilities, spatial abilities). In most ways, it makes sense to assume that the male and female nervous systems are constructed to solve the same problems, and therefore ought to be the same. The genetic interests of both males and females are well served by showing altruism toward kin, protecting the self

EVOLUTIONARY SOCIAL PSYCHOLOGY

113

against predators, accurately discriminating friend from foe, and so on. However, there are sex differences in sensory detectors (e.g., Cain, 1982), in hormones that interact with the brain’s development (e.g., Ellis, 1986), and in the structure of the brain itself (e.g., Allen, Hines, Shryne, & Gorski, 1989; Kimura, 1983). From an evolutionary perspective, it is to be expected that those differences will be linked to differential natural selection and to differential sexual selection pressures on the two sexes. That is, the differences in male and female neurophysiology should be linked to differences in the behaviors that ancestral males and females were designed to perform to survive and to reproduce. Most of the attention to evolved sex differences within cognitive psychology has focused on different survival tasks (e.g., hunting versus gathering). However, the abundant evidence of cross-culturally robust differences in mating behavior, and the theoretical evidence amassed from cross-species comparisons, suggests that as many, if not more, of the differences between human males and females will be linked to sociosexual tasks as to nonsocial ones.

VI. Conclusion: Toward a Unified Science of Behavior In this article I have focused on differences in mating behavior and on consequent differences in cognition that might be expected based on those behavioral differences. However, there are current and potential applications of evolutionary theory to every realm of social behavior, including aggression, altruism, social development, and intergroup conflict and cooperation (e.g., Cunningham, 1981; Daly & Wilson, 1988a, 1988b; Draper & Belsky, 1990; Rushton, 1989). Several journals regularly include studies of adult human social behavior from an evolutionary perspective (e.g., Behavioral and Brain Sciences; Ethology and Sociobiology; Human Nature; Journal of Social und Biological Structures). Evolutionary models are not simply of academic interest but also have important implications for applied social problems. An evolutionary perspective can provide important insights into the causes of behaviors including child abuse (Daly, 1989; Lenington, 1981), jealousy (Buss, Larsen, Westen, & Semmelroth, 1992), homicide (Daly & Wilson, 1988a; Kenrick, Dantchik, & MacFarlane, 1983; Kenrick & Sheets, in press), rape (Thornhill & Thornhill, 1983), and sexual harassment in organizations (Studd & Gattiker, 1991). An evolutionary perspective thus provides a framework that can connect the disparate threads of social psychology. An examination of any textbook in social psychology reveals the extent to which the field has been dominated by “minitheories,” and it is disappointing that the mini-theories used in one chapter of such a text are left unconnected to the mini-theories in another chapter. To an evolutionary theorist, however, there are inherent connections between attribu-

114

DOUGLAS T. KENRICK

tions, person perceptions, aggressive behaviors, mating choices, altruistic acts, and intergroup conflicts. Given that Darwin’s scheme is built on individual differences, it should come as no surprise that the perspective also assumes inherent connections between social processes and personality structure (Buss, 1990; Gangestad & Simpson, 1990; Hogan, 1982; Kenrick, Montello, & MacFarlane, 1985; Simpson & Gangestad, 1991). Thus, instead of viewing the research findings reported in the three sections of the Journal ofPersonality and Social Psychology as descriptions of mutually exclusive domains, an evolutionary perspective elucidates the absurdity of an epistemological apartheid between researchers studying social cognition, interpersonal processes, and individual differences. Beyond these connections within the sociaUpersonality area, an evolutionary perspective can also place our work in the context of research in other areas of psychology. As new findings emerge on the structure and function of the human brain, experimental psychologists are increasingly placing their findings within an evolutionary context, asking how the physical and functional constraints of the brain are designed to solve adaptive problems. As I noted earlier, many evolutionary theorists believe that the human brain is designed largely to facilitate survival in social groups and, like all organic systems, to foster successful reproduction. Reproduction is, of course, an explicitly social task in all vertebrate species, and it is a particularly social task in species like ours, where two parents form long-term bonds. Thus, it is not only possible to connect social psychology to neurophysiology and cognitive processes, it is essential for physiological and cognitive psychologists that their work be considered within the context of social psychology. An evolutionary perspective also leads us across the arbitrary boundaries that separate the discipline of psychology from biology in the one direction and from sociology and anthropology in the other. The foremost pioneering figures in general psychology were Wilhelm Wundt and William James; their counterpart in social psychology was William McDougall. It should come as no surprise that, before turning to psychology, each of these three men was trained in biology, and each viewed Charles Darwin’s evolutionary theory as an essential foundation for the understanding of human nature. What should be surprising is that it has taken us most of the intervening century to begin seriously to build on that foundation.

Acknowledgments The author thanks Michael Cunningham, Steve Gangestad, Dan Montello, Steve Neuberg, Jeff Simpson, Melanie Trost, and Mark Zanna for their helpful comments on an earlier draft of this manuscript.

EVOLUTIONARY SOCIAL PSYCHOLOGY

115

References Abbey, A. (1987). Misperceptions of friendly behavior as sexual interest: A survey of naturally occumng incidents. Psychology of Women Quarterly, 11, 173- 194. Alexander, R. D. (1987). The biology of moral systems. New York: Aldine deGruyter. Allen, L. S . , Hines, M., Shryne, J. E., & Gorski, R. A . (1989). Two sexuallydimorphic cell groups in the human brain. Journal of Neuroscience, 9, 497. Alley, T. R. (1992). Perceived age, physical attractiveness, and sex differences in preferred mates’ ages. Behavioral and Bruin Sciences, I S , 92. Andersson, M. (1982, October 28). Female choice selects for extreme tail length in a widowbird. Nature, pp. 818-820. Antill, J. K. (1983). Sex-role complimentarity versus similarity in mamed couples. Journal of Personality and Social Psychology, 45, 145- 155. Bargh, J. A., & Pietromonaco, P. (1982). Automatic information processing and social perception: The influence of trait information presented outside of conscious awareness on impression formation. Journal of Personality and Social Psychology, 43, 437-449. Barkow, J., Cosmides, L., & Tooby, J. (1992). The adapted mind: Evolutionary psychology and the generation of culture. New York: Oxford University Press. Bolig, R., Stein, P. J., & McKenry, P. C. (1984). The self-advertisement approach to dating: malefemale differences. Family Relations, 33, 587-592. Brehm. S . S . (1985). Intimate relationships. New York: Random House. Brehm, S. S. (1992). Intimate relationships (2nd ed.). New York: McGraw-Hill. Brewer, M. B. & Lui, L. N. (1989). The primacy of age and sex in the structure of person categories. Social Cognition, 3 , 262-274. Broude, G. J. (1992). The May-September algorithm meets the 20th century actuarial table. Behavioral and Brain Sciences, I S , 94-95. Buss, D. M. (1989). Sex differences in human mate preference: Evolutionary hypothesis tested in 37 cultures. Behavioral and Brain Sciences, I2, 1-49. Buss, D. M. (1990). Toward a biologically informed psychology of personality. Journal of Personality, 58. 1-16. Buss, D. M. & Barnes, M. F. (1986). Preferences in human mate selection. Journal of Personality and Social Psychology. 50, 559-570. Buss, D. M., Larsen, R. J., Westen, D., & Semmelroth, J. (1992). Sex differences in jealousy: Evolution, physiology, and psychology. Psychological Science, 3 , 25 1-255. Byrne, D. (1971). The utrruction paradigm. New York: Academic Press. Cain, W. S . (1982). Odor identification by males and females: Predictions versus performance. Chemical Senses, 7 , 129-142. Cameron, C . , Oskamp, S . & Sparks, W. (1977). Courtship American style-newspaper ads. Family Coordinator, 26, 27-30. Clark, M. S . & Mills, J. (1979). Interpersonal attraction in exchange and communal relationships. Journal of Personality and Social Psychology. 37, 12-24. Clark, M. S. & Reis, H. T. (1988). Interpersonal processes in close relationships. Annuui Review of Psvchology, 39, 609-672. Clutton-Brock, T. H. (1984). Reproductive effort and terminal investment in iteroparous animals. American Naruralisr, 123, 212-229. Clutton-Brock, T. (1991, October). Lords of the lek. Natural History, pp. 34-41. Cooper, W. S. (1987). Decision theory as a branch of evolutionary theory: A biological derivation of the Savage axioms. Psychological Review, 94: 395-41 1.

116

DOUGLAS T. KENRICK

Critelli, J. W. & Waid, D. R. (1980). Physical attractiveness, romantic love, and equity restoration in dating relationships. Journal of Personality Assessment, 44: 624-629. Cunningham, M. R. (1981). Sociobiology as a supplementary paradigm for social psychological research. In L. Wheeler (Ed.), Review of personality and social psychology: Vol. 2 (pp. 69106). Beverly Hills, CA: Sage. Cunningham, M. R. (1986). Measuring the physical in physical attractiveness: Quasi-experiments on the sociobiology of female beauty. Journal of Personality and Social Psychology, 50, 925-935. Cunningham, M. R . , Barbee, A. P., & Pike, C. L. (1990). What do women want? Facialmetric assessment of multiple motives in the perception of male physical attractiveness. Journal of Personality and Social Psychology, 59, 61 -12. Daly, M. (1989). Parent-offspring conflict and violence in evolutionary perspective. In R. W. Bell and N. J. Bell (Eds.) Sociobiology and the Social Sciences. Texas Tech University Press. Daly, M., & Wilson, M. (1983). Sex, evolution, and behavior. (2nd ed.) Belmont, CA: Wadsworth. Daly, M., & Wilson, M. (1988a). Evolutionary social psychology and family homicide. Science, 242, 519-524. Daly, M., & Wilson, M. (1988b). Homicide. New York: Aldine deGruyter. Damasio, A. R., Damasio, H., & Van Hoesen, G . W. (1982). Prosopagnosia: Anatomic basis and behavioral mechanisms. Neurology, 32, 331-341. Darwin, C. (1958). The origin of species by natural selection or the presentation of favoured races in the struggle for life. New York: Mentor. (Original work published in 1859) Darwin, C. (1965). The expression of emotions in man andanimals. Chicago: University of Chicago Press. (Original work published in 1872) Dawkins, R. (1976). The selJish gene. Oxford: Oxford University Press. DeSoto, C. B., Hamilton, M. M., & Taylor, R. B. (1985). Words, people, and implicit personality theory. Social Cognition, 3 , 369-382. Deutsch, F. M . , Zalenski, C. M., & Clark, M. E. (1986). Is there a double standard of aging? Journal of Applied Social Psychology, 16: 711-775. Devine, P. G. (1989). Stereotypes and prejudice: Their automatic and controlled components. Journal of Personality and Social Psychology, 56, 5-18. Dewsbury, D. A. (1981). Effects of novelty on copulatory behavior: The Coolidge effect and related phenomena. Psychological Bulletin, 89, 464-482. Draper, P., & Belsky, J. (1990). Personality development in evolutionary perspective. Journal of Personality, 5 8 , 141. Eibl-Eibesfeldt, I. (1975). Ethology: The biology of behavior (2nd ed.). New York: Holt, Rinehart & Winston. Ellis, L. (1986). Evidence of neuroandrogenic etiology of sex roles from a combined analysis of human, nonhuman primate, and nonprimate mammalian studies. Personality and Individual Diyerences, 7 , 519-552. Ekman, P., & Friesen, W. V. (1971). Constants across cultures in the face and emotion. Journal of Personality and Social Psychology, 17, 124- 129. Ekman, P., Friesen, W. V., O’Sullivan, M., Chan, A,, Diacoyanni-Tarlatzis, I., Heider, K., Krause, R . , LeCompte, W. A., Pitcairn, T., Ricci-Bitti, P. E., Scherer, K., Tomita, M., & Tzavaras, A. (1981). Universals and cultural differences in the judgments of facial expressions of emotion. Journal of Personality and Social Psychology, 53, 7 12-117. Feingold, A. (1992). Gender differences in mate selection preferences: A test of the parental investment model. Psychological Bulletin, 112, 125-139. Fiske, S . T., & Neuberg, S. L. (1990). A continuum of impression formation, from category-based to individuating processes: Influences of information and motivation on attention and interpretation. Advances in Experimental Social Psychology, 23, 1-74.

EVOLUTIONARY SOCIAL PSYCHOLOGY

117

Fiske, S. T., & Taylor, S . E. (1991). Social cognition (2nd ed.). New York: McGraw-Hill. Frank, R. H. (1991). Economics. In M. Maxwell (Ed.) The sociobiological imagination. Albany, NY: SUNY Press. Frisch, R. E. (1988, March). Fatness and fertility. Scientijc American, pp. 88-95. Gangestad, S., & Simpson, J. A. (1990). Toward an evolutionary history of female sociosexual variation. Journal of Personality, 58, 69-96. Gardner, H. (1985). The minds new science: A history of the cognitive revolution. New York: Basic Books. Glass, A. L., & Holyoak, K. J. (1986). Cognition (2nd ed.). New York: Random House. Goldberg, L. R. (1981). Language and individual differences: The search for universals in personality lexicons. In L. Wheeler (Ed.), Personality and social psychology review: Vol. 2. Beverly Hills, CA: Sage. Gross, M. (1984). Sunfish, salmon, and the evolution of alternative reproductive strategies and tactics in fishes. In G. Potts & R. Wooton (Eds.), Fish reproduction: Strategies and tactics. New York: Academic Press. Hansen, C. H., & Hansen, R. D. (1988). Finding faces in the crowd: An anger superiority effect. Journal of Personaliry and Social Psychology, 54, 917-924. Harpending, H. (1992). Age differences between mates in southern African pastoralists. Behavioral and Brain Sciences, 15, 102-103. Harrison, A. A., & Saedd, L. (1977). Let's make a deal: Analysis of revelations and stipulations in lonely hearts advertisements. Journal of Personality and Social Psychology, 35: 257-264. Hastie, R., & Park, B. (1986). The relationship between memory and judgment depends on whether the judgment task is memory based or on-line. Psychological Review, 93, 258-268. Hatfield, E., Traupmann, J., Sprecher, S . , Utne, M., & Hay, J. (1985). Equity and intimate relations: Recent research. In W. Ickes (Ed.), Compatible and incompatible relationships. New York: Springer-Verlag. Hendrick, S. S. (1981). Self-disclosure and marital satisfaction. Journal of Personality and Social Psychology, 40: 1 150- 1 159. Hill, J. (1984). Prestige and reproductive success in man. Ethology and Sociobiology, 5, 77-95. Hinde, R. A. (1984). Why do the sexes behave differently in close relationships? Journal of Social and Personal Relationships, 1: 471-501. Hogan, R. (1982). A socioanalytic theory of personality. In M. Page (Ed.), Nebraska symposium on motivation (pp. 55-89). Lincoln, NE: University of Nebraska Press. Johnson, D. J.. & Rusbult, C. E. (1989). Resisting temptation: Devaluation of alternative partners as a means of maintaining commitment in close relationships. Journal of Personality and Social Psychology, 57, 967-980. Keating, C. F., Mazur, A., Segall, M. H. (1981). A cross-cultural exploration of physiognomic traits of dominance and happiness. Ethology and Sociobiology, 2 , 41-18. Kenrick, D. T. (1987). Gender, genes, and the social environment: A biosocial interactionist perspective. In P. Shaver & C. Hendrick (Eds.), Review of Personality and Social Psychology: Vol. 7. Newbury Park, CA: Sage. Kenrick, D. T., Dantchik, A , , & MacFarlane, S. (1983). Personality, environment, and criminal behavior: An evolutionary perspective. In W. S. Laufer & J. M. Day (Eds.), Personaliry theory, moral development and criminal behavior. Lexington, MA: D. C. Heath. Kenrick, D. T., & Dengelegi, L. (1991). Gender diferences in spontaneous recall of faces. Unpublished manuscript, Arizona State University, Tempe. Kenrick, D. T., Engstrom, C., Cornelius, C., & Keefe, R. C. (1992). Age preferences for dating partners among teenagers: Support for an evolutionary model. Unpublished manuscript, Arizona State University, Tempe.

I18

DOUGLAS T. KENRICK

Kenrick, D. T., Groth, G., Trost, M. R., & Sadalla, E. K. 1993. Integrating evolutionary and social exchange perspectives on relationships: Effects of gender, self-appraisal, and involvement level on mate selection. Journal of Personality and Social Psychology, 64, 951-969. Kenrick, D. T., & Gutierres, S. E. (1980). Contrast effects and judgments of physical attractiveness: When beauty becomes a social problem. Journal of Personality and Social Psychology, 38, 131140.

Kenrick, D. T., Gutierres, S. E., & Goldberg, L. (1989). Influence of popular erotica on judgments of strangers and mates. Journal of Experimental Social Psychology, 25, 159- 167. Kenrick, D. T., & Hogan, R. (1991). Cognitive psychology. In M. Maxwell (Ed.), The sociobiological imagination. Albany, NY: SUNY Press. Kenrick, D. T., & Johnson, G. A. (1979). Interpersonal attraction in aversive environments: A problem for the classical conditioning paradigm? Journal of Personalify and Social Psychology. 37, 572-579. Kenrick, D. T., & Keefe, R. C. (1992). Age preferences in mates reflect-sex differences in human reproductive strategies. Behavioral and Brain Sciences, I S , 75- 133. Kenrick, D. T., Montello, D., & MacFarlane, S. (1985). Personality: Social learning, social cognition, or sociobiology? In R. Hogan & W. Jones (Eds.), Perspectives in Personaliry (Vol. 1). Greenwich, CT: JAI Press. Kenrick, D. T., Neuberg, S. L., Zierk, K., & Krones, J. (in press). Evolution and social coghition: Contrast effects as a function of sex, dominance, and physical attractiveness. Personalify and Social Psychology Bulletin. Kenrick, D. T., Sadalla, E. K., Groth, G . , & Trost, M. R. (1990). Evolution, traits, and the stages of human courtship: Qualifying the parental involvement model. Journai of Personakty, 58, 97117. Kenrick, D. T., & Sheets, V. L. (in press). Homicidal fantasies. Ethology and Sociobiology. Kenrick, D. T., & Trost, M. R. (1987). A biosocial model of heterosexual relationships. In K. Kelly (Ed.), Males, females, and sexuality: Research and theory. Albany, NY: SUNY Press. Kenrick, D. T., & Trost, M. R. (1989). A reproductive exchange model of heterosexual relationships: Putting proximate economics in ultimate perspective. In C. Hendrick (Ed.), Review of Personality and Social Psychology: Vol. 10, Close relationships. Newbury Park, CA: Sage Kenrick, D. T., & Trost, M. R. (in press). The evolutionary model of gender differences. In R. J. Stemberg & A. E. Beall (Eds.), Perspectives on the Psychology of Gender. NY: Guilford Press. Kimura, D. (1983). Sex differences in cerebral organization for speech and praxic functions. Canadian Journal of Psychology. 37, 19-35, Lagerspetz, K. (1979). Modification of aggressiveness in mice. In S. Feshbach & A. Fraczek (Eds.), Aggression and behavior change: Biological and social processes. New York: Praeger. Lakoff, G. (1987). Women, fire, and dangerous things: Whur categories reveal about the mind. Chicago: University of Chicago Press. Lancaster, J. B. (1976). Primate behavior and the emergence of human culture. New York: Holt, Rinehart, & Winston. Lancaster, J. B. (1985). Evolutionary perspectives on sex differences in the higher primates. In A. S. Rossi (Ed.), Gender and the life course. New York: Aldine. Lancaster, I. B., & King, B. (1985). An evolutionary perspective on menopause. In J. Brown & V. Kerns (Eds.), Her prime sev. Westport, CT: Bergin & Garvey. Lenington, S. (1981). Child abuse: The limits of sociobiology. Ethologyand sociobiology, 2, 17-29. Leonard, J. L. (1989). Homo sapiens: A good fit to theory, but posing some enigmas. Behavioral and Brain Sciences, 12: 26-27. Levinger, G. (1974). A three-level approach to attraction: Toward an understanding of pair relatedness. In T. L. Huston (Ed.), Foundations of Interpersonal Artracrion. New York: Academic Press.

EVOLUTIONARY SOCIAL PSYCHOLOGY

119

Levinger, G. (1983). Development and change. In H. H. Kelley, E. Berscheid, A. Christensen, J. H. Harvey, T. L. Huston, G. Levinger, E. McClintock, L. A. Peplau, & D. R. Peterson (Eds.), Close relationships (pp. 315-359). New York: Freeman. Lorenz, K. Z. (1963). On aggression. New York: Harcourt, Brace and World. Lumsden, C. I., & Wilson, E. 0. (1981). Genes, mind, and culture: The coevolutionary process. Cambridge, MA: Harvard. Markus, H., & Zajonc, R. B. (1985). The cognitive perspective in social psychology. In G. Lindzey & E. Aronson (Eds.), Handbook of Social Psychology, (Vol. I , pp. 137-230). New York: Random House. Martindale, C. (1991). Cognitive psychology: A neural-network approach. Pacific Grove, CA: BrookslCole. Mazur, A,, Booth, A,, & Dabbs, J. M., Jr. (1992). Testosterone and chess competition. Social Psychology Quarterly. 5 5 , 70-77. McDougall, W. (1908). Social psychology: An introduction. London, Methuen. Mealey, L. (1985). The relationship between social status and biological success: A case study of the Mormon religious hierarchy. Ethology and Sociobiology, 6 , 249-257. Murstein, B. (1970). Stimulus-value-role: A theory of marital choice. JournaI of Marriage and the Family, 32, 465-481. Nieschlag, E., & Michel, E. (1986). Reproductive functions in grandfathers. In L. Mastroianni, Jr. & C. A. Paulsen (Eds.), Aging, reproduction, and the climacteric (pp. 59-71). New York, NY: Plenum Press. Pavlov, I. (1927). Conditional reflexes. London: Oxford University Press. Plomin, R., DeFries, J. C., & McClearn, G. E. (1990). Behavioral genetics: A primer (2nd ed.). New York, NY: W. H. Freeman. Plutchik, R. (1980). A general psychoevolutionary theory of emotion. In R. Plutchik & H. Kellerman, (Eds.), Emotion: Theory, research, and experience: Vol. 1 . New York: Academic Press. Presser, H. B. (1975). Age differences between spouses: Trends, patterns, and social implications. American Behavioral Scientist. 19, 190-205. Ralls, K. (1976). Mammals in which females are larger than males. Quarterly Review ofBiology, 5 1 , 245-276.

Rajecki, D. W., Bledsoe, S. B., & Rasmussen, J. L. (1991). Successful personal ads: Gender differences and similarities in offers, stipulations, and outcomes. Basic and Applied Psychology, 12, 457-469. Rock, I . , & Gutman, D. (1981). The effect of inattention on form perception. Journal of Experimental Psychology: Human Perception and Performance, 7 , 275-285. Rosenblatt, P. C. (1974). Cross-cultural perspective on attraction. In T. L. Huston (Ed.), Foundations of interpersonal attraction. New York: Academic Press. Rusbult, C. E. (1980). Commitment and satisfaction in romantic associations: A test of the investment model. Journal of Experimental Social Psychology, 16, 172- 186. Rushton, I. P. (1985). Differential K theory: The sociobiology of individual and group differences. Personality and Individual Diyerences, 6, 44 1-452. Rushton, I. P. (1989)rrGenetic similarity, human altruism, and group selection. Behavioral andBrain Sciences, 12. 503-559. Sadalla, E. K., & Fausal, M. (1980). Dominance and heterosexual attraction: A field study. Unpublished manuscript, Arizona State University, Tempe. Sadalla, E. K., Kenrick, D. T., & Vershure, B. (1987). Dominance and heterosexual attraction. Journal of Personality and Social Psychology, 52, 730-738. Salzano, F. M., Neel, J. V., & Maybury-Lewis, D. (1967). Further studies on the Xavante Indians. I. Demographic data on two additional villages: Genetic structure of the tribe.American Journal of Human Genetics. 19, 463-489.

120

DOUGLAS T. KENRICK

Scott, J. P. (1980). The function of emotions in behavioral systems: A systems theory analysis. In R. Plutchik & H. Kellerman (Eds.), Emotion: Theory, research, and experience. (Vol. I). New York: Academic Press. Sheny, D. F., & Shachter, D. L. (1987). The evolution of multiple memory systems. Psychological Review, 94, 439-454. Sidanius, J., Cling, B. J., & Pratto, F. (1991). Ranking and linking as a function of sex and gender role attitudes. Journal of Social Issues, 47, 131-150. Simpson, I. A,, & Gangestad, S. (1991). Individual differences in sociosexuality: Evidence for convergent and discriminant validity. Journal of Personality and Social Psychology, 60, 870883.

Simpson, J. A,, & Gangestad, S . W. (1992). Sociosexuality and romantic partner choice. Journal of Personality#60, 3 1-5 1. Simpson, I. A,, Gangestad, S . W., & Lerma, M. (1990). Perception of physical attractiveness: Mechanisms involved in the maintenance of romantic relationships. Journal of Personality and Social Psychology, 59, 1192-1201. Sloman, S . , & Sloman, L. (1988). Mate selection in the service of human evolution. Journal of Social & Biological Structures, 11: 457-468. Srull, T. K., & Wyer, R. S., Ir. (1979). The role of category accessibility in the interpretation of information about persons: Some determinants and implications. Journal of Personality and Social Psychology, 37, 1660- 1672. Srull, T. K., & Wyer, R. S., Jr. (1980). Category accessibility and social perception: Some implications for the study of person memory and interpersonal judgments. Journal of Personality and Social Psychology, 38, 841-846. Studd, M. V., & Gattiker, U. E. (1991). The evolutionary psychology of sexual harassment in organizations. Ethology and sociobiology, 12, 249-290. Symons, D. (1979). The evolution of human sexuality. New York Oxford University Press. Symons, D. (1992). What do men want? Behavioral and Brain Sciences, 15, 113-1 14. Thornhill, N. W., & Thornhill, P. A. A. (1992). The preferred age of a potential mate reflects evolved male sexual physiology. Behavioral and Brain Sciences, 15, 114-1 15. Thornhill, R., & Thornhill, N. W. (1983). Human rape: An evolutionary analysis. Ethology and Sociobiology, 4 , 137- 173. Tooby, J., & Cosmides, L. (1989). Evolutionary psychology and the generation of culture. Part I: Theoretical considerations. Ethology and Sociobiology, 10. 29-49. Townsend, J. M. (1992). Measuring the magnitude of sex differences. Behavioral and Brain Sciences, 15, 115-116. Townsend, J. M., & Levy, G. D. (1990). Effects of potential partner's physical attractiveness and socioeconomic status on sexuality and partner selection. Archives of Sexual Behavior, 19, 149164.

Tranel, D., & Damasio, A. R. (1985). Knowledge without awareness: An autonomic index of facial recognition by prosopagnosics. Science, 228, 1453- 1454. Trivers, R. L. (1972). Parental investment and sexual selection. In B. Campbell (Ed.), Sexual selection and the descent of man. Chicago: Aldine. Trivers, R. L. (1985). Social evolution. Menlo Park, CA: BenjaminlCummings. Turke. P. W., & Betzig, L. L. (1985). Those who can do: Wealth, status, and reproductive success on Ifaluk. Ethology and Sociobiology. 6 , 79-87. Van den Berghe, P. L. (1992). Wanting and getting ain't the same. Behavioral and Brain Sciences, 15. 116-117.

Walster, E., Aronson, V., Abraham, D., & Rottmann, L. (1966). Importance of physical attractiveness in dating behavior. Journal of Personality and Social Psychology. 4 , 508-516.

EVOLUTIONARY SOCIAL PSYCHOWGY

121

Warner, R. R. (1984). Mating behavior and hermaphroditism in coral reef fishes. American Scienrist, 72. 128-134. White, G . M. (1980). Conceptual universals in interpersonal language. American Anthropologist. 82, 759-781. Wilson, E. 0. (1975). Sociobiology: The new synrhesis. Cambridge, MA: Harvard University Press. Wilson, E. 0. (1978). On human nature. Cambridge, MA: Harvard University Press. Wilson, M . , & Daly, M. (1985). Competitiveness, risk taking, and violence: The young male syndrome. Efhology and Sociobiology, 6, 59-73.

This Page Intentionally Left Blank

JUDGMENT IN A SOCIAL CONTEXT: BIASES, SHORTCOMINGS, AND THE LOGIC OF CONVERSATION

Norbert Schwarz

I. Introduction The mainstream of social cognition research can be described as the application of principles and methods of cognitive psychology to social stimuli. Although the information-processingparadigm to which social cognition research is committed (Ostrom, 1984), stimulated an enormous research productivity in social psychology, it has frequently been criticized as being asocial in nature. In the eyes of many critics, its concentration on individuals as isolated information processors fostered a neglect of the social context in which human judgment occurs, prompting Schneider (1991, p. 553) to ask, “Where, oh where, is the social in social cognition?’ As Forgas (1981, p. 3) observed, following the adoption of the information-processingparadigm, “Social psychology found itself transformed into a field now mainly concerned not with human social action, but with human beings as thinkers and information processors about social stimuli.” However, even the study of “human beings as thinkers and information processors” is likely to suffer from this neglect. On close inspection, it seems that much of what we consider to reflect biases in human judgment, artifacts in attitude measurement, and so on, may actually reflect researchers’ ignorance regarding the conversational context of human judgment, rather than serious shortcomings on the side of our subjects. Accordingly, social cognition research may greatly benefit from a fuller consideration of the social context in which humans conduct much of their thinking about social as well as non-social stimuli. A promising starting point for this endeavor is provided by psycholinguistic work into the tacit assumptions that govern the conduct of conversation in everyday life. The current article reviews these tacit assumptions and explores some of their implications for social cognition research. At present, social cognition reI23 ADVANCES IN EXPERIMENTAL SOCIAL PSYCHOIDGY. VOL 26

Copyright 0 1994 by Academic Press, Inc. All rights of reproduction in any form reserved.

124

NORBERT SCHWARZ

searchers have primarily paid attention to conversational aspects of human judgment in exploring attribution processes (e.g., Edwards & Potter, 1993; Hilton & Slugoski, 1986) and the impact of different audiences on the encoding and recall of person information (e.g., Higgins, McCann, & Fondacaro, 1982; Higgins & Rholes, 1978). Given that excellent reviews of that work are available (see Hilton, 1990, 1991; McCann & Higgins, 1992), the current article focuses on the contribution of conversational processes to the emergence of biases and shortcomings in human judgment, drawing on research in judgment and decision making, attitude measurement, and questionnaire construction.

11. The Logic of Conversation

As Clark and Schober (1992, p. 15) noted, it is a “common misperception that language use has primarily to do with words and what they mean. It doesn’t. It has primarily to do with people and what they mean. It is essentially about speakers’ intentions. Determining a speaker’s intention, however, requires extensive inferences on the part of listeners. Similarly, designing an utterance to be understood by a given listener requires extensive inferences on the side of the speaker. In making these inferences, speakers and listeners rely on a set of tacit assumptions that govern the conduct of conversation in everyday life. In their most widely known form, these assumptions have been expressed as four maxims by Paul Grice (1975), a philosopher of language. Subsequent researchers have elaborated on these assumptions, specifying their implications for speakers and listeners (see Clark 8z Brennan, 1991; Clark & Schober, 1992; Higgins, 1981; Higgins, Fondacaro, & McCann, 1982; Levinson, 1983). ”

A. THE LOGIC OF CONVERSATION IN EVERYDAY LIFE A maxim of manner asks speakers to make their contribution so that it can be understood by their audience. To do so, speakers not only need to avoid ambiguity and wordiness but have to take the characteristics of their audience into account, designing their utterances so that the audience can figure out what they mean-and speakers are reasonably good at doing so (Krauss & Fussel, 1991). At the heart of this process are speakers’ assumptions about the information that they share with recipients, that is, the common ground (Schiffer, 1972, Stalnaker, 1978). Listeners, in turn, assume that the speaker observes this maxim and

LOGIC OF CONVERSATION

125

interpret the speaker’s utterance against what they assume to constitute the common ground (e.g., Clark, Schreuder, & Buttrick, 1983; Fussel & Krauss, 1989a, 1989b). Whereas the initial assumptions about the common ground are based on the participants’ assumptions about their cultural and personal background, each successful contribution to the conversation extends the common ground of the participants, reflecting that “in orderly discourse, common ground is cumulative” (Clark & Schober, 1992, p. 19). This cumulative nature of the common ground reflects, in part, the operation of a maxim ofrelation that enjoins speakers to make all contributions relevant to the aims of the ongoing conversation. This maxim entitles listeners to use the context of an utterance to disambiguate its meaning by making bridging inferences (Clark, 1977). Moreover, this maxim implies that speakers are unlikely to assume that a contribution to a conversation is irrelevant to its goal, unless it is marked as such. As Sperber and Wilson (1986, p. vi) noted, “Communicated information comes with a guarantee of relevance,” and if in doubt, it is the listener’s task to determine the intended meaning of the utterance by referring to the common ground or by asking for clarification. In addition, a maxim of quantity requires speakers to make their contribution as informative as is required, but not more informative than is required. That is, speakers should respect the established, or assumed, common ground by providing the information that recipients need, without reiterating information that recipients already have (Clark & Haviland, 1977; Prince, 1981).Finally, a maxim ofquality enjoins speakers not to say anything that they believe to be false or for which they lack adequate evidence. Table I, adapted from McCann and Higgins (1992), summarizes the implications of these maxims in the form of “rules” that speakers and listeners are supposed to follow. These rules apply most directly to situations in which participants attempt to exchange information or to get things done. Obviously, conversations may be characterized by other goals (see Higgins, Fondacaro, & McCann, 1982), such as entertaining one another, in which case participants may not assume that the usual conversational maxims are observed. Given that the current article is concerned with conversational processes in research settings, however, the adjustments required by different conversational goals do not need further elaboration. In general, research participants are likely to perceive the research situation as a task-oriented setting in which participants attempt to exchange information as accurately as possible, thus rendering the assumptions underlying task-oriented conversations highly relevant. In summary, according to the tacit assumptions that govern the conduct of conversation in daily life, “Communicated information comes with a guarantee of relevance” (Sperber & Wilson, 1986, p. vi), and listeners are entitled to assume that the speaker tries to be informative, truthful, relevant, and clear.

126

NORBERT SCHWARZ TABLE I RULESOF THE COMMUNICATION GAME”

Communicators should: 1. take the recipient’s characteristics into account; 2. try to be understood (i. e., be coherent and comprehensible); 3. give neither too much nor too little information; 4. be relevant; 5 . produce a message that is appropriate to the context and the circumstances; 6. produce a message that is appropriate to their communicative intent or purpose; 7. convey the truth as they see it; 8. assume that the recipient is trying, as much as possible, to follow the rules of the communication game. Recipients should: 1. take the communicator’s characteristics into account; 2. determine the communicator’s communicative intent or purpose; 3. take the context and circumstances into account; 4. pay attention to the message and be prepared to receive it; 5 . try to understand the message; 6. provide feedback, when possible, to the communicator concerning their understanding of the message. OAdapted with permission, from McCann, C. D., & Higgins, E. T., “Personal and contextual factors in communication: A review of the ‘communication game’.’’ In G. R. Semin & K. Fiedler (Eds.), Language. interacrion, and social cognition (pp. 144-172). Newbury Park, CA: Sage. 0 1992.

Moreover, listeners interpret the speakers’ utterances “on the assumption that they are trying to live up to these ideals” (Clark & Clark, 1977, p. 122).’

B. THE LOGIC OF CONVERSATION IN RESEARCH SETTINGS What, then, are the implications of these tacit assumptions for communicative processes in research situations, most notably psychological experiments and survey interviews? As many researchers have noted (e.g., Clark & Schober, 1992; Strack, in press; Strack & Schwarz, 1992), “conversations” in research settings differ from natural conversations by being highly constrained. Whereas speakers and addresses collaborate in unconstrained natural conversations “to establish intended word meanings, intended interpretations of full utterances, implications of utterances, mutually recognized purposes, and many other such ‘For an introduction to other aspects of conversational conduct, including nonverbal behavior, see Chap. 2 of Argyle, 1992.

LOGIC OF CONVERSATION

127

things” (Clark & Schober, 1992, p. 25), their opportunity to do so is severely limited in research settings because of the researcher’s attempt to standardize the interaction. Most importantly, the standardization of instructions, or of the questions asked, precludes the tailoring of utterances to meet different common grounds. Moreover, when research participants ask for clarification, they may often not receive additional information. Rather, the previously given instructions may be repeated or a well-trained interviewer may respond, “Whatever it means to you,” when asked to clarify a question’s meaning. In some cases, as when a respondent is asked to complete a self-administered questionnaire, there may also be nobody who can be asked for clarification. As a result, a mutual negotiation of intended meaning is largely precluded in many research situations. Nevertheless, research participants will attempt to cooperate by determining the intended meaning of the researcher’s contributions to the constrained conversation. To do so, they will rely even more on the tacit assumptions that govern the conduct of conversation in daily life than they would under less constrained conditions. Moreover, these assumptions grant them every right to do so. That communicators are supposed to design their utterances so that they will be understood by addressees implies an interpretability presumption, as Clark and Schober (1992, p. 27) noted. This presumption is emphasized by the fact that the researcher as communicator obviously does not foresee any difficulties with the comprehensibility of his or her utterances, or else he or she would have taken appropriate precautions. As a result, research participants will refer to the conversational maxims in inferring the researcher’s intended meaning. Hence, they will assume that every contribution of the researcher is relevant to the aims of the ongoing conversation; that every contribution is informative, truthful, and clear; and they will refer to the context of the conversation to resolve any ambiguities that may arise. Unfortunately, however, research participants are bound to miss one crucial point: Whereas the researcher is likely to comply with conversational maxims in almost any conversation that he or she conducts outside of a research setting, the resexcher is much less likely to do so in the research setting itself. In fact, the researcher may violate each and every maxim of conversation by providing information that is neither relevant nor truthful, informative, and clear-and may have explicitly designed the situation to suggest otherwise. Research participants, however, have no reason to suspect that the researcher is not a cooperative communicator and are hence likely to work hard at finding meaning in the researcher’s contributions. The findings reviewed below suggest that this basic misunderstanding about the cooperative nature of communication in research settings has contributed to some of the more puzzling findings in social and psychological research and is, in part, responsible for the less than flattering picture of human judgmental abilities that has emerged from social cognition research.

128

NORBERT SCHWARZ

III. Communicated Information Comes with a Guarantee of Relevance One of the key assumptions underlying the conduct of conversation holds that all information contributed by participants is relevant to the goal of the ongoing conversation. As noted above, research participants have no reason to assume that this maxim of relevance does not hold in a formal research setting. Accordingly, they assume that all information offered to them by the researcher is relevant to their task-and they will work hard at making sense of it. This implicit guarantee of relevance contributes in important ways to several pervasive biases that have been documented in judgment research and underlies many apparent “artifacts” in attitude and opinion measurement that have captured the attention of survey methodologists.

A. THE CONVERSATIONAL RELEVANCE OF “IRRELEVANT” INFORMATION: IF THE EXPERIMENTER PRESENTS IT, I SHOULD USE IT

Social psychologists have long been intrigued by subjects’ readiness to rely on individuating information of little diagnostic value at the expense of more diagnostic information. Prominent examples of this general bias include the neglect of information about situational factors in explaining the behavior of an actor and the underutilization of base-rate information. As is true of most robust phenomena, both of these biases are likely to have many determinants, as Ross and Nisbett (1991) noted. Nevertheless, several studies indicate that the conversational guarantee of relevance contributes to a considerable degree to the size of the typically obtained effects.

I . The Underutilization of Base Rates Numerous studies have demonstrated a pronounced bias to rely on individuating information of little diagnostic value at the expense of more diagnostic baserate information (see Nisbett & Ross, 1980, for a review). Although the initial conclusion that individuating information will typically overwhelm the impact of base-rate information (e.g., Kahneman & Tversky, 1973; Nisbett & Borgida, 1975) has been called into question by subsequent studies (see Ginossar & Trope, 1987, for a review), the frequently observed underutilization of base-rate information has continued to be a key topic in judgment and decision research. An analysis of the experimental procedures used indicates, however, that the often

LOGIC OF CONVERSATION

129

dramatic findings are, in part, a function of conversational processes rather than of features naturally inherent to base-rate or individuating information. In what is probably the best-known demonstration of base-rate neglect, Kahneman and Tversky (1973) told their subjects that a target person described to them “shows no interest in political and social issues and spends most of his free time on his many hobbies, which include home carpentry, sailing, and mathematical puzzles.” These subjects predicted that the target person was most likely an engineer, independently of whether the base-rate probability for any target being an engineer was .30 or .70. An analysis of the instructions used in this study proves informative. Specifically, the instructions read (emphases added): A panel of psychologists have interviewed and administered personality tests to 30 (resp., 70) engineers and 70 (resp. 30) lawyers, all successful in their respective fields. On the basis of this information, thumbnail descriptions of the 30 engineers and 70 lawyers have been written. You will find on your forms five descriptions, chosen at random from the 100 available descriptions. For each description, please indicate your probability that the person described is an engineer, on a scale from 0 to 100. The same task has been performed by a panel of experts who were highly accurate in assigning probabilities to the various descriptions. You will be paid a bonus to the extent that your estimates come close to those of the expert panel. (p. 241)

The first part of the instructions informed subjects that the individuating information was compiled by psychologists on the basis of respected procedures of their profession, namely interviews and tests. Given that lay persons assume psychologists to be experts on issues of personality (rather than on base rates), this introduction emphasizes the relevance of the individuating information. Moreover, other experts-most likely psychologists as well, given the current context-are said to be highly accurate in making these judgments, thus further increasing the relevance of the individuating information. The subjects’ task is then defined as determining a probability that matches the judgments of the experts. If these experts are assumed to be psychologists, subjects can infer that the experimenter wants them to use the same information that these experts used-which is most likely the personality information compiled by their colleagues. Finally, as the experiment proceeds, subjects are asked to judge several target persons for whom different individuating information is presented. The base-rate information about the sample from which the targets are drawn, on the other hand, is held constant. This further suggests that the individuating information is of crucial importance because this information provides different clues for each judgment and in the absence of this information all tasks would have the same solution. Thus, the instructions and procedures of Kahneman and Tversky’s classic study allowed subjects to infer (however incorrectly) the experimenter’s intention that they should base their judgment on the individuating information.

130

NORBERT SCHWARZ

It therefore comes as little surprise that subjects relied on it when making their judgments. After all, they had no reason to assume that the experimenter violated each and every of the Gricean maxims by providing information that is neither relevant, truthful, informative, nor clear. To test this conversational analysis of the base-rate fallacy, Schwarz, Strack, Hilton, and Naderer (1991, Experiment 1) tried to undermine the guarantee of relevance that characterizes human communication in a modified partial replication of Kahneman and Tversky’s study. Some subjects were told that the person description was written by a psychologist, replicating the instructions used by Kahneman and Tversky. This entitles the recipient to assume that the presented information obeys the normative rules of communication and reflects a particular communicative intention on the part of the experimenter. Other subjects were told that the (identical) description was compiled by a computer that drew a random sample of descriptive sentences bearing on the target person. Obviously, the co-operative principle does not directly apply to the resulting communication, and the communicative intention cannot be unambiguously inferred. Although the data base from which the computer drew the sentences was said to have been compiled by psychologists, the collection drawn by the computer is of dubious relevance. As expected, undermining the implicit guarantee of relevance greatly attenuated subjects’ reliance on the individuating information. Specifically, subjects in the replication condition estimated the likelihood of the target being an engineer as .76, despite a low base rate of .30. However, when the same information was allegedly selected by a computer, their likelihood estimate dropped to .40. This attenuation indicates that subjects’ reliance on individuating information at the expense of base-rate information reflects, in part, their assumption that the experimenter is a cooperative communicator who does not present information that is irrelevant to the task at hand. Accordingly, they tried to find “relevance” in the information provided to them unless the implicit guarantee of relevance was called into question. In a similar vein, Krosnick, Li, and Lehman (1990) observed that the utilization of base-rate information varied as a function of the order in which the base rate and the individuating information were presented. In seven studies, using a variety of different problems, subjects were most likely to use base-rate information when this information was presented after, rather than before, the individuating information. On first glance, such a recency effect may suggest that the information presented last was more accessible in memory. However, recall data and other manipulations rendered this interpretation implausible (see Krosnick et al., 1990). Rather, the emergence of this recency effect could be traced to the operation of conversational conventions. As Krosnick et al. (1990) suggested, subjects who first receive base-rate information and who are subsequently provided with individuating information may reason,

LOGIC OF CONVERSATION

13 1

‘The first piece of information 1 was given (i.e., the base-rate) has clear implications for my judgment, so it was sufficient. A speaker should only give me additional information if it is highly relevant and informative, so the experimenter must believe that the individuating information should be given special weight in my judgment.” (p. 1141)

This reasoning would not only imply the general guarantee of relevance addressed above but would also reflect the conversational convention to present the more informative and important information later in an utterance, in part to direct listeners’ attention to it (see Clark, 1985, pp. 222-224). In turn, listeners may assume that “information presented later is more important and should be the focus of their attention” (Krosnick et al., 1990, p. 1141). If so, the individuating information may be rendered particularly relevant in the conversational context if presented after, rather than before, the base-rate information. Several findings support this assumption. In one of their studies, Krosnick et al. (1990, Experiment 4) observed that base-rate information had a more pronounced impact when the base rate was presented after, rather than before, the individuating information. However, this recency effect was largely eliminated when subjects were informed that the order in which both pieces of information were presented was determined at random. Thus, subjects were likely to rely on the information presented last, unless the informational value of presentation order was called into question. Moreover, this order effect emerged only when the base-rate information contradicted the implications of the individuating information. In this case, subjects gave more weight to whatever information was presented last, suggesting that the presentation order may carry information about the relative importance that the communicator wants to convey. To provide direct evidence for this assumption, Krosnick et al. (1990, Experiment 7) asked subjects to complete the blanks in a transcript of a conversation concerned with buying a car. In the base-rate last condition, the crucial part of this transcript read, My brother-in-law has had one problem after another with his Saab. _ _ a car magazine survey found Saabs have a better repair record than Volvos. Considering all this, I decided to buy a -. I think that is the better choice. (p. 1149)

In contrast, in the base-rate first condition the transcript read, A car magazine survey found Saabs have a better repair record than Volvos. -my brother-in-law has had one problem after another with his Saab. Considering all this, I decided to buy a -. I think that is the better choice. (p. 1150)

As expected on the basis of conversational conventions, most subjects completed the blanks in a way that implied that the speaker considered the second piece of information as more relevant than the first piece. Moreover, most of

132

NORBERT SCHWARZ

these subjects assumed that the speaker decided to buy the car favored by the second piece of information. In combination, the findings of the Schwarz, Strack, Hilton, and Naderer (1991) and Krosnick et al. (1990) studies indicate that subjects based their judgment primarily on the information that corresponded to the inferred communicative intention of the communicator. In the Schwarz et al. (1991) study, subjects were more likely to rely on the individuating information if it was presented by a human communicator whom they could expect to comply with conversational norms, rather than if it was drawn by a computer. Similarly, in the Krosnick et al. (1990) studies, subjects gave differential weight to base rate and to individuating information depending on its apparent importance to the communicator, as conveyed by the presentation order chosen. In both cases, it was not the nature of the presented information per se that determined its impact, but rather its perceived relevance in a given conversational context. The same theme is echoed in research on another well-known judgmental bias, namely the fundamental attribution error.

2 . The Fundamental Attribution Error Numerous studies in the domain of person perception have documented a pronounced readiness to account for an actor’s behavior in terms of his or her dispositions, even under conditions where the actor has responded to obvious situational pressures (see Jones, 1990; Ross & Nisbett, 1991, for reviews). Following a classic study by Jones and Harris (1967), this so-called “correspondence bias” (Jones, 1990) or “fundamental attribution error” (Ross, 1977) is typically investigated in a attitude-attribution paradigm. In most studies, subjects are provided an essay that advocates a certain position and are asked to infer the author’s attitude. Depending on experimental condition, they are informed either that the position taken in the essay was freely chosen by the author or that it was assigned by the experimenter. Whereas the content of the essay is presumably diagnostic for the author’s attitude under free-choice conditions, it is not under assignment conditions. Nevertheless, subjects typically attribute attitudes to the author that reflect the position taken in the essay, even under conditions where this position was assigned. Whereas findings of this type are usually interpreted as evidence for a pervasive “dispositional bias,” subjects seem quite aware that the essay is of little diagnostic value under no-choice conditions. For example, Miller, Schmidt, Meyer, and Colella (1984) observed that a majority of their subjects explicitly reported that the essay written under no-choice conditions did not provide useful information about the author’s true attitude. Nevertheless, the same subjects proceeded to make attitude attributions in line with the assigned position advo-

U X I C OF CONVERSATION

133

cated in the essay. As Wright and Wells (1988) suggested, a violation of conversational norms on the side of the experimenter seems to contribute to this finding. Specifically, Wright and Wells (1988, p. 184) noted that “the direction and content of the essay in the critical no-choice condition are irrelevant to the correct solution of the attribution task because the external constraints are sufficient to account for the essayist’s behavior.” The experimenter nevertheless provides subjects with an essay, thus violating the maxim of relevance. However, subjects have no reason to expect that this maxim is violated and are thus likely to assume “that the experimenter believes that the essay has some diagnostic value (otherwise, why were they given the essay?)” (Wright & Wells, 1988, p. 184). Accordingly, they take the essay into consideration in making attitude attributions, resulting in an apparent dispositional bias. To test this conversational account, Wright and Wells conducted an attitudeattribution study in which subjects were exposed to a pro or a con essay, allegedly written under choice or no-choice conditions. However, in addition to the standard procedure, their study involved a condition designed to undermine the implicit guarantee of relevance. Subjects in this condition were told that the information packages and questionnaire items being given to subjects . . . were being randomly selected from a larger pool of information packages and questions. Thus, their pool might not include sufficient information for them to answer some of their questions. Moreover, their information package might contain some information that was not germane to some of their questions. (Wright & Wells, 1988, p. 185)

As expected, this manipulation significantly reduced the emerging dispositional bias relative to the standard conditions in which subjects could assume that all the information provided to them is relevant to the task at hand. Moreover, the impact of undermining the guarantee of relevance was limited to the theoretically relevant no-choice conditions, and the above manipulation did not affect subjects’ inferences from essays written under free choice. Hence, undermining the guarantee of relevance did not result in generally more cautious judgments. Rather, it set subjects free to rely on the information that they themselves considered diagnostic, without having to find “relevance” in the information provided by the experimenter. Similar processes are likely to contribute to some extent to findings that suggest that the impact of highly diagnostic information is “diluted” by the addition of less diagnostic information (e.g., Nisbett, Zukier, & Lemley, 1981; see Tetlock & Boettger, 1993). To the extent that information presented by the experimenter comes with a guarantee of “relevance,” subjects are likely to consider it in forming a judgment-not because they would find the information per se utterly informative, but because the sheer fact that it was presented to them indicates that it is somehow “relevant.”

134

NORBERT SCHWARZ

3 . Conclusions

As the reviewed studies indicate, some of the biases that have received considerable attention in social cognition research may be less likely to reflect genuine shortcomings of the judgmental process than has typically been assumed. In fact, subjects often seem quite aware that the normatively irrelevant information is of little informational value. Nevertheless, they typically proceed to use it in making a judgment. As the above studies suggest, however, this may often reflect a violation of conversational norms by the experimenter, rather than any inherently flawed reasoning by subjects. Subjects have no reason to assume that the experimenter would intentionally provide information that is uninformative and irrelevant to the task at hand and thus violate the tacit rules that govern the conduct of conversation in everyday life. Accordingly, they proceed on the assumption that the experimenter is a cooperative communicator, and they work hard at making sense of the information provided to them. Once the implicit guarantee of relevance is called into question, however, the impact of normatively irrelevant information is largely reduced. This analysis suggests that the typical procedures used in social cognition research are likely to result in an overestimation of the size and the pervasiveness of judgmental biases. This analysis does not imply, however, that violations of conversational norms are the sole source of judgmental biases. Like most roust phenomena, judgmental biases are likely to have many determinants (see Ross & Nisbett, 1991; Kahneman & Lovallo, 1993). If we are to understand their operation in natural contexts, however, we need to ensure that their emergence in laboratory experiments does not reflect the operation of determinants that are unlikely to hold in other settings.

B. MAKING SENSE OF AMBIGUOUS QUESTIONS Whereas the preceding research examples pertained to the impact of explicit assertions of little informational value, subjects’ judgments have been found to be equally biased by the presuppositions conveyed by a researcher’s question. This misleading impact of questions has received considerable attention in research on public-opinion measurement and eyewitness testimony. As the studies reviewed in this section illustrate, however, the biasing effects of questions are again mediated by researchers’ violations of conversational norms and respondents’ erroneous assumption that the questioner is a cooperative communicator.

1 . Answering Questions about Fictitious issues Public-opinion researchers have long been concerned that the “fear of appearing uninformed” may induce “many respondents to conjure up opinions even

LOGIC OF CONVERSATION

135

when they had not given the particular issue any thought prior to the interview” (Erikson, Luttberg, & Tedin, 1988, p. 44). To explore how meaningful respondents’ answers are, survey researchers introduced questions about highly obscure or even completely fictitious issues, such as the “Agricultural Trade Act of 1978” (e.g., Bishop, Oldendick, & Tuchfarber, 1986; Schuman & Presser, 1981). Resumably, respondents’ willingness to report an opinion on a fictitious issue casts some doubt on the reports provided in survey interviews in general. In fact, about 30% to 50% of the respondents do typically provide an answer to issues that are invented by the researcher. This has been interpreted as evidence for the operation of social pressure that induces respondents to give answers that are presumably based on a “mental flip of coin” (Converse, 1964, 1970). Rather than providing a meaningful opinion, respondents are assumed to generate some random response, apparently confirming social scientists’ wildest nightmares. From a conversational point of view, however, these responses may be more meaningful than has typically been assumed in public opinion research. From this point of view, the sheer fact that a question about some issue is asked presupposes that this issue exists-or else asking a question about it would violate the norms of cooperative conduct. Respondents, however, have no reason to assume that the researcher would ask meaningless questions, and hence they try to make sense of it (see Strack & Martin, 1987, for a general discussion of respondents’ tasks). If the question is highly ambiguous, and the interviewer does not provide additional clarification, respondents are likely to turn to the context of the ambiguous question to determine its meaning, much as they would be expected to do in any other conversation. Once respondents have assigned a particular meaning to the issue, thus transforming the fictitious issue into a better defined issue that makes sense in the context of the interview, they may have no difficulty in reporting a subjectively meaningful opinion. Even if they have not given the particular issue much thought, they may easily identify the broader set of issues to which this particular one apparently belongs. If so, they can use their general attitude toward the broader set of issues to determine their attitude toward this particular one. A study by Strack, Schwarz, and Wanke (1991, Experiment 1) illustrates this point. In this study, German college students were asked to report their attitude toward an “educational contribution.” For some subjects, this target question was preceded by a question that asked them to estimate the average tuition fees that students have to pay at U.S. universities (in contrast to German universities, where university education is free). Others had to estimate the amount of money that the Swedish government pays every student as financial support. As expected, the attitude of students toward an “educational contribution” was more favorable when the preceding question referred to money that students receive from the government than when it referred to tuition fees. Subsequently, respondents were asked what the “educational contribution” implied. Content analyses

136

NORBERT SCHWARZ

of respondents’ definitions of the fictitious issue clearly demonstrated that respondents used the context of the “educational-contribution” question to determine its meaning. Thus, respondents turned to the content of related questions to determine the meaning of an ambiguous one. In doing so, they interpreted the ambiguous question in a way that made sense of it and subsequently provided a subjectively meaningful response to their definition of the question. Accordingly, it comes as no surprise that responses to fictitious issues do not conform to a model of mental coin flipping as Converse (1964, 1970) and other early researchers hypothesized, but do show a meaningful and systematic pattern, as Schuman and Kalton (1985) observed. What is at the heart of reported opinions about fictitious issues is not that respondents are willing to give subjectively meaningless answers, but that researchers violate conversational rules by asking meaningless questions in a context that suggests otherwise. But, much as has been observed in response to useless information presented in psychological experiments, survey respondents work hard at finding meaning in the questions asked. 2 . Leading Questions In a highly influential program of research, Loftus and collaborators (e.g., Loftus, 1975; see Loftus, 1979, for a review) demonstrated a pronounced impact of the presuppositions conveyed by leading questions on subjects’ memory. For example, subjects were shown a brief film clip and subsequently had to answer questions about what they saw. For some subjects, the questions included, “Did you see the children getting on the school bus?’ although no school bus was shown in the film. One week later, these subjects were more likely erroneously to remember having seen the school bus presupposed in the leading question than were subjects who were not exposed to the question (Loftus, 1975, Experiment 4). Findings of this type have typically been interpreted as indicating that “a presupposition of unknown truthfulness will likely be treated as fact, incorporated into memory, and subsequently ‘known’ to be true” (Dodd & Bradshaw, 1980, p. 695). Not surprisingly, such biasing effects of leading questions received considerable attention in applied research into eyewitness testimony. Several studies suggest, however, that the applied implications of this line of work may be more limited than has been assumed. In most experiments, the leading question is asked by the experimenter, and subjects have no reason to assume that the experimenter may lead them astray by knowingly introducing unwarranted presuppositions and violating conversational norms. In an actual courtroom setting, on the other hand, people may be quite aware that communicators may follow their own agenda, may be motivated to introduce misleading information, and may be all but cooperative. Hence, the impact of leading questions may be

LOGIC OF CONVERSATION

137

restricted to conditions under which the questioner is assumed to be a cooperative communicator. In line with this assumption, Dodd and Bradshaw (1980) observed biasing effects of leading questions about an observed car accident when the source of the question was the researcher, but not when the source was said to be the defendant’s lawyer (Experiment 1) or the driver of the car who caused the accident (Experiment 2). Thus, the otherwise obtained biasing effects of leading questions were “canceled by attributing the verbal material to a source that may be presumed to be biased” (Dodd & Bradshaw, 1980, p. 701), calling the source’s cooperativeness into question. Similarly, Smith and Ellsworth (1987) obtained a biasing effect of leading questions only when the questioner was assumed to be familiar with the event that the subject had witnessed. When the questioner was assumed to be unfamiliar with the event, the presupposition was discounted and no impact of the leading question was obtained. Whereas Loftus’s research program focused mainly on the impact of leading questions on reconstructive memory, other researchers explored the impact of leading questions on impression formation. Their findings reiterate the same theme. For example, in an exploration of incrimination through innuendo, Wegner, Wenzlaff, Kerker, and Beattie (198 1) observed that media questions of the type, “Is Jane using drugs?” may quickly become public answers. Again, recipients infer that there must be some evidence that triggered the question in the first place-or why else would someone raise it? Here, as well as in Loftus’s research program, the impact of the presupposition conveyed by the question rests on the implicit assumption that the communicator is cooperative, as a study by Swann, Giuliano, and Wegner (1982) illustrates. In their study, subjects observed how a questioner asked a respondent a leading question of the type, “What would you do to liven things up at a party?’ As expected, subjects considered the question to provide conjectural evidence that the person asked is an extrovert-unless they were told that the questions had been drawn from a fishbowl, thus undermining the implicit guarantee of relevance.* In combination, these studies indicate that theoretical accounts of the processes underlying the impact of leading questions have to take the assumed cooperativeness of the questioner into account. By themselves, the implications conveyed by the leading question are not sufficient to affect subjects’ judgments or recollections. Rather, subjects rely on the conveyed presuppositions only if they can assume that the speaker has access to the relevant knowledge and is a cooperative communicator who complies with the Gricean maxims. Only under those conditions can they expect the communicator to provide information that is informative, truthful, relevant, and clear. From this perspective, the robustness *Similar considerations apply to the inferences drawn from explicit assertions or denials (see Gruenfeld & Wyer, 1992; Harris & Monaco, 1978; Wegner et al., 1981).

138

NORBERT SCHWARZ

of leading question effects under laboratory conditions is not surprising: As the preceding sections of this article illustrated, subjects typically assume that the experimenter is a cooperative communicator and are hence likely to rely on the implications conveyed by the experimenter’s questions. Moreover, the experimenter is presumably a particularly knowledgeable source-after all, who would be more likely to know what was presented in the stimulus materials? By the same token, however, leading questions may provide less of a problem in natural settings, in which “there is often a basis to believe that the interrogator does not know the facts and is likely to have reasons to mislead” (Dodd & Bradshaw, 1980, p. 696).

C. FORMAL FEATURES OF QUESTIONNAIRES Whereas the preceding examples pertained to explicitly presented verbal material, research participants apply the tacit assumptions of daily conversations not only to the verbal content of questions. Rather, they also assume that any other aspect of the question-asking process is “relevant” to the task at hand, including formal features of the questionnaire. As a result, formal features of questionnaire construction, such as the specific numeric values presented as part of a rating scale or the range of response alternatives presented as part of a behavioral frequency question, may strongly influence the obtained responses, as the following examples illustrate (see Schwarz & Hippler, 1991, for a more extended review).

I . The Numeric Values of Rating Scales According to measurement theory, a 7-point rating scale is a 7-point rating scale, independent of how the seven points are graphically represented in the layout of the questionnaire. What psychologists care about is the wording of the question and the nature of the labels used to anchor the end points of the scale (see Dawes & Smith, 1985, for a review), but not the layout in which the scale is presented. For example, a 7-point scale that ranges from 1 to 7 should result in the same data pattern as a 7-point scale that ranges from -3 to +3, or a scale that presents seven unnumbered boxes, as long as the same verbal end-point labels are used. Empirically, however, the specific numerical values used may strongly affect the obtained responses, as Schwarz, Knauper, Hippler, Noelle-Neumann, and Clark (1991, Experiment 1) observed. As part of a larger survey, a representative sample of German adults was asked, “How successful would you say you have been in life?” This question was accompanied by an 11-point rating scale, with the end points labeled not at all successful and extremely successful. In one

139

LOGIC OF CONVERSATION TABLE I1 IMPACT OF NUMERIC SCALEVALUESON REFORTSALONG RATINGSCALES~ 0 to 10 Scale ( N = 480)

-5 to +5 Scale (N = 552)

Scale value

Percentage

Cumulative

6 7 8 9 10 Undecided

2 5 7 20 14 20 20 6 3 3

2 7 14 34 48 68 88 94 97 100

0 I 2 3 4

5

-

Scale value -5 -4 -3 -2 -I 0 +I +2 +3 +4 +5 Undecided

Percentage 1

1 1

I 9 9 23 35 14 4 2

Cumulative I 1

2 3 4 13 22 45 80 94 98 100

~ ~ ( 1 =0 )1 0 5 . 1 , ~< .0001

“Percentages rounded. Data based on a quota sample of 1032 German adults, randomly assigned to conditions. Adapted, with permission, from Schwarz, N., Knauper, B . , Hippler, H. .I.Noelle, Neumann, E . , & Clark, F. (1991). “Rating scales: Numeric values may change the meaning of scale labels.” Public Opinion Quarterly, 55, 570-582.

condition the numeric values of the rating scale ranged from 0 (not at all success-

ful) to 10 (extremely successful), whereas in the other condition they ranged from -5 (not at all successful) to $ 5 (extremely successful). The results showed a dramatic impact of the numeric values used, as shown in Table 11. Whereas 34% of the respondents endorsed a value between 0 and 5 on the 0 to 10 scale, only 13% endorsed one of the formally equivalent values between -5 and 0 on the -5 to + 5 scale. Coding both scales from 0 to 10, this pattern resulted in mean ratings of M = 6.4 on the 0 to 10, but M = 7.3 on the -5 to $ 5 version of the scale. In addition, an inspection of the distributions along both scales indicated that the responses were dislocated toward the high end of the - 5 to +5 scale, as compared to the 0 to 10 scale. This is also reflected in markedly different standard deviations (sds = 1.03 and .56 for the 0 to 10 and -5 to + 5 scale, respectively). Subsequent experiments (Schwarz, Knauper, Hippler, Noelle-Neuman, & Clark, 1991) indicated that the impact of numeric values is mediated by differential interpretations of the ambiguous end-point label not at all successful. When this label is combined with the numeric value 0, respondents interpret it to refer to the absence of noteworthy success. However, when the same label is com-

140

NORBERT SCHWARZ

bined with the numeric value - 5 , they interpret it to refer to the presence of explicit failure. This differential interpretation reflects that a minus-to-plus format emphasizes the bipolar nature of the dimension that the researcher has in mind, implying that one end-point label refers to the opposite of the other. Hence, not at all successful is interpreted as reflecting the opposite of success, that is, failure. In contrast, a rating scale format that presents only positive values suggests that the researcher has a unipolar dimension in mind. In that case, the scale values reflect different degrees of the presence of the crucial feature. Hence, not at all successful is now interpreted as reflecting the mere absence of noteworthy success, rather than the presence of failure. This differential interpretation of the same term as a function of its accompanying numeric value also affects the inferences that judges draw on the basis of a report given along a rating scale. For example, in a follow-up experiment (Schwarz, Knauper, Hippler, Noelle-Neumann, & Clark, 1991, Experiment 3), a fictitious student reported his academic success along one of the described scales, checking either a -4 or a formally equivalent 2. As expected, judges who were asked to estimate how often this student had failed an exam assumed that he failed twice as often when he checked a -4 as when he checked a 2, although both values are formally equivalent along the rating scales used. In combination, these findings illustrate that “even the most unambiguous words show a range of meaning, or a degree of ‘semantic flexibility’ . . . that is constrained by the particular context in which these words occur” (Woll, Weeks, Fraps, Pendergrass, & Vanderplas, 1980, p. 60).Assuming that all contributions to an ongoing conversation are relevant, respondents turn to the context of a word to disambiguate its meaning, much as they would be expected to do in daily life. In a research situation, however, the contributions of the researcher include apparently formal features of questionnaire design, rendering them an important source of information of which respondents make systematic use (see Bless, Strack, & Schwarz, 1993; Schwarz & Hippler, 1991, for more detailed discussions). Far from demonstrating superficial and meaningless responding, findings of this type indicate that respondents systematically exploit the information available to them in an attempt to understand their task and to provide a meaningful answer. At the same time, these findings emphasize that researchers must be sensitive to the informational implications of their research instrument to use them to their advantage. For example, the current findings suggest that rating scales that provide a continuum from negative to positive values may indicate that the researcher has a bipolar cbnceptualization of the respective dimension in mind, whereas scales that present only positive values may indicate a unipolar conceptualization. If so, the choice of numeric values may either facilitate or dilute the polarity implications of the end-point labels that are provided to respondents. Accordingly, researchers may be well advised to match the numeric values that

LOGIC OF CONVERSATION

141

they provide to respondents with the intended conceptualization of the underlying dimension as uni- or bipolar. 2 . Frequency Scales A related line of research explored the impact of response alternatives on behavioral frequency reports (see Schwarz, 1990; Schwarz & Hippler, 1987, for reviews). In survey research, respondents are typically asked to report the frequency with which they engage in a behavior by checking the appropriate value from a set of frequency-response alternatives provided to them. Again, the range of response alternatives may serve as a source of information for respondents. Specifically, respondents assume that the researcher constructed a meaningful scale that reflects his or her knowledge about the distribution of the behavior. Accordingly, values in the middle range of the scale are assumed to reflect the “average” or “typical” behavior, whereas the extremes of the scale are assumed to correspond to the extremes of the distribution. These assumptions influence respondents’ interpretation of the question, their behavioral reports, and related judgments. a . Question Interpretation. Suppose, for example, that respondents are asked to indicate how frequently they were “really irritated” recently. Before the respondent can give an answer, he or she must decide what the researcher means by “really irritated.” Does this refer to major irritations such as fights with one’s spouse, or does it refer to minor irritations such as having to wait for service in a restaurant? If the respondent has no opportunity to ask the interviewer for clarification, or if a well-trained interviewer responds, “Whatever you feel is really irritating,” he or she might pick up some pertinent information from the questionnaire. One such piece of information may be the frequency range provided by the scale. For example, respondents who are asked to report how often they are irritated on a scale ranging from “several times daily” to “less than once a week” may relate this frequency range to their general knowledge about the frequency of minor and major annoyances. Assuming that major annoyances are unlikely to occur “several times a day,” they are more likely to consider instances of less severe irritation to be the target of the question than are respondents who are presented a scale ranging from “several times a year” to “less than once every three months.” Experimental data support this assumption (Schwarz, Strack, Miiller, & Chassein, 1988). Respondents who reported their experiences on the former scale subsequently reported less extreme examples of annoying experiences than respondents who were given the latter scale. Thus, the type of annoying experiences that respondents reported was determined by the frequency range of the response alternatives in combination with respondents’ general knowledge, rather than by the wording of the question per se. Accordingly, the same

142

NORBERT SCHWARZ

question combined with different frequency scales is likely to assess different experiences. Theoretically, the impact of the response alternatives on respondents’ interpretation of the question should be more pronounced the less clearly the target behavior is defined. For this reason, questions about subjective experiences may be particularly sensitive to the impact of response alternatives because researchers usually refrain from providing a detailed definition of the target experience so as not to interfere with its subjective nature. Ironically, assessing the frequency of a behavior with precoded response alternatives may result in doing just what is avoided in the wording of the question. b. Frequency Estimates. Even if the behavior under investigation is reasonably well defined, however, the range of response alternatives may strongly affect respondents’ frequency estimates. This result reflects the fact that mundane behaviors of a high frequency, such as watching TV for example, are not represented in memory as distinct episodes (see Bradburn, Rips, & Shevell, 1987; Schwarz, 1990, for reviews). Rather, the various episodes blend together in a generic representation of the behavior that lacks temporal markers. Accordingly, respondents cannot recall the episodes to determine the frequency of the behavior but have to rely on an estimation strategy (see Menon, in press, for a more detailed discussion). In doing so, they may use the range of the scale presented to them as a frame of reference. This results in higher frequency estimates along scales that present high rather than low frequency-response alternatives. The results of a study on TV consumption, shown in Table 111, illustrate this effect (Schwarz, Hippler, Deutsch, & Strack, 1985, Experiment 1). In this study, 37.5%of a quota sample of German adults reported watching TV for 2.5 hr or more a day when presented the high-frequency response alternatives shown in Table 111, whereas only 16.2% reported doing so when presented the lowfrequency response alternatives. Not surprisingly, respondents’ reliance on the frame of reference suggested by the response alternatives increases as their knowledge about relevant episodes decreases (Schwarz & Bienias, 1990) or the complexity of the judgmental task increases (Bless, Bohner, Hild, & Schwarz, 1992). More importantly, however, the impact of response alternatives is completely eliminated when the informational value of the response alternatives is called into question. For example. telling respondents that they participate in a pretest designed to explore the adequacy of the response alternatives, or informing student subjects that the scale was taken from a survey of the elderly, wiped out the otherwise obtained impact of response alternatives (Schwarz & Hippler, unpublished data). Again, these findings illustrate that respondents assume the researcher to be a cooperative communicator whose contributions are relevant to the ongoing conversation unless the implicit guarantee or relevance is called into question. c. Comparative Judgments. In addition, the frequency range of the response

143

LOGIC OF CONVERSATION TABLE 111 CONSUMFTION AS A FUNCTION OF RESFWNSE ALTERNATIVES“

REFWRTED DAILY TV

Reported daily TV consumption (hr) Low frequency alternatives u p to ‘/2 ‘/2 to 1 1 to I ‘/2 1’/2 to 2 2 to 21/2 More than 2 ‘ / 2 High frequency alternatives u p to 21/2 2‘/2 to 3 3 to 3‘12 3’12 to 4 4 to 4’12 More than 4’/2

%

7.4 17.7

26.5 14.7 17.7

16.2 62.5 23.4

7.8 4.7 1.6 0.0

ON= 132. Adapted, with permission from Schwarz, N., Hippler, H. J., Deutsch, B . , & Strack, F. (1985). “Response scales: Effects of category range on reported behavior and comparative judgments.” Public Opinion Quarterly. 49, 388-395.

alternatives has been found to affect subsequent comparative judgments. Given the assumption that the scale reflects the distribution of the behavior, checking a response alternative is the same as locating one’s own position in the distribution. Accordingly, respondents extract comparison information from their own location on the response scale and use this information in making subsequent comparative judgments. For example, checking “2h” on the low-frequency scale shown in Table I11 implies that a respondent’s TV consumption is above average, whereas checking the same value on the high-frequency scale implies that his or her TV consumption is below average. As a result, respondents in the Schwarz et al. (1985) studies reported that TV plays a more important role in their leisure time (Experiment I), and described themselves as less satisfied with the variety of things they do in their leisure time (Experiment 2), when they had to report their TV consumption on the low- rather than on the high-frequency scale (see also Schwarz & Scheuring, 1988). Moreover, these frame-of-reference effects are not limited to respondents themselves, but influence the users of their reports as well. For example, in a study by Schwarz, Bless, Bohner, Harlacher, and Kellenbenz

144

NORBERT SCHWARZ

(1991, Experiment 2), experienced medical doctors concluded that having the same physical symptom twice a week reflected a more severe medical condition when “twice a week” was a high rather than a low response alternative on the symptoms checklist presented to them.

3 . Conclusions Findings of the type reviewed in this section are usually considered measurement “artifacts.” From a conversational point of view, however, they simply reflect that respondents bring the assumptions that govern the conduct of conversations in daily life to the research situation. Hence, they assume that every contribution is relevant to the goal of the ongoing conversation-and in a research situation, these contributions include apparently formal features of the questionnaire, such as the numeric values presented as part of a rating scale or the response alternatives presented as part of a frequency question. As a result, the scales used are all but “neutral” measurement devices. Rather, they constitute a source of information that respondents actively use in determining their task and in constructing a reasonable answer. Although research methodologists have traditionally focused on the information that is provided by the wording of the question, we do need to pay equal attention to the information that is conveyed by apparently formal features of the questionnaire.

IV. Making One’s Contribution Informative So far, the reviewed research illustrated how many apparent biases and shortcomings in human judgment or artifacts in opinion measurement may, in part, be traced to the implicit guarantee of relevance that characterizes human communication. However, the maxims of cooperative communication do not only determine recipients’ use of the information provided by speakers. Rather, they also determine what information the recipient of a question is expected to provide in turn. Specifically, cooperative speakers are supposed to provide information that is relevant to the goal of the conversation. This not only implies that the provided information should be substantively related to the topic of the conversation. Rather, it also implies that the provided information should be new to the recipient (Clark & Clark, 1977). Hence, the utterance should not reiterate information that the recipient already has, or may take for granted anyway. Accordingly, determining which information one should provide requires extensive inferences about the information that the recipient already has, in order to identify what is, or is not, “informative.” As was the case for information (inadvertently) provided by the researcher,

LOGlC OF CONVERSATION

145

conversational rules that govern the selection of information to be provided by research participants underlie many apparently surprising findings in social and psychological research, ranging from the impact of open versus closed question formats to children’s performance on Piagetian conservation tasks or the use and disuse of easily accessible information in making a judgment.

A. WHAT IS INFORMATIVE? THE IMPACT OF OPEN VERSUS CLOSED QUESTION FORMATS

To answer a question, speakers have to determine what information they are to provide. Suppose that you are asked to report what you have done today. Most likely, you would not include in your report that you took a shower, that you dressed, and so on. If these activities were included in a list of response alternatives, however, you would probably endorse them. This thought experiment reflects a set of standard findings from the survey methodology literature (see Schwarz & Hippler, 1991, for a review). Experimental studies on the impact of open- and closed-response formats have consistently demonstrated that open- and closed-responseformats yield considerable differences in the marginal distribution as well as the ranking of items (e.g., Bishop, Hippler, Schwarz, & Strack, 1988; Schuman & Presser, 1977). On the one hand, any given opinion is less likely to be volunteered in an open-response format than to be endorsed in a closed-response format, if presented. On the other hand, opinions that are omitted from the set of response alternatives in a closed format are unlikely to be reported at all, even if an “other” category is explicitly offered, which respondents in general rarely use (Bradburn, 1983; Molenaar, 1982). Several processes are likely to contribute to these findings. Most importantly, respondents are unlikely to report spontaneously, in an open-answer format, information that seems self-evident or irrelevant. In refraining from these responses, they follow the conversational maxim that an utterance should be informative, as discussed above. This results in an underreporting of presumably self-evident information that is eliminated by closed-response formats where the explicit presentation of the proper response alternative indicates the investigator’s interest in this information. Moreover, respondents may frequently be uncertain whether information that comes to mind does or does not belong to the domain of information the investigator is interested in. Again, closed-response formats may reduce this uncertainty, resulting in higher responses. Finally, a generic “other” response provides little information and would be considered inadequate as an answer in most conversations. Hence, it is rarely checked. In addition, the response alternatives may remind respondents of options that they may otherwise not have considered. The methodological literature has typ-

146

NORBERT SCHWARZ

ically focused on this latter possibility, implying that closed-response formats may suggest answers that respondents would never think of themselves. This assumption is to some degree supported by the observation that less-educated respondents are more likely to refuse to answer in an open-response format, and to provide an answer in a closed-responseformat, than are well-educated respondents (Schuman & Presser, 1981). However, a conversational analysis suggests that the obtained differences are more plausibly traced to the clarification of the questioner’s interest that is provided by a closed-response format. Most importantly, the assumption that respondents may lack the information required for an answer, and hence pick one from the response alternatives, may hold to some degree for complex knowledge questions, but does not hold for questions about daily activities, such as, “What have you done today?’ Nevertheless, the same differences are obtained for questions of this type, and they are most pronounced for activities that the questioner may take for granted anyway, such as taking a shower or having breakfast (see Schwarz, Hippler, and Noelle-Neumann, in press, for a more extended discussion).

B. REPEATED QUESTIONS AND CHANGING INTERPRETATIONS That speakers are supposed to provide new information rather than to reiterate information the recipient already has, has important implications for the interpretation of questions that are highly similar in content or even repeated literally. Unless there is reason to believe that the questioner did not understand the answer already given, the person asked is likely to interpret the second question as a request for new information. The studies reviewed in this and the following section illustrate these shifting interpretations and their impact on respondents’ reports. 1 . Experimenter’s Questions and Children’s Cognitive

Skills: The Piagetian Conservation Task

Much as researchers’ violations of conversational norms have contributed to overestimations of adults’ cognitive biases, they have also contributed to underestimations of children’s cognitive skills. Research on the conservation task introduced by Piaget (1952; Piaget & Inhelder, 1969) may serve as an example. In a typical study (e.g., McGarrigle & Donaldson, 1974), a child is shown two rows of objects, equal in number and aligned in one-to-one correspondence. When asked, “Is there more here or more here, or are both the same number?’ the child usually answers that both rows are the same in number. Next, the experimenter rearranges the objects in one of the rows to extend the row’s length.

LOGIC OF CONVERSATION

147

Following this transformation, the previously asked question is repeated. Many young children now respond that there are more objects in the longer row, suggesting that they did not master number conservation. Give that only the perceptual configuration of the crucial row has changed, explanations of this phenomenon have typically focused on children’s susceptibility to perceptual influences. However, a conversational analysis of the procedures used proves informative. Why would a speaker ask the same question twice within a very short time span, unless he or she inquired about some new aspect? Moreover, what would that new aspect most likely be, following the deliberate and intentional transformation performed by the questioner? As McGarrigle and Donaldson (1974, p. 347) noted, in early stages of language acquisition, children use the behavior of a speaker “to arrive at a notion of speaker’s meaning and this knowledge is utilized to make sense of the language around [them],” much as adults use the context of an utterance to disambiguate its meaning. From this perspective, it could be that the experimenter’s simple direct action of changing the length of the row leads the child to infer an intention on the experimenter’s part to talk about what he has just been doing. It is as if the experimenter refers behaviorally to length although he continues to talk about number. (McGarrigle & Donaldson, 1974, p. 343)

To test this assumption, the authors conducted a study in which they varied the apparent intentionality of the transformation. Whereas they replicated the above standard procedure in one condition, a “naughty teddy bear” appeared in the other and tried to “spoil the game” by rearranging the objects, increasing the length of one row. The results provided strong support for a conversational account: Whereas only 13 out of 80 children showed number conservation when the experimenter manipulated the length of the row, 50 out of the same 80 children showed number conservation when the change was due to the apparently unintended interference of “naughty teddy” (see Dockrell, Neilson, & Campbell, 1980; Light, Buckingham, & Robbins, 1979, for conceptual replications and Donaldson, 1982, for a review). These findings suggest that the children used the behavioral context of the question to infer the speaker’s meaning: When “naughty teddy” changed the arrangement of the objects, the experimenter might indeed want to know whether teddy took an object or whether the number remained the same. However, repeating the previously answered number question makes little sense when the experimenter changed the arrangement himself, leading children to infer that the experimenter apparently wants to talk about what he did, thus changing the reference of the question from number to length. That the changes in children’s interpretation of the question are indeed driven by conversationally inappropriate question reiteration has been nicely demonstrated by Rose and Blank (1974). In their study, children were again shown two

148

NORBERT SCHWARZ

rows of equal length with the same number of objects, but only some of the children had to make an initial judgment at this point. Subsequently, the experimenter changed the arrangement, increasing the length of one row. As usual, many of the children who had already given a judgment when the rows were equal now reported that the longer row had more objects. However, children who had not previously been asked were considerably more likely to respond that the number of objects in both rows was the same (see also Siegal, Waters, & Dunwiddy, 1988). Clearly, it is not children’s confusion of number and length per se that leads them astray. Rather, having already answered the question, the children assume that the experimenter must have something else in mind when the question is reiterated. Hence, they respond to their new interpretation of the reiterated question, unless asking for the same information twice makes sense, as was the case in McGarrigle and Donaldson’s (1974) study.

2 . Are Happiness and Satisfaction the Same Thing? Much as children have been found to change their interpretation of the same question if it is reiterated within a short time span, adults have been observed to change, or not to change, their interpretation of highly similar questions as a function of the conversational context in which the questions are posed. For example, Strack, Schwarz, and Wanke (1991, Experiment 2) asked German students to rate their happiness and satisfaction with life as a whole along 11point scales (1 1 = very happy or very satisfed, respectively). In one condition, both questions were presented at the end of the same questionnaire and were introduced by a joint lead-in that read, “Now, we have two questions about your life.” In the other condition, only the happiness question was presented at the end of the questionnaire, introduced by a parallel lead-in, “Now, we have a question about your life.” The subsequent rating of life satisfaction, however, was presented as the first question in a new and ostensibly unrelated questionnaire about characteristics of research participants, attributed to a different researcher and presented in a different graphic layout. How would these manipulations affect respondents’ interpretation of the two related concepts of happiness and satisfaction? In general, happiness and satisfaction are perceived as closely related concepts and both judgments have typically been found to be affected by the same variables in studies of subjective well-being (see Schwarz, 1987; Schwarz & Strack, 1991). However, when both questions are presented as part of the same conversational context, interpreting them as nearly identical in meaning would result in considerable redundancy. Hence, respondents may infer that the researcher intends both questions to tap different aspects of their subjective well-being and may, accordingly, draw on different information about their lives in making their judgments. Note, however, that this does not apply when the questions are asked by two different communi-

LOGIC OF CONVERSATION

149

cators. In this case, both communicators may simply use somewhat different words to refer to the same thing, and providing identical responses would not violate the norm of nonredundancy, given that each response is directed to a different recipient. As a result, the answers given to both questions should be more similar when the questions are asked by different researchers than when asked by the same researcher. Strack et al.5 (1991, Experiment 2) findings confirmed this expectation. When both questions were asked in ostensibly unrelated questionnaires, subjects’ mean reports of happiness (M = 8.0) and satisfaction (M = 8.2) did not differ, and both measures correlated r = .96. When both questions were presented as part of the same conversational context, however, subjects reported significantly higher happiness (M = 8.1) than satisfaction (M = 7.4), and the correlation between both measures dropped significantly to r = .75. Apparently, respondents inferred from the conversational relatedness of both questions that the researcher must have distinct concepts in mind because asking the same thing twice would make little sense. Accordingly, they presumably based their responses on different aspects of their lives under this condition, a process that is shown more clearly in the studies reviewed in the next section.

C. AVOIDING REDUNDANCY IN ANSWERING QUESTIONS OF DIFFERENTIAL GENERALITY: ITS IMPACT ON THE USE OF HIGHLY ACCESSIBLE INFORMATION

In many studies, respondents are asked to answer several related questions that may vary in generality. For example, they may be asked how satisfied they are with different specific domains of their lives, as well as how satisfied they are with their lives as a whole. Current theorizing in social cognition suggests that answering a specific question increases the accessibility of information used to answer it (see Bodenhausen & Wyer, 1987; Higgins, 1989; Martin & Clark, 1990, for reviews). Hence, this information should be most likely to come to mind when a related general question is asked later on. Depending on the nature of the conversational context, however, using the primed information in answering the question may violate the conversational norm of nonredundancy, as Strack and Martin (1987) pointed out, following related suggestions by Bradburn (1982) and Tourangeau (1984). As a result, conversational processes may determine whether easily accessible information is, or is not, used in making a judgment, thus determining the emergence of priming effects. As the studies reviewed in this section illustrate, the combined operation of cognitive accessibility and conversational norms may strongly influence the relationship ob-

150

NORBERT SCHWARZ

tained between two judgments, or between a behavioral report and a judgment, leading to dramatically different substantive conclusions in social research.

I . General and Specific Judgments: Conversational Norms and the Emergence of Assimilation and Contrast Efects To explore the relationship between related judgments of differential generality, Schwarz, Strack, and Mai (1991) asked respondents to report their marital satisfaction as well as their general life-satisfaction, varying the order in which both questions were asked. The first column of Table IV shows the resulting correlations between marital and life satisfaction. When the life-satisfaction question preceded the marital-satisfaction question, both measures were moderately correlated, r = .32. Reversing the question order, however, increased the correlation to r = .67. This reflects that answering the marital-satisfaction question first increased the accessibility of marriage-related information in memory. As a result, respondents were more likely to consider marriage-related information in evaluating their lives as a whole (see Schwarz & Strack, 1991, for a judgment model of subjective well-being). This interpretation is supported by a highly similar correlation of r = .61 when the general question explicitly asked respondents to include their marriage in evaluating their overall life satisfaction. In a third condition, however, Schwarz, Strack, and Mai (1991) deliberately evoked the conversational norm of nonredundancy. To do so, both questions were TABLE IV CORRELATION OF RELATIONSHIP SATISFACTION AND LIFE SATISFACTION AS A FUNCTION OF QUESTION ORDERAND CONVERSATIONAL CONTEXT‘ Number of specific questions Condition

One

Three

General-Specific Specific-General Specific-General, with joint lead-in Specific-General, explicit inclusion Specific-General, explicit exclusion

.32* .67* .18 .61* .20

.32* .46* .48* .53* .I1

ON = 50 per cell, except for “Specific-General with joint leadin”. for which N = 56. Correlations marked by an asterisk differ from chance, p < .05. Adapted, with permission, from Schwarz, N., Strack, F., & Mai, H. P. (1991), “Assimilation and contrast effects in part-whole question sequences: A conversational logic analysis.” Public Opinion Quarterly, 55, 3-23.

LOGIC OF CONVERSATION

15 1

introduced by a joint lead-in that read, “We now have two questions about your life. The first pertains to your marital satisfaction and the second to your general life-satisfaction.” Under this condition, the same question order that resulted in r = .67 without a joint lead-in, now produced a low and nonsignificant correlation or r = .18. This suggests that respondents deliberately ignored information that they had already provided in response to a specific question when making a subsequent general judgment-$ the specific and the general questions were assigned to the same conversational context, thus evoking the application of conversational norms that prohibit redundancy. In that case, respondents apparently interpreted the general question as if it referred to aspects of their lives that they had not yet reported on. In line with this interpretation, a condition in which respondents were explicitly asked how satisfied they are with “other aspects” of their lives, “aside from their relationship,” yielded a nearly identical correlation of r = .20. In addition, respondents who were induced to disregard their marriage in evaluating their lives as a whole, either by the conversational context manipulation or by explicit instructions, reported higher mean life-satisfaction when they were unhappily married, and lower mean life-satisfactionwhen they were happily married, than respondents who were not induced to exclude this information. Thus, contrast effects were obtained when conversational norms elicited the exclusion of the primed information from the representation formed of one’s life in general, whereas assimilation effects were obtained when the activated information was included in this representation (see Schwarz & Bless’s, 1992, inclusionlexclusion model of assimilation and contrast effects for a more detailed theoretical discussion). In a subsequent study, Schwarz and Hippler (unpublished data) observed that the conversational norm of nonredundancy may be evoked not only by a joint lead-in but also by the graphic layout of a questionnaire. Specifically, the marital satisfaction question and the general question were presented either in separate boxes with a black frame drawn around each question or in a joint box with one frame drawn around both questions. As in the above data, increasing the conversational relatedness of both questions by presenting them in one box significantly reduced the otherwise obtained correlation, again illustrating the conversational relevance of apparently formal features of questionnaires. Note, however, that the applicability of the norm of nonredundancy may vary as a function of the number of specific questions that precede the more general one. If only one specific question precedes the general one, the repeated use of the information on which the answer to the specific question was based results in redundancy in the response to the general question. Hence, this repeated use of the same information is avoided if both questions are assigned to the same conversational context, as the above data demonstrated. Suppose, however, that several specific questions precede the general one. For example, respondents

152

NORBERT SCHWARZ

may be asked to report on their marital satisfaction, their job satisfaction, and their leisure-time satisfaction before a general life-satisfaction question is presented. In that case, they may interpret the general question in two different ways. On the one hand, they may assume that it is a request to consider still other aspects of their lives, much as if it were worded, “Aside from what you already told us. . . .” On the other hand, they may interpret the general question as a request to integrate the previously reported aspects into an overall judgment, much as if it were worded, “Taking these aspects together, how satisfied are you with your life as a whole?’ Note that this interpretational ambiguity of the general question does not arise if only one specific question was asked. In that case, an interpretation of the general question in the sense of “taking all aspects together” would make little sense because only one aspect was addressed, thus rendering this interpretation of the general question completely redundant with the specific one. If several specific questions were asked, however, both interpretations of the general question are viable. In this case, the interpretation of the general question as a request for a final integrative summary judgment is legitimate from a conversational point of view. If several specific questions have been asked, an integrative judgment is informative because it does provide “new” information about the relative importance of the respective domains, which are in the focus of the conversation. Moreover, “summing up” at the end of a series of related thoughts is acceptable conversational practice-whereas there is little to sum up if only one thought was offered. Accordingly, respondents may interpret a general question as a request for a summary judgment if it is preceded by several specific ones, even if all questions are explicitly placed into the same conversational context. To test this theoretical analysis, other respondents of the Schwarz, Strack, and Mai (1991) study were asked three specific questions, pertaining to their leisuretime satisfaction, their job satisfaction, and, finally, their marital satisfaction. As shown in the second column of Table IV, the correlation between marital satisfaction and life satisfaction increased from r = .32 to r = .46 when answering the specific questions first brought information about one’s marriage to mind. However, this increase was less pronounced than when the marital-satisfaction question was the only specific question that preceded the general one (r = .67), reflecting that the three specific questions brought a more varied set of information to mind. More important, introducing the three specific and the general question by a joint lead-in, thus assigning them explicitly to the same conversational context, did not reduce the emerging correlation, r = .48. This result indicates that respondents adopted a “taking-all-aspects-together” interpretation of the general question if it was preceded by three specific questions, rather than one. This interpretation is further supported by a highly similar correlation of r = .53 when the general question was reworded to request an integrativejudgment, and a highly dissimilar correlation of r = . l l when the reworded question required the consideration of other aspects of one’s life.

LOGIC OF CONVERSATION

153

In combination, these findings further emphasize that the interpretation of an identically worded question may change as a function of conversational variables, resulting in markedly different responses. Moreover, the emerging differences are not restricted to the means or margins of the response distribution, as social scientists have frequently hoped. Rather, context variables may result in different correlational patterns, thus violating the assumption that context effects would be restricted to differences in the means, whereas the relationship between variables would be “form resistant” (Schuman & Duncan, 1974; Stouffer & DeVinney, 1949). 2 . Behavioral Reports and Evaluative Judgments This conclusion is further supported by a study reported by Strack, Martin, and Schwarz (1988) that indicates that the above effects are not restricted to order effects between similar evaluative judgments. Rather, the same variables may also affect the observed relationship between behavioral reports and subsequent judgments. Specifically, Strack et al. (1988, Experiment 2) asked American college students to report their general life-satisfaction as well as their dating frequency. When the life-satisfaction question preceded the dating-frequency question, the correlation was weak, r = -. 12, and not significant, suggesting that dating frequency may contribute little to students’ overall well-being. Reversing the question order, however, increased the correlation dramatically to r = .66. This result presumably reflects that the dating-frequency question increased the cognitive accessibility of dating-related information, which was then used in evaluating one’s life as a whole. On the substantive side, this correlation would suggest that dating frequency is a major contributor to life satisfaction for college students. However, placing both questions in the same conversational context by a joint lead-in again reduced the obtained correlation to nonsignificance, r = .15, reflecting that respondents ignored the information they had already provided when the conversational context elicited the norm of nonredundancy. 3. Conclusions

The findings of the Schwarz, Strack, and Mai (1991) and Strack et al. (1988) studies have methodological as well as theoretical implications. On the methodological side, they illustrate the fact that a researcher may draw very different substantive conclusions about the contribution of marital happiness or dating frequency to individuals’ overall well-being, depending on the order in which the questions are asked. To account for the impact of question order, however, it is not sufficient to consider purely cognitive variables in isolation. Whereas preceding questions increase the cognitive accessibility of information used to answer them, this increase does not necessarily result in an increased use of the

154

NORBERT SCHWARZ

primed information in making subsequent judgments, in contrast to what current theorizing in social cognition would suggest. According to current models of information accessibility and use (see Bodenhausen & Wyer, 1987; Higgins & Bargh, 1987; Wyer & Srull, 1986), the use of information is determined by its accessibility in memory and its applicability to the task at hand. In all of this research, applicability has been assumed to be solely determined by the nature of the stimulus materials. As the current studies indicate, however, the conversational context may change the perceived nature of the judgmental task and may lead subjects to deliberately ignore information that is easily accessible and potentially relevant to the judgment at hand. Hence, the emergence of priming effects is determined not only by the nature of the stimulus materials or the literal question asked but also by the conversational context in which subjects are asked to make their judgment. As Strack et al. (1988) emphasized, a full understanding of the use of highly accessible information therefore requires not only a consideration of its applicability to the task at hand but also of its appropriateness in the conversational context (see Martin & Achee, 1992; Schwarz & Bless, 1992; Strack, 1992a, 1992b, for further general discussions of information use).

V. General Conclusions Social cognition research has typically ignored the fact that assessing a judgment in a research situation is part of an ongoing conversation between the subject and the researcher. In essence, social cognition researchers have conceptualized the judge as doing his or her cognitive work in social isolation. In many instances, the obtained findings provided a less than flattering portrait of our respondents. Apparently, people are happy to offer meaningless opinions on nonexistent issues and are biased by irrelevant material such as the numeric values of a rating scale or the response alternatives of a frequency question. Moreover, they are apparently more than willing to use worthless personality information and to ignore more meaningful base rates, or to draw strong dispositional inferences despite obvious situational pressures, and so on. As soon as we conceptualize the assessed judgments as part of an ongoing conversation, however, the often dramatic findings seem less surprising. Rather, research participants seem to do what they would rightly be expected to do in any other conversation: They assume that our utterances as researchers are meaningful, that we do not ask questions about things that don’t exist, that we do construct meaningful rather than arbitrary scales, and so on. Moreover, they try to make sense of our utterances and of our research instruments on the basis of these assumptions. In many cases, people do not lack the ability to make adequate judgments.

LOGIC OF CONVERSATION

155

Rather, what they lack is the insight that we as researchers do not live up to the standards that we would typically observe in any other conversation that we conduct. They simply give us more credit than we deserve, by assuming that the information we provide is relevant to the task at hand, truthful, informative, and clear. Unless we learn to observe these standards in our conduct of research and in our interpretation of results, we may run the risk of painting a rather inadequate picture of human judgment by severely overestimating the size and pervasiveness of judgmental biases and shortcomings. Certainly, violations of conversational norms are nor the only source of judgmental biases. Like most robust phenomena, these biases are likely to be overdetermined, as Ross and Nisbett (1991) emphasized. To fully understand their operation in natural contexts, however, we have to ensure that their operation in our laboratories does not reflect determinants that are unlikely to be similarly powerful outside of our labs. Obviously, this concern with possible “side-effects” of research procedures is not new. Rather, it has been at the heart of a research tradition concerned with the social psychology of the psychological experiment (see Kruglanski, 1975, for an overview). However, following Orne’s (1962, 1969) seminal discussion of demand characteristics, this research has been guided by the assumption that subjects are motivated to look for cues in the experimental situation that provide them with the experimenter’s hypothesis. Depending on their motivation to play the role of a “good subject,” they may then react in line with the suspected hypothesis. Accordingly, most of that early research focused on subjects’ motivation to detect and to act according to the experimenter’s hypothesis, rather than on the process by which subjects extract information from the research procedures used. In contrast, the current analysis suggests that we do not need to make special assumptions about motivations that may be germane to the participation in an experiment to account for the reviewed findings. Rather, the current analysis indicates that subjects’ behavior in an experiment or research interview is guided by the same assumptions and motivations that govern the conduct of conversation in any other setting (see Bless, Strack, & Schwarz, 1993, for a detailed comparison of Orne’s analysis and a conversational perspective). From a conversational point of view, the key difference between experiments and conversations in natural settings is only that the experimenter is less likely to comply with conversational rules in conducting an experiment than in conducting any other conversation, although subjects have no reason to suspect that the experimenter is not a cooperative communicator. As a result, subjects apply the tacit assumptions that usually govern the conduct of conversation to the research setting and go beyond the literal information provided to them by drawing inferences on the basis of the conversational context. The apparent biases and errors that subjects commit by relying on conversational maxims are less likely to result in mistakes in everyday contexts in which communicators try to conform to conversational norms, provide information that

156

NORBERT SCHWARZ

is relevant to the judgment at hand, and make the task one that is clear rather than ambiguous-and in which recipients are indeed expected to use contextual cues to disambiguate the communication, should the communicator not live up to the ideal. Thus, the behavior that may lead to errors in the experimental context may be adaptive in everyday settings. As Funder (1987, p. 82) noted in a related context, It seems ironic that going beyond the information given in this way is so often interpreted by social psychologists as symptomatic of flawed judgment. Current thinking in the field of artificial intelligence is that this propensity is exactly what makes people smarter than computers.

To acknowledge this special potential of human information processors, social cognition research will need to pay closer attention to the social context in which much of our cognitive work is conducted. To do so, social cognition research will eventually need to extend the “flowchart model of information processing that presents us only with a unilateral input/output paradigm that stops short of reciprocity” (Markus & Zajonc, 1985, p. 212). Taking the tacit assumptions that govern the conduct of conversation into account is likely to provide a good starting point for this endeavor.

Acknowledgments Some of the reported research was supported by grant SWF 0044 6 from the Bundesminister fur Forschung und Technologie of the Federal Republic of Germany to the author; by grants Schw 278/2 and Str 26412 from the Deutsche Forschungsgemeinschaft to the author and Fritz Strack, and Schw 278/5 from the Deutsche Forschungsgemeinschaft to the author, Herbert Bless, and Gerd Bohner. Thanks are due to Fritz Strack and Denis Hilton, who greatly contributed to my understanding of conversational issues, and to Herbert Bless, Gerd Bohner, Michael Conway, Tory Higgins, Hans-J. Hippler, Richard Nisbett, Michaela W h k e , and Mark Zanna for stimulating discussions and/or comments on previous drafts.

References Argyle, M. (1992). The social psychology of everyday life. London, England: Routledge. Bishop, G . . Hippler, H. J., Schwarz, N., & Strack, F. (1988). A comparison of response effects in self-administered and telephone surveys. In R. M. Groves, P. Biemer, L. Lyberg, J. T. Massey, W. L. Nicholls, & J. Waksberg (Eds.), Telephone survey methodology (pp. 321-340). New York: Wiley.

LOGIC OF CONVERSATION

157

Bishop, G. F., Oldendick, R. W., & 'hchfarber, R. J. (1986). Opinions on fictitious issues: The pressure to answer survey questions. Public Opinion Quarterly, 50, 240-250. Bless, H., Bohner, G., Hild, T., & Schwarz, N. (1992). Asking difficult questions: Task complexity increases the impact of response alternatives. European Journal of Social Psychology, 22. 309312. Bless, H., Strack, F., & Schwarz, N. (1993). The informative functions of research procedures: Bias and the logic of conversation. European Journal of Social Psychology, 23, 149-165. Bodenhausen, G . V., & Wyer, R. S. (1987). Social cognition and social reality: Information acquisition and use in the laboratory and the real world. In H. J. Hippler, N. Schwarz, & S. Sudman (Eds.), Social information processing and survey methodology (pp. 6-4 I). New York: Springer Verlag. Bradburn, N. (1982). Question wording effects in surveys. In R. Hogarth (Ed.), Question framing and response consistency (pp. 65-76), San Francisco: Jossey-Bass. Bradburn, N. M. (1983). Response effects. In P. H. Rossi, J. D. Wright, & A. B. Anderson (Eds.), The handbook of survey research (pp. 289-328). New York: Academic Press. Bradburn, N. M., Rips, L. J., & Shevell, S. K. (1987). Answering autobiographical questions: The impact of memory and inference on surveys. Science, 236, 157-161. Clark, H. H. (1977). Inferences in comprehension. In D. La Berge & S. Samuels (Eds.), Basic processes in reading: Perception and comprehension (pp. 243-263), Hillsdale, NJ: Erlbaum. Clark, H. H. (1985). Language use and language users. In G . Lindzey, & E. Aronson (Eds.), Handbook of social psychology (Vol. 2, pp. 179-232). New York: Random House. Clark, H. H., & Brennan, S. E. (1991). Grounding in communication. In L. Resnick, J. Levine, & S. D. Teasley (Eds.), Perspectives on socially shared cognition (pp. 127-149). Washington, DC: American Psychological Association. Clark, H. H., & Clark, E. V. (1977). Psychology and language. New York Harcourt, Brace, Jovanovich. Clark, H. H., & Haviland, S. E. (1977). Comprehension and the given-new contract. In R. 0. Freedle (Ed.), Discourse production and comprehension (pp. 1-40). Hillsdale, NJ: Erlbaum. Clark, H. H., & Schober, M. F. (1992). Asking questions and influencing answers. In I. M. Tanur (Ed.), Questions about questions (pp. 15-48). New York: Russel Sage. Clark, H. H., Schreuder, R., & Buttrick, S. (1983). Common ground and the understanding of demonstrative reference. Journal of Verbal Learning and Verbal Behavior. 22, 245-258. Converse, P. E. (1964). The nature of belief systems in mass politics. In D. Apter (Ed.), Ideology and discontent (pp. 238-45). New York: Free Press of Glencoe. Converse, P. E. (1970). Attitudes and non-attitudes: Continuation of a dialogue. In E. R. Tufte (Ed.), The quantitative analysis of social problems (pp. 188- 189). Reading, MA: Addison-Wesley. Dawes, R. M., & T. Smith (1985). Attitude and opinion measurement. In G. Lindzey & E. Aronson (Eds.), Handbook of social psychology (Vol. 2, pp. 509-566). New York: Random House. Dockrell, J., Neilson, I., & Campbell, R. (1980). Conservation accidents revisited. International Journal of Behavioral Development, 3 , 423-439. Dodd, D. H., & Bradshaw, J. M. (1980). Leading questions and memory: Pragmatic constraints. Journal of Verbal Learning and Verbal Behavior, 19, 695-704. Donaldson, M. (1982). Conservation: What is the question? British Journal of Psychology, 73, 199207. Edwards, D., & Potter, J. (1993). Language and causation: A discursive model of description and attribution. Psychological Review, 100, 23-41. Erikson, R. S., Luttberg, N. R., & Tedin, K. T. (1988). Americanpublic opinion (3rd ed.). New York: Macmillan. Forgas, J. P. (1981). What is social about social cognition? In J. P. Forgas (Ed.), Social cognition. Perspectives on everyday understanding (pp. 1-26). New York: Academic Press.

158

NORBERT SCHWARZ

Funder, D. C. (1987). Errors and mistakes: Evaluating the accuracy of social judgment. Psychological Bulletin, 101, 75-90. Fussel, S. R., & Krauss, R. M. (1989a). Understanding friends and strangers: The effects of audience design on message comprehension. European Journal of Social Psychology, 19, 509-525. Fussel, S. R., & Krauss, R. M. (1989b). The effects of intended audience on message production and comprehension: Reference in a common ground framework. Journal of Experimental Social Psychology, 25, 203-219. Ginossar, Z., & Trope, Y. (1987). Problem solving and judgment under uncertainty. Journal of Personality and Social Psychology, 52, 464-474. Grice, H. P. (1975). Logic and conversation. In P. Cole & J. L. Morgan (Eds.), Syntax and semantics: Vol. 3: Speech acts. pp. 41-58. New York: Academic Press. Gruenfeld, D. H., & Wyer, R. S. (1992). Semantics and pragmatics of social influence: How affirmations and denials affect beliefs in referent propositions. Journal of Personality and Social PSycholOgy, 62, 38-49. Harris, R. J., & Monaco, G. E. (1978). Psychology of pragmatic implication: Information processing between the lines. Journal of Experimental Psychology: General, 107, 1-27. Higgins, E. T. (1981). The “communication game”: Implications for social cognition and communication. In E. T. Higgins, M. P. Zanna, & C. P. Herman (Eds.), Social cognition: The Ontario Symposium (Vol. 1, pp. 343-392). Hillsdale, NJ: Erlbaum. Higgins, E. T. (1989). Knowledge accessibility and activation: Subjectivity and suffering from unconscious sources. In J. S. Uleman & J. A. Bargh (Eds.), Uninrended thought (pp. 75-123). New York: Guilford Press. Higgins, E. T. & Bargh, J. A. (1987). Social cognition and social perception. Annual Review of Psychology, 38, 369-425. Higgins, E. T., Fondacaro, R. A., & McCann, C. D. (1982). Rules and roles: The “communication game” and speaker-listener processes. In W. P. Dickson (Ed.), Children’s oral communication skills (pp. 289-312). New York: Academic Press. Higgins, E. T., McCann, C. E., & Fondacaro, R. A. (1982). The “communication game”: Goaldirected encoding and cognitive consequences. Social Cognition, 1 , 21 -37. Higgins, E. T., & Rholes, W. S. (1978). Saying is believing: Effects of message modification on memory and liking for the person described. Journal of Experimental Social Psychology, 14, 363-378. Hilton, D. J. (1 990). Conversational processes and causal explanation. Psychological Bulletin, 107, 65-81. Hilton, D. J. (1991). A conversational model of causal explanation. In W. Stroebe & M.Hewstone (Eds.), European review of social psychology (Vol. 2, pp. 3 1-50), Chichester, England: Wiley. Hilton, D. J., & Slugoski, B. R. (1986). Knowledge based attribution: The abnormal conditions focus model. Psychological Review, 93, 75-88. Jones, E. E. (1990). Interpersonal perception. New York: Freeman. Jones, E. E., & Harris, V. A. (1967). The attribution of attitudes. Journal of Experimental Social Psychology. 3 , 1-24. Kahneman, D., & Lovallo, D. (1993). Timid choices and bold forecasts. Management Science, 39, 17-31. Kahneman, D., & Tversky, A. (1973). On the psychology of prediction. Psychological Review, 80, 237-251. Krauss, R. M., & Fussel, S. R. (1991). Perspective-taking in communication: Representations of others’ knowledge in reference. Social Cognition, 9. 2-24. Krosnick, J. A., Li, F., & Lehman, D. R. (1990). Conversational conventions, order of information on, and the effect of base rates and individuating information on social judgment. Journal of Personality and Social Psychology, 59, 1140-1 152.

LOGIC OF CONVERSATION

159

Kruglanski, A. W. (1975). The human subject in the psychology experiment: Fact and artifact. Advances in Experimental Social Psychology, 8, 101-147. Levinson, S . C. (1983). Pragmatics. Cambridge, England: Cambridge University Press. Light, P., Buckingham, N., & Robbins, A. H. (1979). The conservation task as an interactional setting. British Journal of Educational Psychology, 49, 304-310. Loftus, E. F. (1975). Leading questions and the eyewitness report. Cognitive Psychology, 7 , 560572. Loftus, E. F. (1979). Eyewitness testimony. Cambridge, MA: Harvard University Press. Markus, H., & Zajonc, R. B. (1985). The cognitive perspective in social psychology. In G. Lindzey & E. Aronson (Eds.), The handbook of social psychology (Vol. I , pp. 137-230). New York: Random House. Martin, L. L., & Achee, J. W. (1992). Beyond accessibility: The role of processing objectives in judgment. In L. L. Martin & A. Tesser (Eds.), The construction of social judgments (pp. 195216). Hillsdale, NJ: Erlbaum. Martin, L. L., & Clark, L. F. (1990). Social cognition: Exploring the mental processes involved in human social interaction. In M. W. Eysenck (Ed.), Cognitive psychology: An international review (pp. 265-310). Chichester, England: Wiley. McCann, C. D., & Higgins, E. T. (1992). Personal and contextual factors in communication: A review of the “communication game.” In G. R. Semin & K. Fiedler (Eds.), Language, interaction, and social cognition (pp. 144-172). Newbury Park, CA: Sage. McGarrigle, J., & Donaldson, M. (1974). Conservation accidents. Cognition, 3, 341-350. Menon, G. (in press). Judgments of behavioral frequency: Memory search and retrieval strategies. In N. Schwarz & S. Sudman (Eds.), Autobiographical memory and the validity of retrospective reports. New York: Springer Verlag. Miller, A. G., Schmidt, D., Meyer, C., & Colella, A. (1984). The perceived value of constrained behavior: Pressures toward biased inference in the attitude attribution paradigm. Social Psychology Quarterly, 47, 160-171. Molenaar, N. J. (1982). Response effects of formal characteristics of questions. In W. Dijkstra & J. van der Zouwen (Eds.), Response behavior in the survey interview (pp. 49-90). New York: Academic Press. Nisbett, R. E., & Borgida, E. (1975). Attribution and the psychology of prediction. Journal of Personality and Social Psychology, 32, 932-943. Nisbett, R. E., & Ross, L. (1980). Human inference: Strategies and shortcomings of social judgment. New York: Prentice Hall. Nisbett, R. E., Zukier, H., & Lemley, R. H. (1981). The dilution effect. Cognitive Psychology, 13, 248-277. Ome, M. T. (1962). On the social psychology of the psychological experiment: With particular reference to demand characteristics and their implications. American Psychologist, 17, 776783. Ome, M. T. (1969). Demand characteristics and the concept of quasi-controls. In R. Rosenthal & R. L. Rosnow (Eds.), Artifact in behavioral research. New York: Academic Press. Ostrom, T. M. (1984). The sovereignty of social cognition. In R. S. Wyer & T. K. Krull (Eds.), Handbook of social cognition. (Vol. 1, pp. 1-38). Hillsdale: Erlbaum. Piaget, J. (1952). The child‘s conception of number. London, England: Routledge & Kegan Paul. Piaget, J., & Inhelder, B. (1969). The psychology of the child. New York: Basic Books. Prince, E. F. (1981). Towards a taxonomy of given-new information. In P. Cole (Ed.), Radical pragmatics (pp. 223-256). New York: Academic Press. Rose, S. A,, & Blank, M. (1974). The potency of context in children’s cognition: An illustration through conservation. Child Development, 45, 499-502.

160

NORBERT SCHWARZ

Ross, L. (1977). The intuitive psychologist and his shortcomings: Distortion in the attribution process. Advances in Experimental Social Psychology, 10, 174-221. Ross, L., & Nisbett, R. E. (1991). The person and the situation. New York: McGraw-Hill. Schiffer, S. (1972). Meaning. Oxford, England: Clarendon Press. Schneider, D. (1991). Social cognition. Annual Review of Psychology, 42, 527-561. Schuman, H., & Duncan, 0. D. (1974). Questions about attitude survey questions. In H. L. Costner (Ed.), Sociological Methodology. San Francisco: Jossey-Bass. Schuman, H., & Kalton, G. (1985). Survey methods. In G. Lindzey & E. Aronson (Eds.),Handbook of social psychology (Vol. I). New York: Random House. Schuman, H., & Presser, S. (1977). Question wording as an independent variable in survey analysis. Sociological Methods and Research, 6 , 151-176. Schuman, H.,& Presser, S. (1981). Questions and answers in attitude surveys. New York Academic Press. Schwarz, N. (1987). Stimmung als Information. Untersuchungen zum EinJuJ von Srimmungen auf die Eewertung des eigenen Lebens. [Mood as information. The impact of moods on the evaluation of one’s life.]. Heidelberg: Springer Verlag. Schwarz, N. (1990). Assessing frequency reports of mundane behaviors: Contributions of cognitive psychology to questionnaire construction. In C. Hendrick & M. S. Clark (Eds.), Research methods in personality and social psychology: Review of personality and social psychology (Vol. 1 I , pp. 98- 119). Beverly Hills, CA: Sage. Schwarz, N., & Bienias, J. (1990). What mediates the impact of response alternatives on frequency reports of mundane behaviors? Applied Cognitive Psychology, 4 , 61-72. Schwarz, N., & Bless, H. (1992). Constructing reality and its alternatives: An inclusion/exclusion model of assimilation and contrast effects in social judgment. In L. Martin & A. Tesser (Eds.), The construction of social judgment (pp. 217-245). Hillsdale, NJ: Erlbaum. Schwarz, N., Bless, H., Bohner, G., Harlacher, U . , & Kellenbenz, M. (1991). Response scales as frames of reference: The impact of frequency range on diagnostic judgment. Applied Cognitive PSyChOlOgy, 5 , 37-50. Schwarz, N., & Hippler, H. J. (1987). What response scales may tell your respondents: Informative functions of response alternatives. In H. J. Hippler, N. Schwarz, & S. Sudman (Eds.), Social information processing and survey methodology (pp. 163- 178). New York: Springer Verlag. Schwarz, N., & Hippler, H. J. (1991). Response alternatives: The impact of their choice and ordering. In P. Biemer, R. Groves, N. Mathiowetz, & S. Sudman (Eds.), Measurement error in surveys (pp. 41-56). Chichester: Wiley. Schwarz, N., & Hippler, H. J. [Context effects in surveys]. Unpublished raw data. Schwarz, N., Hippler, H. J., Deutsch, B., & Strack, F. (1985). Response scales: Effects of category range on reported behavior and subsequent judgments. Public Opinion Quarterly, 49, 388-395. Schwarz, N., Hippler, H. J., & Noelle-Neumann, E. (in press). Retrospective reports: The impact of response alternatives. In N. Schwarz & S. Sudman (Eds.), Autobiographical memory and the validity of retrospective reports. New York: Springer Verlag. Schwarz, N., Knauper, B., Hippler, H. J., Noelle-Neumann, E., & Clark, F. (1991). Rating scales: Numeric values may change the meaning of scale labels. Public Opinion Quarterly, 55, 570582.

Schwarz. N., & Scheuring, B. (1988). Judgments of relationship satisfaction: Inter- and intraindividual comparison strategies as a function of questionnaire structure. European Journal of Social Psychology, 18, 485-496. Schwarz, N., & Strack, F. (1991b). Evaluating one’s life: A judgment model of subjective wellbeing. In F. Strack, M. Argyle, & N. Schwarz (Eds.), Subjective well-being. An interdisciplinary perspective (pp. 27-47). Oxford, England: Pergamon.

LOGIC OF CONVERSATION

161

Schwarz, N., Strack, F., Hilton, D. J., & Naderer, G . (1991). Judgmental biases and the logic of conversation: The contextual relevance of irrelevant information. Social Cognition, 9, 67-84. Schwarz, N., Strack, F., & Mai, H. P. (1991). Assimilation and contrast effects in part-whole question sequences: A conversational logic analysis. Public Opinion Quarterly, 55, 3-23. Schwarz, N., Strack, F., Miiller, G., & Chassein, B. (1988). The range of response alternatives may determine the meaning of the question: Further evidence on informative functions of response alternatives. Social Cognition, 6, 107- 117. Siegal, M., Waters, L. J., & Dinwiddy, L. S. (1988). Misleading children: Causal attributions for inconsistency under repeated questioning. Journal of Experimental Child Psychology, 45, 438456. Smith, V., & Ellsworth, P. C. (1987). The social psychology of eyewitness accuracy: Leading questions and communicator expertise. Journal of Applied Psychology, 72, 292-300. Sperber, D., & Wilson D. (1986). Relevance: Communication and cognition. Cambridge, MA: Harvard University Press. Stalnaker, R. C. (1978). Assertion. In Cole, P. (Ed.), Syntax and semantics: Vol. 9. Pragmutics (pp. 315-332). New York: Academic Press. Stouffer, S. A,, & DeVinney, L. C. (1949). How personal adjustment varied in the army-By background characteristics of the soldiers. In S. A. Stouffer, E. A. Suchman, L. C. DeVinney, S. A. Star, & R. M. Williams (Eds.), The American soldier: Adjustment during army life. Princeton, NJ: Princeton University Press. Strack, F. (1992a). Order effects in survey research: Activative and informative functions of preceding questions. In N. Schwarz & S. Sudman (Eds.), Context effects in social andpsychological research (pp. 23-34). New York: Springer Verlag. Strack, F. (1992b). The different routes to social judgments: Experiential versus informational strategies. In L. Martin & A. Tesser (Eds.), The construction of social judgment (pp. 249-275). Hillsdale, NJ: Erlbaum. Strack, F. (in press). Urteilsprozesse in standardisierten Befragungen: kognitive und kommunicative EinJusse. [Judgmental processes in standardized interviews: cognitive and communicative influences.]. Heidelberg: Springer Verlag. Strack, F., & Martin, L. (1987). Thinking, judging, and communicating: A process account of context effects in attitude surveys. In H. J. Hippler, N. Schwarz, & S. Sudman (Eds.),Social information processing and survey methodology (pp. 123- 148). New York: Springer Verlag. Strack, F., Martin, L. L., & Schwarz, N. (1988). Priming and communication: The social determinants of information use in judgments of life-satisfaction. European Journal of Social Psychology, 18, 429-442. Strack, F., & Schwarz, N. (1992). Communicative influences in standardized question situations: The case of implicit collaboration. In K.Fiedler & G. Semin (Eds.),Language, interaction and social cognition (pp. 173-193). Beverly Hills, CA: Sage. Strack, F., Schwarz, N., & Whke, M. (1991). Semantic and pragmatic aspects of context effects in social and psychological research. Social Cognition. 9, 1 11-125. Swann, W. B., Giuliano, T., & Wegner, D. M. (1982). Where leading questions can lead: The power of conjecture in social interaction. Journal of Personality and Social Psychology, 42, 10251035. Tetlock, P. E., & Boettger, R. (1993). The dilution effect: Judgmental bias, conversational convention. Unpublished manuscript, University of California, Berkeley. Tourangeau, R. (1984). Cognitive science and survey methods: A cognitive perspective. In T.Jabine, M. Straf, J. Tanur, & R. Tourangeau (Eds.), Cognitive aspects of survey methodology: Building a bridge between disciplines (pp. 73- 100). Washington, DC: National Academy Press. Wegner, D. M., Wenzlaff, R., Kerker, R. M., & Beattie, A. E. (1981). Incrimination through

162

NORBERT SCHWARZ

innuendo: Can media questions become public answers? Journal of Personality and Social Psychology, 40, 822-832. Woll, S . B . , Weeks, D. G . , Fraps, C. L., Pendergrass, J., & Vanderplas, M. A. (1980). Role of sentence context in the encoding of trait descriptors. Journal of Personality and Social Psychology, 39, 59-68. Wright, E. F., & Wells, G. L. (1988). Is the attitude-attribution paradigm suitable for investigating the dispositional bias? Personality and Social Psychology Bulletin, 14, 183-190. Wyer, R. S., & SNH, T. K. (1986). Human cognition in its social context. Psychological Review, 93, 322-359.

A PHASE MODEL OF TRANSITIONS: COGNITIVE AND MOTIVATIONAL CONSEQUENCES

Diane N. Ruble

I. Introduction The question of continuity and discontinuity in development is among the most basic as well as controversial issues in developmental psychology today. Human lives are characterized by periods of relative stability and periods of change or transition. During times of transition, major reorganization, disruption, and discontinuity may occur, often affecting the personal and interpersonal adjustment of the individuals involved. The literature offers numerous examples of such transition phenomena. One widely known illustration, and one of my own research interests, concerns the transition to parenthood (Belsky, Rovine, & Fish, 1989; Fleming, Ruble, Flett, & Van Wagner, 1990; Ruble, Fleming, Hackel, & Stangor, 1988). Two particularly well documented disruptions associated with this transition are postpartum depression and marital problems. A second illustration is the drop in self-esteem and self-evaluation of children across the transition to junior high school. Two major programs of research-Eccles and her colleagues (Eccles & Midgley, 1990; Wigfield, Eccles, MacIver, Reuman, & Midgley, 1991) and Simmons and her colleagues (Simmons & Blyth, 1987; Simmons, Rosenberg, & Rosenberg, 1973) have shown that the transition to junior high school is often marked by a sharp drop in positive attitudes toward school, negative self-concept and body image, increases in anxiety and worry, and negative changes in motivational orientations. Curiously, in spite of the significance of these periods of change, there has been relatively little theorizing about why the changes occur. There is evidence pointing to certain moderator variables (e.g., adolescent girls are more likely to show a drop in self-esteem than are boys), but little is known about what factors mediate the impact of a period of transition on the individual. In a provocative ADVANCES IN EXPkRIMENTAL SOCIAL PSYCIIOLOGY. VOI. 26

163

Copyright 0 1994 by Academic Press, Inc. All rights of reproduciion in any form reserved.

164

DIANE N. RUBLE

chapter published a few years ago, Connell and Furman (1984) suggested that although a great deal of attention has been paid to describing outcomes of transitions, the mechanisms responsible are virtually unknown. Of course, a number of mechanisms may be involved. Transitions during puberty may be partly mediated by changes in levels and fluctuations in hormones. Certain kinds of social pressures are also presumed to play a role. Lewin (1951), for example, suggested that the effects of adolescent transitions may be mediated, in part, by the conflicting forces of shifting reference groups. My research has concerned a different type of mediating process, one emanating from the changing motivations and cognitions of the individuals themselves. The focus has been on how the individual’s orientation toward relevant information changes across different phases of a transition. Specifically, what is it that the individual is doing during a period of transition that may result in the kinds of changes that have been observed at the end? The present consideration of this question will begin with the specific graphic example of the transition to first motherhood. Literatures related to the proposed transition processes will then be reviewed. Next, I will suggest that the same analysis might apply to certain kinds of social-cognitive changes that occur in childhood that are not usually conceptualized as transitions-changes in concepts about gender, ability, and the characteristicsof people. Finally, the possible implications of this analysis for understanding changes that often accompany well-known transitions, such as pregnancy and the move to junior high school, will be described.

11. Illustration from Data on Transition

to First Motherhood Figure 1 shows some cross-sectional data on women’s information seeking as a function of stage in the transition to motherhood (Deutsch, Ruble, Fleming, Brooks-Gunn, & Stangor, 1988). Six groups of women were compared, representing three stages of pregnancy, two stages postpartum, and a “prepregnant” comparison (i.e., women who expected to become pregnant within a few years). The women were asked to report how much information they had received in the previous month on 12 topics related to motherhood, pregnancy, or labor and delivery. The lines in the figure represent the sum of many different types of information sources (books, friends, doctors, and so on). The solid line reflects the mean amount of information received about motherhood; the dashed line, about pregnancy; and the dash-and-dot line, about labor and delivery. Several interesting observations can be made about these data. First, women clearly anticipate the event by engaging in active information seeking as soon as

165

PHASE MODEL OF TRANSITIONS I

I

I

I

-motherhood ----+pregnancy -.--.+labor and delivery

I

I

4

,,%. /'

i /'

8. 8.

'.

/' /'

'.

'-\ '-\

-. D

J I

I

prepregnant

Is' trimester

I

2nd

trimester

I

I

I

I

3rd I month 3months trimester postpartum postpartum

PHASE IN TRANSITION TO MOTHERHOOD Fig. 1. Mean ratings of amount of information received as a function of phase in transition to motherhood. From "Information-Seekingand Maternal Self-definition during the Transition to Motherhood" by F. M. Deutsch, D. N. Ruble, A. Fleming, J. Brooks-Gunn, and C. s. Stangor, 1988, Journal of Personality and Social Psychology, 55, p. 425. Copyright 1988 by the American Rychological Association.

they enter a transitional state. The difference between pregnant and prepregnant phases was significant for all topics. In addition, the women are clearly seeking the information; it is not imposed upon them. Books are the most frequently cited source on every topic for all groups of women except prepregnant, who are more likely to refer to friends. Second, the women are gathering information about all three topics in a way that shows clear peaks of interest at particular times (e.g., peak in the first trimester for information about pregnancy). This information may be used to construct a set of expectations about what each of these experiences will be like. For example, we found that the amount of information seeking about motherhood during pregnancy was related to the formation of mothering self-definitions (a woman's expectations and confidence about how good a mother she will be). Such expectations may have long-term consequences; the maternal expectations formed during pregnancy carried over to affect postpartum mothering selfdefinitions (Deutsch, Ruble, Fleming, Brooks-Gunn, & Stangor, 1986). Interestingly, Eleanor Maccoby published a study 30 years ago that came to similar conclusions concerning critical periods in receptivity to information among new mothers. She and her colleagues (Maccoby, E., Maccoby, N., Ro-

166

DIANE N. RUBLE

mney, & Adams, 1961; Maccoby, N., Romney, Adams, & Maccoby, E., 1959) found that women with young babies were much more likely to read a booklet on child care that they got through the mail, to retain information from it, and to discuss it with others, relative to women with older children or women who did not yet have a baby. The goal of presenting these data sets is only to illustrate some of the dramatic changes in individuals’ orientations toward information during a transition. The main point with respect to this particular transition is that the expectant mother is very active in anticipating this transition and constructing a new identity for herself. Moreover, the specific information acquired and the expectations formed at this point will influence how she subsequently adapts, so these changes may be viewed as reflecting a kind of critical period. This conclusion leads to the main questions in this paper: How do individuals perceive and respond to information as transitions unfold? Are there any regularities in these processes across different types of transitions? In what ways do shifts in orientations to information occurring across a transition influence the changes often observed at the end? These are the questions considered next.

111. A Phase Model of Transitions

The present model of transition phases is shown in Table I. The basic premise of the model is that different phases of transitions can be identified that are associated with distinct orientations toward social information. The analysis is organized around three core phases proposed to occur in most transitions (Construction, Consolidation, and Integration, as shown across the top of the table). At each of the three phases, an individual’s orientation toward information shifts in predictable and meaningful ways. As shown in Table I, each phase is characterized by different levels and types of knowledge representations and differences in three types of cognitive-motivational orientations: information seeking, information processing, and impact. The first orientation variable refers to level of interest in and focus of social information seeking. Information processing refers to how new information is perceived and remembered. Impact refers to the importance and affective significance of this new information and its application to self-regulatory and interpersonal behaviors. Table I1 illustrates these variables in terms of the transition to motherhood. Phase 1, Construction, is proposed to occur once an individual enters a new psychological situation in which his or her old categories and expectations may not apply, such as becoming pregnant for the first time. Knowledge is low and generally limited to superficial, observable aspects of the topic (e.g., that parenting involves learning numerous caretaking activities). This phase is characterized

167

PHASE MODEL OF TRANSITIONS TABLE I PHASEMODELOF TRANSITIONS Transition phases

Variable

Phase 1 (Construction)

Phase 2 (Consolidation)

Phase 3 (Integration)

Representation

Level of knowledge Type of knowledge

Information seeking (Level of interest and focus)

Information processing (Accessibility, organization, and memory) Impact (Importance, affective significance, and application)

Low

Med-High

High

Superficial, procedural, concrete, observable

Structural, abstract, inferential

Individual differences in accessibility and structure

Cognitive-Motivational Orientation Active interest; focus Active interest; focus on defining feaon drawing infertures and proceences, conclusions dures and on relevance to self Based on relevance Based on topic releto topic, but not on vance and specific any specific conconclusion clusion

Low

High

Passive; reduced receptivity; flexible focus, depending on individual differences Same bases, but increased individual variation

Variable, depending on context and individual differences

by active information seeking to construct new images, as illustrated by the ups and downs of the information-gathering data shown in Fig. 1. Phase 2, Consolidation, is reached when the fundamental knowledge has been acquired and the individual attempts to draw some conclusions about the topic and, when relevant, how it applies to the self (e.g., what will I be like as a mother?). Because the basic content has been mastered, knowledge representation has an organizational structure that allows inferences about more abstract properties associated with the topic (e.g., areas of competence associated with mothering). Given the high level of uncertainty at both of the first two phases (first about the nature of the topic, then about inferences to be drawn), the individual has a high level of interest in gathering information at both times. In order to support the new conclusions being formed, however, orientations toward information are distinctly different at the two phases. Whereas the first phase is characterized by a general interest and openness to information (e.g., a willing-

168

DIANE N. RUBLE TABLE I1 EXAMPLE OF PHASES USING THE TRANSITION TO FIRSTMOTHERHOOD Transition phases Variable

Level of knowledge

5 p e of knowledge

Information seeking

Information processing

Impact

Phase 1 (Construction)

Phase 2 (Consolidation)

Phase 3 (Integration)

Representation Incomplete often inExtensive knowledge Mothering knowledge accurate knowlof mothering near asymptote edge of mothering Specific behaviors Inferences about Individual differences and stereotypes traits and abilities in mothering sche(e.g,, feeding; nurin relation to self mata turing) Cognitive-Motivational Orientation Active information Active information Little information seeking focused on seeking, focused seeking; unreceptive to inconsistent defining mothering on forming expectation about self as information mother Information perceived Information perceived Individual differences and remembered in and remembered in in processing information in terms of terms of applicaterms of personal tion to mothering conclusions about mothering mothering Little affective or beMothering informaReactions depend on havioral reactions tion has affective individual differto information significance; beences in mothering havior monitored self-schema about mothering to be consistent with new conclusions

ness to consider information that suggests I may have problems in the mothering role), information seeking becomes more focused during the second phase as conclusions are formed. Perception and memory processes are also expected to shift in relation to growing knowledge about the topic. As the concept or experience becomes defined at Phase 2, information processing is likely to become schema driven (Bjorklund, 1987; Fiske & Taylor, 1991) so that information is organized in terms of existing conclusions and information consistent with these conclusions is more easily retrieved. To illustrate, information inconsistent with the mothering conclusions already formed may be ignored or forgotten. A woman who has

PHASE MODEL OF TRANSITIONS

169

decided to quit her job and be a full-time mother, for example, may show preferential attention and memory for information showing deleterious, as opposed to positive, effects of child-care facilities. In contrast, during the Construction phase, the same woman may have been equally responsive to the two kinds of information. Consistent with this general idea, additional analyses of the Deutsch et al. (1988) data shown in Fig. 1 suggest that women’s global maternal self-definitionsdo not change as a function of new information received after the birth of the baby. Finally, the impact of topic-relevant information varies across phases. During the Construction phase, information is primarily relevant to definition and has few self-relevant implications. Thus, new information serves only to enlarge and modify the data base and has little affective significance, even if the information implies something negative about becoming a mother. Once the individual is engaged in drawing inferences about how the topic applies to the self, however, some kinds of information will be more welcome than others. Moreover, investment in the conclusions being drawn may lead those conclusions to become imperatives (e.g., believing that there is only one way to be a good mother), and behavior may be modified to match these conclusions (e.g., a woman may attempt to appear more nurturing in interpersonal interactions). Phase 3, Integration, refers to processes that maintain and elaborate upon the conclusions drawn, allowing the individual to integrate them with other identities and to function more flexibly in terms of them. Although the knowledge base continues to grow, new information is not actively sought, and information that could change conclusions is resisted, though not necessarily for defensive reasons. As noted in Ruble and Frey (1991), this is a time when individuals may avoid discrepant information strategically in order to maintain goal-directed activity. Continued emphasis on evaluation, for example, may undermine mastery efforts (Deci & Ryan, 1987). Because the basic knowledge base and conclusions remain, information processing should still be schema driven, though individual differences in knowledge and specific conclusions drawn should contribute to variations in schematic processing. In addition, individual differences in knowledge, conclusions, and external constraints or opportunities should contribute to variations in goals at this time, which in turn influence the impact of relevant information. In brief, then, individual differences in “schematicity” (Bem, 1984; Markus, Crane, Bernstein, & Sidali, 1982) or importance (Costanzo & Fraenkel, 1987) may determine the extent to which inconsistent information is resisted, negative information is distressing, or deviations from expectations are tolerated. Relative to the Consolidation phase of late pregnancy, for example, the likelihood that a woman at this phase would interpret new information about herself in terms of mothering depends on the relative significanceof mothering in her selfdefinitional hierarchy (Deaux, 1991). A major aim of the current article is to suggest that this kind of transition

170

DIANE N. RUBLE

analysis may be a useful way to conceptualize certain kinds of social-cognitive developmental changes and their implications for behavior. As children acquire new social concepts, related information may become personally significant for the first time, just as becoming a mother introduces new, personally significant categories to a woman. This idea that the transition to parenthood is in any way similar to socialcognitive transitions may seem surprising. After all, the former involves fundamental life changes in social roles, whereas the latter involves internal changes in cognitive structures or social knowledge. Of course, there are many obvious disparities between these different types of transitions. The present argument is that there is also a fundamental similarity, involving individual interpretive and epistemological features. In particular, different types of transitions are similar because one’s existing knowledge structures are no longer the best guides for effective behavior, and new guides need to be constructed. How or why the transition begins may not matter. The distinguishing characteristic is the realization of a new category or event with social regulatory implications. Whether the realization is impending motherhood or the importance of gender distinctions, it changes how individuals must deal with the world and motivates them to find out more about it. At this point, individuals proceed through the transition phases, represented by predictable shifts in orientation toward relevant information.

IV. Theoretical Background The model of transition phases being proposed builds upon a number of previous conceptions of transition phenomena and the association between knowledge acquisition and information processing. Related conceptions can be found across diverse literatures, including cognitive development, social psychology, and sociology. The discussion of these ideas in this section is illustrative rather than comprehensive, with the intent of placing the present model in context.

A. COGNITIVE-DEVELOPMENTAL PERSPECTIVES The literature on cognitive development has been concerned with a number of similar questions: What instigates transitions in levels of cognitive development? What are the mechanisms of change? How does level of knowledge or structure influence information processing? How does the individual move from less to more advanced knowledge states? Despite the large and diverse literature describing the acquisition of increasingly sophisticated rules, skills, or structures

PHASE MODEL OF TRANSITIONS

171

during development (e.g., Piaget, 1971; Siegler, 1981), there is little agreement on the mechanisms of change (Acredolo & O’Connor 1991; Nelson, 1986). The generality of the mechanisms proposed by Piaget (assimilation and accommodation) has been met with considerable dissatisfaction, yet only a few attempts to describe more specific mechanisms have been offered (Keil, 1984; Nelson, 1986). Thus, a clear answer to the question raised by Connell and Furman (1984) about the mechanisms of social transitions does not appear in the extensive literature on cognitive developmental change. There is a point of agreement that is relevant, however. The need for change is instigated by incongruity or dissonance, termed disequilibrium by Piaget (1975/ 1985). Specifically, cognitive growth is promoted by the recognition of cognitive conflicts or contradictions, which leads to a state of uncertainty, instability, and possibly anxiety (Acredolo & O’Connor, 1991; Berlyne, 1970). Such a state may occur for a number of reasons. It may be generated either internally via the recognition of incompatible cognitions (Berlyne, 1970; Keil, 1984) or externally during social interaction (Doise & Mugny, 1984; Doise & Palmonari, 1984). The latter is particularly relevant to many types of life transitions because such shifts are likely to alter the probability of encountering people whose viewpoints differ from one’s own (Doise & Palmonari, 1984). Internal cognitive conflict may also come in many different forms (Acredolo & O’Connor, 1991; Keil, 1984). Keil (1984), for example, describes several types of mechanisms that can initiate cognitive change and indicates how the exact nature of the mechanism depends upon the content of the knowledge it is influencing (e.g., acquisition of metaphor; change in grammatical structures). There appear to be sufficient similarities in the different mechanisms, however, to propose some general principles regarding uncertainty mechanisms and cognitive change. In some domains, uncertainty may take the form of a sense of inadequacy of present reasoning (e.g., that a particular grammatical form does not work or that existing assumptions about reality are contradicted). In other domains, it may take the form of a sense of possibili9, as when knowledge of a particular domain has become sufficiently differentiated to reveal new relations to other domains or new meanings about the essential features of that domain. The interesting question is whether these same two principles may represent the mechanisms instigating non-cognitive transitions, such as major life transitions. Certainly, both the sense of inadequacy and the sense of possibility could stimulate the information gathering and mental activity (Acredolo & O’Connor, 1991) that characterize the first (Construction) phase of transitions, as we have defined it. In either case, the individual is motivated to resolve the uncertainty, and the resulting cognitive activity leads to new levels of cognitive understanding. In addition to providing guidance about what initiates a transition, some recent research in the cognitive-developmental literature is relevant to considerations of mechanisms mediating change across transitions. Most notably, recent litera-

172

DIANE N. RUBLE

ture has pointed to the impact of knowledge base (e.g., familiarity; expertise) on a number of information-processingindices (Bedard & Chi, 1992). If, as argued above, transitions are instigated by attempts to resolve uncertainty by gathering information, then an individual’s knowledge about the content involved may increase quite rapidly during the transition. Thus, changes that occur across a transition may be mediated, at least in part, by the influence of a changing base of knowledge on memory and inferential processes. Considerable evidence supports the idea that knowledge base has wideranging effects, independent of age. Some research shows that level of knowledge affects ways of representing category-relevant information (Chi & Ceci, 1987; Keil, 1984). Keil (1984), for example, suggests that the characteristic-todefining shift, in which children progress from an exemplar-based to a rule-based representation of a category (Vygotsky, 1934/1962), may be more related to level of knowledge about the category than to age-related structural change. There is now sufficient evidence that the shift occurs at different times for different concepts, and not at a particular age or stage. Other research shows that memory processes change as a direct function of knowledge base (Bjorklund, 1987; Schneider, 1990). Bjorklund (1987), for example, argues that increased knowledge results in more efficient processing of a particular content as a result of the frequency with which that content has been activated in the past and the number of features and inter-item connections associated with it. Because of the increased processing efficiency, a child’s limited capacity can be allocated more to the execution of other cognitive operations, including deliberate and sophisticated memorizing strategies (Omstein & Naus, 1985). Indeed, traditional age differences in memory may be eliminated when knowledge base is taken into account (Chi, 1977). Taken together, these findings suggest that regardless of age, as an individual proceeds across a transition, incoming information is more likely to be transformed at encoding in relation to the emerging semantic organization, and items that are part of that organization are more easily stored and accessed. B . SOCIAL-COGNITIVE AND MOTIVATIONAL PERSPECTIVES Several lines of research in the social cognition literature provide guidance about the kinds of cognitive-motivational changes likely to occur across phases of a transition. In this section, four such lines of research are discussed as illustrations: category accessibility, schematic processing, level of engagement with information as a function of epistemic goals, and type of information of interest as a function of epistemic goals. The first two research areas are particularly relevant to the information-processing variable in the model, whereas the

PHASE MODEL OF TRANSITIONS

173

last two are relevant primarily to level and focus of information seeking and secondarily to impact.

I . Category Accessibility Considerable evidence suggests that the initial interpretation of information is affected by expectations or accessible schemata (Higgins & Bargh, 1987). If subjects are primed with a particular trait construct, even subliminally, they are likely to use that construct in interpreting subsequent information received in a different context (Bargh & Pietromonaco, 1982; Higgins, Rholes, & Jones, 1977). More relevant to the present model, one’s current goals or needs can also increase the accessibility of particular constructs (Bruner, 1957; Higgins & King, 1981). Individual differences in construct accessibility (e.g., with respect to gender or particular-person characteristics) lead to a greater sensitivity to the presence of such information, regardless of whether or not that construct was initially primed in that context (Higgins, King, & Mavin, 1982). This literature would thus lead to the prediction that incoming information will be “seen” or interpreted in terms of constructs relevant to an ongoing transition, at least when the information is ambiguous. A woman expecting her first child, for example, may focus on the relationship between mothers and children when seeing movies or reading novels, even if those relationships are irrelevant to the development of the story.

2 . Schematic Processing Research based on the concept of the cognitive structure or schema (Bartlett, 1932; Bruner, 1951; Neisser, 1976) has generated a considerable amount of relevant research. In a manner similar to the ideas discussed above with respect to knowledge base in the developmental literature, research in social cognition has shown that highly developed knowledge structures (schemata) represent interpretive frameworks that influence how social information is perceived, encoded, and remembered (Higgins & Bargh, 1987; Markus & Zajonc, 1985; Taylor & Crocker, 1981). Although there are a number of different schema models, they all agree that information relevant to existing schemata are more easily processed and remembered. There has been considerable debate, however, concerning the processing of schema-congruent versus incongruent information (cf., Alba & Hasher, 1983; Graesser, 1981; Hastie & Kumar, 1979). Information inconsistent with the schema may be filtered or ignored, leading to poorer memory, or it may be processed more extensively in order to resolve the discrepancy; evidence in support of both effects has been found (Higgins & Bargh, 1987). Recently, suggestions for resolving this debate have been offered that have direct relevance to changes in information processing that might be anticipated

174

DIANE N. RUBLE

across a transition. Specifically, schema-congruent processing would be expected to increase for more developed or stronger schemata (Fiske & Neuberg, 1990; Higgins & Bargh, 1987; Ruble & Stangor, 1986). Thus, one possible reason for discrepancies in findings across studies is that some studies examine the effects of expectations based on a well-established schema such as gender (Ruble & Stangor, 1986), whereas others examine the effects of short-term experimentally manipulated expectations (e.g., Hastie & Kumar, 1979). Because the former should represent stronger and more developed expectations than the latter, schema-congruent processing should be higher in the former than in the latter. Indeed, Higgins and Bargh (1987) suggest that many of the studies manipulating expectations might properly be considered to be studying impression (or schema) formation rather than impression testing. Thus, it may be suggested that memory processes shift as a function of the level of development of a schema from favoring schema-inconsistentto schema-consistent information, an idea of obvious importance for identifying changes in information processing across phases of a transition. Several recent studies support this hypothesis. In a study that experimentally manipulated the “development” of a schema by providing greater or fewer instances of schema-relevant behavior, schema-congruent memory was found to be stronger for subjects with a more developed schema (Stangor & Ruble, 1989b). Schema-congruentprocessing also appears to be greater in older than in younger children with respect to a content (gender) that clearly develops with age (Stangor & Ruble, 1989a). Finally, a recent meta-analysis provides strong support for this hypothesis (Stangor & McMillan, 1992). Increases in both expectancy strength and age were associated with a tendency to recall expectancy-congruent information. In order to consider how these findings might apply to an analysis of transitions, it is important to understand why information processing changes in relation to schema-strength development. So far, the focus on cognitive structures implies that such changes represent a relatively automatic and pervasive cognitive process by which a schema provides meaning to information to make it more comprehensible and to facilitate memory. Moreover, schematic processing is adaptive because it makes information processing more efficient, thereby freeing the individual to engage in more complex cognitive tasks. Thus, one set of mechanisms that account for schematic processing effects appears to be inherent to the functioning of the structure itself, so that irrelevant or inconsistent information is discounted or never “seen” or processed (Fiske & Taylor, 1991; Higgins & Bargh, 1987). More motivational mechanisms may also be involved. There is considerable evidence that the tendency to engage in theory-based (schematic) versus databased processing varies as a function of processing goals (Fiske & Neuberg, 1990; Higgins & Bargh, 1987). When people are highly motivated to be accu-

PHASE MODEL OF TRANSITIONS

175

rate, for example, they are less likely to rely on previously formed categories or theories and instead will expend the effort to carefully process the data (Chaiken, 1980; Erber & Fiske, 1984). One possible reason for the differences in information processing observed for short-term versus well-formed expectancies is that an individual’s processing goals depend on the level of development of relevant schemata (Ruble, 1987; Stangor & Ruble, 1989a). During the initial stages of expectancy formation, individuals have relatively little knowledge and considerable uncertainty. Processing goals are thus likely to be more oriented toward accuracy, very much in an “effectance” or mastery-motivation sense, and individuals should engage in careful processing of inconsistent information in an attempt to reconcile it with the developing expectancy. In contrast, once expectations are well established, such effort is less justified because the uncertainty has been resolved. Some direct evidence for this latter contention was provided by Stangor and Ford (1992), using an information-seeking paradigm. Perceivers with stronger expectations showed an interest in avoiding exposure to expectancy-incongruent information. In short, changes in knowledge representation across phases of a transition would be expected to change information processing for both cognitive (schematic processing) and motivational reasons. First, knowledge structures impose their own order on the perception and memory of incoming information so that once a structure is in place, consistent information is more easily processed and remembered. Second, as individuals acquire enough knowledge to form conclusions, goals shift from trying to resolve uncertainty to trying to confirm conclusions. Thus, processing information inconsistent with current expectations should decline across phases of a transition. 3. Level of Engagement with Information as a Function of Epistemic Goals

Motivational processes have been directly implicated in a related aspect of relevant social-cognitive research-how orientations toward information change in relation to one’s knowledge base and goals. One general conclusion that seems to hold across a variety of analyses is that the level of seeking and engagement with information changes as a function of level of certainty. For example, Festinger’s (1954) original statement of social comparison emphasized the concept of uncertainty-that people are motivated to compare themselves with others in order to resolve uncertainties about themselves. More generally, Trope (1983) has shown that uncertainty drives a search for diagnostic information, even if that information is likely to result in unflattering conclusions about one’s competence. In contrast, once a relatively certain conclusion is reached, diagnostic information search is reduced. In addition, several findings support dissonance-theory predictions that once a commitment to a particular point of

176

DIANE N. RUBLE

view has been made, active information seeking is reduced and nonsupportive information avoided (Frey, 1986). Similarly, a program of research on making and acting upon decisions shows that a predecision mind set is characterized by an open, impartial approach to information, whereas a postdecision mind set is more closed and biased toward an optimistic analysis of relevant information (Gollwitzer, 1986; Heckhausen & Gollwitzer, 1986). Also, research on confirmation biases suggests that a clear, relatively certain set of beliefs about oneself or others often leads to information seeking that is biased toward confirming those beliefs (Swann, 1986). Finally, as Abelson (1986) suggests, “Beliefs are like possessions”; once one has grown comfortable with a set of ideas, as with an old lounging chair, one may be reluctant to try anything else. Such findings are directly pertinent to changes across transition phases because almost by definition, knowledge and certainty are relatively low in the initial phase but should both increase as the transition proceeds. Most recent models dealing with the concept of uncertainty and its implications for orientations toward information portray the individual as an active seeker and constructor of meaning, strategically selecting information to maximize diagnostic value. Such a portrait is consistent with theories that depict human nature as intrinsically driven toward mastery of the environment (Berlyne, 1966; Deci & Ryan, 1985; White, 1959). Not surprisingly, then, the driving force underlying uncertainty effects are characterized by such terms as need for control and predictability, mastery, insufficient confidence, and need for closure. Four models of epistemic activity, oriented toward determining different types of social judgments, nicely illustrate this point. Chaiken and her colleagues (Chaiken, Liberman, & Eagly, 1989; Maheswaran & Chaiken, 1991), for example, emphasize the principle of the su#ciency of judgmental confidence. They argue that the efficient information processor must balance the need to conserve processing energy (i.e., the idea of the cognitive miser; see Fiske, 1993) with the need for judgmental confidence. Thus, according to this model, changes in the confidence sufficiency threshold, rather than in level of confidence per se, are responsible for the well-documented finding that information is utilized more extensively and systematically when motivation is enhanced by making the judgment personally relevant or stressing the need for accuracy (e.g., Cacioppo, Petty, & Morris, 1983; Neuberg & Fiske, 1987). In other words, contextual increases in motivation raise the sufficiency threshold, and processing efforts must continue longer in order to reach it. Kruglanski and his colleagues (1990; Kruglanski, Peri, & Zakai, 1991) argue that epistemic activity is a function of the need for closure. Conditions of complete uncertainty or ambiguity should promote epistemic “unfreezing” (enhanced information seeking and hypothesis generation). In- contrast, the generation, or ready availability, of a plausible basis for judgment promotes epistemic “freezing” (a reduction in the generation of further hypotheses and the need for

PHASE MODEL OF TRANSITIONS

177

further information seeking). For example, Jamieson and Zanna (1989) showed that under time pressure (which promotes the need for closure), individuals were relatively more inclined to base judgments on existing stereotypes rather than on new information. Recently, Kruglanski et al. (1991) provided direct support for the model by manipulating both confidence and need for closure and by showing that individuals high in need for closure tended to search for information more when confidence was low than when it was high. Trope (1983; 1986) emphasizes predictability and control (utility value of information for making decisions about the future) in his model of epistemic activity with respect to achievement-related behavior. Complete and realistic self-knowledge allows individuals to form attainable goals, including withdrawal from activities for which they are ill-suited. An information search attempting to assess accurately one‘s abilities improves chances of predicting which activity and degree of time expenditure will yield the highest expected utility. According to the model, the value of predecision information equals the resulting improvement in expected utility, much like a cost-benefit analysis, which is a function of the importance of the outcome and the uncertainty as to whether it will be achieved. To the extent that ability level is known, the value of diagnostic predecision information seeking declines relative to other possible goals, such as maintaining self-esteem. Finally, Pittman and his colleagues (Pittman & D’Agostino 1985; Pittman & Heller, 1987) argue that control motivation drives epistemic activity. Individuals who have been deprived of control spend greater effort in information gathering, prefer diagnostic information, and process it in a more careful and deliberate fashion than individuals who have not experienced a loss of control. In addition, information likely to elicit concerns about control, such as negative or unexpected information, also evokes a more extensive analysis. This basic idea is supported, as well, by other types of analyses suggesting that control motivation may underlie individual differences in need for cognition (Thompson, Chaiken & Hazlewood, 1993). In brief, social-cognition theory and research suggests that level of engagement with information changes in relation to basic motives and needs that are likely to show important shifts across a transition, including confidence in one’s judgments, need for closure, adequacy of information for future decisions, and sense of control. Although the models deal with different content and emphases, they converge around similar themes. First, there are costs involved in gathering information and effortful processes, so that particular motivational contexts are required. Second, those motivational contexts involve subjective uncertainty, lack of confidence, and loss of control. Finally, the epistemic activity is reduced as subjective certainty, confidence, and control are regained, particularly given competing contexts (e.g., need for closure) and motives (self-enhancement).The argument with respect to the present phase model is that these motivational

178

DIANE N. RUBLE

processes change systematically across phases of a transition. They thus provide a basis for predicting’that level of engagement with topic-relevant information will shift from high interest and careful, deliberate processing to reduced interest and less extensive processing. 4 . Type of Information of Interest as a Function of Epistemic Goals

Different goals not only affect how actively information is sought and considered; they also affect what kind of information is of interest. Fazio (1979), for example, has shown that the nature of interest in others’ opinions changes in relation to the stage of formation of an attitude. In the early stages, our interest in others’ opinions is concerned with developing a conception of the object or issue in question. At a later time, the goal of information seeking changes to confirming the image that has been constructed (see also Olson, Ellis, & Zanna, 1983). Gollwitzer (1986) argues that the type of information of interest varies as a function of one’s particular “mind-set’’ associated with the phases of making or acting upon a decision. In the predecisional phase, a “deliberative” mind-set should lead individuals to be most interested in information concerned with the expected benefits (or costs) of an action decision; in the preactional phase, an “implemental” mind-set should enhance interest in information concerned with planning; and in the postactional phase, an “evaluative” mind-set should lead to an interest in the quality of the outcome. Trope and Thompson (1993) have demonstrated that people with well-developed expectancies or knowledge structures seek more extreme information, presumably because such inquiries are more diagnostic for them. Several examples of different orientations toward information in relation to disparate goals may be found in the literature on achievement. A number of researchers have examined differences between accurate self-assessment as opposed to self-enhancementgoals and their consequences for type of information sought and utilized (Ruble & Frey, 1991; Trope, 1986; Wood, 1989). The former goal, for example, might lead to an interest in choosing diagnostic tasks or selecting similar others or standard setters for social comparisons, whereas the latter goal might lead to an interest in selecting easy tasks or social comparisons with others who are worse off. Other researchers have examined how mastery or learning goals versus self-evaluation goals affect the selection and use of information (Butler, 1992; Dweck & Leggett, 1988; Nicholls, 1984; Veroff, 1969; Wood, 1989). The former goals are more likely to be Concerned with temporal comparison indicating progress or information about optimal task procedures, whereas the latter goals are more likely to be concerned with comparison of outcomes with similar others. Recently, Ruble and Frey (1991) have proposed that such goals are temporally organized so that level of knowledge and experi-

PHASE MODEL OF TRANSITIONS

179

ence with a particular skill determines which goals are paramount and thus what kinds of information are of greatest interest. To illustrate, individuals need to determine the key features and procedures of a skill by engaging in upward comparison (standard setting and observational learning) before engaging in similarity comparison, because they need to define the key parameters of the task before accurate competence assessment can be meaningful. Such temporal changes in goals may be directly applied to the Construction versus Consolidation phases of the present transition model. That is, initial goals to define the new category or event give way to goals of supporting the conclusions that have formed, and the type of information of interest is likely to change in relation to these different goals.

C. SUMMARY There is substantial convergence across diverse literatures about the cognitive and motivational consequences of differences in what we assume to be two critical and inter-related features of transitions: (1) level of uncertainty and associated processes (loss of confidence and control); and (2) level of knowledge. Uncertainty is high at the beginning of a transition and subsequently declines, whereas the reverse is true for level of knowledge. Indeed, much of the cognitive-developmental literature suggests that some form of uncertainty instigates a transition. Although knowledge and uncertainty might be considered similar variables and are obviously related, they are often treated separately in the literature, in part because uncertainty and its associated processes (loss of control) have clear motivational qualities that level of knowledge does not. What does this literature suggest about the cognitive and motivational variables included in the phase model of transitions? With respect to information seeking, research deriving from vastly different models and perspectives converge on a clear conclusion: Uncertainty elicits an active engagement with relevant information, but once uncertainty is resolved, active epistemic motivation declines. In Table I, active information seeking is depicted at both Phase 1 and Phase 2 because uncertainty is high in both phases, though with respect to different issues and goals. What appears to change is the type of information of interest. A number of lines of research suggest that interest changes with time from definitional concerns (What are the characteristics associated with this category?) to confirmation or evaluation (Was it done correctly? How good was it?). With respect to information processing, most researchers agree that increasing levels of knowledge lead to more efficient processing and certain associated biases in encoding and retrieval of information. An important point for the transition model is that such changes in knowledge and expertise (and corre-

180

DIANE N. RUBLE

sponding effects on processing) can occur at any age. In general, both the cognitive-developmental and social-cognitive literature suggests that memory is increasingly likely to favor information consistent with developing expectations and conclusions as the transition proceeds. In addition, assuming that the topic of a transition is personally significant to the individual, the literature on category accessibility would suggest that this topic will be used to interpret incoming information. Thus, as shown in Table I, the transition model predicts that information will begin to be categorized in terms of the topic and that any information relevant to it will be better remembered than irrelevant information as soon as the transition begins. Finally, impact variables are assumed to change in parallel with information seeking and processing. The exact nature of the individual’s choices, inferences, motivation, and attitude change depend on what information is sought and how it is processed. Moreover, changes in level of uncertainty across phases is associated with changing goals that should be reflected in affective and behavioral reactions. Need for closure and control, for example, should be paramount at Phase 2 as the individual is consolidating conclusions. Thus, the individual will be concerned with verifying conclusions by behaving correctly with reference to them. In addition, whether or not new information challenges current conclusions has affective implications, both because of changes in personal relevance and because the comfort or satisfaction of reaching closure may have to be relinquished.

V. Application to Social-Cognitive Transitions

I have argued that the present transition phase model may be a useful way to conceptualize certain kinds of social-cognitive developmental changes and their implications for behavior. Although externally imposed life transitions, such as becoming a parent or beginning junior high school, are obviously not the same as cognitive and social-cognitive transitions, there are some notable similarities that, if highlighted, may be of heuristic value to both literatures. In both cases, conflicts and uncertainties change an individual’s way of viewing the world, and different types of information become salient and important. The present proposal is that this change in orientation toward information proceeds in a predictable sequence, regardless of the reason for the initial discontinuity. Here, this point will be illustrated in terms of three social-cognitive transitions: (1) the understanding of one’s identity as a male or female (with a focus on gender constancy), (2) the understanding of one’s (school-related) abilities, and (3) the understanding of individual differences in people’s characteristics (with a

PHASE MODEL OF TRANSITIONS

181

focus on stable, dispositional traits, such as helpfulness, meanness, competence).

VI. The Gender Transition Even fairly brief observations of young children will often result in charming illustrations of their attempts to construct the meaning of gender categories. Consider the observations of Vivian Paley, a teacher of preschool and kindergarten children who records the things her children say and who keeps careful notes. Paley (1984) writes that kindergarten children search everywhere for clues to create clear and distinct conceptions of girls and boys. It does not seem to matter whether they match conventional gender stereotypes, only that there are clear and consistent differences. In one of her kindergarten classes, for example, boys hop to get their milk, whereas girls skip to the paper shelf. Boys clap out the rhythm of certain songs; girls sing louder. One day the teacher asks, “How would you feel if Goldilocks came to your house to play?’ The first girl welcomes the prospect. The next eight respondents are girls, who also nod their approval. The first two boys in the circle also agree. Andrew can barely wait for his turn. “No way!” he yells. The boys jump in response. Paul and Jeremy laugh loudly. “She better not, Oh, no!” The first boy then waves his arms. “I changed my mind!” Kohlberg (1966) initially proposed that such attempts to organize gender categories followed an important social-cognitive development, gender constancy, which he related to Piaget’s notion of conservation. This new understanding, which occurs at 4 to 6 years of age in most studies, involves the realization that one’s sex is constant across time, situation, and superficial changes such as hair length or type of clothing (see Stangor & Ruble, 1987, for a review). Gender constancy is measured by asking children a series of questions about their own sex or that of hypothetical other children under a variety of situations (e.g., “When you were little, were you a baby boy or a baby girl?’; “If Jack was wearing a dress, would he be a boy or a girl?”) (Slaby & Frey, 1975; Wehren & De Lisi, 1983). Prior to understanding gender constancy, when asked whether Jack, who is now wearing a dress, or is playing with dolls, is a girl or a boy, young children are likely to say he is a girl. Although it is reasonably clear that knowledge of gender constancy occurs in a predictable developmental progression across cultures (De Lisi, & Gallagher, 1991; Munroe, Shimmin, & Munroe, 1984), much as Kohlberg described, it is also quite clear that children engage in sex-typed behavior and have considerable knowledge of gender stereotypes long before they understand gender constancy (Huston, 1983). Thus, gender constancy does not seem to initiate children’s

182

DIANE N. RUBLE

organized representation of gender. Instead, the onset of the “gender transition” is marked by children’s recognition that gender is an important dimension in their immediate social world. Just as a pregnant woman needs to construct an image of what it means to be a mother, these children need to construct an image of this categorical distinction in gender. Nevertheless, we would argue that the attainment of gender constancy is an important transition phase (Stangor & Ruble, 1987). Knowledge that one’s gender is immutable affords a new power to gender categories. Rather than just being one of many dimensions of the social world to which children might pay attention, it takes precedence. In terms of the present model, it reflects the second (Consolidation) phase. Children not only know about it but they also have to get it right; they have to consolidate their knowledge and apply it to their behavior to make sure it is sex appropriate. Thus, we should expect to see a focus on information relevant to gender-appropriate behavior and a greater value given to such information. Studies of attention to same-sex models support this idea. In a pioneering study of gender constancy, Slaby and Frey (1975) examined 35-year-old children’s attention to same-sex models shown on the right or left side of a movie screen. A man and woman were shown engaging in a number of activities (e.g., playing music, popping corn, drinking juice), and children’s point of fixation on the screen was observed. The data supported the hypothesis that attention to same-sex models would be greater for children who had attained gender constancy, though the differences were significant for boys only. Gender constant boys looked at male models more than preconstant boys, who looked at the two models an equal amount of time. Constant girls also looked at female models somewhat more, as expected, but not significantly. Why boys only? Girls may be less motivated than boys to adopt sex-typed behavior because they encounter less strong norms-that is, broader definitions of sex appropriateness or weaker prohibitions. Consequently, gender constancy may be less of a transition phase for girls. In addition, there was an overall tendency for both sexes to look at male models somewhat more than half of the time. Perhaps there are competing motives for girls, pitting social power against same-sex modeling. This sex difference does not always occur, however. In the study described next, boys and girls showed a similar pattern. Does the greater attention to same-sex models associated with constancy have any impact on behavior? In one study (Ruble, Balaban, & Cooper, 1981), 46-year-old children viewed a commercial in the middle of a cartoon; the comrnercia1 showed either two boys or two girls playing with an attractive toy pretested to be gender neutral (a movie viewer). Subsequently, the children had a chance to play with a set of toys, one of which was the same one they had seen in the commercial. The amount of time spent with this toy was observed from behind a one-way mirror. Only gender-constant children were differentially affected by the sex of the model. Consistent with predictions, gender-constant children seemed

PHASE MODEL OF TRANSITIONS

183

to avoid the toy when other-sex children had played with it, whereas pre-genderconstant children played with the toy regardless of whether boy or girl models were shown in the commercial. Because age was a covariate in these analyses, the differences are not a simple reflection of age. Perceived sex appropriateness of the toy showed the same pattern. The children were asked, “Who would like this toy more, my brother or my sister?” Gender constant children responded in accordance with the information presented in the commercial, whereas pre-gender-constant children did not differentiate. Finally, it is interesting that a standard measure of sex stereotyping showed no difference as a function of constancy. Both pre-gender-constant and constant children labeled toys in accordance with stereotypic labels. This pattern of results highlights the distinction across different phases of a gender transition. Once children realize the significance of gender categories, they begin to construct the meaning associated with gender, and may even develop quite sophisticated sets of gender stereotypes at a fairly young age. Moreover, differential experience with and reinforcement for same-sex activities may lead to sex differentiated behavior at quite a young age. In contrast, the Consolidation phase, as represented by gender constancy, seems associated with an active, self-initiated focus on information regarding gender-appropriate activities, and self, rather than other, regulation. To illustrate, consider the actions of one of the subjects in the study: One preconstant boy labeled the dish set female, but then asked to play with it. In terms of the model, the impact of gender information was low; it did not dictate how he himself should be behaving, as it would at the Consolidation phase. The label was not viewed as a prohibition against an otherwise attractive behavior-a point discussed later with respect to a subsequent study. I should note at this point that the conclusion about the effects of gender constancy being proposed here is controversial. Many studies have shown no association between constancy and toy preference or memory for gender-related information (Bussey & Bandura, 1992; Levy & Carter, 1989; Marcus & Overton, 1978). These findings would seem to contradict the present hypothesis that constancy is associated with changes in orientations toward gender information. How do we explain these apparent discrepancies? One clue comes from a clever study by Eaton, Von Bargen, and Keats (1981) who varied two dimensions of toy choice. Two identical boxes were presented to young children; they were told that each box contained a toy and were asked to indicate which one they would like of the following pairs: ( I ) Boys play with this toy; girls play with this toy; (2) Kids run with this toy; kids sit with this toy. After determining the child’s preference for running or sitting, the preferred mode was placed in conflict with the child’s sex (e.g., girls run with this toy; boys sit with this toy). The differential interest of gender-constant children was shown clearly when

184

DIANE N. RUBLE

the two dimensions were in conflict. Constant children overwhelmingly selected on the basis of gender; preconstant children used the other dimension about equally often. When there was no conflict, both high and low constant children selected the sex-appropriate toy. These data suggest why some studies fail to show a same-sex bias associated with gender constancy. That is, ideas about salience of gender information can be tested only when there are conflicting dimensions. We have recently replicated these findings in a study that covaried age and thereby clearly linked the differential choice directly to gender constancy (Stangor & Ruble, 1989a). When gender appropriateness was the only dimension, all of the children chose the toy associated with their gender; when the preferred activity conflicted with the gender cue, gender-constant children were more likely to select on the basis of gender than were non-constant children, controlling for age. These data raise a more general statement of what we have labeled the “conflict” hypothesis in a recent study (Frey & Ruble, 1992). The idea is that it is only when children understand gender constancy that they become willing to incorporate negative dimensions of gender into their choices and behaviors-that in consolidating a gender identity for themselves, it is more important to be “correct” than to select the more attractive option. To examine this hypothesis, we varied toy attractiveness and sex “appropriateness.” As part of a cartoon show, children saw a program host (Bonzo the Clown) interviewing two boys and two girls about their toy preference. In the conflict condition, the less attractive toy was preferred by same-sex models. In the noconflict condition, the two toys were equal in attractiveness. We predicted, first, that all children would play with the same-sex preferred toy in the no-conflict condition-that is, there would be no constancy effect. Second, in the conflict condition, we predicted that constant children would play with the same-sex preferred unattractive toy, but that preconstant children would not. These predictions were supported, but once again for boys only. In the conflict condition, constant boys played with the same-sex but unattractive toy much longer than did preconstant boys, whereas in the no-conflict condition, both constant and preconstant boys showed considerable interest in the same-sex toy. Taken together, these studies suggest that closer attention to the processes assumed to link gender-related cognitions to behaviors is needed. One reason for inconsistent findings across studies is that children may engage in same-sex play for many reasons. They may exhibit same-sex preferences with little thought because of habits established through early reinforcement histories. Only when a same-sex preference conflicts with other attractive features has a true choice been made that should show a relation to changing knowledge. One other possible reason for inconsistent findings concerns the age of the subjects. Some studies failing to show a relation between gender constancy and

PHASE MODEL OF TRANSITIONS

185

responsiveness to gender information have used a very young sample (mean age of approximately 3.5 years). Why should this matter? There are now several indications that gender constancy shows a curvilinear (U- or J-shaped) function with age, so that 3- or 4-year-old may have higher scores than children one year older (De Lisi & Gallagher 1991; Emmerich, 1982; Gouze & Nadelman, 1980; Wehren & De Lisi, 1983). This decrease in constancy appears to be linked to the kind of explanation children offered for their responses. Wehren and De Lisi (1983) found that 5-year-olds tended to explain their errors in making judgments of constancy in terms of categorical norms (e.g., “If Jack wore lipstick, Jack would be a girl; boys can’t wear lipstick”). It may be that children need to resolve this apparent conflict between sex identity and social-role norms before achieving a true understanding of constancy. Thus, studies examining associations with constancy among very young children may be combining two very different types of constancy, only one of which may be related to an active choice on the basis of gender. If so, the interpretation of failures to find associations between gender constancy and behavior needs to be reconsidered. A related possibility is that some attainments in the acquisition of constancy are more critical than others. Martin and Little (1990), for example, show that beliefs in the stability of gender over time relates more closely to various indices of impact than do beliefs in the consistency of gender across superficial transformations, such as a boy wearing a dress (see also Eaton et al., 1981). Such data are important for understanding differences across studies in the cognitive and motivational consequences of attaining gender constancy. In brief, the adoption of sex-role norms is a kind of transition that occurs between 3 and 7, which is perhaps somewhat more powerful for the choices of boys than girls at that age, though perhaps not in the long run. The four studies described above suggest that the attainment of gender constancy is associated with changes in the focus of information-seeking and impact of gender information. According to the phase model being proposed here, these changes correspond to a shift across the Construction and Consolidation phases. Preconstant children (at the ages included in these studies) are in the Construction phase in the sense that they have developed extensive knowledge of gender categories. The consolidation of that knowledge associated with constancy, however, leads them to make clear and consistent use of that information. Specifically, these four studies show systematic differences associated with level of constancy in attention to same-sex models and self-regulation in terms of those norms. High constant children (Phase 2) look more at same-sex models (first study), play more with the toy that same-sex models advocate (second study), and are more responsive to gender information relative to other kinds of information when selecting play activities (last two studies) than low-constant children (Phase 1). What about the changes in information-processing suggested by the model? Although a few studies have tested the hypothesis that gender constancy is

186

DIANE N. RUBLE

associated with gender memory, there has been no support for this idea (Carter, 1987; Levy & Carter, 1989). One possible reason for these null findings is that the nature of changes in cognitive-motivational orientations across phases is associated with different aspects of knowledge representation associated with the phase. As discussed earlier with respect to general principles of schematic processing, changes in knowledge may affect both processing efficiency and motivation to process. Understanding that gender is invariant may change the personal relevance of same-sex information, with concomitant changes in motivational orientations toward gender (i.e., information-seeking focus and impact). In contrast, the amount of knowledge may affect the ease with which gender stereotypic information is stored and recalled. Stangor and Ruble (1989a) provided support for this hypothesis. As noted above, gender constancy was associated with an impact variable (selecting a same-sex toy, even when it was less attractive), but it was not associated with memory for gender-consistent information. Instead, such memory increased with age, suggesting a closer association to the growth of knowledge about gender, rather than to its personal relevance, per se. Thus, these data suggest that a possible reason for discrepant findings across studies is that different aspects of gender knowledge representation differentially influence particular types of dependent variables. Longitudinal research is needed to link such differences directly to changes in forms of representation (i.e., gender knowledge vs. constancy). Finally, Phase 3, Integration, should be characterized by reduced interest in gender-related information, and greater flexibility in its impact. There are no data directly testing this hypothesis, but some findings provide indirect support. Several studies suggest that the application of gender stereotypes during the early years of school becomes much more flexible by middle to late elementary school (see Ruble, 1988, for a review). Although measures of stereotyping that require a forced choice between males and females show an increase in stereotyping with age, measures that allow for a “both or neither” response often show a decrease in stereotyping after approximately 5 to 7 years of age (Signorella, Bigler, & Liben, 1993; Signorella & Liben, 1985). This pattern of results suggests that although stereotypic knowledge increases, older children may apply this knowledge more flexibly or attach less importance to it. Findings from interview studies lead to a similar conclusion (Carter & Patterson, 1982; Damon, 1979; Stoddart & Turiel, 1985). For example, in the Stoddart and Turiel study, children showed awareness of sex-role violations, such as men wearing dresses, by age 5, and from 5 to 7, they were quite insistent that such acts were wrong. After this age, however, children began to make allowance for individual flexibility in the face of sex-role conventions, at least until adolescence. A similar curvilinear age trend was found in a study of person perception (Berndt & Heller, 1986). When predicting an actor’s future behavior, subjects in third grade paid more attention to gender stereotypes than did those in kindergarten, sixth grade, or college.

PHASE MODEL OF TRANSITIONS

I87

Martin (1989) also found that older children were more flexible in their use of gender as a cue for making predictions, though the exact age at which flexibility changed was difficult to assess because the 4- to 10-year-old children were combined into younger (4-6) and older (7-10) groups. Phase 3 should also be characterized by individual differences in information processing and impact. The large literature on gender “schematicity” (e.g., Bem, 1981; Crane & Markus, 1982) in adults may be construed as representing Phase 3 responding, but there is little evidence about when such individual differences emerge. A series of studies by Liben, Signorella, and their colleagues (see Liben & Signorella, 1987, for a review) indicates that highly stereotyped children recall more gender-consistent pictures than do less stereotyped children, but researchers have not examined when this individual difference first emerges. Once again, longitudinal research is needed to examine these implications of changes across phase in knowledge acquisition and representation. Table 111 summaries the combination of predictions and findings for gender transitions in terms of the phase model. The changes in level and type of knowledge during early childhood are well demonstrated (Martin, Wood, & Little, 1990; Ruble, 1988). Indeed, Martin et al.’s (1990) sophisticated study about the components of stereotype development concluded that they develop in a series of stages that are quite compatible with the present analysis. During Stage 1 (2-4 years), children learn what kinds of things are associated with each sex (e.g., boys play with cars). During Stage 2 (4-6 years), children develop more complex associations for information relevant to their own sex. During subsequent years, children’s knowledge expands to incorporate information relevant to the other sex. The four studies described in detail above primarily illustrate the link between differences in knowledge (preconstant to constant) and two of the cognitivemotivational variables shown in Table 111: Focus of information-seeking and application. The remaining cells of the table represent predictions based on agedifference findings in the literature, such as memory for gender-consistent information (Stangor & Ruble, 1989a), categorization (Martin, 1989; Serbin & Sprafkin, 1986), or stereotype flexibility (Signorella et al., 1992), but specific associations to different levels of knowledge remain to be demonstrated.

VII. The Transition in the Meaning of Ability Another kind of social-cognitive transition concerns children’s increasing understanding of personal competence. When children first enter school, at around age 5, they learn that there is a whole new world to understand. According to the current phase model, this social-role transition, just like all other

188

DIANE N. RUBLE TABLE III EXAMPLE OF PHASES USINGGENDER TRANSITIONS Transition phases Variable

Level of knowledge

S p e of knowledge

Information seeking

Phase 1 (Construction)

Incomplete often inaccurate knowledge of stereotypes Specific behaviors and objects

Impact Importance and affective significance Application

Representation Extensive knowledge of stereotypes Constancy of gender; roles and traits

Phase 3 (Integration)

Stereotypic knowledge near asymptote Individual differences in gender- or selfschemas

Cognitive-morivarionl orientation Little information Active information Active information seeking; unreceptive seeking, focused seeking focused on on expanding to inconsistent infordefining gender same-sex knowlmation categories edge

Information processing Gender often used in Accessibility categorization Memory

Phase 2 (Consolidation)

Gender primary basis used

Recall better for gender-relevant vs. irrelevant

Recall better for gender-consistent vs. inconsistent

Little reaction to violation of stereotypes

Inflexible application of stereotypes and negative reaction to violations Same-sex preferences maintained even if unattractive

Same-sex preferences easily modified

Inflividual differences in use of gender as a category Individual differences in processing of gender relevant and consistent information Flexibility in stereotyping and reactions to violations Same-sex preferences depend on individual differences

transitions, should be accompanied by a heightened interest in information about this new system. Indeed, in two studies, we have shown that children exhibit the most extensive overt discussion about abilities and relative performance in the first year or two of school (kindergarten and first grade), but then such open discussion declines (Frey & Ruble, 1985; Pomerantz, Ruble, Frey, & Greulich, 1993). Moreover, young children are more likely to perceive such discussion in terms of general interest, rather than as indicative of performance assessment.

PHASE MODEL OF TRANSITIONS

189

According to the present argument, the young children’s behavior reflects an open search for information about procedures and what is important in the classroom environment, unencumbered at this time by concerns with demonstrating competence or coming to a particular conclusion about school. In Kruglanski’s (1990) terms, they are oriented toward nonspecific, rather than specific closure. Once they have the information they need, this kind of information search should decline. There are alternative reasons for this decline, however. As children progress through school, they learn that it is not all fun and games-that significant others care about task performance and grades. They also learn that their grades and classroom standing are relatively stable, reflecting a general level of ability at tasks like reading and math. This realization is important because it changes the significance of day-to-day performance in school. Such performances are now indicative of general ability, with implications for future performance and opportunities, as well as for regard by significant others. Thus, the decline in overt exchanges about performance may reflect the increased sensitivity of the information. Indeed, this decline was associated with a measure of children’s awareness of ability as a trait in the Frey and Ruble (1985) study. Children who referred to general ability as a reason to select a particular peer when seeking assistance were less likely to engage in overt performance statements. This new understanding of ability represents a dramatic change that may be viewed as corresponding to the shift from Phase 1 to Phase 2 and would be predicted to affect orientations toward information in the ways shown in Table I. Like gender, it represents a realization that not only is competence a characteristic that is important in the present social world but it is also a meaningful and stable aspect of oneself that has a number of significant implications for one’s desires and goals. Thus, it now becomes crucial not only to understand how it is represented and evaluated but also to assess oneself. Although the drop in overt exchanges may appear to contradict our suggestions of continued high information seeking at Phase 2, there is evidence that the nature and method of information seeking are changing instead. That is, our classroom observational studies suggest that at these ages, children begin to seek ability-assessment information through a more subtle means-asking for information about peer progress (Frey & Ruble, 1985; Pomerantz et al., 1993). In a public setting, this is the self-assessment method of choice because the information can be obtained without necessarily revealing one’s own standing, thereby protecting the self-esteem of both oneself and others (Brickman & Bulman, 1977). Subsequent interviews indicate, further, that at these ages, children increasingly refer to social comparison as a reason for seeking such information. Moreover, references to social comparison during the first year of a 3-year longitudinal study do predict actual assessment behaviors in the classroom, but only in subsequent years (Pomerantz et al., 1993). One possible reason for this

190

DIANE N. RUBLE

lag is that it may take time to figure out how to get the information in a socially acceptable way once one has decided that such information is necessary or, in other words, once this transition phase has begun. There are a number of other indications that this is an important time of transition. First, there are age differences in self-evaluative processes and outcomes that correspond to the ages at which this shift occurs. For example, children older than 7 years of age are more likely to utilize social-comparison information when evaluating their ability (Ruble, 1987; Ruble & Frey, 1991). These results do not imply that social comparison is unfamiliar to young children, however. In a series of elegant studies, Butler (1989a; 1989b; 1990) has shown that younger and older children (i.e., before or after this shift) engage in social comparison in the service of different goals. Specifically, younger children are more likely to seek information relevant to mastering the task, whereas older children are more likely to seek information relevant to competence assessment, consistent with our portrayal of the shift from the Construction to the Consolidation phase. Interestingly, recent data suggest that these different concerns may lead younger and older children to show different forms of evaluative biases (Ruble, Eisenberg, & Higgins, in press). Younger children were less likely to incorporate failure information when evaluating their own specific perjormance than when evaluating another, whereas older children showed this self-other difference for the utilization of failure infomation for evaluations of general ability. Second, measures of conceptions of ability or stable traits, more generally, have been associated with differences in the focus of information seeking or the impact of evaluative information. Ruble and Flett (1988) reported that perceiving ability as constant was associated with greater interest in social comparison, especially among older children. Rholes, Jones, and Wade (1988) showed that 7to 8-year-old children who perceived behaviors in stable terms were more likely to show a learned helplessness response to failure than children who did not perceive such behavioral stability. Similarly, Miller (1985) reported that maladaptive responses to failure feedback among second- and sixth-graders occurred only for children with conceptions of ability as capacity, most of whom were sixth-graders. Thus, taken together, these studies suggest that this shift in the meaning of competence as a stable trait is associated with an increased interest in self-evaluative information, with a particular focus on the form that would be most relevant in most classroom settings (i.e., social comparison). Moreover, it appears to be related to a greater impact of such evaluative feedback on subsequent choices and experiences. Indeed, Benenson and Dweck (1986) have shown that as references to trait explanations of performance increase with grade, positive self-evaluations decline, suggesting that the two events are likely to be related. Recent analyses of our longitudinal data have provided more direct evidence of

PHASE MODEL OF TRANSITIONS

191

this association, indicating further the significance of this period of transition. Our own previous research (Frey & Ruble, 1987; Parsons & Ruble, 1977), as well as that of a number of others (see Stipek & MacIver, 1989, for a review) have shown this decline in positive self-evaluations across the first several years of school. What accounts for it? A number of hypotheses have been proposed, including changes in cognitive processing abilities and widespread changes in the educational environment that occur as children progress through elementary school. Exactly what these increasingly pessimistic self-evaluations represent, however, has to date remained a mystery. One of the problems with the data base presently available is the lack of longitudinal evidence. In order to understand the cause of such changes, it is important to pinpoint when they occur. Like the hypothesis of Benenson and Dweck (1986), one of our own hypotheses has been that this drop is related to changes in perceptions of the nature of ability as a trait at this age-that current performance can be used to predict behavior across time and situation. This realization is important because it implies that poor performance in school can no longer be dismissed as a momentary event with no future implications or to temporary lack of effort that can be easily remedied. The unrealistically high self-perceptions of young children, then, may exist because negative feedback in school or failure experiences can be readily dismissed as irrelevant to obtaining desired outcomes in the future. Thus, we would predict that as children recognize the stable nature of school-related skills, their perceptions of their own competence must be assessed. The conclusions drawn during this Phase 2 assessment process depend on the nature of the information available at this time. Because in school relatively few children are at the top of their class, what this implies is that ratings of competence should decline. This would mean that we would expect to see a sharp decline in self-ratings sometime between 2nd and 4th grades, rather than the gradual decline implied by most previous cross-sectional studies. Moreover, one might expect this realization to be depressing for children who perceive their competence as relatively low. Thus, we predicted further that, only among children who perceive ability as constant, a decline in self-ratings of competence should be associated with increased depressive affect. These predictions were examined as part of our longitudinal research employing a cross-sequential design (Ruble, Pomerantz, Frey, & Greulich, 1993). The children were initially in Grades K, 1, and 2, and they were followed for three years. Consistent with previous research, children’s self-perceptions of cognitive competence (as measured by a slightly modified version of the cognitive subscale of the Harter, 1982, Perceived Competence scale for children) declined over time and with increasing grade. This decline was particularly marked after second grade, as predicted. Ability constancy was assessed using a revised version of the ability-constancy scale described by Ruble and Flett (1988). This measure is based on measures of

192

DIANE N. RUBLE

gender constancy and asks children about another’s reading and math ability over time (e.g., whether a child who is very good in reading now will be very good in two years) and over situation (e.g., whether a child who is good in addition will be good in subtraction). As predicted, changes in ability-constancy scores corresponded to the timing of the decline in self-ratings, notably in beginning to increase after second grade. In order to examine the relation between these two developmental changes, we conducted a hierarchical regression analysis predicting change in perceived competence from change in perceived ability constancy. As expected, an increase in perceiving ability as constant predicted the decline between the second and third grades, but for girls in the second-grade cohort only. Although the reasons for this gender difference are not yet clear, we have speculated elsewhere (Ruble, Greulich, Pomerantz & Gochberg, in press) that girls may receive greater pressure to do well in elementary school and are thus more attentive to performance feedback. Several studies suggest that females at various ages are more responsive to failure feedback (Parsons & Ruble, 1977; Roberts, 1991), and more sensitive than boys to situations that emphasize competence (Butler, 1992). Finally, we examined the relation between perceived ability constancy and the emergence of depressive affect in school-aged children in a separate longitudinal sample (Gochberg, Ruble, & Higgins, 1992). Not surprisingly, children who evaluated themselves as lower in cognitive competence (again using a modified version of the Harter scale) reported more depressive affect. More importantly, this effect was qualified by perceived ability constancy. As scores on the measure increased, the relation between perceived incompetence and depressive affect increased. Thus, not only might the emergence of ability constancy lead to a decline in perceived academic competence, but for those who retain a relatively strong belief that ability remains stable across time and situation, concluding that they are low in ability is particularly depressing. Taken together, these data suggest important shifts in the information-seeking focus and impact of performance information across Phases 1 and 2. There appear to be differences in the type of information sought and in reactions to evaluative feedback corresponding to the perception of ability constancy. What happens at Phase 3? According to the model, we should see a reduced emphasis on seeking information relevant to competence assessment and evidence of individual differences in achievement-related goals. One study of age differences in information-seeking behavior in children reports some data consistent with this hypothesis (Ruble & Flett, 1988). When given a chance to evaluate their performance on a math-related task, high-ability children showed an increase between second and sixth grades in relative interest in mastery-oriented versus selfassessment-oriented (social-comparison) information, whereas low-ability children maintained a relative preference for assessment-oriented information across ages. One interpretation is that high-ability sixth-graders have reached a satisfac-

PHASE MODEL OF TRANSITIONS

193

tory conclusion about their abilities and can now focus on task-relevant goals (Phase 3 concerns), whereas low-ability children are not yet convinced that they have reached a conclusion about their level of math ability. Indeed, high-ability children were more certain of their ability level than were low ability children. With respect to individual differences, there is a large literature showing variations in adults’ and older children’s interest in and responsivity to performancerelated information. As discussed earlier, individuals differ in the extent to which they emphasize learning versus self-evaluation goals (Dweck & Leggett, 1988; Nicholls, 1984), and these differences influence the impact of performance feedback on affect and subsequent task motivation. Whether such differences can be shown to emerge from the outcomes of Phase 2 conclusions is a question that must await further research.

VIII. The Transition in Perceptions of People’s Characteristics A third kind of social-cognitive transition concerns changes in children’s perceptions of people’s characteristics. These changes are related to those discussed above with respect to ability constancy, but they have implications for children’s interpersonal interactions rather than achievement-related motivation and self-esteem. Phase 1 is marked by children’s awareness that there are individual differences in people’s characteristics, a kind of initial implicit personality theory involving types of people. The beginning of this transition is more difficult to identify than the others because it does not have a clear marker, such as the entry to school. In addition, there has been little research on the origins of children’s awareness of individual differences. It is likely to depend upon how early children have regular experiences with multiple peers, allowing the awareness of differences (Higgins & Wells, 1986). Certainly by kindergarten, however, children exhibit an interest in people’s characteristics. In her book Bad Guys Don’t Have Birthdays, Paley (1988) reports, for example, that preschool and kindergarten children are obsessed with differences between “good guys” and “bad guys,” and recent research by Dweck (1991) indicates that this is a critical distinction for young children. Informal observations suggest as well that children of this age exhibit considerable interest in a newcomer’s age, number of siblings, favorite toys, and other “critical” items of identification. The significance of these changes as a period of transition becomes more obvious at Phase 2. Children’s expanding experiences with peer interaction, coupled with cognitive developmental advancements, lead them to view people as having stable, dispositional characteristics, allowing for predictability over time (Rholes, Newman, & Ruble, 1990). This change is important because once

194

DIANE N. RUBLE

children realize that behaviors are invariably linked to individuals, there is a compelling reason to try to organize and understand the characteristicsof others. Such an understanding should allow for predicting another’s behavior in the future, thereby promoting appropriate choices of future playmates and facilitating a smooth interaction. Like other transitions, then, Consolidation is the phase when this general awareness of others’ characteristics becomes personally relevant. Thus, we would expect to see some of the same changes in orientation toward social information that mark the shift to consolidation in other transitions, such as a focus on particular characteristics and a rigid adherence to the rule that people’s behaviors reflect underlying, stable traits. There is considerable evidence that in our culture, children’s awareness of stable, dispositional characteristics emerges between 7 and 9 years of age. Children younger than this age tend to focus on concrete, physical aspects of people (e.g., he has dark hair) instead of the more psychological and predictively useful aspects mentioned by older children and adults (e.g., she is very friendly). Such trends are shown quite clearly in free-description studies, such as the pioneering work of Livesley and Bromley (1973). They showed that the proportion of references to central characteristics (e.g., personality) relative to peripheral (e.g., appearance) increases dramatically with age until 8 years, at which point the proportion is about 50%, where it remains through adolescence. This social-cognitive change is due not just to verbal skills. A number of studies using forced choice selections or rating scales, not relying on verbal production, have now shown that young children typically do not make predictions about an individual’s behavior in a new situation on the basis of inferences about enduring, underlying characteristics (see Rholes et al., 1990, for a review). Our conclusion is that young children do not view people’s behaviors as emanating from an internal force that would lead them to do the same thing in other circumstances. How does this change in social knowledge affect cognitive-motivational orientations toward peers? One prediction is that it would affect the focus of information search upon meeting a new peer. That is, a shift from Phase 1 to Phase 2 would be accompanied by an increased interest in gathering information relevant to predicting behavior in future situations. Camhy and Ruble (in press) looked at this issue directly. Subjects were ninety-six 7-8-year-olds, who were high rather than low on understanding stable, dispositional concepts (SDC). The measure of SDC consisted of a set of items in which children were asked to make behavioral predictions on the basis of previous disposition-relevant behaviors-for example, whether a boy was timid in one situation (climbing a tree) would be timid in another (entering a spooky house). The main dependent measure of the study involved information seeking relevant to getting to know a new child. The subjects had the chance to select possible questions to ask someone they might meet. These questions were drawn from pilot testing involving actual interactions of 7- to 8-year-old children, and

PHASE MODEL OF TRANSITIONS

195

questions roughly corresponding to Livesley and Bromley’s (1973) centralperipheral distinction were selected. To illustrate, a peripheral question was, “How do you get to school?’ whereas a central question was, “Do you like to play tricks?” There was also a manipulation of expected interaction. In one condition, children were led to think that meeting another child was likely. These children were told that there might be a visitor to the school and that they should think of questions they might like to ask him or her. In a second condition, an instrumental component was added to expected interaction. In this condition, children were led to think not only that it was likely they would meet another child but also that they would have a chance to select the target as a possible partner for some games. In the third condition, children expected no interaction. It was predicted that expected interaction would heighten the need to know as much about the other as possible, that is, that it would lead children to focus more on disposition rather than on peripheral information, at least for children high in SDC. This prediction was supported, though only for the instrumental expectation condition. High SDC children who expected to be making decisions based on their knowledge of the other selected relatively fewer peripheral (and more central) questions than did low SDC children. In a conceptual replication, the dependent measure consisted of the actual questions that children asked of another during a free play situation. Only the instrumental expectation condition was compared to the no expectation condition. Although the type of question asked was not affected by the independent variables, as in the first study, the results showed that high SDC children who expected to make decisions about the target’s suitability as a partner for different games asked more questions overall than did low SDC children. Thus, as in the previous study, only high SDC children were affected by the expectation manipulation, though in a somewhat different way. In retrospect, it is probably not surprising that they did not show a preference for disposition-related questions. In an actual interaction with a new child, it would seem to violate social norms to ask personal questions prior to exchanging certain basic information about background and interests. Thus, taken together, the results of these two studies suggest that a variation in person perception from peripheral to stable, dispositional is associated with differences in the focus of information seeking corresponding to the first two phases of the transition model. Understanding stable, dispositional characteristics appears to be related to greater interest in information relevant to future interactions, although the exact behavioral manifestation of this motivation appears to depend on the constraints inherent in the social context. A few other studies are consistent with the predicted cognitive-motivational changes, though not directly related to a change in social knowledge. Graziano, Brody, and Bernstein (1980) showed that third-grade children were more likely than first-grade children to attempt to placate a potentially aggressive peer by

196

DIANE N. RUBLE

providing them tokens that could be exchanged for prizes when they expected future interaction with that peer. In contrast, there were no differences across age in token distribution to an aggressive peer when no future interaction was expected. These findings suggest that increasing social knowledge about peers may result in more strategic orientations toward them. Given previous findings of age differences in perceptions of stable dispositions in others, it is reasonable to assume that the older children expected the aggressive child to behave aggressively toward themselves in the future interaction and were attempting to strategically influence the nature of that interaction. Findings from some of our research supports the idea that there is an increase in strategic orientations at the ages that children typically begin to show an understanding of dispositional characteristics. In one study (Feldman & Ruble, 1988), children at two age levels (5-6 years and 9-10 years) watched videotapes of different target children portraying behaviors relevant to some games that would be played subsequently. Older children were more likely than younger children to choose partners who had demonstrated game-relevant skills, especially when they anticipated future interaction with the target. In contrast, younger children consistently chose partners based on their likability even though they exhibited an awareness of which choices would be personally advantageous. Thus, the bases of interpersonal choices appeared to shift across this age range, with younger children making decisions more on the basis of immediate affective reactions, and older children being more influenced by future instrumental considerations. A recent study reported by Thompson, Boggiano, Costanzo, Matter, and Ruble ( 1993) provided additional information about the processes underlying these age differences. The procedure was basically the same as that used by Feldman and Ruble (1988), with one major exception. Children were provided with either an affective or descriptive encoding focus. Before seeing the videos, children were told to focus either on how the targets made them feel or how they would describe the targets to someone who did not know them. It was reasoned that if older children’s belief in dispositional stability underlies their strategic, instrumental orientations toward interpersonal choices, then older but not younger children would be more likely to respond strategically when instructed to form an impression of the targets’ personalities. In addition, we examined the prediction that children would make more instrumental choices when they expected consistency in the targets’ behavior. As expected, older children made more strategically “correct” paitner choices than did younger children, and only in the descriptive-focus condition. Moreover, a measure of the tendency to see the target’s behavior as relevant to predicting subsequent behavior appeared to mediate these effects. Taken together, these various findings suggest that important changes in children’s orientations toward peers accompany the transition in social knowledge about individual characteristics that occurs between 7 and 9 years of age. Once

PHASE MODEL OF TRANSITIONS

197

children perceive others as possessing stable dispositional characteristics, such as abilities at games, gathering information about them promotes future predictability and control. Thus, attention to such characteristics is personally useful (Phase 2) when forming impressions of a newcomer and in making interpersonal choices. Other changes in cognitive-motivational orientation that should be associated with shifts in social knowledge are the salience of dispositional characteristics in others and the rigidity with which they are inferred. One should expect to see a curvilinear pattern across age, with dispositional characteristics most salient and most rigidly employed at Phase 2, when an image of others as possessing stable individual differences is consolidated, after 8 to 9 years of age. 'Iko studies report findings consistent with this hypothesis (Josephson, 1977; Newman, 1991). Children at this age were less likely than either younger children or adults to consider the possibility that the behavioral implications of an inferred trait could change (e.g., that a generous individual could behave in a stingy manner). In addition, Newman (1991) found that 8-9-year-old children were more likely than adults to infer traits spontaneously by using a reaction-time paradigm. Taken together, these findings are consistent with the shifts in information processing and impact proposed to occur between Phases 2 and 3. The latter, integration, phase should be characterized by a decreased likelihood of interpreting novel information in terms of trait constructs and an increased flexibility in the application of them. Instead, there should be an increased recognition that dispositional stabilities may be qualified by situational constraints (Wright & Mischel, 1988). In addition, individual differences in beliefs about trait constancy (Dweck & Leggett, 1988) or in strategic versus affective orientations toward relationships (e.g., exchange vs. communal orientations, Clark & Mills, 1979) may become evident at this phase. One interesting issue about the shift from Phase 2 to Phase 3 concerns the nature of the experiences at Phase 2 that give rise to individual differences at Phase 3. A question we are addressing in our current research, for example, is whether parental emphasis on instrumental aspects of interactions at the time children become aware of behavioral stabilities leads subsequently to individual differences related to exchange versus communal orientations.

IX. Implications of the Phase Model A. IMPLICATIONS FOR COMMON TRANSITION PHENOMENA How might the present analysis explain some well-known discontinuities and disruptions occurring during two common transitions that were described at the beginning of this chapter? First, consider the well-documented decline in marital satisfaction that accompanies the birth of a first child. According to the present

198

DIANE N. RUBLE

analysis, pregnancy involves Construction and Consolidation-that is, expectations about what it will be like to be a family are constructed and new conclusions are formed. Many women expect that these new family responsibilitieswill be shared, that the husband will contribute substantially to the housework and child care. Generally, the model would predict that expectancy-confirmation processes would operate, leading to a perceived reality that matches the expectation. For many families, however, postpartum reality is too discrepant from expectations of shared responsibilities, leading to a sense of violated expectations (Belsky, 1985; Ruble, et al., 1988). Our research suggests that a decline in marital satisfaction is most pronounced for the women who experience the greatest expectancy violation or for whom confirmation of expectations of sharing are particularly important (Hackel & Ruble, 1992). This is one way, then, that changes in orientations toward information during a period of transition may lead to the disruptions often observed to follow. Information available during the early phases of the transition may lead to expectations that are not supported by the reality of the subsequent phases. A second illustration concerns the drop in positive self-perceptions in early adolescence. A combination of cross-sectional and longitudinal studies have shown that developmental decreases in perceived competence and corresponding motivational orientations are particularly marked at the transition to junior high school. Simmons et al. (1973) found that children’s self-perceptions (e.g., selfesteem, confidence in academic abilities) were most negative during early adolescence, particularly for those who had just begun junior high school. Harter (198 1) reported a sharp decline in intrinsic motivational orientations toward classroom activities between sixth and seventh grades, corresponding to a transition to junior high school. Eccles et al. (1989) showed a decline between sixth and seventh grades, corresponding to a transition to junior high school, in selfesteem, and self-concept of ability in math, English, sports, and social competence. Peterson and Crockett (1985) showed a decline in body image and school grades between Grades 6 and 7, associated with a school transition. Finally, Simmons and Blyth (1987) found in two large studies a decline in self-esteem associated with the move to junior high school, particularly for girls. These two studies are particularly noteworthy, because Simmons and Blyth were able to control for age by comparing school districts with different structures-those with junior high schools beginning in seventh grade and those without (i.e., kindergarten through eighth grade elementary schools). What accounts for these changes? There have been two general categories of explanations. First, Simmons and Blyth (1987) refer to these negative effects in terms of stress reactions. They favor a developmental readiness hypothesis-that a school transition occurring during the vulnerable age of early adolescence is more stressful than if it occurred a few years later. Moreover, they argue that many significant changes (e.g., puberty and a school transition) occurring close

PHASE MODEL OF TRANSITIONS

199

to the same time create a particularly stressful situation and detrimental reactions. Second, Eccles et al. (1984) emphasize changes in the nature of the school environments. Positive self-perceptions decline because junior high school environments are larger, more impersonal, and place a greater emphasis on evaluation and social comparison than do elementary schools. Other changes associated with the transition, such as a disruption of children’s social networks, may also explain the decline. Declines in perceived competence and self-esteem are not shown consistently across studies, however. In a review of a large number of such studies, Eccles and Midgley (1989) describe a wide range of reactions. For example, Thornburg and Glider (1984) suggest that age per se produced more significant effects than did grade or school transitions. Eccles, Midgley, and Adler (1984) observed a decline in self-concept of math but not English ability. Others have found little evidence of any decline at all (e.g., Harter, 1982; Nottelmann, 1987). How can we predict when changes will occur and what directions they will take? Eccles and Midgley (1989) suggest it depends on the specific context in which the adolescents are changing from and to. To translate into the current analysis, it depends on two factors. First, it depends upon whether the transition involves suficiently dramatic changes to elicit a need to redefine old conclusions. If the format of the school is similar to the format of the old school, and if all the school chums are moving together, there may be less uncertainty that would elicit a need for information. Alternatively, if, as is usually the case, there is a change in social relationships, classroom structure, grading practices, and so on, then old conclusions can no longer hold, and transition processes are set in motion. Several findings consistent with this hypothesis are reported in the Eccles and Midgley (1989) review. Simmons and Blyth (1987), for example, found that negative effects were more likely when students moved to a larger school with greater ethnic heterogeneity. Moreover, as Eccles and Midgley (1989) point out, there may be greater changes in classroom practices for some subjects than for others, possibly accounting for the greater changes they found in students’ perceptions of their math versus their English competence. In brief, this analysis suggests that the greater the change, the more the individual is likely to decide that old standards and conclusions about competence no longer apply. Consistent with this suggestion, Connell and Tero (1982, unpublished, cited in Eccles & Midgley, 1989, p. 158) found that students undergoing a transition are more likely to report that they do not understand the causes of their outcomes. As Eccles and Midgley (1989) note, this finding is noteworthy because children typically show evidence of understanding better the bases of their outcomes as they progress through school. Second, it depends on the nature of the information available during this period of increased information seeking accompanying reassessment of conclusions about competence. Eccles and Midgley (1989) review data suggesting that

200

DIANE N. RUBLE

there is an increased focus on social comparison and tougher standards, accompanied by more negative teacher views, associated with the transition from elementary school to junior high school. Thus, it is hardly surprising that many come to a more negative conclusion than they did earlier. A study of the impact of transitions during the junior high school/high school period on gender stereotype flexibility (Alfieri, Ruble, 81 Higgins, 1993) is relevant to this last point. If, as I have argued, dramatic transitions such as these usher in a need to redefine oneself and one’s social world, then even previously well-defined categories, such as gender, may be subject to scrutiny. Such transitions may open up a window of change as individuals enter once again the Construction phase of active information seeking. Consistent with these ideas, we found that gender-stereotype flexibility showed an increase associated with a transition to junior high school. Because the study involved two school districts in which this transition occurred at different ages, it was possible to distinguish the effect of transition from grade. The findings showed increased gender flexibility whether the transition occurred at seventh grade or eighth grade. Interestingly, the findings were virtually identical across cross-sectional and longitudinal analysis. Thus, the outcomes of transition changes are not inevitably negative but may depend on the exact information that is acquired. The important point is that transitions may lead individuals to redefine old conclusions, such as those concerning gender and one’s competence. The negative outcomes found by many may not be intrinsic to the transition itself. Rather they may depend upon the results of the information gathering and construction process that the transition initiates.

B. GENERAL IMPLICATIONS

From the present perspective, differences across phase in orientations toward social information are key to understanding the personal changes that result from transitions. If, as I am arguing, individuals are maximally open to, and motivated to acquire, certain kinds of information during relatively circumscribed periods, then the particular information available at that time will largely determine the nature of the conclusions drawn. Moreover, at any point in the life-span, once conclusions are drawn, they may be difficult to change because of the shifts from active to passive information seeking and from relatively open to closed information processing. For example, although the particular toys young children play with may be sex typed for mostly passive reasons (i.e., adult selections), subsequent choices become actively determined. The gender-typed division of interests and skills that become consolidated at this time, along with the incorporation of gender stereotypes into one’s self-definition, seems a likely source of well-

PHASE MODEL OF TRANSITIONS

20 1

known gender differences in adults in status, occupations, and perhaps even emotional functioning. Similarly, a girl in a classroom of high-achieving children who has concluded erroneously that she is poor at math may subsequently distort discrepant information (e.g., attribute a high score to luck), or simply stop trying. Even in a new context, then, her negative self-perception is maintained, and discouragement and low self-esteem are a likely consequence. Thus, once again, the cognitive-motivational changes that occur at this time may have farreaching implications for girls’ subsequent educational and occupational choices. The idea that the seeking, processing, and impact of information varies systematically across phase implies that there are particular windows of change or vulnerability. In addition, the focus on active, constructiveelements of cognitivemotivational orientations makes the present model particularly relevant to these potential intervention opportunities. Personal and interpersonal problems may arise because of negative implications derived from the context in which a period of transition is initiated or when a shift from one phase to another occurs. In an achievement setting, for example, new concerns with the evaluation of one’s ability may create special affiliation problems for a child low in ability. Such children may isolate themselves because of embarrassment experienced during negative social comparisons with friends. Understanding why such isolation occurs is important to designing appropriate interventions. Because abilityrelated comparisons are less meaningful between friends of different ages, interventions that place isolated children with younger children (Furman, Rahe, & Hartup, 1979) may be particularly appropriate. A number of intervention-oriented research has been directed at the gender stereotypes of young children. Several studies have shown that the stereotyped attitudes of 4- to 6-year-olds become more egalitarian in response to nontraditional books or television (e.g., Berg-Cross & Berg-Cross, 1978; Flerx, Fidler, & Rogers, 1976). More important for the present argument, De Lisi and Johns (1984) found that children at the stage of gender constancy were particularly responsive to nontraditional material. The phase model also has implications for theoretical developments in other areas. A phase perspective on self-evaluation may shed light on questions raised recently about social comparison theory. Butler (1989a) explores possible reasons for mixed conclusions about whether or not children under 7 to 8 years of age engage in social comparison. Goethals (1986) argues that a key proposition of Festinger’s (1954) theory-that there is a drive to evaluate one’s opinions and abilities-remains unresolved. He suggests that research findings indicate that “there is more of a drive to think that our opinions are correct and that our abilities are good than to find out the truth” (p. 274). Previous considerations of such questions have suggested that social comparison shows varying forms and patterns because it serves different functions (e.g., mastery vs. self-assessment; Butler, 1989a). The present analysis pushes this line of reasoning one step

202

DIANE N. RUBLE

further. Not only does social comparison serve different goals but these goals emerge systematically as part of a single, developmental process. Level of knowledge and experience with a particular domain determine which goals are paramount. Thus, for example, young children may tend to emphasize comparisons that help them to understand task requirements rather than self-assessment because they are at early phases for most skills. Similarly, drive-like, diagnostic properties of social comparison may rarely be seen because they are timedelimited. Once an individual knows how a dimension is defined (Phase 1) and has inferred his or her likely capacity at the task (Phase 2), additional assessment is unnecessary as long as plans and subgoals are met with some regularity during Phase 3. Thus, because most social comparison research involves adults likely to be at Phase 3 for most activities, Goethals (1986) may be correct in suggesting that there is little evidence supporting Festinger’s proposed motive to engage in self-evaluation. (See Ruble & Frey, 1991, for further discussion.)

X. Conclusions To reiterate, the main point of this paper is to suggest that transitions of many different types may be characterized by a set of phases representing changing cognitive-motivational orientations toward information. The transition to motherhood and social-cognitive transitions in childhood have been emphasized here, but similar changes are expected across diverse types of transitions, such as religious conversions, hospitalization, going to prison, and retirement. Indeed, in a charming account of the transition to becoming a brother, Mendelson (1990) describes a set of four phases that include some of the cognitive-motivational changes emphasized here, such as anticipatory learning during the first phase and personal style during the final phase. To illustrate, the discussion of the uncertainty and information seeking of the first phase includes the questions and confusions of Mendelson’s 4-year-old son, who, anticipating the birth of a brother or sister, wonders how the baby comes out. He first thought the baby would come out of the back, but then decided that it climbs up and sneaks out either the nose or the mouth. The present model of transitions includes three phases. Construction involves the initiation of active information seeking. To illustrate with respect to data presented in this paper, the onset of pregnancy is marked by an increased interest in books about the experience relative to prepregnant information gathering. Shortly after entering school, children exhibit overt interest in performance and evaluative information, an interest that subsequently declines. Young children also show considerable interest in constructing definitions of gender and other categories of people (good/bad). As adolescents undergo a transition to junior

PHASE MODEL OF TRANSITIONS

203

high school, even well-established conclusions about gender and personal competence may be reassessed. It is important to be cautious about the nature of this initial phase, however. With the exception of the pregnancy data, there have been no adequate comparisons to pretransition levels of information seeking. Thus, the suggestion that information seeking becomes active at the Construction phase is more properly considered a hypothesis than a fact. The present data are clearer about changes that accompany the shift from Construction to Consolidation. This phase is marked by the structuring of a set of expectations and new aspects of identity. The focus of information seeking narrows. Gender-constant children are particularly oriented toward same-sex information. Ability constancy is accompanied by an increased interest in social comparison information. Perceptions of stable, dispositional characteristics lead to a heightened interest in information relevant to future interactions. In addition, there is evidence of a greater personal impact of this focused information. Gender-constant children base decisions on gender information, rather than on other alternatives. Ability constancy is associated with a drop in perceived competence and self-esteem, which, in turn, increases reports of depressive symptoms. The emerging understanding of behavioral consistency is accompanied by an increased tendency to utilize trait information in selecting interaction partners, as well as a relative rigidity in the application of the consistency rule. One way to characterize the shift from Construction to Consolidation, more generally, is in terms of goals, as discussed earlier with respect to socialcognitive and motivational theoretical perspectives. That is, the Construction phase is comparable to a predecisional orientation, subjective uncertainty and low control, and elicits goals of mastery and understanding. In contrast, the Consolidation phase is comparable to a postdecisional orientation, a sense of confidence and commitment, and elicits goals of adoption, need for closure, and assessment. During the Construction phase, the individual asks, “What is it?” whereas during Consolidation, the question turns to, “How does it apply to me?” and “How well am I doing?’ In terms of gender, this shift is away from simply asking what it is that boys and girls do, and toward valuing only one of the alternatives and making sure that one’s behavior is consistent with it. This phase may also be considered one of personal investment in the new category. As part of a transition to a drug-free state, for example, one might wear a pin or post a bumper sticker to announce the adoption of this new identity. Partly because many of the changes across phase are motivational in nature, there are certain empirical difficulties in examining these changes. Several of the studies of Phase 1 to Phase 2 shifts suggested that indices of motivational changes must take into account context variation. For example, changes in social comparison thought to accompany the transition in the meaning of ability may not be detected if observations are limited to a public setting (Frey & Ruble, 1985; Pomerantz et al., 1993). Similarly, the nature of the impact of children’s

204

DIANE N. RUBLE

awareness of stable, dispositional characteristics in people on information search may depend on whether the measure involves a hypothetical or an actual interaction (Camhy & Ruble, in press). Finally, what is defined as personally significant is culturally determined and might be best pursued through cross-cultural or subcultural analyses. As D’Andrade (1981) notes, social pressure to acquire particular items of knowledge varies across cultures. In a culture in which gender is an unimportant category, for example, there may be little evidence that children pass through a phase of same-sex bias, similar to the effect reported here. Indeed, as mentioned above, this “cultural” dimension may be one way to interpret sex differences in the impact of gender constancy-that is, that the culture places greater emphasis on gender distinctions for boys than for girls. Phase 3, Integration, occurs after the uncertainties of the “What?” and “How?’ questions have been answered, and the individual now adapts to the answers and behaves in terms of them. There are two types of integration that occur. First, the conclusions drawn must be integrated with ongoing activities. It is no longer functional to seek information and make general assessments, because it distracts from other, now more significant, activities. Thus, consistencymaintenance processes begun toward the end of Phase 2 should continue. That is, biased perception, memory, and attributions, as well as various social-influence processes (behavioral confirmation, self-fulfilling prophecy), should all operate to maintain the conclusions that were formed in Phase 2 (Crocker, Hannah, & Weber, 1983; Swann, 1990). According to the present model, individuals at early phases, particularly at Phase 1, should show little evidence of consistency biases. The research showing a shift from schema-inconsistent to schema-consistent memory biases as a function of schema development (Stangor & MacMillan, 1992; Stangor & Ruble, 1989a, 1989b) supports this suggestion, but longitudinal data are needed for a more definitive test. Second, the conclusions drawn must be integrated with other aspects of knowledge about the self and the social world. We have argued that individual differences observed in late childhood and in adults reflect a shift to Phase 3. Emphases in the ongoing familial or cultural context may influence whether or not, for example, strong gender differentiations will be maintained as an individual difference at this phase. Similarly, individual differences in achievement goals reflecting mastery orientation versus demonstrating competence (Dweck & Leggett, 1988; Nicholls, 1984) may emerge in Phase 3 as a result of processes occurring during Phase 2. Moreover, findings of cultural differences in references to person-versus-situation factors as primary sources of causal inference (Miller, 1984; 1986) provide further evidence of such individual differences becoming evident subsequent to attaining a Phase 2 level of social knowledge. Once again, however, to more conclusively document this point, longitudinal analyses are needed to show a divergence in behaviors and goals across individuals after conclusions have been formed. Finally, preexisting individual differ-

PHASE MODEL OF TRANSITIONS

205

ences may influence reactions during Phase 3 to conclusions drawn at Phase 2, as suggested by findings of the differential impact of expectancy violations on marital satisfaction after the birth of a first child. At the present point, this proposal for three phases in the course of responding to transitions is by no means the answer to all the questions one might ask. It is not clear, for instance, whether the cognitive-motivational differences observed across phases are specific to particular content areas (gender, ability) or to more general cognitive changes that are occurring in these age ranges, such as representational differences discussed in the literatures on theory of mind (e.g., Frye & Moore, 1991; Frye, Zelazo, & Palfai, 1993) or categorization (Bigler & Liben, 1992; Gelman, Collman, & Maccoby, 1986). Currently, we are involved in longitudinal research addressing that type of question. It is also not clear how well the present proposal would apply if certain assumptions did not hold. What would happen if individuals were not particularly interested in reducing uncertainty or, to take a point from Goodnow (1990), if the information needed to redefine old categories or develop new identities was difficult to obtain instead of being assumed readily accessible? Similar questions arise when there is a strong attachment to existing identities. Individuals who are strongly committed to their work, for example, may shun a new identity as a mother or as a retired person, especially if these identities were involuntary transitions. Other transitions (e.g., homelessness) are unwelcome and may lead to avoidance from the beginning. How do individuals approach transitions that are unwanted or, at least, not actively chosen? Given the importance of choice as a variable in related theoretical constructs, such as decision making and dissonance, future research might usefully compare cognitivemotivational responses to a particular transition, such as becoming a parent, between individuals who chose it or did not. Finally, questions need to be asked about precisely what it is that prompts the move toward reconstructing an old schema. Whether the critical spur is the simple presence of doubt or uncertainty (e.g., Acredolo & O’Connor, 1991), the presence of interpersonal conflict in the form of a difference in opinion from one’s peers (e.g., Doise & Mugny, 1984; Doise & Palmonari, 1984), or the presence of some social pressure to think in a new way (e.g., D’Andrade, 1981), demonstrating the processes involved remains a challenging empirical task. These are, however, questions for the future. The present analysis has the virtue of provoking such questions. It also offers a way of bringing together a number of perspectives and findings from the previous literature, and of adding to these in two unique ways. First, changes in orientations toward information are viewed as natural outgrowths of a kind of developmental change occurring at various points throughout the life-span. Uncertainty, for example, changes systematically in relation to an individual’s emerging knowledge about and experience with a particular transition. Although in some cases this approach overlaps

206

DIANE N. RUBLE

with traditional notions of development (e.g., acquisition of gender in children), in others it is quite distinct in that any new major experience (e.g., hospitalization) may instigate these processes. Second, the present approach examines commonalities in these developmental processes across different types of social and individual change in an attempt to build a more general conceptualization of what is changing and how it is doing so.

Acknowledgments This research was supported by a research grant (MH 37215) and a Research Scientist Development Award (MH 00484), both from the National Institute of Mental Health. I am grateful to Douglas Frye, Jacqueline Goodnow, E. Tory Higgins, Leonard Newman, Eva Pomerantz, Joel Szkrybalo, Charles Stangor, and Erik Thompson for many insightful and challenging comments on an earlier version of this chapter.

References Abelson, R. P. (1986). Beliefs are like possessions. Journal for the Theory of Social Behuviour, 16, 223-250. Acredolo, C . , & O’Connor, J. (1991). On the difficulty of detecting cognitive uncertainty. Human Development, 34, 204-223. Alba, J. W., & Hasher, L. (1983). Is memory schematic? Psychology Bulletin, 93, 203-231. Alfieri, T. J., Ruble, D. N., & Higgins, E. T. (1993). Gender stereotypes during adolescence: Developmental changes and the transition to Junior High School. Manuscript submitted for publication. Bargh, J. A,, & Pietromonaco, P. (1982). Automatic information processing and social perception: The influence of trait information presented outside of conscious awareness on impression formation. Journal of Personality and Social Psychology, 43, 437-449. Bartlett, F. C. (1932). Remembering. Cambridge, England: Cambridge University Press. Bedard, J., & Chi, M. T. H. (1992). Expertise. Current Directions in Psychological Science, 1(4), 135- 139. Belsky, 3. (1985). Exploring individual differences in marital change across the transition to parenthood: The role of violated expectations. Journal ofMarriage and the Family, 47, 1037-1044. Belsky, J., Rovine, M., & Fish, M. (1989). The developing family system. In M. Gunnar & E. Thelen (Eds.), Systems and development: Vol. 22. Minnesota Symposium on Child Psychology (pp. 119-166). Hillsdale, NJ: Erlbaum. Bem, S. L. (1981). Gender schema theory: A cognitive account of sex typing. Psychology Review, 88, 354-364. Bem, S. L. (1984). Androgyny and gender schema theory: A conceptual and empirical integration. In T. B. Sonderegger (Ed.), Nebraska Symposium on Motivation (Vol. 32, pp. 179-226). Lincoln: University of Nebraska Press. Benenson, J. F., & Dweck, C. S. (1986). The development of trait explanations and self-evaluations in the academic and social domains. Child Development, 57, 1179- 1187.

PHASE MODEL OF TRANSITIONS

207

Berg-Cross, L., & Berg-Cross, G . (1978). Listening to stones may change children’s social attitudes. The Reading Teacher, 31, 659-663. Berlyne, D. E. (1966). Curiosity and exploration. Science, 153, 25-33. Berlyne, D. E. (1970). Children’s reasoning and thinking. In P. H. Mussen (Ed.), Carmichuel’s manual of child psychology: Vol. 1. New York: Wiley. Berndt, T. J., & Heller, K. A. (1986). Gender stereotypes and social inferences: A developmental study. Journal of Personality and Social Psychology, 50(5), 889-898. Bigler, R. S . , & Liben, L. S. (1992). Cognitive mechanisms in children’s gender stereotyping: Theoretical and educational implications of a cognitive-based intervention. Child Development, 6 3 , 1351-1363. Bjorklund, D. F. (1987). How age changes in knowledge base contribute to the development of children’s memory: An interpretive review. Developmental Review, 7, 93- 130. Brickman, P., & Bulman, R. J. (1977). Pleasure and pain in social comparison. In J. M. Suls & R. L. Miller (Eds.), Social comparison processes: Theoretical and empirical perspectives (pp. 149186). Washington, DC: Hemisphere. Bruner, I. S. (1951). Personality dynamics and the process of perceiving. In R. R. Blake & G . V. Ramsey (Eds.), Perception: An approach to personality. New York: Ronald Press. Bruner, J. S . (1957). Going beyond the information given. In J. S. Bruner, E. Brunswik, L. Festinger, F. Heider, K. F. Muenzinger, C. E. Osgood, & D. Rapaport (Eds.), Contemporary approaches to cognition. Cambridge, MA: Harvard University Press. Bussey, K . , & Bandura, A. (1992). Self-regulatory mechanisms governing gender development. Child Development, 63, 1236-1250. Butler, R. (1989a). Mastery versus ability appraisal: A developmental study of children’s observations of peer’s work. Child Development. 60, 1350-1361. Butler, R. (1989b). Interest in the task and interest in peer’s work in competitive and noncompetitive conditions: A developmental study. Child Development, 60, 562-570. Butler, R. (1990). The effects of mastery and competitive conditions on self-assessment at different ages. Child Development, 61, 201-210. Butler, R. (1992). What young people want to know when: Effects of mastery and ability goals on interest in different kinds of social comparisons. Journal of Personality and Social Psychology, 62, 934-943. Caciopp, I. T., Petty, R. E., & Moms, K. J. (1983). Effects of need for cognition on message evaluation, recall, and persuasion. Journal of Personality and Social Psychology, 45, 805-81 8 . Camhy, M., & Ruble, D. N. (in press). Becoming acquainted with a peer: Social-cognitive and motivational influences. Social Development. Carter, D. B. (1987). Current conceptions of sex roles and sex typing: Theory and research. New York: Praeger. Carter, D. B., & Patterson, C. 3. (1982). Sex-roles as social conventions: The development of children’s conceptions of sex-role stereotypes. Developmental Psychology, 18, 8 12-824. Chaiken, S. (1980). Heuristic versus systematic information processing and the use of source versus message cues in persuasion. Journal of Personality and Social Psychology, 39, 752-766. Chaiken, S., Liberman, A , , & Eagly, A. H. (1989). Heuristic and systematic information processing within and beyond the persuasion context. In J. S. Uleman & J. A. Bargh (Eds.), Unintended thought (pp. 212-252). New York: Guilford Press. Chi, M. T. H. (1977). Age differences in memory span. Journal of Experimental Child Psychology, 23, 266-281. Chi, M. T. H., & Ceci, S. J. (1987). Content knowledge: Its role, representation, and restructuring in memory development. Advances in child developmenr and behavior, 20, 91- 142. Clark, M.S., & Mills, J. (1979). Interpersonal attraction in exchange and communal relationships. Journal of Personality and Social Psychology, 37, 12-24.

208

DIANE N. RUBLE

Connell, J. P., & Furman, W. (1984). The study of transitions: Conceptual and methodological considerations. In R. Emde & R. Harmon (Eds.), Continuity and discontinuity in development (pp. 153-173). New York: Plenum. Connell, J. P., & Tero, P. F. (1982). Aspects of continuity and change in children’s self-regulated cognitions and afects within the academic domain. Unpublished manuscript, University of Rochester, Rochester, NY. Costanzo, P. R., & Fraenkel, P. (1987). Social influence, socialization, and the development of social cognition: The heart of the matter. In N. Eisenberg (Ed.), Contemporary topics in developmental psychology. New York: Wiley. Crane, M., & Markus, H. (1982). Gender identity: The benefits of self-schema approach. Journal of Personality and Social Psychology, 43, 1195- 1197. Crocker, J., Hannah, D. B., & Weber, R. (1983). Person memory and causal attributions. Journal of Personality and Social Psychology, 44, 55-66. Damon, W. (1979). The social world of the child. San Francisco: Jossey-Bass. D’Andrade, R. G. (1981). The cultural part of cognition. Cognitive Science, 5, 179-195. Deaux, K. (1991). Social identity: Thoughts on structure and change. In R. C. Curtis (Ed.), The relational self: Theoretical convergences in psychoanalysis and social psychology (pp. 77-93). New York: Guilford Press. Deci, E. L., & Ryan, R. M. (1985). Intrinsic motivation and self-determination in human behavior. New York: Plenum. Deci, E. L., & Ryan, R. M. (1987). The support of autonomy and the control of behavior. Journal of Personality and Social Psychology, 53, 1024-1037. De Lisi, R., & Gallagher, A. M. (1991). Understanding of gender stability and constancy in Argentinean children, Merrill-Palmer Quarterly, 37, 483-502. De Lisi, R., & Johns, M. L. (1984). The effects of books and gender constancy on kindergarten children’s sex-role attitudes. Journal of Applied Developmental Psychology, 5, 173- 184. Deutsch, F. M., Ruble, D. N., Fleming, A., Brooks-Gunn, J., & Stangor, C. S. (1986, April). Becoming a mother: Information-seeking and self-definitionalprocesses. Paper presented at the meeting of the Eastern Psychological Association, New York, NY. Deutsch, F. M., Ruble, D. N., Fleming, A., Brooks-Gunn, J., & Stangor, C. S. (1988). Information seeking and maternal self-definition during the transition to motherhood. Journal of Personality and Social Psychology, 55, 420-43 I . Doise, W., & Mugny, G. (1984). The social development of the intellect. Oxford: Pergamon Press. Doise, W., & Palmonari, A. (1984). (Eds.). Social interaction in individual development. Cambridge: Cambridge University Press. Dweck, C. S. (1991). Self-theories and goals: Their role in motivation, personality, and development. In R. Dienstbier (Ed.), Nebraska symposium on motivation: Vol 38, Perspective on motivation (pp. 199-235). Lincoln, NE: University of Nebraska Press. Dweck, C. S., & Leggett, E. L. (1988). A social-cognitive approach to motivation and personality. Psychological Review, 95* 256-273. Eaton, W. O., Von Bargen, D., & Keats, J. G. (1981). Gender understanding and dimensions of preschooler toy choice: Sex stereotype versus activity level. Canadian Journal of Behavioral Science, 13, 203-209. Eccles, J . S., & Midgley, C. (1989). Stage-environment fit: Developmentally appropriate classrooms for young adolescents. In C. Ames & R. Ames (Eds.), Research on motivation in education, Vol 3. Goals and Cognition (pp. 139-186). San Diego, CA: Academic Press, Inc. Eccles, J. S., & Midgely, C. (1990). Changes in academic motivation and self-perception during adolescence. In R. Montemayor, G. Adams, & T. Gullotta (Eds.), From childhood to adolescence: A transitional period? (pp. 134-155). Newbury Park, CA: Sage Publications.

PHASE MODEL OF TRANSITIONS

209

Eccles, J., Midgley, C., & Adler, T. F. (1984). Grade-related changes in the school environment: Effects on achievement motivation. In J. G. Nicholls (Ed.), The development of achievement motivation (pp. 283-332). Greenwich, CT: JAI Press. Eccles, J. S., Wigfield, A,, Flanagan, C. A., Miller, C., Reuman, D. A., & Yee, D. (1989). Selfconcepts, domain values, and self-esteem: Relations and changes at early adolescence. Journal of Personality, 57, 283-310. Emmerich, W. (1982). Nonmonotomic developmental trends in social cognition: The case of gender constancy. In S. Strauss (Ed.), (I-shaped behavioral growth (pp. 249-269). New York: Academic Press. Erber, R., & Fiske, S. T. (1984). Outcome dependency and attention to inconsistent information. Journal of Personality and Social Psychology, 47, 709-726. Fazio, R. H. (1 979). Motives for social comparison: The construction-validation distinction. Journal of Personality and Social Psychology, 37, 1683-1699. Feldman, A. S . , & Ruble, D. N. (1988). The effect of personal relevance on dispositional inference: A developmental analysis. Child Development, 59, 1339- 1352. Festinger, L. (1954). A theory of social comparison processes. Human Relations, 7 , 117-140. Fiske, S . T. (1993). Social cognition and social perception. AnnualReview of Psychology, 44, 155194. Fiske, S. T., & Neuberg, S. L. (1990). A continuum of impression formation, from category-based to individuating processes: Influences of information and motivation on attention and interpretation. Advances in Experimental Social Psychology, 23, 1-74. Fiske, S . T., & Taylor, S. E. (1991). Social cognition. New York: McGraw-Hill. Fleming, A., Ruble, D. N., & Flett, G. L., & Van Wagner, V. (1990). Adjustment in first-time mothers: Changes in mood and mood content during the early postpartum months. Developmental Psychology, 26, 137-143. Flerx, V. C., Fidler, D. S., & Rodgers, R. W. (1976). Sex role stereotypes: Developmental aspects and early intervention. Child Development, 47, 998- 1007. Frey, D. (1986). Recent research on selective exposure to information. Advances in Experimental Social Psychology. 41-80. Frey, K. S., & Ruble, D. N. (1985). What children say when the teacher is not around: Conflicting goals in social comparison and performance assessment in the classroom. Journal of Personality and Social Psychology, 48, 550-562. Frey, K.S., & Ruble, D. N. (1987). What children say about classroom performance: Sex and grade differences in perceived competence. Child Development, 58, 1066- 1078. Frey, K . S . , & Ruble, D. N. (1992). Gender constancy and the “cost” of sex-typed behavior. Developmental Psychology, 28, 7 14-721. Frye, D., & Moore, C. (Eds.) (1991). Children’s theories of mind. Hillsdale, NJ: Erlbaum. Frye, D., Zelazo, P. D., & Palfai, T. (1993). The cognitive basis of rheory of mind. Manuscript submitted for publication. Furman, W., Rahe, D. F., & Hartup, W. W. (1979). Rehabilitation for socially withdrawn preschool children through mixed-age and same-age socialization. Child Development, 50, 9 15-922. Gelman, S. A,, Collman, P., & Maccoby, E. E. (1986). Inferring properties from categories versus inferring categories from properties: The case of gender. Child Development, 57, 396-404. Gochberg, B . , Ruble, D. N., & Higgins, E. T. (1992). The development of depressive symptoms in young school children. Manuscript in preparation. Goethals, G. R. (1986). Social comparison theory: Psychology from the lost and found. Personality and Social Psychology Bulletin, 12, 261-278. Gollwitzer, P. M. (1986). Action phases and mind-sets. In E. T. Higgins and R. M. Sorrentino (Eds.), Handbook of motivation and cognition: Foundations of social behavior (Vol. 2., pp. 5392). New York: Guilford Press.

210

DIANE N. RUBLE

Goodnow, J. J. (1990). Using sociology to extend psychological accounts of cognitive development. Human Development, 33. 81-107. Gouze, K., & Nadelman, L. (1980). Constancy of gender identity for self and others in children between the ages of three and seven. Child Development, 51, 275-278. Graesser, A. C. (1981). Prose comprehension: Beyond the word. New York: Springer. Graziano, W. G., Brody, G. H., and Bernstein, S. (1980). Effects of information about future interaction and peer’s motivation on peer reward allocations. Developmental Psychology, 16, 475-482. Hackel, L. S.,& Ruble, D. N. (1992). Changes in the marital relationship after the first baby is born: Predicting the impact of expectancy disconfirmation. Journal of Personality and Social Psychology, 62, 944-957. Harter, S. (1981). A model of mastery motivation in children: Individual differences and developmental change. In A. Collins (Ed.), Minnesota Symposium on Child Psychology (Vol. 14, pp. 215-255). Hillsdale, NJ: Erlbaum. Harter, S. (1982). The perceived competence scale for children. Child Development, 53, 87-97. Hastie, R., & Kumar, P. A. (1979). Person memory: Personality traits as organizing principles in memory for behavior. Journal of Personality and Social Psychology, 37. 25-38. Heckhausen, H., & Gollwitzer, P. M. (1986). Information processing before and after the formation of an intent. In F. Llix & H. Hagendorf (Eds.), In Memoriam Hermann Ebbinghuus: Symposium on the Structure and Function of Human Memory (pp. 107I- 1082). Amsterdam: Elsevier/North Holland. Higgins, E. T., & Bargh, J. A. (1987). Social cognition and social perception. Annual Review of Psychology, 38, 369-425. Higgins, E. T., & King, G. (1981). Accessibility of social constructs: Information processing consequences of individual and contextual variability. In N. Cantor & J. Kihlstrom (Eds.), Personality, cognition, and social interaction (pp. 69- 121). Hillsdale, NJ: Erlbaum. Higgins, E. T., King, G., & Mavin, G. H. (1982). Individual construct accessibility and subjective impressions and recall. Journal of Personality and Social Psychology, 43, 35-47. Higgins, E. T., Rholes, W. S . , & Jones, C. R. (1977). Category accessibility and impression formation. Journal of Experimental Social Psychology, 13, 144- 154. Higgins, E. T., & Wells, R. S . (1986). Social construct availability and accessibility as a function of social life phase: Emphasizing the “how” versus the “can” of social cognition. Social Cognition, 4 , 201-226. Huston, A. C. (1983). Sex-typing. In E. M. Hetherington (Ed.), Social development: Vol. in P. H. Mussen (Gen. Ed.), Carmicheal’s manual of child psychology (4th ed.). New York: Wiley. Jamieson, D. W., & Zanna, M .P. (1989). Need for structure in attitude formation and expression. In A. R. Pratkanis, S. J. Breckler, & A. G. Greenwald (Eds.), Attitude structure and function. Hillsdale, NJ: Erlbaum. Josephson, J. (1977). The child’s use of situational and personal information in predicting the behavior of another. Unpublished doctoral dissertation, Stanford University, Stanford, CA. Keil, F. C. (1984). Mechanisms of cognitive development and the structure of knowledge. In R. J. Stemberg (Ed.), Mechanisms ofcognitive development (pp. 81-99). New York: W. H. Freeman. Kohlberg, L. (1966). A cognitive-developmental analysis of children’s sex-role concepts and attitudes. In E. E. Maccoby (Ed.), The development of sex differences. Stanford, CA: Stanford University Press. Kruglanski, A. W. (1990). Motivations for judging and knowing: Implications for causal attribution. In E. T. Higgins & R. M. Sorrentino (Eds.), Motivation and cognition (pp. 333-368). New York: Guilford Publications. Kruglanski, A. W., Peri, N., & Zakai, D. (1991). Interactive effects of need for closure and initial confidence on social information seeking. Social Cognirion. 9, 127- 148.

PHASE MODEL OF TRANSITIONS

21 1

Levy, G. D., & Carter, B. D. (1989). Gender schema, gender constancy, and gender-role knowledge: The roles of cognitive factors in preschoolers’ gender-role stereotyped attribution. Developmental Psychology, 25, 444-449. Lewin, K. (1951). Field theory in social science. New York: Harper and Brothers. Liben, L. S., & Signorella, M. L. (Eds.). (1987). Children’s gender schemata. San Francisco: Jossey-Bass. Livesley, W., & Bromley, D. (1973). Person perception in childhood and adolescence. London: Wiley. Maccoby, E. E., Maccoby, N., Romney, A. K., & Adams, J. S. (1961). Social reinforcement in attitude change. Journal of Abnormal and Social Psychology, 63, 109-1 15. Maccoby, N., Romney, A. K., Adams, J. S . , & Maccoby, E. E. (1959).“Critical periods” in seeking and accepting information. American Psychologist. 14, 358. Maheswaran, D., & Chaiken, S. (199 I). Promoting systematic processing in low-motivation settings: Effects of incongruent information on processing and judgment. Journal of Personality and Social Psychology, 61, 13-25. Marcus, D. E., & Overton, W. F. (1978). The development of cognitive gender constancy and sex role preferences. Child Development. 49. 434-444. Markus, H., Crane, M., Bernstein, S., & Sidali, M. (1982). Self-schemas and gender. Journal of Personality and Social Psychology, 42, 38-50. Markus, H., & Zajonc, R. B. (1985). The cognitive perspective in social psychology. In G. Lindzey & E. Aronson (Eds.), Handbook of Social Psychology (Vol. I , pp. 137-230). New York: Random House. Martin, C. L. (1989). Children’s use of gender-related information in making social judgments. Child Development, 25, 80-88. Martin, C . L., & Little, J. K . (1990). The relation of gender understanding to children’s sex-typed preferences and gender stereotypes. Child Development, 61, 1427- 1439. Martin, C. L., Wood, C. H., & Little, J. K. (1990). The development of gender stereotype components. Child Development, 61, 1891-1904. Mendelson, M. J. (1990). Becoming a brother. Cambridge, MA: MIT Press. Miller, A. (1985). A developmental study of the cognitive basis of performance impairment after failure. Journal of Personality and Social Psychology, 49, 529-538. Miller, J. G. (1984). Culture and the development of everyday social explanation. Journal of Personality and Social Psychology, 46, 96 1-978. Miller, J. G. (1 986). Early cross-cultural commonalities in social explanation. Developmental Psychology, 22, 514-520. Munroe, R. H., Shimmin, H. S., & Munroe, R. L. (1984). Gender understanding and sex role preference in four cultures. Developmental Psychology, 20, 673-682. Neisser, U. ( 1976). Cognition and reality: Principles and implications of Cognitive psychology. San Francisco: W. H. Freeman. Nelson, K. (1986). Event knowledge and cognitive development. In K . Nelson (Ed.), Event knowledge: Structure and function in development (pp. 1- 19). Hillsdale, New Jersey: Erlbaum. Neuberg, S. L., & Fiske, S. T.(1987). Motivational influences on impression formation: Outcome dependency, accuracy-driven attention, and individuating processes. Journal of Personality and Social Psychology, 5 3 , 431-444. Newman, L . S. (1991). Why are traits inferred spontaneously? A developmental approach. Social Cognition, 9, 221-253. Nicholls, J. G. (1984). Achievement motivation: Conceptions of ability, task choice, and performance. Psychological Review, 91, 328-346. Nottelmann, E. (1987). Competence and self-esteem during the transition from childhood to adolescence. Developmental Psychology, 23, 44-450.

212

DIANE N. RUBLE

Olson, J. M., Ellis, R. J., & Zanna, M. P. (1983). Validating objective versus subjective judgments: Interest in social comparison and consistency information. Persondig and Social Psychology Bulletin, 9, 427-436. Ornstein, P. A., & Naus, M. J. (1985). Effects of the knowledge base on children’s memory strategies. In H. W. Reese (Ed.), Advances in child development and behavior (Vol. 19). New York: Academic Press. Paley, V. G. (1984). Superheroes in the doll corner. Chicago: The University of Chicago Press. Paley, V. G. (1988). Bad Guys Don’t Have Birthdays. Chicago: The University of Chicago Press. Parsons, J. E., & Ruble, D. N. (1977). The development of achievement-related expectancies. Child Development, 48, 1075- 1079. Petersen, A. C., & Crockett, L. (1985). Pubertal timing and grade effects on adjustment. Journal of Youth and Adolescence. 14, 191-206. Piaget, J. (1985). The equilibration of cognitive structures: The central problem of intellectual development. Chicago: University of Chicago Press. (Original work published 1975) Piaget, J. (1971). The theory of stages in cognitive development. In D. R. Green, M. P. Ford, & G. B. Flamer (Eds.), Measurement and Piager (pp. 1-11). New York: McGraw-Hill. Pittman, T. S . , & D’Agostino, P. R. (1985). Motivation and attribution: The effects of control deprivation on subsequent information processing. In G. Weary & J. Harvey (Eds.), Attribution: Basic issues and applications (pp. 117- 141). New York: Academic Press. Pittman, T. S . , & Heller, J. F. (1987). Social motivation. Annual Review of Psychology, 38, 461489.

Pornerantz, E. M., Ruble, D. N., Frey, K. S . , & Greulich, F. (1993). Meeting goals andconfronting conjlicr: Children’s changing perceptions of social comparison. Manuscript submitted for publication. Rholes, W. S . , Jones, M., &Wade, C. (1988). Children’s understanding of personal dispositions and its relation to behavior. Journal of Experimental Child Psychology, 45, 1-17. Rholes, W. S . , Newman, L. S., & Ruble, D. N. (1990). Understanding self and other: Developmental and motivational aspects of perceiving persons in terms of invariant dispositions. In R. M. Sorrentino & E. T. Higgins (Eds), Handbook of motivation and cognition, Vol. 2 . New York: Guilford Press. Roberts, T. (1991). Gender and the influence of evaluations on self-assessments in achievement settings. Psychological Bulletin, 109, 297-308. Ruble, D. N. (1987). The acquisition of self-knowledge: A self-socialization perspective. In N. Eisenberg (Ed.), Contemporary topics in development psychology. New York: Wiley. Ruble, D. N. (1988). Sex-role development. In M. H. Bomstein & M. E. Lamb (Eds.), Developmental psychology: An advanced rexrbook. Hillsdale, NJ: Lawrence Erlbaum. Ruble, D. N., Balaban, T., & Cooper, J. (1981). Gender constancy and the effects of sex-typed televised toy commercials. Child Development, 52, 667-673. Ruble, D. N., Eisenberg, R., & Higgins, E. T. (in press). Developmental changes in achievement evaluation: Motivational implications of self-other differences. Child Development. Ruble, D. N., Fleming, A. S . , Hackel, L., & Stangor, C. S . (1988). Changes in the marital relationship during the transition to motherhood: Effects of violated expectations concerning division of household labor. Journal of Personaliry and Social Psychology, 55, 78-87. Ruble, D. N., & FIett, G. L. (1988). Conflicting goals in self-evaluative information seeking: Developmental and ability level analyses. Child Development, 59, 97- 106. Ruble, D. N., & Frey, K. S. (1991). Changing patterns of comparative behavior as skills are acquired: A functional model of self-evaluation. In J. Suls and T. A. Wills (Eds.), Social Comparison: Contemporary rheory and research. (pp. 79- 1 13). Hillsdale, NJ: Erlbaum. Ruble, D. N., Greulich, F., Pomerantz. E. M., & Gochberg, B. (in press). The role of gender-related processes in the development of sex differences in self-evaluation and depression. Journal of Affective Disorders.

PHASE MODEL OF TRANSITIONS

213

Ruble, D. N., Pomerantz, E. M.,Frey, K. S., & Greulich, F. (1993). Explaining the decline in perceived competence: Grade and sex direrences in the impact of perceptions of ability as stable. Manuscript submitted for publication. Ruble, D. N., & Stangor, C. (1986). Stalking the elusive schema: Insights from developmental and social analyses of gender schemas. Social Cognition, 4. 227-261. Schneider, W. (1990). The development of memory in children. Newsletter of the lnternational Society for the Study of Behavioral Development (No. 2, Serial No. 18, pp. 1-3). Serbin, L. A,, & Spratkin, C. (1986). The salience of gender and the process of sex typing in threeto seven-year-old children. Child Development, 57, 1188-1 199. Siegler, R. (1981). Developmental sequences within and between concepts. Monographs of the Society for Research in Child Development, 46, ( 2 , Whole No. 189). Signorella, M. L., Bigler, R. S . , & Liben, L. S. (1993). Developmental differences in children’s gender schemata about others: A meta-analytic review. Developmental Review, 13(2), 147-183. Signorella, M. L., & Liben, L. S. (1985). Assessing children’s gender-stereotyped attitudes. Psychological Documents, 15, 7. Simmons, R. G . , & Blyth, D. A. (1987). Moving into adolescence: The impact of pubertal change and school context. Hawthorn, NY: Aldine de Gruyler. Simmons, R. G., Rosenberg, F., & Rosenberg, M. (1973). Disturbance in the self-image at adolescence. American Sociological Review, 38, 553-568. Slaby, R., & Frey, K. (1975). Development of gender constancy and selective attention to same-sex models. Child Development, 46, 849-856. Stangor, C., & Ford, T. E. (1992). Accuracy and expectancy-confirming processing orientations and the development of stereotypes and prejudice. European Review of Social Psychology, 3, 5789. Stangor, C., & McMillan, D. (1992). Memory for expectancy-congruent and expectancy-incongruent information: A review of the social and social developmental literatures. Psychology Bulletin. I1 I , 432-461. Stangor, C.. & Ruble, D. N. (1987). Development of gender role knowledge and gender constancy. In L. S. Liben and M. L. Signorella (Eds.), New directions for child development: Vol. 39. Children’s gender schemata. San Francisco: Jossey-Press. Stangor, C., & Ruble, D. N. (1989a). Differential influences of gender schemata and gender constancy on children’s information processing and behavior. Social Cognition, 7, 353-372. Stangor, C., & Ruble, D. N. (1989b). Strength of expectancies and memory for social information: What we remember depends on how much we know. Journal of Experimental Social Psychology, 25, 18-35. Stipek, D., & Maclver, D. (1989). Developmental changes in children’s assessment of intellectual competence. Child Development. 60. 521-538. Stoddart, T., & Turiel, E. (1985). Children’s concept of cross-gender activities. Child Development, 56, 1241-1252. Swann, W. B., Jr. (1990). To be adored or to be known?: The interplay of self-enhancement and selfverification. In E. T. Higgins & R. M. Sorrentino (Eds.), Handbook of motivation and cognition: Foundations ofsocial behavior (Vol. 2, pp. 408-448). New York: Guilford Press. Taylor, S. E., & Crocker, J. (1981). Schematic bases of social information processing. In E. T. Higgins, C. P. Herman, & M. P. Zanna (Eds.), Social cognition: The Ontario Symposium (Vol. 1, pp. 89- 134). Hillsdale, New Jersey: Erlbaum. Thompson, E. P., Boggiano, A. K., Costanzo, P. E., Matter, J. A., & Ruble, D. N. (1993). Agerelated changes in children‘s orientation toward strategic peer interaction; lmplications for social perception and behavior. Manuscript submitted for publication. Thompson, E. P., Chaiken, S., & Hazlewood, J. E. (1993). Need for cognition and desire for control as moderators of extrinsic reward effects: A person x situation approach to the study of intrinsic motivation. Journal of Personality and Social Psychology, 64, 987-999.

214

DIANE N. RUBLE

Thornburg, H. D., & Glider, P. (1984). Dimensions of early adolescent social perceptions and preferences. Journal of Early Adolescence, 4 , 387-406. Trope, Y. (1983). Self-assessment in achievement behavior. In J. Suls & A. G. Greenwald (Eds.), Psychological perspectives on the self (Vol. 2 , pp. 93-121). Hillsdale, NJ: Erlbaum. Trope, Y. (1986). Self-enhancement and self-assessment in achievement behavior. In R. Sorrentino & E. T. Higgins (Eds.), Handbook of motivation and cognition (Vol. 1, pp. 350-378). New York: Guilford. Trope, Y., & Thompson, E. P. (1993). Seeking the truth in all the wrong places: The effects of accuracy motivation and social category knowledge in social information seeking. Manuscript Submitted for publication. Veroff, J. (1969). Social comparison and the development of achievement motivation. In C. P. Smith (Ed.), Achievement-related motives in children. New York, NY: Russell Sage. Vygotsky, L. S. (1962). Thought and language. Cambridge, MA: MIT Press. (Original work published 1934) Wehren, A,, & De Lisi, R. (1983). The development of gender understanding: Judgments and explanations. Child Development, 54, 1568- 1578. White, R. W. (1959). Motivation reconsidered: The concept of competence. Psychological Review, 66, 297-333. Wigfield, A., Eccles, I. S., MacIver, D., Reuman, D. A., & Midgley, C. M. (1991). Transitions at early adolescence: Changes in children’s domain-specific self-perceptions and general selfesteem across the transition to junior high school. Developmental Psychology, 27, 552-565. Wood, J:V. (1989). Theory and Research Concerning Social Comparisons of Personal Attributes. Psychological Bulletin, 106, 231-249. Wright, J. C., & Mischel, W. (1988). Conditional hedges and the intuitive psychology of traits. Journal of Personality and Social Psychology, 55, 454-469.

MULTIPLE-AUDIENCE PROBLEMS, TACTICAL COMMUNICATION, AND SOCIAL INTERACTION: A RELATIONAGREGULATION PERSPECTIVE John H. Fleming

I. Introduction In this article I have three agendas. First, I seek to resurrect a metaphor deeply woven into the fabric of social psychology. Regrettably, the metaphor has languished neglected by contemporary approaches to interpersonal perception and social interaction. Second, I describe a class of social dilemmas that current theories of communication, impression management, self-presentation, and interpersonal perception largely ignore. In so doing, I offer the metaphor as a heuristic and useful organizational framework for considering several varieties of human tactical communication and social interaction behavior. Specifically, I examine the tactical-communicativeand self-concept maintenance properties of relational-regulation * behaviors (distancing and embracing behaviors, and mixed or hidden message communications). Finally, I survey a program of experimental research that examines both the inter- and intrupersonal consequences of relational regulation. This research suggests that applying a relational-regulation perspective to interpersonal behavior may help organize and synthesize research and theory from a number of disparate areas including dispositional inference, self-perception, impression management, psycholinguistics, symbolic interactionism, and interpersonal communication. 'The term relarional has been used in several different contexts in the psychological and communication literatures. I use the term here to refer to the ways in which elements in the social field are "positioned" in relation to one another (i.e., their proximity, relatedness, clustering, association, etc.) rather than as a term to refer to the dynamics of interpersonal relationships per se (i.e., as in relational communication in close relationships). 215 ADVANCES IN EXPERIMENTAL SOCIAL PSYCHOUXiY. VOL. 26

Copyright 0 1994 by Academic Press, Inc. All rights of reproduction in any form reserved.

216

JOHN H. FLEMING

11. Conceptual Overview of the Relational-Regulation Perspective

The twin notions that one can “distance” oneself from an objectionable association or “embrace” a desired one are both familiar and intuitively appealing. Political commentators frequently speak of candidates “distancing themselves” from some unfortunate outcome, a scandalous event, an unseemly act, or an unsavory individual. Newscasters speak of politicians “embracing” a worthy cause, a fashionable ideology, or an influential group. In terms of this “relational” metaphor, the proximity of an embrace implies relatedness or association; distance implies separation, distinctiveness, or lack of association. The relationship between the observed proximity of objects and their inferred association easily ranks as one of the fundamental elements of human perception and judgment. For example, in his Critique of Pure Reason (1781/1965), Kant considered relation and quantity to be among the primary (or a priori) organizing principles through which individuals experience the world. James (1890) accorded similar status to the psychological processes of association (grouping) and discrimination (discerning). The temporal and physical proximity of paired stimuli formed the core of Pavlov’s pioneering work on classical conditioning, and the work of several generations of researchers who followed him (Razran, 1938, 1940). The Gestalt theorists, too, considered the proximity-distance continuum one of the elemental dimensions of human perception (e.g., Kof€ka, 1935; Kohler, 1940; Wertheimer, 1923). The “relational” metaphor, however, extends well beyond the bounds of physical object perception. Sociological and social-psychologicaltheorists have also embraced the metaphor as a useful tool for describing abstract and symbolic associations. Indeed, the metaphor has a distinguished history within these disciplines. Among sociologists, the sheer explanatory power of the relational metaphor in describing the tactical2 elements of human social behavior has been long recognized, most notably within symbolic interactionism (for overviews see Meltzer, Petras, & Reynolds, 1970; Stryker, 1980; Stryker & Statham, 1985) and its offshoot, the dramaturgical perspective in role theory and ethnomethodology (Berger, 1963/1975; Garfinkel, 1967; Goffman, 1961; Messinger, Sampson, & Towne, 1962). Perhaps nowhere is this recognition more strongly realized than in the theoretical work of Goffman, possibly the best-known purveyor of the rela-

2Following Tedeschi and Norman (1985), the term rncrical refers to behaviors enacted to achieve a specific short-term goal within a relatively circumscribed interaction. Srruregic behaviors, on the other hand, refer to those behaviors that are focused more on long-term goals across a variety of behavioral episodes.

RELATIONAL REGULATION

217

tional metaphor, whose The Presentation of Self in Everyday Life (1959), Encounters ( 1961), and Strategic Interaction ( 1969) are exemplars of the approach. Among social-psychological theorists, relative distance and proximity figured prominently in Lewin’s (1935) field theoretical formulations, in Tagiuri (1952) and Peak’s (1958) early approaches to person perception, and in Heider’s (1958) conception of unit formation. These ideas also informed Bogardus’s (1925) work on prejudice and social distance. Somewhat later, they formed the basis of Mehrabian and his colleagues’ work on immediacy and the implicit communication of attitudes (Mehrabian, 1972; Wiener & Mehrabian, 1968). More recently, variants of the relational (physical, psychological, or symbolic association) metaphor have been invoked both implicitly and explicitly to describe techniques for privacy regulation (Burgoon, Parrott, LePoire, Kelley, Walther, & Perry, 1989; Harris, 1990); maintaining social status hierarchies (Aronsson & SatterlundLarsson, 1987; Brown & Levinson, 1978; Ferguson, 1985; Lott, 1987, 1989; Word, Zanna, & Cooper, 1974); terror management (Pyszczynski, Greenberg, Solomon, Sideris, & Stubing, 1993); dyadic communication and interaction (Kaplan, 1975; Kaplan, Firestone, Klein, & Sodikoff, 1983; Markus-Kaplan & Kaplan, 1984); coping with stress (Folkman & Lazarus, 1985; Folkman, Lazarus, Dunkel-Schetter, DeLongis, & Gruen, 1986; Harris, 1990; Lazarus, 1993; Showers & Ruben, 1990; Stevens, 1989); reacting to one’s own or another person’s success or failure (Hirt, Zillmann, Erickson, & Kennedy, 1992; Snyder, Lassegard, & Ford, 1986); and responding to social anxiety and embarrassment (Bond & Lai, 1986; Buss, 1980; Edelmann & Hampson, 1979, 1981; Patterson, 1987). These contributions represent only a sample of the relational constructs that researchers and theorists have offered to describe a range of social phenomena. At an abstract level, each suggests that the way in which elements in the social field are “positioned” in relation to one another exerts a tremendous influence on people’s reactions to those elements. That is, the relative positioning of elements in the social field provides an organizing framework that helps structure and channel social perception and helps guide social inference and self-perception. In attributional terms, the inferred association between a person and his or her own behavior, the behavior of another person, or a situation is at least partly a function of the distance or proximity implied to exist between those elements (Heider, 1958). Thus, implicit distance or proximity can be a powerful determinant of the attributions observers will make about these various elements. I would like to pick up this metaphorical thread and revive the relational metaphor as a useful perspective from which to describe and understand the nature of tactical social communication processes, their antecedents, and their consequences for interpersonal perception, self-concept maintenance, self-regulation, and social interaction.

218

JOHN H. FLEMING

A. THE (APOSTLE) PETER PRINCIPLE AND SEQUELAE As C. R. Snyder et al. (1986, p. 383) note, the ability to communicate tactically (or to “package” impressions and information for specific audiences) hinges on attribution theorists’ (Heider, 1958; Jones & Nisbett, 1972) presumption that individuals function simultaneously as both actor and observer of their own and others’ behavior. Thus actors are aware of the associations that observers use in their evaluations, and, in knowing the rules of evaluation, they are aware of the means by which they can influence these evaluations; that is by claiming (or severing . . . ) associations with others, actors may influence how they (as “objects”) are valuated in social situations.

This observation and the foregoing discussion suggest that the relational (proximity-distance) continuum is one fundamental dimension to which observers are particularly sensitive and along which they may attempt to package or to adjust information or impressions. Thus, under a variety of circumstances, people may actively attempt to regulate how they stand in relation to some aspect of their own behavior, the behavior of others, or the social situations in which they find themselves to manipulate or manage the impressions others may form (e.g., Cialdini, Borden, Thorne, Walker, Freeman, & Sloan, 1976; Schlenker & Weigold, 1992; Tesser & Rosen, 1975). The Christian gospels, for example, tell the story of how the Apostle Peter three times denied any knowledge of Christ in the hours immediately before Jesus’s crucifixion. Peter distanced himself from Jesus to avoid being punished for their association. Like latter-day incarnations of Peter the Apostle, those who use the proximity-distance continuum to “distance themselves” from some objectionable event, person, situation, or behavior attempt to convince others (and perhaps themselves) of their limited association with a disavowed entity or action by regulating (i.e., obscuring, ambiguating, or diminishing) the implied psychological, symbolic, or physical distance between themselves and the entity or action (e.g., Archibald & Cohen, 1971; Davis & Jones, 1960; Milgram, 1963, 1974; Pyszczynski et al., 1993; C. R. Snyder et al., 1986; M. L. Snyder & Wicklund, 1981; Tesser & Rosen, 1975). These disavowals may be either explicit, such as when a person overtly denies an association, reads a contrived or constrained statement in a blatantly insincere or stilted fashion, or communicates a concealed message, or implicit, such as when a person feigns dissimilarity with a disliked, but similar other person, or retains an uninvolved and detached demeanor when interacting with another person.3 31n contrast, those who “embrace” a desired entity wish to be strongly associated with that entity (Goffman, 1959, 1961). These individuals attempt to reduce symbolically the implied distance

RELATIONAL REGULATION

219

Obviously, these attempts can take a number of different forms. By applying a relational-regulation perspective, however, a collection of disparate findings from a variety of different theoretical contexts can be reconceptualized. For example, in a series of investigations of the mum effect, Rosen, Tesser, their colleagues, and others have found that although people do not hesitate to communicate good news to another person (and seem to revel in the fact that the news is good, including that fact in their statements), they are extremely reluctant to communicate bad news, even when they are not its originators (Bond & Anderson, 1987; Conlee & Tesser, 1973; Fisher, 1979; Ilgen & Knowlton, 1980; R. Johnson, Conlee, & Tesser, 1974; Rosen & Tesser, 1970, 1972; Tesser & Rosen, 1975; Tesser, Rosen, & Batchelor, 1972; Tesser, Rosen, & Tesser, 1971). When they finally do communicate the bad news, people tend to do so only peripherally and indirectly. From a relational-regulation perspective, these findings make intuitive sense: Communicators attempt to regulate the extent to which they are associated with the news to protect or enhance their own self-images and to manage the impressions others may form of them (Manis, Cornell, & Moore, 1974). Communicators of bad news try to institute and maintain some distance from it, whereas communicators of good news wish to associate themselves closely with it. At some level, communicators are apparently aware of both the benefits that accrue to those who carry good tidings and the opposing ancient prescription to slay the messenger. Moreover, they use this knowledge to manage their impressions. Similarly, Cialdini, Snyder, and their colleagues (Cialdini et al., 1976; Cialdini & DeNicholas, 1989; Cialdini & Richardson, 1980; Folkes, 1982; C. R. Snyder, Higgins, & Stucky, 1983; C. R. Snyder et al., 1986) have found that college students engage in tactical relational-regulation behaviors by basking in the reflected glory (BIRGing) of their respective football teams. A greater percentage of students, for example, donned school shirts the Monday following a victory by their team than following a defeat. Moreover, students embraced the team in their descriptions when their team won: “We killed ‘em!” In contrast, students distanced themselves from their football teams when describing a loss by depersonalizing their relationship to the team (by CORFing-cutting off reflected failure; C. R. Snyder et al., 1983): “They threw away our chance for a national championship” (Cialdini, 1988, pp. 188-190; also see Wiener & Mehrabian, 1968, for a description of various verbal embracing and distancing between themselves and the desired entity so that they and others will see them as highly associated. Although embracing behaviors and their effects are interesting tactical communication phenomena, much of our research and theorizing has focused on their behaviorally richer and arguably more interesting relational counterparts, distancing behaviors. Consequently, in the remainder of this article I will focus primarily on distancing behaviors, except in those cases in which we have specifically included and examined embracing behaviors in our research, or when their description adds to the discussion of distancing.

220

JOHN H.FLEMING

techniques). Recently, researchers have linked the relational-regulation processes of BIRGing and CORFing to changes in individuals’ expectations for their own future success or failure (Hirt et al., 1992). Specifically, Hirt et al. (1992) found that when a team won, sports fans’ estimates of their team’s and their own future performances were significantly greater than estimates made by fans whose teams had lost. Besides the mum effect and the BIRGing literature, the relational-regulation perspective can also be used to reconceptualize a third impression-management phenomenon from the interpersonal-attraction literature. That research has generally found that similarity leads to attraction, except when the similar other possesses some undesired or negative attribute (Cooper & Jones, 1969; Lerner & Agar, 1972; Novak & Lerner, 1968; Taylor & Mettee, 1971; Wagner, Wicklund, & Shaigan, 1990). Under these circumstances, people try to “distance” themselves from the disliked or negatively evaluated, but similar, other by altering the implied association (similarity) between themselves and the other. Novak and Lerner (1968), for example, found that subjects physically distanced themselves from similar others who they thought were mental patients. Subjects in the Cooper and Jones (1969) research altered their attitudes to make them appear less similar to a disliked other; Taylor and Mettee (1971) found a similar effect for judgments of liking. Wagner et al. (1990) found that subjects devalued a similar, but incompetent, other on dimensions related to the incompetence. Finally, an intriguing finding from the cognitive dissonance literature suggests that relational-regulation behaviors may play a role in neutralizing attitude change resulting from counter-attitudinal behavior-as when one delivers an untrue and extremely derogatory evaluation of another person (Davis & Jones, 1960). Davis and Jones had subjects provide such an evaluation of a “phantom other” student under conditions of either high or low choice. The phantom student was ostensibly listening over an intercom. They noted that subjects could cancel or “take back” their dissonance-arousing behavior (i.e., the derogatory evaluation) by signaling to the phantom student or by distancing themselves from their evaluations-giving “cues” that their performances were other than genuine. However, rather than examining such behavioral “neutralization” techniques directly, Davis and Jones operationalized cancellation by leading subjects to believe that they would have an opportunity to meet with the derogated student at the conclusion of the experiment. Davis and Jones found that only subjects in the high-choice condition who believed their behavior to be irrevocable experienced enough dissonance to create the predicted opinion shift. Subjects who expected to be able to take back their derogatory behavior, however, did not show dissonance-induced attitude change; the opportunity to undo or cancel out the behavior they had freely chosen prevented these subjects’ negative behavior from producing its self-concept altering effects. As Davis and Jones noted, it follows that on-line behavioral cancellations should be just as effective in reducing

RELATIONAL REGULATION

22 1

dissonance-induced attitude change as their irrevocability manipulation. This small sample of striking examples illustrates how social actors can use relationalregulation behaviors tactically for impression management and self-presentational purposes in a wide range of contexts. The relational-regulation metaphor not only provides a useful framework with which to reconceptualize and organize a number of mainstream social psychological phenomena (such as those above) but it can also serve as a springboard from which to extend thinking about some of the central concerns in contemporary social psychology-tactical communication behavior, social interaction, self-regulation, and interpersonal perception. In the remainder of this article I would like to explore a number of these possibilities by describing a program of experimental research that focuses on the inter- and intrapersonal implications of two broad but interrelated classes of relational-regulation behaviors: verbal and nonverbal distancing cues; and techniques for communicating mixed or hidden messages.

B. TACTICAL COMMUNICATION AND MULTIPLEAUDIENCE PROBLEMS: IMPLICIT RELATION AS INFERENTIAL HEURISTIC It is now considered a virtual maxim of human social behavior that people engage in tactical communication-they “package” information or impressions for consumption by various audiences to create conditions that will further their own purposes or achieve their goals in relation to those audiences. A variety of techniques for doing so have been advanced (Jones & Pittman, 1982; Schlenker & Weigold, 1992). Thus, in its broadest sense, tactical communication is social behavior designed to communicate something (a self-image, an opinion, an explanation, an internal state, or a piece of information) to a specific audience4 in order to further the communicator’s purposes with that audience. Tactical communication thus defined encompasses a broad spectrum of social behavior from lying to persuasion; from concerns over politeness to the offering of accounts, excuses, disclaimers, and explanations; and from impression-management tactics to self-presentational techniques, among a host of others (Cody & McLaughlin, 1990; Schwartz & Strack, 1991). In broad terms, all tactical communication and impression-management techniques can be conceptualized as attempts to negotiate situated identities with those with whom individuals inter4By an audience I mean a group comprising any number of potential observers who share some common attributes and differ from other audiences in these attributes, irrespective of whether the behavior is intended for them. Although an audience generally refers to a collection of observers, an audience can also have as few as one member. Consistent with Greenwald and Breckler’s (1985) view, one of these audiences may be (and, in fact, frequently is) the self.

222

JOHN H. FLEMING

act (Goffman, 1959). In more familiar attributional terms, these tactics are all characterized by actors’ desires to have observers make specific dispositional attributions about their behavior, or perhaps to prevent observers from drawing dispositional inferences altogether. Obviously, no single theoretical account could reasonably encompass and organize all of these varieties of tactical communication processes. In this article I shall restrict myself to a discussion of a limited subset of these processes: The inter- and intrupersonal consequences of relational-regulation behaviors (i.e., distancing and embracing) brought on by the presence of multiple audiences. To foreshadow the main themes I will address, I would like to suggest that a variety of these verbal and nonverbal behaviors (sometimes intentional and sometimes inadvertent) can signal to observers that a person wishes to place psychological or symbolic distance between himself or herself and some undesirable or otherwise self-concept-threateningbehavior or association. By engaging in relational regulation, individuals can shield themselves both from the possible sanctions of others and from the potentially self-altering effects of their own (and possibly others’) negative self-discrepant behavior.

C. HOW THE WEST(ERN) WAS WON: DISTANCING AND DIALOGUE

My work on the tactical uses of relational regulation has grown out of a more specific interest in multiple-audienceproblems (Fleming, Darley, Hilton, & Kojetin, 1990; Fleming & Darley, 1989, 1991; Fleming & Kojetin, 1993; Fleming & Rudman, 1993; Fleming & Scott, 1991). Multiple-audience problems are acute social communication conflicts in which social actors may be motivated to try to convey different impressions or pieces of information to different audiences when the actors’ behavior is constrained by the presence of a particular audience. The constraint may stem from role demands, self-presentationalconcerns, or in some cases, coercion. Whatever its origins, the constraint imposed by one of the witnessing audiences may cause actors to behave in potentially contradictory and/or self-discrepant ways. That is, social actors in multiple-audience situations may attempt to communicate different messages to simultaneously present audiences. I offer the multiple-audience problem both as a paradigmatic example of a complex social communication dilemma and as a platform for examining the antecedents and consequences of relational regulation. Consider the following example to illustrate the utility and effectiveness of relational regulation in a real-world multiple-audience problem. During the heyday of the Hollywood western, filmmakers frequently stereotyped native Americans as the “bad guys” of the American West who invariably suffered defeat at the hands of the cavalry (“Hollywood’s war,” 1990). True native Americans

RELATIONAL REGULATION

223

rarely received the opportunity to play the roles of “Indians.” When they did, it was customary for them to say a few lines of unscripted and untranslated dialogue in their native language as they talked among themselves to lend a veneer of “authenticity” to the film. This situation also presented something of a multiple-audience problem for the native American actors in the film. Although their roles largely constrained their on-screen behavior, they needed to indicate the constrained nature of their performances to other native Americans who might see the film. In a recent letter to the New York Times, a film-going minister from South Dakota noted that in spite of their caricatured depictions, the Indians may have actually won a few battles even in the “bad old days” of John Wayne westerns, and certainly more often than most people realize (Garvey, 1990). In his letter, the minister described how as a teenager he puzzled over the audience’s reaction to the most recent Hollywood western being shown at the local cinema. Many in the audience were Lakota and were conversant in their native language. The minister’s puzzlement arose because the Lakota in the audience would unpredictably break into riotous laughter at various points during the film; points that neither he nor other non-Lakota in the audience found particularly amusing. It was only many years later that he realized that the actors had used the unscripted native dialogue as an opportunity to lampoon the film and simultaneously to distance themselves from their prescribed roles, reducing the stereotyped celluloid caricatures of native Americans to little more than farce. Lines that had originally been intended to imply “I see the smoke of their camp fire. We attack at dawn,” for example, were transformed into sarcastic send-ups: “This will be the third time I’ve died in this picture . . . it’s a good thing we all look alike,” and so on. Because they spoke the language, the Lakota in the audience understood these jabs and laughed. Those who did not speak the language were excluded from these “inside jokes .”

D. PRESENCE OF MULTIPLE AUDIENCES This example from films about the “epic West” illustrates in a general way some of the themes central to understanding how social actors can use relational regulation tactically to manage the impressions that observers form from the actor’s public behavior. One important consideration contained in the example concerns the nature of the audience members who witness (or are likely to witness) one’s behavior. As Jones and Pittman (1982) note, impressionmanagement attempts are usually intended to accomplish a specific purpose with a particular audience. Self-presenters exploit some desired image to manipulate the inferences that a particular and single target audience will make about them. In relational terms, self-presenters wish to “embrace” a particular set of attributes

224

JOHN H.FLEMING

for a specific audience. As Goffman and others aptly point out, frequently more than one audience may be present to witness a person’s self-presentational“performance” at a given time (i.e., Clark & Schaefer, 1987). To complicate the situation further, the purposes individuals may have for a particular audience may be at odds with their purposes for another witnessing audience; all performances are not appropriate for all audiences. Under these circumstances individuals confront an impression-management dilemma characterized by a conflict between the various performances they desire to give and the consequences of those performances within the various audiences (Barker, 1942; Dollard & Miller, 1950; Lewin, 1935, 1938). In other words, these kinds of situations are ones in which social actors may wish to regulate the extent to which one or all of the present audiences dispositionally associates them with their actions.

E. EXPERIMENTAL INVESTIGATIONS OF THE MULTIPLE-AUDIENCE PROBLEM

To examine experimentally the ways in which communicators attempt to resolve multiple-audience problems like those faced by native American film actors, and to assess the effectiveness of those tactics, my colleagues and I have conducted a number of studies in which we manipulated the nature of the subjects’ communication task and the audiences addressed by a constrained communication (Fleming et al., 1990; Fleming & Darley, 1991). Several of these studies have examined how senders communicate mixed messages-messages that contain both an explicit surface content and a covert (and usually contradictory) hidden content. 1 . Formulating and Communicating Hidden Messages

In the initial study in the series (Fleming et al., 1990, Study l), randomly assigned introverted and extroverted author-subjects wrote pairs of essays presenting themselves as either extreme introverts or extreme extroverts. Authors initially prepared an essay under typical assigned essay instructions designed to convince an undefined audience that they held personalities that were the opposite of their true personalities (the “standard” instructions condition). After completing the first essay, authors prepared a second essay in which, on a surface level, they were to leave the original content the same as in their first essays. At the same time, however, they tried to communicate their true personalities in the essays (the “communicate” instructions condition). A group of reader-subjects then read and rated these essays. Results suggested that under the “standard” set of instructions, authors constrained to present a personally discrepant self-image reported no particular

RELATIONAL REGULATION

225

motivation to try to communicate an accurate image of themselves. When we instructed authors to try to get their true personalities across, however, they reported greater motivation to do so and reported that they had manipulated the credibility of their essays to convey their true personalities. The unspecified audience of readers reported that they relied on their perceptions of essay credibility in arriving at estimations of the authors’ true personalities for essays written under both the “standard” and “communicate” sets of instructions. Authors and readers, however, did not agree on what constituted credibility, and the approaches reported by these subjects failed to converge. Consequently, authors were unable to communicate their true personalities successfully. Readers, in turn, were unsuccessful at estimating the authors’ true personalities, and, on average, judged the authors to hold personalities consistent with those presented in the essays. This was true even for those essays in which the authors had attempted to communicate their true personalities surreptitiously. At least in written communications, authors apparently have difficulty manipulating credibility in ways that readers can easily understand and in ways that facilitate communicating hidden content. These subjects’ communicative shortfalls stand in rather stark contrast to other findings that suggest that communicators can sometimes display great skill in such situations. For example, observers in a study conducted by Zuckerman, DeFrank, Hall, Larrance, and Rosenthal (1979) detected posed deception from the facial and vocal cues of videotaped targets at levels significantly better than chance. Zuckerman et al. (1979) instructed subjects to tell a lie, but to communicate their deception simultaneously through their facial expressions. Observers were apparently quite capable of deciphering these kinds of embedded cues and of using them to make inferences about the targets’ behavior. What can account for readers’ failure to detect and decode the hidden messages that authors reported embedding in their essays? First, the medium used (i.e., written essays) may not have been ideal for covert communication in that particular situation. In addition, taking a cue from theories of language use, another possible explanation was that despite their valiant attempts to manipulate the credibility of their written essays, authors knew very little about the specific audience they addressed in their covert communication attempts. These two constraints severely restricted authors’ covert communication repertoires and thereby limited the effectiveness of their attempts.

2 . The Role of Audience Specification and “Shared Knowledge” In a follow-up study (Fleming et al., 1990, Study 2), author-subjects prepared constrained essays in which they told the truth, lied, or attempted to convey a mixed message about a set of items. To create conditions analogous to those

226

JOHN H . FLEMING

confronted by communicators in multiple-audience problems, the authors were explicitly told that a group of strangers and a group of their friends would read the essays. One of the items in each set was randomly preselected as the “target” item, and authors were told which item had been so designated. The authors’ task in the truth trials was to convince all readers to select the target item. In the lie trial they were to convince all readers to select an item other than the target. Finally, in the mixed or multiple-message trial they were to convince their friends to select the target while simultaneously persuading the strangers to select some other item. In other words, the communicator’s task was to construct a single message that would be interpreted differently by each of the witnessing audiences. The readers’ task was to guess which of the four items on each trial had been randomly assigned as the “target.” These groups of readers then read the essays. Results are presented in the upper panel of Fig. 1 and showed that when they were motivated to do so and when the audiences were explicitly described, authors communicated hidden messages in their written essays that the strangers neither detected nor decoded, but that their friends were able to detect and to interpret accurately. As the film example suggests, authors manipulated some aspect of knowledge they shared with the friends-but not the strangers-to convey their hidden messages and to distance themselves from the constrained content of their essays. In an additional study that used virtually identical procedures (Fleming et al., 1990, Study 3), senders made sets of videotaped truth, lie, and multiple-message communications about collections of popular songs. A group of the sender’s friends and two groups of strangers, all of whom were naike about the task the senders had been given, later watched these tapes. One of the groups of strangers was matched to the friends audience on a variety of demographic variables (i.e., age, academic status, etc.). Results are presented in the lower panel of Fig. 1 and indicated that the three groups of observers did not differ in the frequency with which they believed the communicators were lying, nor did they differ in their abilities to correctly select the target song on the truth and lie trials. On the mixed-message trial, however, the friends correctly selected the target song significantly more often than did either group of strangers, despite the fact that the communicator overtly claimed an incorrect song in his or her presentation for that trial. Again, senders generally manipulated some aspect of knowledge they shared with the friends-but not with the strangers-to convey their hidden messages. This research clearly demonstrates that ordinary constrained communicators, like their native American counterparts in western films, can construct messages that convey a specific hidden content to an intended subset of those who receive the communication. Individuals not intended to receive the hidden message did not. Instead, they took the explicit surface content of the message at face value, neither detecting nor decoding its concealed content. Taken together, these studies suggest that a multiple-audience approach to examining tactical relational

227

RELATIONAL REGULATION

100

-

STUDY 2

-L. FRIENDS

80-

a-

-

-

STRANGERS

-

60-

3%

ES

q

40-

-

200-

-

100

STUDY 3

-

Y FRIENDS

80-

+MATCHED STRANGERS UNMATCHED STRANGERS

60-

-

% 340-

-

R E

95

w"S

O 2

-

200. I

I

I

I

r

MESSAGE TRIAL Fig. 1. Perception of hidden messages. Hidden messages were detected and decoded by their intended audiences, but not by others; paradoxically, the intended audience was more easily deceived. Adapted from Fleming, Darley, Hilton, & Kojetin, 1990. Copyright 1990 by the American Psychological Association.

regulation is not only workable but that it also has the potential to produce a substantial empirical and theoretical yield. I shall return to a detailed discussion of covert communication techniques in subsequent sections, but it should now be apparent that multiple-audienceproblems can arise in a number of situations and that a variety of such problems exist. From the stereotyped performances of

228

JOHN H. FLEMING

motion picture actors to propaganda broadcasts made by POWs, all multipleaudience problems have as a central component an acute conflict between the competing motives to conform to the situationally constrained or role-based expectations of one set of witnesses, and to avoid being miscast by another, different set of witnesses for conforming to those expectations. Thus, multipleaudience problems may motivate actors to communicate tactically and covertly the constrained nature of their behavior to some subset of the audiences who witness the behavior.

111. A Tactical Relational-RegulationAnalysis of Role Conflict

A. ROLE THEORY AND ROLE ENACTMENTS Much of the seminal theoretical work on tactical communication comes out of a sociological, rather than a social-psychological tradition, epitomized by the writings of Goffman and other dramaturgical theorists. Briefly, by using a pervasive theatrical metaphor, role theorists construe individuals as “actors” or “performers” enacting roles for the consumption of an “audience” of witnesses who have been “admitted” to the performance. In role-theoretical terms, managing impressions entails enacting the role that is appropriate for a particular audience, while avoiding enacting a role that conflicts with one’s goals with a particular audience. People’s public performances are often specific to both the roles they want to claim for themselves and to the particular audience for whom they claim the role. That is, individuals attempt to control the self-images they present to others to make themselves appear as though they possess specific personalities, attitudes, or attributes, and as though they have legitimate claims to the roles with which they wish to be identified. 1. Multiple Roles and Role Conflict

The task of managing the public performances that people present to others is usually relatively simple. Much of the time individuals can manage the numerous roles they enact by taking advantage of situated activity systems (Goffman, 1961)-behaving “in role” only when one is in the presence of the audience for which that role is appropriate and by avoiding situations in which one’s roles may be contradictory or may conflict. There are times, however, when people’s performances (i.e., the roles they enact or the impressions they wish to convey) do conflict with the audiences who are present to witness the performance. In

RELATIONAL REGULATION

229

other words, there are times when a performance is witnessed not only by those who have been admitted but also by those who have not. These kinds of conflicts sometimes ensue because of the variety of settings and the complex situational pressures that lead individuals to adopt different roles. Accordingly, the roles that people adopt in the presence of the various others with whom they interact may overlap completely, partially, or not at all. Some individuals adopt distinct and separate roles in different situations; others adopt a relatively constant role (or set of roles) in almost all situations (see Markus, 1977; M. Snyder, 1987). Goffman (1959, 1961) refers to situations in which a given role would be inappropriate for one of the audiences who could witness it as role conjlicts or role dilemmas. In these cases managing one’s impressions becomes extremely complicated. The more individuals enact different performances in different situations for different audiences, the greater will be the possibility for role conflict (Burchard, 1954; Marks, 1977). 2 . Normative Role-Based Behavior and the Nai‘ve Understanding of Constraint Rightly or wrongly, people are often assumed to have a fairly well developed understanding of the nature of role-based behaviors. A defense attorney for an accused pornographer is constrained by her profession to play the role of advocate for her client independent of either her personal feelings toward pornography or of the audiences who witness her present her defense. Similarly, international news correspondents, whose reports are frequently censored by local authorities, may be constrained to report a version of a story that lacks important details or that is rife with propaganda, independent of either their personal feelings about comprehensive coverage or of the audiences who will receive their report. Conventional wisdom suggests that virtually everyone understands the role demands of certain occupations and activities (e.g., a defense lawyer, a member of a debate team, or an international correspondent whose reports are censored). Thus, individuals occupying these roles presumably need not fear, for example, that their role compliance will be seen as diagnostic of some dispositional approval of pornography or tacit support of the government’s positions.

B. CORRESPONDENCE BIAS It follows that observers of such constrained behavior will routinely take the constraining influences into account and avoid using constrained behavior as a basis for making dispositional inferences. How realistic is this assessment and exactly how prevalent is an accurate understanding of role-based constraint? A large body of research suggests that people’s naive understanding of constraint is

230

JOHN H. FLEMING

less than complete and that even highly constrained behavior often serves as the basis for strong dispositional inferences (Fleming & Scott, 1991; Jones, 1990). Thus, even in situations customarily regarded as well understood from a role perspective (e.g., defense attorneys, debaters, correspondents, and POWs), the need frequently arises for individuals to signal that their behavior is constrained. A substantial body of research and theory underscores this need for communicators to signal that their behavior was enacted under constraint rather than to simply assume that observers will acknowledge the constraint and adjust their inferences accordingly. This literature focuses on biased causal attributions made when situational constraint is high and has been variously labeled thefundamental atfribution error (L. Ross, 1977), observer bias (Jones & Nisbett, 1972), overattribution (Quattrone, 1982), or correspondence bias (Gilbert & Jones, 1986). This research shows repeatedly that observers who watch communicators deliver messages will attribute those messages to some dispositional characteristic of the communicator, such as an attitude or personality trait, even when they are vividly aware that the communicators were constrained in some way to deliver the message (for reviews see Fiske & Taylor, 1991; Jones, 1979, 1986, 1990; Nisbett & L. Ross, 1980; L. Ross, 1977). Observers who read pro- and anti-Castro essays, for example, assume that the author endorses the position advocated in the essay even when they know that the direction of the essay was assigned by the experimenter (Jones & Harris, 1967). Even subjects whose constrained self-referent essays (describing themselves as extreme introverts or extroverts) were shuffled and redistributed to others who had just completed the very same task inferred that the essays were diagnostic of the author’s underlying personality (A. G. Miller, Jones, & Hinkle, 1981). Although originally unexpected, this finding has proven both robust (e.g., Jones, Worchel, Goethals, & Grumet, 1971) and generalizable across variations (Jones & Harris, 1967; A. G. Miller et al., 1981; Schneider & R. S. Miller, 1975). It has also survived suggestions that it was produced by experimental artifact (Kelley, 1972; M. L. Snyder & Jones, 1974). What can account for observers’ apparent inability to adjust adequately for the kinds of constraints present in the typical assigned-essay paradigm? One possible explanation is that observers expect constrained behavior to differ in several important ways from freely chosen behavior. For example, A. G. Miller and Rorer (1982) found that observers expect qualitative differences, such as differences in credibility, between essays that correspond to an author’s true position on an issue and essays that do not. Research by A. G. Miller, Baer, and Schonberg (1979) and Jones et al. (1971) has provided evidence that authors given no choice over the position they were to support nonetheless generated essays that were moderately persuasive and virtually indistinguishable from those generated by subjects who freely chose the position they would support. Insofar as constrained behavior does not differ from freely chosen behavior, observers have no reason to doubt the association between the actor and the act, and will use the

RELATIONAL REGULATION

23 1

constrained behavior to infer correspondent dispositions (J. T. Johnson, Jemmott, & Pettigrew, 1984). These results suggest that if observers are to take situational constraints into account, communicators must somehow signal the fact that situational constraint is operating and that they do not wish to be seen in close relation to their behavior. Unless communicators explicitly acknowledge the presence of situational constraints, observers may take their constrained behavior at face value and draw correspondent dispositional inferences. A number of recent studies have shown that under certain circumstances a number of relational-regulation cues can influence the extent to which observers will take a target’s behavior at face value (Fein, Hilton, & D. T. Miller, 1990; Fleming, 1993; Fleming & Darley, 1989; Fleming & Stukas, 1991; Ginzel, Jones, & Swann, 1987; Jones et al. 1971; Kraut, 1978; Quattrone, 1982). For example, Jones et al. (1971) found that if an observer saw a target randomly assigned to prepare a given essay, and perceived the essay to be a weak one, the correspondence bias effect was reduced. In relational terms, observers used the strength of authors’ essays as an indicator of authors’ relation to the essays and ultimately, as an index of the strength of their actual support for the issue. Kraut (1978) found that inserting a 7-s pause between an interviewer’s question and a target’s self-serving response also attenuated the bias. In that study, an unusually long pause presumably led observers to infer that the target had ulterior motives, and thus to discount the target’s response.

C. THE ALLO- AND AUTO-COMMUNICATIVE PROPERTIES OF DISTANCING BEHAVIORS

Distancing cues such as the kinds just described may be defined as signals that are given off either intentionally or unintentionally . . . that signal that [a target’s] observed behavior was not freely chosen, intended, or foreseen . . . and could alert the perceiver that the target’s behavior is constrained . . . and should be discounted. What these cues are depends on the context, and they can be subtle. [They] point to constraints in the situation that may not be recognized or given sufficient weight by the perceiver. (Fleming & Darley, 1989, p. 32)

Distancing cues are so named because they imply that targets wish to place (or wish observers to place) psychological distance between themselves and their behavior. In Jones’s apt attributional terminology, distancing cues are a type of “plea for noncorrespondence”; “distancers” do not want to be dispositionally associated with their constrained behavior.5 What situations precipitate these 5Similarly, embracing cues are “pleas for correspondence”; the embracer wants to be dispositionally associated with his or her behavior.

232

JOHN H. FLEMING

behaviors and exactly what are these cues? And what are their effects on how others perceive us and on how we see ourselves? I would like to suggest that relational-regulation behaviors (specifically, distancing behaviors) serve two distinct but interlocking functions (Fleming & Rudman, 1993). An example based on Goffman’s (1961) work illustrates the first function. A child requests a merry-go-round ride from his parents. In the midst of the ride, he sees a group of his peers who would undoubtedly consider this activity rather puerile. To resolve the dilemma posed by the need to present to both audiences, the child alternately smiles at his parents to show his gratitude as he rides by them, and grimaces, rolling his eyes upward at his friends as he passes by them. Thus, one function served by distancing behaviors is allocommunicative (or self-presentational)-to signal to others who may witness a performance that the person’s behavior was constrained and should be discounted. In the merry-go-round example, the signaling was intentional. Under other circumstances, however, the allo-communicative aspect of distancing behaviors may be inadvertent or accidental. For example, the person whose out-oftown friend makes a particularly off-color remark in front of her colleagues may enact a sequence of behaviors that “disassociate” her from her friend’s remark: looking or turning away, looking up or down, shuffling her feet, and hiding her face, among others. These behaviors may also inadvertently signal to others that she does not wish to be associated with her friend’s behavior. Distancing behaviors function at one level by signaling the extent to which an individual wishes to be associated with his or her own behavior, the behavior of another person, or a specific situation. As I shall describe in detail in subsequent sections, this allocommunicative function can also serve as a powerful contextual discounting prime for observers, the effects of which can carry over into other, unrelated judgment situations. A second and equally important function served by distancing behaviors is a two-fold auto-communicative and self-regulating function-to signal to oneself that one’s own behavior was situationally demanded. For instance, consider what many children have long known: A lie isn’t really a lie if you tell it with your fingers crossed (Desmond, Collet, Marsh, & O’Shaughnessy, 1979). Usually hidden from the view of others, this schoolyard ritual allows children symbolically to cancel the effects of telling a lie. I would like to suggest that distancing behaviors, many of the essential features of which this simple act captures, also fulfill an important self-concept-regulating function. They allow people to symbolically take back, “undo,” or distance themselves from potentially selfconcept-threatening behavior by disassociating themselves from that behavior. Paradoxically, they also permit people to engage in egregious self-discrepant behavior more extensively or for longer periods (when the situation demands it; e.g., McGraw, 1987; Milgram, 1963, 1974). Numerous researchers have shown that engaging in potentially negative, self-concept-discrepant or otherwise con-

RELATIONAL REGULATION

233

flictual behavior increases a person’s level of arousal (Festinger, 1957; Milgram, 1963; Zanna & Cooper, 1974), which can lead, in turn, to changes in various aspects of the self. Like the “regulator valve” on a pressure cooker, distancing behaviors may help to restore and maintain an individual’s internal equilibrium by safely diffusing this arousal and the accompanying self-altering effects of negative self-discrepant behavior.

D. RELATIONAL CUES AND ATTITUDE ATTRIBUTION I have suggested that relational-regulation behaviors serve both allo- and autocommunicative functions and that both functions signal the implicit relation between actors and their behavior, others’ behavior, or the situation. The allocommunicative function is directed at the external audience of witnesses, whereas the auto-communicative function is reserved for consumption by the internal audience of the self. Distancing behaviors should thus signal to witnesses that the person’s behavior was constrained and should be discounted. In contrast, embracing behaviors signal that actors and their acts should be similarly described (Jones & Davis, 1965, p. 223) and indicate that actors wish to place no distance between themselves and their behavior. Both of these functions, however, are context dependent; these cues should affect observers’ judgments only when there is some reason to doubt whether the target freely chose the behavior. They should exert no systematic effects on inference when the target emitted the observed behavior freely and without apparent constraints (Fleming & Darley, 1991). Before embarking on a detailed exploration of the auto-communicative properties of relational-regulation behaviors, I would like to begin by briefly examining how distancing behaviors affect attitude attribution. To examine the relationship between contextual and tactical-relational cues and correspondent inferences, John Darley and I had subjects watch and rate a videotape of a male target preparing and reading an essay in support of personal choice in abortion (Fleming & Darley, 1989, Experiment 2). In the conditions most relevant to this discussion, some subjects saw an additional segment in which the target freely chose the pro-choice position. Other subjects saw a videotaped segment in which a randomly assigned target prepared the pro-choice essay under typical assigned essay instructions. Cross-cutting this manipulation, a third of the subjects saw a version of the tape that included a segment showing the target smile with delight (an embracing cue) when the direction of his essay was determined. An additional third of the subjects saw a version of the tape that included a segment that showed the target frown with disappointment (a distancing cue) as the essay was determined. The videotape watched by the remaining subjects contained a neutral reaction to the essay assignment.

234

JOHN H. FLEMING

12

-

10

-

16 14

8-

6-

4 .

-

HIGHCHOICE

I LOWCHOICE

I

I

EMBRACING

NEUTRAL

1

1

DISTANCING

TARGET'S FACIAL EXPRESSION Fig. 2. Effects of distancing and embracing cues on judgments of a target's attitudes under conditions of high and low choice. Cues affected attributions when behavior was constrained, but not when it was freely chosen. Adapted from Fleming & Darley, 1989; Experiment 2. Copyright 1989 by the American Psychological Association.

The relevant results are presented graphically in Fig. 2 and indicated that when these nonverbal distancing and embracing cues were embedded in the target's videotape, observers used them to determine the extent to which the target actually agreed with the essay. Observers who saw the target frown in response to the assigned essay rated him as significantly less pro-choice than did observers who saw a neutral-expression target. In contrast, observers who saw the target smile in response to the assigned essay rated him as significantly more pro-choice than did observers who saw the neutral-expression target. Moreover, the distancing and embracing facial expressions were emergent-they took on a significance only when the context suggested that the target might not have had any choice in determining which essay he prepared and read (also see Trope, 1986). When the context suggested the target had choice, the facial expressions either went unnoticed by the observers or were considered accidental and uninformative. These findings suggest that observers will use relational cues embedded in a target's constrained behavior and will abandon their tendency to make dispositional attributions based on that constrained behavior. Consequently, they suggest that by using relational-regulation behaviors, constrained communicators may sometimes succeed in conveying a message that is different from the apparent one being sent.

RELATIONAL REGULATION

235

E. RESOLVING ROLE CONFLICT: ROLE DISTANCING AND MIXED MESSAGES How can social actors resolve or neutralize role conflicts such as the multipleaudience problems just described? Findings from a number of relatively disparate lines of research suggest that people can often exhibit great skill in resolving complex social communication dilemmas such as those imposed by the role conflict inherent in multiple-audience problems. The tactical task confronting those facing role conflict is somehow to prevent their respective role enactments from being witnessed by the audience or audiences for which they are not intended. Under slightly different circumstances, however, the task is to signal to one or more of the witnessing audiences that a particular performance was either constrained or was otherwise situationally induced.

1 . Audience Segregation As a first line of defense, performers can strive to separate the various audiences who witness their performances. Some individuals actively seek to segregate the different groups with whom they interact (M. Snyder, 1987). However, this is a more appealing tactic to some individuals than to others. M. Snyder and his colleagues, for example, have demonstrated that compared to low selfmonitors (SMs), high SMs tend to choose clearly defined and clearly roleappropriate social situations (M. Snyder & Gangestad, 1982; M. Snyder & Kendzierski, 1982) and partners who are specific to the activity in which they engage (M. Snyder, Gangestad, & Simpson, 1983). Variations on this general theme are also possible. For example, performers can wait until one of the audiences is no longer present, or they can find protected or free spaces to which the undesired audience does not have access (Goffman, 1974). Alternately, they can take time-outs,during which they are along with one of the audiences, and then enact their performances (Darley, 1984). Because these options are not always available, performers may also try to restrict the physical channels they use when communicating in the presence of multiple audiences (i.e., by whispering). All of these tactics are examples of audience segregation (Goffman, 1959, pp. 137-140). 2 . Role Distancing Although audience segregation can be an effective technique for neutralizing role conflict in some situations, frequently multiple audiences are all present at the same time. Thus, regardless of how actively they attempt to keep their audiences separate, it is not always possible for performers to segregate those for

236

JOHN H. FLEMING

whom a performance is intended from those for whom it is not. In fact, attempts to separate the audiences physically can frequently arouse the very suspicions and conflicts that the tactic was initially intended to prevent. It may also be physically impossible to separate audiences. As a second line of defense, and depending on the circumstances, individuals are likely to engage in one of two possible role distancing (Archibald & Cohen, 1971; Goffman, 1959, 1961; Stebbins, 1969) strategies to resolve the dilemma imposed by these multiple-audience problems. When few details are known about the constitution of the various witnessing audiences, when the communication medium does not lend itself to hidden messaging, when the audiences are hidden, or when the constraining influences are unusually coercive, communicators may be partially able to resolve multipleaudience problems by engaging in verbal and nonverbal distancing behaviors. Although these behaviors fall short of communicating a distinct and identifiable message to the intended audience, they may alert the intended audience to the constrained nature of the performer’s behavior and that they should therefore discount, disregard, or interpret it cautiously. Clearly, though, a variety of situations lend themselves to role-conflict resolutions that are even more elaborate than the usual distancing tactics. As the Fleming et al. (1990) results suggest, in these situations, communicators may try to resolve multiple-audience problems by communicating complex hidden or mixed messages. In subsequent sections I describe the theoretical, anecdotal, and empirical evidence relevant to this mixed or hidden message approach. The enactment of specific relational-regulation behaviors depends on a complex set of situational variables, including the availability of resources in the situation, the nature of the situational constraint, the relative costs associated with being “unmasked,” and the nature of the performer’s relationships to the witnessing audiences, to name a few. For example, sometimes the constraint flowing from the performer’s relationship with one of the present audiences is apparent to all, such as when a hostage or POW is forced to sign a confession or make a propaganda videotape (cf. Fleming & Scott, 1991, 1993). At other times the performer must alert one audience to the existence of the constraint before his or her complex signals will be recognized and understood. Sometimes the specific behaviors will constitute “acts of commission”active attempts to signal directly to one audience. A facial expression of disgust or disappointment, for example, is an exemplar of this category of hypergenerative relational-regulationbehaviors. Hypergenerative behaviors will predominate when the costs of being detected or “unmasked” are low and the need to communicate effectively (by discriminating between audiences) is high. At other times the behaviors will instead resemble “acts of omission.” This class of behaviors includes more passive alterations such as speaking in a muted or monotone voice, inhibiting physical movement and dampening facial activity, or feigning

RELATIONAL REGULATION

237

disinterest (Fleming & Scott, I99 1). These hypogenerative relational-regulation behaviors will predominate when the costs of being “unmasked” are high (such as when a hostage risks physical injury or death if detected) and the need for the communication to discriminate between audiences is relatively low.

IV. Distancing and Embracing: Properties and Consequences of Relational-Regulation Behaviors A. WHAT ARE RELATIONAL-REGULATION BEHAVIORS? In the following sections I would like to describe several experiments that explore the dual nature of relational-regulation behaviors-their allocommunicative role in interpersonal perception and interaction, and their autocommunicative role in self-concept maintenance and change. My hope is that this research will allow us to better understand exactly how people distance themselves from their constrained or role-based behavior, and will extend our understanding of the implications of distancing for self- and other-perception. Many of the ideas regarding the relational concepts of distancing and embracing raised here are extensions of themes integral to Goffman’s (1959, 1961, 1969) and Mehrabian’s (1972, pp. 31-53; Wiener & Mehrabian, 1968) formulations. Mehrabian’s scheme, for example, has served as a useful starting point for developing both a typology of verbal relational transformations and a taxonomy of nonverbal distancing and embracing behaviors. Mehrabian (1972) and Wiener and Mehrabian (1 968) have described a number of behaviors that a communicator can use to distance himself or herself when ”[he] has negative feelings toward [his] addressee, toward the contents of [his] communication, or toward the act of communicating those contents” (p. 31). In spite of its original promise as a broad-brush approach to social interaction, however, the bulk of research on the immediacy construct has focused primarily on the inadvertent behaviors that signal one’s implicit orientation toward an interaction partner. Very little of that research has examined either communicator’s orientations to objects other than another person, or the ways in which interactants can use nonimmediacy intentionally for tactical purposes. The relational-regulation perspective, in contrast, suggests that distancing and other relational-regulation behaviors comprise a broader class of behaviors (including non-immediacy) that regulate one’s relation to one’s own behavior, the behavior of others, or the situation in relation to the self. Relational-regulation behaviors can occur spontaneously and inadvertently, but social actors can also use them tactically to shape the

238

JOHN H. FLEMING

impressions that others form. Thus, the relational-regulation analysis is an attempt to extend many of these ideas to a broader spectrum of social behavior. 1. Nonverbal Behaviors Exactly what behaviors constitute distancing and embracing behaviors? As they have been described, distancing behaviors are attempts to undo symbolically some offending behavior and to signal that the behavior does not accurately reflect some underlying personal disposition. Thus, distancing can be conceptualized as a response to an acute avoidance conflict (Dollard & Miller, 1950; Lewin, 1935, 1938). Embracing behaviors, in contrast, communicate the absence of any kind of avoidance conflict. In some cases, they go beyond communicating the absence of conflict; sometimes affinity is implied. For example, gays and lesbians who have not openly admitted their homosexuality may use passing strategies-embracing techniques designed to allow them to move and interact undetected in the larger community of “straights” (Pearlman, 1989; Ponse, 1978). Consequently, the search for distancing and embracing behaviors has focused on behaviors that seem to communicate this kind of avoidance conflict or its absence. A number of specific nonverbal distancing behaviors have previously been suggested and described, including grimacing or making other kinds of negative facial expressions; using defiant or scatological gestures and emblems (such as the middle-finger gesture); rolling one’s eyes to convey sarcasm; sticking out one’s tongue to communicate disgust; deliberately looking down or away; hiding or shielding one’s face; and behaving in a less-than-spontaneous,stilted, or overly controlled fashion (Fleming & Darley, 1989; Fleming & Rudman, 1993; Fleming & Scott, 1991). Among the behaviors considered likely candidates for embracing behaviors are smiling and other expressions of delight or pleasure, head nodding, spontaneity and an absence of overt behavioral control, and other immediacy cues. Interestingly, situational cues to purposeful action and neutral behavior (the absence of direct distancing) may also serve as “sufficient” or default embracing cues (Fleming & Darley, 1989, 1991; Johnson et al., 1984; Jones et al., 1971; A. G. Miller et al., 1979). In contrast to these arguably more deliberate distancing behaviors, the kind of avoidance conflict generated by multiple audience problems can also spawn any of a number of inadvertent, protective avoidance behaviors that could also serve to insulate subjects from the effects of their own negative actions. Although numerous behaviors could be so classified, among the inadvertent behaviors that are likely candidates for distancing behaviors are those that have previously been related to the expression or interpersonal communication of avoidance, hesitation, defensiveness, shame, or embarrassment, such as head nodding, head shaking, smiling, frowning, gesticulating, fidgeting, “hiding,” and looking down or away (Buss, 1980; Edelmann, 1987; Edelmann & Hampson, 1979, 1981;

RELATlONAL REGULATION

239

Ekman & Friesen, 1969, 1972; Exline, Gray, & Schuette, 1965); social distance, dislike, and non-immediacy, such as backward body lean, body orientation, and the use of verbal reinforcers (Lott, 1987, 1989; Mehrabian, 1968a, 1968b, 1969, 1972; Mehrabian & Friar, 1969; Wiener & Mehrabian, 1968; Word et al., 1974); cognitive effort and arousal, such as silent speech, speech “dysfluencies,” lip and mouth movements, and other facial actions (e. g., Cacioppo & Petty, 1981; Kasl & Mahl, 1965; Mahl, 1959; Rosenfeld, 1966a, 1966b; Zajonc & Markus, 1983); stress and threat (Archibald & Cohen, 1971; Riskind & Gotay, 1982) and persuasiveness, such as an “open,” “closed,” or slumped body posture (H. McGinley, LeFevre, & P. McGinley, 1975; Mehrabian & Williams, 1969), including a relatively neglected class of behaviors called adapters (Ekman & Friesen, 1969, 1972). According to Ekman and Friesen (1969), adapters are vestigial fragments of previously instrumental behaviors that “were first learned as part of adaptive efforts to satisfy self or bodily needs, or to perform bodily actions, or to manage emotions, or to develop or maintain prototypic interpersonal contacts, or to learn instrumental activities” (p. 84). That is, people emit adapters with little conscious mediation or awareness. They are not intended to communicate an emotional state to others, although they are informative about the state and may sometimes communicate it inadvertently. Actors (usually) emit these behaviors under stress or threat; the behaviors have presumably degenerated from their original instrumental instantiations into symbolic fragments. For example, a person may respond to a physical threat by throwing up his or her arms to “ward off” the threatening agent, or to cower. When faced with a nonphysical but psychological threat, a person may enact the corresponding adapter of (symbolic) selfembracing-wrapping the arms around the upper body in a closed or selfprotective way (see McGinley et al., 1975). These adapters (as well as the other inadvertent distancing behaviors described above) may be nonconscious analogues of consciously “crossing one’s fingers” to undo or ward off the ill effects of a lie. That is, they may allow a person to “disown” (both for themselves and for others) behaviors that could otherwise lead to unwarranted, unfortunate, or negative dispositional attributions. 2 . Verbal Relational Transformations Implicit in any discussion of verbal distancing and embracing is the assumption that the easiest way to imply a direct connection between an agent, an action, and an object is to use direct reference (Searle, 1975). Communicators can manipulate the psychological or symbolic distance (or lack thereof) between an agent, an action, or an object, however, through the use of a number of linguistic transformations that run the gamut from very simple to extremely complex. These transformations either obfuscate or accentuate the agent-action-

240

JOHN H. FLEMING

object relationship. I shall briefly describe a few of these, although a complete treatment of this issue is well beyond the scope of this article. For example, Wiener and Mehrabian (1968) identify six varieties of verbal distancing or nonimmediacy transformations, including spatio-temporal indicators (e.g ., using “that” instead of “this” to indicate distance), denotative specificity (e.g., using more ambiguous or inclusive descriptors, or passive constructions such as “One might like to fly’’ instead of “I like to fly”), selective emphasis (e.g., using sequencing or emphasizing a specific attribute over others, such as “Let’s go visit Pam and Mike” instead of “Let’s go visit Mike and Pam” or “Let’s go visit the Booths”), agent-action-object relationships (e.g ., emphasizing different elements in a statement, such as “I went to the mall with my sister” instead of “My sister took me to the mall”), modifiers (including qualification and objectification-e.g., using hypotheticals, such as “It could be the case that . . .” and statements indicating broad consensus, such as “It should be obvious that . . .”), and automatic phrasing (using phrases like “you know” and “right?” in speech; see Wiener & Mehrabian, 1968, pp. 34-47 for a detailed discussion of these transformations). Other possible distancing transformations include the use of indirect or overly polite speech (Brown & Levinson, 1987; Holtgraves, 1986; Searle, 1975), the use of slang and sarcasm (Goffman, 1959; Schein, 1956; Schein, Schneier, & Barker, 1961), dissenting or engaging in subterfuge (Milgram, 1974), code-switching into different languages (Bond & Lai, 1986); violating conversational maxims (Grice, 1975, 1978), and (more directly) verbally disclaiming or denying that one had any choice about engaging in the behavior (Hewitt & Stokes, 1975).

B. THE AUTO-COMMUNICATIVE PROPERTIES OF DISTANCING: DYNAMICS OF SELF-CONCEPT REGULATION The results of the Fleming and Darley (1989) research suggest that distancing behaviors play an allo-communicative role by altering observers’ inferences about constrained behavior. The relational-regulation perspective, however, suggests that distancing behaviors may also play a self-concept-regulating role. How can people protect themselves from the potential ill effects of negative selfdiscrepant behavior? We have all witnessed situations in which people have tried to pass themselves off in ways that did not accurately reflect who they really were or what they truly believed. How do these self-presenters insulate themselves from the knowledge that they are really emperors with no clothes? As I noted in the overview, the twin relational notions of “distancing” and “embracing” can serve as useful heuristics for understanding the ways in which people implicitly regulate their association with some aspect of their own behavior, the behavior of

RELATIONAL REGULATION

24 1

another person, or a particular situation. These relational adjustments can also act as a buffer between one’s outward behavior and its effects on one’s internal self-appraisals.

I . Parameters Afecting Internalization of Self-Discrepant Behavior Understanding the nature of the “self” presents something of a paradox. The considerable research evidence supporting the view of the self as malleable and dependent on external actions (Baumeister, 1982; Baumeister & Tice, 1986b; Bem, 1965, 1967, 1972; Gergen, 1965, 1968, 1981, 1982; Goffman, 1959; Jones & Pittman, 1982; Rhodewalt, 1986; Wicklund & Brehm, 1976), for example, stands in rather stark contrast to two sets of findings. The first set documents situational variables that make counter-attitudinal or self-discrepant performances less likely to be internalized, such as a lack of behavioral freedom, the absence of assumed responsibility, and the revocability of undesired consequences, among others (Cooper & Fazio, 1984; Davis &Jones, 1960). The other suggests that people also possess a rich repertoire of behaviors that can buffer them from the potential ill effects of their own self-discrepant behavior (Darley, 1984; Davis &Jones, 1960; Fleming & Darley, 1989, 1991; Fleming et al., 1990; Goffman, 1959, 1961; Schein, 1956; Schein et al., 1961). What exactly are these mechanisms that prevent one’s outward behaviors from affecting one’s internal self-appraisals? Simply put: How do people insulate themselves from the consequences of their own self-discrepantbehavior and how do they communicate that a particular behavior is discrepant? To begin to answer this question and to examine the intrapersonal consequences of relational regulation in multiple-audience dilemmas, Laurie Rudman and I recruited 60 subjects known to hold moderately extreme to extremely proaffirmative-action attitudes (as measured by a pretest) to write counter-attitudinal essays in favor of a change in university policy toward eliminating affirmative action for women and minorities (Fleming & Rudman, 1993). Half of these subjects were forced to write the essay (the “low choice” conditions; LC).The remaining subjects were induced to comply with the request to write the essay as a favor to the experimenter (the “high choice” conditions; HC). We dekribed the essays as having been requested by a university dean to help with discussions concerning possible changes in university affirmative-action policies. Consequently, they could conceivably have a foreseeable negative impact, conditions necessary for dissonance arousal (Cooper & Fazio, 1984; Cooper & Scher, in press; Scher & Cooper, 1989). Cross-cutting this manipulation, one-third of the subjects tape-recorded their arguments in the presence of a black female confederate ostensibly interested in watching an experiment in progress to help her decide whether to volunteer as a

242

JOHN H. FLEMING

research assistant (the “black-female-confederate”condition; BFC). An additional one-third of the subjects taped their arguments in the presence of a white male confederate (the “white-male-confederate” condition; WMC). The remaining subjects recorded their statements on audiotape without the confederate (the “confederate-absent”conditions; CA). All subjects gave their permission to have the confederate observe. The confederates’ presence served a twofold purpose: first, it created a multiple-audienceproblem for subjects (one audience comprising the dean and the experimenter; the other the confederate), and second, it emphasized the counter-attitudinal nature of their essays. As in previous dissonance research, we assessed subjects’ attitudes toward affirmative-actionpolicies prior to and immediately following the experiment; changes in attitudes constituted the primary dependent measure. Finally, we unobtrusively videotaped subjects’ behaviors throughout the three segments of the experimental session: the pre-essay segment, the period immediately preceding recording the essay; the essay segment, the period during which the subject recorded the essay; and the post-essay segment, the period immediately following recording the essay. Consistent with the multiple audience/communication analysis, we predicted that subjects who recorded their essays in the presence of someone who might be adversely affected as a result of their seemingly freely chosen behavior (i.e., the black woman) and who might believe that the essays accurately reflected their true opinions, would find themselves in the clutches of an acute impressionmanagement dilemma and might be motivated to try to communicate the constrained nature of the speech, and possibly their true attitudes, to the confederate. That is, they would attempt to distance themselves from their self-discrepant behavior and would not show the typical dissonance-induced pattern of attitude change as a result. In contrast, subjects who had no choice in preparing their essays and those who read them either alone or in the presence of the white male confederate would produce a pattern of attitude change that mirrored that found in typical induced-compliance research. Moreover, we expected that subjects’ attitude change would be inversely related to the extent to which they tried to distance themselves from their counter-attitudinal behavior. In other words, the more distancing behaviors subjects emitted, the less dissonance-induced attitude change would result. The results are presented graphically in Fig. 3, and as the figure illustrates, they confirmed the predictions. The HC-BFC subjects did not shift their attitudes in the anti-affirmative-action direction, whereas all other subjects changed their attitudes in ways predicted by dissonance theory. Moreover, the absence of dissonance-related attitude change among HC-BFC subjects was significantly and directly related to the extent to which they engaged in a collection of verbal and nonverbal distancing behaviors: self-embracing adapters (i. e., wrapping the arms around the body in a defensive, “closed,” or self-protective way; see Fleming & Rudman, 1993, p. 51); frowns, grimaces and other negative facial

243

-

RELATIONAL REGULATION

12

10

-

8-

Y

NO CONFEDERATE BLACK FEMALE CONFEDERATE WHITE MALE I CONFEDERATE

6W

Q

42-

0I

1

LOW

HIGH

1

CHOICE CONDITION 0.40

1

tsn

0.30

HIGH CHOICE- BLACK FEMALE CONFEDERATE SUBJECTS ALL OTHER SUBJECTS

I

. I

-

z 3 0.20 -

zg y t: 0.10 2 9 0.00

-0.10

-

!

c I

I

I

PRE-ESSAY

ESSAY

POST-ESSAY

4

PROCEDURAL SEGMENT Fig. 3. Effects of distancing behaviors on attitude change in counter-attitudinal essay-dissonance paradigm. (Top) Attitude change was attenuated when counter-attitudinal essays were read in the presence of a black female confederate. (Bottom) Distancing behaviors occurred more frequently in the black female confederate cell than in other cells, particularly prior to reading the essay. Adapted from Fleming & Rudman, 1993. Copyright 1993 by the American Psychological Association.

expressions; verbal reinforcers; lip and mouth movements (i.e., facial movements involving the lips, mouth, or tongue); verbal disclaimers and excuses; and downward eye gaze and head orientation. This was true, however, only when the distancing was preemptive-the attempts occurred during the pre-essay segment prior to recording the essays. After-the-fact or remedial distancing did not atten-

244

JOHN H. FLEMING

uate attitude change. We have speculated that distancing serves a self-conceptregulating function in a manner similar to other self-regulatory, feedback-loop models (i.e., Carver & Scheier, 1981, 1985; Milgram, 1974).

2 . Perceptions of Internal States Although our data did not speak specifically to the process link between distancing behaviors and attitude change, several lines of research suggest that feeling states and evaluative judgments are influenced, at least in part, by kinesic or proprioceptive cues received from remote locations of the body, primarily the face (Baumeister & Tice, 1986a; Cupchick & Leventhal, 1974; Duncan & Laird, 1977; Izard, 1972, 1977, 1990; Laird, 1974, 1984; Lanzetta, Cartwright-Smith, & Kleck, 1976; Rhodewalt & Comer, 1979; Zajonc & Markus, 1983). In support of the notion of a facial-feedback system in emotional experience, Strack, Martin, and Stepper (1988) have found that subjects rate cartoons as funnier when they have a pen clenched in their teeth (inadvertently simulating a smile) and less funny when they have a pen pursed in their lips (inadvertently inhibiting a smile). Other research suggests that it is not just facial feedback that communicates information about feeling states (Buck, 1980, 1984; Delgado, 1969; Hohmann, 1966); gaze and posture have also been implicated in mood (Kellerman, Lewis, & Laird, 1989; Martin, Harlow, & Strack, 1992; Riskind & Gotay, 1982; Stepper & Strack, 1993), and activity in other nonverbal channels has been shown to influence memory and lexical retrieval as well (Krauss, Morrel-Samuels, & Colasante, 1991; Morrel-Samuels & Krauss, 1992; Zajonc, Adelmann, Murphy, & Niedenthal, 1987; Zajonc, Pietromonaco, & Bargh, 1982). For example, Krauss and his colleagues (Krauss et al., 1991; Morrel-Samuels & Krauss, 1992) have found that illustrators and other gesticulations facilitate lexical searches in conversation, in addition to their role as gestures of emphasis. Moreover, research on interoceptive perception suggests that people have a reasonably accurate sense of what is happening with and to their bodies (Buck, 1984; Markov, 1950, cited in Buck, 1984). Whatever the precise nature of the relationship between affect and cognition, the “hard interface”-motor, kinesic, and proprioceptive cues to bodily states-undoubtedly has an impact on cognition and mental representation (Zajonc & Markus, 1983). 3 . A Proprioceptive-Feedback Model Although it is largely speculative, I would like to sketch out briefly one of several possible processes through which distancing behaviors may produce their self-concept-regulating effects. Determining the exact nature of the process awaits further research, but I would like to suggest that distancing behaviors fill their self-concept regulating functions in much the same way as do the types of kinesic proprioceptive, or interoceptive cue-feedback systems described above (also see Carver & Scheier, 1981, 1985; Scheier & Carver, 1988; Zajonc &

RELATIONAL REGULATION

245

Markus, 1983). That is, they have two related but distinct effects. First, in the terms of the example I offered earlier, distancing behaviors may function much like a pressure cooker’s “regulator valve.” By helping to dissipate a person’s conflict-generatedarousal, relational-regulation behaviors can diffuse the accompanying self-altering effects of arousal precipitated by unfortunate self-discrepant behavior (Notarius & Levenson, 1979). Second, they may provide hard, physical representations of “undoing” (or other relational-regulation) in their own right, which may, in turn, channel or focus subsequent cognitive processing, such as self-perception, labeling, or evaluation (Edelmann, 1985; Zajonc & Markus, 1983). In the Fleming and Rudman (1993) research, for example, the presence of a salient audience may have increased subjects’ self-awareness and focused their attention on their public and unavoidably discrepant behavior (Carver & Scheier, 1978). For the HC-BFC subjects, the fact that their self-discrepantbehavior may have adversely affected (or perhaps insulted) the black female confederate undeniably compounded this audience salience. Wicklund (198 1) has argued that increased self-awareness automatically leads to increased motor activity to refocus attention away from the self; this kind of activity has been shown to reduce self-attributions. In the present context, self-attention and the realization that they had violated personal sets of standards in the presence of a person who may have taken offense, in turn created conflict and generated arousal in the subjects. This arousal then triggered relational-regulation (distancing) behaviors (Buss, 1980; Edelmann & Hampson, 1979; 1981). Subjects then “read” these behaviors either directly (for verbal distancing behaviors) or proprioceptively (for nonverbal behaviors) and interpreted them in the context of the situation (Eiser & M. Ross, 1977; Patterson, 1978; Schachter, 1964; Slane & Leak, 1978). The interpreted feedback thus signaled that the behavior had been “undone” or taken back, and dissonance was thereby reduced or eliminated (Davis & Jones, 1960). Alternatively, distancing behaviors could simply have short-circuited the conflict-generated arousal, or could have provided proprioceptive “evidence” of acute anxiety or shame, which may have led subjects to misattribute their arousal to any number of situational variables. Any of these processes could presumably have resulted in the observed attitude change-distancing behavior relationship. Needless to say, all of these possibilities suggest a set of intriguing relationships between discrepant behavior and relational regulation deserving of further attention. Future research should undoubtedly work to disentangle the processes that account for the insulating effects of relational-regulation (i.e., distancing) behaviors.6 61t is worth noting in passing that a collection of nonverbal behaviors similar to those that comprised the Distancing Index (i.e., facial expressions, head aversion, adapters, postural shifts, and pauses) has been used to predict which depressed patients will respond to antidepressant (amitriptyline) therapy on the basis of a single behavioral interview before rhe patients had received any medications (Ranelli & Miller, 1981).

246

JOHN H. FLEMING

C. THE ALLO-COMMUNICATIVEPROPERTIES OF DISTANCING: DYNAMICS OF IMPRESSION REGULATION

I . Distancing Cues: Effects on Observers’ Judgments Subjects’ distancing behaviors not only dissipated the effects of their counterattitudinal behavior but they also effectively communicated constraint to naive observers. In the final phase of the Fleming and Rudman research, observers watched videotapes of the full experimental procedure, and rated subjects’ true attitudes toward affirmative action. On the basis of a median split on their distancing index scores, we classified subjects into high- and low-distancing groups. Not surprisingly, observers’ ratings of low-distancing subjects’ attitudes (irrespective of the treatment condition to which they had originally been assigned) reflected correspondence. Observers used the manifest content of these subjects’ essays to infer their attitudes. Consistent with the tactical relationalregulation approach, however, observers discounted the content of subjects’ essays in their ratings of high-distancing subjects’ attitudes (irrespective of their treatment condition assignment). The magnitude of the discounting effect was directly related to how much distancing subjects showed. The results of the Fleming and Darley (1989) and Fleming and Rudman (1993) research clearly demonstrated that distancing cues (and to a lesser extent, embracing cues) embedded in the behavior of a constrained target caused observers to adjust their impressions of the target’s true attitudes. Moreover, instead of simply attenuating correspondence bias, distancing cues led observers to make contrasted judgments; attitude inferences in the distancing conditions mirrored (equal in magnitude, but opposite in direction) those made when an embracing cue accompanied the essay. Not only did distancing-cue subjects discount the content of the targets’ essays when inferring their attitudes but in the Fleming and Darley (1989) research observers’ ratings suggested that the distancing cue somehow indicated that the target’s manifest behavior reflected the opposite point of view from the target’s true position. However, how long-lasting and global are the effects of these cues? In other words, do relational-regulation cues exert their effects only for the same target, in the same context or situation in which they occur, or do they generalize to other judgment situations and other targets? 2 . Carry-Over Effects The fact that the context in which a behavior occurs often affects subsequent judgments has intrigued cognitive social psychologists for some time (Darley & Gross, 1983; E. T. Higgins & King, 1981; Martin, 1986; Sherman, Ahlm, Berman, & Lynn, 1978; Tourangeau & Rasinski, 1988; Wyer & Srull, 1981).

RELATIONAL REGULATION

247

These effects-traditionally called context, assimilation, or carry-over effectshave emerged in a variety of situations ranging from how items are arranged in survey research instruments and questionnaires (Tourangeau & Rasinski, 1988; Schwartz & Hippler, 1987; Schwartz & Wyer, 1985) to impression-formation research (Anderson & Hubert, 1963; Asch, 1946; Darley & Gross, 1983; E. T. Higgins & Rholes, 1978; E. T. Higgins, Rholes, & Jones, 1977; Martin, 1986; Quattrone, 1982) and from attitude change (Fazio, Sanbonmatsu, Powell, & Kardes, 1986; Herr, Sherman, & Fazio, 1984; Hovland, Harvey, & Sherif, 1957; Sherman et al., 1978) to the effects of mood on judgment (Schwartz & Clore, 1983; Schwartz, Strack, Kommer, & Wagner, in press; Strack, Schwartz, & Gschneidinger, 1985). Carry-over effects result when the judgment context or subtle induction procedures prime a specific category, dimension, or trait. Once primed, the category or dimension exerts influence over subsequent judgments (Bargh & Pietromonaco, 1982; E. T. Higgins & Stangor, 1988; Martin, 1986). 3 . Appraisal Sets

The Fleming and Darley (1989, Experiment 2) results demonstrated that distancing and embracing cues can serve as powerful proximal cues to constraint and behavioral freedom, respectively. Because of their effectiveness in conveying relational information to observers, however, they may also function as effective contextual primes for a generalized “discounting set” (for distancing cues) or a “correspondence set” (for embracing cues) for observers. For example, in the assigned paradigm, distancing behaviors may alert observers to constraints inherent in the situation and may cause them to discount their inferences on the basis of these situational factors (Martin, 1986). As proximal cues, relational behaviors probably function at the level of conscious awareness and trigger suspicion; observers easily identified the cues and actively used them in the inference process (Fleming & Darley, 1989). Once primed, however, these appraisal sets may continue to exert influence over judgments of subsequent behavior in different contexts on a less conscious level by channeling an observer’s attention toward situational constraints that may be operating and by providing a frame of reference for observers to interpret targets’ behaviors (Fleming & Darley, in press). 4 . Distancing and Embracing Cues: Efects on Subsequent Judgments of Diferent Targets

Arthur Stukas and I (Fleming, 1993; Fleming & Stukas, 1991) conducted two studies to examine how distancing and embracing cues (again, using facial expressions of disappointment and delight, respectively) affect observers’ perceptions and judgments of constrained targets. In addition, we examined how

248

JOHN H. FLEMING

judgments of one target’s behavior (“debiased” by attending to distancing cues) affected judgments of subsequent constrained targets. In the first study, observer-subjects watched a short videotape of a male student (ostensibly a second-year law student) preparing and reading a constrained essay either in favor of or opposed to federal funding for controversial art. Orthogonal to this essay-position manipulation, half of the subjects saw the target distance (grimace and frown) when he received his assigned essay-position assignment and half saw the target embrace (smile) instead. They then rated the target’s attitudes toward funding for controversial art. After this initial rating task, subjects watched a staged videotape of two female students playing a quiz game adapted from one used by L. Ross, Amabile, and Steinmetz (1977). The tape showed a randomly assigned “questioner” pose difficult trivia and general knowledge questions to a “contestant.” The contestant answered 2 questions out of 10 correctly. Thus, this situation provided situationally constrained behavior (contestants’ inability to answer questions posed by persons who had time to formulate unanswerable items) that was sufficiently different from the assigned-essay task in which subjects had previously participated. After watching the mock quiz game, subjects judged both the intelligence and general knowledge of the game players. The results of this experiment are illustrated in the left-hand panel of Fig. 4. Consistent with Fleming and Darley (1989), results revealed that subjects used the facial cues provided by the essay reader to adjust (either by augmenting or discounting) their ratings of the essay reader’s attitudes. Subjects who initially saw the essay-reading target smile and express delight with his essay assignment provided polarized and correspondent ratings of the target’s arts-funding attitudes. Moreover, these subjects’ ratings of the game players mirrored those of observers in the original L. Ross et al. (1977) study. They rated the contestant as being close to average in general intelligence (and no differently than did a group of control subjects who watched only the quiz game reenactment and rated the players), but succumbed to correspondence bias by rating the questioner as having significantly more general intelligence than the contestant, regardless of whether the essay reader had presented a pro- or anti-arts funding opinion. Again, ratings of the questioner did not differ reliably from those of control subjects. Among subjects who initially saw the target grimace, however, ratings of the target’s attitudes revealed a strong contrast effect-that is, subjects judged the target to hold attitudes that were the reverse of the position he advocated in his constrained essay. Moreover, as expected, this contrast effect carried over into these subjects’ ratings of the game players. Specifically, because distancing-cue subjects had been exposed to a powerful discounting cue (the facial expression of displeasure), we expected them to be more attentive to the situational advantages conferred upon the questioner and thus, to discount their ratings of her general

RELATIONAL REGULATION

L

+ ANTI-FUNDING ESSAY * PRO-FUNDING ESSAY ....... CONTROL

............................................

50

40

- d

249

f

A -

..........................................................................

. RELATIONAL CUE

RELATIONAL CUE

Fig. 4. Effects of distancing, embracing, and neutral cues on judgments of a constrained target’s attitudes and on subsequent, unrelated judgments. (Top) Observers discounted target’s behavior when it was accompanied by a distancing cue. (Bottom) Effects of cues carried over into judgments of subsequent targets. Data from Fleming & Stukas, 1991 (left); Fleming, 1993 (right).

250

JOHN H. FLEMING

intelligence. In addition to rating the contestant as being close to average in general intelligence (and, again, no differently than the controls), distancing-cue subjects discounted their inferences of the questioner’s intelligence, rating her significantly less intelligent than did either embracing-cue subjects or control subjects . A second study (Fleming, 1993) replicated and extended the Fleming and Stukas (1991) results. In this study, however, we replaced the embracing cue manipulation with a “no-expression” or neutral cue condition to examine whether the discounting effects of the distancing cue, the augmenting effects of the embracing cue, or a combination of the two could account for the carry-over effects we obtained in the first study. In contrast to the Fleming and Darley (1989) findings, the results, illustrated in the right-hand panel of Fig. 4, suggest that the distancing cue produced the anticipated discounting and carry-over effects, but that the embracing cue yielded judgments that were not substantially different from those produced by subjects who received no cue. Thus it may be the case that rather than enhancing correspondence, embracing cues may simply serve as default cues that reinforce correspondent inferences. These results demonstrate the communicative power of distancing behaviors in altering observers’ judgments of a constrained target. They also suggest the intriguing possibility that distancing behaviors may prime suspicion or “get observers ready” to adjust for situational constraint, and that this discounting carries over into subsequent and unrelated judgmental situations.

5 . Suspicion and Discounting Taken together, the results from these two experiments suggest that distancing cues (such as facial expressions of disappointment or displeasure) not only serve as powerful cues to situational constraint but also produce relatively enduring effects on subsequent social inferences. Appraising the results from these studies, however, one immediately wonders about the ways in which distancing cues are first detected and their meaning decoded by their recipients. What is the particular mechanism or technique that alerts an audience to disregard the surface content of a message? Fleming and Darley (1991) have speculated that distancing cues trigger observers’ suspicions about the motivations underlying the target’s overt behavior. Others have suggested characterizing suspicion as a suspension of unchallenged belief, a skepticism about the face validity of a piece of behavior, and a search for an alternate meaning of, or motivation for, that piece of behavior (Fein et al., 1990; Goffman, 1959; Hewes & Graham, 1989; Kraut, 1978; Pyszczynski & Greenberg, 1981). As Kraut (1978) has noted, for exam’Although significant and substantial, this discounting was not absolute. Distancing-cue subjects still rated the questioner as significantly more intelligent than the contestant.

RELATIONAL REGULATION

25 1

ple, suspicion is a shift from the normal conspiracy mode (Goffman, 1959), in which honoring a person’s claims is more important than challenging those claims, to a sparring or agonistic mode, in which challenging a person’s claims becomes paramount. Being suspicious is thus characterized by uncertainty and by a heightened need and search for information to reduce that uncertainty to arrive at a reasonable and plausible explanation for the observed behavior.

6 . Actuators What precipitates the transition from conspiracy to sparring? Although it would seem crucial to any comprehensive understanding of human inference and judgment, the concept of suspicion per se (especially its instigating conditions) seems to have been largely overlooked. Although attention has been paid to a few crucial elements of suspicion such as ulterior motivation (Fein et al., 1990; Frankel & Morris, 1976; Jones, Davis, & Gergen, 1961; Jones & Wortman, 1973; Walster, Aronson, & Abrahams, 1966) and what Kraut (1978) refers to as the controllability rule-that the best behavioral indicators of deception are those that are least controllable-coherent and focused treatments of the antecedents, process, and consequences of suspicion are virtually absent in the socialpsychological literature. As an essential first step, suspicion requires an instigating event or actuator. The actuator can be a contextual cue, such as awareness of the circumstances under which the performer generated the message or specific information about the performer’s possible motives.Alternatively, it can be a behavioral cue on the part of the communicator, such as a distancing cue-a revealing facial expression, a gesture, a turn of phrase-or some other aspect of the behavior that appears unexpected, unnatural, extreme, or untoward (Ekman & Friesen, 1969, 1974; Krauss, 1981; Kraut, 1978, 1982; Pyszczynski & Greenberg, 1981). Under certain circumstances, an actuator may be some complex emergent combination of a particular behavior that takes on meaning only in a specific context (Fleming & Darley, 1989; Trope, 1986). Whatever its origins, however, suspicion leads observers to suspend the normal attribution process and to apply an alternate set of rules to interpret the behavior. In the two studies just described, for example, it appears that observers may have applied some sort of “mirror-image” rule when making their judgments, and did so in a fairly conscious way. Distancing cues did not simply cause subjects to discount the content of the target’s essay when inferring his attitudes, they also appeared to have led observers to infer that he held opposite and equally extreme attitudes. This non-normative mirror image inferencing is intriguing and demonstrates the power of relational cues in affecting social perception. As a result of the effortful inferential adjustments, suspicious observers’ “readiness” to adjust for situational constraint then persists (possibly at a level outside of awareness),

252

JOHN H.FLEMING

manifesting itself in attributions in other judgment situations that are also adjusted for the effects of situational constraint. How enduring these effects are, however, is a contribution left to future research.

7. Peculiar Peccavis: Correspondence Bias and the Persian Gulf War POW Videotapes The power of distancing and embracing cues as relational-regulation devices has been demonstrated in a number of laboratory studies. However, how do these kinds of signals reveal themselves in the real world? Much of the anecdotal evidence for the existence of distancing behaviors comes from a body of research that emerged following the Korean War; this research documented the techniques that POWs used to distance themselves from the coerced confessions and propaganda statements their Chinese Communist captors routinely forced them to make (see Schein, 1956; Schein et al., 1961; Segal, 1954). These techniques and those used by people in similar situations in Vietnam and in Beirut, Lebanon, have been described elsewhere (Fleming & Scott, 1991). Recently, however, we had an opportunity to examine some of these anecdotal observations firsthand and to test experimentally a number of relational-regulation hypotheses. In this research Britain Scott and I were interested in examining one facet of the instantaneous broadcast of war-related images, specifically the videotapes of eight of the Gulf War POWs released by the Iraqis during the air campaign. In a number of these videotapes (obtained from the local CBS affiliate), the POWs made propaganda statements against U.S. involvement in the conflict. Several other tapes, however, simply showed the POWs responding to fact-based questions or conveying personal messages to family and friends. We used this opportunity to obtain observers’ judgments of these videotapes and of the POWs’ opinions on the Gulf War. Several questions interested us in particular. First, in this real-world situation involving extremely constrained behavior, are observers as vulnerable to correspondence bias as the research evidence implies they are, or are they instead able to make the kinds of adjustments for constraint that the conventional wisdom of many experts suggests they can and do? Second, how do observers’ beliefs about relational communication behaviors affect their judgments of these constrained targets’ (POWs) true attitudes? What impact do these kinds of propaganda statements really have on observers’ impressions, and what do constrained communicators try to do to prevent such statements from being believed? Finally, exactly what kinds of cues do narve observers use to discount these targets’ behavior? The content of the Persian Gulf War POW videotapes mirrored a number of themes common to other POW broadcasts (Schein, 1956; Schemer, 1988). For example, one videotape showed Air Force Major Jeffrey Tice reading a prepared statement. The script from which Tice read was grammatically incorrect and he

RELATIONAL REGULATION

253

read it verbatim, making no attempt to alter or correct the flawed grammar: “I was to an attack an Iraqi airfield . . .” In addition, Tice apparently mocked the interrogator by using a fake accent that grew increasingly stronger during his statement. During a television interview on CNN, Tice’s brother pointed out the accent and indicated that he believed Tice had used it to signal that his statement had been forced. Other POWs gave apparently false statements to their interrogators; some of the details present in the tapes differ from the official information about the pilots released by the Pentagon (“Taped Statements,” 1991). All the POWs spoke in a very stilted, halting fashion typical of contrived statements, presumably to communicate that the statements had been forced. These and other techniques run through many of the Persian Gulf War POW videotapes. In this research, undergraduate observers watched sets of the POW videotapes, providing ratings of the POWs’ attitudes toward the war and a number of other opinions. Interestingly, in half of the tapes in each set, the POWs made some sort of anti-war proclamation (the statement POWs); in the remaining tapes in each set they did not (the no-statement POWs). Thus, the videotapes provided us with a naturally occurring independent variable: the presence or absence of a coerced propaganda statement. As Fig. 5 shows, analysis of observers’ ratings of the POWs’ attitudes about the Gulf War revealed that observers judged statement POWs to be significantly less supportive of the conflict than were no-statement POWs. In other words, observers succumbed to the correspondence bias in their ratings of the POWs’ attitudes. Additional analyses suggested that this effect did not result from a procedural artifact; subjects’ ratings of the POWs’ Gulf War attitudes and of how much they agreed with what they had said were significantly negatively correlated, suggesting that observers were not simply reiterating the content of the POWs’ statements, but were instead trying to infer their true attitudes. Surprisingly, given the demonstrated correspondence bias, observers believed that statement POWs had either tried to communicate a hidden message or had tried to “distance” themselves from their behavior significantly more than did nostatement POWs. As Fig. 5 also illustrates, the analysis of observers’ inferences about the POWs’ attitudes toward the Gulf War revealed that for ratings of statement POWs, observers discounted their inferences if they thought the POW had either tried to communicate a hidden message or had tried to “distance” himself from his statement. In addition, among Statement POWs, hypogenerative distancing-less spontaneous or natural facial activity and speech, more eye blinking, and greater attributed fearfulness-was related to greater observer discounting. Among ratings of the no-statement POWs, however, no discounting effect emerged. In addition to rating the POWs’ Gulf War attitudes, observers provided global ratings of several of the POWs’ nonverbal behaviors. In general, there was less variability in ratings of Statement POWs’ behaviors and attitudes than in ratings

254

JOHN H . FLEMING 80

E3

1

fn

60

40

20

0 I

I

I1

NO PROPAGANDA STATEMENT

PROPAGANDA STATEMENT

PRISONER-OF-WAR RATINGS OF SUBJECTS WHO THOUGHT POW SENT A HIDDEN MESSAGE RATINGS OF SUBJECTS WHO DID NOT THINK POW SENT A HIDDEN MESSAGE Fig. 5 . Judgments of Persian Gulf War POWs’ attitudes toward the war. POWs whose videotapes included a propaganda statement were rated as less in favor of the war than were POWs whose videotapes did not include such a statement. Observers discounted POWs’ statements when they thought the POW had sent a hidden message or “distanced.” Data from Fleming & Scott, 1993.

of no-statement POWs. Compared to the less communicative counterparts, observers judged POWs who they thought had tried to communicate a hidden message or to “distance” to be less natural looking and more controlled. These results demonstrate, first, that observers will use the behavior of even severely constrained communicators to infer their attitudes. Second, however, some people appear to be sensitive to the exigencies of such situations and look for clues to the communicators’ true opinions (or at the very least, for some evidence to discount their manifest behavior). In the case of the Persian Gulf War POWs, these clues included how natural and controlled the behavior appeared. In other words, the kinds of relational behaviors that observers attended to were related to how spontaneous, controlled, and natural the behavior appeared. On the basis of these clues, some observers adjusted their inferences of the POWs’ attitudes.

RELATIONAL REGULATION

255

V. Communicating Hidden or Mixed Messages in Multiple-Audience Situations The western film example and the initial multiple-audience problem studies I described earlier (Fleming et al., 1990; Fleming & Darley, 1991)laid the groundwork for an examination of how people may attempt to cope with multipleaudience problems. They also foreshadowed the kinds of elaborate and complex techniques through which communicators can sometimes resolve multipleaudience problems. As those results suggest, in these situations, communicators may try to resolve multiple-audience problems by employing principles of language use to communicate hidden or mixed messages to some members of the witnessing audience that to other audience members are obscure or opaque. However, exactly how can communicators do this and exactly what are these techniques? To set the stage for a discussion of these issues, 1 would like to return briefly to the earlier example. A. DANCES WITH WOLVES AND WALTZES WITH WORDS: THE ART OF COVERT COMMUNICATION Kevin Costner, the director and star of Dunces With Wolves, went to great lengths to make the film as authentic as possible, including taking a crash course in the Lakota language. Lakota-speaking native American actors played the native characters in the film. Unlike the dialogue in previous films about the American West, however, their lines were subtitled in English. However, Lakota is a language that men and women speak differently. As fate would have it, Costner’s language tutor in the film was a woman. As a result, Costner’s Lakota had an incongruity to it that was obvious (and inadvertently humorous) to native Lakota speakers. This incongruity apparently posed something of a selfpresentational dilemma for the native American actors in the film, a dilemma that needed somehow to be acknowledged and resolved. The native American actors in the film knew that their Lakota dialogue would be translated and subtitled, and this prevented them from simply commenting in Lakota that “the well-meaning but unsophisticated actor picked the wrong voice coach” in a manner similar to the sendups previously used in other films. Instead, they used a different approach. In one scene in which Costner and a number of tribesmen are gathered around a campfire speaking Lakota, an elder comments, “that White man speaks funny.” Very few who saw the film would have found that particular comment unusual. After all, it could simply have referred to a non-native speaker’s attempt to learn a new and difficult language. To native Lakota speakers, however, the comment conveyed something more, and did so

256

JOHN H. FL,EMING

quite effectively. It communicated that the actors were not unaware of the language incongruities in the film, but were doing their best to muddle through in spite of them. This acknowledgment allowed the actors to maintain a sense of credibility among members of their broader culture while simultaneously accommodating the exigencies of their roles, This example suggests one possible mechanism by which a communicator can convey messages to certain audiences that remain hidden from other audiences. Theories of language use and research in communication suggest others. B . EQUIVOCATION When such situations arise, communicators can engage in equivocation (formerly referred to as disqualiJication;Bavelas, 1983; Bavelas, Black, Chovil, & Mullett, 1990; Bavelas & Smith, 1982); that is, through equivocation or the creation of interpretive or attributional ambiguity they can attempt to distance themselves from their behavior by obscuring the conclusions observers may draw from their behavior. For example, suppose that you recently received an unusual gift from a friend, but because it is quite unusual, you are unsure whether your friend intended the gift as a joke or whether it should be taken seriously (Bavelas et al., 1990, p. 143). Now you find yourself on the telephone with the gift giver, who asks you how you liked the gift. According to Bavelas et al., this is an example of an avoidance-avoidance communication conflict (Barker, 1942; Bavelas et al., 1990; Dollard & Miller, 1950; Lewin, , 1935, 1938); all of the direct communication options (responses) carry negative implications. On the one hand, you may not wish to hurt the gift giver’s feelings-an outcome that is likely if you either indicate that you did not like the gift or express confusion about the nature of the gift. On the other hand, you do not wish to tell a lie by claiming that you actually liked the gift. These situations may lead to equivocation-indirect speech acts that obscure or blunt the potential negative impact of a direct response. As Bavelas et al. note (p. 145), a response such as, “Ah? . . . It’s quite intriguing” or “Gee . . . that was very thoughtful. I hope I can return the favor someday” may solve such a dilemma. (For a comprehensive analysis of equivocation, see Bavelas et al., 1990.) Equivocation may be a particularly appealing choice of tactic if one knows little about the constitution or the motivations of the various audiences who may witness one’s behavior. C. AUDIENCE DESIGN AND MESSAGE MODIFICATION

The use of equivocation, however, does not obscure or alter the fact that a person actually responded to the question. Instead, equivocation affects the delivery of the response. Most competent speakers of the language would undoubt-

RELATIONAL REGULATION

257

edly understand the meaning and the implications of an equivocal reply. Thus equivocation may be neither a satisfying nor effective strategy, especially if the stakes are high or if the communicator fears that equivocation may also lead to the same kinds of undesirable consequences its use was intended to avoid. Communicators may still be able to accomplish their purposes, however, by using what they know about audience design (A. Bell, 1984; Clark & Carlson, 1982a; Clark & Murphy, 1983; Garfinkel, 1967; Goffman, 1959; Grice, 1975, 1978; Krauss, 1979, 1987; Rommetveit, 1974; Sacks, Schegloff, & Jefferson, 1974). According to language theorists, a communicator’s goal is to convey meaning, or a sense of his or her intentions, to a particular audience (Clark, 1985). Frequently communicators direct their communications at a single audience. E. T. Higgins (198 1) notes that effective communicators not only should configure their communications to suit the receiving audience but that research has demonstrated that communicators do tailor their messages to suit their audiences (Flavell, Botkin, Fry, Wright, & Jarvis, 1968; Glucksberg, Krauss, & Higgins, 1975; E. T. Higgins, 1977; Kraut & Higgins, 1984). In his review, Clark also points out that it is well documented that adult speakers of language are able to engage in audience design-that is, they plan their utterances to be understood by the specific audience to which the utterances are addressed (see Fussell & Krauss, 1989a, 1989b, for examples from the communication literature; see Cialdini, Levy, Herman, & Evenbeck, 1973; Newtson & Czerlinsky, 1974, for examples from the persuasion literature). Even when no specific tactical “plan” is in effect, characteristics of the target audience not only affect the ways in which a message is produced but also affect the way the message is recalled (E. T. Higgins & Rholes, 1978). As I have already suggested, this message modification effect may play a pivotal role in the intrupersonal consequences of roledistancing behaviors. Robert Krauss and his colleagues’ long-standingprogram of research provides a representative look at the ways in which audience design can affect the way one formulates messages. Krauss’s research has frequently used variations of a referential communication task devised some years ago (Krauss & Weinheimer, 1964). In that task, subjects generate descriptions of, or labels for, nonsense figures or other objects that other subjects then use to identify the figures or objects. An early investigation of the role of audience considerations in message formulation demonstrated that subjects instructed to label color chips either for themselves (nonsocial condition) or for someone else (social condition) later identified the chips with a high degree of accuracy when they used the labels they had previously generated for themselves. They exhibited somewhat less accuracy when using other subjects’ social labels, and the least accuracy when using labels that other subjects had generated for their own use (Krauss, Vivekananathan, & Weinheimer, 1968). These results that labels (or messages) differ in their specificity as a function of the audience for whom they were generated.

25 8

JOHN H. E E M I N G

More recently, Fussell and Krauss (1989a, 1989b) have replicated and extended these results to descriptions of nonsense figures prepared for someone else’s or for subjects’ own use. Descriptions prepared for others’ use were less idiosyncratic, less diverse, and longer than descriptions prepared for subjects’ own use. They have also found differences when they further specified the “other” group. Specifically, messages prepared for friends are more efficient, effective, and less redundant than are messages prepared for strangers (Fussell & Krauss, 1989b). These findings all suggest that communicators actively configure their messages to fit the known attributes of the receiving audience.

D. PROBLEMS OF EAVESDROPPERS As I noted, language theorists have generally been content to consider only those messages that communicators formulate for a single audience. In their research on overheard communications, however, Clark and Schaefer (1987) point out that traditional approaches to message formulation and audience design ignore eavesdroppers-other audiences who are not intended to receive the communication, but who may be in a position to do so. The presence of eavesdroppers necessarily changes the nature of the task confronting communicators. Rather than simply communicating directly with the addressee, communicators who are aware of the presence of an eavesdropper must somehow get the addressee to understand them without the eavesdropper doing so. In an investigation of just this sort of communication dilemma, Clark and Schaefer (1 987) had pairs of previously acquainted subjects engage in a referential communication task. They designated one member of the pair as the communicator and the other as the addressee. The communicator’s task was to get the addressee to identify photographs of a series of local landmarks by means of clues provided by the communicator. The communicator and the addressee were free to communicate back and forth, but the presence of an eavesdropper complicated their task. The eavesdropper’s task was also to try to identify the photographs of landmarks from the communicator’s clues. The communicator and the addressee both knew about the presence of the eavesdropper and tried to prevent that person from identifying the landmarks. As Clark and Schaefer (1987) note, communicators have a number of solutions at their disposal to deal with the problem of eavesdroppers. Chief among these solutions is for communicators to engage in concealment tactics or in Goffman’s (1969) terms, covering moves. In essence, concealment involves rendering part or all of the communication unintelligible for the eavesdroppers so that it is understandable only by those for whom it is intended. Communicators can do this by using principles of audience design to determine what references will be understood by the intended audience and, importantly, what references will not be understood by the undesired audience.

RELATIONAL REGULATION

259

Thus, when confronted with multiple audiences (some subset of which are not intended to receive a communication), communicators must provide sufficient references from the material assumed to be held in common with the intended audience so that that audience can recognize the concealed meaning. At the same time, however, the communicator must avoid providing information that would allow the other audience(s) to recognize or even conjecture the concealed meaning. How can this be done? How is it possible for James to communicate a message to Peter, while avoiding having its meaning decoded by Chuck?

E. CONCEALING THE CONTENT OF A COMMUNICATION: COMMON GROUND AND PRIVATE KEYS One solution is for James to determine the shared knowledge* that exists between himself and his addressees-“the knowledge, beliefs, and/or suppositions that a pair of communicators share, and know that they share” (Fussell & Krauss, 1989a, p. 204). He then must select certain private keys (Clark & Schaefer, 1987) from the set of material shared only with Peter. These private keys are specific references drawn from two broad classes of shared information: communal and personal common ground. 1 . Communal Common Ground

Communal common ground is made up of the common background that exists between James and Peter based on their membership in certain groups, cultures, or communities-knowledge that they share because of similar backgrounds or familiar environments that Chuck has not shared or experienced. Foreign languages provide a particularly effective covering tactic for concealing the content of a communication from those for whom it is not intended. For example, if James, Peter, and Chuck are all native English speakers, but only James and Peter speak French, one solution to the communication dilemma immediately becomes apparent. If James wants to communicate with Peter without Chuck understanding, he can shift the communication from English (their shared tongue) to French (which he knows Peter understands but Chuck does not). In his discussion of conversational implicature, Grice (1975) relates an example of a particularly clever concealed message: Take the complex example of the British General who captured the town of Sind [now part of Pakistan] and sent back the message Peccavi. The ambiguity involved (‘I have Sind’/‘I sThis concept has been described by others as common ground (Clark, 1985; Clark & Carlscn, 1981, 1982a, 1982b; Clark & Marshall, 1981). common knowledge (Lewis, 1969), and mutual knowledge (Krauss & Fussell, 1990; Schiffer, 1972).

260

JOHN H. FLEMING

have sinned’) is phonemic, not morphemic; and the expression actually used is unambiguous, but since it is in a language foreign to speaker and hearer, translation is called for, and the ambiguity resides in the standard translation into native English. (p. 55)

2 . Personal Common Ground Personal common ground is similar to communal common ground. It consists of the common ground that exists between James and Peter based on the common personal experiences they have shared-what they have said in previous conversations with each other, but also what they have experienced in each other’s presence, such as trips or parties, or movies, that excluded Chuck. In many ways, personal common ground provides the most effective means for covert communication. Because it is uniquely suited to differentiate among the various audiences who may receive a particular communication, it has the least uncertainty and highest likelihood of recognition by the intended audience. Moreover, the use of personal common ground for covert communications, especially such idiosyncratic “languages” as those developed in close personal relationships, carries with it the inadvertent benefit of increased closeness among its users.

F. CONCEALING THE FACT THAT COMMUNICATION IS TAKING PLACE: ENCRYPTION SYSTEMS Although changing languages can conceal the content of a communication, it cannot hide the fact that covert communication is taking place. Subspecies or variants of languages-ingroup jargon, slang, or argot, however, may obscure both the content of a communication and the fact that covert communication is taking place. These language variants can provide powerful tactical solutions to multiple-audienceproblems. Moreover, certain communities, such as jazz musicians, and con artists, among others, may have prearranged encryption systems designed to allow communicators to convey meaning to one audience while concealing it from other audiences. These systems are hidden (or at least disguised and therefore uninterpretable) and allow their users to converse openly with one audience while simultaneously transmitting different information to those who know the system. Encryption systems can run the gamut from culturally defined emblems (Ekman, 1985)-such as the middle-finger gesture used by the imprisoned officers of the spy ship Pueblo to invalidate the celebrated propaganda photograph they were forced to make-or consensually agreed-upon gestures, such as specially arranged hand signals (Fleming & Darley, 1991). They can also be more easily disguised systems such as slang-a variant version of the common language, filled with novel words and expressions and with ordinary

RELATIONAL REGULATION

26 1

words given other than their ordinary meanings, that facilitate disguised communication (Bourhis & Giles, 1977; Drake, 1978; Dumas & Lighter, 1978; Nunberg, 1978; Sledd, 1965). Slang is particularly well suited for covert communication. During a recent interview, for example, Irving R. Levine, veteran correspondent for NBC News related the following example from his stint as the Moscow bureau chief during the 1950s. At that time Cold War tensions were at their peak and Soviet officials required all members of the American news media inside the Soviet Union to have their daily reports censored. One of the stories that Levine desperately wanted to report concerned the rampant use and abuse of alcohol among senior Soviet officials and members of the Politburo. The censors, however, would never have allowed any direct mention of alcohol abuse. One night, the Soviet leaders held a gala celebration in Moscow complete with fireworks over the Kremlin. The next day Levine’s report contained the following concealed message: “The night sky over Moscow was lit up, and so were the Russian leaders.” In an examination of the role of slang and other encryption systems in resolving multiple-audience problems (Fleming & Darley, 199l), we replicated our earlier findings and extended the multiple-audience effect to situations in which communicators (in this case, teenage insiders) collaborated in their covert signaling strategies. After discussing various encryption systems, insiders made two sets of videotapes. Four observer audiences (the senders’ fellow insiders, a group of the insiders’ friends, a group of adult strangers, and a group of the insiders’ parents) watched the videotapes, showing the communicator making a truthful, a deceptive, and two hidden-message communications. The first of the hiddenmessage communications was directed only at the communicators’ fellow insiders. The second was to signal the correct choice to both the insiders and the insiders’ friends. Results for two different replications used in the study were presented in the upper and lower panels of Fig. 6 and showed first that in general, the teenagers employed their own in-group slang to communicate their hidden messages. Second, with the exception of the insiders, none of the audiences who viewed the tapes differed in lie-detection ability. On the trial containing a hidden message directed only at the insiders, however, the insiders performed significantly better than any of the other three observer audiences. On the trial containing a hidden message directed at both groups of friends, both audiences performed significantly better than did either the parents or the adults. As our research has shown, the task of the communicator confronted with the presence of multiple audiences is thus a tactical one: to say enough to have the message recognized by one audience but not accurately conjectured by the second audience. Moreover, determining what references to use in any given situation requires a high degree of recursive awareness-not only must communicators know something about the various audiences but they must also anticipate what the audiences are likely to know about them. They must “take the role

262

JOHN H . FLEMING

SECRET MEETING TAPES

I

I

TRUTH

LIE

I

I

SENDERS FRIENDS AND ONLY SENDERS

MESSAGE TRIAL Fig. 6. The role of encryption systems in resolving multiple-audience problems. Hidden messages were detected and decoded by different subsets of the witnessing audience; the use of encryption systems such as hand gestures and in-group slang facilitated covert-message transmission. Adapted from Fleming & Darley, 1991.

RELATIONAL REGULATION

263

of the other” and anticipate the likely responses their behavior will elicit from, and the impact their behavior will have on, those others (Elkind, 1980; Flavell et al., 1968; Fussell & Krauss, in press b; Goffman, 1969; Selman, 1971).9

G. BACK-CHANNELING AND “OFF-LINE’ COMMUNICATION ATTEMPTS Communicators face further complications in their covert communication attempts. Because the information that constitutes “shared knowledge” can usually only be approximated, the communicator’s task also contains a high degree of uncertainty and potential bias. For example, recent research by Fussell and Krauss (Fussell & Krauss, 1991, 1992; Krauss & Fussell, 1991) suggests that communicators base their estimations of what others are likely to know on what the communicators know themselves. Thus these estimates of shared knowledge are likely to be egocentrically biased, which can complicate, if not disrupt, the effectiveness of covert communication attempts. These estimates of common ground can affect not only the level of specificity of the initial reference in a referent communication task (how obscure or how comprehensive the initial reference is) but also the number of iterations required for pairs to complete the communication exchange. Sometimes, however, the situation allows the negotiation of communication dilemmas to be an interactive process: The communicator receives cues “on line” from the intended audience that he or she is catching on. These cues exist at two levels: Direct verbal indications and less direct nonverbal “back-channel” cues that the communication is proceeding smoothly (Krauss, Garlock, Bricker, & McMahon, 1977; Krauss & Weinheimer, 1966; Kraut & Lewis, 1982). For example, in the Clark and Schaefer (1987) “overhearers” research previously described, communicators frequently initiated the information exchange by requesting an estimate of the addressee’s shared knowledge base through direct questioning: “Are you from San Francisco?’ Direct probing and back-channel responding can help to reduce the ambiguity of the situation, but if the situation requires concealment of the act of sending a hidden communi’Although the dilemmas facing communicators who are attempting to avoid eavesdroppers are similar in some ways to multiple-audience problems, there are also several important points of difference. It may be useful to distinguish between the two dilemmas by noting that with multipleaudience problems, the primary task of the communicator is not only to communicate a hidden message to the target audience but also to do so without the existence of the message being detected by the constraining audience. The overheard communicator desires accurately to convey a message to one hearer, but need not necessarily mask from the eavesdropper that a hidden message is being transmitted. It is this difference-that all audiences of the communication should receive the message they are intended to receive and no others-that distinguishes multiple-audience problems from the complex, but nonetheless tactically simpler case of overheard communications.

264

JOHN H.FLEMING

cation (as in multiple-audience situations), then these probes and back-channel messages must be equally hidden. However, if the situation simply requires the content of the communication to be concealed rather than its existence (as is the case with overhearers), the receiver is free to back-channel. Clark’s research situation was of this latter sort, and the communicator and audience interacted frequently and overtly during the decoding task. Other situations require communicators to make their best guesses about a communication strategy that will work and to use it “off line” or in the absence of feedback. In these situations, the uncertainty of the shared knowledge can be substantial and can play a pivotal role in the success or failure of such attempts.

H. INDIVIDUAL DIFFERENCES IN COVERT COMMUNICATION

I . Extending the Concept of “Shared Knowledge” Multiple-audience studies suggest that shared knowledge is an effective avenue through which to communicate covertly. Most conceptualizations of shared knowledge, however, view it as the set of explicit knowledge, beliefs, and or suppositions that is assumed to be shared by two communicators on the basis of common experience, background, or group membership. It may be useful, however, to extend the concept of shared knowledge to more implicit and intangible domains, such as shared personality attributes. That is, shared personality attributes, especially those related to interpersonal orientations such as selfmonitoring (SM), may provide communicators with implicit shared knowledge that they can also use effectively to communicate covertly. Ickes and Barnes (1977) have provided evidence suggesting that similarity in self-monitoringorientation facilitates interpersonal interaction. Ickes and Barnes paired previously unacquainted same-sex subjects for a 5-min unstructured interaction during which they unobtrusively recorded both subjects’ verbal and nonverbal interaction behaviors. Some of these dyads comprised mismatched pairs of subjects (i.e., a high SM paired with a low SM) whereas other pairs contained matched subjects (i.e., a low SM paired with a low SM or a high SM paired with another high SM). Because the interpersonal orientations of high SMs and low SMs are presumed to be quite different (and consequently non-complementary; M. Snyder, 1987), Ickes and Barnes reasoned that mismatched subjects would have more difficulty interacting than would the matched pairs. Consistent with this reasoning, their results suggested that the severely mismatched dyads experienced significantly more difficulty interacting (i.e., there were more and longer silences during the interaction) than did the matched jobs. These data suggest one way in which self-monitoring similarity may facilitate

RELATIONAL REGULATION

265

social interaction (cf. Dabbs, Evans, Hopper, & Purvis, 1980). It remains to be seen, however, whether self-monitoring similarity yields similar sorts of effects when the focus moves from the arena of social interaction to that of interpersonal perception and tactical communication.

2 . Self-Monitoring Similarity and Covert Communication The goal in this research was to examine the role that self-monitoring plays in tactical social-communication situations by examining how high SMs and low SMs send and receive messages when confronted with a multiple-audience problem. One would expect high SMs to exhibit significantly more tactical facility (as both senders and receivers) in social communication settings and that they would benefit from a personality-conferred advantage in a variety of such settings. vpically, self-monitoring research has simply examined how high SMs differ from low SMs at some task, such as lie detection (in which only the observer in the situation is taken into account) or emotional expressiveness (in which only the actor in the situation is taken into account). These studies have rarely, if ever, examined the interaction of actors and observers in these situations. In line with Douglas’s (1984) and Ickes and Barnes’s (1977) assertions, however, communication is probably best considered as an endeavor in which successful communication is a joint product of senders and receivers, each group accomplishing its respective communicative tasks. In complex communication tasks, such as that posed in multiple-audienceproblems, even the most adroit communicator can be stymied by an inattentive or socially insensitive observer, and the most perceptive observer can come up empty handed at the whim of an even modestly ineffective communicator. Thus, this research was designed to examine how selfmonitoring similarity between communicators and observers affects tactical social communication and the inferences drawn from those communications. In two studies Brian Kojetin and I examined the tactical communicationfacilitating properties of self-monitoringsimilarity (Fleming & Kojetin, 1993). In the first study, observers more accurately detected and correctly decoded mixed messages (those with a surface content intended for one audience and a hidden content intended for a second audience) when they were similar to the communicators in their self-monitoring propensities, regardless of whether the senderreceiver pairs were matched high SMs or matched low SMs. Dissimilar observers evaluated mixed messages significantly less effectively. These results are presented graphically in the left-hand panel of Fig. 7. As the figure shows, however, this differential facilitation effect did not extend to non-mixed (i.e., either truthful or deceptive) messages. A second experiment employed a traditional personality-attribution paradigm in which author-subjects composed assigned essays describing themselves as either extreme introverts or extreme extroverts under typical constrained essay

STUDY 1 0.6

-

0.4

-

0.3

-

0.5

02

0.1

SIMILAR OBSERVERS

q,

--t DISSIMILAROBSERVERS

STUDY 2

"1

-

EXTRAVERTEDPORTRAYALS

--C DISSIMILAR OBSERVERS

17

INTROVERTED PORTRAYALS MESSAGE

Fig. 7. Effects of self-monitoring similarity on information transmission and personality attribution. (Left) Hidden messages were detected and decoded by observers who were similar to sender in terms of self-monitoring propensities more often than by dissimilar observers. (Right) Similar observers were more accurate on mixed message/personality-inconsistent essays than were dissimilar observers. Similar observers were less accurate on straightforward constrained essays than were dissimilar observers. Data from Fleming & Kojetin (1993).

22

17

I

t

E

1 STANDARD

COMMUNICATE

ESSAY INSTRUCTIONS

RELATIONAL REGULATION

267

instructions and recorded them on videotape (“standard” instructions condition). Next, authors composed a second essay in which they attempted to subtly (tactically) communicate their true personalities while still following the assigned easy instructions to present themselves as extreme introverts or extroverts (“communicate’’ instructions condition; see also Fleming et al., 1990). After writing the second essays, authors also recorded them on videotapes. Observers who were either similar to or discrepant from the authors in self-monitoring later viewed the resulting tapes. We expected observers in similar author-observer pairs to more accurately interpret mixed-message essays than would observers in dissimilar pairs (i.e., essays prepared under the “communicate” instructions). Moreover, we expected observers in similar pairs to perform better when the surface content of the essay and its mixed message were inconsistent (i.e., actual extrovert claiming to be an extreme introvert-the surface content and the hidden message differ both in valence and degree) rather than when the two were consistent (Le., actual extrovert claiming to be an extreme extrovert-the surface content and the hidden message differ in degree, but not in valence). In contrast to these mixed-message predictions, however, we anticipated that observers in similar author-observer pairs would be significantly less accurate than were observers in dissimilar pairs when interpreting constrained behavior in which no mixed message was embedded. We made this latter prediction for a number of reasons. First, we selected the personality dimension used in this research because of its direct relationship to interpersonal interaction and perception. Therefore, we assumed that authors and observers in similar pairs possessed some interpersonal shared knowledge with which to communicate and interpret their mixed messages. In previous research, however, shared knowledge was a double-edged sword-the very attributes that led to enhanced tactical communication also led to more effective lying (Fleming et al., 1990; Fleming & Darley, 1991). That is, if communicators intend to mislead an audience and share knowledge with that audience, they can use the shared knowledge to make the lies they tell more convincing. Second, we instructed authors to “persuade another person” that they possessed an extreme personality. Insofar as authors intended to fulfill their tasks successfully, we expected similar pairs to be significantly more vulnerable to correspondence bias than were dissimilar pairs, irrespective of the personality that authors portrayed. The results, presented in the right-hand panel of Fig. 7, supported the predictions. As the figure shows, we replicated the general findings from the first study. Although matching senders and observers on their SM scores facilitated the communication of content-inconsistent mixed messages, similar observers were significantly more likely than were dissimilar observers to fall prey to correspondence bias when the messages were not mixed. Although the designs were complex, the main findings to emerge from these

268

JOHN H. FLEMING

studies are relatively straightforward. It is worth reiterating, however, that the results suggest that self-monitoring similarity confers a special status on social interactants who possess it, a status that has not only been shown to facilitate dyadic interaction (Ickes & Barnes, 1977) but that also appears to facilitate the tactical communication of information and personality attributes as well. As such, this research joins the ranks of a small but growing body of similar findings that demonstrate that examining individual-difference similarities among interactants or communicators may prove to be a fruitful avenue of investigation and may shed some light on the complexities of interpersonal perception and inference. The similarity-conferred advantage demonstrated in this research, however, does not come without a cost. Just as in previous research in which blatantly misleading statements misled observers who possessed shared knowledge with a communicator more than those statements misled observers who did not possess shared knowledge (see Fleming et al., 1990, Studies 2 and 3). similar observers fell prey to correspondencebias significantly more often when making inferences based on constrained behavior than did dissimilar observers. Thus, personality similarity, like shared knowledge, can be a double-edged sword: Under certain circumstances it can enhance communicators’ abilities to communicate accurately and observers’ abilities to interpret accurately those communications. Under other circumstances, however, personality similarity places observers at a decided disadvantage, leading them to make more biased (and presumably, less accurate) inferences.

1. THE ROLE OF SUSPICION IN MULTIPLEMESSAGE COMMUNICATION: FLOUTING QUALITY AND EXPLOITING MANNER In the case of hidden-message communications, what is the process whereby the audience becomes aware that the surface content of a message should be disregarded in favor of some concealed and yet-to-be-delivered content? As with the interpretation of distancing cues described in preceding sections, the communicator’s task is to make the intended audience suspicious about the motivations that underlie the surface content of the message so that they will disregard it. This can be done by using audience design, message-modification techniques, and such other linguistic devices as suspicion actuators. The communicator must then convey the concealed content in a way that the intended audience can recognize and decode. One key to understanding the techniques that senders and authors used to signal the presence of a hidden message in the multiple-audience studies previously described can be drawn from Grice’s (1975) work on conversational max-

RELATIONAL REGULATION

269

ims. Grice draws a distinction between blatant violations of a particular maxim, which he callsflouring the maxim, and blatant violations that in and of themselves carry a particular implication, which he calls enploiring a maxim. Two of his maxims are especially relevant to discussing how suspicion can be triggered in advance of a hidden message: quality and manner. Not surprisingly, a communicator can exploit the maxim of quality by relating incongruous private keys or other inaccurate or false information from the common ground shared with the intended audience. For example, when describing which of two pencils he preferred, one sender in our research noted that the pencil he preferred was yellow and thus reminded him of his favorite football team, the Los Angeles Rams. In fact, his favorite team was the Miami Dolphins, an incongruity that the members of the intended audience immediately noticed, but that members of the unintended audience did not. These discrepancies may trigger suspicion among members of the intended audience and may cause them either to attend more closely to what follows (Atkinson & Allen, 1993; Pyszczynski & Greenberg, 1981) or, paradoxically, to completely disregard the manifest content of the communication (Fein et al., 1990; Zuckerman, Kernis, Driver, & Koestner, 1984). In addition to violating the maxim of quality, however, communicators can also engage in an alternate and more subtle approach to triggering suspicion in the intended audience by exploiting the maxim of manner, or its supermaxim. Be perspicuous.” Grice describes three forms that violations of the maxim of manner can take: ambiguity, obscurity, and lack of brevity or succinctness. Ambiguity arises when more than one specific meaning is conveyed by an utterance. The previous example of the British general’s message about the capture of Sind is an example of ambiguity generated by flouting the maxim of manner. According to Grice, obscurity arises in the case of potentially overheard communications. That is, a communicator may become intentionally obtuse or telegraphic to indicate that the message is not intended for consumption by the unintended audience. For instance, in the presence of a child, parents may spell out the name of a possible Christmas gift to prevent the child from understanding the discussion: “Let’s buy little Joshua a T-R-I-K-E for C-H-R-I-S-T-M-A-S.” Finally, a failure to be brief-or an overspecification of an obviously well-understood conceptcan also trigger suspicion. For example, rather than stating “Mary is at home working on her dissertation,” one could say, “Mary is at the location that she considers her domicile working on that very detailed and comprehensive written work for which she will ultimately receive an advanced degree.” All of these tactics work by violating the generally expected rules of communication in a blatant way that arouses the attention and curiosity of the intended audience. Noticing the violation and conjecturing the implication of the violation lead to a readiness to receive the hidden content. Interestingly, all of these conversational exploitations come together in the

270

JOHN H. FLEMING

forced confession prepared by Captain Lloyd Bucher, commander of the U.S. Navy spy ship USS Pueblo, while imprisoned by the North Koreans: We who have been rotating upon the fickle finger of fate for such long languid months give our word to the Great Speckled Bird that we will heretofore in all sincerity cleanse ourselves of rottenness and vituperations. We solemnly await our retum to our loved ones so that the fickle finger can be replaced by the rosy fingers of dawn and salvation. So help me, Hanna. (Byme, 1991)

Once a communicator arouses suspicion within the intended audience, the next step is to convey the concealed content. Again, a communicator can accomplish this through any of the mechanisms previously described. Taken together, the results of the hidden-messageresearch suggest that communicators with differing degrees sophistication (i.e., high school students, undergraduates, graduate students, and other adults) can effectively resolve multiple-audience problems by using their knowledge about language and about the witnessing to construct hidden messages that accomplish the desired communication, avoiding detection by another simultaneously present audience. As Clark’s analysis predicted, the preferred means of constructing hidden messages was to use private keys.

VI. Conclusions and Future Directions In this article I have offered a relational perspective as a framework for understanding the ways in which tactical communication and impression management can frame and channel social interaction and self-perception processes. From Beirut hostages and Persian Gulf War POWs forced to make propaganda videotapes to politicians being interviewed on the news to children at play in the schoolyard, at one time or another we all mentally (and sometimes physically) “cross our fingers” and engage in behaviors that distance us from our actions. At other times we cling to our opinions and desired affiliations. As I have argued, these tactical relational-regulation techniques serve a pair of interrelated regulatory functions: They help insulate individuals from the ill effects of negative or self-discrepant behavior while simultaneously signaling observers that the individuals do not wish to be dispositionally associated with some aspect of their own behavior, the behavior of others, or the social situation. Under certain circumstances individuals can even construct communications that convey distinct and separate messages to different simultaneously present audiences. Multiple-audienceproblems serve both as a prototype of the kind of communication conflict that precipitates relational-regulationbehaviors and as a focal point for a relational analysis of tactical communication processes. Using this perspec-

RELATIONAL REGULATION

27 1

tive, I have described a number of experimental investigations that suggest that the paradigm and this approach are both fruitful and promising. Ultimately, I expect this research to allow us to better understand how people regulate their relative “positions” with respect to other elements in the social field, and the implications of these behaviors for self- and other perception. The relational-regulation perspective I have proposed can address a number of issues of both applied and theoretical interest. For example, on a theoretical level, the adoption of a relational-regulation approach may help to provide a framework for better understanding the relationship between verbal and nonverbal behavior on the one hand, and embarrassment (Edelmann, 1985; Modigliani, 1968) and social anxiety (Leary, 1982, 1983) on the other. The repertoire of behaviors that people use to regulate their relation to their own behavior is remarkably similar across these domains, yet there is little agreement at present about the roles that these behaviors play in the intrapersonal dynamics of self-concept maintenance and modification. On an applied level, adopting a relational-regulation approach may shed light on issues as varied as the tactical nature of electronic communication in organizations and at work, relational techniques in written communications such as recommendation letters, and factors that influence the credibility of witnesses in court. In employment and organizational settings, for instance, the increased use of electronic media for routine correspondence has not only changed the nature of even the most mundane kinds of communications (Gutek & Bikson, 1985; Huff,Sproull, & Kiesler, 1989; Iacono & Kling, 1984; National Research Council, 1987) but it has also expanded individuals’ access to those communications. One might expect to find an increased use of tactical relational-regulation devices in such correspondence, especially when that correspondence is accessible to a larger audience made up of individuals with diverse motives and purposes within an organization. In other words, the expanded access to even routine inter-ofice and inter-departmental memos means that authors of such correspondence must be aware of and manage the impact of those communications on the numerous individuals who may see them. Moreover, even these subtle alterations in individuals’ communication styles may have profound ripple effects throughout an organization (National Research Council, 1987). In the clinical domain, it would be useful to understand the ways in which outward behavior affects one’s internal self-representations. If the therapeutic benefits of self-concept modification (such as those suggested by behavior or cognitive therapy) are to be fully realized, it is essential to understand the intrapersonal mechanisms that may prevent overt behavior from affecting, and ultimately modifying, the self. Similarly, understanding the ways in which people insulate themselves from the consequences of their role-based behavior (i.e., “emotional work” required by the roles of service professionals or individuals involved in “nontraditional” relationships, or by roles generated by dysfunction-

272

JOHN H. FLEMING

al, inequitable, or multigenerational family dynamics) may lead to more effective therapeutic interventions and applications (Hochschild, 1983; Masten, Best, & Garmezy, 1990; Masten & O’Connor, 1989; Pearlman, 1989). On a broader level, understanding the ways in which people indicate that their behavior either should or should not be used to draw dispositional inferences about them is central to an understanding of interpersonal remedial tactics such as disclaimers (Hewitt & Stokes, 1975; Stokes & Hewitt, 1976), excuses (C. R. Snyder, 1985; C. R. Snyder & Higgins, 1988; C. R. Snyder et al., 1983), accounts (Gonzales, Manning, & Haugen, 1992; Gonzales, Pederson [Haugen], Manning, & Wetter, 1990; Scott & Lyman, 1968), and apologies. All of these potential applications share the common metric of relational regulation and all await a thoroughgoing appreciation for inter- and intrapersonal relational-regulation behaviors such as distancing and embracing.

A. ON THE AUTOMATIC-DELIBERATE DISTINCTION IN RELATIONAL REGULATION Like those who have also struggled with the vocabulary of tactical behavior, I do not mean to suggest that any of the tactical relational-regulation behaviors I have discussed are necessarily, primarily, or predominantly consciously mediated or deliberate (although they may at times be), as terms such as tactics, place distance, embrace, take back, and undo, generally imply. Instead, I acknowledge that the vocabulary is less than perfect and use these terms to simplify the description of how these collections of conceptually similar behaviors manifest themselves. Much of the previous research on distancing behaviors and multiple-audience problems has focused on intentionally communicative, tactical behaviors such as grimaces, smiles, eye-rolls, hidden verbal and nonverbal messages, hand gestures, and the like. However, as impression-management theorists have noted, it is not necessarily useful to cast relational-regulation behaviors or other sorts of impression-management behaviors as belonging exclusively either to a class of overtly deliberate behaviors or to a class of unintentional automatic behaviors; they are probably best considered as falling along an “intentional-automatic’’ continuum. Extreme kinds of signaling behaviors clearly define the “intentional” or “tactical” end of that continuum. With the exception of extremely obvious cases (such as blinking Morse code with one’s eyelids or sarcastically rolling one’s eyes), however, in ongoing interaction an attempt to draw a clear distinction between behaviors that are “automatic,” “spontaneous,” “controlled,” or “intentional” may best be artificial and unnecessarily limiting. At worst, it may be impossible (for discussions of this and related points, see DePaulo, 1992;

RELATIONAL REGULATION

273

Ekman & Friesen, 1969; Leary & Kowolski, 1990; Schlenker & Weigold, 1992). As self-concept regulating devices, both deliberate and inadvertent behaviors probably hold equal stature, although they may differ substantially in their exact methods of influence. It is not necessarily awareness of the act, but the act itself that generates the self-insulating effect (Stepper & Strack, 1993; Strack et al., 1988). Thus, it may not be necessary for a person to be aware of the enactment of a specific behavior for it to exert systematic effects on his or her mood and judgment. Nor is it necessary for those behaviors to be identifiable as intentionally generated communicative acts for them to be used as discounting cues by observers. Clarification of these issues, however, remains an issue left for future research.

B. THE MILGRAM OBEDIENCE STUDIES: A RELATIONALREGULATION ANALYSIS As I have noted throughout this article, I believe a relational-regulation perspective can be used to integrate and consolidate a number of seemingly unrelated findings in the social-psychological literature under one theoretical “umbrella.” I would like to conclude by exploring briefly a small sample of research that this approach can help elucidate. Several of these studies were alluded to in the overview. Although the Tesser and Rosen (1973, Cialdini et al. (1976), and Davis and Jones (1960) research is consistent with a relational-regulation analysis, other research not usually so construed can also be reconceptualized in tactical-relational terms. For example, although it is not usually regarded as such, Milgram’s (1974) obedience paradigm provided an early demonstration of the power of relational-regulation behaviors to communicate constraint and to insulate the self from the potential consequences of negative behavior. By way of review, male subjects in that research were induced to deliver a series of electric shocks to a learner-subject at the urging of an authority figure (i.e., an experimenter). The main findings of that research, namely, that subjects will obey a person in authority well beyond the limits most people would consider reasonable, have found their way into the folklore of social psychology. Another set of findings, however-equally intriguing-have remained in relative obscurity. According to Milgram (1974), subjects in the obedience experiments confronted a variant of a multiple-audience problem: In this situation the subject must resolve a conflict between two mutually incompatible demands from the social field. He may continue to follow the orders of the experimenter and shock the learner with increasing severity, or he may refuse to follow the orders of the experimenter and heed the learner’s pleas. (p. 24)

274

JOHN H. KEMING

Milgram noted further: A conflict develops between the deeply ingrained disposition not to harm others and the equally compelling tendency to obey others who are in authority. The subject is quickly drawn into a dilemma, and the presence of high tension [among the subjects] points to the considerable strength of each of the antagonistic vectors. (p. 43)

Rather than regarding the situation as one in which competitive drive-like states were at work (as Milgram did), the obedience situation can be reconstructed both as an impression-management or self-presentationaldilemma and as a self-perception dilemma. On one hand, subjects may wish to be regarded as compliant and respectful of authority by the experimenter. This course of action, however, puts them at risk for being seen as callous and uncaring by the learner, by themselves, and by some potential (and as yet undefined) external audience. On the other hand, subjects may also have wished to see themselves and to be regarded by the learner as decent, gentle, and sympathetic. This course of action, in contrast, puts them at risk for being seen as a deviant or as a “boat rocker” by the experimenter. From the multiple-audience perspective, Milgram’s subjects experienced an acute role conflict. In light of the Fleming and Rudman (1993) results, we might have expected Milgram’s subjects to engage in greater amounts of verbal and nonverbal distancing behaviors as the severity of the shocks they administered increased. Such a perspective would have helped explain some observations that to Milgram were a curiosity-that even as subjects proceeded to deliver stronger and stronger shocks to the learner, they disassociated themselves from their actions through outright dissent. Indeed, Milgram noted that as the tension subjects experienced increased with each successive shock, so too did their tendency to try to verbally distance themselves from their actions. For example, Milgram cites numerous instances when subjects dissented: They expressed confusion about what they should do next, denounced the experiment as stupid or senseless, indicated disapproval of the procedure, disagreement that the experiment should continue, and suggested that they were not the kinds of people who would engage in that sort of behavior. Milgram notes that dissent can serve a “dual and conflicting” multiple-audience function (p. 160); it can reduce the apparent discrepancy between a person’s self-concept and his or her actions while simultaneously allowing him or her to perform a self-discrepant act. This notion is key to understanding the potential paradoxical properties of distancing behaviors-the more emphatically people signal their disavowal of responsibility for performing a negative self-discrepant action, even though they are clearly performing that act, the longer the act can be performed or the more selfdiscrepant the action can be (Kelman & Hamilton, 1989). Similarly, Milgram’s observers, in their contemporaneous notes, recorded a

RELATIONAL REGULATION

275

wide variety of behaviors that indicated extreme anxiety, shame, and embarrassment. Specifically, observers noted a tendency for subjects to avert their gaze from the learner (p. 34), to look down or away (p. 34), to hide their faces (p. 34), to heave sighs of relief (p. 33), and to rub their fingers over their eyes (p. 33). These behaviors are remarkably similar to those distancing behaviors that attenuated dissonance-induced attitude change in previous research (Fleming & Rudman, 1993). Finally, Milgram noted that some subjects engaged in subterfuge. That is, they tried to signal the correct answer to the learner or tried to deliver a lower shock than they reported (see Davis and Jones’s 1960, discussion of similar issues). Again, these tactics served a multiple-audience function; they allowed subjects to “appear as if” they were obeying, while engaging in behaviors designed to minimize the impact of that obedience on the learner. Similarly, some subjects tried to regulate the duration of the shocks by pressing the shock lever very gingerly. As Milgram notes, all of these tactics serve to deemphasize the act as a basis for self-judgment; by trying to do something (however minimal or tangential), subjects gave themselves and others something with which to discount their obedience as a basis for dispositional judgments. Given these observations of subjects’ distancing behaviors in the obedience situation, we would predict that those subjects who showed the greatest amount of distancing would be least affected by their obedience. In contrast, among subjects who showed very little distance, we would expect a great deal of selfdiscrepancy (dissonance) induced opinion shift. Similarly, we would expect greater self-reported anxiety as a reconstructive defense among subjects who obeyed to the end than among subjects who ultimately disobeyed. This latter result received some tentative support in Milgram’s data: Obedient subjects (those who continued to the end of the shock generator) reported themselves as having been more tense and nervous at their highest level of tension than did subjects who disobeyed.

C . RELATIONAL REGULATION IN PUBLIC AND PRIVATE LIFE I . Relational Regulation in the Media and the Courtroom

The implications of tactical relational-regulation behaviors extend well beyond the boundaries of simple face-to-face interaction. Their effects can be found whenever communicators face the presence of multiple audiences. For example, newspaper reporters (especially those who cover political issues) may sometimes find themselves personally supporting positions or candidates that are in disagreement with either the stated policies and positions of their editorial boards, or with their respective readership. In these situations, reporters may attempt to

276

JOHN H. FL,EMING

distance themselves from the content of their stories. In fact, in an examination of a similar issue (Hess & Gossett, 1974), a sample of Northern newspaper editorials on Nixon’s “Vietnamization”policies contained more distancing (nonimmediacy) transformations than did comparable Southern editorials, even though the manifest content of the regional editorials did not differ. Along somewhat different lines, the effectiveness of witnesses’ testimony may be influenced by the extent to which they alter their relation to their testimony. For example, courtroom testimony offered in the presence of an accused person whom the witness may fear (a multiple-audience problem) may differ in qualitative ways from testimony that is offered when the accused is absent (such as when a deposition is taken). In other words, witnesses who must testify in open court may try to distance themselves from their testimony, and a jury may, in turn, find the testimony less compelling, believable, and persuasive (Erickson, Lind, Johnson, & O’Barr, 1978). These and other potential applications of the relational-regulationperspective are as yet largely unexplored. Nonetheless, they hold out the promise of more fully understanding the ways in which audiences affect the ways in which communications are delivered and the impact those communications will ultimately have on those who witness them.

2 . Tactical Relational Regulation in Political Campaigns: “Spin Control” and Image Management In the wake of contradictory newspaper headlines concerning the United States’s “official” reaction to the then-smoldering Persian Gulf crisis, National Public Radio reported on the Bush administration’s need to “speak with different voices”-to send mixed messages domestically and abroad, by issuing military threats for hard-liners on one hand, while simultaneously advocating diplomatic solutions for moderates on the other. On National Public Radio’s “All Things Considered” program, Ted Clark wondered whether such mixed messages were possible and questioned the wisdom of such an approach if indeed they were. The program of research described here suggests that mixed messages are indeed possible and are perhaps even an essential part of day-to-day social interaction. President Bush, however, is neither alone in speaking, nor the first in public office to “speak with different voices” on sensitive social and political issues. As most adroit politicians quickly realize, learning to “distance one’s self” from unfortunate people and events, “embracing” patriotic and other symbolic images and values, and mastering the art of “spin control” are essential ingredients of contemporary public life. Relational regulation, however, is just as much a part of the fabric of everyday life as it is a concern for those in the public eye; public awareness of the notion of “distancing” is so widespread and intuitive that the term has become a part of the vernacular.

RELATIONAL REGULATION

277

3. Tactical Relational Regulation in Academe In the tactical-communication arena, few topics are as intriguing and potentially useful as letters of recommendation. They can assure access to higher levels of education for some, and deny that access to others. At any number of levels in education and professional life, a sterling recommendation letter can be the key to advancement; a poor one can be a guarantee of rejection. As Siskind (1966) has noted, however, unlocking the truth buried within such a letter is a more complicated task than simply reading and evaluating its surface content. Recommendation letters offer a veritable gold mine for examining the ways in which people use relational devices to regulate the association between what they say and what they really mean. Preparing a recommendation letter is much like facing a multiple-audience situation. On the one hand, the recommender may have an implicit obligation not to betray the trust of a student. On the other hand, a recommender may not wish to compromise his or her “track record” by strongly recommending someone who shows only marginal promise. Providing a better sense of how to “read between the lines” to determine whether an evaluator’s glowing comments should be taken as face-valid, or whether they should be discounted instead because “the flag is up,” would be a useful contribution in its own right. If such a contribution also permitted more effective evaluation of this essential assessment tool as a by-product, it would be all the more useful. 4 . Interpersonal Benefits of Covert Signaling Systems

Beyond the media, courtroom, and political arenas, close relationships provide an idea setting in which to examine the consequences of tactical relationalregulation behaviors. Specifically, close relationships are a fertile source of shared knowledge from which can be drawn the private keys that make covert communications possible. They are also likely to create the motivation for sending hidden messages: Those in the relationship may wish to continue their particular dialogue without having it understood by others, to arrange and communicate plans for future meetings without including those who may overhear those plans, and to experience the sheer pleasure of sharing secret references that do not communicate anything of substance to outsiders. All of these reasons may lead to frequent covert messaging by those in close relationships. Conversely, because covert messaging is likely to increase the bonding between those who participate in it, and is also likely to increase the social distance from those who do not participate in it, it may be one mechanism that contributes to the increased feelings of closeness of those in such relationships (R. Bell, Buerkel-Rothfuss, & Gore, 1987; Drake, 1978; Pearlman, 1989; Schein, 1956). It may also contribute to the increased isolation of those in such relationships from the other groups

278

JOHN H. FLEMING

from whom they hide various communications. Evidence for these possibilities has been provided by R. Bell and his colleagues (1987). They found that individuals involved in close relationships use a variety of idiomatic expressions to communicate covertly and that the number and frequency of use of such expressions was directly related to how loving, committed, and close they rated their relationships. This was particularly true of expressions used to refer to affection, sexual matters, and the initiation of sexual activities. One particularly noteworthy example (p. 47) allowed one subject to determine whether his partner had remembered to insert her contraceptive diaphragm: “Did you bring the yarmulke for the cabbage patch kid?’ The kinds of issues I have explored here, including self-conception, social communication, interpersonal perception, and self-regulation, are some of the focal concerns of contemporary social psychological research. The multipleaudience problem paradigm and the relational-regulation perspective I have suggested are potentially useful vantage points from which to examine these concerns. Moreover, the analysis of the tactical aspects of relational regulation I have proposed is intended to reawaken a dormant, though useful, metaphor in social psychology. Through the relational metaphor, we may begin to understand the interrelationships between a number of different basic aspects of tactical communication, self-presentation, impression management, and selfconstruction. Not only will an examination of these issues provide a framework for a more comprehensive understanding of self-concept regulating and tactical interpersonal behavior but will also lay the groundwork for a more thorough examination and synthesis of issues from a number of relatively disparate areas of research and theory within social psychology, such as strategic selfpresentation, verbal and nonverbal communication, lying and the detection of deception, correspondence bias, and social communication. For these reasons, multiple-audience problems and the tactical relational-regulation behaviors they engender seem to be a potentially promising and intriguing arena for future research.

Acknowledgments

The generous support of the Graduate School of the University of Minnesota facilitated the preparation of this article through a Research Fellowship to John H. Fleming. I gratefully acknowledge the helpful suggestions of John Darley, Joel Cooper,Marti Hope Gonzales, Mark Snyder, Ellen Berscheid, Howard Lavine, Laurie Rudman, Brian Kojetin, and the members of my 1992 research group-Britain Scott, Scott Elkins, Julia Brock, Pamela Regan, and Brian Smith-who commented on earlier drafts of this article.

RELATIONAL REGULATION

279

References Anderson, N., & Hubert, S. (1963). Effects of concomitant verbal recall on order effects in personality impression formation. Journal of Verbal Learning and Verbal Behavior, 2. 379-391. Archibald, W. P., & Cohen, R . L. (1971). Self-presentation, embarrassment, and facework as a function of self-evaluation, conditions of self-presentation, and feedback from others. Journal of Personality and Social Psychology, 20, 287-297. Aronsson, K.,& Satterlund-Larsson, U.(1987). Politeness strategies, and doctor-patient cornmunication: On the social choreography of collaborative thinking. Journal of Language and Social Psychology, 6, 1-27. Asch, S. (1946). Forming impressions of personality. Journal of Abnormal and Social Psychology, 41, 258-290. Atkinson, M . L., & Allen, V. L. (1983). Perceived structure of nonverbal behavior. Journal of Personality and Social Psychology, 45, 458-463. Bargh, J. A., & Pietromonaco, P. (1982). Automatic information processing and social perception: The influence of trait information presented outside of conscious awareness on impression formation. Journal of Personality and Social Psychology, 43. 437-449. Barker, R. G . (1942). An experimental study of the resolution of conflict by children: Time elapsing and amount of vicarious trial-and-error behavior occurring. In Q. MacNemar & M .A. Merrill (Eds.), Studies in personality (pp. 13-34). New York: McGraw-Hill. Baumeister, R. F. (1982). A self-presentational view of social phenomena. Psychological Bulletin, 91, 3-26. Baumeister, R. F., & Tice, D. M. (1986a). Emotion and self-presentation. In R. Hogan & W. H. Jones (Eds.), Perspecrives in personality (Vol. 2, pp. 181-199). Greenwich, CT: JAI Press. Baumeister, R. F., & Tice, D. M. (1986b). Four selves, two motives, and a substitute process selfregulation model. In R. F. Baumeister (Ed.), Public selfandprivare self(Chap. 3, pp. 63-74). New York: Springer-Verlag. Bavelas, J. B. (1983). Situations that lead to disqualification. Human Communicarion Research, 9, 130- 145. Bavelas, J. B., Black, A,, Chovil, N., & Mullett, J. (1990). Equivocal communication. Newbury Park, CA: Sage. Bavelas, J. B., & Smith, B . J. (1982). A method for scaling verbal disqualifications. Human Communicarion Research, 8, 214-227. Bell, A. (1984). Language style as audience design. Language in Society, 13, 145-204. Bell, R. A,, Buerkel-Rothfuss, N. L., & Gore, K. E. (1987). “Did you bring the yarmulke for the cabbage patch kid?’ The idiomatic communication of young lovers. Human Communication Research, 14, 47-67. Bem, D. J. (1965). An experimental analysis of self-persuasion. Journal of Experimental Social Psychology, I , 199-2 18. Bem, D. J. ( 1967). Self-perception: An alternative interpretation of cognitive dissonance phenomena. Psychological Review, 74, 183-200. Bem, D. I. (1972). Self-perception theory. Advances in Experimenral Social Psychology, 6 , 1-62. Berger, P. (1975). Sociological perspectives-Society as drama. In D. Brissett & C. Edgley (Eds.), Life as theater: A dramaturgical sourcebook (pp. 13-22). Chicago: Aldine. (Original work published 1963) Bogardus, E. S . (1925). Measuring social distances. Journal of Applied Sociology, 9. 299-308. Bond, C. F., Jr., & Anderson, E. L. (1987). The reluctance to transmit bad news: Private discomfort or public display? Journal of Experimental Social Psychology, 23, 176- 187.

280

JOHN H. FLEMING

Bond, M. H., & Lai, T. (1986). Embarrassment and code-switching into a second language. Journal of Social Psychology, 126, 179-186. Bourhis, R. Y., & Giles, H. (1977). The language of intergroup distinctiveness. In H. Giles (Ed.), Language, erhniciry, and intergroup relations. New York: Academic Press. Brown, P., & Levinson, S . (1978). Universals in language use: Politeness phenomena. In E. N. Goody (Ed.), Questions and politeness. Cambridge, England: Cambridge University Press. Brown, P., & Levinson, S . (1987). Politeness: Some universals in language usage. New York: Cambridge University Press. Buck, R. (1980). Nonverbal behavior and the theory of emotion: The facial feedback hypothesis. Journal of Personality and Social Psychology, 38, 81 1-824. Buck, R. (1984). The communication of emotion. New York: Guilford. Burchard, W. W. (1954). Role conflicts in military chaplains. American Sociological review, 19, 528-535. Burgoon, J. K., Parrott, R., LePoire, B. A., Kelley, D. L., Walther, J. B., & Perry, D. (1989). Maintaining and restoring privacy through communication in different types of relationships. Journal of Social and Personal Relationships, 6 . 131-158. Buss, A. H. (1980). Self-consciousness and social anxiety. San Francisco: Freeman. Byrne, C. (1991, January 28). For Peublo commander, “confessions” trigger memories of pain and survival. Minneapolis Star-Tribune, pp. 1, 11A. Cacioppo, J. T., & Petty, R. E. (1981). Electromyographic specificity during covert information processing. Psychophysiology, 18, 518-523. Carver, C. S . , & Scheier, M. F. (1978). Self-focusing effects of dispositional self-consciousness, mirror presence, and audience presence. Journal of Personality and Social Psychology, 36,324332. Carver, C. S . , & Scheier, M . F. (1981). Attention and self-regulation: A control rheory approach to human behavior. New York: Springer-Verlag. Carver, C. S . , & Scheier, M. F. (1985). Aspects of self and the control of behavior. In B. Schlenker (Ed.), The self and social life (pp. 146-174). New York: McGraw-Hill. Cialdini, R. B. (1988). Injluence: Science andpractice (2nd ed.). Glenview, IL: Scott, Foreman and

c o. Cialdini, R. B., Borden, R. J., Thorne, A., Walker, M. R.,Freeman, S . , & Sloan, L. R. (1976). Basking in reflected glory: Three (football) field studies. Journal of Personality and Social Psychology, 34, 366-375. Cialdini, R. B., & DeNicholas, M. E. (1989). Self-presentation by association. Journal of Personality and Social Psychology, 57, 626-631. Cialdini, R. B., Levy, A., Herman, C. P., & Evenbeck, S . (1973). Attitudinal politics: The strategy of moderation. Journal of Personality and Social Psychology, 25, 100-108. Cialdini, R. B., & Richardson, K. D. (1980). ’hvo indirect tactics of image management: Basking and blasting. Journal of Personality and Social Psychology, 39, 406-415. Clark, H. H. (1985). Language use and language users. In G. Lindzey & E. Aronson (Eds.), The handbook of social psychology (3rd ed., Vol. 2, pp. 179-231). New York: Random House. Clark, H. H., & Carlson, T. B. (1981). Context for comprehension. In J. Long & A. D. Baddeley (Eds.), Artenrion and performance IX (pp. 313-330). Hillsdale, NJ: Erlbaum. Clark, H. H., & Carlson, T. B. (1982a). Hearers and speech acts. Language, 58, 332-373. Clark, H. H.,& Carlson, T. B. (1982b). Speech acts and hearers’ beliefs. In N. V. Smith (Ed.), Mutual knowledge (pp. 1-36). London: Academic Press. Clark, H. H., & Marshall, C. R. (1981). Definite reference and mutual knowledge. In A. K.Joshi, B. L. Webber, & 1. A. Sag (Eds.), Elements of discourse and undersianding (pp. 10-63). Cambridge, England: Cambridge University Press.

RELATIONAL REGULATION

28 1

Clark, H. H.,& Murphy, G. L. (1983). Audience design in meaning and reference. In J. F. LeNy & W. Kintsch (Eds.), Language and comprehension. Amsterdam: North-Holland Publishing. Clark, H. H., & Schaefer, E. F. (1987). Concealing one’s meaning from overhearers. Journal of Memory and Language, 26, 209-225. Cody, M. J., & McLaughlin, M. L. (Eds.). (1990). The psychology of tactical communication. Philadelphia: Multilingual Matters. Conlee, M. C., & Tesser, A. (1973). The effects of recipient desire to hear on news transmission. Sociometry, 36, 588-599. Cooper, J., & Fazio, R. H. (1984). A new look at dissonance theory. Advances in Experimental Social Psychology 17, 229-266. Cooper, J., & Jones, E. E. (1969). Opinion divergence as a strategy to avoid being miscast. Journal of Personality and Social Psychology, 13, 23-30. Cooper, J., & Scher, S. J. (in press). Actions and attitudes: The role of responsibility and aversive consequences in persuasion. In T. Brock & S. Shavitt (Eds.), The psychology ofpersuasion. San Francisco: Freeman. Cupchick, G. C., & Leventhal, H. (1974). Consistency between expressive behavior and the evaluation of humorous stimuli: The role of sex and self-observation. Journal of Personality and Social PSyChOlOgy, 30, 429-442. Dabbs, J. M., Jr., Evans, M. S., Hopper, C. H., & Purvis, J. A. (1980). Self-monitors in conversation: What do they monitor? Journal of Personality and Social Psychology, 39, 278-284. Darley, J. M. (1984). Role distance. Unpublished manuscript, Princeton University, Princeton, NJ. Darley, J. M., & Gross, P. H. (1983). A hypothesis-confirming bias in labeling effects. Journal of Personality and Social Psychology, 44, 20-33. Davis, K. E., & Jones, E. E. (1960). Changes in interpersonal perception as a means of reducing cognitive dissonance. Journal of Abnormal and Social Psychology, 61, 402-410. Delgado, J. M. R. (1969). Physical control of the mind. New York: Harper & Row. DePaulo, B. M. (1992). Nonverbal behavior and self-presentation. Psychological Bulletin, I I I , 203-243. Desmond, D., Collet, P., Marsh, P., & O’Shaughnessy, M. (1979). Gestures: Their origins and distribution. New York: Stein and Day. Dollard, J., & Miller, N. E. (1950). Personality andpsychotherapy. New York: McGraw-Hill. Douglas, W. (1 984). Initial interaction scripts: When knowing is behaving. Human Communication Research, 11, 203-219. Drake, G. F. (1978). The social role of slang. In H. Giles, W. P. Robinson, & P. M. Smith (Eds.), Language: Social psychological perspectives (pp. 63-70). New York: Pergamon Press. Dumas, B. K., & Lighter, J. (1978). Is slang a word for linguistics? American Speech, 53, 5-17. Duncan, J. W., & Laird, J. D. (1977). Cross-modality consistencies in individual differences in selfattribution. Journal of Personality, 45, 191-196. Edelmann, R. J. (1985). Social embarrassment: An analysis of the process. Journal of Social and Personal Relationships, 2 , 195-213. Edelmann, R. J. (1987). The psychology of embarrassment. New York: Wiley. Edelmann, R. J., & Hampson, S . E. (1979). Changes in nonverbal behavior during embarrassment. British Journal of Social and Clinical Psychology, 18, 385-390. Edelmann, R. J., & Hampson, S. E. (1981). Embarrassment in dyadic interaction. Social Behavior and Personality, 9, 17I- I77 Eiser, J. R., & Ross, M. (1977). Partisan language, immediacy, and attitude change. European Journal of Social Psychology, 7. 477-489. Ekman, P. (1985). Telling lies: Clues to deceit in the marketplace, politics, and marriage. New York: Norton.

282

JOHN H. FLEMING

Ekman, P., & Friesen, W. V. (1969). The repertoire of nonverbal behavior: Categories, origins, usage, and coding. Semiotica, I , 49-98. Ekman, P., & Friesen, W. V. (1972). Hand movements. Journal of Communication, 22, 353-374. Ekman, P., & Friesen, W. V. (1974). Detecting deception from the body and the face. Journal of Personality and Social Psychology, 29, 288-298. Elkind, D. (1980). Strategic interactions in early adolescence. In J. Adelson (Ed.), Handbook of adolescent psychology (Chap. 13, pp. 432-444). New York: Wiley-Interscience. Erickson, B., Lind, E. A , , Johnson, B. C., & O’Barr, W. M. (1978). Speech style and impression formation in a court setting: The effects of “powerful” and “powerless” speech. Journal of Experimental Social Psychology, 14. 266-279. Exline, R. V., Gray, D., & Schuette, D. (1965). Visual behavior in a dyad as affected by interview content and sex of respondent. Journal of Personality and Social Psychology, I , 201209. Fazio, R. H., Sanbonmatsu, D. M., Powell, M. C., & Kardes, F. R. (1986). On the automatic activation of attitudes. Journal of Personality and Social Psychology, 5 0 , 229-238. Fein, S . , Hilton, J. L., & Miller, D. T. (1990). Suspicion of ulterior motivation and correspondence bias. Journal of Personality and Social Psychology, 5 8 , 753-764. Ferguson, C. A. (1985). Special language registers. Discourse Processes, 8, 391-394. Festinger, L. (1957). A theory of copitive dissonance. Stanford, CA: Stanford University Press. Fisher, C. D. (1979). Transmission of positive and negative feedback to subordinates: A laboratory investigation. Journal of Applied Psychology, 64, 533-540. Fiske, S. T., & Taylor, S. E. (1991). Social cognition (2nd ed.). Reading, MA: Addison-Wesley. Flavell, J. H., Botkin, P. T., Fry, C. L., Wright, J. W., & Jarvis, P. E. (1968). The development of role-taking and communication skills in children. New York: Wiley. Fleming, J. H. (1993). [Distancing behaviors and correspondent inferences]. Unpublished raw data, University of Minnesota, Minneapolis, MN. Fleming, J. H., & Darley, J. M. (1989). Perceiving choice and constraint: The effects of contextual and behavioral cues on attitude attribution. Journal of Personality and Social Psychology, 56. 27-40. Fleming, I. H., & Darley, J. M. (1991). Mixed messages: The multiple-audience problem and strntegic communication. Social Cognition, 9, 25-46. Fleming, J. H . , & Darley, J. M. (in press). Actors and observers revisited: Correspondence bias, counterfactual surprise, and discounting in successive judgments of constrained behavior. Social Cognition. Fleming, J. H., Darley, J. M., Hilton, J. L., & Kojetin, B. A. (1990). Multiple audience problem: A strategic communication perspective on social perception. Journal of Personality and Social Psychology, 58, 593-609. Fleming, J. H., & Kojetin, B. A. (1993). It takes one to know one, sometimes: The eflects ofselfmonitoring similarity on tactical communication. Unpublished manuscript, University of Minnesota, Minneapolis. Fleming, J. H., & Rudman, L. A. (1993). Between a rock and a hard place: The self-concept regulating and communicative properties of distancing behaviors. Journal of Personality and Social Psychology, 64, 44-59. Fleming, J. H., & Scott, B. A. (1991). Costs of confession: An analysis of the Persian Gulf War POW videotapes in historical and theoretical perspective. Contemporary Social Psychology, 15, 127- 138. Fleming, J. H . , & Scott, B. A. (1993). [Analysis of the Persian Gulf War POW videotapes]. Unpublished raw data, University of Minnesota, Minneapolis, MN. Fleming, J. H., & Stukas, A. A,, Jr. (1991, May). Distancing and embracing cues and their effects on judgments of subsequent behavior. Paper presented at the meeting of the Midwestern Psychological Association, Chicago, IL.

RELATIONAL REGULATION

283

Folkes, V. S. (1982). Communicating the reasons for social rejection. Journal of Experimental Social Psychology, 18, 235-252. Folkman, S., & Lazarus, R. S. (1985). If it changes it must be a process: A study of emotion and coping during three stages of a college examination. Journal of Personality and Social F‘sychology, 48, 150-170. Folkman, S . , Lazarus, R. S., Dunkel-Schetter, C., DeLongis, A,, & Gruen, R. J. (1986). Dynamics of a stressful encounter: Cognitive appraisal, coping, and encounter outcomes. Journal of Personality and Social Psychology, 5 0 , 992- 1003. Frankel, A , , & Moms, W. N. (1976). Testifying in one’s own defense: The ingratiator’s dilemma. Journal of Personality and Social Psychology, 34, 475-480. Fussell, S. R., & Krauss, R. M. (1989a). The effects of intended audience on message production and comprehension: Reference in a common ground framework. Journal of Experimental Social Psychology. 25, 203-219. Fussell, S. R., & Krauss, R. M. (1989b). Understanding friends and strangers: The effects of audience design on message comprehension. European Journal of Social Psychology, 19. 509525. Fussell, S . R., & Krauss, R. M. (1991). Accuracy and bias in estimates of others’ knowledge. European Journal of Social Psychology, 2 1 , 44-454. Fussell, S. R., & Krauss, R. M. (1992). Coordination of knowledge in communication: Effects of speakers’ assumptions about what others know. Journal of Personality and Social Psychology, 62, 378-391. Garfinkel, H. (1967). Studies in ethnornethodology. Englewood Cliffs, NJ: Prentice-Hall. Garvey, J. (1990, November 25). [Letter to the editor]. New York Times, p. H4. Gergen, K. I. (1965). Interaction goals and personalistic feedback as factors affecting the presentation of self. Journal of Personality and Social Psychology, I , 413-424. Gergen, K. J. (1968). Personal consistency and the presentation of self. In C. Gordon & K. J. Gergen (Eds.), The self in social interaction (Vol. I , pp. 299-308). New York: Wiley. Gergen, K. J. (1981). The function and foibles of negotiating self-conception. In M. D. Lynch, A. A. Norem-Hebersin, & K. J. Gergen (Eds.), Self-concept: Advances in theory and research (pp. 59-73). Cambridge, MA: Ballinger. Gergen, K. J. (1982). From science to self What is there to know? In J. Suls (Ed.), Psychological perspectives on the self (pp. 129-149). Hillsdale, NJ: Erlbaum. Gilbert, T. D., & Jones, E. E. (1986). Perceiver-induced constraint: Interpretations of self-generated reality. Journal of Personality and Social Psychology, 50, 269-280. Ginzel, L. E., Jones, E. E., & Swann, W. B., Jr. (1987). How “naive” is the naive attributor? Discounting and augmentation in attitude attribution. Social Cognition, 5 , 108-130. Glucksburg, S . , Krauss, R. M., & Higgins, E. T. (1975). The development of referential communication skills. In F. Horowitz, E. Hetherington, S. Scarr-Salapaek, & G . Siege1 (Eds.), Review of child development research (Vol. 4). Chicago: University of Chicago Press. Goffman, E. (1959). The presentafion of self in everyday life. Garden City, NY: Doubleday. Goffman, E. (1961). Encounters. Indianapolis: Bobbs-Merrill. Goffman, E. (1969). Strategic interaction. Philadelphia: University of Pennsylvania Press. Goffman, E. (1974). Frame analysis. New York: Harper and Row. Gonzales, M. H., Manning, D. J., & Haugen, J. A. (1992). Explaining our sins: Factors influencing offender accounts and anticipated victim responses. Journal of Personality and Social Psychology, 62, 958-971. Gonzales, M. H., Pederson [Haugen], J. H., Manning, D. J . , & Wetter, D. W. (1990). Pardon my gaffe: Effects of sex, status, and consequence severity on accounts. Journal of Personality and Social Psychology, 58, 610-621. Greenwald, A. G . , & Breckler, S. J. (1985). To whom is the self presented? In B. Schlenker (Ed.), Self and social life (pp. 126-145). New York: McGraw-Hill.

284

JOHN H. FLEMING

Grice, H. P. (1975). Logic and conversation. In P. Cole & J. L. Morgan (Eds.), Syntax and semantics: Vol. 3. Speech acts (pp. 41-48). New York: Academic Press. Grice, H. P. (1978). Some further notes on logic and conversation. In P. Cole (Ed.), Syntax and semantics: Vol. 9. Pragmatics (pp. 113-128). New York: Academic Press. Gutek, B. A., & Bikson, T. K. (1985). Differential experiences of men and women in computerized offices. Sex Roles, 13, 123-136. Harris, J. L. (1990). Self-care actions of chronic schizophrenics associated with meeting solitude and social interaction requisites. Archives of Psychiatric Nursing, 4 , 298-307. Heider, F. (1958). The psychology of interpersonal relations. New York: Wiley. Herr, P., Sherman, S. J., & Fazio, R. H. (1984). On the consequences of priming: Assimilation and contrast effects. Journal of Experimental Social Psychology. 19. 323-340. Hess, A. K., & Gossett, D. (1974). Nixon and the media: A study of non-immediacy in newspaper editorials as reflective of geographic attitude differences. Psychological Reports, 34, 10551058.

Hewes, D. E., &Graham, M. L. (1989). Second-guessing theory: Review and extension. Communication Yearbook, 12, 213-248. Hewitt, J. P., & Stokes, R. (1975). Disclaimers. American Sociological Review, 40, 1-11. Higgins, E. T. (1977). The varying presuppositional nature of comparatives. Journal of Psycholinguistic Research, 6 , 203-222. Higgins, E. T. (1981). The “communication game”: Implications for social cognition and persuasion. In E. T. Higgins, C. P. Herman, & M. P.Zanna (Eds.), Social cognirion: The Ontario Symposium (Vol. 1, pp. 343-392). Hillsdale, NJ: Erlbaum. Higgins, E. T., & King. G. (1981). Accessibility of social constructs: Information processing consequences of individual and contextual variability. In N. Cantor & J. F. Kihlstrom (Eds.), Personality, cognition, and social interaction (pp. 69-121). Hillsdale, NJ: Erlbaum. Higgins, E. T., & Rholes, W. (1978). “Saying is believing”: Effects of message modification on memory and liking for the person described. Journal of Experimental Social Psychology, 14, 363-378.

Higgins, E. T., Rholes, W., &Jones, C. R.(1977). Category accessibility and impression formation. Journal of Experimental Social Psychology, 13, 141-154. Higginr, E. T., & Stangor, C. (1988). Context-driven social judgment and memory: When “behavior engulfs the field” in reconstructive memory. In D. Bar-Tal & A. W. Kruglanski (Eds.), The social psychology of knowledge (pp. 262-298). New York: Cambridge University Press. Hirt, E. R., Zillmann, D., Erickson, G. A., & Kennedy, C. (1991). Costs and benefits of allegiance: Changes in fans’ self-scribed competencies after team victory versus defeat. Journal of Personality and Social Psychology, 63, 724-738. Hochschild, A. R. (1983). The managed heart: Commercialization of human feeling. Berkeley, CA: University of California Press. Hohmann, G. (1966). Some effects of spinal cord lesions on experimental emotional feelings. Psychophysiology, 3, 143-156. Hollywood’s war on Indians draws to a close. (1990, October 7). New York Times, pp. 15, 18-19. Holtgraves, T. (1986). Language structure in social interaction: Perceptions of direct and indirect speech acts and interactants who use them. Journal of Personality and Social Psychology, 51, 305-314.

Hovland, C., Harvey, 0.. & Sherif, M. (1957). Assimilation and contrast effects in communication and attitude change. Journal of Abnormal and Social Psychology, 55, 244-252. Huff, C. W., Sproull, L., & Kiesler, S. (1989). Computer communication and organizational commitment: Tracing the relationship in a city government. Journal of Applied Social Psychology, 19, 1371-1391. Iancono, S.,& Kling, R. (1 984). Office routine: the automated pink collar. IEEE Spectrum, 21,73-76.

RELATIONAL REGULATION

285

Ickes, W. J., & Barnes, J. D. (1977). The role of sex and self-monitoring in unstructured dyadic interactions. Journal of Personality and Social Psychology, 35, 3 15-330. Ilgen, D. R., & Knowlton, W. A., Jr. (1980). Performance attributional effects on feedback from superiors. Organizational Behavior and Human Performance, 25, 441-456. Izard, C. E. (1972). The face of’ emotion. New York: Appleton-Century-Crofts. Izard, C. E. (1977). Human emotions. New York: Plenum. Izard, C. E. (1990). Facial expressions and the regulation of emotion. Journal of Personality and Social Psychology, 58. 487-498. James, W. (1890). Principles of psychology (Vols. 1 & 2). New York: Henry Holt. Johnson, J. T., Jemmott, J. B. 111, & Pettigrew, T. F. (1984). Causal attribution and dispositional inference: Evidence of inconsistent judgments. Journal of Experimental Social Psychology, 20, 567-585. Johnson, R., Conlee, M. C., & Tesser, A. (1974). Effects of similarity of fate on bad news transmission: A reexamination. Journal of Personality and Social Psychology. 29, 644-648. Jones, E. E. (1979). The rocky road from acts to dispositions. American Psychologist. 34, 107117.

Jones, E. E. (1986). Interpreting interpersonal behavior: The effects of expectancies. Science, 234, 41-46. Jones, E. E. (1990). Interpersonal perception. New York: Freeman. Jones, E. E., & Davis, K. E. (1965). From acts to dispositions: The attribution process in person perception. Advances in Experimental Social Psychology, 2 , 2 19-266. Jones, E. E., Davis, K. E., & Gergen, K. J. (1961). Role playing variations and their informational value for person perception. Journal of Abnormal and Social Psychology, 63, 302-310. Jones, E. E., & Harris, V. A. (1967). The attributions of attitudes. Journal ofExperimenta1 Social Psychology, 3 , 1-24. Jones, E. E., & Nisbett, R. E. (1972). The actor and the observer: Divergent perceptions of the causes of behavior. In E. E. Jones, D. E. Kanouse, H. H. Kelley, R. E. Nisbett, S. Valins, & B. Weiner (Eds.), Atrribution: Perceiving the causes of behavior (pp. 79-94). Morristown, NJ: General Learning Press. Jones, E. E., & Pittman, T. S . (1982). Toward a general theory of strategic self-presentation. In J. Suls (Ed.), Psychological perspectives on the self (Vol. 1, pp. 231-262). Hillsdale, NJ: Erlbaum. Jones, E. E., Worchel, S., Goethals, G.R., & Grumet, J. (1971). Prior expectancy and behavioral extremity as determinants of attitude attribution. Journal of Experimental Social Psychology, 7 , 59-80. Jones, E. E., & Wortman, C. B. (1973). Ingratiation: An artriburional approach. Morristown, NJ: General Learning Press. Kant, I. (1965). Critique ofpure reason (N. K. Smith, Trans., pp. 11 1-119). New York: St. Martin’s Press. (Original work published 1781) Kaplan, K. J. (1975). Structure and process in interpersonal “distancing.” Environmental Psychology and Nonverbal Behavior, I , 104- 121 . Kaplan, K. J., Firestone, I. J., Klein, K. W., & Sodikoff, C. (1983). Distancing in dyads: A comparison of four models. Social Psychology Quarterly, 46, 108-1 15. Kasl, S. V., & Mahl, G . F. (1965). The relationship of disturbances and hesitations in spontaneous speech to anxiety. Journal of Personality and Social Psychology, I . 425-433. Kellerman, J., Lewis, J., & Laird, J. D. (1989). Looking and loving: The effects of mutual gaze on feelings of romantic love. Journal of Research in Personality, 23, 145- 161. Kelley, H. H. (1972). Attribution in social interaction. In E. E. Jones, D. E. Kanouse, H. H. Kelley, R. E. Nisbett, S. Valins, & B. Weiner (Eds.), Attribution: Perceiving the causes of behavior (pp. 1-26). Momstown, NJ: General Learning Press.

286

JOHN H. FLEMING

Kelman, H. C., & Hamilton, V. L. (1989). Crimes of obedience. New Haven, C T Yale University Press. Kofika, A. (1935). The principles of Gestalt psychology. New York: Harcourt, Brace. Kohler, W. (1940). Dynamics in psychology. New York: Liveright. Krauss, R. M. (1979). Communication models and communicative behavior. In R. L. Schiefelbusch & J. H. Hollis (Eds.), Language intervention from ape to child. Baltimore: University Park Press. Krauss, R. M. (1981). Impression formation, impression management, and nonverbal behaviors. In E. T. Higgins, C. P. Herman, & M. P. Zanna (Eds.), Social cognition: The Ontario Symposium (Vol. 1, pp. 323-341). Hillsdale, NJ: Erlbaum. Krauss, R. M. (1987). The role of the listener: Addressee influences on message formulation. Journal of Language and Social Psychology, 6 , 81-97. Krauss, R. M., & Fussell, S. R. (1990). Mutual knowledge and communicative effectiveness. In J. Galegher, R. E. Kraut, & C. Egido (Eds.), Intellectual teamwork: Social and technical bases of collaborative work. Hillsdale, NJ: Erlbaum. Krauss, R. M., & Fussell, S. R. (1991). Perspective-taking in communication: Representations of others’ knowledge in reference. Social Cognition. 9, 2-24. Krauss, R. M., Garlock, C. M., Bricker, P. D., & McMahon, L. E. (1977). The role of audible and visual back-channel responses in interpersonal communication. Journal of Personality and Social Psychology, 35, 523-529. Krauss, R. M., Morrel-Samuels, P., & Colasante, C. (1991). Do conversational hand gestures communicate? Journal of Personality and Social Psychology, 61 743-754. Krauss, R. M., Vivekananathan, P. S., & Weinheimer, S. (1968). “Inner speech and “external speech: Characteristics and communication effectiveness of socially and nonsocially encoded messages. Journal of Personality and Social Psychology, 35, 523-529. Krauss, R. M., & Weinheimer, S. (1964). Changes in the length of reference phrases as a function of social interaction: A preliminary study. Psychonomic Science, 1 , 113-1 14. Krauss, R. M . , & Weinheimer, S . (1966). Concurrent feedback, confirmation, and the encoding of referents in verbal communication. Journal of Personality and Social Psychology, 4, 342346. Kraut, R. E. (1978). Verbal and nonverbal cues in the perception of lying. Journal of Personality and Social Psychology, 36, 380-391. Kraut, R. E. (1982). Social presence, facial feedback, and emotion. Journal of Personality and Social Psychology, 42, 853-863. Kraut, R. E., & Higgins, E. T. (1984). Communication and social cognition. In R. S. Wyer, Jr., & T. K. Srull (Eds.), Handbook of social cognition (Vol. 3, pp. 87-127). Hillsdale, NJ: Erlbaum. Kraut, R. E., & Lewis, S. H. (1982). Feedback and the coordination of conversation. In H. Sypher & J. Applegate (Eds.), Cognition and communication. Hillsdale, NJ: Erlbaum. Laird, J. D. (1974). Self-attribution of emotion: The effects of expressive behavior on the quality of emotional experience. Journal of Personality and Social Psychology, 29, 475-486. Laird, J. D. (1984). The real role of facial response in the experience of emotion: A reply to Tourangeau and Ellsworth, and others. Journal of Personality and Social Psychology, 47, 909917. Lanzetta, J. T., Cartwright-Smith, J., & Kleck, R. E. (1976). Effects of nonverbal dissimulation on emotional experience and autonomic arousal. Journal of Personality and Social Psychology, 33, 354-370. Lazarus, R. S. (1993). From psychological stress to the emotions: A history of changing outlooks. Annual Review of Psychology, 44, 1-21. Leary, M. R. (1982). Social anxiety. In L. Wheeler (Ed.), Review ofpersonality and socialpsychology (Vol. 3, pp. 97-120). Beverly Hills, CA: Sage. ~

RELATIONAL REGULATION

287

Leary, M. R. (1983). Understanding social anxiety. Beverly Hills, CA: Sage. Leary, M. R., & Kowolski, R. M. (1990). Impression management: A literature review and twocomponent model. Psychological Bulletin, 107, 34-47. Lerner, M. J., & Agar, E. (1972). The consequences of perceived similarity: Attraction and rejection, approach and avoidance. Journal of Experimental Research in Personality. 6 , 69-75. Lewin, K. (1935). A dynamic theory of personality. New York: McGraw-Hill. Lewin, K. ( 1938). The conceptual representation and measurement of psychological forces. Contributions to Psychological Theory, I , (4, Serial No. 4). Lewis, D. K. (1969). Convention. Cambridge, MA: Harvard University Press. Lott, B. (1987). Sexist discrimination as distancing behavior: I. A laboratory study. Psychology of Women Quarterly, I I , 47-58. Lott, B. (1989). Sexist discrimination as distancing behavior: 11. Primetime television. Psychology of Women Quarterly, 13, 341-355. Mahl, G. F. (1959). Measuring the patient’s anxiety during interviews from “expressive” aspects of his speech. Transactions of the New York Academy of Sciences, 21, 249-257. Manis, M., Cornell, S. D., & Moore, J. C. (1974). Transmission of attitude-relevant information through a communication chain. Journal of Personality and Social Psychology, 30, 81-94. Markov, P. 0. (1950). A study of interoception in human subjects. Uchenye Zapiski Leningradskogo Gosudarstvennogo Universiteta. Seriye Biologicheskikh Nauk, 22. 345-368. Marks, S. R. (1977). Multiple roles and role strain: Some notes on human energy, time, and commitment. American Sociological Review, 42, 921-936. Markus, H. (1977). Self-schemata and processing information about the self. Journal of Personality and Social Psychology, 35, 63-78. Markus-Kaplan, M., & Kaplan, K. J. (1984). A bidimensional view of distancing: Reciprocity versus compensation, intimacy versus social control. Journal of Nonverbal Behavior, 8, 3 15-326. Martin, L. L. (1986). Set/Reset: Use and disuse of concepts in impression formation. Journal of Personality and Social Psychology, 51, 493-504. Martin, L. L., Harlow, T. F., & Strack, F. (1992). The role of one’s own bodily sensations in the evaluation of social events. Personality and Social Psychology Bulletin, 18, 412-4 19. Masten, A. S., Best, K. M., & Garmezy, N. (1990). Resilience and development: Contributions from the study of children who overcome adversity. Development and Psychopathology, 2, 425444. Masten, A. S., & O’Connor, M. I. (1989). Vulnerability, stress, and resilience in the early development of a high risk child. Journal of the American Academy of Child and Adolescent Psychiatry, 28, 274-278. McGinley, H. LeFevre, R., & McGinley, P. (1975). The influence of a communicator’s body position on opinion change in others. Journal of Personality and Social Psychology, 31, 686-690. McGraw, K. M. (1987). Guilt following transgression: An attribution of responsibility approach. Journal of Personality and Social Psychology, 53, 247-256. Mehrabian, A. (1968a). Inference of attitudes from the posture, orientation, and distance of a communicator. Journal of Consulting and Clinical Psychology, 32, 296-398. Mehrabian, A. (196813). Relationship of attitude to seated posture, orientation, and distance. Journal of Personality and Social Psychology, 10, 26-30. Mehrabian, A. (1969). Some referents and measures of nonverbal behavior. Behavior Research Methods and Instrumentation, I 203-207. Mehrabian, A. (1972). Nonverbal communication. Chicago: Aldine-Atherton. Mehrabian, A., & Friar, J. T. (1969). Encoding of attitude by a seated communicator via posture and position cues. Journal of Consulting and Clinical Psychology, 33, 330-336. Mehrabian, A., & Williams, M. (1969). Nonverbal concomitants of intended and perceived persuasiveness. Journal of Personality and Social Psychology, 13, 37-58, I

288

JOHN H. FLEMING

Meltzer, B. N., Petras, J. W., &Reynolds, L. T. (1975). Symbolic interactionism: Genesis, varieties, and criticism. Boston: Routledge & Kegan Paul. Messinger, S. E., Sampson, H., & Towne, R. D. (1962). Life as theater: Some notes on the dramaturgic approach to social reality. Sociometry, 25, 98- 110. Milgram, S. (1963). Behavioral study of obedience. Journal of Abnormal Psychology, 67, 371-378. Milgram, S . (1974). Obedience to authority. New York: Harper & Row. Miller, A. G., Baer, R.,& Schonberg, P. (1979). The bias phenomenon in attitude attribution: Actor and observer perspectives. Journal of Personality and Social Psychology, 37, 1421-143 I . Miller, A. G., Jones, E. E., & Hinkle, S. (1981). A robust attribution error in the personality domain. Journal of Experimental Social Psychology, 17, 587-600. Miller, A. G., & Rorer, L. C. (1982). Toward an understanding of the fundamental attribution error: Essay diagnosticity in the attitude-attribution paradigm. Journal of Research in Personality, 16, 41-59. Modigliani, A. (1968). Embarrassment and embarrassability. Sociometry, 31, 313-326. Morrel-Samuels, P., & Krauss, R. M. (1992). Word familiarity predicts the temporal asynchrony of hand gestures and speech. Journal of Experimental Psychology: Learning, Memory, and Cognition, 18, 615-622. National Research Council. (1987). Computer chips and paper clips: Technology and women’s employment (Vol. 1). Washington, DC: National Academy Press. Newtson, D., & Czerlinsky, T. (1974). Adjustment of attitude communications for contrasts by extreme audiences. Journal of Personality and Social Psychology, 30, 829-837. Nisbett, R. E., & Ross, L. (1980). Human inference: Strategies and shortcomings of social judgment. Englewood Cliffs, NJ: Prentice-Hall. Notarius, C. I., & Levenson, R. W. (1979). Expressive tendencies and physiological response to stress. Journal of Personality and Social Psychology, 3 , 1204-1210. Novak, D. W., & Lerner, M. J. (1968). Rejection as a consequence of perceived similarity. Journal of Personality and Social Psychology, 9, 147- 152. Nunberg, G. (1978). Slang, usage-conditions, and l’arbitraire du signe. In D. Farkas, W. M. Jacobsen, & K. W. Todrys (Eds.), Papers fmm the parasession on the lexicon (pp. 301-31 1). Chicago: University of Chicago Press. Patterson, M. L. (1978). Arousal change and cognitive labeling: Pursuing the mediators of intimacy exchange. Environmental Psychology and Nonverbal Behavior, 3 , 17-22. Patterson, M. L. (1987). Presentational and affect-management functions of nonverbal involvement. Journal of Nonverbal Behavior, 11, 110-122. Peak, H. (1958). Psychological structure and person perception. In R. Tagiuri & L. Petrullo (Eds.), Person perception and interpersonal behavior (pp. 337-352). Stanford, CA: Stanford University Press. Pearlman, S. F. (1989). Distancing and connectedness: Impact on couple formation in lesbian relationships. Women and Therapy, 8. 77-88. Ponse, B. (1978). Identities in the lesbian world: The social construction of self(Chap. 3, pp. 5788). Westport, CT: Greenwood Press. Pyszczynski, T. A., & Greenberg, J. (1981). Role of disconfinned expectancies in the instigation of attributional processing. Journal of Personality and Social Psychology, 40, 3 1-38. Pyszczynski, T., Greenberg, J., Solomon, S., Sideris, J., & Stubing, M. J. (1993). Emotional expression and the reduction of motivated cognitive bias: Evidence from cognitive dissonance and distancing from victims’ paradigms. Journal of Personality and Social Psychology, 64, 177186. Quattrone, G. A. (1982). Behavioral consequences of attributional bias. Social Cognition. 1 , 358378. Ranelli, C. J., & Miller, R. E. (1981). Behavioral predictors of amitriptyline response in depression. American Journal of Psychiatry, 138, 30-34.

RELATIONAL REGULATION

289

Razran, G. H. S. (1938). Conditioning away social bias by the luncheon technique. Psychological Bulletin. 35, 693. Razran, G. H. S. (1940). Conditioned response changes in rating and appraising sociopolitical slogans. Psychological Bulletin, 37, 481. Rhodewalt, F. T. (1986). Self-presentation and the phenomenal self On the stability and malleability of self-conceptions. In R. F. Baumeister (Ed.), Public selfandprivate self(pp. 117-142). New York: Springer-Verlag. Rhodewalt, F. T., & Comer, R. (1979). Induced-compliance attitude change: Once more with feeling. Journal of Experimental Social Psychology, 15, 35-47. Riskind, J. H., & Gotay, C. C. (1982). Physical posture: Could it have regulatory or feedback effects on motivation and emotion? Motivation and Emotion, 6 , 273-298. Rommetveit, R. (1974). On message structure: A framework for the study of language and communication. New York: Wiley. Rosen, S., & Tesser, A. (1970). On reluctance to communicate undesirable information: The mum effect. Sociometry, 33, 253-263. Rosen, S . , & Tesser, A. (1972). Fear of negative evaluation and the reluctance to transmit bad news. Journal of Communication, 22, 124- 141. Rosenfeld, H. M. ( 1966a). Approval-seeking and approval-inducing functions of verbal and nonverbal responses in dyads. Journal of Personality and Social Psychology, 4 , 597-605. Rosenfeld, H. M. (1966b). Instrumental affiliative functions of facial and gestural expressions. Journal of Personality and Social Psychology, 4 , 65-72. Ross, L. (1977). The intuitive psychologist and his shortcomings. Advances in Experimental Social Psychology, 10, 173-200. Ross, L. Amabile, T. M., & Steinmetz, J. L. (1977). Social roles, social control, and biases in socialperception processes. Journal of Personality and Social Psychology, 35, 485-494. Sacks, H . , Schegloff, E., &Jefferson, G. (1974). A simplest systematics for the organization of tumtaking in conversation. Language, 50, 696-735. Schachter, S . (1964). The interaction of cognitive and physiological determinants of emotional state. Advances in Experimental Social Psychology, I 49-80. Scheier, M. F., & Carver, C. S. (1988). A model of behavioral self-regulation: Translating intention into action. Advances in Experimental Social Psychology, 21, 303-346. Schein, E. H . (1956). The Chinese indoctrination program for prisoners of war: A study of attempted “brainwashing.” Psychiatry, 19, 149-172. Schein, E. H., Schneider, I., & Barker, C. H . (1961). Coercive persuasion-A socio-psychological analysis of the “brainwashing” of American civilian prisoners by Chinese Communists. New York: Norton. Schemmer, B. F. (1988). The secret war of the POWs. In S. Lipsman (Ed.), The Vietnam experience: War in the shadows (pp. 150-165). Boston: Boston Publishing. Scher, S. J., & Cooper. J. (1989). Motivational basis of dissonance: The singular role of behavioral consequences. Journal of Personality and Social Psychology, 56, 899-906. Schiffer, S. (1972). Meaning. Oxford: Clarendon Press. Schlenker, B. R., & Weigold, M. F. (1992). Interpersonal processes involving impression regulation and management. Annual Review of Psychology, 43. 133-168. Schneider, K. J . , & Miller, R. S. (1975). The effects of enthusiasm and quality of arguments on attitude attribution. Journal of Personality, 43, 693-708. Schwartz, N., & Clore, G. (1983). Mood, misattribution, and judgments of well-being: Informative and directive functions of affective states. Journal of Personality and Social Psychology, 45, 5 13-523. Schwartz, N., & Hippler, H. J. (1987). What response scales may tell your respondents: Informative functions of response alternatives. In H.J. Hippler, N. Schwartz, & S. Sudman (Eds.), social information processing and survey methodology (pp. 163- 178). New York: Springer-Verlag. I

290

JOHN H. FLEMING

Schwartz, N., & Strack, F. (Eds.) (1991). Social cognition and communication: Human judgment in its social context. Social Cognition, 9, 1-125. Schwartz, N., Strack, F., Kommer, D., & Wagner, D. (in press). Soccer, rooms, and the quality of your life: Further evidence on informative functions of affective states. European Journal of Social Psychology. Schwartz, N., & Wyer, R. S. (1985). Effects of rank ordering stimuli on magnitude ratings of these and other stimuli. Journal of Experimental Social Psychology, 21, 30-46. Scott, M. B., & Lyman, S . M. (1968). Accounts. American Sociological Review, 33, 46-62. Searle, J. R . (1975). Indirect speech acts. In P. Cole & J. L. Morgan (Eds.), Synrux and semantics: Vol. 3. Speech acts (pp. 59-82). New York: Academic Press. Segal, H. A. (1954). Initial psychiatric findings of recently repatriated prisoners of war. American Journal of Psychiarry, 61, 358-363. Selman, R . (1971). The relation of role-taking to the development of moral judgment in children. Child Developmenr, 42, 79-91. Sherman, S. I., Ahlm, K., Berman, L., & Lynn, S. (1978). Contrasts effects and their relationship to subsequent behavior. Journal of Experimental Social Psychology, 14, 340-350. Showers, C., & Ruben, C. (1990). Distinguishing defensive pessimism from depression: Negative expectations and positive copying mechanisms. Cognitive Therapy andResearch, 14, 385-399. Siskind, G. (1966). Mine eyes have seen a host of angels. American Psychologisr, 21, 804-806. Slane, S . , & Leak,G . (1978). Effects of self-perceived nonverbal immediacy behaviors on interpersonal attraction. Journal of Psychology, 98, 241-248. Sledd, J. (1965). On not teaching English usage. English Journal, 54, 698-703. Snyder, C. R. (1985). The excuse: An amazing grace? In B. Schlenker (Ed.), Selfand social life (pp. 235-260). New York: McGraw-Hill. Snyder, C. R., & Higgins, R. L. (1988). Excuses: Their effective role in the negotiation of reality. Psychological Bullerin, 104, 23-35. Snyder, C. R., Higgins, R. L., & Stucky, R. J. (1983). Excuses: Masquerades in search of grace. New York: Wiley-Interscience. Snyder, C. R., Lassegard, M., & Ford, C. E. (1986). Distancing after group success and failure: Basking in reflected glory and cutting off reflected failure. Journal of Personality and Social Psychology, 51, 382-388. Snyder, M. (1987). Public appearancesIPrivare realities: The psychology of self-monitoring. New York: W. H. Freeman. Snyder, M., & Gangestad, S. (1982). Choosing social situations: Two investigations of selfmonitoring processes. Journal of Personality and Social Psychology, 43, 123- 135. Snyder, M., Gangestad, S . , & Simpson, I. A. (1983). Choosing friends as activity partners: The role of self-monitoring. Journal of Personality and Social Psychology, 45, 1061- 1072. Snyder, M., & Kendzierski, D. (1982). Acting on one’s attitudes: Procedures for linking attitude and behavior. Journal of Experimenral Social Psychology, 18, 165- 183. Snyder, M. L., & Jones, E. E. (1974). Attitude attribution when behavior is constrained. Journal of Experimenral Social Psychology, 10, 585-600. Snyder, M. L., & Wicklund, R. A. (1981). Attribute ambiguity. In J. A. Harvey, W. J. Ickes, & R. F. Kidd (Eds.), New directions in arrriburion research (Vol. 3, pp, 197-221). Hillsdale, NJ: Erlbaum. Stebbins, R. A. (1969), Role distance, role distance behavior, and jazz musicians. British Journal of Sociology, 20, 406-414. Stepper, S . , & Strack, F. (1993). Proprioceptive determinants of emotional and non-emotional feelings. Journal of Personality and Social Psychology, 64, 21 1-220. Stevens, M. (1989). Coping strategies of hospitalized adolescents. CHC, 18, 163-169. Stokes, R., & Hewitt, J. P. (1976). Aligning actions. American Sociological Review, 41, 838-849.

RELATIONAL REGULATION

29 1

Strack, F., Martin, L. L., & Stepper, S. (1988). Inhibiting and facilitating conditions of the human smile: A nonobtrusive test of the facial feedback hypothesis. Journal of Personaliry and Social Psychology, 54# 768-777. Strack, F., Schwartz, N., & Gschneidinger, E. (1985). Happiness and reminiscing: The role of time perspective, affect, and mode of thinking. Journal of Personality and Social Psychology, 47, 1460- 1469. Stryker, S . (1980). Symbolic interactionism: A social structural version. Menlo Park, CA: Benjamin/Cummings. Stryker, S., & Statham, A. (1985). Symbolic interaction and role theory. In G. Lindzey & E. Aronson (Eds.), The handbook of socialpsychology (3rd ed., Vol. 1, pp. 31 1-378). New York: McGraw-Hill. Tagiuri, R. (1952). Relational analysis: An extension of sociometric method with emphasis on social perception. Sociometry, 15, 91-104. Taped statements assail war. (1991, January 21). Minneapolis Star-Tribune, pp. IA, 16A. Taylor, S. E., & Mettee, D. R. (1971). When similarity breeds contempt. Journal offersonaliry and Social Psychology, 20, 75-81. Tedeschi, J. T., & Norman, N. (1985). Social power, self-presentation and the self. In B. Schlenker (Ed.), Self and sociul life (pp. 293-322). New York: McGraw-Hill. Tesser, A., & Rosen, S . (1975). The reluctance to transmit bad news. Advances in Experimental Social Psychology, 8, 193-232. Tesser, A., Rosen, S . , & Batchelor, T. (1972). On the reluctance to communicate bad news (the mum effect): A role play extension. Journal of Personaliry and Social Psychology, 40, 83-103. Tesser, A., Rosen, S., & Tesser, M. (1971). On the reluctance to communicate undesirable messages (the mum effect): A field study. Psychological Reports, 29, 651-654. Tourangeau, R., & Rasinski, K. A. (1988). Cognitive processes underlying context effects in attitude measurement. Psychological Bulletin, 103, 299-314. Trope, Y. (1986). Identification and inference processes in dispositional attribution. Psychological Review, 93, 239-257. Wagner, U., Wicklund, R. A,, & Shaigan, S . (1990). Open devaluation and rejection of a fellow student: The impact of threat to a self-definition. Basic and Applied Social Psychology, 1I , 6176. Walster, E., Aronson, E., & Abraham, D. (1966). On increasing the persuasiveness of a low prestige communicator. Journal of Experimental Social Psychology, 2 , 325-342. Wertheimer, M. (1923). Untersuchingen zur Lehre von der Gestalt, I1 [Investigations of teaching from Gestalt, 111. Psychologische Forschung, 4 , 301-350. Wicklund, R. A. (1981). How society uses self-awareness. In J. Suls (Ed.), Psychological perspectives on the self (pp. 209-230). Hillsdale, NJ: Erlbaum. Wicklund, R. A , , & Brehm, J. W. (1976). Perspectives on cognitive dissonance. Hillsdale, NJ: Erlbaum. Wiener, M., & Mehrabian, A. (1968). Language within language. New York: Appleton-CenturyCrofts. Word, C. O., Zanna, M. P., & Cooper, J. (1974). The nonverbal mediation of self-fulfilling prophecies in interracial interaction. Journal of Experimental Social Psychology, 10, 109- 120. Wyer, R. S., & Srull, T. K. (1981). Category accessibility: Some theoretical and empirical issues concerning the processing of social information. In E. T. Higgins, C. P. Herman, & M. P. Zanna (Eds.), Social cognirion: The Ontario Symposium (Vol. 1, pp. 161-197). Hillsdale, NJ: Erlbaum . Zajonc, R. B . , Adelmann, P., Murphy, S . T., & Niedenthal, P. M. (1987). Convergence in the physical appearance of spouses. Motivation and Emotion, 11, 335-346. Zajonc, R. B., & Markus, H. (1983). Affect and cognition: The hard interface. In C. E. Izard,

292

JOHN H. FLEMING

J. Kagan, & R. B. Zajonc (Eds.), Emotions, cognition, and behavior. New York: Cambridge University Press. Zajonc, R. B., Pietromonaco, P., & Bargh, J. (1982). Independence and interaction of affect and cognition. In M. S . Clark & S. T.Fiske (Eds.), Affect and cognition: The 17th Annual Carnegie Symposium on Cognition (pp. 21 1-228). Hillsdale, NJ: Erlbaum. Zanna, M. P., & Cooper, J. (1974). Dissonance and the pill: An attribution approach to studying the arousal properties of dissonance. Journal of Personality and Social Psychology, 29, 703-709, Zuckerman, M., DeFrank, R. S., Hall, J. A., Lanance, D. T., & Rosenthal, R. (1979). Facial and vocal cues of deception and honesty. Journal of Experimental Social Psychology, 15, 378-396. Zuckerman, M., Kernis, M. R., Driver, R., & Koestner, R. (1984). Segmentation of behavior: Effects of actual deception and expected deception. Journal of Personality and Social Psychology. 46, 1173-1182.

FROM SOCIAL INEQUALITY TO PERSONAL ENTITLEMENT: THE ROLE OF SOCIAL COMPARISONS, LEGITIMACY APPRAISALS, AND GROUP MEMBERSHIP Brenda Major

I. Introduction Inequality among social groups is characteristic of virtually all societies. In the United States, despite the creed that “all men are created equal,” groups vary widely in the outcomes, treatment, and access to social rewards that they experience. Social, economic, and political discrimination against women, AfricanAmericans, Native-Americans, and Latino- and Latina-Americans, for example, is widely documented (cf. Crosby, Bromley, & Saxe, 1980; Hoiberg, 1982; Jones, 1986; U.S. Government, 1978). Over the past few decades, social psychologists have intensively examined the processes of stereotyping, prejudice, and discrimination directed against stigmatized or oppressed groups (cf. Dovidio & Gaertner, 1986). To a lesser extent, social psychologists also have examined the reactions of those who are the targets of oppression and discrimination. One of the most robust findings of this latter research is that the objective conditions of peoples’ lives often bear only minimal relation to their subjective satisfaction with those conditions, or with their subjective well-being in general (see Andrews & Withey, 1976; Campbell, Converse, & Rodgers, 1976; Crocker & Major, 1989; Crosby, 1976; Diener, 1984; and Pettigrew, 1967, for reviews). Often, people who are objectively disadvantaged report themselves to be just as content and satisfied with their circumstances as are people who are objectively more privileged (e.g., Martin, 1986a; Stouffer, Suchman, DeVinney, Star and Williams, 1949). Sometimes, however, they do not. The major premise of this chapter is that beliefs about entitlement are a critical determinant of how members of social groups react affectively, evaluatively, and 293 ADVANCES IN EXPERIMENTAL SOCIAL PSYCHOICGY. VOI. 26

Copyright 0 1994 by Academic Press. Inc. All rights of reproduction in any form reseived.

294

BRENDA MAJOR

behaviorally to their socially distributed outcomes (i.e., outcomes in which another person or social system is involved). It is argued that one of the most far reaching consequences of discrimination and social inequality is that they can alter what people feel they deserve, or are entitled to receive, from their social relationships. The disadvantaged often come to believe they deserve their lesser outcomes, whereas the overprivileged often come to believe that they are entitled to their position of relative advantage. This argument is not new. Concepts of entitlement and deserving figure centrally in theories of social justice and relative deprivation, and have frequently been invoked to explain the reactions of the disadvantaged (e.g., Cook, Crosby, & Hennigan, 1977; Crosby, 1976; 1982; Davis, 1959; Deutsch, 1985; Gurr, 1970; Martin, 1981; 1986a). Despite the prominence of entitlement and deserving in theories of social justice, however, there have been few attempts to organize what is known about the antecedents and consequences of beliefs about entitlement. Furthermore, attempts to examine empirically how beliefs about entitlement develop among individuals who are members of chronically disadvantaged groups, or how beliefs about entitlement affect subsequent affect, evaluations, or behavior are relatively rare. These topics are the focus of this article. This essay highlights the role that social comparison processes and attributions of responsibility play in translating social inequality into beliefs about personal and collective entitlement. The first half of this article addresses how social comparison processes and attributions contribute to the development of a lesser sense of personal entitlement among members of objectively disadvantaged groups. It is argued that social comparison biases tend to prevent awareness of disadvantage, and attribution biases tend to legitimize disadvantage. As a result, what “is” has a marked tendency to become what “ought” to be. These processes are illustrated through a program of research on the origins of gender differences in personal entitlement to pay. Gender differences in entitlement are proposed to underlie the finding that women and men typically do not differ in their life, job, or marital satisfaction, despite situations at work and at home that are, by objective criteria, disadvantageous for women compared to the situations of men. The second half of this article considers why members of other disadvantaged groups (e.g., African-Americans) do express discontent with their objectively unjust situations. It is argued that situational and personal factors that prompt people to compare with advantaged outgroups and that lead them to question the legitimacy of outcome distributions result in elevated entitlement among the disadvantaged and correspondingly higher levels of discontent (Folger, 1987; Gurin, 1985).

ENTITLEMENT AND SOCIAL INEQUALITY

295

11. Social Groups and Reactions to Social Inequality: The Cases of Gender and Race Women and African-Americans' are two demographic groups whose relative disadvantage in American society has been widely documented, both in terms of economic opportunities and outcomes and in terms of interpersonal outcomes (cf. Crocker & Major, 1989; Crosby, Bromley, & Saxe, 1980; Jones, 1986 for reviews). A substantial amount of research has examined stereotypes about women and blacks, and has documented the existence of prejudice and discrimination against these groups. This is not the concern of the present article. Rather, it is concerned with how members of these groups evaluate their circumstances. A brief review of this literature is presented here.

A. GENDER AND THE TOLERATION OF INEQUALITY The job and pay satisfaction of working women provides an illustrative example of paradoxical contentment in the presence of objective disadvantage. It is well documented that full-time working women in the United States are concentrated in lower paying, lower prestige jobs than are men, and are paid less than similarly qualified men doing comparable work (cf. Marini, 1989; Treiman & Hartmann, 1981; Wittig & Lowe, 1989). Evidence of disparities in earnings between working women and working men is abundant and consistent (e.g., Boulgarides, 1984; Crosby, 1982; Dreher, Dougherty, & Whitely, 1989; Frieze, Olson, & Good, 1990; Godley, Hafer, Riggar, & Maki, 1983; Sigelman, Milward, & Shepard, 1982). This difference in pay between men and women holds even when men and women are matched on numerous job characteristics that typically covary with sex, such as age, education, organizational tenure, tenure in present position, and occupational prestige. Common sense suggests that women should feel angry, resentful, and dissatisfied as a result of receiving fewer valued rewards than do similar men doing similar work. Most field studies of job and pay satisfaction, however, find little evidence that women are subjectively more dissatisfied than are men with their pay or with their jobs (e.g., Crosby, 1982; Dreher, 1981; Sauser & York, 1978). Crosby (1982), for example, surveyed 345 male and female full-time workers, matched for occupational prestige, on a number of dimensions relevant to satis'Because the terms blacks and African-Americans are both used as a way of self-identification among members of this ethnic group, these terms will be used interchangeably in this article.

296

BRENDA MAJOR

faction at work. Consistent with earlier research, she found that the women in her sample were underpaid relative to men with jobs of similar prestige. Paradoxically, however, she found that these women reported no evidence of personal deprivation or dissatisfaction with their own pay, jobs, or treatment, although they were aware of and aggrieved about the underpayment of women in general. In other words, although these women realized that other women were being discriminated against, they did not believe that this inequitable situation applied to themselves as well. Crosby dubbed this pattern the “paradox of the contented female worker” (see Crosby, 1982; Major, 1987 for a review of this literature). Inequality of experience because of gender is not limited to the economic and public domains but extends even into the private domain of the family. Research on the division of the unpaid labor of the family consistently indicates that compared to their husbands, wives do a disproportionate share of the housework and child care, even in families where both spouses work full time (see Spitze, 1988; Thompson & Walker, 1989 for reviews). Furthermore, husbands’ lesser contribution to family work is relatively unaffected by whether or not their wives work, the number of hours their wives spend in paid work, or wives’ earnings, either personal or relative to their husbands’. Rather, gender, more than any other factor, accounts for how family work is allocated (e.g., Berk, 1985; Biernat & Wortman, 1991; Pleck, 1985; see reviews by Spitze, 1988; Thompson & Walker, 1989). Scholars have coined terms such as the second shift to describe this pattern (e.g., A. Hochschild, 1989). Recognition of the extra burden that family responsibilitiesplace on working women relative to working men is illustrated by the highly publicized suggestion that there be a “mommy track” for corporate women (Schwartz, 1989). Given this situation, one might think that women would express feelings of relative deprivation about the division of labor in the home. This does not seem to be the case for most women. In general, women are not more dissatisfied with their relationships than men are, and most studies indicate that although both spouses recognize the inequality in the division of labor in their marriages, they do not perceive it as unfair (see Ferree, 1990; Major, 1993; Spitze, 1988; Thompson, 1991 for reviews). Rather, the majority of women seem satisfied with the smaller amounts of housework their husbands do (relative to themselves). For example, in a recent study of couples of relatively equal economic and professional status, Biernat and Wortman (1991) found that wives contributed a greater share of child care and had more responsibility for the performance of household tasks than did their husbands. Nevertheless, 62% of the wives felt that their husbands did a satisfactory amount of child care and household chores, and 14% actually indicated that their husbands did too much. Thus, despite the unequal distribution of family work between husbands and wives, most studies suggest that there is not a great deal of conflict over the division of labor within families. Studies of women’s and men’s subjective well-being and perceived quality of

ENTITLEMENT AND SOCIAL INEQUALITY

297

life based on national probability samples echo the findings from these two research areas. Little difference in global happiness or satisfaction is usually found between the sexes (e.g., Andrews & Withey, 1976; Campbell et al., 1976; see Diener, 1984 for a review). Thus, despite women’s disadvantaged status in economic, political, and even family domains relative to men, there is little evidence that women in general are more dissatisfied with their situations than are men.

B. RACE AND REACTIONS TO INEQUALITY Consider, in contrast, the life circumstances and subjective well-being of African-Americans. African-Americans as a group arguably experience more systematic and severe disadvantage than any other group in American society. Compared to whites, blacks are disadvantaged in health, education, housing, jobs, income, and positions of political and economic power (cf. Jones, 1986). Research documenting the nature of racial attitudes toward black Americans, and the extent of blacks’ disadvantage compared to whites is abundant (e.g., Dovidio & Gaertner, 1986; Schuman, Steeh, & Bobo, 1985; U.S. Government, 1978). Research examining how blacks compare to whites in their satisfaction with their outcomes or subjective well-being, in contrast, is relatively rare. Most available studies are based on national probability surveys conducted prior to the 1980s. Furthermore, as several authors have pointed out (e.g., Diener, 1984; J. S. Jackson, Chatters, & Neighbors, 1986), the samples of blacks in these studies have been small and of questionable representativeness. In general, these studies indicate that blacks typically (although not consistently) score lower than whites on measures of subjective well-being and satisfaction (see Diener, 1984; J. S. Jackson et al., 1986 for reviews). Campbell et al. (1976), for example, found that compared to the 2,146 whites in their sample, the 223 blacks they surveyed were no more likely to report feelings of stress, but were substantially less willing to call themselves “very happy,” were less satisfied with their own lives as a whole, were slightly less positive on an index of general affect, were less self-confident that they could control their own lives, and were substantially less trusting of people. Blacks were also less satisfied than whites in all 15 different life domains assessed, and especially so in those domains in which blacks are most obviously disadvantaged compared to whites (standard of living, savings, housing, amount of education, and neighborhood). Statistically equating blacks to whites on six control measures of social, economic, residential, and age status did not wholly remove these race differences in satisfaction. Thus, unlike women as a group, black Americans as a group appear aware of, and aggrieved about, their disadvantaged status in American life.

298

BRENDA MAJOR

111. Toward a Psychology of Entitlement

A number of different explanations have been proposed for why members of some objectively disadvantaged groups are satisfied and accepting of their situations whereas others are not. A review of this vast literature is beyond the scope of this paper.* The current paper focuses on one precondition to a sense of discontent or grievance: the psychological experience of entitlement. It is argued that beliefs about entitlement are key psychological mediators of how people react affectively, evaluatively, and behaviorally to their socially distributed outcomes. In the following section, a brief review of how the concepts of entitlement and deserving have been used and defined in the literature is first presented. Three antecedents to beliefs about entitlement are then discussed: comparisons with reference standards, legitimacy appraisals, and goals.

A. ENTITLEMENT: AN OVERVIEW

Entitlement and deservingness are central constructs in most psychological theories of social justice. (Most justice scholars use these terms interchangeab l ~ . Equity )~ theory (e.g., Adams, 1965; Walster, Walster, & Berscheid, 1978) and relative-deprivation theory (cf. Crosby, 1976; Davis, 1959; Gurr, 1970; Martin, 1981), for example, share the assumption that people feel unjustly treated when they do not receive valued outcomes that they feel legitimately entitled to receive (see Crosby, 1982, for a review). According to Lerner (1987), “the experience of entitlement is the essential psychological ingredient of an entire family of human events associated with social justice” (p. 108). Attempts to define these terms or to organize what is known about their antecedents and consequences are, however, rare. Lay definitions of entitlement and deservingness equate these terms with having a right to something. Webster’s New World Dictionary (1970), for example, defines entitled as “having a right or legal title to something,” and deserving as 2The reader is referred to Cook,Crosby, and Hennigan (1977), Crosby (1976; 1982; 1984), Martin (1981), and Taylor and Moghaddam (1987) for excellent reviews of some of this literature. 3Steil (in press) has observed that in common usage, the term entitlemenr is more frequently used to imply having a right to something on the basis of an ascribed characteristic (a quality), whereas the term deserving is more frequently used to imply having a right to something on the basis of an achieved characteristic (an act). For example, individuals are said to be entitled to various forms of government assistance, to vote, to public education, and so forth, simply by virtue of their citizenship or humanity. In contrast, people are said to be deserving of something because of what they have done (e.g.. a hard day’s work). Because the psychological implications of deserving and entitlement are similar, however, I will follow the tradition of most justice scholars and use these terms interchangeably in this paper.

ENTITLEMENT AND SOCIAL INEQUALITY

299

“having a right to (something) because of acts or qualities.” The sociologist Singer (198 1) defines entitlement as “an expectation with normative force,” (p. 88); that is, an expectation that one should or ought to receive something. According to Lerner (1987), the sense of deservingness or entitlement is a cognitive judgment with affective and motivational implications, and refers to the relationship between a person and his or her outcomes. The cognitive component of the sense of entitlement is “the judgment, often tacit, that someone, or some category of people, is entitled to a particular set of outcomes by virtue of who they are or what they have done” (p. 108). Thus, if an individual fulfills certain preconditions, he or she will feel entitled to certain outcomes. These preconditions may be either ascribed (e.g., she is of a particular race or sex) or earned (e.g., he has made particular contributions). Lerner (1987) further states that entitlement is “experienced affectively and motivationally as an imperative, a sense of requiredness between the actor’s perceived outcomes and the person’s attributes or acts” (pp. 107-108). This sense of a moral imperative is what sets entitlement apart from related concepts like wants and expectations. It also produces a unique set of emotional and behavioral consequences relative to related concepts. People may feel sad if they don’t get what they want, and disappointed or frustrated if they don’t get what they expect. If their sense of entitlement is violated, however, they will experience a sharper emotion, anger, which produces a more consistent motivation to seek change (e.g., Shaver, Schwartz, Kirson & O’Connor, 1987).

B. ANTECEDENTS TO ENTITLEMENT A comprehensive psychology of entitlement has proved elusive, in part as a result of the complexity of the concept and the breadth of domains and resources to which it can be applied. Much of the social-psychological research on entitlement has examined the impact of distributive justice rules, and to a lesser extent, perceptions of procedural justice, on beliefs about entitlement. Other factors also influence beliefs about entitlement, including the presence of laws, commitments, and contracts (cf. Furby, 1986), early experienced relationships between preconditions and outcomes (Lerner, 1980; 1987), and bargaining power (Deutsch, 1985). Three antecedents to beliefs about entitlement are discussed here: comparisons with reference standards, legitimacy appraisals, and goals (see Deutsch, 1985, for an excellent overview). 1 . Comparisons with Reference Standards Comparison processes are critical determinants of beliefs about entitlement (e.g., Adams, 1965; Crosby, 1982; Lerner, Miller, & Holmes, 1976; Messick &

300

BRENDA MAJOR

Sentis, 1983; Singer, 1981). Most justice theories assert that beliefs about entitlement are based on comparisons of outcomes (and in some theories, inputs) with some standard of reference. These standards may include laws, societal rules or customs, the outcomes received by another person or group, or outcomes received by the self (cf. Folger, 1987; Furby, 1986; Levine & Moreland, 1987). Of these various comparison standards, most justice theories accord a central role to those that are socially derived (Adams, 1965; Lerner et al., 1976; Levine & Moreland, 1987; Pritchard, 1969; Walster et al., 1978). Many philosophers and justice scholars have argued that there is no absolute justice, only justice relative to others (see Shepelak & Alwin, 1986). Different justice rules and preconditions for determining entitlement exist within societies, making a precise specification of deserving difficult (Lerner et al., 1976). The importance of social comparison processes in arriving at judgments of personal entitlement is due in part to the considerable ambiguity that exists as to what a person deserves. In social situations where ambiguity exists, people turn to others for a stable referent to determine what they deserve or are entitled to receive (see Festinger, 1954; Pettigrew, 1967). Social comparisons (with others or self) provide several important types of information relevant to judgments of entitlement. First, they provide information about what types of outcomes and/or treatments it is possible to achieve (e.g., salary, promotion, respect). Hence they affect wants, aspirations, and judgments of what is feasible for oneself. Second, they provide information as to the relevant preconditions for obtaining those outcomes or treatment (e.g., research, teaching, service in academia). Third, and most important, social comparisons provide information about the relationship, or covariation, between preconditions and outcomes or treatment. They provide information, for example, on what the relationship should be between academic contributions such as research, teaching, or service and outcomes such as promotion or tenure. They specify what people who have met certain preconditions can legitimately expect to receive in terms of outcomes or treatment. Because people make inferences about inputs on the basis of outcomes (e.g., Lerner et al., 1976), social comparisons also provide information about what people who have received certain outcomes are likely to have contributed (i.e., how well they have met the preconditions for obtaining those outcomes). Despite the acknowledged centrality of social comparison processes in judgments of entitlement, surprisingly little empirical research has directly addressed how social comparison processes affect judgments of outcome entitlement. Most of the theoretical and empirical work on social comparison has involved assessments of abilities, opinions, and emotions (Festinger, 1954; Suls & Miller, 1977; Suls & Wills, 1991). As a result, knowledge of how people use social comparison to estimate their entitlements and to evaluate their outcomes is incomplete. According to Adams’ (1965) and Walster et al.’s (1978) versions of equity

ENTITLEMENT AND SOCIAL INEQUALITY

30 1

theory, it is the specific, “local” comparison others who are salient in the immediate environment who most strongly affect beliefs about entitlement. According to the status-value formulation of equity theory (e.g., Berger, Zelditch, Anderson, & Cohen, 1972) and reference-group theory (e.g., Kelley, 1952), it is comparisons with reference groups, or abstracted “generalized others,” that determines entitlement. Berger et al. (I972), for example, speculate that entitlement is based on a “referential reward structure” that consists of stereotypic information about the typical relation between levels of characteristics (e.g., inputs, contributions) possessed and levels of rewards (e.g., pay, status) received by a generalized reference other (e.g., the starting salaries generally received by new PhDs in social psychology). These stereotyped conceptions about what is typically the case are assumed to become the basis for defining what ought to be the case. Yet other justice theorists emphasize the importance of intrapersonal comparisons, such as comparisons of present outcomes with past outcomes, as determinants of beliefs about entitlement (Goodman, 1974; Pritchard, 1969; Runciman, 1966). This view is consistent with the proposition that people consider internal standards derived from their past experience when making social judgments (Festinger, 1954; Thibaut & Kelley, 1959). In an attempt to systematize knowledge on outcome comparisons, Levine and Moreland (1987) identified three major dimensions of the outcome-comparison process. The first dimension, type of comparison, refers to the identities of the source and target of comparison. Three types were identified: (1) self/self, in which the outcomes of a single individual are compared; (2) selfother, in which the outcomes of different individuals are compared; and (3) grouplgroup, in which the outcomes of one or more groups are compared. It is useful to add a fourth type to their taxonomy, selfgroup, in which the outcomes of a single individual are compared with those of a reference group or “generalized other” (cf. Berger et al., 1972). The second dimension refers to the group identification of the source and the target of comparison. In intragroup comparisons, the source and the target are identified with the same group, whereas in intergroup comparisons, the source and target are identified with two different groups. The third dimension refers to the time period or periods during which the two outcomes under consideration occur. These may be either intratemporal (occurring during the same time period) or intertemporal (occumng during different time periods). Thus, this taxonomy suggests 16 different kinds of comparison standards that may be used to evaluate outcomes. Which of these comparison standards are most important to the development of beliefs about personal entitlement? The inability of justice theories to answer this question is often cited as a major weakness of these theories (cf. Austin, 1977; Pettigrew, 1967). Recent research on social comparison processes, however, suggests that three major factors are likely to affect the choice of comparison targets: (1) structural factors in the environment, such as proximity and

302

BRENDA MAJOR

salience, that affect which comparison targets are readily available for comparison purposes; (2) the perceived similarity (broadly defined) of comparison targets to the self; and (3) the goals or motivations of the person who is making the comparison (Levine & Moreland, 1987; Wood, 1989). Proximity or salience of comparison targets in the immediate environment affects whom people compare with when estimating what they deserve (Austin, 1977; Levine & Moreland, 1987; Major, 1987; 1989; Singer, 1981). People in natural settings report making comparisons with others who are physically close or readily accessible, such as family, friends, or co-workers (Patchen, 1961; Singer, 198 1). Laboratory research has demonstrated the powerful influence that proximal-comparison others have on fairness and entitlement judgments (e.g., Austin, McGinn & Susmilch, 1980). Because of segregation and affiliation preferences, intragroup others, or others like ourselves, tend to be more proximal in the immediate environment, and hence most readily accessible for comparison (see Crocker & Major, 1989). Consequently, both selflother and self/group comparisons are likely to be biased toward intragroup comparisons. Perceived similarity is also an important determinant of whom people compare with when estimating what they, themselves, deserve (cf. Adams, 1965; Austin, 1977; Lemer et al., 1976; Pettigrew, 1967). In general, people tend to compare with others like themselves (Festinger, 1954). This bias to compare with similar, or intragroup, others has been demonstrated in a large number of studies on social comparisons of abilities and outcomes (see Crocker & Major, 1989; Wood, 1989). Sociologists working in the tradition of reference-group theory (e.g., Hyman & Singer, 1968; Merton, 1957; Runciman, 1966; Stouffer et al., 1949), for example, find that people spontaneously report comparing their living standards and social status with others whose class or role situation is similar to their own (see Singer, 198 1 , for a review). Experiments have demonstrated that people prefer to compare their abilities and performances with others who are similar to themselves in attitudes, values, or personality traits (see Suls & Miller, 1977), who are similar to themselves on attributes perceived to be related to the dimension under evaluation (Goethals & Darley, 1977), or who are similar to themselves even on dimensions unrelated to the dimensions under evaluation (see Wood, 1989 for a review). Even trivial similarities in group membership between self and comparison others can influence social comparisons (Major, Schiacchitano, & Crocker, in press). These findings underscore the degree to which basic social categorization processes underlie comparison processes and the development of entitlement beliefs (Atkinson, 1986). Once groups, or social categories, are psychologically created, a number of processes promote the perception of greater homogeneity within groups than between groups (cf. Tajfel & Turner, 1986). It follows that a favorite target of comparison should be people (either specific individuals or

ENTITLEMENT AND SOCIAL INEQUALITY

303

groups) we view as in the same category as ourselves, because they are likely to be seen as most similar to us. Finally, goals also affect comparison choices. When the comparer’s goal is to estimate accurately his or her entitlement, others who are similar to the self are most likely to be chosen as comparison targets (Wood, 1989). When external comparison others are unavailable, not salient in the environment, or are regarded as too dissimilar to self, estimates of entitlement are likely to be made on the basis of intrapersonal (self/self) comparisons. The self is a readily available and highly similar comparison target. Consequently, reference standards derived from intrapersonal comparisons with one’s own past experiences in similar domains are likely to be perceived as highly diagnostic of personal deserving (cf. Austin et al., 1980; Desmarais, 1993). Self/group comparisons, or comparisons of self with abstract stereotypes of what similar generalized others typically obtain in similar circumstances, will also affect estimates of personal deserving when specific others are not available for comparison (Berger et al., 1972; Major, McFarlin, & Gagnon, 1984). In summary, people tend to make intragroup rather than intergroup comparisons when estimating what they deserve because of the greater availability, and assumed greater similarity and diagnosticity, of the former. People also estimate their deserving on the basis of intrapersonal comparisons, especially when external-comparison others are unavailable or perceived as too dissimilar to self. As a result, people typically feel they deserve the same treatment or outcomes that they have received in the past or that others like themselves receive. 2 . Legitimacy Appraisals Appraisals of the legitimacy of observed relationships between preconditions (e.g., attributes, qualities, acts, needs) and outcomes also are a crucial factor in the development of beliefs about entitlement. Outcome comparisons occur in the context of distributions that are appraised as legitimate or illegitimate. Perceptions of legitimacy are what give entitlements the experience of “ought” and what set entitlement apart from wants and expectations. Comparisons with others are likely to affect expectations irrespective of the appraised legitimacy of the context. They are likely to translate into beliefs about personal entitlements, however, only if a distribution is appraised as legitimate (Martin, 1986b). People must believe, for example, that their own or others’ outcomes are legitimate before they will infer that they deserve the same outcomes or treatment that they received in the past, or that others like themselves have received (Bylsma, Major, & Cozzarelli, 1992). Furthermore, as long as people appraise discrepancies between their outcomes and those of their comparison standard to be legitimate, they will not feel entitled to the same outcomes or treatment as the compar-

304

BRENDA MAJOR

ison standard and will not experience anger as a result of not obtaining the same (e.g., Della Fave, 1980; Folger, 1987; Gurin & Townsend, 1986; Mark & Folger, 1984). Perceptions of legitimacy may be derived in several ways.4 These include: the belief that an observed distribution of outcomes conforms to an accepted rule of distributive justice, such as equity, equality, or need (see Deutsch, 1985); the belief that it resulted from fair procedures (see Tyler & Lind, 1992); and the attribution that individuals or groups are personally responsible for their level of outcomes (see Gurin, 1985). Each of these is briefly discussed here. a. Distributive Justice Norms. Appraisals of legitimacy are affected by the extent to which an outcome distribution is believed to conform to an accepted rule of distributive justice, and by the extent to which that rule of distributive justice is assumed to be appropriate in a given context. All societies have rules that dictate what people should do to obtain certain outcomes, and that specify a set of preconditions that ought to lead to these outcomes (Lerner et al., 1976). These rules, however, are often vague, contradictory, and under dispute. For example, in the United States there is a strong endorsement of the idea that all citizens are entitled to certain basic rights guaranteed by the Constitution (e.g., freedom of speech, freedom of religion, equal opportunity). Americans also endorse humanitarian principles, such as the principle that all people are entitled to at least some minimum standard of living. Most strongly of all, however, Americans endorse the belief that individuals who have made greater contributions, or who have higher investments, merit (deserve) more desirable outcomes (e.g., J. L. Hochschild, 1981; Kluegel & Smith, 1986; Sampson, 1975). The existence of differing justice rules provides the opportunity for disputes over which rule is most legitimate in what circumstance. When the equity rule is operative, those who contribute more to an exchange relative to others are believed to be entitled to proportionately higher outcomes from that exchange. When the need rule of justice is operative, those who need more than others are believed to be entitled to receive more. When the equality rule of justice is applied, equality of outcomes, irrespective of contributions or need is believed to be legitimate (e.g., Deutsch, 1985; Sampson, 1975). Research has illustrated that the perceived legitimacy of differing distributive justice rules is highly dependent on contextual factors, such as the nature of the relationship between parties and the type of outcome being allocated (e.g., Clark & Mills, 1979; Deutsch, 1985; Leventhal, 1976). J. L. Hochschild (1981), for example, observes that Americans tend to endorse the principle of equality in political and socializing (e.g., family) domains, but equity in economic domains. 4A full discussion of the determinants of legitimacy is beyond the scope of this article. The reader is referred to Kluegel and Smith (1986) for a discussion of the legitimacy of income inequality, and to Qler and Lind (1992) for a discussion of the legitimacy of authorities.

ENTITLEMENT AND SOCIAL INEQUALITY

305

Americans tend to endorse equality in the distribution of opportunity, but equity in the distribution of outcomes (Kluegel & Smith, 1986). Individuals and cultures also differ in the extent to which they endorse different justice rules. Men, for example, tend to give greater weight to contributions, and less weight to need than women do when allocating rewards between themselves and others (Major & Deaux, 1982; Major, Romero, CastBno, Barbera, Martinez-Benlloch, & Pastore, 1993). Countries that have a strong individualistic and meritocratic ideological bias, such as the United States, tend to emphasize the legitimacy of contributions as a precondition to entitlement, and endorse the equity rule of justice, more than do non-Western countries (Kluegel & Smith, 1986; Sampson, 1975). b. Procedural Justice Norms. A second factor that affects appraisals of the legitimacy of outcome distributions is the extent to which the distribution is perceived to have resulted from fair procedures. A growing body of research suggests that procedural justice considerations are an important determinant of the perceived legitimacy of distributions and of authorities (Lind & Tyler, 1988; Qler & Lind, 1992). Perceptions of procedural justice are affected by several factors. One is the degree of control a person believes he or she has over the distribution: Greater perceived control leads to greater perceived procedural fairness (Leventhal, 1976). Closely related to this factor is the finding that people who have “voice,” (i.e., who have an opportunity to state their case) regard procedures as fairer than do people who do not have a voice (Lind & Tyler, 1988). Perceptions of procedural fairness are also affected by the extent to which the procedures used in distributing outcomes are seen as being biased, prejudiced, applied inconsistently, or based on inadequate information (see Leventhal, 1976; Lind & Tyler, 1988). c . Attributions of Responsibility. Attributions about why outcome differences exist among individuals or groups are among the most important, and least studied, determinants of appraisals of legitimacy and entitlement (Cohen, 1982; Gurin, 1985; Gurin & Townsend, 1986). Although none of the major psychological theories of justice make attributions a central and explicit concern, a number of justice scholars have noted the close association between attributions of responsibility and feelings of entitlement (e.g., Cook et al., 1977; Homans, 1974; Lerner et al., 1976; Taylor & Moghaddam, 1987). People who take personal responsibility for not having something they want are less likely to feel entitled to it (Homans, 1974; Cook et al., 1977). Those who do not feel responsible for lacking something they want, in contrast, are more likely to feel entitled to it and more likely to direct anger at external agents thought to be responsible for depriving them of it (see Smith & Ellsworth, 1985). Disparities among people or groups that are attributed to factors internal to those individuals or groups (e.g., differences in abilities, qualifications, personal choices, needs) are more likely to be perceived as legitimate than are disparities that are seen to be externally

306

BRENDA MAJOR

caused (Gurin & Townsend, 1986; Miller, Gurin, Gurin, & Malanchunk, 1981). This link between internal attribution and entitlement is particularly likely to be true in contexts where equity is the endorsed rule of distributive justice, or among individuals who strongly endorse the equity principle (see Cohen, 1982). 3 . Goals

Goals or motives also affect beliefs about entitlement. Although several justice scholars have made this observation (e.g., Deutsch, 1985; Levine & Moreland, 1987; Reis, 1987), few researchers have examined it empirically. Most research has been guided by the assumption that an individual’s goal is an accurate estimation of his or her entitlement. Research on social comparison processes indicates that others who are similar to the self are most likely to be chosen as comparison targets under these circumstances (see Wood, 1989, for a review). Beliefs about entitlement are also influenced by other motives, such as the desire to protect self-esteem from threat (self-protection) and the desire to improve the outcomes of self or group (self-improvement). When the comparer’s goal is self-esteem protection, people are more likely to choose to compare with others who are worse off than themselves (Wills, 1981; Wood, 1989). This strategy allows comparers to protect themselves from the affective distress engendered by knowing that others are receiving more than they are (i.e., protects them from experiencing relative deprivation or inequity distress) (Brickman & Bulman, 1977; Crocker & Major, 1989; Major, 1989). Wills (1981) suggests that social comparisons motivated by self-protective concerns are particularly likely to occur when the comparer is in a situation that involves “frustration or misfortune . . . that is difficult to remedy through instrumental action” (p. 145). This is a situation in which many disadvantaged and stigmatized groups find themselves with regard to socially distributed outcomes (see Crocker & Major, 1989). By comparing with others who are not doing so well, however, the comparer may not only feel good but may also come to feel entitled to less herself. The desire to protect self-esteem may also affect appraisals of the legitimacy of outcome discrepancies between self and others. The desire to protect self-esteem from threat, for example, may lead an individual to attribute poor outcomes to a biased external agent or procedurally unjust system rather than to acts or qualities of himself or herself. Although this external attribution may buffer self-esteem from threat, however, it may be psychologically costly in other ways. This strategy requires acknowledging oneself as a victim of injustice; it necessitates abandoning the illusions that the world is a distributively or procedurally just place and that one can personally control one’s own outcomes. Attributions of poor outcomes to external agents may also lessen a sense of personal responsibility for effecting change, even when such change is possible (Crocker & Major, 1993; Major & Crocker, 1993). When people desire to enhance their outcomes and feel there is a possibility of

ENTITLEMENT AND SOCIAL INEQUALITY

307

doing so, they may be motivated to make social comparisons that are unfavorable to themselves or their group (Levine & Moreland, 1987; Major, Testa, & Bylsma, 1991; Testa & Major, 1990). Such upward comparisons, if they can be legitimated, may then act as a lever for change (Patchen, 1961; Singer, 1981). Self-improvement goals may also increase the likelihood that members of disadvantaged groups will attribute observed,outcome disparities between themselves and others to illegitimate rather than legitimate factors. Claims that one is personally a victim of injustice or that one’s group is collectively a victim of injustice can, under some circumstances, carry considerable moral weight in negotiating for better treatment or more favorable outcomes. Messick and Sentis (1983) and Reis (1987) note that in disputes over scarce resources, contending parties often present their own claims and preferences as fair, deserved, and fully justified. In this way entitlement may be used strategically to further personal or collective desires.

IV. The Impact of Social Inequality on Personal Entitlement: How Disadvantage Becomes Deserved Beliefs about entitlement, and the processes that lead to these beliefs, do not occur within a social vacuum, but within social contexts that often reward people differently on the basis of the social groups to which they belong. In this section, the impact of group-based social inequality on beliefs about personal entitlement is discussed. It is argued that social comparison processes and attributions of responsibility for outcomes play a central role in translating group-based social inequality into group differences in beliefs about personal entitlement. In particular, it is argued that tendencies to make intragroup and intrapersonal social comparisons result in members of disadvantaged groups frequently being unaware of their personal disadvantage. Furthermore, it is argued that because outcome disparities between self and advantaged outgroup members often are attributed to internal causes or causes under personal control, even when members of disadvantaged groups do become aware of their disadvantaged status, they often appraise it as legitimate. As a result, the disadvantaged often come to believe that they are personally entitled to less than do members of more advantaged groups. A. COMPARISON PROCESSES IN CONTEXTS OF GROUP-BASED INEQUALITY One of the most important psychological consequences of social inequality is that it produces group-based differences in the comparison standards, or standards of reference, that are derived from social comparison processes. This

308

BRENDA MAJOR

occurs in two primary ways (cf. Major, 1987; 1989). First, as noted previously, people tend to make intragroup rather than intergroup comparisons when estimating what they deserve, in part because of the greater availability and diagnosticity of the former. As a result, people typically feel they deserve the same outcomes or treatment that others like themselves have received. If members of one’s own group have been systematically deprived in society, a tendency to make intragroup comparisons will result in lower standards of reference than if members of one’s own group have enjoyed a privileged status in society. Second, as discussed previously, people also estimate their deserving on the basis of intrapersonal comparisons, especially when external comparison others are unavailable or perceived as too dissimilar to self. Consequently, people also typically feel entitled to the same treatment or outcomes that they have received in the past. Members of less privileged groups are likely to have personally experienced less positive outcomes in response to their actions or qualities than are members of more privileged groups. As a result, intrapersonal comparisons will result in different standards of reference for members of underprivileged versus overprivileged groups. Specifically, if individuals have been personally deprived or underrewarded in a given domain in the past, comparisons of current circumstances with own past circumstances will lead to lower reference standards, however, than if individuals have been personally privileged or overrewarded in the past. A general model of how social comparison processes contribute to group differences in personal entitlement within a context of (legitimated) social inequality is illustrated in the upper path of Fig. 1. First, structural inequalities exist between groups in access to and receipt of socially distributed rewards (Box A). For example, members of some social categories (men) typically receive higher rewards (income) for their acts or qualities (work, education) than do members of other social categories (women). These structural inequalities affect the outcomes of members of the ingroup as well as the self. Second, social, cognitive, and motivational factors exist that enhance the tendency to socially compare oneself and one’s outcomes with the self or others who are similar to the self, and hence with “ingroup” members (Box B). Third, because groups differ in their outcomes, this tendency to compare with ingroup members or self gives rise to group differences in reference standards (Box C). Fourth, these reference standards translate into beliefs about personal entitlement (Box F). The model presented in Fig. 1 describes how social comparison processes influence beliefs about entitlement when the procedures and distributive-justice rules underlying an outcome distribution are appraised as legitimate. Why outcome distributions tend to be appraised as legitimate is discussed in Section IV, B. How comparison processes operate when outcome distributions are appraised as illegitimate, and the factors that lead to perceptions of illegitimacy, will be discussed later in the paper.

ENTITLEMENT AND SOCIAL INEQUALITY B

309

C

~

lntragroup

Group

B. LEGITIMIZING PROCESSES IN CONTEXTS OF GROUP-BASED INEQUALITY A second important consequence of social inequality is that it is often appraised as legitimate. One of the more intriguing phenomena of social justice is that people tend to legitimate the status quo, even when it is disadvantageous to the self. Homans (1974, p. 250), for example, observed that “what people say ought to be is determined in the long run and with some lag by what they find in fact to be the case.” Heider (1958, p. 235) observed that “the ‘is’ takes on the character of the ‘ought.”’ And Berger et al. (1972, p. 139) wrote, “As a consequence of beliefs about what is typically the case, expectations . . . come to be formed about what one can legitimately claim ought to be the case.” Several widely held beliefs contribute to this tendency to legitimize the status quo by structuring attributions about the causes of successful and unsuccessful outcomes (cf. Deutsch, 1985; Kluegel & Smith, 1986). These include the belief in a just world (Lerner et al., 1976), the belief in personal causation (e.g., Nisbett & Ross, 1980), and the belief in personal control (e.g., Langer, 1977). Each of these beliefs tends to legitimize the observed relations between preconditions and outcomes by locating the cause of a person’s outcomes within that person. Lerner and his colleagues (e.g., Lerner, 1975; Lerner & Miller, 1978; Lerner et al., 1976), for example, have argued that people are motivated to believe in a

3 10

BRENDA MAJOR

just world. As a result of this belief, people infer that those who are advantaged (e.g., who have more rewards than others, or who are treated better than others) must be more deserving (e.g., they are smarter, more qualified, nicer). Likewise, those who are disadvantaged must be less deserving. Lerner and Miller (1978) note that the need to believe in a just world also affects self-evaluations. That is, victims often blame themselves for their misfortune or disadvantage (see Bulman & Wortman, 1977; Comer & Laird, 1975). A similar outcome results from the widespread tendency (at least among North Americans) to locate the cause of behaviors (or outcomes) within individuals rather than within situations (or systems) (Nisbett & Ross, 1980). As a result of this bias to make internal attributions, individuals are seen as causal agents in their fates, and hence as deserving of their fate. This bias toward making internal attributions (as well as self-serving biases) contributes to those who are advantaged taking credit for their positive outcomes. It also contributes to self-blame rather than system-blame among the disadvantaged. The bias toward internality, not surprisingly, is stronger for positive than for negative outcomes. Nevertheless, national surveys suggest that even among the most disadvantaged in society, the majority do not blame their situations on external causes such as society or government (J. L. Hochschild, 1981; Kluegel & Smith, 1986). People also need to feel that they are in control of their own outcomes, and often report greater control than they objectively have (e.g., Alloy & Abramson, 1979). The more people believe that the relationship between their actions and their outcomes is personally controllable, the more likely they are to believe that their outcomes (positive or negative) are deserved (Crocker & Major, 1993). Allocation procedures that give participants “voice,” and hence that provide them with greater feelings of personal control over outcome distributions, are rated as more fair than those that do not give voice (cf. Lind & Qler, 1988). North American college students of both sexes believe that selection procedures based on individually controllable factors (e.g., grades, test scores) are more fair and legitimate than are selection procedures based on uncontrollable factors (e.g., group membership) (Major, Feinstein, & Crocker, in press; Nacoste, 1990; Taylor & Dube, 1986). For example, men and women who are selected or rejected for a desirable leadership position on the basis of their good (or poor) scores on a pretest of leadership ability rate the selection procedure as more fair and legitimate than do men and women who are selected or rejected solely on the basis of their sex (Major et al., in press). Although beliefs in personal causation, personal control, and a just world are held individually, they gain their power to legitimize social and personal inequalities through their collective endorsement within a culture. Each of these beliefs is widely prevalent among North Americans, including those who are economically disadvantaged (cf. J. L. Hochschild, 1981; Kluegel & Smith, 1986; Nisbett & Ross, 1980). Furthermore, Americans value these beliefs (Sears & Funk,

ENTITLEMENT AND SOCIAL INEQUALITY

31 1

1991). Participants in experiments like people who believe that behavior is caused by internal factors more than they do people who believe that behavior is caused by external factors, and claim to believe in internal control more than the “average person” does (Jellison & Green, 1981). These beliefs contribute to a dominant ideology among Americans (Feagin, 1975; Kluegel & Smith, 1986; Nisbett & Ross, 1980). The dominant ideology directly endorses the notion that people deserve what they get. With regard to income inequality, Kluegel and Smith (1986) observe that this dominant ideology includes the beliefs that (1) the opportunity for economic advancement is widespread in America today; (2) individuals are personally responsible for their positions in American society, and (3) the overall system of inequality in America is, therefore, equitable and fair. As a consequence of the collective nature of these beliefs, the legitimacy of observed distributions is often never questioned, even when it is disadvantageous to the self. Some authors (e.g., Kluegel & Smith, 1986) locate the source of these legitimizing beliefs in basic psychological processes and cultural socialization practices that have a strong individualistic bias. Others locate the source of these beliefs in attempts to manage existential terror (e.g., Greenberg, Pyszczynski, & Solomon, 1987) or to maintain positive self-esteem and psychological well-being (e.g., Taylor & Brown, 1988). Yet others suggest that the origins of these beliefs reside in the strategic attempts of those in power to maintain their position of advantage (e.g., Cohen, 1986). Sidanius and Pratto (1993), for example, argue that beliefs such as belief in a meritocracy are essentially legitimizing myths promoted by the hegemonic group. Legitimizing myths often make reference to the lower inputs of oppressed groups, and by so doing, locate the cause of victims’ misfortune within themselves (cf. Sidanius & Pratto, 1993). The argument that blacks were simply incapable of attending to their own affairs was an attempt to legitimize slavery (and indeed make it benevolent). Likewise, the beliefs that women are less capable than men of being leaders, are too emotional to hold jobs of great responsibility, or are too invested in their families to devote themselves completely to work, locate the cause of women’s inferior status within women and hence legitimize denying women positions of power. Biases toward internal attributions and personal control are highly adaptive in some areas of life, such as coping with accidents, illnesses, or traumas, and are argued by some to be hallmarks of good mental health (cf. Taylor & Brown, 1988). These biases can also, however, lead to acceptance of injustice and exploitation of the disadvantaged. Members of disadvantaged groups who endorse these legitimizing myths (i.e., who internalize stereotypes of their group) may devalue their own or their group’s acts or qualities in domains in which they are disadvantaged. As a result, even when they do compare their own situations with members of an advantaged outgroup, they may nonetheless be content

312

BRENDA MAJOR

because they believe they deserve their lower outcomes. Likewise, members of privileged groups may believe they deserve their greater outcomes because they come to overvalue their inputs. This process of inferring relative inputs from relative outcomes is especially likely to occur when equity is endorsed as the legitimate rule of distributive justice. According to equity theory, the lower an individual’s personal inputs (e.g., qualifications, contributions, performance) relative to others, the fewer positive outcomes that individual is entitled to receive relative to others. If group membership (e.g., sex, race) is believed to be associated with attributes that somehow detract from one’s contributions or “inputs,” then it can appropriately influence one’s outcomes, as long as equity is the prevailing rule of distributive justice that is endorsed.5 The bottom path of Fig. 1 illustrates this process. Inequality in the distribution of social outcomes between groups (Box A) is assumed to be accompanied by legitimizing beliefs and attributions (Box D), some of which refer to the lower inputs of the disadvantaged compared to the advantaged (Box E). This belief in lower “inputs,” in turn, is assumed to lead to lesser evaluations of personal entitlement among the disadvantaged (Box F). As illustrated by the bidirectional arrow between Boxes A and D, this model also assumes that legitimizing attributions and myths can contribute to social inequality, for example by justifying discrimination against the disadvantaged group.

C. THE INTERACTION AMONG COMPARISONS AND ATTRIBUTIONS

Outcome comparisons and legitimizing attributions, the topics of the last two sections (IV, A and IV, B), are reciprocal processes (cf. Cohen, 1982; Gurin, 1985; Homans, 1974; Patchen, 1961). This is illustrated by the bidirectional arrow between Boxes B and D of Fig. 1. Appraisals of legitimacy affect social comparison processes. Likewise, social comparisons affect appraisals of the legitimacy of distributions. When outcome differences among individuals or groups are appraised as legitimate (e.g.. attributed to internal causes), people are more likely to estimate their personal deserving on the basis of intragroup than intergroup comparisons. When outcome differences are appraised as illegitimate, in contrast, people are more likely to make intergroup comparisons and to estimate their personal deserving on the basis of these standards (cf. Bylsma et al., 1992; Gurin & Townsend, 1986). Patchen (1961), for example, noted that people who are disadvantaged and who blame themselves for their lesser outcomes 5A similar process may occur if need is believed to be the operative rule of distributive justice. People may infer that those who have higher outcomes had greater needs.

ENTITLEMENT AND SOCIAL INEQUALITY

313

avoid comparisons with others who are better off, whereas people who are disadvantaged and who reject personal responsibility choose to compare with advantaged others. These latter comparisons, he argued, lead to righteous indignation and justify claims to larger outcomes. Intragroup comparisons shield members of disadvantaged groups from awareness of their relative disadvantage. As a result, they help to maintain a belief that the distribution is legitimate. Intergroup comparisons with members of advantaged groups, in contrast, make the disadvantaged aware of their lower outcomes. Consequently, intergroup comparisons are apt to increase the extent to which the legitimacy of distributions is questioned.

D. SUMMARY As a result of tendencies to make intragroup and intrapersonal social comparisons, members of disadvantaged groups often are unaware of their personal disadvantage. Furthermore, because outcome disparities between themselves and disadvantaged outgroup members tend to be attributed to internal causes or causes under personal control, members of disadvantaged groups often appraise these disparities as legitimate. Consequently, the disadvantaged often come to believe they are personally entitled to less than do members of more advantaged groups.

V. From Social Disadvantage to Personal Deserving: The Case of Gender This section applies the theoretical framework that is outlined above and illustrated in Fig. 1 to understanding the “paradoxical contentment” of working women with their jobs and pay. It presents a program of research on the antecedents and consequences of gender differences in perceived entitlement to pay. This research empirically verifies the role that intragroup and intrapersonal social comparison processes and legitimizing attributions play in producing a lesser sense of entitlement among members of an objectively disadvantaged group (working women), and shows how a lesser sense of entitlement can account for women’s paradoxical satisfaction with their pay.6 This section starts by docu%ender differences in values have also been proposed as an explanation for the “paradoxical contentment” of working women with their pay. According to this perspective, women are contented with their jobs and pay despite inequitable treatment because there is little discrepancy between what women want and what they receive (cf. Crosby, 1982; Sauser & York, 1978). This perspective tends

314

BRENDA MAJOR

menting the existence of gender inequalities in the paid work force and demonstrating the existence of gender differences in perceived entitlement to pay. It then illustrates how social comparison processes and legitimizing attributions translate inequality into entitlement.

A. STRUCTURAL INEQUALITY AND DISCRIMINATION AGAINST WOMEN As applied to gender, the model presented in Fig. 1 begins with the assumption that there are systematic differences in how men and women are treated and rewarded in the domain of paid work (Box A). A substantial amount of research, some of which was cited earlier in this paper, supports this assumption. Genderbased structural inequalities in the work force provide the context in which intragroup and intrapersonal social comparison processes and legitimizing attributions lead to gender differences in entitlement to pay. Four aspects of men’s and women’s experiences in the work force are particularly influential determinants of gender differences in pay entitlement. The first is the pervasive and pronounced gender segregation of work. A number of studies have demonstrated that the labor market is profoundly sex-segregated both across and within occupations and organizations (Bielby & Baron, 1984). Indeed, across societies, although the nature of the work may differ, most tasks and work are sharply divided into “women’s work” and “men’s work” (Eagly, 1987). As a consequence, most men and women tend to work and affiliate with co-workers who are the same sex as themselves, and to compare with these coworkers. A second structural inequality that gives rise to gender differences in entitlement in the work domain is the undervaluing of women’s work. Again, almost irrespective of the job’s content, if a job is regarded as women’s work, it is viewed as being less difficult and as worth less monetarily than if the job is regarded as men’s work (e.g., McArthur & Obrant, 1986). In the United States, for example, the more a job is dominated by women, the less it pays (Treiman & Hartmann, 1981). Longitudinal studies have confirmed this pattern: Jobs, such as bank teller and secretary, that have increasingly come to be performed by womto minimize how structural factors, such as the nature of women’s and men’s jobs and opportunities, shape women’s and men’s goals and values. It also suggests greater cross-situational consistency in gender-linked behaviors than is typically observed. The research presented here, for example, illustrates that women’s and men’s sense of entitlement to pay and satisfaction with pay is strongly determined by proximal comparison standards and attributions (e.g., Bylsma & Major, 1992b). Although gender differences in values are likely to exist (cf. Eagly, 1987; Gilligan, 1982; Markus & Oyserman, 1989), they have yet to be established as a mediator of gender differences in outcome evaluations (see Major, 1987 for a discussion).

ENTITLEMENT AND SOCIAL INEQUALITY

315

en, have shown corresponding drops in their pay scale (Treiman & Hartmann, 1981). A third structural inequality that affects women’s and men’s entitlement is the underpayment of women. Even within the same jobs, there is substantial evidence that women workers are often paid less than men workers (Treiman & Hartmann, 198 1). A fourth factor that affects women’s and men’s entitlement in the work domain is gender inequality in job opportunities and home responsibilities. Numerous studies have demonstrated that women face more barriers to fulfilling their occupational aspirations than do men, in part because of sex discrimination in employment opportunities (cf. Blau, 1984), in part because women are more likely to accommodate their own work to their husbands’ career needs than their husbands are to accommodate to their work (Steil & Weltman, 1991), and in part because of women’s greater home and child-care responsibilities (cf. O’Neill, 1985). As illustrated in Fig. 1, these structural inequalities are proposed to translate into gender differences in beliefs about personal entitlement to pay. This is hypothesized to occur through two processes: (1) social comparison biases to estimate entitlement on the basis of intragroup and intrapersonal comparisons, and (2) beliefs and attributions that legitimize unequal distributions. In the following sections (V, B-G), I first demonstrate the existence of gender differences in personal entitlement to pay, and then illustrate how comparison processes and legitimizing beliefs contribute to these differences.

B. GENDER DIFFERENCES IN PERCEIVED PERSONAL ENTITLEMENT TO PAY Do women and men differ in their beliefs about what they are personally entitled to receive in terms of pay for work (Box F of Fig. l)? Several lines of evidence suggest that they do. A finding that has appeared with some regularity in the social-psychological literature on distributive-justice preferences is that women and men allocate rewards differently between themselves and others (for reviews, see Kahn & Gaeddert, 1985; and Major & Deaux, 1982). Specifically, when asked to divide a joint reward between themselves and a co-worker, men tend to take more of the reward for themselves, and give correspondingly less to the co-worker, than do women having the same performance level. One explanation for this pattern is that it reflects women’s greater concern for the feelings of others (cf. Sampson, 1975). Some studies provide modest support for this explanation (e.g., Major & Adams, 1983; Watts, Messe, & Vallacher, 1982). Women’s lower self-allocations, however, are not restricted to situations in which they must choose between rewarding themselves and rewarding another, or indeed to situations where another person is directly involved. Callahan-Levy

316

BRENDA MAJOR

and Messe (1979, Experiment l), for example, examined how women and men allocate rewards to themselves and to others when these decisions are independent or noncontingent. Men and women did not differ in the amount of money they paid to others, male or female. They did, however, differ in how much they paid themselves. After working on the same task for the same time period, women paid themselves less money and reported that less money was fair pay for their work than men did. Several additional studies provide further evidence that women and men differ in what they expect or feel they deserve to be paid for their work. In one study (Major, McFarlin, & Gagnon, 1984, Experiment 1) male and female undergraduates were asked to pay themselves privately, from a $4.00 pot, what they thought was fair pay for their work (our “operationalization” of entitlement). One fourth of the subjects took what they thought was fair in the absence of any social comparison information. The remaining subjects were given bogus information indicating how much money other subjects had presumably paid themselves for equivalent work on the same task. One fourth of the subjects saw that comparison men had paid themselves more (M = $2.50) than comparison women (M = $1.50);one fourth saw that women (M = $2.50) had paid themselves more than men (M = $1SO); and one fourth saw information indicating that men and women had paid themselves the same average amounts (M = $2.00). As expected, we found that when no social comparison standards were available, women paid themselves less (M = $1.95) than men did (M = $3.18); p < .01). Women in this condition also reported that less money was fair pay for their work (M = $1.89) than did men (M = $3.61). In contrast, no gender differences in self-pay occurred in any of the conditions in which clear social comparison cues (i.e., the self-pay of others) were present. Both men and women in these conditions matched their self-pay to the average of the comparison standards provided ($2.00). Thus, unless they were in the presence of salient comparison standards identical to those to which men were exposed, women felt that they deserved less pay for their work than did men. A recent pair of studies by Desmarais (1993) replicated this pattern. When asked to pay themselves (from an $8.00 pot) what they thought they deserved for their work, women paid themselves significantly less than men did when no comparison standards were made salient. Our second study (Major, McFarlin, & Gagnon, 1984; Experiment 2) explored the corollary hypothesis: If women have a lesser sense of entitlement than do men in terms of pay for work, then women should work longer and do more work for a fixed amount of money than men when no social comparison standards are salient. Male and female university students in this experiment worked on a lengthy task. Prior to beginning work on the task, they were paid $4.00 and were asked to do as much work as they thought was fair for the amount of money they were paid. They were told that the $4.00 was theirs to keep regardless of how long they worked or how much work they completed. As predicted, women

ENTITLEMENT AND SOCIAL INEQUALITY

317

worked significantly longer, did more work, completed more correct work, and worked more efficiently than did men. These gender differences were significant both among participants who believed their work was being monitored by the experimenter and among participants who believed their work was unmonitored. Thus, this study also indicates that in the absence of salient social comparison standards, men and women differ in their sense of entitlement with respect to pay for work. These findings are not limited to laboratory settings. Desmarais and Curtis (199 1) found gender differences in perceived income entitlement among a nationally representative sample of Canadian workers. Full-time workers were asked, “What was the income you deserved (in the previous year) all things considered?’ Women workers responded that they deserved significantly lower incomes than did men, even when “human-capital factors” (age, educational level, and years of experience in the same occupation) and job factors were controlled. Furthermore, in a survey of perceived wage entitlements among firstyear undergraduates, Desmarais (1993) found that college women reported deserving lower wages in their summer jobs than did college men, and that women’s estimates of the yearly salary they would deserve for their first full-time job after graduation were significantly lower than were men’s, even after controlling for college major and career aspirations.

C. GENDER DIFFERENCES IN REFERENCE STANDARDS FOR PAY

The theoretical model outlined above and illustrated in the top path of Fig. 1 predicts that women’s and men’s standards of reference (Box C of Fig. 1) are an important determinant of these gender differences in entitlement. Women’s and men’s reference standards are proposed to be biased by a tendency to make intragroup and intrapersonal social comparisons of outcomes (Box B). Three types of social comparisons are discussed in this section: (1) selflother comparisons, (2) self/group comparisons, and (3) self/self comparisons. Each of these types of comparisons is proposed to lead to gender differences in reference standards for pay.

I . SevIOther Comparisons Social comparisons between self and specific others are an important and powerful determinant of beliefs about personal entitlement. According to the model presented in Fig. 1, if women and men tend to make intragroup self/other comparisons (e.g., selectively compare with others who are the same sex or who do the same type of work as themselves), gender differences in comparison

318

BRENDA MAJOR

standards for pay will result. This will occur because women’s standards are more likely to be based on comparisons with specific others who are disadvantaged (e.g., other underpaid women, people working in lower-paid femaledominated jobs). Several field studies have demonstrated that women and men are more likely to name same-sex than cross-sex others and same-job than different-job others when asked with whom they compare their wages or jobs (Crosby, 1982; Goodman, 1974; Moore, 1990; Oldham et al., 1982; Patchen, 1961). Zanna, Crosby and Loewenstein (1987), however, demonstrated that working married women without children were an exception to this pattern. These women were more likely to compare their jobs with those of men. They were also more dissatisfied with their pay than were other women, as would be predicted from a reference standard perspective. Both proximity and similarity contribute to the tendency to make intragroup (within-sex and within-job) comparisons in field settings. Because of the sexsegregation of the labor market, same-sex co-workers and others doing the same job are more proximal and visible, and hence more accessible for comparison purposes. Women are also more likely to compare their pay with that of other women or others also working in “women’s jobs” because these individuals are assumed to be more similar to themselves on attributes related to pay (e.g., qualifications, type of job held, abilities or family constraints) (cf. Festinger, 1954; Goethals & Darley, 1977), and hence more appropriate for comparison purposes and more diagnostic of personal entitlement. Because most people work in jobs dominated by their own sex, however, the effects of proximity versus similarity on comparison choices cannot be disentangled. We conducted several experiments to examine whether intragroup comparisons of outcomes are preferred to intergroup comparisons of outcomes, and how these outcome comparison preferences affect comparison standards, entitlement, and satisfaction. These experiments examined comparison preferences in “legitimated” contexts (i.e., contexts in which there was no reason for participants to question the legitimacy of the distributive or procedural rules underlying the distribution of rewards). Our first experiment examined whether people prefer intragroup (same-sex and/or same-job) pay comparisons over intergroup pay comparisons when both types of comparisons are equally available (Major & Forcey, 1985). College students were assigned randomly to work on a job described as masculine, feminine, or sex-neutral (all students actually worked on the same sex-neutral task). They were informed in advance that they would be paid according to their performance and job assignment. After working on the task, all were given privately an ambiguous performance score and paid an identical wage. They were then given the opportunity to rank order their preference for seeing the average wage paid to nine different groups of individuals-the average male, average female, and average combined-sex wage for each of the three jobs.

ENTITLEMENT AND SOCIAL INEQUALITY

319

(They did not actually see this information.) They also indicated how fair they thought the wage they received was and how satisfied they were with it. Consistent with an intragroup bias in comparison choices, most men and women preferred to maximize similarity in their choice of wage comparisons, with the majority (63%) choosing to see the pay of the same-sex and same-job group first. This preference occurred despite the equal availability of the crosssex wage and combined-sex wage. Same-job others were chosen first regardless of the sex linkage of the subject’s assigned job. Sex linkage of job did, however, affect same-sex preferences: 80% of participants assigned to “sex-appropriate” jobs, and 67% assigned to sex-neutral jobs, chose to compare with same-sex others first, in contrast to only 47% assigned to “sex-inappropriate” jobs. In a follow-up study (Major & Testa, 1989, Experiment I), we tested the hypotheses (implied in Fig. 1) that: ( I ) intragroup biases will occur in social comparison choices; (2) these biases will lead to group differences in reference standards when some groups (e.g., women) are paid less than other groups (e.g., men); (3) these standards will affect group differences in judgments of personal entitlement, which (4) will in turn affect ratings of pay satisfaction. Men and women in this experiment worked on one of two sex-neutral tasks (which were in fact identical) and were then privately paid an identical wage. They were then given the opportunity to see the wage received by one other student who had presumably participated in the experiment earlier. They could select this student from a list of eight-two males and two females for each of the two jobs. Unbeknown to the participants, in one job the pay to males was higher than that to females, whereas in the other job this pattern was reversed. Participants were given the wage information corresponding to whatever comparison target they had selected. They then indicated how much pay they thought others had received (comparison standards), how much pay they felt was fair for them to receive (personal entitlement), and how satisfied they were with their own pay. Once again there was a strong intragroup bias in choice of comparison other; 91% of subjects chose to see the wage of a same-sex/same-job other. Second, as a result of this intragroup comparison bias, those assigned to a job in which their own sex was disadvantaged had lower reference standards for pay than did those assigned to a job in which their own sex was advantaged. The former believed that a lower wage standard existed for everyone (e.g., in general, in their own job, in the other job, for men, and for women) than did the latter. Third, those in the former group, regardless of sex, also tended to feel personally entitled to a lower wage, and were significantly more satisfied with the wage they received than were those in the latter group. To summarize, these studies support the model presented in Fig. 1. First, a strong preference to make intragroup (same-sex) rather than intergroup (crosssex) comparisons in judging entitlement or evaluating outcomes was observed

320

BRENDA MAJOR

(Box B of Fig. 1). These same-sex biases occurred in contexts in which infonnation about the pay of cross-sex others was as accessible as was information about the pay of same-sex others, and in contexts in which sex was ostensibly unrelated to performance. Second, this research showed that intragroup biases in choice of comparison others can lead, by quite logical means, to gender differences in (1) reference standards for pay (Box C), (2) perceptions of personal entitlement (Box F), and (3) pay satisfaction (Box G ) when they occur in a context in which women and men are differentially rewarded for their work. If people compare only with others like themselves (who are also disadvantaged), they may not become aware that better alternatives exist. 2 . SelflGroup Comparisons

Selflgroup comparisons also produce gender differences in reference standards for pay. The status-value formulation of equity theory (e.g., Berger et al., 1972) asserts that people make comparisons with similar generalized others, and by so doing come to feel entitled to the same reward level that is typically allocated to people like themselves. Most people are aware that women typically are paid less than men for their work, and that “women’s jobs” are lower paying than “men’s jobs” (cf. Crosby, 1982). Several of our studies illustrate the nature of stereotypes about the relationship between gender and pay and gender-linked work and pay. Major, McFarlin, & Gagnon (1984) found that in the absence of salient social comparison standards, male and female college students thought that men would pay themselves significantly more money (M = $3.19) than women would pay themselves (M = $2.55) for doing the same work (see also Desmarais, 1993), and that women would work significantly longer (M = 51.00 min) than men would (M = 39.33 min) for the same pay. Furthermore, in the study by Major and Forcey (1985) described above, both men and women assigned to do the “feminine job” expected to earn significantly less money prior to doing the job, and thought that the (identical) pay they subsequently earned was significantly more fair and more satisfactory, than did men and women assigned to do the identical job labeled as “masculine.” To the extent that women and men estimate their entitlement or evaluate their outcomes on the basis of their stereotypes of what members of their own sex typically receive, women will have lower reference standards for pay than will men. Several lines of evidence suggest that women’s “reference group” standards for pay are lower than are men’s. Major and Konar (1984) surveyed male and female management students enrolled in a work internship program regarding their own career-entry and career-peak pay expectations, their expectations as to what others in their field (sex-unspecified) would earn at career entry and career peak, and other factors that might predict pay expectations (e.g., specialty area, qualifications, job values). Male management students expected to earn approx-

ENTITLEMENT AND SOCIAL INEQUALITY

32 1

imately $2,600 more than female management students expected to earn at career entry ( p < .05), and almost $20,000 more at career peak (p < .01). Men also thought that others in their field would earn significantly more at career entry and at career peak than women did. These gender differences in comparison standards accounted for more of the gender difference in career-entry and career-peak pay expectations than did any other factor. Subsequent studies have replicated these gender differences in pay expectations and in social reference standards for pay, as well as the finding that gender differences in reference standards (either perceived pay of others or estimates of “fair pay”) were by far the strongest predictor of gender differences in pay expectations (L. A. Jackson, Gardner, & Sullivan, 1992; McFarlin, Frone, Major, & Konar, 1989). Why are women’s estimates of what sex-unspecified “others” earn lower than men’s? In part because women’s and men’s general reference standards are biased by their stereotypes of what members of their ingroup (e.g., others of their sex or co-workers) earn. We have found that stereotype-basedintragroup comparisons exert a stronger influence on pay expectations and estimates of deserving than do stereotype-based intergroup comparisons. McFarlin et a!. (1989), for example, found that management students’ own pay expectations were more strongly related to their perceptions of what same-sex others earn than to their perceptions of what cross-sex others earn. Major, McFarlin, & Gagnon (1984; Experiment 1) found that among women and men asked to pay themselves what they thought they deserved in the absence of social comparison information, selfpayments were more highly correlated with what they thought same-sex others had paid themselves than with what they thought cross-sex others had paid themselves. A study by Biernat, Manis, and Nelson (1991; Experiment 2) illustrates how stereotypes about women’s and men’s typical rates of pay influence the judgment standards used to subjectively evaluate men’s and women’s financial status. Participants in this study were shown slides of male and female faces and were asked to indicate the financial success of these targets. Half were asked to estimate the dollars earned per year by the target (an objective rating). The other half were asked to rate the targets’ financial success on a scale from 1 (very unsuccessful) to 7 (very successful) (a subjective rating). Participants using the objective scale judged the men to be earning substantially more dollars per year than were the women, thus confirming the existence of the stereotype that men earn more than women. Participants who rated financial success on a subjective scale, however, reversed this effect; these judges rated female targets as being more financially successful than were male targets. This result illustrates that in domains where group (sex) stereotypes exist (such as earnings), subjective ratings of males and females are made against a sex-stereotypedjudgment standard. A similar phenomenon should apply to subjectivejudgments about the self (e.g., satisfaction with own pay). Women and men estimate their personal deserving

322

BRENDA MAJOR

against a (same) sex-stereotyped judgment standard. As a consequence, even though women earn objectively less (in dollars per year) than do men, they report themselves to be just as subjectively satisfied as men. Summing up, women and men have similar stereotypes about the relationship between gender and pay and gender-linked work and pay. Both sexes base their pay expectations and estimate their entitlement against judgment standards that are biased by these sex-stereotyped mental representations. Stereotype-based intragroup comparisons (e.g., assumptions about the rewards typically received by members of the same sex) influence the reference standards against which men and women evaluate their own outcomes. Because women and people doing “women’s jobs” are typically paid less than men and people doing “men’s jobs,” women estimate their personal deserving and evaluate their outcomes against a lower reference standard for pay than do men. 3 . Self1Self Comparisons

Because of gender-based inequality in the work force and the family, and discrimination against women in the domain of paid work, selflself comparisons are also likely to result in lower reference standards for pay among women than among men. As noted above (Section II,A), there is substantial evidence that female workers are paid less than male workers, even when women and men are matched on numerous job characteristics that typically covary with sex (e.g., Marini, 1989; Treiman & Hartmann, 1981). When we have assessed prior pay history in our own research (e.g., Major, McFarlin, & Gagnon, 1984), we have consistently found that female college students report receiving less money in their current or prior jobs than do male college students. Personal experiences with past pay are an important determinant of women’s and men’s sense of entitlement to pay. A series of studies by Desmarais (Desmarais, 1993; Desmarais & Curtis, 1991) nicely illustrates this point. Recall that Desmarais and Curtis (1991) found that among full-time workers in Canada, women’s estimates of the incomes they deserved were significantly lower than were men’s estimates. These women workers, however, also had earned significantly less money than had male workers during the prior year, and when gender differences in prior income were controlled for, gender differences in entitlement were eliminated. Desmarais (1993) replicated these findings in a survey of first-year undergraduates: Gender differences in income entitlements were eliminated when prior gender differences in experiences with income were statistically controlled. Furthermore, Desmarais (1993) demonstrated in a laboratory experiment that gender differences in self-pay could be eliminated if women who had previously earned as much as men were reminded of their income just prior to making their self-payments. Together, these studies indicate that when women have been paid the same as men (and hence have the same self/self comparison standards available as men

ENTITLEMENT AND SOCIAL INEQUALITY

323

do), and this self-comparison standard is salient, they think about the income they deserve in the same way as do men. Social inequality and discrimination are also likely to depress what women, relative to men, regard as feasible or realistic alternatives to their current situation. Thibaut and Kelley (1959) observed that evaluations of existing conditions are based on comparison levels for alternatives, as well as on absolute comparison standards. Because of sex discrimination in employment opportunities, less job mobility, and greater home responsibilities, women are likely to have lower comparison standards for alternative pay than men. That is, women are less likely than men to regard high pay or a high-paying job as feasible options for themselves. In sum, as a result of gender differences in what women and men have personally experienced with regard to work and pay, or realistically expect they could obtain in this regard, women’s reference standards for pay derived from intrapersonal comparisons are typically lower than are men’s.

D. GENDER, COMPARISON STANDARDS, AND ENTITLEMENT TO PAY: A SUMMARY The research reviewed above empirically supports the social comparison path from inequality to entitlement illustrated in the top portion of Fig. 1. Women and men estimate what they deserve on the basis of intragroup and intrapersonal comparisons that occur in a social context in which women are disadvantaged (e.g., a sex-segregated labor market, lower pay to women and women’s jobs than to men and men’s jobs, and unequal opportunities and home responsibilities). On the basis of their self/other, self/group, and self/self comparisons, women derive a lower reference standard for pay than do men. These lower reference standards, in turn, result in women having a lesser sense of entitlement to pay than do men. Thus, the underpaid female worker feels she is getting from her job what she deserves, based on (1) the people with whom she typically compares (primarily other underpaid women); (2) her stereotypes of what other women in general and people doing “women’s jobs” are typically paid, and (3) what she has earned in the past or realistically expects to obtain given her restricted job opportunities and greater home responsibilities (see Major, 1989). Gender differences in entitlement for pay are an important reason why women are no less satisfied than men with their jobs or pay despite their objective disadvantage.

E. LEGITIMIZING GENDER INEQUALITY The comparison processes described above occur in the context of distributions that are appraised as legitimate. That is, people believe they deserve the same as similar others, or what they have received in the past, when they believe

324

BRENDA MAJOR

that the distribution of outcomes across individuals or groups is legitimate. As discussed above, unequal social distributions have a powerful tendency to be legitimated. One way they are legitimated is through a set of beliefs that locate the cause of a person’s outcomes within that person (cf. Cohen, 1982; Lerner, 1980). In general, differences in outcomes among people or groups that are attributed to factors internal to those individuals or groups (e.g., differences in abilities, qualifications, personal choices, needs) are more likely to be perceived as legitimate than are differences among people or groups that are attributed to an external agent (e.g., another person or a system). This analysis suggests that if outcome distributions are appraised as legitimate (e.g., attributed to internal factors), people will estimate their personal entitlement more on the basis of intragroup than intergroup outcome comparisons even if they are aware of their group’s disadvantaged status relative to the outgroup. That is, people will assume that they deserve the same outcomes or treatment as others like themselves (the ingroup), even if intergroup outcome comparisons are salient and even if this information makes clear that the ingroup is disadvantaged relative to the outgroup. Bylsma and Major (1992a) tested this hypothesis. Participants in this experiment completed a (sex-neutral)task and were told that how much they earned would depend on their performance on the task (an internal, legitimating attribution). While performing the task, they saw a bogus “cash receipt form” that contained the comparison information manipulation. One third of the participants saw that on average, women who had participated in the experiment had been paid more (M = $4.45) than men (M = $2.55), whereas one third saw the reverse. The cash receipt form was blank for the other third. Subsequent to completing the task, participants indicated how much they deserved to be paid (entitlement), and how well they thought they had done (perceived inputs). All then received an identical payment ($3.50),and were asked how much they thought other subjects had earned (reference standards) and how satisfied they were with their payment. We predicted that reference standards, entitlement, and satisfaction would be based more on intragroup comparisons than on intergroup comparisons. Results were consistent with predictions (see Table I). Men and women who saw that their own sex was paid less than the other sex: (1) had lower reference standards for pay (thought others in general had earned less in the experiment), (2) said that they personally deserved less money, and (3) were more satisfied with their own payment than were men and women who saw that their own sex was paid more than the other sex. Responses of men and women in the no-comparison condition were intermediate between these two groups. Furthermore, subjects who saw that members of their own sex had been paid more than members of the other sex rated their own performance significantly higher (M = 5.21) than did subjects who saw that members of their own sex had been paid less than members of the other sex (M = 4.28). Thus, both sexes

ENTITLEMENT AND SOCIAL INEQUALITY

325

TABLE I COMPARISON STANDARDS, ENTITLEMENT, A N D PAYSATISFACTION BY RELATIVEEARNINGS OF INGROUP (SAME-SEX) TO OUTGROUP (OTHER-SEX)" Comparison condition Judgment

lngroup > outgroup

No information

Outgroup > ingroup

Estimated earnings of average other Personal entitlement Pay satisfactionb

$3.74, $3.37, 5.82,

$3.41, $3.29, 6.00,

$3.37, $2.61, 6.59,

OFrom Bylsma and Major (1992a). Within each dependent variable, means with different subscripts differ at p < .05. bPay satisfaction was rated on a I to 7 scale with 7 indicating very sarisjed.

inferred their own performance inputs from observations of the relative outcomes received by women versus men. Although this pattern occurred for both sexes, it was stronger for women that it was for men. Finally, among participants who saw no social comparison information, women thought they had performed less well on the task than did men ( p < .05) and thought they deserved significantly less money (M = $2.90) than men thought they, themselves, deserved (M = $3.59). This latter set of findings provides evidence of the alternate route by which women may come to have a lesser sense of entitlement to pay than men. This is the route that is illustrated in the bottom part of Fig. 1. As noted earlier, people often infer that those who possess greater outcomes have better met the preconditions for obtaining those outcomes than did those who have lower outcomes (Box D in Fig. 1). Because of this inference process, when people do become aware of outcome disparities between themselves and others, or between their ingroup and an outgroup, they may nonetheless infer that these differences were caused by personal or group-related differences in preconditions (e.g., inputs). Furthermore, observations of group-related differences in outcomes may lead to corresponding inferences about personal levels of inputs (broadly defined). That is, a person who infers that her group has lower inputs (because of their lesser outcomes) may also infer that she, as a member of that group, has lesser inputs. This inference process from group-related outcomes to personal inputs is illustrated by Box E in Fig. 1. Lower evaluations of personal inputs, in turn, are related to a lesser sense of personal deserving (Box F of Fig. 1). Thus, as a result of the observation that women are paid less for their work than men are, women (and men) may infer that women have not met the preconditions necessary to obtain high pay to the same extent as men. For example, women may be assumed to hold less responsible jobs, be less willing to make career-related sacrifices, be less qualified, and so forth. Furthermore, women

326

BRENDA MAJOR

may make similar inferences about their own personal contributions or preconditions. Women’s lower evaluations of their own contributions, in turn, may lead them to feel they personally deserve less pay for their work than men do. Several lines of evidence are consistent with this analysis. In the absence of explicit performance feedback, women tend to have less confidence in their abilities and performances, and to evaluate them lower, than men do (see Lenney, 1977 for a review). Roberts and Nolen-Hoeksema (1989), for example, found that women’s productivity estimates were lower than men’s in conditions in which participants received feedback that was irrelevant to their competence or performance. Major and Forcey (1985) found that in the absence of explicit external social comparison standards, women’s evaluations of their own performances were lower than men’s. These women also thought they deserved significantly less pay for their work than did men. (This pattern was observed regardless of whether the task was described as masculine, feminine, or gender neutral). A similar pattern was observed in the study by Bylsma and Major (1992a) described above (Section V,E). Women are also sometimes (although not consistently), more likely to attribute objectively poor outcomes to themselves than are men (cf. Deaux, 1984; but see Frieze, Whitley, Hartman Hanusa, & McHugh, 1982). In addition, women see performance evaluations as more diagnostic of their abilities than do men (Major, Osborne, & Crocker, 1992; Roberts, 1991). These latter tendencies should heighten women’s belief that variations in outcomes between themselves and others are diagnostic of variations in contributions between themselves and others.

F. TWO ROUTES TO ELIMINATING GENDER DIFFERENCES IN ENTITLEMENT The theoretical framework and research presented so far suggest that group differences in both comparison standards and self-evaluations of inputs contribute to group differences in personal entitlement. Accordingly, this framework predicts that controlling for either of these initial differences can eliminate or at least attenuate group differences in personal entitlement. As applied to gender, this model predicts that gender differences in perceived entitlement to pay should be eliminated by providing women and men with either: (1) the same salient social (or self) comparison standards, and/or (2) the same explicit feedback about their contributions. Bylsma and Major (1992b) tested this hypothesis in an analogue study. Participants read a scenario in which they were asked to imagine themselves in the position of a research assistant who had been hired by a professor for an undetermined wage (a potential hourly wage of $3.50 to $7.50 was specified). Social comparison information and performance feedback were manipulated. Depend-

ENTITLEMENT AND SOCIAL INEQUALITY

327

ing on the condition to which they had been assigned, participants either read that three other research assistants (sex-unspecified) had received high wages (an average of $6.50) or low wages (an average of $4.50), or read no wagecomparison information. In addition, participants either read that their performance as an assistant had been evaluated favorably by the professor or unfavorably by the professor, or they received no performance feedback from the professor. Participants were asked to indicate how much they deserved to be paid. Subsequent to making this judgment they were asked to imagine that they had been paid $5.50 and were asked to indicate how fair and satisfactory this wage was. Consistent with our earlier studies, in the absence of both comparison information and explicit performance feedback, women’s estimates of how well they had performed were significantly lower than were men’s, as were women’s estimates of what they deserved to be paid (M = $5.30) compared to men’s estimates (M = $6.03). As predicted, however, women and men exposed to high social comparison standards felt they deserved significantly more ( M = $5.64) than did those exposed to low social comparison standards (M = $4.78), and those told they had performed well thought they deserved significantly more (M = $5.59) than did those told they had performed poorly (M = $4.86). No gender differences were observed in these conditions. Thus, exposing women and men to either the same social comparison standards or the same performance feedback eliminated gender differences in personal deserving.

G. FROM ENTITLEMENT TO OUTCOME EVALUATIONS The theoretical framework presented here not only assumes that perceived entitlement is an important determinant of outcome evaluations but also assumes that entitlement mediates the effects of antecedent factors, such as comparison standards and perceived contributions, on outcome evaluations. This assumption, although shared by most justice theories, has rarely been tested. Bylsma and Major (1992b) examined this assumption by conducting path analyses on data from the analogue study, described in Section V,F. We tested both the direct and mediated (through entitlement) effects of social comparison condition (high vs. low wages to others) and contribution condition (good vs. poor performance feedback) on ratings of pay fairness and satisfaction. Results of the path analysis are shown in Fig. 2. As expected, both contributions and comparison standards significantly predicted entitlement. People exposed to high social comparison standards or who received good performance feedback felt they deserved more pay than did people exposed to lower social comparison standards or who received poor performance feedback. Perceived entitlement, in

328

BRENDA MAJOR

-.79 Entitlement

-64 Entitlement

Satisfaction

b

-

Fairness

Fig. 2. The effects of comparison information, perceived performance, and entitlement on satisfaction and fairness. All betas significant at p < .05.

turn, was significantly related to outcome evaluations. The more people felt they deserved to be paid, the less fair and satisfactory they rated the (moderate) wage they received. Importantly, neither performance feedback nor comparison information was directly related to judgments of pay fairness. That is, both of these types of information affected judgments of fairness only through their impact on entitlement. Social comparison information also was not directly related to pay satisfaction. A small direct relationship was observed between performance feedback and pay satisfaction, indicating that those who thought they had performed better were more satisfied with their pay. This unexpected relationship may have been due to a spillover of positive affect in the positive feedback condition to satisfaction ratings (an affectively tinged measure). In summary, consistent with the theoretical framework presented here, contributions and comparisons both predicted entitlement, and entitlement was a significant predictor of outcome evaluations. Furthermore, the effects of social comparison information on outcome evaluations (ratings of both pay fairness and satisfaction) were entirely mediated through perceived entitlement. Likewise, the effects of contributions (performance feedback) were either entirely (in the case of ratings of pay fairness) or predominately (in the case of pay satisfaction) mediated through perceived entitlement.

ENTITLEMENT AND SOCIAL INEQUALITY

329

VI. When Inequality Becomes Undeserved The theoretical framework and research on gender and entitlement presented thus far in this paper illustrate that often those who are disadvantaged do not perceive themselves as disadvantaged or as receiving less than they legitimately deserve. As a result, they may appear “paradoxically content” despite objective disadvantage. This analysis may cause one to wonder how social protest movements and revolutions ever occur. Clearly, some groups do report considerable discontent, and attempts at both personal mobility and social change do occur. In comparison to women as a group, for example, African-Americans as a group are discontent with their situations, regard them as illegitimate, and are committed to collective action (see Gurin, 1985). This section considers what causes some individuals and social groups to believe that what “is” is not the way it “ought” to be. Two essential conditions must be met before people who are objectively deprived of valued outcomes will believe that their situations are undeserved. First, disadvantage must be perceived to exist, and second, it must be perceived as illegitimate (cf. Folger, 1986; 1987; Gurin, 1985; Gurin & Townsend, 1986; Miller et al., 1981). Thus, personal and situational factors that: (1) encourage awareness of discrepancies between one’s own outcomes and those of others and that (2) foster the perception that discrepancies are illegitimate will heighten a sense of personal entitlement and increase feelings of personal deprivation and discontent. Social comparison processes are an important determinant of awareness of disadvantage. The more that members of disadvantaged groups make intergroup rather than intragroup comparisons, the more likely they are to become aware of discrepancies between their outcomes and those of advantaged others. Attributions are an important determinant of the appraised legitimacy of perceived discrepancies. The more that discrepancies among groups are attributed to external agents, such as the prejudices of another person or a biased system, rather than to internal factors, the more likely it is that outcome distributions will be appraised as illegitimate (Gurin, 1985). As noted in Section IV,C, comparison processes and appraisals of legitimacy are reciprocal processes. Appraisals of legitimacy affect whom people regard as appropriate comparison targets; when outcomes are regarded as illegitimate, intergroup comparisons are more likely to be made. Likewise, comparisons affect appraisals of legitimacy; when intergroup comparisons foster awareness of group discrepancies, the legitimacy of the distribution is more likely to be questioned (Gurin, 1985; Gurin & Townsend, 1986). In this section, several factors that heighten awareness of intergroup disparities in outcomes, and/or that enhance perceptions that outcomes and interpersonal

330

BRENDA MAJOR

treatments are illegitimate, are discussed. These include: (1) personal and situational factors that lead to a collective rather than a personal perspective on deprivation and disadvantage, (2) personal dispositions and collective representations that locate the cause of disadvantage in external agents (e.g., the system) rather than in individual attributes, and (3) factors that increase the perception that the procedures that produced the current distribution of outcomes are unfair (e.g., biased, inconsistent, prejudiced).

A. FROM PERSONAL TO COLLECTIVE ENTITLEMENT: GROUPlGROUP COMPARISONS Perhaps one of the most important determinants of whether people become aware of and aggrieved about inequalities is whether they are estimating their entitlement and evaluating their outcomes from a personal (individual) or a collective (group) perspective. In terms of Levine and Moreland’s (1987) taxonomy of outcome comparisons, this latter comparison is a grouplgroup comparison. In general, personal or situational factors that cause people to shift from thinking about themselves as individuals to thinking about themselves as a member of a group should also shift their comparison targets away from the ingroup and toward outgroups. Comparisons with advantaged outgroups, in turn, should elevate a sense of collective entitlement among the disadvantaged because it enhances awareness of disadvantage. This distinction between personal and collective entitlement borrows from the important distinction, made originally by Runciman (1966), between personal (or egoistic) and group (or fraternal) deprivation. Personal deprivation refers to a type of personal discontent that occurs when an individual compares his or her own situation to that of other individuals or groups. Group deprivation refers to a more social or group discontent that results when an individual compares the situation of his or her own reference group as a whole with that of some other reference group (cf. Crosby, 1982; Martin, 1986a; Runciman, 1966). The domains and consequences of personal and group deprivation differ. Personal deprivation is most relevant to evaluations of oneself or one’s own situation, whereas group deprivation is more pertinent to understanding the role of perceived injustice in bringing about structural changes in society (e.g., Dion, 1986; Dube & Guimond, 1986; Vanneman & Pettigrew, 1972). It is those who experience group deprivation, rather than personal deprivation, who are most likely to be involved in social protests and to push for social change (e.g., Dion, 1986; Dube & Guimond, 1986). What causes people to take a collective rather than an individual perspective on entitlement? Group identity is one critical factor (Crosby, Pufall, Snyder, O’Connell, & Whalen, 1989; Gurin, 1985; Gurin & Townsend, 1986; Miller et

ENTITLEMENT AND SOCIAL INEQUALITY

33 I

al., 1981).7 Group identity refers to a person’s awareness of membership in a group and feelings attached to being a member of that group (Miller et al., 1981; Tajfel & Turner, 1986). Gurin and Townsend (1986) conceptualize group identity as a multidimensional concept, consisting of the extent to which group members are perceived to be similar in their personal characteristics, are perceived to share a common fate (are treated similarly in society), and how central group membership is in the way a group member thinks of the self. Because group identity sharpens group boundaries, it fosters intergroup comparisons. As a result, group identity heightens awareness of between-group inequalities in outcomes and increases the likelihood that discontent will be experienced (e.g., Clayton & Crosby, 1992; Gurin & Townsend, 1986). Consistent with this hypothesis, Gurin and Townsend (1986) found that women who scored high on all three dimensions of group identity were considerably more discontent about the situations of women as a group than were women who did not. Furthermore, they found that women who scored high on the “common fate” dimension of group identity were more likely to believe that status disparities between women and men were illegitimate, and were more likely to endorse collective efforts to seek change in the status of women. Shifting people’s perspective from their personal situation to that of their group enhances feelings of discrimination. Crosby (1982), for example, found that although the women in her study did not believe that their personal situations were unjust, they were aware of and aggrieved about the underpayment of women as a group. This same pattern-perceiving a lower level of personal than of ingroup deprivation or discrimination-has since been observed across a number of racial, ethnic, and social groups (Crocker, Luhtanen, Blaine, & Broadnax, 1993; Crosby et al., 1989; Taylor, Wright, Moghaddam, & Lalonde, 1990; Taylor, Wright, & Porter, 1993). The size of the disparity between personal and group judgments, however, varies markedly by group (Crocker et al., 1993; Taylor et al., 1993). The women in Crosby’s (1982) sample, for example, reported almost no personal deprivation, despite perceiving substantial group deprivation. In contrast, the black college students in Crocker et al.3 (1993) sample perceived high levels of both personal and group deprivation, although judgments of the former were lower than the latter. White students in Crocker et al.’s (1993) sample reported relatively low levels of both group and personal discrimi-

’Group identification has been conceptualized both as a component of group consciousness (e.g., Gurin, 1985; Miller et al., 1981) and as distinct from group consciousness (e.g., Gurin &

Townsend, 1986). Gurin and Townsend (1986) define group consciousness as the following: a sense of collective discontent over the group’s position, an appraisal that the stratification system producing the group’s position is illegitimate, and the belief that collective action is required to realize the group’s interests. These three components of group consciousness are treated as different variables in the current framework.

332

BRENDA MAJOR

nation on the basis of race. But like other groups, white students perceived higher levels of group than of personal discrimination, illustrating that the personal/group discrepancy extends beyond disadvantaged groups to dominant, or hegemonic groups. Several different explanations exist for this personal versus group discrepancy. Some focus on motivational factors, whereas others focus on cognitive factors (see Crosby et al., 1989; Taylor et al., 1990; Taylor et al., 1993, for reviews). One explanation focuses on denial (Clayton & Crosby, 1992; Crosby, 1982). According to this perspective, it is too threatening for members of disadvantaged groups to acknowledge that they are personally discriminated against, because this belief requires the recognition that they are personally a victim and that others in their immediate environment are harm doers. This type of explanation has difficulty accounting for why some groups (e.g., African-Americans; Haitian immigrants) perceive substantiaj amounts of personal discrimination (Crocker et al., 1993; Taylor et al., 1993). A second explanation focuses on the types of comparisons prompted by evaluating the outcomes of an individual versus a group. As illustrated by the research reviewed in this article, judgments of personal deserving and evaluations of personal outcomes are typically arrived at through infrugroup comparisons. Members of disadvantaged groups typically compare their personal situations with those of others who are also members of disadvantaged groups. As a result, they see little disparity between their own outcomes and those of their comparative referents. Evaluations of the situation of a group as a whole, in contrast, by necessity require making intergroup comparisons. As a result of these betweengroup comparisons, disparities between the situation of one’s own group and that of others become more apparent. Hence, intergroup comparisons heighten awareness of relative disadvantage. A third explanation for differences in the experience of personal versus group deprivation focuses on the attributions that can be made for the lower outcomes of an individual versus a group. Whereas there are many different factors that may be perceived as legitimately accounting for the poorer outcomes of one individual (including oneself) relative to others (e.g., personal choices, ability level, motivation), these types of internal attributions become less legitimate and less plausible when applied across a group as a whole. It is easier to see discrimination on the collective level than on the individual level (Clayton & Crosby, 1992; Crosby, Clayton, Hemker, & Alksnis, 1986). The perceived legitimacy of allocating outcomes on the basis of ascribed or group characteristics such as race, religion, or sex can change in a culture over time (Crocker & Major, 1993; Gurin, 1985). It was only a few decades ago when exclusion of blacks from public facilities, schools, and jobs was considered legal and fair by many Americans. It was considered justifiable to pay men more than

ENTITLEMENT AND SOCIAL INEQUALITY

333

women for the same work. These practices were justified by the widely shared belief that whites and men, relative to blacks and women, were deserving of greater rewards because they make (or have the potential to make) greater contributions (e.g., they are smarter, more rational, better leaders, etc.). Men were also considered to be entitled to more pay than were women because of their greater need-men were judged to need higher pay in order to support their families (Kessler-Harris, 1982). The Civil Rights movements of women and African-Americans, as well as other groups, dramatically challenged the legitimacy of these procedures. The assumptions that men or whites as a group make higher contributions came under attack. Correspondingly, the belief that men or whites as a group are collectively entitled to greater benefits (because of their greater contributions or need) came under dispute (see Gurin, 1985). S. Steele (1992) has argued that in the late 1960s the U.S. federal government began to legitimize a different kind of collective entitlement-entitlement accorded to racial, ethnic, and other groups on the basis of redress of past grievances. For example, affirmative action policies were instituted as a way of attempting to rectify historical advantages given to some people (white males) over others (blacks, females). ironically, resistance to affirmative action policies is in part a result of the fact that the policy is seen to entitle some people to desired outcomes on the basis of group membership (an ascribed characteristic), rather than on the basis o’f individually controllable (achieved) characteristics (Nacoste, 1990). Both beneficiaries and nonbeneficiaries regard selection policies that give greater weight to achieved characteristics as fairer than policies that give greater weight to ascribed characteristics (Major, Feinstein, & Crocker, in press; Taylor & Dube, 1986). The social factors that shape the perceived legitimacy of collective versus individual entitlements and that determine whether and under what circumstances ascribed and achieved characteristics are considered legitimate preconditions to entitlement are extremely important issues for future research.

B. THE PERCEPTION OF PROCEDURAL INJUSTICE Perceptions of the fairness of the procedures underlying distributions are an important determinant of the perceived legitimacy of distributions and authorities (Tyler & Lind, 1992). Perceptions of procedural fairness are affected by the extent to which the procedures used in distributing rewards are seen as being biased, prejudiced, applied inconsistently, or based on irrelevant or inadequate information (see Leventhal, 1976; Lind & Tyler, 1988). In general, any event, situation, or personal factor that leads to the attribution that prejudice, bias, or other types of unfair procedures produced one’s own or one’s ingroup’s lesser

334

BRENDA MAJOR

outcomes heightens the perception of illegitimacy. The perception of illegitimacy, in turn, leads to the belief that one’s own or one’s ingroup’s treatment or outcomes were undeserved. Because they affect legitimacy appraisals, perceptions of procedural unfairness affect social comparison processes. In particular, beliefs about procedural fairness should affect the extent to which people estimate their entitlement on the basis of intragroup versus intergroup comparisons. If ingroup members’ outcomes are believed to be based on unfair procedures (e.g., a biased allocator), there is less reason to assume that ingroup outcomes are more diagnostic of personal deserving than are outgroup outcomes. Bylsma, Major, and Cozzarelli (1992) tested this hypothesis in a study that replicated and extended the study by Bylsma and Major (1992a) described in Section V,E. Women were asked to complete a (gender-neutral) task. While working on the task, half saw social comparison information indicating that prior male participants in the study had been paid more than prior female participants, whereas the other half saw social comparison information indicating that women had been paid more than men. In addition, women were randomly assigned to one of three allocation-procedure conditions. One third were told that pay was determined by objective performance (this objective condition replicated the study described above by Bylsma and Major (1992a). One third (the subjective condition) were told that pay would be based not only on the objective correctness of their performance but also on the experimenter’s subjective assessment of their performance. Women in the third (biased) condition received the same instructions as did participants in the subjective condition, but in addition, during the experiment a confederate posing as another subject spoke up and said that she had heard that the experimenter was unfair, that it probably didn’t matter how well they did on the task, and that she “bet the experimenter paid girls and guys different all the time.” After completing the task, women were asked how much they felt they deserved to be paid for their work. As expected, allocation procedures affected appraisals of the legitimacy of the distribution. Women in the subjective and biased conditions rated the experimenter as significantly more biased than did women in the objective condition. Furthermore, women’s estimates of personal deserving were affected by an interaction between social comparison condition and allocation-procedurecondition. As predicted, women in the objective condition based their estimates of their own entitlement more on the outcomes of their ingroup than on the outgroup. That is, replicating results of our earlier study (Bylsma & Major, 1992a), women in the objective allocation condition felt that they deserved less when they were exposed to social comparison information indicating that women had been underpaid than when they were exposed to comparison information indicating that women had been overpaid relative to men. In contrast, judgments of own deserving made by women in the subjective and biased procedure conditions were

ENTITLEMENT AND SOCIAL INEQUALITY

335

unaffected by the relative outcomes of comparison women versus men. That is, these women did not regard the outcomes of the ingroup (other women) to be more diagnostic of their own deserving than outcomes of the outgroup (other men). Indeed, the group of women who felt entitled to the highest pay were those who saw that males were paid more than females and who were led to believe that the procedure was biased. Thus, when members of disadvantaged groups begin to question the legitimacy of the procedures, or the larger system, that produced their own or their group’s objectively poorer outcomes or treatment, they are unlikely to feel that they or their group deserve their poorer outcomes or treatment. In addition, they are more likely to react to their poorer outcomes with anger, moral outrage, and attempts to seek change. Specific, salient events that highlight procedural injustice may trigger this process. This was illustrated by the 1992 Rodney King verdict, in which a group of Los Angeles police were acquitted of police brutality despite the existence of a widely publicized videotape highly suggestive that the police used excessive force. Reactions to the verdict included not only disillusionment with the fairness of the judicial system but also widespread rioting and looting.

C. PERSONAL IDEOLOGY AND COLLECTIVE REPRESENTATIONS The beliefs in a just world, in personal control, and in personal causation are robust within North American culture. These beliefs legitimize inequality by locating the cause of successful and unsuccessful outcomes within individuals. As noted above, these beliefs represent a “dominant ideology” of American life (Kluegel & Smith, 1986). Nevertheless, individuals and groups differ in the strength with which they endorse these beliefs. The less strongly individuals and groups endorse the dominant ideology, the less likely they should be to regard inequality between self and others, or own group and other groups, as legitimate (Martin, Scully, & Levitt, 1990). Results of several studies are consistent with this observation. Martin (1986a), for example, found that blue-collar workers who had an external locus of control were less accepting of large differences in pay between blue-collar technicians and management than were blue-collar workers with an internal locus of control. Similarly, secretaries who endorsed a “feminist” ideology were less accepting of and more angry about pay disparities between themselves and managers than were secretaries who endorsed a more “traditional” ideology (Martin, 1986b). Hafer and Olson (1989) found that the less strongly participants endorsed a belief in a just world, the less likely they were to perceive that their own (rigged) failure to earn points toward a desirable goal was fair.

336

BRENDA MAJOR

Groups also differ in their collective view of the reasons underlying inequalities between the social rewards received by members of the ingroup versus outgroups. The more that groups embrace an ideology that blames the system, rather than personal deficiencies, for group differences in outcomes or social status, the more likely they should be to question the legitimacy of inequalities in the distributions of social outcomes. Greater questioning of the legitimacy of the system, in turn, should be accompanied by the belief that inequality is undeserved, and by consequent feelings of discontent. Crocker and Major (1989; Major & Crocker, 1993) proposed that people who are members of stigmatized groups, that is, groups that are chronically and systematically exposed to prejudice and discrimination, experience considerable attributional ambiguity with regard to the cause of their social outcomes and treatment. This attributional ambiguity occurs because stigmatized group membership provides a readily available and plausible alternative explanation for the outcomes and treatment that the stigmatized receive. For members of such groups, it is often unclear whether negative treatment or outcomes (e.g., lower pay, lack of advancement) are deserved (e.g., result from their own personal acts or qualities) or are a result of prejudice and discrimination against their group. As a result, members of such groups may be sensitized to look for signs of prejudice, bias, and/or discrimination when they observe themselves or their ingroup receiving poor outcomes or negative treatment. To test this hypothesis, Crocker, Voelkl, Testa and Major (1991) conducted a study in which African-American and white students were led to believe they were participating in a study on friendship development with a white same-sex evaluator. The evaluator either could see them (blinds on a one-way mirror were up) or could not see them (blinds on a one-way mirror were down). After exchanging self-description forms with their partner, participants received either very favorable or very unfavorable interpersonal feedback from the “other subject.” Black students were more likely than white students to attribute the interpersonal feedback to the evaluator’s prejudice when the feedback was negative, rather than positive, and when the blinds were up (and the evaluator could see them, and hence knew their race) than when the blinds were down. (Both effects were significant for black students but not for whites). Furthermore, black students were more likely to attribute the feedback (both positive and negative) to their personality when they could not be seen than when they could be seen, whereas this effect was nonsignificant and in the opposite direction for white students. These data suggest that black students were less likely than were white students to feel that they deserved the feedback they received if they knew the evaluator was aware of their race, and especially if the feedback was negative. As noted earlier in this paper, in contrast to women as a group, AfricanAmericans as a group are more aware of and aggrieved about the extent of personal and group deprivation and discrimination that they experience (e.g.,

ENTITLEMENT AND SOCIAL INEQUALITY

337

Crocker et al., 1993). Several lines of evidence suggest that attributions of outcomes and/or interpersonal treatment to the system, or to external agents, contribute to this pattern. Relative to whites, African-Americans typically score more external on Rotter's Locus of Control scale (Porter & Washington, 1979), and are less confident that they have the ability to control their own lives (Campbell et al., 1976). Gurin and Epps (1975) suggest that these higher feelings of external control among African-Americans do not reflect a belief in a world subject to fate or chance, but rather reflect system-blame. Consistent with this assertion, African-Americansas a group are much more likely than are whites to claim that their race and prejudice have held them back from achieving success in life (Crocker et al., 1993), and to attribute race disparities in outcomes to structural forces (e.g., discrimination) than to personal deficiencies (Gurin, 1985). Gurin and Epps (1975) note that blaming the system indicates a healthy sensitivity to the real world among African-Americans, because they experience discrimination and prejudice systematically and reliably. Greater system-blame should lead to more questioning of the legitimacy of inequalities in the distributions of social rewards. African-Americans are less likely to endorse the "dominant ideology" regarding income inequality than are whites as a group. Blacks on average see more inequality of opportunity than do whites, they see income inequalities as less just, and are less confident that the social stratification system is fair (Kluegel & Smith, 1986). They stop short, however, of denying the justice of economic inequality in principle and of dismissing the ideas that the rich and the poor as individuals are deserving of their fate (Kluegel & Smith, 1986). In contrast, although women as a group express more doubt about the justice of economic inequality than do men as a group, sex differences in beliefs about economic inequality are much smaller than parallel race differences (Gurin, 1985; Kluegel & Smith, 1986). As noted in Section V,E, women as a group are somewhat more willing to attribute objectively poor outcomes to themselves than are men (cf. Deaux, 1984), and to see performance evaluations (Major et al., 1992; Roberts, 1991) and socially distributed outcomes that are presumed to covary with performance (Bylsma & Major, 1992a) as more diagnostic of their abilities than are men. There are a number of reasons for a lack of parallelism in the belief structures of African-Americans as a group and women as a group (see Gurin, 1985; Kluegef & Smith, 1986 for a discussion). One important reason is that gender inequality is not as marked as racial inequality, and perception of inequality is more difficult when it is not extreme. A second important reason is that compared to African-Americans, women are much less conscious of themselves as a group in relation to inequality in society (Gurin, 1985). On the basis of national surveys completed in 1972 and 1978, Gurin (1985) concluded that women were not as group conscious as blacks, people over 60 years of age, or people in blue-collar

338

BRENDA MAJOR

jobs. Relative to other subordinate categories in America, Gurin (1985) argues, women lack many of the structural conditions that promote bonds to others in the same category. The economic and social status of women and men are closely intertwined, and women and men share fundamental social values, including individualism and the belief in a just world. Most importantly, the structure of women’s relations with other women and men inhibit gender consciousness among women. Women and men interact frequently and intimately with each other as lovers, husbands and wives, mothers and sons, and brothers and sisters. These structural conditions can be contrasted with the situations of other subordinate groups, such as African-Americans. African-Americans experience considerably greater objective racial inequality and racial segregation, conditions that can promote awareness of disparity on the basis of the category of race. Furthermore, African-Americans’ relationships with the outgroup (whites) are more emotionally distant, and the black community has been organized for social change for a longer time (Gurin, 1985). All of these factors are likely to enhance group consciousness among African-Americans.

D. SUMMARY Individuals who are objectively disadvantaged with regard to valued outcomes will believe that their entitlements have been violated only if they: (1) become aware that a discrepancy exists between their outcomes and those of others and (2) believe that the discrepancy is illegitimate. Three factors were discussed as enhancing the perception of illegitimacy: (1) personal or situational factors that cause a person to take a collective rather than a personal perspective on deprivation and disadvantage; (2) factors that enhance the perception that the procedures underlying the current distribution of outcomes are unfair (e.g., biased, inconsistent, prejudiced), and (3) personal ideologies and collective representations that locate the cause of disadvantage in external agents (e.g., the system) rather than in individual attributes.

VII. The Consequences of Entitlement Beliefs about entitlement are critical and proximal determinants of evaluative, affective, and behavioral responses to socially distributed outcomes. The bulk of research on the consequences of entitlement has assessed its implications for outcome evaluations and, in particular, outcome satisfaction. This research, some of which has already been reviewed in this paper, indicates that individuals

ENTITLEMENT AND SOCIAL INEQUALITY

339

who believe they are receiving the valued outcomes that they deserve are more satisfied with their outcomes than are those who believe they are receiving less than they deserve. Furthermore, perceived entitlement mediates the effects of antecedent factors, such as social comparisons and perceived contributions, on outcome satisfaction (Bylsma et al., 1992). The implications of entitlement for affect, self-evaluations, and behavior have been less frequently examined. Although a full discussion of the consequences of entitlement is well beyond the scope of this article, a few general implications are briefly discussed here.

A. ENTITLEMENT, AFFECT, AND SELF-EVALUATIONS Perceptions of entitlement and deserving are hypothesized to be an important determinant of how people respond affectively and/or emotionally to outcome distributions. According to equity theory, the experience of inequity is affectively distressing-receiving more than one deserves results in guilt, whereas receiving less than one deserves results in anger (Adams, 1965; Walster et al., 1978). The experience of relative deprivation is often operationalized as the emotion of anger, resentment, moral outrage, or righteous indignation (Crosby, 1976; Martin, 1981; Taylor & Moghaddam, 1987). Despite the emphasis on affect and emotion inherent in many justice theories, researchers have rarely examined the affective and/or emotional implications of beliefs about entitlement (Greenberg, 1984). Theoretically, when negative outcomes or treatments are believed to be deserved, people will feel sad, disappointed, or depressed, but will not feel anger or resentment. In contrast, when negative outcomes or treatment are believed to be undeserved, people will experience anger, resentment, and the particular constellation of negative emotional responses associated with relative deprivation (moral outrage, righteous indignation). The hypothesized links among beliefs about personal entitlement, affect, and emotional experience are consistent with cognitive theories of emotion (e.g., Shaver et al., 1987; Smith & Ellsworth, 1985; Weiner, 1985). These theories posit that the experience of specific emotions (e.g. anger, sadness) is associated with specific sets of cognitive antecedents. Shaver et al. (1987), for example, inferred from participants’ accounts of emotional experiences that the cognitive antecedents to the emotion of anger include the judgment that frustration, interruption, power reversal, or harm (usually caused by another person) is illegitimate, or contrary to what it ought to be. Shaver et al. (1987) note that the perception of illegitimacy was the most frequent feature of the anger prototype, occurring in 95% of accounts of the experience of anger. Although infrequently examined, beliefs about entitlement also have implica-

340

BRENDA MAJOR

tions for self-evaluations and self-esteem. Crocker and Major (1989; Major & Crocker, 1993) proposed that the attribution of negative outcomes or poor treatment to prejudice or bias short-circuits the belief that negative outcomes are deserved and buffers self-esteem and affect from negative outcomes. Data from several studies are consistent with this perspective. Crocker et al. (1991, Experiment l), for example, asked women to write an essay that the women believed would be subsequently evaluated by a male peer. Half of the women were led to believe that the evaluator had positive attitudes toward women, the other half were led to believe he held negative, prejudicial, attitudes toward women. Women later received either positive or negative feedback on their essay from the evaluator. As predicted, women who received negative feedback from an evaluator they believed was prejudiced were significantly less depressed than were women who received negative feedback from a nonprejudiced evaluator. Data from a second study (Crocker et al., 1991, Experiment 2) revealed that AfricanAmerican students who could attribute a negative interpersonal evaluation to prejudice (because the evaluator was aware of their race) tended to have higher self-esteem than did African-American students who could not make this attribution (because the evaluator did not know their race). In both of these cases, we would argue that the attribution to prejudice (a procedural injustice) weakened the perception that negative outcomes were deserved. This reduction of personal deserving attenuated the threats to self-esteem and affect inherent in negative outcomes (see Crocker & Major, 1993). Beliefs about personal deserving also moderate affective responses to positive outcomes or treatment (Major & Crocker, 1993). Positive outcomes that are believed to be deserved are likely to result in more positive affect, and in enhanced self-evaluations and self-esteem. In contrast, receiving positive outcomes that are believed to be undeserved is likely to attenuate positive affect, and perhaps result in negative affect, guilt, and negative self-evaluations. Crocker et al. (1991, Experiment 2) obtained findings consistent with this hypothesis. African-American students who received a favorable evaluation from a white student who was aware of their race were more likely to attribute the evaluation to prejudice, and were less likely to attribute it to their personality, than were African-American students who received the same favorable evaluation from a white student who was unaware of their race. Furthermore, compared to baseline levels of self-esteem, the former group showed a drop in self-esteem, whereas the latter group showed an increase in self-esteem. The difference between these two groups was significant. We have hypothesized that this pattern results from the attributional ambiguity that members of chronically stigmatized groups (such as African-Americans) experience about whether positive treatment or outcomes they receive are personally deserved (e.g., due to their own personal qualities or merits) or are due to

ENTITLEMENT AND SOCIAL INEQUALITY

34 1

their stigmatized group membership (Crocker et al., 1991; Major & Crocker, 1993). Positive outcomes might be seen as being caused by stigmatized group membership for several reasons, including pity or sympathy, a desire of others to avoid the appearance of prejudice, or might be seen as a consequence of special programs and policies targeted toward members of stigmatized or “protected” groups. The belief that positive outcomes are not personally deserved is likely to minimize or detract from the extent to which the stigmatized feel good about themselves or their accomplishments (Major & Crocker, 1993). Several studies on women’s reactions to preferential selection policies are consistent with this hypothesis. Compared to men or women who believe they were selected for a desirable position solely on the basis of merit, women who believe they were preferentially selected for a position solely on the basis of their sex subsequently evaluate their skills in that domain more harshly (Heilman, Simon, & Repper, 1987; Turner, Pratkanis, & Hardaway, 1991); show less interest in persisting in the position (Heilman et a]. , 1987); value the position less (Major, Feinstein, & Crocker, in press); and choose less challenging tasks for themselves (Heilman, Rivera, & Brett, 1991). Men in these studies generally were unaffected by their method of selection, or in some cases, evaluated their skills more favorably after being selected on the basis of sex than on merit. These gender differences in response to preferential selection appear to be due to men’s initially more favorable appraisals of their own abilities and/or the abilities of their gender (Heilman, Lucas, & Kaplow, 1990; Major, Feinstein, & Crocker, in press) in the domain for which they were selected. Men appeared more confident that they (or men in general) deserved the position because of their greater abilities. If someone is confident that he, or members of his group, deserves a position because of superior abilities, his evaluation of his own competence is unlikely to suffer as a result of being preferentially selected on the basis of a “nonability” attribute. In general, theory and research on the antecedents and consequencesof entitlement would benefit from greater cross-fertilization with theories of selfevaluation and self-esteem. Crocker and Major (1989), for example, reviewed a substantial amount of evidence indicating that contrary to popular belief and most theories of self-esteem, members of many stigmatized and disadvantaged groups (e.g., African-Americans, Latino- and Latina-Americans, physically disabled, homosexuals) do not have lower self-esteem than do members of more advantaged or dominant groups. We argued in that paper that three selfprotective strategies underlie the “paradoxical” high self-esteem among members of stigmatized and disadvantaged groups: ( 1) Members of disadvantaged groups devalue those dimensions or attributes on which they fare poorly and value more highly those dimensions or attributes on which they fare well relative to more advantaged groups; (2) members of disadvantaged groups compare within the

342

BRENDA MAJOR

ingroup, and hence avoid comparisons with more advantaged groups, and (3) members of disadvantaged groups attribute negative outcomes and treatment they receive to prejudice and discrimination based on their membership in a stigmatized group, and hence buffer their self-esteem from the negative implications of this feedback. We further hypothesized that the more disadvantaged individuals think of themselves as a member of a stigmatized group (i.e., the more they adopt a collective perspective), the more likely they are to use these three self-protective strategies. The three central processes underlying Crocker and Major’s analysis of self-esteem among the stigmatized-values, comparison standards, and attributions for treatment-are very similar to the processes highlighted in this paper and in theories of social justice as responsible for the paradoxical contentment of the disadvantaged. Two key differences between these perspectives are the more explicit assumptions of Crocker and Major (1989) that (1) the valuing, comparison, and attributional strategies are motivated, at least in part, by the desire of the disadvantaged to maintain or protect their self-esteem from threat, and that (2) these processes have implications for the self-esteem of members of disadvantaged and stigmatized groups. Attention to social justice concerns is also likely to enrich theories of selfesteem. In their theory of self-esteem maintenance, for example, Tesser and Campbell (1983) propose that poor performance relative to a close (similar) comparison other on a self-relevant (valued) dimension produces an “esteemthreat” that is accompanied by negative affect. Likewise, equity theory (e.g., Walster et al., 1978) proposes that poor outcomes relative to a close (similadwithin reference group) comparison other when the outcome is important (valued) produces a “justice threat” (e.g., inequity distress). Furthermore, both theories propose that comparers are motivated to resolve the threat (i.e., to maintain self-esteem or restore justice) through a variety of cognitive and behavioral mechanisms (e.g., changing comparison other, distancing oneself from comparison other, devaluing the domain, altering inputs or outcomes). A key difference between these perspectives is the attributional ambiguity or clarity surrounding the reason for the “threat,” that is, the comparatively low outcomes or poor performances. Tesser and Campbell’s ( 1983) self-esteem maintenance theory primarily addresses personally produced performances where there is little ambiguity as to who was responsible for one’s performance. Hence, the implications of this theory focus almost exclusively on self-directed affect and selfevaluations. In contrast, theories of social justice deal with socially distributed outcomes where there is potential ambiguity as to whether oneself or another person or system is responsible for one’s poor outcomes or treatment. As noted above, however, most social justice research has focused on outcome-directed affect and outcome evaluations. Recognition of the attributional ambiguity inherent in social interactions suggests that both theoretical perspectives would benefit from expanding their horizons to consider a broader array of affective reactions,

ENTITLEMENT AND SOCIAL INEQUALITY

343

including affect directed toward the self, outcomes, others, and the system (cf. Mark & Folger, 1984).

B. ENTITLEMENT AND BEHAVIOR Beliefs about entitlement are also associated with behavioral responses to outcome distributions. The perception of violated entitlement is thought to be an important motivator of attempts to seek change. This theoretical link between violated entitlement and motivation to seek change is consistent with research on the behavioral implications of emotions. Shaver et al. (1987), for example, found that unlike the experiencing of other negative emotions (e.g., sadness, fear), reports of the experiencing of anger were accompanied by reports of becoming stronger, or higher in potency, and reports of becoming energized in order to fight or rail against the cause of anger. According to Shaver et al. (1987), the angry person’s responses “seem designed to rectify injustice-to reassert power or status, to frighten the offending person into compliance, to restore a desired state of affairs” (p. 1078). The link between entitlement and behavior is a weak spot of theory and research on social justice (cf. Crosby, 1976; Martin, 1986a). Behavior is multiply determined. As a consequence, although perceptions of violated entitlement are strongly linked to the motivation to seek change, this motivation may or may not translate into change-oriented behaviors. Whether it does depends on situational constraints, pragmatic considerations, personal dispositions, and so forth (cf. Martin, 1986a). When motivation does translate into behavior, attempts to seek change may be directed either individually (e.g., at one’s personal situation) or collectively (e.g., at the situation of one’s group as a whole), and may be negative (e.g., alcoholism, rioting) or positive (self-improvement efforts; working for change within the limits of the law) (Crosby, 1976; Mark & Folger, 1984). Recent research suggests that efforts will be directed at self-improvement unless access to higher status groups is completely blocked (Wright, Taylor & Moghaddam, 1990). The behavioral consequences of entitlement beliefs is an area sorely in need of further research. One consequence of beliefs about entitlement is that they may create their own self-fulfilling prophecies. Those who feel entitled to less may ask for less, and as a consequence receive less, than those who feel entitled to more. A study by Major, Vanderslice, and McFarlin (1984) illustrates how this may occur. We gave management students the credentials of a job applicant and asked them to evaluate the applicant’s qualifications, make a hiring decision, and recommend a starting salary for the applicant. The only factors that varied were the applicant’s sex and the applicant’s salary expectations. Depending on condition, applicants indicated one of five different expected salary levels on their r6sumCs. These

344

BRENDA MAJOR

ranged from slightly below the specified salary range for the job to well above the specified salary range. Results were clear. Even though all applicants were rated as equally qualified, and were equally as likely to be hired, applicants who asked for more money were offered more money. This was true irrespective of the sex of the applicant. Thus, beliefs about entitlement can be self-perpetuating.

VIII. Conclusions One goal of this article was to illustrate the importance of entitlement as an explanatory construct in understanding how members of different social groups react to their socially distributed outcomes. I argued that the paradoxical contentment of some objectively disadvantaged groups, such as working women, can be understood by considering how social inequality produces group differences in personal entitlement. Social inequality was proposed to lead to group differences in entitlement via the mediational processes of intragroup and intrapersonal social comparisons, and attributions and beliefs that legitimize the status quo. Social comparison biases tend to prevent awareness of disadvantage, whereas attribution biases tend to legitimize disadvantage. Individuals who are objectively disadvantaged of valued outcomes will believe that their entitlements have been violated only if they become aware of a discrepancy between their outcomes and those of others and believe that the discrepancy is illegitimate. Factors such as perceptions of procedural injustices, personal and collective ideologies that challenge the legitimacy of the status quo, and that shift people’s perspective from personal entitlement to collective entitlement, increase awareness of disadvantage and enhance perceptions of illegitimacy. Disadvantaged individuals and groups who feel they are not receiving the valued outcomes or treatment that they or their group deserve are most likely to be dissatisfied, experience the emotions of anger and resentment, and be motivated to seek change. Although this theoretical framework is meant. to be broadly applicable to understanding how members of various social groups react evaluatively, affectively, and behaviorally to their outcomes, several potential limitations to the generality of this theoretical perspective deserve comment. One issue concerns whether and to what extent the theoretical framework developed here generalizes across different types of outcomes. People feel entitled to many different types of things, in many different types of contexts. Feelings of entitlement, for example, may be concerned with equal access to housing, education, and political influence, with rewards like honors, roles, and job assignments, and to treatment by parents and lovers. Most research on entitlement, including my own, has exam-

ENTITLEMENT AND SOCIAL INEQUALITY

345

ined feelings of entitlement to tangible rewards or outcomes, such as money, that are easily compared across individuals. The theoretical framework developed here may not generalize well to less tangible types of outcomes, or to outcomes that are less amenable to social comparison. A second issue concerns the generalizability of the theoretical framework developed here to groups other than gender and race. Gender and race are important, socially salient categories. They are the focus of much individual and societal attention, and these groups are the targets of strongly held stereotypes. The impact of intragroup versus intergroup social comparison processes and legitimizing attributions may be especially pronounced for socially salient categories of this type. An important question for future research is how comparison processes and legitimacy appraisals affect beliefs about entitlement among people who are disadvantaged because of other types of social characteristics (e.g., sexual preference, disabilities). Intragroup social comparison biases may be less pronounced among members of less salient groups, unless group members have developed a strong sense of group identity. Legitimizing attributions, in contrast, may be more common among members of such groups, especially if it is believed that one’s disadvantage is the result of a controllable factor (e.g., overweight). A third issue concerns the extent to which the entitlement framework developed here generalizes to nonvalued outcomes. This article, as well as most psychological theory and research on the topic, has focused on antecedents and consequences of feelings of entitlement to valued outcomes or rewards (cf. Adams, 1965; Crosby, 1982). Feelings of entitlement, social comparisons, and legitimacy appraisals are also likely to be important processes in how people react to distributions of negative outcomes such as punishments. Punishments are likely to be accepted and accompanied by sadness and depression if they are perceived as deserved. Punishments are likely to lead to moral outrage, righteous indignation, and motivation to seek redress, in contrast, if they are perceived as undeserved. Social comparisons and appraisals of the legitimacy of the distribution of negative outcomes are also apt to be critical determinants of whether punishments are believed to be deserved. Finally, the generalizability of the theoretical framework outlined here to domains other than economic, such as family and political domains (e.g., J. L. Hochschild, 1981) warrants attention. Most psychological theory and research on entitlement has focused on exchange relationships (e.g., relationships between acquaintances or people doing business with each other) rather than on communal relationships (e.g., relationships with friends or family members) (cf. Clark & Mills, 1979). Some scholars have argued that social comparison processes and entitlement issues are muted or irrelevant in intimate relationships (e.g., Prentice & Crosby, 1987). Other scholars, in contrast, have argued that

346

BRENDA MAJOR

comparison, justification, and entitlement processes are also important in understanding the dynamics of intimate relationships (e.g., Ferree, 1991; Major, 1993; Thompson, 1991). I have argued, for example, that women and men have a different sense of personal entitlement with regard to what they should put into and receive from marriage in the domain of family work (Major, 1993). These gender differences in entitlement are hypothesized to result from societal norms regarding women’s and men’s roles within the family, intrapersonal and interpersonal comparison processes that lead women and men to have different reference standards for household labor, and attributions and justifications that legitimize an unequal division of family labor. Gender differences in entitlement, in turn, are hypothesized to underlie the finding that despite contributing a disproportionate share of the unpaid labor of the family (e.g., housework and child care) compared to their husbands, employed wives nonetheless report relative contentment with this unequal distribution (see Major, 1993, for a more complete discussion). A second goal of this article was to organize into a systematic framework current knowledge about the psychological antecedents and consequences of beliefs about entitlement. Entitlement is a core issue in social conflict and public discourse on social policy. Many social conflicts center on who is entitled to what, and on what basis. Controversy over affirmative action, for example, centers on whether ascribed characteristics (e.g., race, sex) are legitimate preconditions to entitlement to jobs, promotions, and other valued outcomes. Debates over welfare versus workfare hinge on whether people should have to contribute work in order to be entitled to benefits. Attempts to cut any program that has come to be regarded as an “entitlement program,” such as welfare, public education, or social security, are met with public outcry and resistance. Despite the centrality of entitlement in public discourse and theories of social justice, however, the psychology of entitlement is still in its infancy. Development of a comprehensive psychology of entitlement has been hampered by the complexity of the construct of entitlement and the breadth of domains to which it can be applied. It has also been hindered by the narrow focus of much of the research and theory on this topic. Given the importance of the construct of entitlement to both personal well-being and collective behavior, it is time for social psychologists to take a broader look at this construct.

Acknowledgments Preparation of this manuscript was supported by National Science Foundation grant BNS 9010487. I am deeply grateful to Wayne Bylsma, Jennifer Crocker, Kay Deaux, Carol Miller, Dean h i t t , Harry Reis, and Mark Zanna for their insightful comments on an earlier draft of this manuscript.

ENTITLEMENT AND SOCIAL INEQUALITY

347

References Adams, J. S . (1965). Inequality in social exchange. Advances in Experimental Social Psychology, 2, 267-299. Alloy, L. B., & Abramson, L. Y. (1979). Judgment of contingency in depressed and nondepressed students: Sadder but wiser? Journal of Experimental Psychology: General, 108, 441-485. Andrews, F. M., & Withey, S. B. (1976). Social indicators of well-being. New York: Plenum. Atkinson, M. L. (1986). The perception of social categories: Implications for the social comparison process. In J. M. Olson, C. P. Herman, & M. P. Zanna (Eds.), Relative deprivation and social comparison: The Ontario Symposium (Vol. 4, pp. 117-134). Hillsdale, NJ: Erlbaum. Austin, W. (1977). Equity theory and social comparison processes. In J. Suls, & R. Miller (Eds.), Social comparison processes: Theoretical and empirical perspectives. Washington, DC: Hemisphere. Austin, W., McGinn, N. C., & Susmilch, C. (1980). Internal standards revisited: Effects of social comparisons and expectancies on judgments of satisfaction. Journal of Experimental Social Psychology, 16, 426-441. Berger, J., Zelditch, M., Jr., Anderson, B., & Cohen, B. P. (1972). Structural aspects of distributive justice: A status value formulation. In J. Berger, M. Zelditch, Jr., & B. Anderson (Eds.), Sociological theories in progress (Vol. 2 , pp. 119- 146). Boston: Houghton Mimin. Berk, S. F. (1985). The gender factory: The apportionment of work in American households. New York: Plenum. Bielby, W. T., & Baron, J. N. (1984). A woman's place is with other women: Sex segregation within organizations. In B. F. Reskin (Ed.), Sex segregation in the workplace: Trends. explanations, remedies (pp. 27-55). Washington, DC: National Academy Press. Biernat, M., Manis, M., & Nelson, T. E. (1991). Stereotypes and standards ofjudgment. Journal of Personality and Social Psychology, 60, 485-499. Biernat, M., & Wortman, C. B. (1991). Sharing of home responsibilities between professionally employed women and their husbands. Journal of Personality and Social Psychology, 60, 844860. Blau, F. D. (1984). Occupational segregation and labor market discrimination. In B. F. Reskin (Ed.), Sex segregation in the workplace: Trends, explanations. remedies (pp. 117- 143). Washington, DC: National Academy Press. Boulgarides, J. D. (1984). A comparison of male and female business managers. Leadership and Organizational Development Journal, 5 , 27-31. Brickman, P., & Bulman, R. J. (1977). Pleasure and pain in social comparison. In J. M. Suls & R. L. Miller (Eds.), Social comparison processes: Theoretical and empirical perspectives (pp, 149186). Washington, DC: Hemisphere. Bulman, R. J., & Wortman, C. B. (1977). Attributions of blame and coping in the "real world": Severe accident victims react to their lot. Journal of Personality and Social Psychology, 35, 35 1-363. Bylsma, W. H., & Major, B. (1992a). Social comparisons and the perpetuation of disadvantage. Unpublished manuscript, State University of New York at Buffalo, Department of Psychology. Bylsma, W. H., & Major, B. (1992b). Two routes to eliminating gender differences in personal entitlement: Social comparisons and performance evaluations. Psychology of Women Quarterly, 16, 193-200. Bylsma, W. H . , Major, B., & Cozzarelli, C. (1992). The role of social comparisons andprocedural fairness in the perpetuation ofdisadvantage. Unpublished manuscript, State University of New York at Buffalo, Department of Psychology. Callahan-Levy, C. M., & Messe, L. A. (1979). Sex differences in the allocation of pay. Journal of Personality and Social Psychology, 37. 443-446.

348

BRENDA MAJOR

Campbell, A., Converse, P. E., & Rodgers, W. L. (1976). The quality ofAmerican life: Perceptions. evaluations, and satisfactions. New York: Russell Sage Foundation. Clark, M. S., & Mills, J. (1979). Interpersonal attraction in exchange and communal relationships. Journal of Personality and Social Psychology, 37, 12-24. Clayton, S. D., & Crosby, F. J. (1992). Justice, gender, and afirmarive action. Ann Arbor, MI: University of Michigan Press. Cohen, R. L. (1982). Perceiving justice: An attributional perspective. In J. Greenberg & R. L. Cohen (Eds.), Equity and justice in social behavior (pp. 119-160). New York: Academic Press. Cohen, R. L. (1986). Power and justice in intergroup relations. In H. W. Bierhoff, R. L. Cohen, & J. Greenberg (Eds.), Justice in social relations. New York: Plenum. Comer, R., & Laird, J. D. (1975). Choosing to suffer as a consequence of expecting to suffer: Why do people do it? Journal of Personality and Social Psychology, 32, 92-101. Cook, T. D., Crosby, F., & Hennigan, K. M. (1977). The construct validity of relative deprivation. In J. M. Suls & R. L. Miller (Eds.),Social comparison processes: Theoretical and empirical perspecrives (pp. 307-333). Washington, DC: Hemisphere. Crocker, J., Luhtanen, R., Blaine, B., & Broadnax, S . (1993). Judgments of personal and group disadvantage and discrimination:African-American,Asian, and white college students evaluate their lot. Unpublished manuscript. Crocker, J., & Major, B. (1989). Social stigma and self-esteem: The self-protective properties of stigma. Psychological Review, 96, 608-630. Crocker, J., & Major, B. (1993). Reactions to stigma: The moderating role of justifications. In M. Zanna & J. Olson (Eds.), The psychology of prejudice: The Ontario Symposium. (Vol. 7, pp. 289-314). Hillsdale, NJ: Erlbaum. Crocker, J., Voelkl, K.,Testa, M., &Major, B. (1991). Social stigma: The affective consequences of attributional ambiguity. Journal of Personality and Social Psychology, 60, 21 8-228. Crosby, F. (1976). A model of egotistical relative deprivation. Psychological Review, 83, 85-1 13. Crosby, F. (1982). Relative deprivation and working women. New York: Oxford University Press. Crosby, F. (1984). Relative deprivation in organizational settings. Research in Organizational Behavior, 6 , 51-93. Crosby, F., Bromley, S . , & Saxe, L. (1980). Recent unobtrusive studies of black and white discrimination and prejudice: A literature review. Psychological Bulletin, 87, 546-563. Crosby, F. J., Clayton, S. D., Hemker, K., & Alksnis, 0. (1986). Cognitive biases in the perception of discrimination: The importance of format. Sex Roles, 14. 637-646. Crosby, F. J., Pufall, A., Snyder, R. C., O’Connell, M., & Whalen, P. (1989). The denial of personal disadvantage among you, me, and all the other ostriches. In M. Crawford & M. Gentry (Eds.), Gender’s thought: Psychological perspectives (pp. 79-99). New York: Springer-Verlag. Davis, J. A. (1959). A formal interpretation of the theory of relative deprivation. Sociometry, 22, 280-296. Deaux, K. (1984). From individual differences to social categories: Analysis of a decades’ research on gender. American Psychologist, 39, 105-1 16. Della Fave, L. R. (1980). The meek shall not inherit the earth: Self-evaluation and the legitimacy of social stratification. American Sociological Review, 45, 955-971. Desmarais, S. (1993). Gender, income history, and perceived income entitlements. Unpublished doctoral dissertation, University of Waterloo, Waterloo, Ontario, Canada. Desmarais, S., & Curtis, J. (1991). Gender differences in perceived income entitlements. Paper presented at the International Conference on Social Justice Research, Utrecht, the Netherlands. Deutsch, M. (1985). Distributivejustice: A social-psychological perspective. New Haven, CT: Yale University Press. Diener, E. (1984). Subjective well-being. Psychological Bulletin, 95, 542-575. Dion, K. L. (1986). Responses to perceived discrimination and relative deprivation. In J. M. Olson,

ENTITLEMENT AND SOCIAL INEQUALITY

349

C. P. Herman, & M. P. Zanna (Eds.), Relative deprivation and social comparison: The Ontario Symposium (Vol. 4, pp. 159-180). Hillsdale, NJ: Erlbaum. Dovidio, 1. F., & Gaertner, S. L. (Eds.) (1986). Prejudice, discrimination, and racism. San Diego, CA: Academic Press. Dreher, G. F. (1981). Predicting the pay satisfaction of exempt employees. Personnel Psychology, 34, 579-589. Dreher, G . F., Dougherty, T. W., & Whitely, W. (1989). Influence tactics and salary attainment: A gender-specific analysis. Sex Roles, 20. 535-550. DuM, L., & Guimond, S. (1986). Relative deprivation and social protest: The personal group issue. In J. M. Olson, C. P. Herman, & M. P.Zanna (Eds.), Relative deprivation and social comparison: The Ontario Symposium (Vol. 4, pp. 201-216). Hillsdale, NJ: Erlbaum. Eagly, A. H. (1987). Sex differences in social behavior: A social-role interpretation. Hillsdale, NJ: Erlbaum. Feagin, J. R. (1975). Subordinating the poor: Welfare and American beliefs. Englewood Cliffs, NJ: Prentice-Hall. Ferree, M. M. (1990). Beyond separate spheres: Feminism and family research. Journal of Marriage and the Family, 52. 866-884. Ferree, M. M. (1991). Gender, conflict and change: Family roles in biographical perspective. In W. Heinz (Ed.), Theoretical advances in life course research (pp. 144- 161). Weinheim, FRG: Deutscher Studien Verlag. Festinger, L. (1954). A theory of social comparison processes. Human Relations, 7, 117-140. Folger, R. (1986). Rethinking equity theory: A referent cognitions model. In H. W. Bierhoff, R. L. Cohen, & J. Greenberg (Eds.), Justice in social relations. New York: Plenum. Folger, R. (1987). Reformulating the preconditions of resentment: A referent cognitions model. In J. C. Masters & W. P. Smith (Eds.), Social comparison, social justice, and relative deprivation: Theoretical, empirical, and policy perspectives (pp. 183-2 15). Hillsdale, NJ: Erlbaum. Frieze, I. H., Olson, J. E., &Good, D. C. (1990). Perceived and actual discrimination in the salaries of male and female managers. Journal of Applied Social Psychology, 20, 46-67. Frieze, 1. H., Whitley, B. E., Hartman Hanusa, B., & McHugh, M. C. (1982). Assessing the theoretical models for sex differences in causal attributions for success and failure. Sex Roles, 8, 333-343. Furby, L. (1986). Psychology and justice. In R. L. Cohen (Ed.), Justice: Views from the social sciences (pp. 153-203). New York: Plenum. Gilligan, C. (1982). In a direrent voice: Psychological theory and women's development. Cambridge, MA: Harvard University Press. Godley, S. H., Hafer, M., Riggar, T. F., & Maki, D. R. (1983, February). Gender and job function: Differences in salary, educational levels, and experience. Journal of Rehabilitation Administration, pp. 22-26. Goethals, G. R., & Darley, J. (1977). Social comparison theory: An attributional approach. In J. M. Suls & R. L. Miller (Eds.), Social comparison processes: Theoretical and empirical perspectives (pp. 259-278). Washington, DC: Hemisphere. Goodman, P. S. (1974). An examination of referents used in the evaluation of pay. Organizational Behavior and Human Performance. 12, 170-195. Greenberg, J. (1984). On the apocryphal nature of inequity distress. In R. Folger (Ed.), The sense of injustice: Social psychological perspectives. New York: Plenum. Greenberg, J., Pyszczynski, T., & Solomon, S . (1987). The causes and consequences of a need for self-esteem: A terror management theory. In R. Baumeister (Ed.), Public selfand private self. New York: Springer-Verlag. Gurin, P. (1985). Women's gender consciousness. Public Opinion Quarterly, 49, 143- 163. Gurin, P., & Epps, E. (1975). Black consciousness, identity, and achievement. New York: Wiley.

350

BRENDA MAJOR

Gurin, P., & Townsend, A. (1986). Properties of gender identity and their implications for gender consciousness. British Journal of Social Psychology, 25, 139- 148. Gurr, T. R. (1970). Why men rebel. Princeton, NJ: Princeton University Press. Hafer, C. L., & Olson, J. M. (1989). Beliefs in a just world and reactions to personal deprivation. Journal of Personality, 57, 799-823. Heider, F. (1958). The psychology of interpersonal relations. New York: Wiley. Heilman, M. E., Lucas, J. A., & Kaplow, S. R. (1990). Self-derogating consequences of sex-based preferential selection: The moderating role of initial self-confidence. Organizational Behavior gnd Human Decision Processes, 46, 202-216. Heilman, M. E., Rivera, J. C., & Brett, J. (1991). Skirting the competence issue: The effects of sexbased preferential selection on the task choices of women and men. Journal of Applied Psychology, 76, 99-105. Heilman, M. E., Simon, M. C., & Repper, D. P. (1987). Intentionally favored, unintentionally harmed? Impact of sex-based preferential selection on self-perceptions and self-evaluations. Journal of Applied Psychology. 72. 62-68. Hochschild, A. (with Machung, A). (1989). The second shift. New York: Avon Books. Hochschild, J. L. (1981). What’sfair? American beliefs abour distributive justice. Cambridge, MA: Harvard University Press. Hoiberg, A. (1982). Women and rhe world of work. New York: Plenum Press. Homans, G. (1974). Social behavior: Its elemenrary forms (rev. ed.). New York: Harcourt Brace Jovanovich. Hyman, H. H., & Singer, E. (Eds.). (1968). Readings in reference group rheory and research. New York: Free Press. Jackson, J. S., Chatters, L. M.,& Neighbors, H. W. (1986). The subjective life quality of Black Americans. In F. M. Andrews (Ed.), Research on rhe quality of life (pp. 193-213). Ann Arbor, MI: Institute for Social Research. Jackson, L. A,, Gardner, P. D., & Sullivan, L. A. (1992). Explaining gender differences in self-pay expectations: Social comparison standards and perceptions of fair pay. Journal of Applied Psychology, 77, 65 1-663. Jellison, J. M., &Green, J. (1981). A self-presentation approach to the fundamental attribution error: The norm of internality. Journal of Personality and Social Psychology. 40, 643-649. Jones, J. M. (1986). Racism: A cultural analysis of the problem. In J. F. Dovidio & S. L. Gaertner (Eds.), Prejudice. discrimination, and racism (pp. 279-314). San Diego, CA: Academic Press. Kahn, A. S., & Gaeddert, W. P. (1985). From theories of equity to theories of justice: The liberating consequences of studying women. In V. E. O’Leary,R. K. Unger, & B. S. Wallston (Eds.), Women, gender, and social psychology (pp. 129- 148). Hillsdale, NJ: Erlbaum. Kelley, H. H. (1952). Attitudes and judgments as influenced by reference groups: W o functions of reference groups. In G. E. Swanson, T. M. Newcomb, & E. L. Hartley (Eds.), Readings in socialpsychology (rev. ed., pp. 410-414). New York: Holt & Co. Kessler-Harris, A. (1982). Our to work: A history of wage-earning women in the United Stares. New York: Oxford University Press. Kluegel, J. R., & Smith, E. R. (1986). Beliefs abour inequality: Americans’ view of whar is and what oughr to be. Hawthorne, NJ: Aldine de Gruyter. Langer, E. J. (1977). The psychology of chance. Journal for rhe Theory of Social Behavior, 7. 185208.

Lenney, E. (1977). Women’s self-confidence in achievement settings. Psychological Bulletin, 84, 113.

Lerner, M. J. (1975). The justice motive in social behavior: An introduction. Journal of Social Issues, 31, 1-19. Lerner, M. J. (1980). The belief in a jusr world: A fundamental delusion. New York: Plenum Press. Lerner, M. J. (1987). Integrating societal and psychological rules of entitlement: The basic task of

ENTITLEMENT AND SOCIAL INEQUALITY

35 1

each social actor and fundamental problem for the social sciences. Social Justice Research, I , 107- 125. Lerner, M. J., & Miller, D. T. (1978). Just world research and the attribution process: Looking back and ahead. Psychological Bulletin, 85, 1030- 1051 . Lemer, M. J., Miller, D. T., & Holmes, J. G. (1976). Deserving and the emergence of forms of justice. Advances in Experimental Social Psychology, 3 , 133- 162. Leventhal, G. S. (1976). Fairness in social relationships. In J. W. Thibaut, J. T. Spence, & R. C. Carson (Eds.), Contemporary topics in social psychology (pp. 21 1-239). Morristown, NJ: General Learning Press. Levine, J. M., & Moreland, R. L. (1987). Social comparison and outcome evaluation in group contexts. In J. C. Masters & W. P. Smith (Eds.), Social comparison, social justice, and relative deprivation: Theoretical, empirical. and policy perspectives. Hillsdale, NJ: Erlbaum. Lind, E. A,, & Qler, T. R. (1988). The social psychology ofprocedural justice. New York: Plenum Press. Major, B. (1987). Gender, justice and the psychology of entitlement. In P. Shaver & C. Hendricks (Eds.), Review ofpersonality and socialpsychology (Vol. 7, pp. 124-148). Newbury Park, CA: Sage. Major, B. (1989). Gender differences in comparisons and entitlement: Implications for comparable worth. Journal of Social Issues, 45, 99-1 15. Major, B. (1993). Gender, entitlement, and the distribution of family labor. Journal of Social Issues, 49, 141-159. Major, B., & Adams, J. B. (1983). Role of gender, interpersonal orientations, and self-presentation in distributive-justice behavior. Journal of Personality and Social Psychology, 3 , 598-608. Major, B., & Crocker, J. (1993). Social stigma: The consequences of attributional ambiguity. In D. M. Mackie & D. L. Hamilton (Eds.), Affect, cognition, and stereotyping: Interactive processes in group perception (pp. 345-370). San Diego: Academic Press. Major, B., & Deaux, K. (1982). Individual differences in justice behavior. In J. Greenberg & R. L. Cohen (Eds.), Equity and justice in social behavior (pp. 43-76). New York: Academic Press. Major, B., Feinstein, J., & Crocker, J. (in press). The attributional ambiguity of affirmative action. In Basic and Applied Social Psychology. Major, B., & Forcey, B. (1985). Social comparisons and pay evaluations: Preferences for same-sex and same-job wage comparisons. Journal of Experimental Social Psychology. 21, 393-405. Major, B., & Konar, E. (1984). An investigation of sex differences in pay expectations and their possible causes. Academy of Management Journal. 27, 777-792. Major, B., McFarlin, D., & Gagnon, D. (1984). Overworked and underpaid. Journal of Personality and Social Psychology, 47, 1399-1412. Major, B., Osborne, J., and Crocker, J. (1992). Gender differences and self-esteem in response to positive and negative outcomes. Unpublished manuscript, State University of New York at Buffalo, Department of Psychology. Major, B., Romero, O., CastaAo, D., Barbera, E., Martinez-Benlloch, I., & Pastore, R. (1993). Gender, culture, and distributive culture: Reconciling need vs. equity. Manuscript submitted for publication. Major, B. Schiacchitano, A. M., & Crocker, J. (in press). Ingroup vs. out-group comparisons and self-esteem. Personality and Social Psychology Bulletin. Major, B., & Testa, M. (1989). Social comparison processes and judgments of entitlement and satisfaction. Journal of Experimental Social Psychology, 25, 101- 120. Major, B., Testa, M., and Bylsma, W. H. (1991). Responses to upward and downward social comparisons: The impact of esteem-relevance and perceived control. In J. Suls & T. Wills (Eds.), Social comparison: Contemporary theory and research (pp. 237-260). Hillsdale, NJ: Erlbaum.

352

BRENDA MAJOR

Major, B., Vanderslice, V., & McFarlin, D. B. (1984). Effects of pay expected on pay received: The confirmatory nature of initial expectations. Journal of Applied Social Psychology, 14, 399-4 12. Marini, M. M. (1989). Sex differences in earnings in the United States. Annual Review of Sociology, 15, 343-380. Mark, M. A., & Folger, F. (1984). Responses to relative deprivation: A conceptual framework. In P. Shaver & C. Hendrick (Eds.), Review ofpersonaliry and socialpsychology (Vol. 5 , pp. 217242). Beverly Hills, CA: Sage. Markus, H. & Oyserman, D. (1989). Gender and thought: The role of self concept. In M. Crawford & M. Gentry (Eds.), Gender and thought (pp. 100-127). New York: Springer-Verlag. Martin, J. (1981). Relative deprivation: A theory of distributive injustice for an era of shrinking resources. Research in Organizational Behavior, 3 , 53- 107. Martin, J. (1986a). The tolerance of injustice. In J. M. Olson, C. P. Herman, & M. P. Zanna (Eds.), Relative deprivation and social comparison: The Ontario Symposium (Vol. 4, pp. 217-242). Hillsdale, NJ: Erlbaum. Martin, J. (1986b). When expectations and justice do not coincide: Blue-collar visions of a just world. In H. W. Bierhoff, R. L. Cohen, & J. Greenberg (Eds.), Justice in social relations (pp. 317-335). New York: Plenum. Martin, J., Scully, M., & Levitt, B. (1990). Injustice and the legitimation of revolution: Damning the past, excusing the present, and neglecting the future. Journal of Personality and Social Psychology, 59, 281-290. McArthur, L. Z . , & Obrant, S. W. (1986). Sex biases in comparable worth analyses. Journal of Applied Social Psychology, 16, 757-770. McFarlin, D. B., Frone, M., Major, B., & Konar, E. (1989). Predicting career-entry pay expectations: The role of gender-based comparisons. Journal of Business and Psychology, 3 , 331-340. Merton, R. K. (1957). Social theory and social structure. Glencoe, IL: Free Press. Messick, D. M., & Sentis, K. (1983). Fairness, preference, and fairness biases. In D. M. Messick & K. S. Cook (Eds.), Equity theory: Psychological and sociological perspectives. New York: Praeger. Miller, A. H., Gurin, P., Gurin, G., & Malanchuk, 0. (1981). Group consciousness and political participation. American Journal of Political Science, 25, 494-5 1 1. Moore, D. (1990). Discrimination and deprivation: The effects of social comparisons. Social Justice Research, 4 , 49-64. Nacoste, R. B. (1990). Sources of stigma: Analyzing the psychology of affirmative action. Law and Policy, 12, 175-195. Nisbett, R. E., & Ross, L. (1980). Human inference: Strategies and shortcomings of social judgment. Englewood Cliffs, NJ: Prentice-Hall. Oldham, G. R., Nottenburg, G., Kassner, M. W.,Ferris, G., Fedor, D., & Masters, M.(1982). The selection and consequences of job comparisons. Organizational Behavior and Human Performance, 29, 84- I 1 1. O'Neill, J. (1985). Role differentiation and the gender gap in wage rates. In L. Lanvood, A. H. Stromberg, & B. A. Gutek (Eds.), Women and work: An annual review (Vol. 1, pp. 50-75). Beverly Hills, CA: Sage. Patchen, M. (1961). The choice of wage comparisons. Englewood Cliffs, NJ: Prentice-Hall. Pettigrew, T. F. (1967). Social evaluation theory: Convergences and applications. In D. Levine (Ed.), Nebraska symposium on motivation (Vol. 15, pp. 241-31 1). Lincoln: University of Nebraska Press.

Pleck J. (1985). Working wives/working husbands. Beverly Hills, CA: Sage. Porter, J. R., & Washington, R. E. (1979). Black identity and self-esteem: A review of studies of black self-concept, 1968- 1978. Annual Review of Sociology, 5 , 53-74. Prentice, D. A., & Crosby, F. (1987). The importance of context for assessing deservingness. In J. C.

ENTITLEMENT AND SOCIAL INEQUALITY

353

Masters & W. P. Smith (Eds.), Social comparison, social justice, and relative deprivation: Theoretical, empirical, and policy perspectives (pp. 165- 182). Hillsdale, NJ: Erlbaum. Pritchard, R. D. (1969). Equity theory: A review and critique. Organizational Behavior and Human Performance, 4 , 176-21 1. Reis, H. T. (1987). The nature of the justice motive: Some thoughts on operation, internalization, and justification. In J. C. Masters & W. P. Smith (Eds.), Social comparison, social justice, and relative deprivation: Theoretical, empirical, and policy perspectives (pp. 13 1- 150). Hillsdale, NJ: Erlbaum. Roberts, T. A. (1991). Gender and the influence of evaluations on self-assessments in achievement settings. Psychological Bulletin, 109. 297-308. Roberts, T., & Nolen-Hoeksema, S. (1989). Sex differences in reactions to evaluative feedback. Sex Roles, 21, 725-747. Runciman, W. G . (1966). Relative deprivation and social justice: A study of the attitudes to social inequality in 20th century England. Berkeley: University of California Press. Sampson, E. E. (1975). Justice as equality. Journal of Socia[ Issues, 31, 45-61. Sauser, W. I., & York, M. (1978). Sex differences in job satisfaction: A reexamination. Personnel Psychology, 31, 537-547. Schuman, H., Steeh, C., & Bobo, L. (1985). Racial attitudes in America: Trends and interpretations. Cambridge, MA: Harvard University Press. Schwartz, F. N. (1989). Management women and the new facts of life. HarvardBusinessReview, 89, 65-76.

Sears, D. O., & Funk, C. L. (1991). The role of self-interest in social and political attitudes. Advances in Experimental Social Psychology, 24, 1-9 1. Shaver, P. Schwartz, J., Kmon, D., & OConnor, C. (1987). Emotion knowledge: Further explorations of a prototype approach. Journal of Personality and Social Psychology, 52, 1061-1086. Shepelak, N. J., & Alwin, D. F. (1986). Beliefs about inequality and perceptions of distributive justice. American Sociological Review, 51, 30-46). Sidanius, J., & Pratto, F. (1993). The inevitability of oppression and the dynamics of social dominance. In P. Sniderman, P. E. Tetlock & E. G. Carmines (Eds.), Prejudice, politics, and race in America today (pp. 87- 103). Stanford, CA: Stanford University Press. Sigelman, L., Milward, H. B., & Shepard, I. M. (1982). The salary differential between male and female administrators: Equal pay for equal work? Academy of Management Journal, 25, 664671.

Singer, E. (1981). Reference groups and social evaluations. In M. Rosenberg & R. H. lbrner (Eds.), Social psychology: Sociological perspectives (pp. 66-93). New York: Basic Books. Smith, C. A,, & Ellsworth, P. C. (1985). Patterns of cognitive appraisal in emotion. Journal of Personality and Social Psychology, 48, 813-838. Spitze, G . (1988). Women's employment and family relations: A review. Journal of Marriage and the Family, 50, 595-618. Steele, S. (1992, July). The new sovereignty: Grievance groups have become nations unto themselves. Harper's Magazine, pp. 47-54. Steil, J. M. (in press). Equality and entitlement in marriage. In M. Lerner & G. Mikula (Eds.), Entitlement to the affectional bond. Steil, I. M., & Weltman, K. (1991). Marital inequality: The importance of resources, personal attributes, and social norms on career valuing and the allocation of domestic responsibilities. Sex Roles, 24, 161-179. Stouffer, S. A,, Suchman, E. A., DeVinney, L. C., Star, S. A., &Williams, R. M., Jr. (1949). The American soldier. Princeton, NJ: Princeton University Press. Suls, J., & Miller, R., Eds. (1977). Social comparison processes: Theoretical and empirical perspectives. Washington: Hemisphere.

354

BRENDA MAJOR

Suls, J., and Wills, T. A. (Eds.). (1991). Social comparison: Contemporary theory and research. Hillsdale, NJ: Erlbaum. Tajfel, H., & lhrner, J. C. (1986). The social identity theory of intergroup behavior. In W. Austin & S . Worchel (Eds.), The social psychology of intergroup relations (pp. 7-24). Monterey, CA: BrookslCole. Taylor, D. M., & Dube, L. (1986). ’nvo faces of identity: The “I” and the “We.” Journal ofSocial Issues, 42, 81-98. Taylor, D. M., & Moghaddam, F. M. (Eds.). (1987). Theories of intergroup relations: International social psychological perspectives. New York: Praeger. Taylor, D. M., Wright, S. C., Moghaddam, F. M., & Lalonde, R. N. (1990). The personal/group discrimination discrepancy: Perceiving my group, but not myself, to be a target for discrimination. Personality and Social Psychology Bulletin, 16, 254-262. Taylor, D. M., Wright, S . C., & Porter, L. E. (1993). Dimensions of perceived discrimination: The personal/group discrimination discrepancy. In M. Zanna & J. Olson (Eds.), The psychology of prejudice: The Ontario Symposium (Vol. 7, pp. 233-255). Hillsdale, NJ: Erlbaum. Taylor, S. E., & Brown, J. D. (1988). Illusion and well-being: A social psychological perspective on mental health. Psychological Bulletin, 103, 193-210. Tesser, A., & Campbell, J. (1983). Self-definition and self-evaluation maintenance. In I. Suls & A. G. Greenwald (Eds.), Psychological perspectives on the self(Vo1. 2, pp. 1-31). Hillsdale, NJ: Erlbaum. Testa, M., & Major, B. (1990). The impact of social comparison after failure: The moderating effects of perceived control. Basic and Applied Social Psychology, I I , 205-2 18. Thibaut, J. W., & Kelley, H. H. (1959). The social psychology of groups. New York: Wiley. Thompson, L. (1991). Family work: Women’s sense of fairness. Journal ofFamily Issues, 12, 181196. Thompson, L., & Walker, A. J. (1989). Gender in families: Women and men in marriage, work, and parenthood. Journal of Marriage and the Family, 51, 845-871. Treiman, D. J., & Hartmann, H. I. (1981). Women, work and wages: Equal pay for jobs of equal value. Washington, DC: National Academy Press. Turner, M. E., Pratkanis, A. R., & Hardaway, T. J. (1991). Gender differences in responses to preferential selection: Towards a model of preferential selection as help. Journal of Social Behavior and Personality, 6 , 797-820. b l e r , T. R., & Lind, E. A. (1992). A relational model of authority in groups. Advances in Experimental Social Psychology, 25, 115- 191, U S . Government. (1978). Income and earnings diyerentials between black and white Americans (Document No. TF-6-95-0). Washington, DC: U.S. Government Printing Office. Vanneman, R. D., & Pettigrew, T. G. (1972). Race and relative deprivation in the urban United States. Race, 13, 461-486. Walster, E., Walster, G. W., & Berscheid, E. (1978). Equity theory and research. Boston: Allyn & Bacon. Watts, B. L., Messe, L. A,, & Vallacher, R. R. (1982). Toward understanding sex differences in pay allocations: Agency communion, and reward distribution behavior. Sex Roles, 12, 1175- 1 188. Websrer’s New World Dictionary of the American Language. (1970). New York: World. Weiner, B. (1985). An attributional theory of achievement motivation and emotion. Psychological Review. 92, 548-573. Wills, T. A. (198 1). Downward comparison principles in social psychology. Psychological Bulletin, 90,245-271. Wittig, M. A., & Lowe, R. H. (1989). Comparable worth theory and policy. Journal of Social Issues, 45, 1-21. Wood, J. V. (1989). Theory and research concerning social comparisons of personal attributes. Psychological Bulletin. 106, 231-248.

ENTITLEMENT AND SOCIAL INEQUALITY

355

Wright, S. C . , Taylor, D. M., & Moghaddam, F. M. (1990). Responding to membership in a disadvantaged group: From acceptance to collective protest. Journa/of Personality and Social Psychology, 58, 994- 1003. Zanna, M. P., Crosby, F., & Loewenstein, G. (1987). Male reference groups and discontent among female professionals. In B . Gutek & L. Larwood (Eds.), Women’s career development. Newbury Park: Sage.

This Page Intentionally Left Blank

MENTAL REPRESENTATIONS OF SOCIAL GROUPS: ADVANCES IN UNDERSTANDING STEREOTYPES AND STEREOTYPING

Charles Stangor James E. Lange

I. Introduction [We] fill in the . . . picture by means of stereotypes we carry around in our heads. (Lippmann, 1922, p. 59) Until . . . clear commitments are made to representations and their contents, no unequivocal predictions can be made concerning processing of stereotype information. (Hastie, 1983, p. 520)

The study of stereotypes and stereotyping is one of the oldest interests of social psychologists and one of the most avidly researched. Of course an expansive interest in stereotypes is not surprising given that one of the fundamental goals of social psychologists is to discover how people understand and react to the other people in their environment. Perhaps the most important way people reduce their uncertainty about others is by using their previously stored knowledge about people. Just as we categorize cars and teacups, we order our social worlds according to perceived similarities among people who share skin color, ethnic or religious group membership, social class, physical appearance, age, height, weight, the presence of social stigmata, and other social features. People assume (and are certainly often quite correct in doing so) that those who share overt physical features also share underlying beliefs, interests, personal habits, and (most importantly in the present context) personality characteristics. Thus stereotypes about social groups represent an important and commonly used set of expectations about others. These expectancies help us understand what others are like, predict how they will behave, and thus determine how we respond to them. That social expectations are a basic determinant of interpersonal judgments and behavior has been the major theme of most social-psychological models of 357 ADVANCES IN EXPERIMENTAL SOCIAL PSYCHOLOGY. VOL. 26

Copyright 0 1994 by Academic Press, Inc. All rights of reprnduclion in any form reserved.

358

CHARLES STANGOR AND JAMES E. LANGE

person perception (Asch, 1946; Fiske & Taylor, 1991; Heider, 1958; Markus & Zajonc, 1985; D. T. Miller & Turnbull, 1986; but see McArthur & Baron, 1983). Over the years evidence has accumulated showing that the physically apparent social-category memberships of others are spontaneously attended to and used as a basis of encoding (e.g., Brewer, 1988; Stangor, Lynch, Duan, & Glass, 1992; Taylor, Fiske, Etcoff, & Ruderman, 1978; Zlirate & E. R. Smith, 1990), that perceiving these categories quickly activates personality characteristics that are associated with the categories in memory (e.g., Diehl & Jonas, 1991; Dovidio Evans, & vier, 1986; Gaertner & McLaughlin, 1983; Gilbert and Hixon, 1991), and that these stereotypical expectancies influence both judgments of (e.g., Bodenhausen & Lichtenstein, 1987; Darley & Gross, 1983; Landy & Sigall, 1974; Leyens, Yzerbyt, & Schadron, 1992; Snyder & Uranowitz, 1978) and behavior toward others (e.g., Darley & Fazio, 1980; Jussim, 1991; Snyder, Tanke, & Berscheid, 1977; Word, Zanna, & Cooper, 1974). Expectancies are also known to influence what information is sought out (Bassok & Trope, 19831984; Snyder & Cantor, 1979), attended to (Belmore, 1987), and remembered about other people (Hastie & Kumar, 1979; Rothbart, Evans, & Fulero, 1979; Srull & Wyer, 1989; Stangor & McMillan, 1992), how that information is interpreted (Duncan, 1976; Peabody, 1968; Sagar & Schofield, 1980), and the ease with which it is processed (Macrae, Stangor, Griffiths, & Milne, 1993) as well as the affective and emotional responses that we experience in the presence of others (Cooper & Singer, 1956; Cooper, 1969; Fiske, 1980; Fiske & Pavelchak, 1986). There would be no reason for social psychologists to be concerned about, rather than merely interested in, a tendency to arrange and process one’s world in terms of shared social features if the process wasn’t often unfair to those who are judged on the basis of these expectations. However, when people base their judgments of and behavior toward others on their beliefs about the groups that those people belong to (and it may well be impossible ever to react to another person without using some expectation, either consciously or unconsciously), the process is often unfairly detrimental to the person being judged. Thus stereotypes have profound influences upon an extremely broad range of important decisions made about people in their everyday lives, including hiring and promotional decisions (e.g., Swim, Borgida, Maruyama & Myers, 1989), legal judgments (e.g., Bodenhausen & Lichtenstein, 1987, Ugwuegbu, 1979), and school performance (e.g., Jussim, 1991). It is the insidious effects of expectancy-based processing upon stigmatized individuals that, more than anything else, has fueled the persistent interest in studying stereotypes. Stereotyping (the use of stereotypical expectations about social groups in judging individual members of those groups) may be harmful regardless of whether the person using them is or is not aware of doing so. When the perceiver makes judgments of others on the basis of expectations without being aware of

GROUP REPRESENTATIONS AND STEREOTYPING

359

using them, there will be a tendency to apply the expectation indiscriminately to every individual member who possesses the feature with which the expectancy is associated. This may be unfair to the target person when these expectations are evaluatively negative, because even the most indiscriminating perceiver would admit that not everyone in a given social group has the same underlying personality (e.g., Mann, 1967; McCauley, Stitt & Segal, 1980). Furthermore, social expectations also indirectly influence responses to others by biasing information processing in ways that tend to strengthen existing expectations, such as when they influence memory, information seeking, and interpretation. These expectancyconfirming processes often remain out of the awareness of the perceiver. However, even when the individual is aware of using his or her expectations, realizes that they may be unwarranted, and tries to adjust for them, he or she may not be fully aware of the actual impact of these expectancies on responses. It is well known that people are generally insufficient in their adjustment of responses away from their initial characterizations (Gilbert & Krull, 1988; Kahneman, Slovic & Tversky, 1982; Trope, 1986). Thus expectancy-based processing, while potentially informative, also has the strong potential for leading to inaccurate and biased responses to others. Given the immense literature that has accumulated over the past decades, one would expect that progress in understanding the nature of stereotypes and the process of stereotyping would have accrued. The primary goal of this article is to review the major advance in conceptualizing stereotypes that has occurred over the past 20 years. This advance has resulted from the explicit recognition that stereotypes, like any other type of expectation, must be stored in long-term memory. As it has become increasing clear that a full understanding of the process of stereotyping requires a correspondingly full understanding of the nature of the mental representations of social groups, many researchers, taking many different approaches, have begun to direct their research programs toward an understanding of this issue. Becoming more explicit about how social groups are represented in memory has produced an abundance of insights into how stereotypes should be conceptualized and measured, in predicting the influence they will have on social judgment and behavior, and in understanding how they develop and why they are so resistant to change. Conceptualizing social stereotypes as mental representations thus represents a true advance in experimental social psychology.

11. Stereotypes as Mental Representations

One of the concomitants of an avid interest in a domain over a period of time is that the topic is approached by many researchers from many different directions.

360

CHARLES STANGOR AND JAMES E. LANGE

Stereotypes have been defined in so many different ways over the years that it has often seemed as if no two writers could agree upon even the most basic aspectsof the concept. In a comprehensivereview of the literature published about 20 years ago, John Brigham (1971a) bemoaned the current state of research and theorizing about stereotypes:“There is littleagreementbetween theorists and researchers concerning the function and importance of ethnic stereotypes in social perception and behavior . . . [and] . . . there is little consensus on just what a stereotype is” (p. 15). Brigham’s pessimistic conclusions are understandable if one assesses the definitions of stereotypes that were available at the time of his review. The top panel of Table I presents a sampling of definitions coined before 1975. These involve an array of different characterizations, no two of which seem to be in agreementa veritable Tower of Babel.

TABLE I DEFINITIONS OF STEREOTYPES: 1950 TO 1991 I . Years 1950 to 1975 “a relatively rigid and oversimplified or biased perception or conception of an aspect of reality, especially of persons or social groups’’ (English & English, 1958, p. 523). “Stereotyping represents the attributing of similar characteristics to different members of the same group, without sufficient account being taken of possible differences among the members of the group.” (Doise, 1969, p. 138). “an exaggerated belief associated with a category. Its function is to justify (rationalize) our conduct in relation to that category.” (Allport, 1954, p. 191). “A stereotype is commonly thought of as involving a categorical response, i. e., membership is sufficient to invoke the judgment that the stimulus person possesses all the attributes belonging to that category.” (Secord, 1959, p. 309). “a generalization made about an ethnic group, concerning a trait attribution, which is considered to be unjustified by an observer” (Brigham, 1971a, p. 31). “a collection of trait-names upon which a large percentage of people agree as appropriate for describing some class of individuals” (Vinacke, 1957, p. 229). 11. Years 1975 10 1991 “a collection of associations that link a target group to a set of descriptive characteristics” (Gaertner & Dovidio, 1986, p. 81). “a set of beliefs about the personal attributes of a group of people” (Ashmore & Del Boca, 1981, p.

16).

“a set of expectations held by the perceiver regarding members of a social group” (Hamilton, 1979, p. 65).

“sets of traits attributed to social groups” (Stephan, 1985, p. 600). “attributes that are strongly associated with a particular . . . concept” (Hamilton & Trolier, 1986, p. 141).

“highly organized social categories that have properties of cognitive schemata” (Andersen, Klatzky, & Murray, 1990, p. 192). “those generalizations about a class of people that distinguish them from others” (McCauley, Stitt, & Segal, 1980, p. 197).

GROUP REPRESENTATIONS AND STEREOTYPING

36 1

At about the time that Brigham was expressing his exasperation, however, a new approach to conceptualizing stereotypes was beginning to emerge. Beliefs about social groups were increasingly being considered in terms of social categories, prototypes, or person schemas. The view that social categorization was a major determinant of social perception was not a new one. It had been proposed several decades earlier by Allport (1954) and Vinacke (1957), and the influence of categorization upon social judgments had been extensively elaborated in seminal work by Tajfel and others (cf. Eiser & Stroebe, 1972; Tajfel, 1981). Although these earlier approaches had predicted and confirmed many detrimental outcomes of thinking about people in terms of social categories (e.g., overestimating similarities among and differences between category members; responding more positively to ingroup than to outgroup members), they had not explicitly addressed the issue of how social categories were mentally structured. The emerging focus upon the structure of mental representations of social groups was a new and important direction, and one that immediately resulted in a paradigm shift in the way stereotypes were defined. Even by 1975 the definition of stereotypes within social psychology had been consolidated around a shared understanding. Researchers had found a common language for conceptualizing stereotypes, a vernacular based upon principles of cognitive representation. The pervasiveness of this change is evident in the bottom panel of Table I where it can be seen that, in opposition to earlier discord, contemporary theorists are now in virtually unanimous agreement that stereotypes are characteristics that are descriptive of, attributed to, or associated with members of social groups or categories.' The agreement is now so pervasive that, unlike earlier writers, current authors rarely bother to define what they mean by a stereotype; it is assumed that all readers acknowledge the basic representational conceptualization. The multitude of definitions of stereotypes that prevailed before 1975 has been consolidated, replaced by a simple and apparently consensual way of thinking about stereotypes. A. AN ASSOCIATIVE MODEL OF SOCIALGROUP REPRESENTATION The seminal assumption of this new approach is that people possess abstract knowledge representations of the characteristics that they perceive to be true of the members of social groups. Although different theorists have considered these representations in somewhat different ways (cf. Fiske, 1982; Fiske, Neuberg, Beattie, & Millberg, 1987; Stephan & Stephan, 1993 for examples), there is a 'There appears to be some disagreement in the literature as to whether the term stereotype refers to the mental representation itself or to the characteristics associated with the category label within the representation. Most commonly, however, the term is used to refer to the stereotypical characteristics, and we will use the term in this sense.

-

362

CHARLES STANGOR AND JAMES E. LANGE

College Professors

I Absentminded I

Fig. 1.

Hypothetical mental representation of a social group.

basic agreement that stereotypical characteristics are associated with group labels in semantic memory through mental associations. These basic assumptions are derived from work in cognitive psychology (e.g., Anderson’s 1983 propositional network theory; the Collins 8z Loftus, 1975 “spreading-activation”model). Let us consider a concrete version of a simple associational model, as shown in Fig. 1. In this model a social-group representation consists of a category label (e. g., “college professors”) and a number of mentally-associated stereotypes. These associated characteristics are “featural” in the sense that they represent the salient or most typical characteristics of the category (cf. E. E. Smith 8z Medin, 1981, Chap. 4). In this case the stereotypes are trait terms, but although this has traditionally been the focus of stereotype research, stereotypes do not necessarily need to be limited to traits. As we will discuss in a later section, other types of characteristics (and particularly affective and emotional states) may also be associated with the category label. The power of the associational model is that it explains how stereotypes become activated in memory upon exposure to the category label. Excitation (activation in memory) of a construct is assumed to spread along links from an activated node (the category label) to associated nodes, which in turn become activated.* In addition to the actual semantic content of the label and the stereotypes (which varies, of course, from group to group), the representation itself can be specified in terms of a very limited number of parameters. One set of parame2Whether the stereotypical characteristics are themselves structurally interconnected and whether activation can spread upward from the characteristics to the label is an interesting question but one that we do not address here (cf. Diehl &Jonas, 1991). It is sufficient in the current model that activation be able to spread downward from the label to the characteristics.

GROUP REPRESENTATIONS AND STEREOTYPING

363

ters specifies the strength of association between the group label and each associated characteristic (represented graphically in Fig. 1 by the thickness of the connecting lines). The only other parameter is the number of characteristics within the representation (i.e., the number of non-zero linkages-in this case three). The following sections focus upon the determinants of associative strength between the category label and the characteristics. We turn to a fuller discussion, including the effects of the number of associated characteristics, in Section V. B . ASSESSING GROUP REPRESENTATIONS Associational models of group representation were naturally not developed in the absence of supporting evidence. On the contrary, several lines of research strongly support the assumption that characteristics are associated with group labels in memory. 1. Memory for Information about Social Groups

Perhaps the most straightforward prediction that can be derived from an associational model is that subsequently encountered material that is relevant to expectations about the group will be interpreted in terms of the existing knowledge (cf. Hastie, 1981; Srull & Wyer, 1989). This interpretation process serves to increase associations between the group label and the interpreted items, with the result that expectancy-relevant items will be particularly strongly linked to the group concept in memory. Furthermore, because material that “fits” the expectation (either evaluatively or descriptively) will be particularly easy to encode into the knowledge structure, expectancy-congruent information should be particularly strongly associated with the category label in memory. This latter assumption leads to two predictions: First, it is expected that expectancy-congruent information will be relatively easy to recall because it is highly associated with the category label that provides a cue for its retrieval. Secondly, we can predict that because expectancy-congruent material is strongly activated by the category label, it will thus seem familiar and should be more likely to be erroneously reported as having been seen on recognition memory tasks that do not control for guessing. To test the strength of these memory effects, Fyock and Stangor (1993) combined the results of 26 experiments that had presented subjects with expectancycongruent and -incongruent information describing social groups about which the subjects had well-developed representations (e.g., men and women; ethnic groups; occupational roles), before testing their memory for this material. Both

364

CHARLES STANGOR AND JAMES E. LANGE

children and adults were included as subjects in these studies, and a variety of social groups were used as stimuli. Significant congruency effects both for recall memory and for recognition response biases were found, so that stimulus items that were congruent with expectations about the groups were significantly more likely to be remembered than were stimulus items that were incongruent with the group expectation. This pattern held regardless of whether the to-be-remembered items were behaviors, traits, or photographs, and across a variety of social groups. Demonstrating the robustness of this effect, 23 of 26 studies showed a congruency effect in memory. In an experiment that investigated the memory hypothesis in a dynamic manner, Stangor and Ruble (1989a) studied the development of stereotypes about men and women in children. Children from three age groups (4 years; 7 years; 10 years) were shown a series of line drawings of men and women performing behaviors that are stereotypical of men (e.g., fixing a car) and women (e.g., doing the ironing). After a brief delay, the children were asked to remember as many of the pictures as they possibly could. The youngest children showed no preference toward remembering stereotype-congruent information, but this effect increased monotonically with age, so that 10-year-oldchildren remembered significantly more stereotype-congruent than -incongruent drawings. These results are quite supportive of the hypothesis that children develop mental representations of men and women over time and that these mental structures facilitate,the encoding and retrieval of expectancy-congruent information about the groups. Parallel results have been found in adults who learn about fictitious social groups in the lab (Stangor & Ford, 1992; Stangor & Ruble, 1989b).3 2. Priming Measures

The finding of preferential memory for expectancy-congruent information about well-known social groups is consistent with the hypothesis that group information is stored in mental representations, but is only an indirect test. More direct tests have been made using priming paradigms. If group characteristics are mentally associated with group labels in memory, then this association should be detectable through standard measures designed to assess mental linkages (cf. Anderson, 1983; Anderson & Bower, 1973; Kintsch, 1974; Meyer & Schvan31t should be noted that memory for social targets is much more complicated than this analysis suggests. Many variables influence whether people preferentially remember expectancy-congruent or -incongruent information, including whether the target is an individual or a group, the amount of experience those people have had with the target, the amount of time available to process the information, and the type of information to be remembered. In fact, memory for groups of individuals about which expectations are already strongly developed represents a rather unique case in which strong congruency effects are the rule (cf. F'yock & Stangor, 1993; Rojahn & Pettigrew, 1992; Stangor & McMillan, 1992).

GROUP REPRESENTATIONS AND STEREOTYPING

365

evelt, 1971). Specifically, when the category label is primed, other associated parts of the structure (stereotypes) should also be activated through spreading activation. Dovidio and his colleagues (Dovidio et al., 1986) used a modified version of the Lexical Decision Task (Meyer & Schvanevelt, 1971) to test directly whether stereotypical trait terms were associated with category labels in memory. Subjects participated in a series of trials in which a lexical label of a well-known social category (“blacks,” “whites”) was first presented, immediately followed by either a positive or negative trait term. On the basis of other measures, these traits were known to be either stereotypical or nonstereotypical of the primed social groups. The subjects’ task was to answer, both accurately and as quickly as possible, whether the presented trait could “ever be true of” the primed group. Results showed that very few errors were made on the task, and that, as expected, decisions about stereotypical traits (and particularly those that were evaluatively consistent with the perceivers’ prior attitudes toward the social group) were made more quickly in comparison to nonstereotypical traits. Similar findings have been reported by Gaertner and McLaughlin (1983). One limitation of the Dovidio et al. (1986) experiment is that the results are potentially subject to a demand interpretation. When subjects view social-group primes for several seconds and then are asked to respond to terms that may easily be perceived as stereotypical, they may modify their response criteria according to their expectations about the researcher’s hypothesis, and this could affect their response times. Macrae, Stangor, Griffiths, and Milne (1993) used an unrelatedexperiments design in an attempt both to avoid potential demand effects and to show the influence of priming on a measure other than reaction time. We expected that if exposure to a category label primes associated trait terms, then this activation should increase the ease with which these terms can be processed in a subsequent task. Subjects were first asked to write a short paragraph describing either a “child molester” or their “preferred inner-city living accommodations” (the latter is a control task pretested to be equally negative as, but unrelated to, the childmolester stereotype). Then, in a second experiment, ostensibly completely unrelated to the first, subjects viewed a list of 30 words presented for several seconds each on a computer screen, with the goal of memorizing them. One half of the presented words were known to be stereotypical of child molesters, whereas the other 15 words were fruits. While viewing the words, subjects also performed a secondary task-listening to a factual description of the geography and economy of Indonesia. It was predicted that because stereotypical traits were activated for the subjects who originally wrote about child molesters, they would more easily learn the stereotypical words (but not the fruits), and would also (because of increased residual capacity) demonstrate better memory for the Indonesia passage, as assessed by a multiple-choice questionnaire regarding the passage (the

366

CHARLES STANGOR AND JAMES E. LANGE

questionnaire was given at the end of the session). Demonstrating that activation of the traits had these beneficial effects, both predictions were supported. One limitation of the Macrae, Stangor, Griffiths, & Milne (1993) experiment is that, because subjects were given plenty of time to think about the category, it is not clear how quickly exposure to the category label primed the associated trait terms. Because spreading activation between nodes in a cognitive network is expected to occur automatically (e.g., Meyer & Schvaneveldt, 1971), exposure to a category label should effortlessly and invariably activate the associated trait terms. In one experiment that challenges this expectation, Gilbert and Hixon (199 1) had subjects perform a word-fragment-completion task on fragments that could be completed as either stereotypical of Asians or as stereotype unrelated. Gilbert and Hixon found that the fragments were more likely to be completed in the stereotypical manner when the person who was presenting the fragments was an Asian. However, suggesting that priming of stereotypical characteristics may not occur as automatically as would be expected by an associational model, subjects who were asked to remember a multi-digit number during the fragmentcompletion task did not show such priming effects. To this point, data are not conclusive about the automaticity of stereotype activation from exposure to category labels or category members. Some studies can be interpreted as showing quick activation (e.g., Devine, 1989; Dovidio et a]., 1986; Perdue & Gurtman, 1990), and others suggest that priming is not automatic (Gilbert & Hixon, 1991). Further research along these lines, preferably using tried-and-true measures of association such as the Stroop color-naming task (Higgins, VanHook, & Dorfman, 1988; Warren, 1972) will be necessary to determine the answer to this question. One possible resolution is that some group stereotypes are automatically activated in all contexts, whereas the activation of other stereotypes may be more dependent upon contextual factors (cf. Barsalou, 1987). Despite some remaining ambiguities, the memory and priming data provide considerable support for the basic assumption of the associational model, namely that stereotypical characteristics are strongly associated with the category label in memory, and are activated when it is activated.

111. A New Way of Defining Stereotypes The question of what makes a personality characteristica stereotypehas been a continuing source of controversy within the area, and was a contributing factor to Brigham’s discouragementabout the stereotype concept. One of the many advantages of considering stereotypes in terms of mental associations is that this approach produces an elegant solution to this question. According to the basic

GROUP REPRESENTATIONS AND STEREOTYPING

367

logic of the structural approach, a characteristic is a stereotype to the extent that it is strongly associated with the relevant category label in memory. Although traditional approaches did not explicitly consider stereotypes in terms of associative networks, they often defined them in terms of constructs that are highly related to the associational idea. Thus theorists variously considered generalization, homogeneity, and group dtrerences to be important features of stereotypes, although disagreeing among themselves regarding the relative importance of each. Considering stereotypes in terms of abstract mental representations produces a reconciliation of these diverging definitions.

A. GENERALIZATION It is not surprising that many theorists considered a characteristic to be a stereotype only if it was generalized to the group as a whole. An intuitively obvious requirement for a characteristic to be considered stereotypical is that it be found throughout the group. If “romantic” is a stereotype of Italians, for instance, then people should believe that Italians are, by and large, romantics. Although both a prominent component of many traditional definitions and an intuitively logical proposition, the generalization notion is problematic because it is routinely found that terms that appear to be stereotypical on the basis of other criteria are not, in fact, generalized to a great percentage of group members. This is particularly true for negative characteristics, the assignment of which is commonly assumed to represent the deleterious aspect of holding stereotypes in the first place. This difficulty is clearly demonstrated in a recent study from our laboratory. We first asked a set of subjects to generate lists of characteristics that they believed were most true of seven different social groups. We then coded these responses by selecting the six traits (including synonyms) that were listed most often for each group by these subjects and asked a second set of subjects from the same population to estimate the percentage of people in each of the groups who possess each of the traits. The results of the percentage estimation task are shown in Table 11. The traits presented to the second group of subjects should be perceived as highly stereotypical because they were those that were most frequently generated by the first group of subjects as descriptive of the target groups. Nevertheless, the average percentage ascription of the stereotypical traits across all groups was only 61% for the positive traits and 52% for the negative traits. Overall, almost 25% of these traits were rated as being true of less than 50% of the group members. The low instance of assignment is particularly ac$for negative traits, for which almost one half of the traits were perceived as being true of less than 50% of the group. Of course (although complete anonymity was guaranteed to

368

CHARLES STANGOR AND JAMES E. LANGE TABLE I1 POSITIVE AND NEGATIVE STEREOTYPES ASCRIBED TO SEVENSOCIALGROWS Social group Arabs

Blacks

Hispanics

Gays

Jews

Asians

Russians

Average

Positive stereotypes

%

Negative stereotypes

%

Religious Powerful Wealthy Athletic Rhythmic Sociable Cheerful Traditional Proud Loving Liberal Sensitive Industrious Thrifty Talkative Intelligent Gentle Industrious Intelligent Obedient Patriotic

61.2 50.8 40.8 62. I 51.8 55.4 56.1 65.0 54.2 62.4 69.9 67.6 55.3 61.9 58.0 78.8 71.3 78.1 67.4 63.1 59.9

Greedy Smelly Violent Aggressive Lazy Inferior LZY Dirty Aggressive Abnormal Confused Fussy Loud Greedy Clannish Selfish Nationalistic Passive Cold Evil Naive

66.5 50.7 40.6 49.7 49.8 36.6 58.6 61.9 54.4 53.8 49.4 63.4 55.7 53.0 56.8 30.8 76.7 60.8 36.1 32.4 47.7

61.5

51.7

our subjects) reporting biases may have lowered the estimates for the negative characteristics, and these biases may be particularly strong for college students (but see Stangor, Sullivan, & Ford, 1992, who found few differences between stereotypes generated by college students and those generated by a sample taken from a broader population). Nevertheless, even if these reporting biases were controlled, it is highly unlikely that these data could ever be taken as demonstrating that these traits are perceived as highly generalized to the relevant groups, in an absolute sense. The finding of low generalization of stereotypical traits is not unique to our laboratory. For instance, McCauley and Stitt (1978) found that traits perceived to be highly characteristicof Germans were ascribed to only 57% of the group overall, in comparison to 46% for nonstereotypical traits, and Locksley, Borgida, Brekke and Hepbum (1980) found that “assertiveness,” a trait presumably highly typical of men, was rated to be true of only 49%of them. Other stereotype theorists were also concerned with generalization, although they expressed this notion in a slightly different form, arguing the importance of group homogeneity as a necessary feature of a stereotype (cf. Doise, 1969;

GROUP REPRESENTATIONS AND STEREOTYPING

369

Doise, Deschamps, & Meyer, 1978). Generalization of a characteristic to a group will be accompanied by high homogeneity of the characteristic within the group because generalization and variability are necessarily negatively correlated (cf. Judd, Ryan, & Park, 1991). Thus if traits are not highly generalized to real social groups, it is unlikely that they will be perceived as highly homogeneous within those groups either. Although few experimenters have reported data about the homogeneity of real social groups, possibly because traditional stereotype measures were not designed to measure perceived group variability, some reports do suggest that groups are not perceived as highly homogeneous on stereotypical traits. In a recent experiment looking at the perceptions of Germans by American students who were visiting in Germany (Stangor, Jonas, Stroebe, & Hewstone, 1991), we again found that stereotypical traits were not perceived as highly generalized to the groups (average assignment of 60 cm on a 100-cm line) and also that the perceived range of the traits within the groups, again estimated using ratings on a 100-cm line, was quite high (average range = 76 cm). Similar findings are reported by Judd et al. (1991).

B. STEREOTYPE DISTINCTIVENESS The previous data suggest that stereotypes of real social groups are often not highly generalized to those groups, nor are they perceived as particularly homogeneous within them. Other theorists who have become aware of these findings have argued that group distinctiveness is a more important criterion for a stereotype (cf. McCauley et al., 1980; C. L. Martin, 1987). Distinctiveness refers to the extent to which a characteristic is perceived to be more highly associated with one social group in comparison to other comparison groups. However, distinctiveness is also a criterion that is not always met in ascription of stereotypes to real groups. Consider the traits in Table 11, which are those that were the most frequently generated for each of the groups in our first sample of subjects. It is clear that the same traits are perceived as stereotypical of many different groups. For instance, “intelligence” was one of the three most often generated traits for two out of the seven groups, “sociability” (sociable, cheerful, talkative) was generated for three out of the seven, and hostility (aggressive, violent, evil) was generated as highly stereotypical of four out of the seven groups. In an insightful analysis of the origin of stereotypes, Brewer and Campbell (1976) argued that groups that share similar occupational roles will be assigned similar characteristics, and a similar argument about social roles has been made for the origin of gender stereotypes by Eagly and Steffen (1984). Thus stereotypes of groups that have similar societal roles, or more generally have similar functional relations to the perceiver, would not be expected to be distinctive. Our

370

CHARLES STANGOR AND JAMES E. LANGE

finding that hostility is assigned to many different outgroups probably represents in part the fact that many outgroups are perceived to be threatening. C . CATEGORY DIFFERENTIATION One solution to the apparent insufficiency of traditional concepts such as generalization, homogeneity and distinctiveness to capture the full flavor of stereotypes is suggested by the associational approach. One of the basic tenets of theories of both object and social categorization is that categorical representations develop to provide useful information about the environment. Moreover, categories are particularly informative when the mean difference between or among categories on a set of attributes is large relative to the variability of those attributes within the categories (cf. Kelley, 1967; Rosch, 1978; n m e r , 1987). According to this logic, one likely determinant of whether a characteristic is strongly associated with a category in memory (and thus stereotypical) is the degree to which it maximizes category informativeness. Because within-group representativenessand homogeneity as well as between-group distinctiveness all contribute to category differentiation, all three of these factors should, as expected by traditional theorists, be important determinants of stereotypicality. Yet none should alone be a necessary criterion. Some recent experiments from our laboratory (Ford & Stangor, 1992) have tested the hypothesis that traits become stereotypical to the extent that they maximize category differentiation, and that differentiation is determined in part by generalization, homogeneity, and distinctiveness. In these experiments subjects read a series of behaviors supposedly performed by members of two social groups. Subjects were told that these behaviors were selected from real social groups at another university, but that for the purposes of the experiment the groups would simply be labeled as the “blue” group and the “red” group. The behaviors were selected so that the red group performed behaviors that demonstrated more intelligence than did those performed by the blue group, whereas the blue group performed behaviors that were more friendly than those performed by the red group. Furthermore, group distinctivenesswas manipulated by varying the size of this between-group difference in extremity to be either greater on the friendliness dimension or on the intelligence dimension (variability of the behaviors did not differ across any of the conditions). In another set of conditions the between-group differences were held constant, while the withingroup variability (homogeneity) of the traits was manipulated to be greater for either intelligence or friendliness. Subjects read the behaviors, interspersed from each group, in booklets for several seconds each. After learning about the groups, stereotypes were assessed by having subjects list four thoughts about each of the groups, and by counting the number of

GROUP REPRESENTATIONS AND STEREOTYPING

37 1

mentions of each of the two relevant trait terms (or their synonyms). This measure showed that the stereotypes developed about the groups were more likely to consist of the trait dimension that more highly differentiated the two groups, regardless of whether that differentiation was the result of high between-group distinctiveness or low within-group variability. Consistent with the categorydifferentiation hypothesis, these data suggest that, when within-group variability is controlled, a trait is more stereotypical if it differentiates groups, and when between-group differences are controlled, traits that show lower within-group variability become more stereotypical. Also suggesting that these stereotypes developed to help differentiate groups from each other, these effects did not occur when the extremity and variability of the traits were manipulated in the same manner, but subjects learned about only a single group. Diehl and Jonas (1991) took a different approach to the same question, but found quite similar results. They first measured, at an individual-difference level, both generalization of traits to existing national groups (Germans; Italians) and perceived distinctiveness between the groups. They found that both of these measured variables significantly and independently predicted the speed with which subjects could identify whether the stereotypical traits were descriptive of the groups on a subsequent reaction-time task. Because this task appears to be a good measure of the strength of association between the group label and a trait term (cf. Dovidio et al., 1986; Fazio, Powell & Herr, 1983; Ford, 1992), these data also demonstrate that the extent to which characteristics are associated with social groups in memory is predicted by both their generalization and their distinctiveness. In summary, considering stereotypes in terms of strength of association between a characteristic and a category label explains and confirms the intuitions of many previous conceptions of stereotypes, and well summarizes the contemporary definitions in the lower panel of Table I. A characteristic does not necessarily need to be either generalized, homogeneous, or distinctive to be stereotypical, but characteristics that jointly maximize these features (and thus category differentiation) will be more strongly associated with the category label in memory, and thus stereotypical.

D. OTHER DETERMINANTS OF ASSOCIATIVE STRENGTH Our discussion to this point might lead to the conclusion that because generalization, homogeneity and distinctiveness all contribute to making a characteristic stereotypical, category differentiation might be the only determinant of stereotype development. However, although the extent to which a trait maximizes category differentiation is one determinant of associative strength, other factors

372

CHARLES STANGOR AND JAMES E. LANGE

unrelated to differentiation are also important. Because these factors are known to be integral to the development of stereotypes, it is worth considering them in some detail. One of the most traditional explanations for the development of stereotypes is that they serve an ego-enhancing function, serving to heighten one’s own selfesteem. This process may involve developing either negative perceptions of outgroups or positive perceptions of ingroups (Tbrner, 1987). To test the hypothesis that a need for developing a positive image of one’s own social group influences associational strength independently of category differentiation, Ford (1992) created situations in which needs to favor the ingroup over the outgroup either coincided with or conflicted with category differentiation. Subjects were led to believe that they were either a member of Group X or a member of Group Z on the basis of a bogus personality test. Then subjects learned about the personality characteristics of members of Group X and Group Y. Group X performed mostly unintelligent and friendly behaviors, whereas Group Y performed mostly unfriendly and intelligent behaviors. Furthermore, friendliness was a more utilitarian stereotype in some conditions because it more highly differentiated the two groups, whereas intelligence was more informative in other conditions. Subjects were then asked to express their stereotypes about Group X and Group Y using a thought-listing task. Ford found that subjects who were members of the irrelevant group (Group Z) were more likely to mention the more differentiating dimension when describing the two groups. However, subjects who thought they were members of Group X were more likely to mention friendliness (the trait that favorably distinguished their own group) than intelligence, to describe both Group X and Group Y, regardless of which trait maximized category differentiation. In a similar demonstration of the importance of self-enhancement motivations, Schaller and Maass (1989) found that motivations to view one’s own group in a positive manner reduced or eliminated the effects of stimulus salience in the illusory-correlation paradigm (Hamilton & Gifford, 1976). In these studies subjects who were led to believe that they were members of one of two target groups formed positive impressions of that group, even though negative ingroup information about that group was highly salient and nonmotivated perceivers (who saw the same information but did not believe they were a member of the group) formed negative impressions of this group. To the extent that illusory correlations occur because subjects perceive infrequently occurring and thus unusual or unexpected information to be particularly informative and thus weight it heavily (cf. Fiske, 1980; Jones & Davis, 1965), this finding is quite parallel to Ford’s (1992) findings. Taken together, the findings of Ford (1992) and Schaller and Maass (1989) show that the need to evaluate one’s own group positively can override the influence of other variables that contribute to stereotype formation.

GROUP REPRESENTATIONS AND STEREOTYPING

373

In sum, we have argued in this section that characteristics should be considered stereotypical to the extent they are strongly associated with a category label in memory. The notion of association is explicit in current conceptualizations of stereotypes and is supported by several lines of research. One determinant of strength of association is the extent to which the characteristic provides differentiation among categories. Characteristics that have a high correlation with category membership will be more strongly associated with the category label in memory. This notion well captures the ideas that category representations of social groups serve a pragmatic function because they help the perceiver simplify the environment (Fiske & Taylor, 1984) and accurately understand others (Fiske, 1993; Oakes & Turner, 1990). Moreover, as suggested by other approaches to stereotype development (e.g., Turner, 1987), strength of association, and thus stereotypicality, is also determined by other motivations, including the desire for positive ingroup identity. Although we have focused here on self-enhancement motives, it is likely that there are many motivational variables that lead to strong associations between category labels and associated characteristics, and an analysis of these factors represents an important topic for future research. These motivations will range from needs to simplify or accurately understand the environment, to maintaining existing expectations intact (cf. Stangor & Ford, 1992), to those of egoenhancement, social acceptance, communication with others (Farr & Moscovici, 1984; Hogg & Abrams, 1988), and justification of existing social situations (Hoffman & Hurst, 1990) or attitudes (Snyder & Miene, 1993). Recent research by Snyder and Miene (1993) has addressed some of these underlying motivations and represents an important contribution in this regard. One important goal of this research will be to determine the underlying processes by which these diverse motivations influence stereotype development. For instance, motivations may act at the perceptual level, leading people to be more or less likely to expose themselves to or to attend to stereotype-congruent versus incongruent information. Furthermore, they may act on memory, making stereotypical information more or less likely to be stored and retrieved. Alternatively, motivational effects may occur by influencing when and how expectations influence behavior toward others. The ultimate goal of research along these lines will be to produce a unifying integration of more traditional motivationally based models of stereotyping with more recent cognitive approaches.

IV. Stereotype Measurement Over the past decades social psychology has become increasingly focused upon issues of process, and issues of content have been relatively neglected. This

374

CHARLES STANGOR AND JAMES E. LANGE

trend has also been apparent within the stereotyping literature, where questions about which traits are stereotypical of which groups have tended to be displaced by concerns such as “What are the influences of social stereotypes on social judgment and behavior?’ and “When do social stereotypes influence responses?’ The neglect of the content issue has been exacerbated by an increasing tendency to use artificial, rather than existing, groups to study the stereotyping process. One negative outcome of a greater interest in process than content is that advances in measuring stereotypes have not kept pace with an understanding of the influence of stereotypes upon social perception. As a result there is still little agreement upon the most appropriate techniques for assessing stereotypes of real social groups. Another benefit of considering stereotypes in terms of mental associations is that it provides direction for doing so.

A. TRADITIONAL MEASUREMENT TECHNIQUES Given the multiplicity of “traditional” definitions of stereotypes, it would naturally be expected that a corresponding multiplicity of measurement techniques would have developed over the years. Surprisingly, this has not been the case. There are only three commonly used stereotyping measures. In the Katz and Braly (1933) stereotype-checklisttechnique, subjects are presented with the names of social groups, followed by a list of trait terms, and are asked to check off those they see as most descriptive of the groups. Brigham (1971a) advocated a percentage estimate technique in which subjects estimate the percentage of people within a group who possess an associated characteristic (What percentage of Italians are romantic?). A modification of this method is the McCauley and Stitt (1978) diagnostic ratio measure in which percentage estimates for groups are divided by percentage estimates of “people in general.” Together, the checklist and percentage estimate approaches have been used in virtually all stereotyping studies over the past 20 years. These two approaches, along with the less commonly used Likert scale approach (cf. Judd et al., 1991; Rosenkrantz, Vogel, Bee, Broverman, & Broverman, 1968; Sigall and Page, 1971) and the Gardner stereotype differential (Gardner, Kirby, Gorospe, & Villamin, 1972; Gardner, 1993) thus represent the traditional approach to stereotype measurement. It does not matter to a great degree which of these stereotype measures are used, because they are highly intercorrelated. Gardner and his colleagues (e.g., Gardner et al., 1972; Gardner, Wonnacott, & Taylor, 1968) have published a number of reports showing that these traditional measures produce a single common underlying factor. These studies provide convergent validation for the traditional approach to measuring social stereotypes, and suggest that all of these measures are tapping some common aspect of stereotyping.

GROUP REPRESENTATIONS AND STEREOTYPING

375

B. THE FREE-RESPONSE APPROACH Traditional stereotyping measures were quite suitable for the descriptive studies that dominated the field for many decades. However, these measures are less than adequate for contemporary work because they were not developed to assess stereotypes as conceptualized in terms of mental associations. As a result these measures are likely to provide only an indirect assessment of associational strength. One way to conceptualize the difference is that traditional stereotype measures are designed to assess the availability of a trait within a representation-that is, does the respondent, when asked, indicate that the trait is true of the group? On the other hand, the current approach suggests that it is the accessibility of a trait (the likelihood of a trait being activated upon exposure to the group label or to a member of a e group) that is critical (cf. Higgins & King, 1981).4 If a characteristic is stereotypical to the extent that it is strongly associated with a group in memory, then any measure of the strength of relationship between the category label and the trait should serve as a direct measure of a characteristic’s stereotypicality. Thus reaction time measures such as the Stroop color-naming task (Warren, 1972) or the Lexical Decision Task would provide the most direct measures, and (as discussed in Section II,B,2) such priming techniques are more commonly being used to assess stereotypes. Of course, because these approaches are unwieldy for everyday use, they are unlikely ever to be routinely used as stereotype measures. One potentially useful alternative measure is based on a methodology in which individuals are asked to indicate what thoughts come to mind when they think about the relevant social groups. The procedure of asking subjects to freely choose their own descriptive terms to describe the groups of concern is currently enjoying a revival of interest (cf. Devine, 1989; Eagly & Mladinic, 1989; Esses, Haddock, & Zanna, 1992, 1993; Ford & Stangor, 1992; Stangor et al., 1992). The free-response technique has a number of potential advantages over more traditional measures, as well as some potential limitations. The most important advantage of free-response measures is that they represent a more direct indication of the strength of association between a category label and a characteristic than do traditional measures. Traditionally, free-response measures have been considered among the best approaches for assessing representational structure (cf. Cantor & Mischel, 1979; Rosch, 1978), and have routinely been employed for this purpose in other domains, such as mapping the self-schema (Markus, 1977; McGuire & McGuire, 1988; Trafimow, Triandis, & Goto, 1991). Given 40f course one would not expect that measures of accessibility and availability would be entirely independent. As we will argue subsequently, the perceived extremity of a trait may well be determined, at least in part, by the strength with which it is activated by the category label.

376

CHARLES STANGOR AND JAMES E. LANGE

the importance of free-response measures in other domains, it is rather surprising to find that this approach has so rarely been used in the area of stereotype measurement. If free-response measures are better indications of associative strength than are more traditional stereotype measures, then they should be more highly correlated with direct measures of mental association. To test this hypothesis, Ford (1992) had subjects learn about two groups by reading behavioral descriptions of group members. The behaviors performed by the “red” group were more friendly than those performed by the “blue” group, whereas behaviors performed by the “blue” group were more intelligent than those performed by the “red” group. After learning about the groups, subjects completed a free-response measure (listing whatever thoughts came to mind about the groups) as well as traditional (percentage estimate and Likert) measures of stereotypes. In addition, subjects participated in a direct measure of associational strength-a reaction-time task consisting of a number of trials in which the category label (“red” group or “blue” group) was presented simultaneously with a trait associated with either friendliness or intelligence. Subjects indicated as quickly as possible whether the trait term could be used to describe the group. Ford found that the thought-listing and reaction-time measures were highly correlated, as were the two traditional (percentage and Likert) measures. Confirmatory factor analyses of the four measures showed that a two-factor solution based on the traditional (availability)versus associational (accessibility)dichotomy provided a significantly better fit to the data than did a single-factor solution. These data are highly supportive of the use of free-response techniques to assess stereotype strength, and suggest that these measures are not entirely redundant with more traditional measures.

C. VALIDITY OF VARIOUS MEASUREMENT STRATEGIES Another potential advantage of free-response measures is that they may be more sensitive gauges of stereotyping than are traditional approaches. In a recent study by Rothbart and John (1992) the stereotypesof U.S. students were assessed when those students first entered college and again 4 years later when they left college. Correlations of over .90 across this period were observed on traditional measures of stereotyping, even though it was expected that the students would have changed their stereotypes over this time as a result of their new college experiences and the fact that many of them first came in contact with members of the relevant social groups when they came to campus. The observed rigidity over time suggests either that the stereotypes are truly very stable or that the measuring instruments were not sensitive enough to detect changes. It is possible that

GROUP REPRESENTATIONS AND STEREOTYPING

377

free-response methodologies might have resulted in more sensitive tests of these changes (see for instance, Barsalou, 1987, for a summary of data suggesting that concepts, when assessed using free-response methodologies, show substantial variability over time). One type of data that might be informative about the validity of different stereotype measures involves looking at how different measuring techniques predict attitudes and behavior toward the relevant groups. If stereotypes represent beliefs about social groups, then they should predict prejudicial attitudes and discriminatory behaviors, just as beliefs about any other attitude object would be expected to predict attitudes and behavior toward that object (cf. Eagly & Mladinic, 1989; Fishbein & Ajzen, 1975). This relationship follows directly from the commonly proposed assumption that stereotypes represent the cognitive “component” of prejudice (e.g., Harding, Proshansky, Kutner, & Chein, 1969). A recent experiment from our laboratory (Stangor, Sullivan, & Ford, 1992) supports the prediction that free-response measures are better indicators of prejudice and discrimination than are more traditional measures. We had subjects express stereotypes, attitudes, and behavioral tendencies toward a set of nine existing social groups (blacks, Russians, Asians, gays, etc.). Stereotypes were assessed using both a percentage-estimate task involving three positive and three negative traits determined on the basis of pretesting to be highly stereotypical of each group, and on a free-response format, in which subjects listed the thoughts about the groups that came to mind. After completing these tasks, subjects also rated both the thoughts and the stereotypes in terms of their perceived positivity, and both measures were weighted by the respective ratings. We found that free-response ratings were significantly more highly correlated with measures of both prejudice (attitudes) and behavioral tendencies (discrimination) than were the traditional stereotype measures. In fact, the traditional measures (regardless of whether they were assessed in terms of generalization or distinctiveness) were only very slightly related to the outcome variables. Of course the magnitude of this difference between measurement types may have been due in part to the fact that subjects produced their own descriptive terms on the free-response measures, whereas in the traditional measures the terms were provided by the experimenter. This difference points out another advantage of free-response approaches-they free the researcher from having to choose the stereotypical traits a priori. This is advantageous in that there is no longer danger that the wrong traits will be chosen for inclusion (see Widiger & Settle, 1987, for an egregious illustration of this type of 5lt could well be argued that subjects should also be allowed to generate the category labels that they see as relevant. One explanation for the low perceived generalization and high perceived variability in stereotypes of real social groups that was noted in Section II,A is that the groups for which subjects are typically asked to generate stereotypes may not be those that are particularly

378

CHARLES STANGOR AND JAMES E. LANGE

D. POTENTIAL DIFFICULTIES Although the above analysis suggests that free-response methodologies may provide more useful measures of stereotypes than do the traditional approaches, their use is not without some potential problems. For one, they do not always predict attitudes better than traditional measures do. For instance, Eagly & Mladinic (1989) found no differences in the degree to which free responses and traditional measures predicted attitudes toward women. Thus the conditions under which free-response approaches are most useful still need to be determined. Free-response methods also pose some other interesting measurement problems. Although the researcher does not have to choose stereotypes a priori, content coding of the generated stereotypes may be necessary post hoc, particularly to assess consensus among respondents. This coding can take time and will require reliability checking of the coding schemes. Nevertheless the information gleaned from this type of analysis can be very revealing (McGuire & McGuire, 1988). Because thought-listing tasks are not pure measures of associative strength, they should not be used to substitute for more direct measures when the use of the latter is possible. When listing thoughts about social groups, there is plenty of time, and potentially motivation, to edit responses. For instance, the rules of communication suggest that individuals may tailor their responses to avoid redundancies. This practice may lead to different attributes being expressed in different contexts, depending, for instance, upon what other groups are being described at the same time.6 However, this latter problem is also found in the use of traditional measures (Diab, 1963a, 1963b; Oakes & Turner, 1990). Despite these potential difficulties, free-response methodologies represent a convenient and relatively direct measure of structural association, and data show that they are highly correlated with reaction-time measures and at least partially independent of more traditional stereotype measures. We propose that researchers consider more frequently employing free-response methodologies in their work, potentially in addition to the collection of traditional measures.

V. Stereotype Accessibility: What “Comes to Mind” We have now considered some evidence related to a conceptualization of stereotypes in terms of cognitive associations and the implications of this apinformative or functional to them, and thus about which they have developed particularly strong stereotypes. 61t is quite possible that characteristics that are more strongly associated with the category label are less subject to this type of contextual effect than are less strongly associated characteristics (cf. Barsalou, 1987, for a review of relevant findings for object categories).

GROUP REPRESENTATIONS AND STEREOTYPING

379

proach for stereotype measurement. However, the true advantages of an associational model become most apparent when applied to understanding phenomena that have traditionally been difficult to account for, but that are parsimoniously explained in terms of mental representations. In this section we consider how the associational model can be used to understand the influence (or lack thereof) of stereotypes upon social judgment and behavior. It is known that social expectations have small (but potentially important) overall effects upon responses to others (Jussim, 1991) and that their use is highly determined by context (Deaux & Major, 1987; E. R. Smith & ZLate, 1992). Thus, predicting when stereotypes will or will not be used as a basis of judgment is not a trivial matter. Most generally the issue involves determining which of many available sources of information are used by the perceiver as a basis of responses to others. Stereotypes represent one source of information that can be used as a basis for responses, but many other sources of information are also, in most cases, likely to be available. One prediction that can be derived from the present approach is that a given stereotype will have an influence upon responses to the degree that it has “come to mind’ (is accessible in memory), in comparison to other competing characteristics, at the time the response is made. We have so far considered one factor that determines the degree to which a characteristic will be activated on exposure to a group member, namely the strength with which that characteristic is associated with the group label in memory. More strongly associated characteristics will be more strongly activated. However, the situation is more complicated than this for several reasons. For one, the likelihood that a given characteristic will be used as a basis of judgment will be decreased to the extent that other competing characteristics are also activated. Thus as the number of characteristics associated with the category label within a given representation increases, the likelihood that any one of them will be used in judgment decreases (the likelihood of their being activated may also decrea~e).~ In the previous discussion we have not addressed this issue, but it forms the basis of some important predictions of the current model. Secondly, we have heretofore considered situations in which only one representation is activated. However, in everyday social interactions targets are likely to be thought of in terms of many different representations (e.g., “woman,” “attractive person,” “Italian,” “mother,” “feminist”), and each of these compet’The present model makes the simplifying assumption that the degree of activation of a characteristic is determined solely by its strength of association with the category label. However, it is not unreasonable to assume that the strength of activation of a Characteristic would be negatively related to the number of other characteristics within that representation because activation from the category label would be spread among all of the characteristics. Thus even strongly associated characteristics might be only weakly activated if the representation in which they are contained has many characteristics. The neglect of this likely possibility does not change any of the important predictions of the present model.

3 80

CHARLES STANGOR AND JAMES E. LANGE

ing representations may be activated to different degrees. Thus a more complete prediction of when a given stereotype will influence responses must simultaneously consider (1) the strength of activation of the group representation containing the stereotype, (2) the relative activation of competing representations, and (3) the strength of association between all of the activated representations and their associated characteristics. This relationship might be conceptualized as: p(Use of Characteristicj contained in Representation i in social judgment)

-

R; * sij r n n

where Riis the strength of activation of Representation i and sii represents the strength of association between Representation i and Characteristicj. The summation in the denominator is across all of the m representations that are currently activated (i.e., Ri> 0) and all of the n associated characteristics within each representation (i.e., sij > 0). Although based on only four parameters (the number of activated representations and the strength of activation of each, and the number of non-zero associations in each representation and the strength of each category label-stereotype association), this equation well captures the flexibility of social judgments by specifying which stereotypes will be activated for use in information processing under which conditions. Specifically, a given characteristic is less likely to be used as a basis of judgment to the extent that other competing representations, with other strongly associated characteristics, are simultaneously active in memory (see footnote 7).

A. STEREOTYPE ASSESSMENT: GROUPS VERSUS INDIVIDUALS Traditionally, stereotypes are assessed by presenting subjects with a group label (“Italians”) and asking them to indicate the characteristics they find to be true of the group. Consider an alternative approach in which subjects make judgments about a series of individual group members rather than about the group as a whole, and these individual ratings are then averaged to create the stereotype measure. Such a procedure is commonly used to assess stereotypes about physically attractive and unattractive individuals, although it is relatively rare in the stereotyping literature overall (but see Secord, 1959; Secord & Berscheid, 1963). Although these two measurement approaches may seem conceptually equivalent, experiments in which both techniques have been used routinely

GROUP REPRESENTATIONS AND STEREOTYPING

38 1

find that there is less relationship between them than might be expected (e.g., Brigham, 1971b; but see Jonas & Hewstone, 1987). In a recent study from our lab (Sigall, Stangor, & Ford, 1990) white subjects first completed a group (percentage estimate) measure of stereotyping toward blacks, Asians and whites. Returning for an ostensibly unrelated session held several weeks later, the same subjects made judgments about a series of facial photographs of black, Asian and white individuals. For each target person, subjects checked off whether each of the stereotypical traits that they had rated in the first session was appropriate for that person. Our results showed highly significant differences between the two measurement techniques. Significantly fewer negative stereotypical traits were assigned in the individual- versus the group-judgment task for both whites and Asians. For black targets, on the other hand, fewer positive and more negative traits were assigned on the individual measures (See Brigham, 1971b and Mann, 1967 for other examples of such discongruities). Although there are a number of potential explanations for these findings (including self-presentationeffects), perhaps the most straightforward is simply that different stereotypes were activated in the different judgmental conditions. When making judgments of group labels, the perceiver’s focus is on the label, and few competing representations will be strongly activated. When a perceptual image of an individual group member is present at the time of judgment, however, any number of competing features (facial features, skin color, hair style, age) are available and may activate associated representations beyond those of supposed interest. To the extent that many representations are simultaneously activated, the influence of any given one would be expected to be limited. Although quite basic, these findings represent the basic problem of predicting when a given representation will be used as a basis of judgment. B . ACTIVATION OF COMPETING SOCIALCATEGORY REPRESENTATIONS We have previously considered some of the factors that produce associations between category labels and stereotypical characteristics, and the methods of assessing these associations. Now we consider the factors that contribute to activating the representations themselves (cf. Jones & Thibaut, 1958; E. R. Smith & ZArate, 1992). Although less is currently known about this important question than would be desired, representation activation is known to be dependent upon both contextual and perceiver characteristics. Most generally, any factor that increases the accessibility of a given social representation will be likely to increase its use in subsequent processing. The variables that determine accessibility are well known and include both data-driven factors (e.g., percep-

382

CHARLES STANGOR AND JAMES E. LANCE

tual salience) and perceiver variables (e.g., recency and frequency of activation, and the current processing goals of the individual). 1 . Salience One variable that is known to increase the momentary activation of a group representation is the salience of the representation in the stimulus array. All other things being equal, more perceptually salient social categories are more strongly activated (cf. Fiske, 1980). Thus Taylor et al. (1978) found that as the number of blacks in a mixed-race group or the number of men in a mixed-sex group decreased, the targets in the numerical minority were increasingly judged in stereotypical terms (cf. Higgins & King, 1981). The known importance of the “primitive” categories of age, sex, and skin color as a basis of social judgment is due to a large extent to their high perceptual salience. These categories are likely to be almost always highly accessible for all perceivers (Brewer, 1988; Fiske & Neuberg, 1990).

2 . Perceiver Variables The basic principles of construct accessibility suggest that there will be both long-term and short-term variation in the accessibility of social categories, and thus that perceivers should differ in their use of different representations in different contexts. In one experiment assessing long-term individual differences in category accessibility, Stangor (1988) first asked subjects to describe other individuals whom they knew using a free-response format (Higgins, King, & Mavin, 1982). It was expected that subjects for whom gender was a highly accessible category would be more likely to describe others using genderstereotypical characteristics. The generated thoughts were coded in terms of their relevance to gender categories, and subjects who used many gender-stereotypical terms were classified as “gender-accessible,” whereas those who used few gender-related terms were classified as “non-gender-accessible.” On a subsequent task held several weeks later, subjects viewed stereotypical and nonstereotypical behaviors performed by male and female targets. Suggesting that the use of gender as a categorization system was stable over time, those subjects classified as highly gender accessible in the first session were more likely, on a subsequent recognition memory task, to falsely report having seen the targets performing stereotypical (versus nonstereotypical) behaviors in comparison to the non-accessible subjects. This finding is consistent with the idea that there are relatively stable long-term individual differences in the extent to which gender representations are chronically activated and used as a basis of information processing. In another test of this hypothesis, Stangor, Lynch, Duan, & Glass (1992) first

GROUP REPRESENTATIONS AND STEREOTYPING

383

selected high- and low-prejudice white subjects according to their responses on a measure of prejudice toward blacks. Several weeks after taking the prejudice measure these subjects viewed a series of photographs of male and female, black and white targets. Each photograph was accompanied by a randomly selected verbal statement (cf. Taylor et al., 1978). After viewing all of the statements, subjects were asked, in a surprise recall test, to match each statement with the person who originally made it. We found that when highly prejudiced subjects made errors, they were significantly more likely to confuse statements made by blacks with other blacks, and whites with other whites, in comparison to lowprejudice subjects. Confusions by gender, however, did not differ for high- and low-prejudice subjects. This pattern of results is consistent with the idea that prejudiced individuals have a chronic concern with race and employ this dimension routinely in categorizing others (see also Allport & Kramer, 1946). Because social-group memberships function to provide information about targets, it would be predicted that representations that are more informative about underlying personality dispositions should have a higher probability of activation than should less informative representations (cf. Nosofsky, 1986 for a similar prediction in object categorization). Two findings in the research reported by Stangor, Lynch, Duan, & Glass (1992) support this general hypothesis. First, we found that our subjects were more likely to confuse targets by sex than by race, suggesting that sex was used to a greater degree in categorization than was race (see Zirate & Smith, 1990, for a parallel finding). In addition, when subjects were directly asked about the informativeness of sex and race in providing information about underlying personality, they indicated that sex was the more informative category.8 Second, we predicted that social-group representations would be activated in parallel (cf. Brewer, 1988; Kahneman & Miller, 1986) and that categorization in terms of the intersection of two categories (e.g., as a black male), rather than in terms of single categories, would be more informative when targets varied on both sex and race (black vs. white). We found that subjects made more withinrace/within-sex confusions than would be expected on the basis of their singlecategory errors. Both of these results suggest that the extent to which a representation is informative about underlying personality is a critical determinant of its activation. Understanding and predicting others is only one of the many processing orientations that perceivers may adopt in social perception. Another important goal is that of simplifying the social environment. It is now well established that grouplevel representations are more likely to be developed, activated, and applied as st[ is not clear exactly why college students find sex more informative than race, nor whether this preference would generalize to other populations. Nevertheless, the effect has been found in a number of experiments and appears worthy of further investigation.

384

CHARLES STANGOR AND JAMES E. LANGE

tasks become more complicated (Gilbert & Hixon, 1991; Macrae, Hewstone, & Griffiths, in press; Rothbart, Fulero, Jensen, Howard, & Birrell, 1978; Stangor & Duan, 1991), require more elaborate decisions (Bodenhausen & Lichtenstein, 1987), or are made under increasing time pressure (Kruglanski & Freund, 1983; Pratto & Bargh, 1991). Perceivers may also operate with particular contextsensitive goals that lead them to activate particular representations in given situations. For instance, Stangor, Lynch, Duan, & Glass (1992) found that subjects categorized others in terms of the style of their clothing when the task demanded it, but did not normally categorize on the basis of this information.

C. USE OF CATEGORY-BASED VERSUS BEHAVIORAL INFORMATION The previous section dealt with the issue of which of many competing socialcategory representations are activated at the time of judgment. However, category-based information represents only one source of data about others. An important competing source is the current behavior of the individual. It is known that both categorical and behavioral information serve as a basis of judgment, and that, depending upon context, either type of information may predominate. Locksley and her colleagues (Locksley, Borgida, Brekke, & Hepbum, 1980; Locksley, Hepburn, & Ortiz, 1982) focused on conditions in which behavioral information was most strongly used, finding that a target woman who acted assertively in a single situation was judged as equally dispositionally assertive as a man who performed the same behavior, even though the subjects clearly possessed the stereotype that men were more likely to be assertive, overall, in comparison to women. However, there are also instances where categorical information dominates. For instance, Neuberg and Fiske (1987) found that perceivers tended to base their judgments on a targets’ category label (“exschizophrenic”), rather than on personal information supplied by the target, unless they were given particular motivation to attend to the latter information. The distinction between using category-based and individuating information as a basis of judgment has been a continual theme within the area of stereotyping, and has recently been explicitly formulated in Fiske and Neuberg’s (1990) continuum model of impression formation. In this model, judgments are assumed to vary continuously between those entirely driven by category-based information and those driven entirely by behavioral (individuating) information. Fiske and Neuberg outline a host of variables that determine which predominates. One potentially informative way to conceptualize this situation is by considering behavioral information as another source of knowledge that may compete for momentary activation. Both category-based and behavioral information are infor-

GROUP REPRESENTATIONS AND STEREOTYPING

385

mative about personality only through reference to previously stored knowledge about the meanings of these observed data for underlying personality disposit i o n ~Knowledge .~ about the meaning of behavior for personality (including the necessity of adjusting for contextual factors in determining this relationship) is learned, just as is knowledge about the meaning of social-category membership. It is well known that children younger than 7 to 8 years do not make use of behavioral information to infer underlying personality to the same extent as do older children and adults, even though they may well understand the meaning of the behavior itself (e.g., Rholes, Newman, & Ruble, 1990; Ruble, in this volume). Although it may be more heuristic to conceptualize the process of deriving trait implications from observed behaviors in terms of procedures (e.g., E. R. Smith, 1989; E. R. Smith & Branscombe, 1987) rather than in terms of application of a knowledge representation, the two processes are in many ways parallel. In both cases a perceived stimulus evokes a mental process that results in the activation of one or more personality-relevant characteristics. Furthermore, it is the strength of activation of these characteristics that determines the extent to which the perceived information influences subsequent responses toward the person. Because the implications for personality of both categorical and behavioral information are dependent upon a stored knowledge base, the question of which is used under which conditions is fundamentally equivalent to that addressed in Section V,B-that is, which of many competing characteristics are most strongly activated at the time of judgment? Given the underlying similarity between the questions, it is expected that the same factors will determine whether categorical or behavioral information has the most influence upon judgment as determine which of many competing category representations are activated. It has frequently been proposed that, because of its high perceptual salience, category-based information will generally have more influence on responses to others than will behavioral information (Brewer, 1988; Fiske & Neuberg, 1990). This seems reasonable given that, when judging an individual, information about many social categories is easily visible and thus readily available for use, whereas behavioral information may or may not be available. On the other hand, when 91n this sense the traditional “top-down” versus “bottom-up” distinction is quite misleading in this domain. Moreover, whereas judging others on the basis of “individuating” rather than “categorical” information may be more unbiased or unprejudiced in many cases, it is not always more accurate. For instance, the well-known tendency to use the behaviors of others as a basis of judgment (cf. Nisbett & Ross, 1980) suggests that in some cases negative behavior may be given more weight than it deserves, leading to “unfair” or “overgeneralized” judgments. The extent to which categorybased versus individuating judgments are accurate is a neglected question worthy of research attention.

386

CHARLES STANGOR AND JAMES E. LANCE

both types of information are present in the stimulus array, behavioral information should be highly salient and thus in many cases can be expected to have a greater impact upon responses than should category information. In addition to salience, the informational value of a stimulus should also determine its impact upon judgments. To test this hypothesis, Krueger and Rothbart ( 1988) orthogonally manipulated the “diagnosticity” of categorical and behavioral information, determined through standard extremity measures, to be either high or low. They found that both types of information were used in judgment in direct relationship to their manipulated extremity. Conceptualizing the situation in terms of activation suggests that the perceiver may determine the diagnosticity of either a category or a behavior by determining how strongly the stimulus activates personality-relevant characteristics. When characteristics are strongly associated with the relevant category and thus strongly come to mind, or when the behavior strongly evokes the relevant trait, the stimuli are highly influential upon judgments, and are also rated as being extreme or highly diagnostic. Another variable that is known to influence the relative use of category-based versus behavioral information is the current processing goal of the individual. It has been shown that people rely more upon behavioral information and less upon categorical information when accuracy goals are increased through either cooperation (Ruscher & Fiske, 1990), competition (Neuberg & Fiske, 1987), or by directly encouraging accuracy (Neuberg & Fiske, 1987) or accountability (Tetlock, 1985). The effects of accuracy goals on judgments are easily derived from the current hypothesis in terms of the number of representations that are activated in a given situation. When accuracy goals are high, the perceiver will be motivated to process the information deeply and carefully by activating and searching many representations for relevant information (Kruglanski, 1989). Because the category label represents only one (although initially highly salient) representation, accuracy goals reduce its influence on judgments. Similarly, when the situation forces a reduction in the number of representations that can be activated, for example, through high perceived time pressure (Kruglanski & Freund, 1983) or fatigue (Bodenhausen, 1990), category-associated stereotypes have bigger effects upon responses. The finding that category-based information is used relatively less than is behavioral information under accuracy or accountability conditions may well be limited to cases where behavioral information activatesjudgment-relevant material more strongly than does the category label. For instance, in the Neuberg and Fiske (1987) experiments, the category label (“ex-schizophrenic”) may have been less informative than was the information the target used to describe himself. On the other hand, in cases where stereotypes are very highly developed, or when only categorical information is available about the target, accuracy goals

GROUP REPRESENTATIONS AND STEREOTYPING

387

would be expected to increase the use of the available information and thus increase reliance upon category information and stereotyping. Testing the hypothesis that accuracy goals increase the use of the most diagnostic information, regardless of whether it is categorical or behavioral, would appear to be an important avenue for future research.

D. THE INFLUENCE OF STEREOTYPE ACTIVATION ON SUBSEQUENT RESPONSES: ACTIVE VERSUS PASSIVE PROCESSING EFFECTS To this point we have argued that the relative accessibility of a stereotype in memory (and thus its likelihood of being used as a basis of responding to others) is multiplicatively determined by the extent to which the representation containing the stereotype is activated and the degree to which the stereotype is strongly associated with the category label. Furthermore, we have argued that the probability that a stereotype will influence judgment is determined by the relative activation of competing constructs in memory. This formulation is admittedly incomplete, however, for two reasons. First it suggests that activated constructs are always used in judgment. Second, it suggests that the influence of an activated construct will always be assimilative in nature. Neither of these assumptions is always true. The issue is best understood by differentiating the potential role of passive versus active processing effects in social judgments. These processes map respectively onto the traditional distinction between “automatic” and “controlled” processes (cf. Neely, 1977; Bargh, 1984; 1990), but in the current context the distinction is additionally informative because it specifies the direction of the likely influence of each type of process on information processing. Passive processes are those that invariably lead to assimilation of new information into the existing representation, rather than to accommodation of the representation to the information. When a category representation and its associated characteristics are activated, they tend to increase exposure to and memory for expectancy-confirming (versus disconfirming) information, lead to distortion of expectancy-disconfirminginformation to be confirming, and produce categoryconsistent affective responses (which produce assimilation-type effects upon judgments). Such “passive” effects have a pervasive influence because they occur regardless of the perceiver’s current goals and often operate out of conscious awareness. Generally, it is expected that more highly activated stereotypes will result in stronger passive processing effects, regardless of whether this activation is due to contextual (e.g., Stangor & Duan, 1991) or individual (e.g., Bem, 1981; Markus, Crane, Bernstein, & Sidali, 1982) differences. If the perceiver becomes aware of the activation of the representation and

388

CHARLES STANGOR AND JAMES E. LANGE

accompanying stereotypes, he or she may decide that the information is not applicable or should not be used as a basis of judgment, or potentially that the stereotypical beliefs should be modified. Active effects occur when the perceiver consciously controls the influence of the activated information. In one relevant study Leyens et al. (1992) gave their subjects categorical information about a target person, and through a bogus dichotic listening task led half of them to believe that they had also received behavioral information about the target (no relevant information was actually presented). The other half of the subjects received exactly the same categorical information and participated in the same listening task but did not think that the information they heard was descriptive of the target. Leyens et al. found that their subjects used the category information as a basis of judgment only in the conditions in which they believed that they had also received behavioral information. They interpreted these results as demonstrating that subjects preferred not to use the category label and the associated stereotypes to describe the target unless they could justify their use on the basis of having other “information” (cf. Darley & Gross, 1983). Thus these subjects, all of whom had an activated category label, actively controlled whether or not they used it as a basis of judgment. In contrast to passive processes, which are always assimilative in nature, active effects may result in either assimilation or contrast effects on judgment (cf. Lombardi, Higgins, & Bargh, 1987; L. L. Martin, Seta, & Crelia, 1990). For instance, when an individual engages in “reverse discrimination” this represents an active effect because the individual has consciously reacted to the influence of the activated representation, avoiding its use. Similarly, under some processing conditions individuals may actively focus upon expectancy-incongruent information, leading to a reversal of the default-congruency effect in memory (cf. Srull & Wyer, 1989; Stangor & McMillan, 1992; Footnote 3). There are also individual differences in the degree to which individuals actively use group representations as a basis of judgment, such as prejudice (Devine, 1989; Stangor, Lynch, Duan, & Glass, 1992) and self-monitoring (Fiske & Von Hendy, 1992). Active effects that occur in parallel with passive effects may, to the extent that they are contrastive in nature, cancel the assimilative effect of the former process. The assumption that people have control over the stereotyping process has led a number of researchers to investigate the role of active processes in stereotyping and prejudice (cf. Brewer, 1988; Devine, 1989; Gilbert & Hixon, 1991; Neuberg & Fiske, 1987). Our analysis of stereotypes in terms of mental representations predicts which of many potential attributes will be available for use in both active and passive processing. Because many of the important influences of stereotypes in everyday life are passive in nature, this represents an important understanding. However, although the current model places limitations upon which representations and stereotypes will be used in active processing (those that are activated), it cannot predict their influence upon judgment.

GROUP REPRESENTATIONS AND STEREOTYPING

389

VI. Stereotype Maintenance and Change The research we have discussed to this point has been primarily concerned with investigating how representations of social groups develop and their use in information processing. However, the structural approach has also proved to be informative in promoting an understanding of why stereotypes maintain themselves even in the face of conflicting information, and in predicting how and when they will change. In terms of stereotype change, the structural approach also suggests methods of changing stereotypes that go beyond those offered by traditional models of stereotyping.

A. STEREOTYPE RIGIDITY The unyielding nature of stereotypes has frequently been evoked as one of their defining criteria and has remained one of the major puzzles of those interested in intergroup relations. If one accepts the hypothesis that beliefs about social groups are represented in abstract summary form, however, their resistance to change is not the least bit surprising. Over time, individuals have direct contact with members of social groups and also learn about the groups through secondary sources. This information tends to increase the strength of association between group representations and stereotypical characteristics because direct contacts are often confirming of the stereotype (the “kernel of truth”) and because secondary source information (e.g., humor; the media) generally supports existing stereotypes. Furthermore, passive processes that tend to produce assimilative effects on information processing also lead to a strengthening of stereotypical associations. As label-stereotype associations become stronger as a result of these processes, they naturally become more difficult to subsequently undo. This of course suggests that preventing stereotypes from initially developing will be a much more successful strategy overall than will trying to change them once developed. The present approach also makes explicit the expectation that, even if stereotype-inconsistent information is encountered, attended to, remembered, and used to update the representation, it is unlikely to have a large effect on overall beliefs about the group. Over time, representations about frequently encountered social groups are strengthened by many stereotype-confirming experiences, each amplified by passive processing effects. Against these strong confirmatory effects, exposure to some stereotype-incongruent individuals is not likely to produce much change in the abstract group representation (cf. Rothbart & John, 1985). In addition to being influenced by cognitive processes that tend to maintain

390

CHARLES STANGOR AND JAMES E. LANCE

stereotypical beliefs, individuals are also likely to be highly motivated to maintain their expectations about social groups intact, thus making those expectations even more resistant to change. Social-group representations (like all cognitive representations) develop in order to meet important functions for the individual. Because perceivers rely upon their representations to inform their everyday social perception, they are naturally reluctant to abandon them. Rather than giving up their stereotypical beliefs, individuals will go through the elaborate processes of “refencing” (Allport, 1954) or “subtyping” (Weber & Crocker, 1983) to keep their representations intact in the face of disconfirming evidence. Thus both active and passive processes work together to maintain stereotypical beliefs, resulting in a formidable resistance to change.

B. CHANGING STEREOTYPES Approaches to changing stereotypes have traditionally been based upon attempts to alter the content of the group representation. In the current approach changes in content involve altering the strength of association between category labels and associated characteristics, either by weakening associations between the category label and stereotypical traits or by strengthening associations between the category label and nonstereotypical characteristics. Of course, as discussed above, contact with a few disconfirming individuals is not likely to produce dramatic changes in the content of the overall group representation. Although we are in agreement that changing stereotype content can, in some cases, be an effective means of changing stereotypes, the current approach makes rather broader predictions about the change process and suggests other approaches that may be more successful. According to our predictions about stereotype activation, the influence of a stereotype on judgment is positively related to its strength of association with the category label and negatively related to the number of characteristics within the representation. Thus increases in the number of characteristics associated with the category label, even if those characteristics are evaluatively neutral, will reduce the likelihood that any one (stereotypical) characteristic will be influential on judgment. This effect, commonly known as stereotype dilution, has been demonstrated in a number of experiments (cf. Hilton & Fein, 1989). Thus the harmful effects of stereotypes may be reduced by adding new beliefs about the group, as well as through the more traditional technique of eliminating existing beliefs. It is possible that the former approach may be more successful overall. In addition to reducing the likelihood that negative characteristicsare activated by the category label, increasing the number of Characteristics associated with the label also reduces the informativeness of the representation as a basis for understanding others (cf. Rothbart & John, 1985). Representations with many

GROUP REPRESENTATIONS AND STEREOTYPING

39 1

weakly associated attributes will be perceived as less informative, and thus will be less strongly activated in judgment situations, in comparison to other representations with fewer and stronger associations. The joint effects of associative strength and number of associated characteristics in determining responses to group members has been well documented in the heavily researched literature concerning differential perceptions of ingroups and outgroups. The associational model provides a parsimonious explanation for these differences, based upon strength of association and number of associated characteristics. Through frequent or more in-depth contact with ingroup members (or potentially as the result of other factors such as stronger motivations to accurately characterize them), ingroup representations come to contain a large number of weakly associated stereotypes, whereas outgroup representations are generally made up of fewer, but more strongly associated, characteristics. As a result, when people are judging the outgroup members (or the outgroup itself), the few stereotypical items come quickly and strongly to mind and do not receive much competition. This leads to highly stereotypical (extreme) judgments of outgroup members on these characteristics and to perceptions of the outgroup itself as extreme and homogenous (cf. Linville, Salovey, & Fischer, 1986; Park, Judd, & Ryan, 1991). In comparison, when thinking about the ingroup or about individual ingroup members, many different characteristics are simultaneously activated, resulting in less extreme judgments and perceptions of high variability among group members. 10 In addition, because outgroup (versus ingroup) representations are perceived as highly informative, they will be more strongly activated as a basis of judgment. For this reason, categorical judgments are more likely for outgroup than for ingroup members overall (cf. Turner, 1987; Wilder, 1986). One of the basic goals of stereotype change, then, is to make outgroup representations more similar to ingroup representations. Although it is not strictly speaking a process of stereotype change, another approach to reducing the negative outcomes of stereotypes on responses to others is to influence which representations are activated upon contact with individuals. A reduction in the likelihood that a given (undesirable) representation will be activated can occur by activating other representations, either at the same level, at a broader level (e.g., as “human beings”; cf. ’hmer, 1987), or at lower (subtype) levels. Many of these processes have been discussed by N. Miller and I01n this sense the associational model provides a means by which group variability can be (indirectly) represented in memory. Note that this model assumes that variability is calculated online, when required, by assessing the strength of activation of stereotypical characteristics. More strongly activated characteristics are judged to be less variable within the group. If strength of association determines perceived variability, then it seems unlikely that perceived variability will contribute to judgments of individual members of social groups over and above that of perceived extremity. This hypothesis awaits a test.

392

CHARLES STANGOR AND JAMES E. LANGE

Brewer (1986), so we do not dwell on them here. The factors that determine representation activation, and thus which should be useful in effecting this type of stereotype change, have been enumerated in Section V. Heretofore, we have proceeded on the assumption that representations change quantitatively in terms of either number of characteristics or their strength of association as they develop and change. However, increasing group contact or other forms of learning about the group may produce qualitative changes as well, specifically relating to whether representations are organized in terms of social groups or on other bases. It is known that increasing familiarity with a target group increases the likelihood that representations will be formed around persons (Brewer, 1988; Park, 1986; Pryor & Ostrom, 1981) or around subgroups (Park, Ryan, 8t Judd, 1992), rather than around the broader level social category. Similarly, it is well known that representations of ingroups (with which one has plenty of experience) are more likely to be organized around persons, whereas outgroup representations are organized around other types of categorization schemes (Ostrom, Carpenter, Sedikides, & Li, 1993). These findings have potential implications for the process of stereotype change, but their role in this process remains to be determined. Again, we recommend that future research focus explicitly upon the functions of stereotypes to the perceiver. Just as group representations develop to meet functional goals, they will only be able to be changed to the extent that the new representations (formed at whatever level and around whatever concepts) are equally or better able to meet these goals.

VII. Future Issues: On the Sufficiency of the Abstractionist Approach The theme of the present review is that being more specific about how social groups are represented in memory has produced substantial progress in understanding the basic processes of stereotypes and stereotyping. We have seen that conceptualizing stereotypes in terms of mental associations between group labels and trait characteristics provides a convenient method of thinking about stereotypes and produces a number of theoretical benefits. These include specification of a standard operational definition of a stereotype and appropriate measurement techniques, and prediction of when stereotypes will or will not be used as a basis of judgment. The approach is also informative about the basic issues of stereotype development and change. Although there are still many areas that are less than perfectly understood, and our current models are undoubtedly overly innocent, much progress has indeed been made. In Section VII we discuss the current limitations of and the possible future for this approach, by working in two directions. In Section VII, A we take a micro-

GROUP REPRESENTATIONS AND STEREOTYPING

393

analytic approach, considering the ways in which future work may get still more specific about representational structure. Then we turn to a more macro-analytic level, returning to a consideration of the relationship between the grouprepresentational approach and more traditional models of stereotyping, noting particularly where the current approach has less than adequately captured traditional notions of stereotypes.

A. GETTING STILL MORE SPECIFIC: REPRESENTATIONAL CONTENT As we have noted, issues of process have recently tended to outweigh issues of content in the area of stereotypes and stereotyping, and as a result less is known than we would like about the exact content of social-group representations. Consequently, we have often been intentionally vague in our discussion of the contents of the representations that form the basis of this review. We have suggested that characteristics are stereotypical to the extent that they are associated with a category label in memory. However, we have not been very specific about what these characteristics and category labels are. Furthermore, we have not speculated about how these representations are related to one another and to other types of knowledge. In terms of category labels, we have focused primarily on stereotypical expectations about social groups or categories. Although this focus has been dominant within the literature, mostly because of the persistent concern with problems of sexual and ethnic stereotyping, social categories come in different forms, and it may prove useful to consider explicitly these differences in future research. Although we know of no research that has addressed this issue, it seems likely that mental representations are packaged together in interrelated sets, which supply our stereotypes for a given domain. Some stereotypes are determined by a set of multiple representations-for instance those that account for our beliefs about similarities within and differences among members of various racial groups. For most U.S. citizens, this would presumably include representations of African-Americans, EuroAmericans, Hispanics, Asians and so forth. Other domains (for instance sex-at least as conceptualized in western societies-and the ingroupourgroup or wethey distinction) are accounted for using only two, presumably interrelated, representations. In still other cases the feature that provides the basis of the stereotype is continuous in nature, such as age, (Brewer, Dull, & Lui, 1981), physical attractiveness (Landy & Sigall, 1974; C. T. Miller, 1988), and “babyfacedness” (Berry & McArthur, 1986). Although it is often implicitly assumed that such continuous features are actually represented by discrete (either multiple or dual) categories (e.g., “old people,” “young people”), it is possible that

394

CHARLES STANGOR AND JAMES E. LANGE

knowledge about continuous features such as age are stored in a single representation reflecting the perceived correlations between the feature and relevant personality characteristics. Although the hypothesis is admittedly speculative, it is possible that different types of representational sets may have substantially different influences upon processing. For instance, dual representationsare likely to be particularly strongly developed because they maximize category distinctiveness (the ability to contrast categories with one another). The utility of dual-category stereotypes could help explain why sex is repeatedly found to be the most common basis of social categorization (Zkate & Smith, 1990; Stangor, Lynch, Duan, & Glass, 1992) and why the we versus they distinction has such strong effects on responses to others (Perdue, Dovidio, Gurtman, & Qler, 1990; Tajfel, 1981; Turner, 1987). On the other hand, continuous categories may be used less routinely, perhaps because it is more difficult to categorize individuals using these features because comparisons are less obvious, or because they are perceived as less informative overall. In any case it seems clear that a truly general model of stereotype representation will be able to account for how characteristics are associated with social-group labels, and how different group representations are themselves interrelated, regardless of whether the feature upon which the stereotype is based is categorical or continuous in nature (cf. Linville et al., 1986). The above discussion suggests that we do not yet have as much evidence as would be desired concerning how different group representations are related to one another in memory. We have argued that representations within the same domain, such as men and women, seem likely to be organized in parallel. However, across domains either parallel or hierarchical organization is possible. For instance, it is not known whether people generally possess a “black men” representation that is nested under the broader representation for “men,” or only separate representations for “blacks” and “men” that are activated in parallel (cf. Stangor, Lynch, Duan, & Glass, 1992). Existing data addressing this issue are not conclusive: Cantor and Mischel ( 1979) hypothesized that the structure of group representations would be hierarchical, but their data were rather equivocal in this regard. Lingle, Altom and Medin (1 984), on the other hand, argued that social representationsare less likely to be hierarchical than are object categories, and research by Deaux, Winton, Crowley, and Lewis (1989, which failed to find evidence for superordinate- and subordinate-level social-group categories, supports this prediction. Devine and Baker (1991) found evidence that people could easily generate subgroups for a broad social category (blacks), but provided no evidence that these subgroups were arranged hierarchically in memory. Brewer et a]. (1981) used a sorting task to demonstrate that subcategories are more informative than are higher level categories, and argued that they should be more frequently used in everyday

GROUP REPRESENTATIONS AND STEREOTYPING

395

perception. Andersen and Klatzky (1987) have provided data concerning the relative hierarchy of stereotypes and trait terms. Recent research by Park et al. (1992) suggests that subgroups are indeed associated with superordinate-level groups in memory because judgments of subgroups are related to judgments about superordinate-levelcategories. Because much speculation about the nature of stereotype change is based on the relationships among categories and subcategories (cf. Johnston & Hewstone, 1992; Rothbart & John, 1985; Weber & Crocker, 1983), more work in this area is sorely needed. The existing data represent a beginning in this regard, but formal models about the relationships among mental representations of real-world social groups are still lacking. The neglect of content has also left us with substantial lacunae in our understanding of what types of characteristics are contained within stereotypical representations of social groups. We have proceeded in this review on the assumption that stereotypes are primarily trait terms. This premise seems justified given that stereotypes have traditionally been considered to be traits and that people primarily generate trait terms when asked to describe both persons (Fiske & Cox, 1979; McGuire & McGuire, 1988) and groups (Stangor, Sullivan, & Ford, 1992). Nevertheless, some theorists have argued that stereotypes may include more than trait terms, for instance, beliefs about others’ attitudes and values (Judd et al., 1991), attitudinal or affective responses to the group (Fiske, 1980; Stephan & Stephan, 1993), behavioral tendencies (Stephan & Stephan, 1993), and pictorial representations (Brewer, 1988). One advantage of considering stereotypes in terms of mental associations is that it provides a method for determining what is stereotypical. If attitudes, emotions and behavioral tendencies are indeed contained in the representation of a social group, they should be activated when the category label is primed and it should be possible to measure this activation using appropriate paradigms (Cooper & Singer, 1956; Fiske et al., 1987). We also do not have much information about when representations will be formed on the basis of social groups, rather than on the basis of other competing categorization systems. We do know that stimulus complexity increases the likelihood of forming group-based (rather than individual) representations (cf. Rothbart et al., 1978; Stangor & Duan, 1991). We expect that a more explicit consideration of the processing goals of individuals when forming social impressions will be informative regarding what types of representations are formed under what circumstances. The pragmatic perceiver can be expected to develop and use representations in a highly flexible manner, for instance to communicate with others (Farr & Moscovici, 1984; Hogg & Abrams, 1988) and to predict other’s behavior (Oakes & Turner, 1990). Some underlying motivations may increase the likelihood of forming group-based representations, whereas other processing goals may favor the formation of representations based upon other classification schemes.

396

CHARLES STANGOR AND JAMES E. LANGE

B. A COMPETING APPROACH: EXEMPLAR MODELS Our review has been based upon the assumption that abstracted representations of social groups are formed through experience and have substantial effects upon subsequent information processing. We have interpreted the results of many experiments as supporting the abstract-associationalmodel, and noted the ability of the model to account for a broad range of findings. It is our belief that associational models (in one form or another) represent the most useful approach currently available to understanding the stereotyping process, and this belief is echoed by many in the field. Recently another broad approach to mental representations of social groups has been proposed in the form of instance-storageor exemplar models of memory. The most well-developed of these is E. R. Smith and ZArate’s (1992) elegant model. The Smith and Zhrate model makes many of the same predictions about social judgment (particularly in terms of its extreme flexibility among perceivers and social contexts) as do associational models, but does so in a way that is directly counter to the “abstractionist” approach. According to the exemplar approach, no abstracted representations of social groups are ever formed or stored in memory; rather, only specific individuals or behaviors themselves are stored. Thus in this formulation stereotypes about social groups do not exist a priori, but are, rather, calculated on-line when called for. Judgments of individuals and groups are made by retrieving stored exemplars from memory, rather than by retrieving an abstract representation. The relative ability of exemplar versus abstraction models to account for observed data has recently received much attention in the cognitive and socialcognitive literatures (See Lingle et al., 1984; E. R. Smith, 1990; Smith & ZArate, 1992; Srull & Wyer, 1990 for summaries). Much (if not all) data that can be accounted for in terms of abstraction models may also be explained in terms of exemplars. For instance, priming experiments showing that exposure to a category label activates category-associated trait terms may be accounted for by assuming that the primed category label activates memories of individual category members, and these exemplars are associated with and activate the relevant traits. Although this explanation appears less parsimonious than the hypothesis that traits are directly associated with category labels in memory, we know of no existing data that support one model over the other in terms of priming effects. Some findings seem more supportive of abstraction than of exemplar models. For one, the associational hypothesis appears to provide the most straightforward account of the robust finding of preferential memory for expectancy-congruent information about social groups. Although it may well be possible to expand exemplar models to provide a mechanism by which a category label can increase the ability to encode and retrieve material that is congruent with the label, such models have not yet been developed in the social domain. Associational models,

GROUP REPRESENTATIONS AND STEREOTYPING

397

however, are already well developed in this regard (cf. Srull & Wyer, 1989). It is also known that there is rarely an observed relationship between recall of observed behaviors and trait judgments about targets (Dreben, Fiske, & Hastie, 1979; Hastie & Park, 1986), except under extreme overload conditions (Bargh & Thein, 1985) when abstracted representations cannot be formed. Although these findings do not necessarily contradict exemplar approaches (because they do not assume that individuals can consciously retrieve the stored exemplars), abstraction models have the advantage of making the relevant prediction explicitlyrecall and judgment will not be correlated when abstractions have been formed, because memory for the original stimulus information is not stored (cf. Hastie & Park, 1986). Individuals have no trouble expressing stereotypes about groups with which they have had only little (Park & Hastie, 1987) or no (Kanungo & Das, 1960) direct contact. Although people may generate and store exemplars from hearsay about group characteristics (one could develop an exemplar of an Albanian from hearsay about Albanians), this process seems more complicated than does forming abstract representations, and the extent to which it occurs is not yet clear. In addition to empirical findings that are currently better accounted for in terms of abstraction models, there are other theoretical reasons to prefer them at this time. For one, stereotypes have traditionally been conceptualized as established beliefs about social groups. The idea that such knowledge is never formed or stored is thus contrary to established conceptualization. Although this does not mean that the established view is correct, it will be necessary to demonstrate that exemplar models make significantly better predictions than do abstraction models if the latter are to be discarded. The abstraction approach also provides a theoretical bond between stereotyping research and that in other areas of social behavior. For instance, much theorizing has been based on the assumption that attitudinal beliefs and values are stored in interrelated cognitive structures (Fazio et al., 1983; Judd 8z Krosnick, 1982). Furthermore, associational or schematic models have traditionally been used to describe the self-concept (Higgins et al., 1988; Markus, 1977; Rogers, Kuiper, & Kirker, 1977), the social situation (Schank & Abelson, 1977), and self-other relationships (Baldwin, 1992). The general approach to person perception within social psychology has been associational in orientation (cf. Ashmore, 1981; Schneider, 1973), and thus abstraction models are derivative of that basic tradition. Exemplar models have not heretofore been commonly used in these domains, and have been shown to be less adequate than abstraction models in terms of understanding the self concept (Klein & Loftus, 1990). Abstraction models also account very well for the rigidity of social stereotypes in contact situations where they might be expected to be less than stable (Rothbart & John, 1992; Stephan, 1985). In abstraction models a recently encountered individual is aggregated into the group representation, whereas exemplar models

398

CHARLES STANGOR AND JAMES E. LANGE

suggest that newly encountered stereotype-disconfirming individuals (who should be highly accessible and thus have disproportionate weight in judgment) should produce substantial changes in group beliefs. This type of stereotype change is, however, notably difficult to effect. Postulating cognitive associations between category labels and stereotypes also accounts for the abundance of accessibility-driven passive processing effects (such as biased memory, information search and interpretation) that occur so regularly in social-information processing when beliefs are spontaneously activated upon contact with group members. The exemplar approach also seems unparsimonious because of the number of pieces of information that must be stored, although these models assume that this storage and retrieval occurs automatically, in parallel, and is not highly effortful. As we have noted earlier, however, abstraction models also require an embarrassing number of stored constructs. For instance, do people really possess unique representations for tall, attractive, 30-year-old women with black hair? (cf. Smith, 1990). Although exemplar models avoid this problem of indefinite nesting by assuming that only the most specific information is stored, it is possible that at least some effects could be accounted for using the opposite approach, namely by assuming that only the highest-order representations are stored and that judgments are generated by simultaneous activation and appropriate combination of multiple representations (cf. E. E. Smith & Osherson, 1984). The present model, which assumes that multiple representations are generally activated in most situations, is quite amenable to such a process. Abstraction approaches are also very consistent with the idea that individuals develop and use theories about the relationships between social categories and personality characteristics as a basis for informing, constraining, and modifying social beliefs (cf. Murphy & Medin, 1985). For instance, beliefs about the elderly may be supported and constrained by knowledge of the physical and mental changes that occur during the aging process. The “package” of stereotypes about the elderly (frail; forgetful; hard-of-hearing) can be derived, at least in part, from knowledge about what happens to people as they age. Similar theories abound about the characteristics of minority groups (for instance, characteristics that develop when people are systematically deprived of important resources) and religious groups (characteristicsbased upon the fundamental tenets of their beliefs). Althmgh the relationship between such theories and stereotype development have unfortunately not yet been fully addressed, abstraction models provide a direct means by which this process could occur, because the relevant knowledge is derivable from, and informed by, the content of the representation itself. Exemplar models, on the other hand, although amenable to the idea that existing theories can guide which exemplars are stored in and retrieved from memory, do not themselves provide a means for explaining how relationships among the various stereotypes of a given group are represented.

GROUP REPRESENTATIONS AND STEREOTYPING

399

Finally, and perhaps most important, exemplar models have not yet fully addressed the issue of stereotype development. We have seen that abstraction models, such as the one proposed here, can be used to explain which characteristics will become stereotypical of which groups, and why (i.e., because of their informational value). Exemplar models, on the other hand, are based on the assumption that stereotypical characteristics are connected or stored with the exemplars in memory, and are activated with the exemplars, but do not address the important issue of exactly how and why one characteristic, rather than another, is so stored. Of course accepting the conclusion that social-group information is stored at least partially in the form of abstract group representations does not preclude the possibility that we also remember specific individuals. We obviously do remember people whom we have encountered in the past, and those memories do influence our responses to others (Andersen & Cole, 1990; Gilovitch, 1981; Lewicki, 1985). For instance, Andersen and Cole (1990) found that subjects used the personality characteristics of a close friend to draw inferences about whether traits had previously been presented in a recognition memory task. Furthermore, exemplar models prove better at predicting categorization decisions than do prototype models in some cases (E. R. Smith & Zirate, 1990). At this point, however, given the basic similarities in the predictions made by exemplar and prototype models, the difficulty (if not impossibility) of distinguishing them both conceptually (E. R. Smith, 1990) and empirically (Barsalou, 1990), and the possibility that representations may themselves be generated on-line when needed (Barsalou, 1987), we find ourselves agreeing with Cantor and Mischel’s (1979) earlier assessment that it is not yet clear “how important this difference [between exemplar and prototype representation] will turn out to be” (p. 29). Assuming that both abstracted and specific information are stored in memory (cf. Kahneman & Miller, 1986; Messick & Mackie, 1989; Park & Hastie, 1987; Sherman, Judd & Park, 1989), one of the important questions that will need to be addressed concerns the recurring issue of how individual exemplars are related to abstract representations of social groups. If I know Bill, who is also a college professor, do I represent “professor” as an attribute of Bill, Bill as a member of the category of college professors, or both? Put another way, the question concerns whether exemplars are stored at all levels of the representation, or whether they always represent the bottom level in the hierarchy (cf. Andersen & Klatzky, 1987). These different possibilities have been well elaborated in an important chapter by E. R. Smith (1990), but the answers remain unknown. Also less than adequately understood is the question of when judgments will be based upon the retrieval of exemplars and when they will be based upon the retrieval of abstracted representations. These issues have obvious implications for stereotype change upon encounters with new individuals (Rothbart & John, 1985). For instance, the finding that

400

CHARLES STANGOR AND JAMES E. LANGE

stereotype change is generally greater when (the same amount of) stereotypedisconfirming information is dispersed throughout many individuals rather than being concentrated in a few highly discrepant individuals (Johnston & Hewstone, 1992; Weber and Crocker, 1983) suggests that some form of exemplar storage is occurring in which the number of disconfirming individuals is being represented. Future research along this line could profitably be framed in terms of the exemplar/abstraction distinction. One potential approach to understanding these relationships is through the use of thought-listing tasks. Individuals, when appropriately prompted, should be able to generate exemplars of social categories, and such data, perhaps analyzed using clustering approaches, should provide information about the relationship between categorical representations and exemplars in memory.

C. REINTEGRATING CLASSICAL CONCEPTIONS: FEELINGS, CONSENSUS, AND ACCURACY It has often been argued that the recent infatuation with cognitiveindividualistic approaches within social psychology has had the unfortunate result of de-emphasizing the truly social aspects of relationships among individuals. Although we have seen that associational models provide a method for integrating many of the traditional “defining features” of stereotypes, such as generalization, distinctiveness, and homogeneity, other traditional notions have not yet been so easily assimilated. For instance, we have not yet considered the traditional notions of overgeneralizations, unjustijied beliefs, exaggeration, and inaccuracy, (cf. Ashmore & Del Boca, 1981), nor have we addressed the issue of societal consensus. The problem is that stereotypical representations are not just generic cognitive representations held by individuals-beliefs about social groups are inherently different than beliefs about houseplants or station wagons, particularly because they have important implications for intergroup behavior. These issues have long been at the heart of theorizing about stereotypes and stereotyping, and are particularly crucial in terms of understanding the relationships among stereotypes, prejudice, and discrimination.

1 . Motivation and Affect: Stereotypes Involve Feelings In contrast to the cognitive emphasis of contemporary models, the motivational needs and goals of the perceiver played a more major role in all traditional models of intergroup relations. In these approaches, stereotypes and prejudice were considered to be the result of such needs as self-esteem, releasing unwanted

GROUP REPRESENTATIONS AND STEREOTYPING

40 1

aggression, and societal acceptance. Considering stereotypes as cognitive representations may seem antithetical to this basic conceptualization, and there is no question that recent theories have neglected motivational and affective variables (outside of the basic motivation of the cognitive miser-that of simplifying the social environment). However, traditional models of stereotyping were also incomplete because they did not provide theoretical mechanisms to account for how motivations influence stereotype development and their use in responding to others. One important advantage of associational models is that they provide a framework for understanding the relationship between motivation and cognition. Most basically, this relationship is derived from the expectation that individuals will develop, maintain, and use as a basis of judgments those representations that best meet their processing goals. Stangor and Ford (1992) have taken one approach to this issue, arguing that many different underlying motivational needs influence stereotyping by leading perceivers to approach information in either an “accuracy-oriented‘’or an “expectancy-confirming” manner (cf. Fiske, 1993). It is well known that these basic motivations influence the nature of representation formation and their use. Broadening our knowledge about the influences of motivations on representation development and use from the avidly researched “simplifying the social environment” to “accurately understanding others,”“protecting self-esteem,” “communicating with others,” “justifying existing attitudes or societal conditions,” and so forth represents an important integrative task, but one to which a structural approach seems very amenable. In addition to providing a link to motivations, the associational model is also integral for understanding the role of affective and emotional states in stereotyping. The link between feelings and cognition is derived in several ways. For one, affective and emotional states influence the accessibility of constructs (such as group representations) in memory (Forgas, 1992; Forgas & Bower, 1987; Stangor, 1990), and thus may influence both which of many competing social representations are primed (Erber, 1991) and potentially which characteristicswithin a given representation become activated (Esses et al., 1993) in any given context. Thus if encounters with outgroup members tend to produce negative affect (perhaps because such encounters are more unusual or uncomfortable than interactions with ingroup members), these emotional states may lead to even more negative responses to the target by priming negative representations and negative stereotypes. Affect may also influence the extent to which the individual exerts information-processingeffort and thus determine the number of different representations that are activated (cf. Mackie & Worth, 1989). Thus under strong affective states, the use of highly salient group representations (rather than, for instance, individuating information) may be increased. The associational approach also provides a clue to the source of these affective states. Affective and emotional reactions become associated with social-group

402

CHARLES STANGOR AND JAMES E. LANGE

labels through both primary and secondary learning processes (Bower & Cohen, 1982; Clark, Milberg, & Erber, 1988; Fiske, 1980; Fiske & Pavelchack, 1986), and are activated upon subsequent exposure to group members (cf. Cooper, 1969; Cooper & Singer, 1956; Fiske et al., 1987). When affect and arousal are elicited upon exposure to group members, they will naturally influence how target information is interpreted, how targets are responded to, and they may affect whether the perceiver interacts with the members of the target group in future situations. Thus the basic conceptualization of stereotypes in terms of abstract-knowledge representations provides a useful and heuristic vehicle for studying the role of motivations and emotions in stereotyping. The chapters in a recent volume by Mackie and Hamilton (1993) present a summary of current research linking affect and stereotyping. Although much is known about the role of mood states as determinants of information processing, in everyday intergroup interactions stronger motivational and emotional states, such as guilt (Devine, Monteith, Zuwerink, & Elliot, 1991; Gaertner & Dovidio, 1986), anger, fear, and pride are probably more relevant. Theorizing about the relationship between emotional and physiological responses to social groups and the development and change of group beliefs represents an extremely important line of inquiry. 2. Stereotypes as Consensual Beliefs Perhaps the most important traditional defining feature of stereotypes is the notion of societal consensus (cf. Gardner, 1973, 1993; Gardner, Kirby, & Findlay, 1973). In all traditional approaches, characteristics were considered stereotypes only if they were shared by the people within a given society. For instance, Katz and Braly (1933) explicitly assessed consensus by computing the average number of traits it took to account for 50% of their subjects’ responses on their checklist. Across groups it took an average of only 8.5 traits (out of 84 presented traits) to reach this criterion, leading them to conclude that stereotypes are indeed highly consensual and thus that they must be learned from social sources (hearsay, media), rather than through face-to-face contact. More recently, it has been argued that societal consensus is not a necessary feature of stereotypes. Ashmore and Del Boca (198 l), for instance, proposed that the consensual or cultural stereotype was no more important than the individual stereotype (unique, individual beliefs about characteristics of social groups), and researchers have more frequently been assessing the latter (Devine, 1989; Eagly & Mladinic, 1989; Esses et al., 1992, 1993; Stangor, Sullivan, & Ford, 1992), particularly when using free-response formats. There is no question that individual beliefs about social groups exist and that these beliefs may differ from the consensual, cultural stereotype (cf. Devine, 1989, and Stangor, Sullivan, & Ford, 1992 for demonstrations of this discongruity). The important question

GROUP REPRESENTATIONS AND STEREOTYPING

403

here, however, is not whether individual beliefs exist, but rather whether these beliefs should be considered to be stereotypes. Despite recent attempts to do away with the consensus criterion, and despite the fact that consensus may not always be observed in perceptions of existing social groups, it is simply not possible to consider stereotypes without assuming that they are shared, at least to some extent, among individuals. For one, the precise danger of stereotypes to members of the stereotyped social group is that, because stereotypes are consensual, outgroup members tend to respond to them in similar ways. If each individual reacted to stigmatized group members according to his or her own idiosyncratic beliefs, the negative aspects of stereotypes for the stigmatized would be greatly reduced. In addition to its important real-world implications, consensus is integral to many of the most important current social-psychological theories about the development and use of stereotypes. According to the models of stereotype development that we have reviewed in this paper, stereotypes are learned through common social processes that serve to augment and distort existing intergroup differences. The assumption that stereotypes are learned and potentially changed through experience (either direct, or potentially indirect through hearsay) with members of social groups underlies otherwise diverse analyses of the stereotype development process proposed by, among others, Crocker, Hannah, and Weber (1983), Eagly and Steffen (1984), Hamilton and Gifford (1976), McArthur (1982), Linville et al. (1986), Rothbart and John (1985) and Smith and Zhrate (1992). This fundamental assumption also forms the basis of the well-known contact hypothesis regarding stereotype change (cf. Cook, 1978; Hewstone & Brown, 1986). According to all of these approaches, stereotypes are the result of actual intergroup encounters. Because these common social processes operate on the same objective data, stereotypes must come to be societally shared. That there be at least some consensus among individuals is of absolute importance for social-psychological models because it allows predicting responses to group members at a group level. If each person developed his or her own unique set of expectations about the personality characteristics associated with a given social category, traditional stereotyping measures would be useless. Given the substantial number of experiments that have been successful in predicting behavior toward others on the basis of shared social stereotypes, it is apparent that there is in fact at least some "Of course this consensus will not be perfect. Individuals disagree substantially even on their perceptions of common object categories (Barsalou, 1987), and they will certainly not be in perfect agreement on their perceptions of social groups. Distinguishing between those beliefs about social groups that are consensual and those that are individual is an extremely important goal for future research, particularly to the extent that consensual beliefs are generally more strongly associated with the category label in memory and thus held with more confidence, have more influence upon responses to others, and are more difficult to change.

404

CHARLES STANGOR AND JAMES E. LANGE

agreement across individuals, although this agreement may be quite small (cf. Jussim, 1991). There are two important implications of the consensus notion in the current context. For one, researchers who use free-response measures to assess stereotyping must consider the extent to which their measures are tapping true stereotypes. Although we have argued that there are many advantages to the freeresponse approach, its individual nature has the danger of de-emphasizing the cultural aspects of stereotypes unless content coding of the responses in terms of consensus is performed. This coding is difficult, but we believe it will produce highly informative data about stereotypes. Secondly, being explicit about consensus will increase our ability to construct more general models of the stereotyping process by connecting social-cognitive approaches with European models of social representation that are explicitly based upon the societal, and thus consensual, aspects of stereotypes (e.g., Farr & Moscovici, 1984; Hogg & Abrams, 1988). This integrative task, long overdue, represents a very important goal for a complete understanding of intergroup relations.

3 . Stereotypes as Inaccurate or Overgeneralized The above discussion of the extent to which current approaches assume (if only implicitly) that stereotypes are learned by observing group members highlights a final conceptual change that has become remarkably apparent over the past 20 years. This is a general renunciation of thinking about stereotypes in terms of accuracy or inaccuracy. Earlier notions of stereotypes in terms of “overgeneralizations,” and as products of “inferior judgmental processes” have disappeared from virtually all recent conceptualizations(see Table I). Up to the time of Brigham’s 1971a paper, however, the notion of inaccuracy was paramount. For instance, Brigham noted that stereotypes could be incorrect either in the sense that they were in the “wrong direction” (i.e., a person who thinks that most members of Group X possess Trait Y when in fact most members of the group do not possess the trait), or in terms of magnitude (a person who estimates that 95% of the members of Group X have the trait when only 65% really do). In fact Brigham developed a measure of unjustijed generalizations to assess inaccuracy in stereotype assignment, arbitrarily designating traits that were assigned to over 80% or less than 20% of group members as inaccurate stereotypes. The subjectivity of this procedure highlights the difficulty of defining criteria for accuracy or inaccuracy. The problems in making inaccuracy a defining feature of stereotypes are severalfold (cf. Ashmore & Del Boca, 1981). First, neither agreement among a group of people regarding the stereotypes of another group, nor demonstrations that stereotypical beliefs come from hearsay rather than from direct observation

GROUP REPRESENTATIONS AND STEREOTYPING

405

imply that such beliefs are incorrect. And even the use of autostereotypes (selfreports of the traits that are characteristic of the individual or the ingroup) as accuracy criteria (e.g., Judd et al., 1991; Vinacke, 1957) provides only a weak test of accuracy. There is no particular reason to believe that people have a more accurate conception of the traits that are true of themselves (or of their own group) than does another group viewing them. People may be just as likely to underestimate or overestimate their own characteristics as others are to do so. Virtually the only real hope of testing the accuracy hypothesis is to investigate assignment of characteristics the prevalence of which can be empirically confirmed. For instance, McCauley and Stitt (1978) compared whites’ estimates of the percentage of blacks who were on welfare or who had dropped out of high school to the actual numbers shown in government statistics. The data showed that the perceptions were indeed inaccurate, but consistently in the direction of underestimating the actual prevalence of the characteristics in the relevant populations. Although the types of characteristics that can be assessed in this manner are limited, this technique does represent an important approach to the accuracy issue. For instance, significant differences among groups on objective behavioral criteria (e.g., that men are significantly more aggressive than women and women more empathic than men) may be taken as indications of accuracy if people hold corresponding beliefs about the prevalence of the traits in the groups. Stereotypes are likely to have pernicious consequences regardless of their accuracy as defined in terms of a match between beliefs about the prevalence of a trait and the true base rate in the population. As Stephan (1985) aptly noted, “Stereotypes are often unjustified overgeneralizations, but even when they are not, they have the same implications for information processing and behavior” (p. 600).Stephan’s point is that the conceptualization of stereotype inaccuracy in terms of overgeneralized ascription of trait characteristics to social group members is only one way of thinking about inaccuracy. More recently, stereotype inaccuracy has been considered in a different way, and it is this approach that now reflects the major concern with accuracy within the field. Even if one’s beliefs about the base rate of a trait within a given category are entirely correct, there may still be an erroneous tendency to judge each and every group member as possessing the trait. Recent research (e.g., Brewer, 1988; Fiske & Neuberg, 1990) has demonstrated the extent to which responses based upon stereotypical expectations tend to dominate in interpersonal interaction at the expense of more individuating characteristics such as behaviors. Consider an extreme case where the correlation between category membership and an associated trait is objectively found to be .99. It is still potentially inaccurate and unfair to respond to any given category member as if he or she has the relevant trait. It is this latter type of inaccuracy (perhaps better labeled as unfairness) that represents the truly pernicious aspect of stereotyping. Even McCauley and Stitt

406

CHARLES STANGOR AND JAMES E. LANGE

(1978), who presented the strongest argument that stereotypes were not generally faulty, incorrect, or inaccurate, did concede that the use of stereotypes to guide behavior toward others was incorrect. Although it has been cogently argued that the use of expectations is not always “errorful” or inaccurate (e.g., Jussim, 1991; Oakes & Turner, 1990) because they sometimes provide accurate information about individuals, these arguments finesse the critical essence of the social scientists’ concern with social stereotypes. Stereotypes may certainly provide useful information about the likely behavior of individuals who possess the characteristic. However, when individuals attribute negative characteristics to others on the basis of category memberships, there is inevitably the potential for unjustified responses. Furthermore, it is this potential cost of stereotypes that remains the central preoccupation of the social scientist interested in stereotyping. Nevertheless, it seems clear that future research in stereotyping is going to focus more directly upon the functional utility of stereotypes for individuals (Fiske, 1993; Ford & Stangor, 1992; Jussim, 1991; Oakes & nrner, 1990; Snyder & Miene, 1993), and this means that we will need to get serious again about the accuracy issue. If people use stereotypes at least in part to gain information about others, how effective is this process? To what extent is stereotyping accurate versus (as more commonly assumed) biasing? Finding answers to these issues will not be easy given the difficulty of defining appropriate standards for assessing accuracy, but stereotype accuracy is a question that is not going to go away, and one that needs to be more frankly addressed.

Acknowledgments We thank Diane Ruble, Neil Macrae, Tom Ford, and Eliot Smith for comments on an earlier version.

References Allport, G. W. (1954). The nature of prejudice. Reading, MA: Addison-Wesley. Allport, G. W., & Kramer, B. W. (1946). Some roots of prejudice. Journal of Psychology, 22, 9-39. Andersen, S . M., & Cole, S . W. (1990). “Do I know you?” The role of significant others in general social perception. Journal of Personaliry and Social Psychology, 59, 384-399. Andersen, S . M., & Klatzky, R. L. (1987). Traits and social stereotypes: Levels of categorization in person perception. Journal of Personality and Social Psychology, 53, 235-246. Andersen, S. M., Klatzky, R. L., & Murray, J. (1990). Traits and social stereotypes: Efficiency differences in social information processing. Journal of Personaliry and Social Psychology, 59, 192-201. Anderson, J. R. (1983). The architecture of cognition. Cambridge, MA: Harvard University Press.

GROUP REPRESENTATIONS AND STEREOTYPING

407

Anderson, J. R., & Bower, G. H. (1973). Human associative memory. Washington, DC: V . H. Winston and Sons. Asch, S. E. (1946). Forming impressions of personality. Journal of Abnormal and Social Psychology, 41, 258-290. Ashmore, R. D. (1981). Sex stereotypes and implicit personality theory. In D. L. Hamilton (Ed.), Cognitive processes in stereotyping and intergroup behavior. Hillsdale, NJ: Erlbaum. Ashmore, R. D., & Del Boca, F. K. (1981). Conceptual approaches to stereotypes and stereotyping. In D. L. Hamilton (Ed.), Cognitive processes in stereoryping and intergroup behavior (pp. 135). Hillsdale, NJ: Erlbaum. Baldwin, M. W. (1992). Relational schemas and the processing of social information. Psychological Bulletin, 112, 461-484. Bargh, J. A. (1984). Automatic and conscious processing of social information. In R. S. Wyer, & T. K. Srull (Eds.), Handbook of social cognition (Vol. 3). Hillsdale, NJ: Erlbaum. Bargh, J. A. (1990). Auto-motives: Preconscious determinants of social interaction. In E. T. Higgins & R. M. Sorrentino (Eds.), Handbook of motivation and cognition: Foundations of social behavior (Vol. 2). New York: Guilford Press. Bargh, J. A., & Thein, D. (1985). Individual construct accessibility, person memory, and the recalljudgment link: The case of information overload. Journal of Personality and Social Psychology, 49, 1129-1146. Barsalou, L. W. (1987). The instability of graded structure: Implications for the nature of concepts. In U. Neisser (Ed.), Concepts and Conceptual Development. Cambridge, England: Cambridge University Press. Barsalou, L. W. (1990). On the indistinguishability of exemplar memory and abstraction in category representation. In T. K. Srull & R. S. Wyer, Jr. (Eds.), Advances in Social Cognition (Vol. 3). Hillsdale, NJ: Erlbaum. Bassok, M., & Trope, Y. (1983-1984). People’s strategies for testing hypotheses about another’s personality: Confirmatory or diagnostic? Social Cognition, 2, 199-216. Belmore, S. (1 987). Determinants of attention during impression formation. Journal of Experimental Psychology: Learning, Memory, and Cognition. 13, 480-489. Bem, S . L. (1981). Gender schema theory: A cognitive account of sex typing. Psychological Review, 88, 354-364. Berry, D. S., & McArthur, L. 2. (1986). Perceiving character in faces: The impact of age-related craniofacial changes on social perception. Psychological Bulletin, 100, 3- 18. Bodenhausen, G . V. (1990). Stereotypes as judgmental heuristics: Evidence of circadian variations in discrimination. Psychological Science, I, 3 19-322. Bodenhausen, G. V., & Lichtenstein, M. (1987). Social stereotypes and information processing strategies: The impact of task complexity. Journal of Personality and Social Psychology, 52, 871-880. Bower, G. H., & Cohen, P. R. (1982). Emotional influences in memory and thinking: Data and theory. In M. S. Clark & S. T. Fiske (Eds.), Afect and cognition. Hillsdale, NJ: Erlbaum. Brewer, M. B. (1988). A dual process model of impression formation. In T.K. Srull & R. S. Wyer (Eds.), Advances in social cognition (Vol. I). Hillsdale, NJ: Erlbaum. Brewer, M. B., & Campbell, D. T. (1976). Ethnocentrism and intergroup attitudes: East African evidence. New York: Sage. Brewer, M. B., Dull, L., & Lui, L. (1981). Perceptions of the elderly: Stereotypes as prototypes. Journal of Personaliw and Social Psychology, 4 1 , 656-670. Brigham, J. C. (1971a). Ethnic stereotypes. Psychological Bulletin, 76, 15-38. Brigham, J. C. (1971b). Racial stereotypes, attitudes, and evaluations of and behavioral intentions toward Negroes and whites. Sociometry, 34, 360-380. Cantor. N., & Mischel, W. (1979). Prototypes in person perception. Advances in Experimental Social Psychology, 12, 3-52.

408

CHARLES STANGOR AND JAMES E. LANGE

Clark, M. S . , Milberg, S., & Erber, R. (1988). Arousal-state-dependent memory: Evidence and implications for understanding social judgments and social behavior. In K. Fiedler & J. P. Forgas (Eds.), Affect, Cognition and social behavior (pp. 63-83). Toronto: Hogrefe. Collins, A. M., & Loftus, E. F. (1975). A spreading-activation theory of semantic processing. Psychological Review, 82, 407-428. Cook, S . W. (1978). Interpersonal and attitudinal outcomes in cooperating interracial groups. Journal of Research in Developmental Education, 12, 97-1 13. Cooper, J. B. (1969). Emotional response to statements congruent with prejudicial attitudes. Journal of Social Psychology, 79, 189- 193. Cooper, J. B., & Singer, D. N. (1956). The role of emotion in prejudice. Journal of Social Psychology, 44, 241-247. Crocker, J., Hannah, D. B., & Weber, R. (1983). Person memory and causal attributions. Journal of Personality and Social Psychology, 44( I), 55-66. Darley, J. M., & Fazio, R. H. (1980). Expectancy confirmation processes arising in the social interaction sequence. American Psychologist, 35, 867-88 1. Darley, J. M., & Gross, P. H. (1983). A hypothesis-confirming bias in labeling effects. Journal of Personality and Social Psychology, 44, 20-33. Deaux, K., & Major, B. (1987). Putting gender into context: An interactive model of gender-related behavior. Psychological Review, 94, 369-389. Deaux, K., Winton, W., Crowley, M., &Lewis, L. L. (1985). Level of categorization andcontent of gender stereotypes. Social Cognition, 3, 145-167. Devine, P. G.(1989). Stereotypes and prejudice: Their automatic and controlled components. Journal of Personality and Social Psychology, 56, 5-18. Devine, P. G.,& Baker, S. M. (1991). Measurement of racial stereotype subtyping. Personality and Social Psychology Bulletin. 17, 44-50. Devine, P. G.Monteith, M. J. Zuwerink, J. R., & Elliot, A. J. (1991). Prejudice with and without compunction. Journal of Personality and Social Psychology, 60, 8 17-830. Diab, L. N. (1963a). Factors affecting studies of national stereotypes. Journal of Social Psychology, 59, 29-40. Diab, L. N. (1963b). Factors determining group stereotypes. Journal of Social Psychology, 61, 3- 10. Diehl, M., & Jonas, K. (1991). Measures of national stereotypes as predictors of the latencies of inductive versus deductive stereotypic judgments. European Journal of Social Psychology, 21, 317-330.

Doise, W. (1969). Intergroup relations and polarization of individual and collective judgements. Journal of Personality and Social Psychology, 12, 136- 143. Doise, W., Deschamps, J. C., & Meyer, G.(1978). The accentuation of intra-category similarities. In H. Tajfel (Ed.), Diyerentiation between social groups: Studies in the social psychology of intergroup relarions. London: Academic Press. Dovidio, J. F., Evans, N. E., & Tyler, R. B. (1986). Racial stereotypes: The contents of their cognitive representations. Journal of Experimental Social Psychology, 22, 22-37. Dreben, E. K., Fiske, S. T., & Hastie, R. (1979). The independence of item and evaluative information: Impression and recall order effects in behavior-based impression formation. Journal of Personality and Social Psychology, 37, 1758-1768. Duncan, B. L. (1976). Differential social perception and attribution of intergroup violence: Testing the lower limits of stereotyping of blacks. Journal of Personality and Social Psychology, 34. 590-598.

Eagly, A. H., & Mladinic, A. (1989). Gender stereotypes and attitudes toward women and men. Personality and Social Psychology Bulletin, 15, 543-558. Eagly, A. H., & Steffen, V. J. (1984). Gender stereotypes stem from the distribution of women and men into social roles. Journal of Personality and Social Psychology, 46, 735-754.

GROUP REPRESENTATIONS AND STEREOTYPING

409

Eiser, R. I., & Stroebe, W. (1972). Categorization and social judgment. New York: Academic Press. English, H. B., & English A. C. (1958). A comprehensive dictionary of psychological andpsychoanalytic terms: A guide to usage. New York: David McKay. Erber, R. (1991). Affective and semantic priming: Effects of mood on category accessibility and inference. Journal of Experimental Social Psychology, 27, 480-498. Esses, V. M., Haddock, G., & Zanna, M. P. (1992). The role of mood in the expression of intergroup stereotypes. In M. P. Zanna & J. M. Olson (Eds.), The psychology of prejudice: The Ontario symposium (Vol. 7). Hillsdale, NJ: Erlbaum. Esses, V. M., Hadd0ck.G.. &Zanna, M. P. (1993).Values, stereotypesandemotionsasdeterminantsof intergroupattitudes. In D. M. Mackie& D. L. Hamilton(Eds.),Affect,cognition, andstereotyping: Interactive processes in group perception. (pp. 137- 166) San Diego: Academic Press. Farr, R. M., & Moscovici, S . (1984). Social Representations. Cambridge, England: Cambridge University Press. Fazio, R. H., Powell, M. C., & Herr, P. M. (1983). Toward a process model of the attitude-behavior relation: Accessing one’s attitude upon mere observation of the attitude object. Journal of Personality and Social Psychology, 4 4 , 723-735. Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention and behavior: An introduction to theory and research. Reading, MA: Addison-Wesley. Fiske, S. T. (1980). Attention and weight in person perception: The impact of negative and extreme behavior. Journal of Personality and Social Psychology, 38, 889-906. Fiske, S. T. (1982). Schema-triggered affect: Applications to social perception. In M. S. Clark & S. T. Fiske (Eds.), Affect and cognition. Hillsdale, NJ: Erlbaum. Fiske. S. T. (1993). Social cognition and social perception. Annual Review of Psychology, 4 4 , 155194. Fiske, S. T., & Cox, M. G. (1979). Person concepts: The effect of target familiarity and descriptive purpose on the process of describing others. Journal of Personality, 47, 136-161. Fiske, S . T., & Neuberg, S. L. (1990). A continuum of impression formation, from category based to individuating processes: Influences of information and motivation on attention and interpretation. Advances in Experimental Social Psychology, 23, 1-74. Fiske, S . T., Neuberg, S. L. Beattie, A. E., & Milberg, S . J. (1987). Category-based and attributebased reactions to others: Some informational conditions of stereotyping and individuating processes. Journal of Experimental Social Psychology, 23, 399-427. Fiske, S. T., & Pavelchak, M. A. (1986). Category-based versus piecemeal-based affective responses: Developments in schema-triggered affect. In R. M. Sorrentino & E. T. Higgins (Eds.), The handbook of motivation and cognition: Foundations social behavior. New York: Guilford Press. Fiske, S. T., & Taylor, S. E. (1984). Social cognirion. Reading, MA: Addison-Wesley. Fiske, S. T., & Taylor, S. E. (1991). Social cognition (2nd ed.). New York: McGraw-Hill. Fiske, S. T., & Von Hendy, H. M. (1992). Personality feedback and situational norms can control stereotyping processes. Journal of Personality and Social Psychology, 62, 577-596. Ford, T. E. (1992). The effect of morivation and attribute diagnosticity on stereotype formation. Unpublished doctoral dissertation, University of Maryland, College Park, MD. Ford, T. E., & Stangor, C. (1992). The role of diagnosticity in stereotype formation: Perceiving group means and variances. Journal of Personality and Social Psychology, 63, 356-367. Forgas, J. P. (1992). Affect in social judgments and decisions: A multiprocess model. Advances in Experimental Social Psychology, 25, 227-275. Forgas, J. P., & Bower, G. H. (1987). Mood effects on person-perception judgments. Journal of Personality and Social Psychology, 53, 53-60. Fyock, J., & Stangor, C. (1993). The role of memory biases in stereotype maintenance. Manuscript under review.

410

CHARLES STANGOR AND JAMES E. LANCE

Gaertner, S. L., & Dovidio, J. F. (1986). The aversive form of racism. In S. L. Gaertner & J. F. Dovidio (Eds.). Prejudice, discrimination, and racism (pp. 1-34). Orlando, FL: Academic Press. Gaertner, S. L., & McLaughlin, J. P. (1983). Changing not fading: Racial stereotypes revealed by a non-reactive, reaction time measure. Social Psychology Quarterly, 46, 23-30. Gardner, R. C. (1973). Ethnic stereotypes: The traditional approach, a new look. Canadian Psychologist, 14, 133-148. Gardner, R. C. (1993). Stereotypes as consensual beliefs. In M. Zanna & J. M. Olson (Eds.). The psychology of prejudice: The Ontario Symposium (Vol. 7). Hillsdale, NJ: Erlbaum. Gardner, R. C., Kirby, D. M., & Findlay, J. C. (1973). Ethnic stereotypes: The significance of consensus. Canadian Journal of Behavioural Science, 5. 4- 12. Gardner, R. C., Kirby, D. M. Gorospe, F. H., & Villamin, A. C. (1972). Ethnic stereotypes: An alternative assessment technique, the stereotype differential. Journal of Social Psychology, 87, 259-267.

Gardner, R. C., Wonnacott, E. J., & Taylor, D. M. (1968). Ethnic stereotypes: A factor analytic investigation. Canadian Journal of Psychology, 22, 35-44. Gilbert, D. T., & Hixon, J. G. (1991). The trouble of thinking: Activation and application of stereotypic beliefs. Journal of Personality and Social Psychology, 60, 509-5 17. Gilbert, D. T., & Krull, D. S. (1988). Seeing less and knowing more: The benefits of perceptual ignorance. Journal of Personality and Social Psychology, 54, 193-202. Gilovich, T. (1981). Seeing the past in the present: The effect of associations to familiar events on judgments and decisions. Journal of Personality and Social Psychology, 40, 797-808. Hamilton, D. (1979). A cognitive-attributional analysis of stereotyping. Advances in Experimental Social Psychology, 12, 53-81. Hamilton, D. L., & Gifford, R. K. (1976). Illusory correlation in interpersonal perception: A cognitive basis of stereotypic judgments. Journal of Experimental Social Psychology, 12, 392407.

Hamilton, D. L., & Trolier, T. (1986). Stereotypes and stereotyping: An overview of the cognitive approach. In J. F. Dovidio & S. L. Gaertner (Eds.), Prejudice. discrimination and racism (pp. 127-164). Orlando: Academic Press. Harding, J., Proshansky, H., Kutner, B., & Chein, I. (1969). Prejudice and ethnic relations. In G. Lindzey & E. Aronson, (Eds.), Handbook of social psychology, (2nd ed., Vol. 5). Reading, MA: Addison-Wesley. Hastie, R. (1981). Schematic principles in human memory. In E. T. Higgins, C. P. Herman & M. P. Zanna. Social cognition: The Ontario Symposium. Hillsdale, NJ: Erlbaum. Hastie, R. (1983). Social inference. Annual Review of Psychology, 34, 51 1-542. Hastie, R., & Kumar, P. (1979). Person memory: Personality traits as organizing principles in memory for behaviors. Journal of Personality and Social Psychology, 37, 25-38. Hastie, R., & Park , B. (1986). The relationship between memory and judgment depends on whether the judgment task is memory-based or on-line. Psychological Review, 93, 258-268. Heider, F. (1958). The psychology of interpersonal relations. New York: Wiley. Hewstone, M., & Brown, R. J. (1986). Contact and conj?ict in intergroup encounters. Oxford, England: Blackwell. Higgins, E. T., & King, G. (1981). Accessibility of social constructs: Information-processing consequences of individual and contextual variability. In N. Cantor & J. F. Kihlstrom (Eds.), Personality, cognition. and social interaction. Hillsdale, NJ: Erlbaum. Higgins, E. T., King, G . , & Mavin, G. (1982). Individual construct accessibility and subjective impressions and recall. Journal of Personality and Social Psychology, 43, 35-47. Higgins, E. T., Van Hook, E., & Dorfman, D. (1988). Do self-descriptive traits form a selfstructure? Social Cognition, 6, 177-206.

GROUP REPRESENTATIONS AND STEREOTYPING

41 1

Hilton, J. L., & Fein, S. (1989). The role of typical diagnosticity in stereotype-based judgments. Journal of Personality and Social Psychology, 57. 201-21 1 . Hoffman, C., & Hurst, N. (1990). Gender stereotypes: Perception or rationalization? Journal of Personality and Social Psychology, 58, 197-208. Hogg, M. A., & Abrams, D. (1988). Social identifications. New York: Routledge. Johnston, L., & Hewstone, M. (1992). Cognitive models of stereotype change: Subtyping and the perceived typicality of disconfirming group members. Journal of Experimental Social Psychology, 28, 360-386. Jonas, K., & Hewstone, M. (1987). The assessment of national stereotypes: A methodological study. Journal of Social Psychology, 126, 745-754. Jones, E. E., & Davis, K. E. (1965). From acts to dispositions: The attribution process in person perception. Advances in Experimental Social Psychology, 2 , 219-266. Jones, E. E., & Thibaut, J. W. (1958). Interaction goals as bases of inference in interpersonal perception. In R. Tagiuri & L. Petrullo (Eds.), Person perception and inrerpersonal behavior. Stanford, CA: Stanford University Press. Judd, C. M., & Krosnick, J. A. (1982). Attitude centrality, organization, and measurement. Journal of Personality and Social Psychology, 42. 436-447. Judd, C. M . , Ryan, C. S., & Park, B. (1991). Accuracy in the judgment of in-group and out-group variability. Journal of Personality and Social Psychology, 61 366-379. Jussim, L. (1991). Social perception and social reality: A reflection-construction model. Psychological-Review,98, 54-73. Kahneman, D., & Miller, D. T. (1986). Norm theory: Comparing reality to its alternatives. Psychological Review, 93, 136- 153. Kahneman, D., Slovic, P., & 'hersky, A. (1982). Judgmenr under uncertainty: Heuristics and biases. Cambridge, England: Cambridge University Press. Kanungo, R., & Das, J. P. (1960). Differential learning and forgetting as a function of the social frame of reference. Journal of Abnormal and Social Psychology, 61. 82-86. Katz, D., & Braly, K. W. (1933). Racial stereotypes of one hundred college students. Journal of Abnormal and Social Psychology, 28, 280-290. Kelley, H. H. (1967). Attribution theory in social psychology. In D. Levine (Ed.), Nebraska Symposium on Motivation, (Vol. 15), Lincoln: University of Nebraska Press. Kintsch, W. (1974). The representation of meaning in memory Hillsdale, NJ: Erlbaum. Klein, S. B., & Loftus, I. (1990). The role of abstract and exemplar-based knowledge in selfjudgments: Implications for a cognitive model of the self. In T. K. Srull & R. S. Wyer (Eds.), Advances in social cognition (Vol. 3). Hillsdale, NJ: Erlbaum. Krueger, J., & Rothbart, M. (1988). Use of categorical and individuating information in making inferences about personality. Journal of Personality and Social Psychology, 55. 187195. Kruglanski, A. W. (1989). Cognitive-motivational bases of human knowledge: A theory of lay episternics. New York: Plenum. Kruglanski, A. W., & Freund, T. (1983). The freezing and unfreezing of lay inferences: Effects on irnpressional primacy, ethnic stereotyping, and numerical anchoring. Journal of Experimental Social Psychology, 19, 448-468. Landy, D., & Sigall, H. (1974). Beauty is talent: Task evaluation as a function of the performer's physical attractiveness. Journal of Personality and Social Psychology, 29, 299-304. Lewicki, P. (1985). Nonconscious biasing effects of single instances on subsequent judgments. Journal of Personality and Social Psychology, 48, 563-574. Leyens, J. P., Yzerbyt, V. Y., & Schadron, G . (1992). The social judgeability approach to stereotypes. European Review of Social Psychology, 3, 91-120. Lingle, J. H., Alton, M. W., & Medin, D. L. (1984). Of cabbages and kings: Assessing the I

412

CHARLES STANGOR AND JAMES E. LANGE

extendibility of natural object concept models to social things. In R. S. Wyer & T. K. Srull (Eds.), Handbook of social cognition (Vol. I , pp. 71-118). Hillsdale, NJ: Erlbaum. Linville, P. W., Salovey, P., & Fischer, G. W. (1986). Stereotyping and perceived distributions of social characteristics: And application to ingroup-outgroup perception. In J. F. Dovidio and S. L. Gaertner (Eds.), Prejudice, discrimination and racism. Orlando, FL: Academic Press. Lippmann, W. (1922). Public opinion. New York: Harcourt & Brace. Locksley, A,, Borgida, E., Brekke, N., 8z Hepbum, C. (1980). Sex stereotypes and judgments of individuals. Journal of Personality and Social Psychology, 39(5), 821-83 1. Locksley, A., Hepburn, C., & Ortiz, V. (1982). Social stereotypes and judgments of individuals: An instance of the base-rate fallacy. Journal of Experimental Social Psychology, 18, 23-42. Lombardi, W. J., Higgins, E. T., & Bargh, J. A. (1987). The role of consciousness in priming effects on categorization. Personalie and Social Psychology Bulletin, 13, 41 1-429. Mackie, D. M., & Hamilton, D. L. (1993). Affect. cognition, and stereotyping: Interactive processes in group perception. San Diego: Academic Press. Mackie, D. M., & Worth, L. (1989). Processing deficits and the mediation of positive affect in persuasion. Journal of Personality and Social Psychology, 57, 21-40. Macrae, C. N., Hewstone, M., & Griffiths, R. J. (1993). Processing load and memory for stereotypebased information. European Journal of Social Psychology, 13, 77-87. Macrae, C. N., Stangor, C., Griffiths, R. J., & Milne, A. B. (1993). Stereotypes and mental life: Every cloud has a silver lining. Manuscript in preparation. Mann, J. W. (1967). Inconsistent thinking about group and individual. Journal of Social Psychology, 71, 235-245. Markus, H. (1977). Self-schemata and processing information about the self. Journal of Personality and Social Psychology, 35, 63-78. Markus, H., Crane, M., Bemstein, S., & Sidali, M. (1982). Self-schemas and gender. Journal of Personality and Social Psychology, 42, 38-50. Markus, H., & Zajonc, R. B. (1985). The cognitive perspective in social psychology. In G. Lindzey and E. Aronson (Eds.), Handbook of socialpsychology (3rd ed.). Reading, MA: Addison-Wesley. Martin, C. L. (1987). A ratio measure of sex stereotyping. Journal of Personality and Social Psychology, 5 2 , 489-499. Martin, L. L. Seta, J. J., & Crelia, R. A. (1990). Assimilation and contrast as a function of people's willingness and ability to expend effort in forming an impression. Journal of Personality and Social Psychology, 59, 27-31. McArthur, L. 2 . (1982). Judging a book by its cover: A cognitive analysis of the relationship between physical appearance and stereotyping. In A. Hastorf & A. Isen (Eds.), Cognitive socialpsychology. New York: Elsevier North-Holland. McArthur, L. Z., & Baron, R. M. (1983). Toward an ecological theory of social cognition. Psychological Review, 90, 215-238. McCauley, C . , & Stitt, C. L. (1978). An individual and quantitative measure of stereotypes. Journal of Personality and Social Psychology, 36, 929-940. McCauley, C., Stitt, C. L., & Segal, M. (1980). Stereotyping: From prejudice to prediction. Psychological Bulletin, 87. 195-208. McGuire, W. J., & McGuire, C. V. (1988). Content and process in the experience of self. Advances in Experimental Social Psychology, 2 1 , 97-144. Messick, D. M., & Mackie, D. M. (1989). Intergroup relations. Annual Review of Psychology, 40. 45-82. Meyer, D. E., & Schvaneveldt, R.,W. (1971). Facilitation in recognition of pairs of words: Evidence of a dependence between retrieval operations. Journal of Experimental Psychology, 90,227-234. Miller, C. T. (1988). Categorization and the physical attractiveness stereotype. Social Cognirion. 6, 23 1-25 1.

GROUP REPRESENTATIONS AND STEREOTYPING

413

Miller, D. T., & 'hrnbull, W. (1986). Expectancies and interpersonal processes. Annual Review of Psychology, 37, 233-256. Miller, N., & Brewer, M. B. (1986). Categorization effects on ingroup and outgroup perception. In S. L. Gaertner & J. F. Dovidio (Eds.), Prejudice, discrimination, and racism (pp. 209-230). Orlando, FL: Academic Press. Murphy, G. L., & Medin, D. L. (1985). The role of theories in conceptual coherence. Psychological Review, 92, 289-316. Neely, J. H. (1977). Semantic priming and retrieval from lexical memory: Roles of inhibitionless spreading activation and limited-capacity attention. Journal of Experimental Psychology, 106, 226-254. Neuberg, S. L., & Fiske, S. T. (1987). Motivational influences on impression formation: Outcome dependency, accuracy-driven attention, and individuating processes. Journal of Personality and Social Psychology. 53, 43 1-444. Nisbett, R. E., & Ross, L. D. (1980). Human inference: Strategies and shortcomings of social judgment. Englewood Cliffs, NJ: Prentice-Hall. Nosofsky, R. M. (1986). Attention, similarity, and the identification-categorization relationship. Journal of Experimental Psychology: General, 115, 39-57. Oakes, P. J., & 'hrner, J. C. (1990). Is limited information processing capacity the cause of social stereotyping? European Review of Social Psychology I , I 11-135. Ostrom, T. M., Carpenter, S. L., Sedikides, C., & Li, F. (1993). Differential processing of ingroup and out-group information. Journal of Personality and Social Psychology, 64,2 1-34. Park, B . (1986). A method for studying the development of impressions of real people. Journal of Personality and Social Psychology, 51, 907-911. Park, B., & Hastie, R. (1987). Perception of variability in category development: Instance versus abstraction-based stereotypes. Journal of Personality and Social Psychology, 53, 621 -635. Park, B., Judd, C. M., & Ryan, C. S. (1991). Social categorization and the representation of variability information. European Review of Social Psychology, 2 , 22 1-245. Park, B., Ryan, C. S., & Judd, C. M. (1992). Role of meaningful subgroups in explaining differences in perceived variability for in-groups and out-groups. Journal of Personality and Social PSycholOgy, 63, 553-567. Peabody, D. (1 968). Group judgments in the Philippines: Evaluative and descriptive aspects. Journal of Personality and Social Psychology, 10. 290-300. Perdue, C. W., Dovidio, J. F., Gurtman, M. B., & 'Qler, R. B. (1990). Us and them: Social categorization and the process of intergroup bias. Journal of Personality and Social Psychology, 59, 415-486. Perdue, C. W., & Gurtman, M. B. (1990). Evidence for the automaticity of ageism. Journal of Experimental Social Psychology, 26, 199-2 16. Pratto, F., & Bargh, J. A. (1991). Stereotyping based on apparently individuating information: Trait and global components of sex stereotypes under attention overload. Journal of Experimental Social Psychology, 27, 26-47. Pryor, J. B . , & Ostrom, T. M. (1981). The cognitive organization of social information: A converging-operations approach. Journal of Personality and Social Psychology, 4 1 , 628-64 I . Rholes, W. S., Newman, L. S., & Ruble, D. N. (1990). Understanding self and other tal and motivational aspects of perceiving persons in terms of invariant disposi Higgins, & R. M. Sonentino (Eds.), Handbook of motivation and cognition: Foundations of social behavior (Vol. 2). New York: Guilford. Rogers, T. B., Kuiper, N. A., & Kirker, W. S. (1977). Self-reference and the encoding of personal information. Journal of Personality and Social Psychology, 35, 677-688. Rojahn, K., & Pettigrew, T. F. (1992). Memory for schema-relevant information: A meta-analytic resolution. British Journal of Social Psychology, 31, 8 I - 109.

414

CHARLES STANGOR AND JAMES E. LANCE

Rosch, E. (1978). Principles of categorization. In E. Rosch & B. B. Lloyd (Eds.), Cognition and categorization (pp. 27-48). Hillsdale, NJ: Erlbaum. Rosenkrantz, P., Vogel, Bee, H., Broverman, I. K., & Broverman, D. M. (1968). Sex-role stereotypes and self concepts in college students. Journal of Consulting and Clinical Psychology, 32, 287-295. Rothbart, M., Evans, M., & Fulero, S. (1979). Recall for confirming events: Memory processes and the maintenance of social stereotypes. Journal of Experimental Social Psychology, 15, 343-355. Rothbart, M., Fulero, S . , Jensen, C., Howard, J., & Birrell, P. (1978). From individual to group impressions: Availability heuristics in stereotype formation. Journal of Experimental Social Psychology, 14. 237-255. Rothbart, M., & John, 0. P. (1985). Social categorization and behavioral episodes: A cognitive analysis of the effects of intergroup contact. Journal of Social Issues, 4 1 , 81-104. Rothbart, M., & John, 0. (1992). Intergroup relations and stereotype change: A social-cognitive analysis and some longitudinal findings. In P. M. Sniderman & P. E. Tetlock (Eds.), Prejudice, politics and race in America. Stanford, CA: Stanford University Press. Ruscher, J. B., & Fiske, S. T. (1990). Interpersonal competition can cause individuating processes. Journal of Personality and Social Psychology, 58, 832-843. Sagar, H. A.. & Schofield, J. W. (1980). Racial and behavioral cues in black and white children’s perceptions of ambiguously aggressive acts. Journal of Personality and Social Psychology, 39, 590-598. Schaller, M., & Maass, A. (1989). Illuso~ycorrelation and social categorization: Toward an integration of motivational and cognitive factors in stereotype formation. Journal of Personality and Social Psychology, 56, 709-721. Schank, R. C., & Abelson, R. P. (1977). Scripts. Plans, Goals and Understanding. Hillsdale, NJ: Erlbaum. Schneider, D. A. (1973). Implicit personality theory: A review. Psychological Bulletin, 79, 294-309. Secord, P. F. (1959). Stereotyping and favorableness in the perception of Negro faces. Journal of Abnormal and Social Psychology, 59, 309-315. Secord, P. F., & Berscheid, E. S. (1963). Stereotyping and the generality of implicit personality theory. Journal of Personality, 31 309-315. Secord, P. F., & Berscheid, E. S. (1963). Stereotyping and the generality of implicit personality theory. Journal of Personality, 31, 65-78. Sherman, S. J., Judd, C. M.. & Park, B. (1989). Social Cognition. Annual Review ofPsychology, 40, 281-326. Sigall, H., & Page, R. (1971). Current stereotypes: A little fading, a little faking. Journal of Personality and Social Psychology, 18, 247-255. Sigall, H., Stangor, C., & Ford, T. E. (1990, June). Measurement of social stereotypes: Group versus individual approaches. Budapest: European Association of Experimental Social Psychology. Smith, E. E., & Medin, D. L. (1981). Categories and concepts. Cambridge, MA: Harvard University Press. Smith, E. E., & Osherson, D. N. (1984). Conceptual combination with prototype concepts. Cognitive Science, 8, 337-361. Smith, E. R. (1989). Procedural efficiency: General and specific components and effects on social judgment. Journal of Experimental Social Psychology, 25, 500-523. Smith, E. R. (1990). Content and process specificity in the effects of prior experiences. In T. K. Srull & R. S. Wyer (Eds.), Advances in social cognition (Vol. 3, pp. 1-59), Hillsdale, NJ: Erlbaum. Smith, E. R., & Branscombe, N. R. (1987). Procedurally mediated social inferences: The case of category accessibility effects. Journal of Experimental Social Psychology, 23. 361-382. Smith, E. R., & Ziirate, M. A. (1990). Exemplar and prototype use in social categorization. Social Cognition, 8, 243-262. ~

GROUP REPRESENTATIONS AND STEREOTYPING

415

Smith, E. R., & Zirate, M. A. (1992). Exemplar-based model of social judgment. Psychological Review, 99, 3-21. Snyder, M., & Cantor, N. (1979). Testing hypotheses about other people: The use of historical knowledge. Journal of Experimental Social Psychology, 15, 330-342. Snyder, M., & Miene, P. (1993). On the functions of stereotypes and prejudice. In M. Zanna & J. M. Olson (Eds.), The psychology of prejudice: The Ontario Symposium (Vol. 7). Hillsdale, NJ: Erlbaum. Snyder, M., Tanke, E. D., & Berscheid, E. (1977). Social perception and interpersonal behavior: On the self-fulfilling nature of social stereotypes. Journal of Personality and Social Psychology, 35, 656-666. Snyder, M., & Uranowitz, S. W. (1978). Reconstructing the past: Some cognitive consequences of person perception. Journal of Personality and Social Psychology, 36, 941-950. Srull, T. K., & Wyer, R. S . (1989). Person memory and judgment. Psychological Review, 96, 58-83. Srull, T. K., & Wyer, R. S. (1990). Advances in social cognition (Vol. 3). Hillsdale, NJ: Erlbaum. Stangor, C. (1988). Stereotype accessibility and information processing. Personality and Social Psychology Bulletin, 14, 694-708. Stangor, C. (1990). Arousal, accessibility of trait constructs, and person perception. Journal of Experimental Social Psychology, 26, 305-32 1. Stangor, C., & Duan, C. (1991). Effects of multiple task demands upon memory for information about social groups. Journal of Experimental Social Psychology, 27, 357-378. Stangor, C., Ford, T. E. (1992). Accuracy and Expectancy-Confirming Processing Orientations and the Development of Stereotypes and Prejudice. European Review of Social Psychology, 3 , 5789. Stangor, C., Jonas, K., Stroebe. W., & Hewstone, M. (1991). Development andchange of national stereotypes and attitudes. Manuscript in preparation. Stangor, C., Lynch, L., Duan, C., & Glass, B. (1992). Categorization of individuals on the basis of multiple social features. Journal of Personality and Social Psychology, 62, 207-281. Stangor, C., McMillan, D. ( 1992). Memory for Expectancy-Congruent and Expectancy-Incongruent Information: A review of the social and social-developmental literatures. Psychological Bulletin, 111. 42-61. Stangor, C . , & Ruble, D. N. (1989a). Differential influences of gender schemata and gender constancy on children’s information processing and behavior. Social Cognition, 7 , 353-372. Stangor, C., & Ruble, D. N. (1989b). Strength of expectancies and memory for social information: What we remember depends on how much we know. Journal of Experimental Social Psychology, 25, 18-35. Stangor, C., Sullivan, L. A , , & Ford, T.E. (1992). Affective and cognitive determinants of prejudice. Social Cognition, 9, 359-380. Stephan, W. G. (1985). Intergroup relations. In G. Lindzey & E. Aronson (Eds.), Handbook of social psychology. New York: Random House. Stephan, W. G., & Stephan, C. W. (1993). Cognition and affect in stereotyping: Parallel interactive networks (pp. I 1 1-136). In D. M. Mackie & D. L. Hamilton (Eds.), Affect, cognition. and stereotyping: Interactive processes in group perception, San Diego: Academic Press. Swim, J., Borgida, E., Maruyama, G., & Myers, D. (1989). Joan McKay versus John McKay: Do gender stereotypes bias evaluations? Psychological Bulletin, 105, 409-429. Tajfil, H. (1981). Human groups and social categories. Cambridge, England: Cambridge University Press. Taylor, S . E., Fiske, S. T., Etcoff, N. L., & Ruderman, A. J. (1978). Categorical and contextual bases of person memory and stereotyping. Journal of Personality and Social Psychology, 36, 778-793.

416

CHARLES STANGOR AND JAMES E. LANCE

Tetlock, P. E. (1985). Accountability:The neglected social context of judgment and choice. Research in Organizational Behavior, 7 , 297-332. Trafimow, D., Triandis, H.C., & Goto, S. G. (1991). Some tests of the distinction between the private self and the collective self. Journal of Personality and Social Psychology, 60, 649-655. Trope, Y.(1986). Identification and inferential processes in dispositional attribution. Psychological Review, 93, 239-257. 'Tbmer, J. C. (1987). Rediscovering the social group: A self-categorization theory. London: Basil Blackwell. Ugwuegbu, D. C. E. (1979). Racial and evidential factors in juror attribution of legal responsibility. Journal of Experimental Social Psychology, 15, 133-146. Vinacke, W. E. (1957). Stereotypes as social concepts. Journal of Social Psychology, 46, 229-243. Warren, R. E. (1972). Stimulus encoding and memory. Journal of Experimental Psychology. 94,90100. Weber, R., & Crocker, J. (1983). Cognitive processes in the revision of stereotypicbeliefs. Journal of Personality and Social Psychology, 45, 961-977. Widiger, T. A., & Settle, S. A. (1987). Broverman revisited: An artifactual sex bias. Journal of Personality and Social Psychology, 53, 463-469. Wilder, D. A. (1986). Social categorization:Implications for creation and reduction of intergroup bias. Advances in Experimental Social Psychology, 19, 291-355. Word, C. 0.. Zanna, M. P., & Cooper,J. (1974). The nonverbal mediation of self-fulfilling prophecies in interracial interaction. Journal of Experimental Social Psychology, 10, 109- 120. m a t e , M. A., & Smith, E. R. (1990). Person categorization and stereotyping. Social Cognition, 8, 16I- 185.

INDEX

A Ability, meaning of, transition in, 187-189 age differences, 190, 202 competence, 187, 189-192 constancy, 191-192, 203 educational factors, 191- 192, 198- 199 individual differences, 190, 202 Abstractionist approach, to stereotyping, 392393 consensual beliefs, 402-404 exemplar models, 396-400 feelings, 400 motivation, 400-402 overgeneralization, 404-406 specific content, 393-395 Academic performance, fall in, reactions to, 24-28 misdemeanor as cause, 28-31 Accessibility information, conversational, 149- 154 stereotypes, 378-380 assessment, 380-381 category-based information, 384-387 category representations, 381-384 processing effects, 387-388 Achievers, attitudes toward, 1-3, 22, 65-69 attribution theory in analysis of, 3, 7-9 correlational studies, 22-23, 46-56 cultural variables, 36-39 deservingness analysis, 4, 13-18 research, 31-36 experimental research, 22-23 Heider’s theory of reactions to another’s lot, 3-7 417

public figures, 56-62, 65 fallen, 62-65 reactions to another’s lot, 3-7 to fallen public figures, 62-65 self-esteem in relation to, 7 academic performance, fall in, 23 authoritarianism, 54 comparison between Australia and Japan, 54-56 cultural context, 45 Rosenberg Self-Esteem Scale, 46, 48-49 Tall Poppy Scale, 43-44, 47-49, 52 social comparison in analysis of envy, 4, 18-22 theory, 4, 7 status, as variable, 5-6, 23-24, 39-40 academic performance, 24-3 I cultural differences, 36-39 deservingness, 3 1-36 Tall Poppy Scale, 1, 40-42 authoritarianism, 44-45, 53-54 correlational studies, 46-56 cultural variables, 45 self-esteem, 42-44 values, 42 values Protestant ethic, 49-53 role of, 42, 46-49 theory, 10- 13 Actuators, in impression regulation, 251 -252 Advertisements, mate preferences in, 88-93 Affect in response to inequality, 339-343 in stereotyping, 401 Age, preferences, in sexual selection, 84-94

418

INDEX

Ambiguity, conversational, 134-138 Appraisal, self-, see Self-appraisal Appraisals, legitimacy, see Legitimacy appraisals Appraisal sets, in impression regulation, 247 Associative model, of group representation, 361-363, 366 Associative strength, determinants, 371-373 Attention, gender differences in, 110-1 11 Attitudes toward achievers, see Achievers, attitudes toward attribution, in relational regulation, 233-234 Attractiveness, as criteria, in sexual selection, 108-1 10 Attribution error, conversational, 132-133 Attribution theory, in study of attitudes toward achievers, 3, 7-9 Audiences, multiple, see Relational regulation Authoritarianism, in attitudes toward achievers, 44-45 Right-Wing Authoritarianism Scale, 53-54 Automatic behavior, in relational regulation, 272-273

B Back-channeling, in communication to multiple audiences, 263-264 Behavior evolutionary characteristics, 77-78, 113 universality, 78-79 in relational regulation, 237-238 automatic, 272-273 deliberate, 272-273 impression regulation, 246-254 nonverbal, 238-239 obedience, 273-275 private life, 277-278 public life, 275-277 self-concept, 240-245 verbal relational transformations, 239-240 responses to entitlement, 343-344 self-discrepant, see Self-discrepant behavior stereotyping of, 384-387 Beliefs, stereotypes as, 402-404 Biases correspondence in impression regulation, 252-254 in observation of role-based constraint, 229-23 I

by expectancy-based processing, 359 judgmental, and logic of conversation, 123124, 134, 144, 155 in perception of disadvantage, 294 Biology, evolutionary, see Evolutionary biology

C Carry-over effects, in impression regulation, 246-247 Category-based information, accessibility, 384387 Category differentiation, in stereotyping, 370372 Category label, in stereotyping, 373 Category representation in stereotyping, 381-384 in transition, 172 Children, conversational cognition in, 146- 148 Cognition conversational in children, 146-148 context, 123 research, 134, 154, 156 evolutionary, 75, 101-105, I14 attractiveness, as criteria, 108-1 10 dominance, 108-1 10 gender differences, 110- I13 paradigms, in study of gender differences, 110-1 13 parental investment, 105 schemata, default, 105-108 gender differences in, I 10- I13 role of feelings in, 401 in transition, 172, 179 ability, meaning of, 187, 191-192 gender, 186, 201 motivation, 177 perception of personal characteristics, 193-195, 197 schematic, 174-175 Cognitive development, in transition, 170- 171 gender, 181 knowledge base, 172 level of uncertainty, 179 memory, 180 perception of personal characteristics, 193 Collective entitlement, 330-333 Common ground, in communication to multiple audiences, 259-260

419

INDEX Communal common ground, in communication to multiple audiences, 259-260 Communal orientation, in attitudes toward achievers, 49- 5 3 Communication conversational relevance of, 128- 144 cooperative, conversational, 137-138, 144 covert, to multiple audiences, 255-256, 264-268 informativeness, speaker’s role, 144-154 off-line, to multiple audiences, 263-264 tactical, in multiple audience situations, 22 1-222 role conflict, 228-237 Comparison, social in entitlement perception, 334, 344 gender, 326-328 reference standards, 300-303 and envy, 4 , 18-22 closeness variable, 2 I justice component, 21-22 self-evaluation, 19-20 in transition, 201-202 Competence, perceived in school performance, 48-49 transition school change, 198-199 school entry, 187 self-assessment, 189- 192 Concepts, dispositional, see Dispositional concepts Conflict, role, see Role conflict Consolidation phase, in transition, 167- 169 postdecisional orientation, 203 Constancy ability, 192 gender, 181-185, 201 Construction phase, in transition, 166-167 information seeking, 202-203 Context conversational, 154, 156 ambiguity, 135 informative contributions, 148- 154 of judgment, 123-124 motivational, in transition, 177 Conversation, logic of, 123-124 context, 155-156 everyday life, 124- 126 information quality, 125 quantity, I25

informative contributions, 144- 145 accessibility, 149- 153 redundancy avoidance, 149- 154 relevance, judgment of, 143- 144 repeated questions, 146- 149 response formats, 145-146 judgment, 123-124, 150-151, 154 accessibility of information, 145 biases, 134, 144, 155 comparative, 143-144 evaluative, 153 relevance, 125, 128-134 ambiguity, 134- 138 questionnaires, 138-144 research settings, 126-127 response formats, 144- 146 Cooperative communication, conversational, 137-138, 144 Correspondence bias in impression regulation, 252-254 in observation of role-based constraint, 22923 I Covert communication, to multiple audiences, 255-256 individual differences, 264-268 Cues distancing, in impression regulation, 246250 embracing, in impression regulation, 247250 relational, in attitude attribution, 233 Cultural differences, reaction to fall of achievers, 36-39, 45 Culture, and age preferences, in mate selection, 86-88 advertisements, 91-93

D Deliberate behavior, in relational regulation, 272-273 Deprivation group, 330-331, 336 personal, 330-331, 336 relative, 339 Deservingness, see also Entitlement in study of attitudes toward achievers analysis, 4, 13-18 research, 31-36 Development, cognitive, see Cognitive development

420

INDEX

Differentiation, category, in stereotyping, 370372 Disadvantage, see Inequality Discounting, in impression regulation, 250 Discrimination, see also Inequality against women, and structural integrity, 3 14-3 15 Disequilibrium, in transition, 171 Dispositional concepts, stable, perception of, 194-197, 203 Distancing, in relational regulation, 216, 218219 behaviors, 237-240 cues, 23 1, 233-234 function, 232 impression management, 246-254 multiple audiences, 222-223 obedience study, 274-275 roles, 235-236 self-concept regulation, 240-245 Distinctiveness, stereotype, 369-370 Distributive justice, norms, 304-305 Dominance as attractiveness criteria, 95 in evolutionary cognition, 108-1 10

E Eavesdroppers, in communication to multiple audiences, 258-259 Embracing, in relational regulation, 216 behavior, 237-240 cues, 233-234 impression management, 246-254 self-concept regulation, 240-245 Emotions in response to inequality, 339, 343-344 in stereotyping, 401-402 Encryption systems, in communication to multiple audiences, 260-263 Engagement, level of, as function of goals, 175-179 Enhancement, self, see Self-enhancement Entitlement, 293-294, 344-346 collective, 330-333 consequences, 338-339 affect, 339-343 behavioral responses, 343-344 gender differences inequality elimination, 326-327 legitimacy appraisals, 323-326

outcome evaluations, 327-328 pay, 313-323 inequality, impact, 307 comparisons, 307-308, 312-313 group membership, 307-312 personal, 330, 344 procedural injustice, perception of, 334 psychology of, 298-299, 346 comparison standards, 299-303 legitimacy appraisals, 303-307 Envy, and attitudes toward high achievers, 4, 6-7, 18-22 closeness variable, 21 justice component, 21-22 self-evaluation, 19-20 Epigamic selection, 82-84 Equalitarianism, in correlational studies of attitudes toward achievers, 49-53 Equity, in sexual selection process, 98, 101 Equivocation, in multiple-audience situations, 256 Esteem, self-, see Self-esteem Evolutionary biology, 85-86 sexual selection age preferences, 86-94 parental investment, 94-101 Evolutionary cognition, 75, 101-105, 114 attractiveness, as criteria, 108-1 10 dominance, 108-1 10 gender differences, 110- I13 paradigms, in study of gender differences, 110-113 parental investment, 105 schemata, default, 105-108 Evolutionary psychology, 75-77, 113 behavior, 77-78 universal, 78-79 cognition in, 75 gender differences, 81 mating strategies, 80-83 natural selection, 77-78 parental investment, differential, 81-84 selection epigamic, 82-84 intrasexual, 82 sexual, 82-85 strategies frequency-dependent, 79 mating, 80-83 species-wide, 80-83 Exemplar models, of group representation,

INDEX compared to abstractionist approach, 396400 Expectations in schematic processing, 175 stereotypical, 357-358 biased information processing, 359 memory, 363 in transition, 180 motherhood, 198

F Feedback effects on perceived entitlement, 327-328, 336 model of self-concept regulation, 244-245 negative, effects on perceived entitlement, 340, 342 Formats, conversational response, 144- 146

G Gender constancy, 181-185, 201 transition, 18I- 188, 201-204 Gender differences attention, 110-1 11 cognition, 110-1 12 division of labor, 81 entitlement, 294, 313-314, 323, 346 inequality, elimination, 326-327 legitimacy appraisal, 323-326 outcome evaluations, 327-328 pay, 314-323 segregation of work, 314 habituation, I 1 1 priming, 112-113 reactions to inequality, 295-297 sexual selection, 113 Generalization, in stereotyping definitions, 367-369 inaccuracy, 404-406 overgeneralization, 404-406 Goals information as function of, 178-179 knowledge as function of, 175-178 Group membership, role in inequality, 293, 329, 344-345 collective entitlement, 330-333 comparisons, 307-309 ideology, 335-338

42 1

legitimizing processes, 309-3 12 procedural injustice, 334-335 Group representation, stereotypes as accessiblity, 380 associative model, 361-363, 366 memory, 363-364 priming, 364-366

H Habituation, gender differences, 1 1 I Heider’s theory, of reactions to another’s lot, 3-7 High achievers, attitudes toward, see Achievers, attitudes toward

I Illusory-correlation paradigm, in group representation, 372 Image management, in political campaigns, 276 Impression management, in relational regulation, 220-221 actuators, 251-252 behaviors, 272 correspondence bias, 252-254 distancing cues, 246-250 embracing cues, 247-250 suspicion, 250-251 Inequality, social, 293-294, 344-346 gender differences, in reaction to entitlement, 295-297, 313-314, 323 elimination of inequality, 326-327 legitimacy, 323-326 outcome evaluations, 327-328 pay, 314-323 impact on entitlement, 307 comparisons, 3 12-31 3 group membership, 307-312 reactions to, effects of race, 297 undeserved, 329-330 collective entitlement, 330-333 ideology, 335-338 procedural injustice, 333-335 Information base-rate, underutilization, 128- 132 category-based, accessibility, 384-387 communicated, and guarantee of relevance, 128-144 Injustice, procedural, perception, 333-335

422

INDEX

Integration phase, in transition, 169, 204 and gender information, 186 Internalization, self-discrepant behavior, 241244 lntrasexual selection, 82 Investment, parental, see Parental investment

J Judgment and logic of conversation, 123- 127 biases, 134, 144, 155 comparative, 143- 144 context, 154 evaluative, 153 general, 150-151 guarantee of relevance, 128-144 information accessibility, 145 in transition, confidence in, 176-177 Justice analysis of, in study of attitudes toward achievers, 4 in legitimacy appraisals of entitlement distributive, 304-305 procedural, 305

K Keys, private, in shared knowledge, 259 Knowledge, see also Shared knowledge as function of goals, 175-179 hidden, in relational regulation, 225-228 social, in transition, 196-197

L Labels, category, in stereotyping, 373 Legitimacy appraisals, of entitlement attributions of responsibility, 305-306 awareness, 329 biases, 294 comparisons, 312-313 future research, 345 gender differences, 314, 323-326 goals, 306-307 group membership, 307-3 12 ideology, 335 norms distributive justice, 304-305 procedural justice, 305 preconditions, 303-304

procedural injustice, 333-334 Logic, of conversation, see Conversation, logic of

M Mate selection, see Sexual selection Memory gender, 186-187 group representations, 363-364 knowledge-based, in transition, 172 schema-congruent, in transition, 174, 204 stereotypical, 373, 380 in transition, 180 Mental representations, stereotypes as, 359361 associative model, 361-363, 366 memory, 363-364 priming, 364-366 Message modification, to multiple audiences, 256-258 Messages, in relational regulation hidden, 224-225, 255-256 mixed, 236, 255 Milgram obedience studies, relational regulation analysis, 273-275 Monitoring, self-, see Self-monitoring Motherhood, transition to, 164-166, 217 Motivation control, 177 in stereotyping, 400-402 in transition, 172, 175, 179 ability, 198 gender, 186, 201 perception of personal characteristics, 193-197 Multiple audiences, see Relational regulation, multiple audiences

N Natural selection, applied to behavior theory, 77-78 Nonredundancy, norm of, in conversation, 151, 153 Nonverbal behavior, in relational regulation, 238-239 Norms conversational, 134, 136, 151, 155 cultural, in sexual selection, 86-87, 91-93

INDEX distributive justice, 304-305 nonredundancy, in conversation, 151, 153 procedural justice, 305

0 Obedience studies, distancing behavior in, 273-275 Off-line communication, to multiple audiences, 263-264 Orientation, communal, in attitudes toward achievers, 49-53 Overgeneralization, in stereotype formation, 404-406

423

Public life, relational regulation in courtroom, 276 media, 275-276 political campaigns, 276 recommendation letters, 277

Q Questionnaires, formal features, 138, 144 frequency scales, 141-144 rating scales, 138-141 Questions, ambiguous, making sense of, 134I38

R

P Paradigms, in study of gender differences, 110-113 Parental investment default schemata for, 105 differential, 81-83 model, 83-84 in sexual selection relationship levels, 94-98 self-appraisals, 98- 101 Parenthood, transition to, see Motherhood, transition to Performance, academic, see Academic performance Personal common ground, in communication to multiple audiences, 260 Priming gender differences, 1 12- 1 13 group representation measures, 364-366 Private keys, in shared knowledge, 259 Private life, covert signaling in, 277-278 Procedural injustice, perception of, 333-335 Procedural justice, norms, 305 Processing, schematic, in transition, 173-175, 204 Processing effects active, 387-388 passive, 387-388 Protestant ethic values, in attitudes toward achievers, 49-53 Psychology, evolutionary, see Evolutionary psychology Public figures attitudes toward, 56-62, 65 fallen, attitudes toward, 62-65

Race, in reactions to inequality, 297, 329 collective entitlement, 331-333 entitlement consequences, 340 ideology, 336-338 Reactions to achievers, 3-7 to fallen public figures, 62-65 to inequality, 295 gender, 295-297 race, 297 Redundancy, avoidance in conversation, 149154 Relational regulation, 215-218, 270-272 audience segregation, 235 behavior, 237-238 automatic, 272-273 deliberate, 272-273 impression regulation, 246-254 nonverbal, 238-239 obedience, 273-275 private life, 277-278 public life, 275-277 self-concept regulation, 240-245 verbal relational transformations, 239-240 covert communication, 255-256, 265-268 distancing, 218-219, 222-223 equivocation, 256 impression management, 220-221, 272 multiple audiences, communication to, 223224, 255-256 back-channeling, 263-264 common ground, 259-260 eavesdroppers, 258-259 encryption systems, 260-263 hidden messages, 224-225

424

INDEX

message modification, 256-258 self-monitoring, 265-268 shared knowledge, 225-228, 259, 264265 suspicion, 268-270 personal relationships, 277-278 political campaigns, 276 recommendation letters, 277 role conflict, 228-237 tactical communication, 221-222 Relevance, conversational, 128, 134 ambiguity, 135- 138 attribution error, 132-133 base-rate information, 128- 132 implicit guarantee of, 144 judgment, 144-145 judgmental biases, 134 questionnaires, 138- 144 Representations, see Category representation; Group representation; Mental representations Response formats, conversational, 144- 146 Role conflict, in tactical relational regulation attitude attribution, 233-234 correspondence bias, 229-231 cues, 233 distancing behaviors, 231-234, 236 embracing behaviors, 233-234 role distancing, 235-237 role enactments, 228-229 Role distancing, in multiple-audience situations, 235-237 Role enactments, in tactical relational regulation, 228-229 Rosenberg Self-Esteem Scale, 46, 49, 56-57 S Salience, of group representation, 382 Scales, of conversational relevance frequency, 141-144 rating, 138-141 Schemata default, for parental investment, 105-108 defined, in evolutionary cognition, 105 in transition, individual differences in, 169 Schematic processing, in transition, 173-175, 204 Schwartz Value Survey, 46, 49, 51 Selection epigamic, 82-84

intrasexual , 82 natural, applied to behavior theory, 77-78 sexual, see Sexual selection Self-appraisal, in sexual selection, 99- 101 Self-concept, regulation by distancing, 232, 240-24 I feedback model, 244-245 internalization of behavior, 241-244 Self-discrepant behavior, internalization, 241 244 Self-enhancement, in stereotyping, 372-373 Self-esteem in attitudes toward achievers, 7, 27 Australia, 56-58 authoritarianism, 54 cultural context, 45 fall in academic performance, 23 Japan, 56-58 Rosenberg Self-Esteem Scale, 46, 48-49 Tall Poppy Scale, 43-44, 47-49, 52 in response to perceived entitlement, 340, 342 in transition ability, 189, 198 gender, 201 Self-monitoring audience segregation by, 235 in covert communication, 265-268 Sets, appraisal, in impression regulation, 247 Sex, see Gender Sexual selection, 75-76 age preferences, 84-94 attractiveness criteria, 108- 110 default schemata study, 105-108 evolutionary studies advertisements for, 88-93 age preferences, 86-94 biology and social psychology, integrated studies, 85-101 culture, 86-88, 91-93 norms, 86-87, 91-93 teenagers, preferences, 93-94 gender differences, 81, 110-1 13 and parental investment differential, 81-83 model, 83-85 relationship levels, 94-98 self-appraisals, 98- 101 within-species, 80-8 1 Shared knowledge, in covert communication, 259, 264-265

INDEX Shortcoming, in judgment, and logic of conversation, 123 Social cognition, see Cognition Social comparison, see Comparison, social Social inequality, see Inequality, social Spin control, in political campaigns, 276 Status preferences for, in sexual selection, 101 varied, and attitudes toward achievers, 5-6, 23-40 academic performance, 24-31 cultural differences, 36-39 deservingness, 3 1-36 Stereotypes, 357-359 abstractionist approach, 392-393 consensual beliefs, 402-404 exemplar models, 396-400 feelings, 400 motivation, 400-402 overgeneralization, 404-406 representational content, 393-395 accessibility, 378-380 assessment, 380-381 category-based information, 384-387 category representations, 381-382 processing effects, 387-388 change, 390-392 definitions, 360 definitions, new, 366-367 associative strength, determinants, 37 I373 category differentiation, 370-372 distinctiveness, 369-370 generalization, 367-369 gender, 183, 186-187, 200-201 measurement, 373-374 free-response approach, 375-376 problems, 379 techniques, 374 validity, 376-377 as mental representations, 359-361 associative model, 361-363, 366 memory, 363-364 priming, 364-366 rigidity, 389-390 sex, in entitlement, 321, 323 Stigma, group, based on perceived entitlement, 340-342 Suspicion in covert communication, 268-270 in impression regulation, 250-251

425 T

Tactical communication, in multiple-audience situations, 221-222 role conflict, 229-234 role distancing, 235-237 role enactments, 228-229 Tall poppies, theory and research, 1-73 Teenagers, preferences in sexual selection, 9394 Transformations, verbal relation, in distancing behavior, 239-240 Transitions, phase model, 163- 164, 166- 170, 197-206 ability, meaning of, 187-193, 202-203 competence, 187, 189, 190-192, 203204 application to social-cognitive transitions, 180- 181 category accessibility, 173, 180 cognitive development, 170- 172, 179-180 consolidation, 167-169, 203 construction, 166- 167, 202-203 dispositional concepts, stable, 194-1 97, 203 engagement level, 175- I78 gender, 181-187, 201-204 implications, 197-202 information, as function of goals, 175-179 integration, 169, 204 motherhood, 164-166, 170 motivation, 172-179 perception of personal characteristics, 193197 schematic processing, 173-175, 204 social cognition and motivation, 172- I79

U Uncertainty, level of, in transition, 179- 180

V Values, in attitudes toward achievers, 4, 42, 46-48 Protestant ethic, 49-53 Value theory, in study of attitudes toward achievers, 10-13 Verbal relation transformations, in distancing behavior, 239-240

This Page Intentionally Left Blank

CONTENTS OF OTHER VOLUMES

Volume 1 Cultural Influences upon Cognitive Processes Harry C. Triandis The Interaction of Cognitive and Physiological Determinants of Emotional State Stanley Schachter Experimental Studies of Coalition Formation William A . Gamson Communication Networks Marvin E. Shaw A Contingency Model of Leadership Effectiveness Fred E. Fiedler Inducing Resistance to Persuasion: Some Contemporary Approaches William J. McGuire Social Motivation, Dependency, and Susceptibility to Social Influence Richard H. Walters and Ross D . Purke Sociability and Social Organization in Monkeys and Apes William A. Mason Author Index-Subject Index

Volume 2 Vicarious Processes: A Case of No-Trial Learning Albert Bandura Selective Exposure Jonathan L. Freedman and David 0. Sears Group Problem Solving L. Richard Hofman Situational Factors in Conformity Vernon L. Allen 421

Social Power John Schopler From Acts to Dispositions: The Attribution Process in Person Perception Edward E. Jones and Keith E. Davis Inequality in Social Exchange J. Stacy Adams The Concept of Aggressive Drive: Some Additional Considerations Leonard Berkowirz Author Index-Subject Index

Volume 3 Mathematical Models in Social Psychology Robert P. Abelson The Experimental Analysis of Social Performance Michael Argyle and Adam Kendon A Structural Balance Approach to the Analysis of Communication Effects N. T. Feather Effects of Fear Arousal on Attitude Change: Recent Developments in Theory and Experimental Research Irving L. Janis Communication Processes and the Properties of Language Serge Moscovici The Congruity Principle Revisited: Studies in the Reduction, Induction and Generalization of Persuasion Percy H. Tannenbaum Author Index-Subject Index

428

CONTENTS OF OTHER VOLUMES

Volume 4 The Theory of Cognitive Dissonance: A Current Perspective Elliot Aronson Attitudes and Attraction Donn Byrne Sociolinguistics Susan M. Ervin-Tripp Recognition of Emotion Nico H. Frijda Studies of Status Congruence Edward E. Sampson Exploratory Investigations of Empathy Ezra Stotland The Personal Reference Scale: An Approach to Social Judgment Harry S. Upshaw Author Index-Subject Index

Volume 5 Media Violence and Aggressive Behavior: A Review of Experimental Research Richard E. Goranson Studies in Leader Legitimacy, Influence, and Innovation Edwin P. Hollander and James W. Julian Experimental Studies of Negro-White Relationships Irwin Katz Findings and Theory in the Study of Fear Communications Howard Leventhal Perceived Freedom Ivan D.Steiner Experimental Studies o f Families Nancy E. Waxler and Elliot G. Mishler Why Do Groups Make Riskier Decisions than Individuals? Kenneth L. Dion, Robert S. Baron, and Norman Miller Author Index-Subject Index

Volume 6 Self-Perception Theory Daryl J. Bem Social Norms, Feelings, and Other Factors Affecting Helping and Altruism Leonard Berkowitz

The Power of Liking: Consequences of Interpersonal Attitudes Derived from a Liberalized View of Secondary Reinforcement Albert J. Lott and Bernice E. Lott Social Influence, Conformity Bias, and the Study of Active Minorities Serge Moscovici and Claude Faucheux A Critical Analysis of Research Utilizing the Prisoner’s Dilemma Paradigm for the Study of Bargaining Charlan Nemeth Structural Representations of Implicit Personality Theory Seymour Rosenberg and Andrea Sedlak Author Index-Subject Index

Volume 7 Cognitive Algebra: Integration Theory Applied to Social Attribution Norman A. Anderson On Conflicts and Bargaining E r i k Apfelbaum Physical Attractiveness Ellen Bersheid and Elaine Wulster Compliance, Justification, and Cognitive Change Harold B. Gerard, Edward S. Connolley, and Roland A. Wilhelmy Processes in Delay of Gratification Walter Mischel Helping a Distressed Person: Social, Personality, and Stimulus Determinants Ervin Sraub Author Index-Subject Index

Volume 8 Social Support for Nonconformity Vernon L. Allen Group Tasks, Group Interaction Process, and Group Performance Effectiveness: A Review and Proposed Integration J. Richard Hackman and Charles G. Morris The Human Subject in the Psychology Experiment: Fact and Artifact Arie W. Kruglanski Emotional Arousal in the Facilitation of Aggression through Communication Percy H. Tannenbaum and Dolf Zillman The Reluctance to Transmit Bad News Abraham Tesser and Sidney Rosen

CONTENTS OF OTHER VOLUMES Objective Self-Awareness Robert A . Wicklund Responses to Uncontrollable Outcomes: An Integration of Reactance Theory and the Learned Helplessness Model Camille B. Wortman and Jack W. Brehm Subject Index

Volume 9 New Directions in Equity Research Elaine Walster, Ellen Berscheid, and G. William Walster Equity Theory Revisited: Comments and Annotated Bibliography 1. Stacy Adams and Sara Freedman The Distribution of Rewards and Resources in Groups and Organizations Gerald S. Leventhal Deserving and the Emergence of Forms of Justice Melvin J. Lerner, Dale T. Miller, and John G. Holmes Equity and the Law: The Effect of a Harmdoer’s “Suffering in the Act” on Liking and Assigned Punishment William Austin, Elaine Walster, and Mary Kristine Utne Incremental Exchange Theory: A Formal Model for Progression in Dyadic Social Interaction L. Lowell Huesmann and George Levinger Commentary George C. Homans Subject Index

Volume 10 The Catharsis of Aggression: An Evaluation of a Hypothesis Russell G. Geen and Michael B. Quanty Mere Exposure Albert A. Harrison Moral Internalization: Current Theory and Research Martin L. Hoffman Some Effects of Violent and Nonviolent Movies on the Behavior of Juvenile Delinquents Ross D. Parke. Leonard Berkowitz, Jacques P. Leyens, Stephen G . West, and Richard Sebastian

429

The Intuitive Psychologist and His Shortcomings: Distortions in the Attribution Process Lee Ross Normative Influences on Altruism Shalom H. Schwartz A Discussion of the Domain and Methods of Social Psychology: Two Papers by Ron Harre and Barry R. Schlenker Leonard Berkowitz The Ethogenic Approach: Theory and Practice R. Harre On the Ethogenic Approach: Etiquette and Revolution Barry R. Schlenker Automatisms and Autonomies: In Reply to Professor Schlenker R. Harre Subject Index

Volume 11 The Persistence of Experimentally Induced Attitude Change Thomas D. Cook and Brian F. Flay The Contingency Model and the Dynamics of the Leadership Process Fred E. Fiedler An Attributional Theory of Choice Andy Kukla Group-Induced Polarization of Attitudes and Behavior Helmut Lamm and David G . Myers Crowding: Determinants and Effects Janet E. Stockdale Salience: Attention, and Attribution: Top of the Head Phenomena Shelley E. Taylor and Susan T. Fiske Self-Generated Attitude Change Abraham Tesser Subject Index

Volume 12 Part I. Studies in Social Cognition Prototypes in Person Perception Nancy Cantor and Walter Mischel A Cognitive-Attributional Analysis of Stereotyping David L. Hamilton Self-Monitoring Processes Mark Snyder

430

CONTENTS OF OTHER VOLUMES

Part 11. Social Influences and Social Interaction Architectural Mediation of Residential Density and Control: Crowding and the Regulation of Social Contact Andrew Baum and Stuart Valins A Cultural Ecology of Social Behavior J. W. Berry Experiments on Deviance with Special Reference to Dishonesty David P. Farrington From the Early Window to the Late Night Show: International Trends in the Study of Television’s Impact on Children and Adults John P. Murray and Susan Kippax Effects of Prosocial Television and Film Material on the Behavior of Viewers J . Phillipe Rushton Subject Index

Volume 13

Dialectic Conceptions in Social Psychology: An Application to Social Penetration and Privacy Regulation Irwin Altman, Anne Vinsel, and Barbara B . Brown Direct Experience and Attitude-Behavior Consistency Russell H. Fazio and Mark P. Zanna Predictability and Human Stress: Toward a Clarification of Evidence and Theory Suzanne M. Miller Perceptual and Judgmental Processes in Social Contexts Arnold Upmeyer Jury Trials: Psychology and Law Charlan Jeanne Nemeth Index

Volume 15

People’s Analyses of the Causes of AbilityLinked Performances John M. Darley and George R. Goethals The Empirical Exploration of Intrinsic Motivational Processes Edward I . Deci and Richard M. Ryan Attribution of Responsibility: From Man the Scientist to Man as Lawyer Frank D. Fincham and Joseph M. Jaspars Toward a Comprehensive Theory of Emotion Howard Leventhal Toward a Theory of Conversion Behavior Serge Moscovici The Role of Information Retrieval and Conditional Inference Processes in Belief Formation and Change Robert S. Wyer, Jr. and Jon Hartwick Index

Balance, Agreement, and Positivity in the Cognition of Small Social Structures Walter H. Crockett Episode Cognition: Internal Representations of Interaction Routines Joseph P. Forgas The Effects of Aggressive-Pornographic Mass Media Stimuli Neil M. Malamuth and Ed Donnerstein Socialization in Small Groups: Temporal Changes in Individual-Group Relations Richard L. Moreland and John M. Levine Translating Actions into Attitudes: An IdentityAnalytic Approach to the Explanation of Social Conduct Barry R. Schlenker Aversive Conditions as Stimuli to Aggression Leonard Berkowitz Index

Volume 14

Volume 16

Verbal and Nonverbal Communication of Deception Miron Zuckerman, Bella M. DePaulo, and Roberr Rosenthal Cognitive, Social, and Personality Processes in the Physiological Detection of Deception William M . Waid and Martin T. Orne

A Contextualist Theory of Knowledge: Its Implications for Innovation and Reform in Rychological Research William J. McGuire Social Cognition: Some Historical and Theoretical Perspectives Janet Landman and Melvin Manis

CONTENTS OF OTHER VOLUMES Paradigmatic Behaviorism: Unified Theory for Social-Personality Psychology Arthur W. Staats Social Psychology from the Standpoint of a Structural Symbolic Interactionism: Toward an Interdisciplinary Social Psychology Sheldon Stryker Toward an Interdisciplinary Social Psychology Carl W. Backman Index

Volume 17 Mental Representations of Self John F. Kihlstrom and Nancy Cantor Theory of the Self Impasse and Evolution Kenneth J. Gergen A Perceptual-Motor Theory of Emotion Howard Leventhal Equity and Social Change in Human Relationships Charles G. McClintock, Roderick M. Kramer, and Linda J . Keil A New Look at Dissonance Theory Joel Cooper and Russell H. Fazio Cognitive Theories of Persuasion Alice H. Eagly and Shelly Chaiken Helping Behavior and Altruism: An Empirical and Conceptual Overview John F. Dovidio Index

Volume 18 A Qpological Approach to Marital Interaction: Recent Theory and Research Mary Anne Fitzpatrick Groups in Exotic Environments Albert A. Harrison and Mary M. Connors Balance Theory, the Jordan Paradigm, and the Wiest Tetrahedon Chester A . Insko The Social Relations Model David A . Kenny and Lawrence La Voie Coalition Bargaining S. S. Komorita When Belief Creates Reality Mark Snyder Index

43 1

Volume 19 Distraction-Conflict Theory: Progress and Problems Robert S. Baron Recent Research on Selective Exposure to Information Dieter Frey The Role of Threat to Self-Esteem and Perceived Control in Recipient Reaction to Help: Theory Development and Empirical Validation Arie Nadler and Jeffrey D. Fisher The Elaboration Likelihood Model of Persuasion Richard E. Petty and John T.Cacioppo Natural Experiments on the Effects of Mass Media Violence on Fatal Aggression: Strengths and Weaknesses of a New Approach David P. Phillips Paradigms and Groups Ivan D. Steiner Social Categorization: Implications for Creation and Reduction of Intergroup Bias David A . Wilder Index

Volume 20 Attitudes, Traits, and Actions: Dispositional Prediction of Behavior in Personality and Social Psychology Icek Ajzen Prosocial Motivation: Is It Ever Tmly Altruistic? C. Daniel Batson Dimensions of Group Process: Amount and Structure of Vocal Interaction James M . Dabbs. Jr. and R. Barry Ruback The Dynamics of Opinion Formation Harold B. Gerard and Ruben Orive Positive Affect, Cognitive Processes, and Social Behavior Alice M. Isen Between Hope and Fear: The Psychology of Risk Lola L. Lopes Toward an Integration of Cognitive and Motivational Perspectives on Social Inference: A Biased Hypothesis-Testing Model Tom Pyszczynski and Jeff Greenberg Index

432

CONTENTS OF OTHER VOLUMES

Volume 21 Introduction Leonard Berkowitz Part I. The Self as Known Narrative and the Self as Relationship Kenneth J. Gergen and Mary M. Gergen Self and Others: Studies in Social Personality and Autobiography Seymour Rosenberg Content and Process in the Experience of Self William J. McGuire and Claire V. McGuire Information Processing and the Study of the Self John F. Kihlstrom, Nancy Cantor, Jeanne Sumi Albright, Beverly R. Chew, Stanley B. Klein, and Paula M. Niedenthal Part 11. Self-Motives Toward a Self-Evaluation Maintenance Model of Social Behavior Abraham Tesser The Self: A Dialectical Approach Carl W. Backman The Psychology of Self-Afirmation: Sustaining the Integrity of the Self Claude M. Steele A Model of Behavioral Self-Regulation: Translating Intention into Action Michael F. Scheier and Charles S. Carver Index

Volume 22 On the Construction of the Anger Experience: Aversive Events and Negative Priming in the Formation of Feelings Leonard Berkowitz and Karen Heimer Social Rychophysiology: A New Look John T. Cacioppo, Richard E. Petty?and Louis G. Tassinary Self-Discrepancy Theory: What Patterns of SelfBeliefs Cause People to Suffer? E. Tory Higgins Minding Matters: The Consequences of Mindlessness-Mindfulness Ellen J. Langer

The Tradeoffs of Social Control and Innovation in Groups and Organizations Charlan Jeanne Nemeth and Barry M. Staw Confession, Inhibition, and Disease James W. Pennebaker A Sociocognitive Model of Attitude Structure and Function Anthony R. Pratkanis and Anthony G. Greenwald Introspection, Attitude Change, and AttitudeBehavior Consistency: The Disruptive Effects of Explaining Why We Feel the Way We Do Timothy D. Wilson, Dana S. Dunn. Dolores Kraft, and Douglas J . Lisle Index

Volume 23 A Continuum of Impression Formation, from Category-Based to Individuating Processes: Influences of Information and Motivation on Attention and lnterpretation Susan T.Fiske and Steven L. Neuberg Multiple Processes by Which Attitudes Guide Behavior: The MODE Model as as Integrative Framework Russell H. Fazio PEAT: An Integrative Model of Attribution Processes John W. Medcof Reading People's Minds: A Transformation Rule Model for Predicting Others' Thoughts and Feelings Rachel Karniol Self-Attention and Behavior: A Review and Theoretical Update Frederick X . Gibbons Counterfactual Thinking and Social Perception: Thinking about What Might Have Been Dale T. Miller, William Turnbull, and Cathy McFarland Index

Volume 24 The Role of Self-Interest in Social and Political Attitudes David 0. Sears and Carolyn L. Funk A Terror Management Theory of Social Behav-

CONTENTS OF OTHER VOLUMES

433

ior: The Psychological Functions of Self- Volume 25 Esteem and Cultural Worldviews Sheldon Solomon, Jeff Greenberg, and Tom Universals in the Content and Structure of Values: Theoretical Advances and Empirical Tests in Pyszczynski Mood and Persuasion: Affective States Influence 20 Countries the Processing of Persuasive Communications Shalom H. Schwartz Norbert Srhwarz, Herbert Bless, and Gerd Motivational Foundations of Behavioral Confirmation Bohner A Focus Theory of Nonnative Conduct: A TheoMark Snyder retical Refinement and Reevaluation of the A Relational Model of Authority in Groups Tom R. Tyler and E. Allan Lind Role of Norms in Human Behavior Robert B. Cialdini, CarlA. Kallgren, andRay- You Can’t Always Think What You Want: Problems in the Suppression of Unwanted Thoughts mond R. Reno The Effects of Interaction Goals on Person PerDaniel M. Wegner Affect in Social Judgments and Decisions: A ception Multiprocess Model James L. Hilton and John M . Darley Studying Social Interaction with the Rochester Joseph Paul Forgas The Social Psychology of Stanley Milgram Interaction Record Thomas Blass Harry T. Reis and Ladd Wheeler Subjective Construal, Social Inference, and Hu- The Impact of Accountability on Judgment and Choice: Toward a Social Contingency Model man Misunderstanding Philip E. Terlock Dale W. Griffin and Lee Ross Index Index

This Page Intentionally Left Blank