Stress and Emotion, Volume 17: Anxiety, Anger, and Curiosity (Stress and Emotion)

RT44367 half title page 8/12/05 3:25 PM Page 1

STRESS AND

EMOTION

The Series in Stress and Emotion: Anxiety, Anger, and Curiosity (Formerly part of the Series in Clinical and Community Psychology) Consulting Editors Charles D. Spielberger and Irwin G. Sarason Spielberger, Sarason, and Defares Stress and Anxiety, Volume 9 Spielberger and Sarason Stress and Anxiety, Volume 10: A Sourcebook of Theory and Research Spielberger, Sarason, and Defares Stress and Anxiety, Volume 11 Spielberger, Sarason, and Strelau Stress and Anxiety, Volume 12 Spielberger, Sarason, Strelau, and Brebner Stress and Anxiety, Volume 13 Spielberger, Sarason, Kulcsár, and Van Heck Stress and Anxiety, Volume 14 Spielberger, Sarason, Brebner, Greenglass, Laungani, and O’Roark Stress and Anxiety, Volume 15 Spielberger, Sarason, Brebner, Greenglass, Laungani, and O’Roark Stress and Anxiety, Volume 16 Spielberger and Sarason Stress and Anxiety, Volume 17

RT44367_C000.indd 2

8/31/05 9:38:02 AM

RT44367 title page 8/12/05 3:22 PM Page 1

STRESS AND

EMOTION ANXIETY, ANGER, and CURIOSITY

VOLUME 17 Edited by

Charles D. Spielberger and Irwin G. Sarason

New York London

RT44367_Discl.fm Page 1 Wednesday, August 31, 2005 9:34 AM

Published in 2005 by Routledge Taylor & Francis Group 270 Madison Avenue New York, NY 10016

Published in Great Britain by Routledge Taylor & Francis Group 2 Park Square Milton Park, Abingdon Oxon OX14 4RN

© 2005 by Taylor & Francis Group, LLC Routledge is an imprint of Taylor & Francis Group Printed in the United States of America on acid-free paper 10 9 8 7 6 5 4 3 2 1 International Standard Book Number-10: 0-415-94437-6 (Hardcover) International Standard Book Number-13: 978-0-415-94437-3 (Hardcover) Library of Congress Card Number BF575.S75A38 1973 616.8'522 4-28292 MARC No part of this book may be reprinted, reproduced, transmitted, or utilized in any form by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying, microfilming, and recording, or in any information storage or retrieval system, without written permission from the publishers. Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe.

Library of Congress Cataloging-in-Publication Data Advanced Study Institute on Stress and Anxiety in Modern Life, Murnau, Ger. 1973 Stress and anxiety: [proceedings]/edited by Charles D Spieberger, Irwin G Sarason. Washington: Hemisphere Publ. Corp. v. :ill.: 24 cm. (v. 1-2: The series in clinical psychology v. 3-5: The series in clinical and community psychology) Includes bibliographical references and indexes. 1. Stress, (Psychology)--Congresses. 2. Anxiety--Congresses. I. Sarason, Irwin G., ed. II Spielberger, Charles Donald, date, ed. III. North Atlantic Treaty Organization, Division of ScientificAffairs. IV. Title. [DNLM: 1. Anxiety. 2. Stress, Psychological. WM 172 S755a] BF575.S75A38 1973 616.8'522 4-28292 MARC

Visit the Taylor & Francis Web site at http://www.taylorandfrancis.com Taylor & Francis Group is the Academic Division of T&F Informa plc.

and the Routledge Web site at http://www.routledge-ny.com

Contents

Contributors

xii

Preface

xiii I STRESS, ANXIETY, AND COPING

1

2

Stress in a Life-Threatening Illness Pittu Laungani

3

Stress in a Life-Threatening Illness Participant Observation Background to the Study Transfer to St. Mary’s Hospital Arriving at a Correct Diagnosis Transfer to Intensive Care Fear of Pain and Death Transfer from the ITU The Nature of Stress in a Life-Threatening Illness Conceptualizing Stress 1. Extrapersonal Factors 2. Interpersonal Relationships with Hospital Staff 3. Intrapersonal Factors Conclusion

3 6 8 10 12 13 16 17 19 19 20 22 25 26

Anxiety, Anger, and Moral Judgment Angela M.B. Biaggio

31

Anxiety, Anger, and Moral Judgment Theory and Research on Anxiety and Moral Development Relation of Moral Judgment to State and Trait Anxiety Measuring the Experience, Expression, and Control of Anger Construction and Validation of the Portuguese STAXI Anxiety, Anger, and Risk-Taking Behavior Appendix A

31 33 35 36 37 40 44



RT44367_C000.indd 5

8/31/05 9:38:03 AM

vi

3

CONTENTS

Trait-State Anxiety, Worry, and Emotionality in Athletic Competition Frontisek Man, F. Táborský, and Charles D. Spielberger

45

Trait-State Anxiety, Worry, and Emotionality in Athletic Competition 45 Study 1 47 48 Method and Procedure 49 Results and Discussion Study 2 51 52 Method and Procedure 52 Results and Discussion Summary and Conclusions 55 4

Evaluation of Test Anxiety Using Pictorial Evaluation of Test Reactions (PETER) Joseph Toubiana Evaluation of Test Anxiety Using Pictorial Evaluation of Test Reactions (PETER) Method Subjects and Instruments Previous Research with the PETER-TQ Procedures Results Cultural and Sex Differences in Test Anxiety Discussion Differences in Anxiety across Cultures Gender Differences Conclusions Acknowledgments

5

Culture and Gender Factors in the Structure of the Test Anxiety Inventory: A Meta-Analysis Brad J. Bushman, Peter R. Vagg, and Charles D. Spielberger Culture and Gender Factors in the Structure of the Test Anxiety Inventory: A Meta-Analysis Test Anxiety, Worry-Emotionality, and Academic Achievement The Test Anxiety Inventory Method and Procedures Literature Search Procedures and Criteria for Inclusion of Studies Results Invariance of the Factor Structure of the TAI Across Culture and Gender Discussion and Conclusions Appendix A A Geometrical Description of Kaiser’s Method for Comparing Factor Structures

RT44367_C000.indd 6

61

61 63 63 65 66 67 67 68 69 70 71 75 77

77 78 80 81 82 83 86 87 91 91

8/31/05 9:38:03 AM

CONTENTS

vii

6

97

Vigilant and Avoidant Coping: Theory and Measurement Heinz Walter Krohne and Boris Egloff Vigilant and Avoidant Coping: Theory and Measurement Model of Coping Modes Measurement of Vigilance and Cognitive Avoidance Design of the Mainz Coping Inventory Participants Analysis 1: Appraisal of the MCI Situations Analysis 2: Statistical and Psychometric Properties of the MCI Analysis 3: Internal Structure of the MCI Analysis 4: Relationships with Other Coping Variables (Construct Validity) Analysis 5: Relationships with Various Criterion Variables Discussion

97 98 102 102 104 104 107 107 108 112 113

II EXPERIENCE, EXPRESSION, AND CONTROL OF ANGER 7

8

The Expression and Management of Anger and Hostility: Psychological Implications Esther R. Greenglass

123

The Expression and Management of Anger and Hostility: Psychological Implications Multiple Emotions and CHD Interactive Approaches to the Study of CHD The Social Context of Anger/Hostility and CHD Social Support and Health Anger, Hostility, and Coping Communicative Expression of Anger and Hostility Anger, Hostility, and Gender Differences Conclusions

123 124 125 128 129 132 133 134 136

School Stress, Anger Expression, and Adolescent Problem Behaviors 143 Chok C. Hiew and Glendon Rayworth School Stress, Anger Expression, And Adolescent Problem Behaviors 143 Problem Behavior and Anger Expression 145 Method 146 Predictor and Criterion Measures 147 Results 148 Discussion and Conclusions 150

RT44367_C000.indd 7

8/31/05 9:38:03 AM

viii

9

CONTENTS

Experience of Anger and Anxiety Emotions during Examinations: Cross-Cultural Evidence Norbert K. Tanzer and Charles D. Spielberger Experience of Anger and Anxiety Emotions during Examinations: Cross-Cultural Evidence Method Instruments Trait and State Instructions Rating Scale Formats Statistical Analyses Subjects and Procedures Results Experience of Anger and Anxiety during Examinations: Results from Trait Studies Temporal and Situational Specificity of Anger and Anxiety during Examinations Academic Domain Specificity and Course (Teacher) Specificity of Trait Anger and Trait Anxiety during Examinations Trait Anger and Anxiety during Regular Classroom Lessons Discussion Appendix A Items of the English Version Appendix B Items of the German Version and Closest English Translation

10

11

155

155 156 156 157 158 158 159 161 161 165 167 168 169 177 177 177 177

Family Cohesiveness, Flexibility, and Maternal Anger: Boon or Detriment to Children’s Adjustment? Solly Dreman

179

Family Cohesiveness, Flexibility, and Maternal Anger: Boon or Detriment to Children’s Adjustment? Method Measures Procedure Results Relation between Cohesion, Adaptability, and Behavior Problems Cohesion and Children’s Behavior Problems Adaptability and Children’s Behavior Problems Custodial Mothers’ Anger and Children’s Behavior Problems Discussion and Conclusions

179 180 181 184 184 184 184 186 187 190

Hindi Anger Expression Scale: Construction and Validation Sagar Sharma, Archna Krishna and Charles D. Spielberger

197

The Hindi Anger Expression Scale: Construction and Validation Adaptation and Validation of the Hindi AX Scale

197 198

RT44367_C000.indd 8

8/31/05 9:38:03 AM

CONTENTS

Psychometric Properties and Equivalence of the Hindi and English AX Scales Internal Consistency Reliability of the Hindi AX Scale Research with the Hindi Anger Expression Scale

ix

199 201 202

III CROSS-CULTURAL RESEARCH ON OCCUPATIONAL STRESS 12

The German Adaptation of the Job Stress Survey: A Multistudy Validation in Different Occupational Settings Volker Hodapp, Norbert K. Tanzer, Christian Korunka, Ehrentraud R. Maier and Iris A. Pestemer

209

The German Adaptation of the Job Stress Survey: A Multistudy Validation in Different Occupational Settings 210 The Original English-Language Job Stress Survey 210 Development of the German Adaptation of the Job Stress Survey 211 Study 1: Primary-School Teachers 213 213 Participants and Procedure 213 Results Study 2: Secondary-School Teachers 216 216 Participants and Procedure 217 Results Study 3: Hospital Personnel 219 219 Subjects and Procedure 220 Results Study 4: Bank Employees 222 222 Participants and Procedure 224 Results Discussion and Conclusions 226 Authors’ Notes and Acknowledgments 231 13

Spanish Adaptation of the Job Stress Survey: All Dressed Up and Ready to Go to Work Maria Jesus González and José M. Prieto

233

Spanish Adaptation of the Job Stress Survey: All Dressed Up and 233 Ready to Go to Work Construction and Development of the JSS-SV 234 235 Choice of Complementary Tests 236 Selecting and Obtaining the Sample 236 Application of the Tests 236 Description of the Sample Results and Discussion 237 237 Psychometric Properties of the JSS-SV

RT44367_C000.indd 9

8/31/05 9:38:03 AM



CONTENTS

Validity of the JSS-SV Factor Structure of the Spanish Job Stress Survey Concurrent Validity of the Spanish JSS Comparison of the Factor Structure of the Spanish JSS-SV and the American JSS Conclusions and Suggestions for Further Research 14

Occupational Level and Gender Differences in Job Stress in Venezuelan Employees Lya Feldman, Grace Chacón-Puignau, Nuri Bagés, and María Gabriela Pérez Occupational Level and Gender Differences in Job Stress in Venezuelan Employees Method Participants Instruments Procedure Results Reliability and Factor Analyses of the Spanish JSS Severity and Frequency Scales Differences in Gender and Organizational Level Social Support and Automatic Thinking Discussion Gender and Organizational Level Differences in Occupational Stress Appendix A

15

Job Stress and Gender: A Comparison of Government and Corporate Employees Lisa Fiksenbaum and Esther R. Greenglass Job Stress and Gender: A Comparison of Government and Corporate Employees Stress and Health Occupational Stress: Its Sources and Consequences Gender Differences in Occupational Stress Method and Procedure Participants Measures Procedures Results Group and Gender Differences in Ratings of Specific JSS Stressor Events Discussion Job Stress and Gender Differences

RT44367_C000.indd 10

238 239 242 246 248 253

253 255 255 255 256 256 256 257 260 261 262 266 271

271 272 273 275 276 276 277 278 278 280 281 286

8/31/05 9:38:04 AM

CONTENTS

16

17

xi

Norwegian Adaptation of the Job Stress Survey Kjell Håseth

293

Norwegian Adaptation of the Job Stress Survey Theoretical Views of Occupational Stress The Job Stress Survey The Norwegian Job Stress Survey Developing the Norwegian JSS The Factor Structure and Psychometric Properties of the JSS-N The JSS-N Lack of Organizational Support and Job Pressure Subscales Intercorrelations among the JSS Scales and Subscales for Males and Females Validity of the Norwegian JSS-N Occupational Group Differences in JSS-N Scores Discussion

293 294 295 297 297 298

Calibrating Productivity in the Workplace: Assessing Stress Judiciously Ann M. O’Roark Calibrating Productivity in the Workplace: Assessing Stress Judiciously Calibration Procedures and Selecting a Pivotal Instrument Individual Calibration: A Special Courts Judge Large-scale Calibration: A Family Courts Agency Stressor Differences in the Five Judicial Community Work Groups Recommendations for Reducing Job Stress in the Judicial Community Gender Differences in Occupational Stress in the Judicial Communities Discussion and Conclusions

301 302 303 306 307 313

313 314 319 323 325 327 329 331

Author Index

333

Subject Index

335

RT44367_C000.indd 11

8/31/05 9:38:04 AM

Contributors NURI BAGÉS, Universidad Simón Bolívar, Caracas, Venezuela ANGELA M. B. BIAGGIO, Universidade Rio Grande du Sul, Brazil BRAD J. BUSHMAN, University of Michigan, Ann Arbor, MI GRACE CHACÓN-PUIGNAU, Universidad Simón Bolívar, Caracas, Venezuela SOLLY DREMAN, Ben Gurion University of the Negev, Israel BORIS EGLOFF, Johannes-Gutenberg-Universität, Mainz, Germany LYA FELDMAN, Universidad Simón Bolívar, Caracas, Venezuela LISA FIKSENBAUM. York University, Toronto, Canada MARÍA JESÚS GONZÁLEZ, Complutense University, Madrid, Spain ESTHER R. GREENGLASS, York University, Toronto, Canada KJELL HÅSETH, University of Oslo, Norway CHOK C. HIEW, University of New Brunswick, Fredericton, Canada VOLKER HODAPP, J.-W. Goethe-Universität, Frankfurt, Germany CHRISTIAN KORUNKA, University of Vienna, Austria ARCHNA KRISHNA, Himachal Pradesh University, Shimla, India HEINZ WALTER KROHNE, Johannes-Gutenberg-Universität, Mainz, Germany PITTU LAUNGANI, Honorary Research Fellow, Manchester University, England EHRENTRAUD R. MAIER, Karl-Franzens-Universität, Graz, Austria FRONTISEK MAN, University of South Bohemia, Ceske Budejovice, Czech Republic ANN M. O’ROARK, Consultant, Management and Leadership Development, St. Petersburg, FL MARÍA GABRIELA PÉREZ, Universidad Simón Bolívar, Caracas,Venezuela IRIS A. PESTEMER, Karl-Franzens-Universität, Graz, Austria JOSÉ M. PRIETO, Complutense University, Madrid, Spain GLENDON RAYWORTH, University of New Brunswick, Fredericton, Canada IRWIN G. SARASON, University of Washington, Seattle SAGAR SHARMA, Himachal Pradesh University, Shimla, India CHARLES D. SPIELBERGER, University of South Florida, Tampa, FL F. TÁBORSKÝ, Charles University, Prague, Czech Republic NORBERT K. TANZER, Alliant International University, San Diego, CA JOSEPH TOUBIANA, Zahav Clinic, Petach Tikva, Israel PETER R. VAGG, University of South Florida, Tampa, FL

xii

RT44367_C000.indd 12

8/31/05 9:38:04 AM

Preface

The publication of Volume 17 of Stress and Emotion: Anxiety, Anger and Curiosity, marks the 30th anniversary of this series. When the series began in 1975, the title was Stress and Anxiety. In keeping with a broader interest in the effects of stress on other emotions, the current title was introduced in 1991, along with the publication of Volume 14. The initiation of this series was stimulated by a grant received by the editors in 1972 from the Scientific Affairs Division of the North Atlantic Treaty Organization to organize an Advanced Study Institute on Stress and Emotion in Modern Life. This Institute was convened in Murnau-am-Staffelsee, West Germany in June 1973, with the goal of facilitating an exchange of information and stimulation of collaborative research efforts among behavioral and medical scientists who were involved in research on stress and anxiety. The first volume in this series was based entirely on the contributions of the invited participants at the Murnau Conference. Volumes 2 through 7 of the series were also based primarily on the contributions of participants in NATO-sponsored Advanced Study Institutes organized by the editors that were held in Norway, Italy, and at Cambridge University in the UK. The contents of Volumes 8 through 13 were based primarily on presentations at international conferences, which were held in Israel, The Netherlands, Poland, and Australia, supported by universities and the governments of these countries. The name of the series was changed with the publication of Volume 14, which was based on presentations at our first conference on Stress and Emotion, held in Visegrad, Hungary and sponsored by the Hungarian Psychological Association and the Eotvos Lorand University of Budapest. Volumes 15 and 16 were based primarily on papers presented at the 1994 International Congress of Applied Psychology in Madrid. Volume 17 differs from the preceding volumes in this series in that it is not based on papers presented at special conferences or professional meetings. Individual authors from a number of different countries who were engaged in research on stress and emotion were invited to contribute chapters based on their current work. Special attention was also given to inviting authors whose research was related to occupational stress, and the experience, expression and control of anger. The contributors to Volume 17 were from 12 different countries located on five continents. The six chapters included in Part I examine the effects of different sources of stress on anxiety and anger. Two of these chapters focus specifically on xiii

RT44367_C000.indd 13

8/31/05 9:38:04 AM

xiv

CT

test anxiety. While most of these chapters report research findings, the author of Chapter 1 reports on his unique personal experience in being hospitalized for a life-threatening illness. The five chapters in Part II are concerned with the assessment of the experience, expression and control of anger and its effects on examinations and problem behaviors. The six chapters in Part III report findings related to job stress and its effects on productivity and performance in six different countries. The contents of this volume will be of interest to social, behavioral and medical scientists who are concerned with stress related effects on emotions, and with the assessment and treatment of stress related disorders. The findings reported in the individual chapters demonstrate important relations regarding the effects of stress, emotions and personality on performance in a wide range of situations, and on family relationships and work performance. For their contributions to Volume 17, the series editors are very pleased to express our appreciation to the authors of each chapter, and to Virginia Berch, Christine Tam and Eric Reheiser of the University of South Florida for their assistance in corresponding with the chapter authors, and in preparing the manuscript for publication. Charles D. Spielberger Irwin G. Sarrason

RT44367_C000.indd 14

8/31/05 9:38:04 AM

I STRESS, ANXIETY, AND COPING

RT44367_C001.indd 1

7/13/05 9:33:02 AM

RT44367_C001.indd 2

7/13/05 9:33:02 AM

1 Stress in a Life-Threatening Illness Pittu Laungani Honorary Research Fellow, Manchester University, England Abstract When life-threatening illnesses sometimes appear full-blown, out of the blue, doctors conduct a series of tests to determine the nature and etiology of the illness. In the case of rare diseases, the investigative procedures may take several weeks, during which the patient’s condition may deteriorate dramatically, and the patient may become severely confused, frightened, and distressed. What effects do such stressful states have upon the patient’s well-being and eventual recovery? For the hospital staff, the care of the patient raises several critical issues. How should a patient be dealt with in such situations? What should patients be told or not told? By whom? How? What ought to be the nature of doctor-patient and nurse-patient relationships? These issues are examined on the basis of personal experiences of the author, who was a patient in a London hospital a few years ago. Factors associated with severe stress and viable strategies for the successful management of a life-threatening illness are considered in detail.

STRESS IN A LIFE-THREATENING ILLNESS This chapter is unconventional and also quite unusual because the style and format are quite different from the other chapters comprising this volume. In keeping with scientific writing, these chapters follow the convention of being written in the third person singular. The only exception to this convention is to be found when a particular chapter, written jointly by co-authors, uses the collective pronoun “we.” Most of the chapters in this volume describe extensive empirical studies designed to test a variety of hypotheses related to stress, health, culture, and personality. The studies reported in this volume also involve the use of carefully constructed and validated psychometric instruments based on large samples, giving careful attention to the gathering and the analyses of data. Powerful statistical tests, including multivariate analyses and factor analytic techniques, have been employed to analyze and interpret the findings, which are explained in accordance with the relevant theoretical models postulated by the researchers. The scientific rigor and objectivity of

3

RT44367_C001.indd 3

7/13/05 9:33:02 AM

4

P. LAUNGANI

the research enterprise provides an elegant uniformity, which binds the chapters in this volume. This chapter conforms to none of the above patterns. What makes the chapter unusual is that, from time to time, direct speech is used, and for reasons that will soon become clear, several sections have been written in the first person singular. The chapter does not describe any large-scale empirical study involving the use of standardized psychometric instruments to measure stress, anger, or individual differences among the subjects. In point of fact, no experimental studies of any kind—on small groups or large—have been undertaken. The study to be described is based on the use of the smallest possible sample: n = 1. The merits and demerits of employing a sample size of 1 have been extensively discussed and commented upon by several distinguished experimental psychologists, including B. F. Skinner, and therefore need not concern us here. Suffice it to say that Skinner (1953) himself was extremely critical of group experimentation designs, using complex statistical methods. He was personally in favor of single-participant research methods, with their emphasis on repeated measures and replications. For a critical understanding of the issues related to the use of single sample studies, the interested reader is referred to McGuigan (1993). This chapter also breaches another important convention. There is a much respected—even hallowed—tradition within the natural and the biological sciences, which involves the separation of experimenters from the objects of their study. Such a separation has always been considered to be of paramount importance to any scientific enterprise. Scientists believed, and many still do, that clear separation from the objects of a study enables the scientist to maintain total objectivity. It was therefore incumbent upon scientists to take great care not to allow the phenomena under observation and, in the case of psychology, the people being studied to influence their observations in any way. The extent to which scientists succeeded in controlling their observations is generally considered indicative of the level of objectivity of a study. It was also considered essential that the scientist’s observations were pure, neutral, uncontaminated, and free from biases. The caveat concerning the purity of observations was emphasized strongly in the 16th century by Francis Bacon (1561–1626), who insisted that scientists should not be influenced by their own preconceptions, predilections, or prejudices, because these biases were likely to affect the purity and influence the quality of their observations. Bacon also argued that the fact that we were “unable to read of the book of nature” was due, in large part, to our own impurities. The doctrine that Bacon proposed might be referred to as veracitas naturae, the truthfulness of Nature; his views were succinctly expressed by Popper (1963): “Nature is an open book. He who reads it with a pure mind cannot misread it. Only if his mind is poisoned by prejudice can he fall into error” (p. 7). What Bacon suggests is that we should purge our minds of all anticipations, conjectures, guesses, and prejudices. Bacon’s views are similar to those expressed by Descartes (1596–1650), according to Popper (1963), who observed that Descartes’s method of systematic doubt “. . . is fundamentally the same: it is a method of destroying

RT44367_C001.indd 4

7/13/05 9:33:03 AM

STRESS, ANXIETY AND COPING

5

all false prejudices of the mind, in order to arrive at the unshakable basis of self-evident truth” (p. 15). Thus, for research to be bias-free, neither the objects of one’s research nor the biases inherent in the scientist’s observations must influence the research findings. Psychologists following in the footsteps of natural scientists have generally conformed to these time-honored conventions, viz., absolute objectivity and neutrality of one’s observations. While these conventions have come to acquire the status of a myth, when examined carefully, they turn out to be untenable. Let us see why. The myth that scientific observations need not be objective was laid to rest by Heisenberg (1930). In formulating his principle of indeterminacy in quantum mechanics, Heisenberg discovered that the presence of the scientist along with his instruments influences the phenomena under observation, which seriously compromises the notion of total objectivity. Moreover, when one turns to the philosophy of science, one learns that, in practice, total scientific objectivity turns out to be a myth (Brush, 1976; Feyerabend, 1975; Gillies, 1993; Kuhn, 1962; Lakatos & Musgrave, 1970; Scheffler, 1969; Veatch, 1972). Although scientists assert the need for objectivity and prescribe normative standards of research (Popper, 1968), the activity of science itself is pursued in a clumsy, subjective, and ad hoc manner (Lakatos, 1970). While debating the notion of the objectivity of science, Brush (1976) notes, with a tone of despair: “Perhaps one must finally ask: Are the standards of objective scientific method worth preserving, even as ideals that are rarely attained in practice? Or do we distort our understanding of the nature of science by paying lip service to such standards?” (pp. 81–82). Feyerabend (1978), for whom scientific activity is largely anarchic, goes a step further in arguing that there are no rules, no guiding principles to which scientists all over the world subscribe. Although they would like to believe they do, what they believe in and what they actually practice are virtually unrelated. The guiding principle in science that Feyerabend points out is “Anything goes!” The second convention, which refers to the purity or neutrality of one’s observations, is also flawed. Popper (1963, 1972) has pointed out that all observations presuppose a conjecture, a hunch, an idea, or in some instances, a clearly formulated set of hypotheses deduced from a theory. According to Popper, observations are made in order to test hypotheses, not to generate one. Popper argues that the belief in the neutrality of observations arises out of an acceptance of the inductivist model in science. The problem of induction was also of great concern to Hume (1711–1776). In his exposé of the inductivist fallacy, Hume emphasized the untenability of inductivist epistemology as a secure basis for objective knowledge. The supposed neutrality and passivity of observations was also questioned by Kant. In the preface to the second edition of his Critique of Pure Reason, Kant (1781/1787) writes: “Our reason can understand only what it creates according to its own design: that we must compel nature to answer our questions rather than cling to Nature’s apron string and allow her to guide us . . .” (p. 49). Similar observations have also been made in psychology. Rosenthal and Rosnow (1969, 1975) described several experiments in which, by a series of subtle, nonverbal cues, the investigators’ hypotheses were tacitly conveyed

RT44367_C001.indd 5

7/13/05 9:33:03 AM

6

P. LAUNGANI

to the study participants, who in turn behaved in a manner that led to the confirmation of the hypotheses. The experimental subjects who participate in an experiment come with their own preconceptions concerning the nature and the purpose of the experiment. They, too, are involved in some sort of a guessing game, trying to figure out what it is that the experimenter is trying to achieve. Thus, the experimental situation creates its own demand characteristics (Orne, 1962). In other words, once they have figured out the nature and the purpose of an experiment, the subjects help the experimenters to obtain the results they are looking for! Rosenthal and Rosnow (1969, 1975) claim that such a finding is true not only of human subjects, but for animals as well! Most psychologists are familiar with the story of Clever Hans, the wonder horse who was thought to possess the ability to solve simple mathematical problems, until this “ability” was discovered to be little more than a conditioned response to a nonverbal cue offered inadvertently by its trainer (Rosenthal, 1964). Psychologists, of course, have not been slow to respond to the concerns expressed by other scientists, and by philosophers of science, which have contributed to a refreshing rethinking within both science and psychology. There is a feeling within certain areas of psychology that the old venerated methods of research, which were an imitation of the methods of the natural sciences, are limited in their usefulness (Alarcon, 1992; Jain & Misra, 1991; Kovac, 1991). Some even see them as outdated, particularly in view of the rethinking that has preoccupied practitioners of the natural sciences (Bickhard, 1992). As was pointed out earlier, the objectivity-subjectivity dichotomy has come to be seriously questioned. Scientists, by their presence, or by the use of instruments, influence the phenomena that they observe. In that sense, complete objectivity is now recognized as an unattainable ideal. Objectivity is best realized in terms of degree, ranging from low to extremely high levels. The chapters in this volume display high levels of objectivity in terms of research design, the selection of large, representative samples, the use of standardized tests, and complex statistical analyses and interpretation of the data. Participant Observation This chapter does not lay any claims to objectivity in the above sense. However, the research methodology that is followed bears some resemblance to a form of participant observation. Participant observation is a popular technique in fieldwork that has been used extensively by anthropologists, sociologists, and more recently by psychologists. As Judd, Smith, & Kidder (1991) point out, many “. . . field workers participate fully in the lives of the people they study, by being or becoming members of the group they study. Others remain outsiders, purely observers” (p. 301). Between these two extremes, there are several gradations in observation and participation. The concept of participant observation carries with it, albeit loosely, the object-subject dichotomy. The experimenter or investigator is part of the group and yet apart. In such a setting, the investigator, while involved in the group

RT44367_C001.indd 6

7/13/05 9:33:03 AM

STRESS, ANXIETY AND COPING

7

activity, is expected to retain a detached distance. The observer, who is part of the ongoing action, at the same time observes the action or phenomena. Objectivity in such a situation is to be understood in terms of the extent to which the observer is able to be part of the group and yet remain apart. Not only are the observers observing the group, but they are also observing themselves while observing the group activity. The potential danger of such an exercise is that it can lead to infinite regress. Another danger with participant observation is its reliance on introspective reports. Observers interpret an ongoing activity of which they are part, and which they are also observing. Thus, the reports of observers often turn out to be introspective accounts of what has transpired; these accounts are the observers’ versions of perceived social and external reality. Given the introspective nature of such reports, there may be some doubts concerning their reliability and validity. The reported observations may indeed be a true reflection of “external reality,” but in the absence of any objective criteria against which the accounts might be systematically assessed, one cannot be certain. In recent years, however, there has been a resurgence of interest in the uses of introspection in psychological research. A number of writers (e.g., Howe, 1991; Laplane, 1992; Morstyn, 1992; Ryback, 1974; Sakamoto, 1984) have also argued that it would be unwise to dismiss introspective reports out of hand. The technique of participant observation, with its reliance on introspective reports, is clearly not without its problems. Its strengths and weaknesses in psychological research have been discussed extensively (e.g., Bauer, 1984; Judd et al., 1991; Mathews & Ridgeway, 1984; Meyer, 1992; Wilson-Barnett, 1992). Yet distinguished psychologists, including Festinger, Reicken, and Schachter (1956), Judd et al. (1991), and others, have shown a great deal of ingenuity in designing field studies that rely on methods of participant observation. In the present case, however, the notion of participant observation, as explicated by writers in this field, has been seriously compromised. As mentioned above, participant observation assumes that the observers are, in a sense, participants in an ongoing social activity or phenomenon, and are also simultaneously involved in observing the very social activity or phenomenon of which they are part. In this chapter, the researcher and the subject of the research are one and the same person. Thus, the outsider becomes the insider; object becomes the subject; investigator becomes the patient. It should be noted, however, that the investigator was not a pseudopatient, as was the case in Rosenhan’s (1973) well-known study, “On being sane in insane places.” Nor did the investigator adopt a variety of deceptively realistic roles, emulating a research strategy popularized by Latanè and Darley (1968), Milgram, Sabini, and Silver (1992), and Zimbardo (1971). Situations where researchers become their own subject are not unknown in psychology. Ebbinghaus, in his highly original experiments on memory, used himself as his own subject. Concentrating on frequency of repetition as the essential condition of association, he showed how repetitions could be used as a measure of memory (see Boring, 1957). Another famous study that involved the investigator studying himself was that of Bruno Bettelheim (1970), who

RT44367_C001.indd 7

7/13/05 9:33:04 AM

8

P. LAUNGANI

was incarcerated by the Nazis in two concentration camps, Dachau and Buchenwald. Bettelheim realized that his chances of surviving the horrors of the concentration camp were very low, unless he adopted a strategy that would enable him to overcome hunger, cold, overcrowding, senseless beatings, illness, and intense humiliation and dehumanization. With a contrived “switch” of his mind, he convinced himself that he was a research scientist sent to the concentration camp to investigate the problems that arose there. This is an “old trick,” often used by psychotherapists working within a psychoanalytic framework, which allowed Bettelheim to distance himself from his own sufferings and from those of the others in the camps. Bettelheim survived his stay in the concentration camps and recorded his experiences in his intensely moving book, The Informed Heart (1970). Although the present author’s situation cannot be compared with that of Bettelheim and the brutalities he experienced, there are parallels to be drawn. The author found it necessary to switch into the role of a research scientist, which enabled him to distance himself from his own illness, pain, and suffering. Using this strategy helped him to transform his life-threatening illness into a fascinating case study that had to be examined with great care and detachment. It should be noted, however, that such a strategy wasn’t always successful. On several occasions, the pain, weakness, fear, and stress were so acutely unbearable that it completely undermined his attempts to treat them as sets of variables associated with a life-threatening illness. Under such circumstances, it was impossible to sustain the role of a research scientist who was investigating the psychological processes involved in a life-threatening illness. The life-threatening illness appeared suddenly and aggressively. The doctors attempted to determine the nature and the etiology of the illness. However, it was a rare disease, and their investigative procedures took several weeks, during which the patient’s condition deteriorated dramatically. He became severely confused, frightened, and distressed. What follows now is a brief account by the author of his illness and his attempts to understand it and examine its impact on his psychological and physical well-being. This account explores the unique techniques that he “invented” in order to cope with the severe stresses induced by his illness. A more comprehensive account of his experiences has been described elsewhere (see Laungani, 1992). BACKGROUND TO THE STUDY In July 1989, I went to India. I spent six weeks in Bombay undertaking a research study related to patterns of socialization in children in Dharavi, West Bombay. Regarded as the largest slum in Asia, Dharavi covers an area of 7.1 square kilometers and has a population of approximately 1 million residents (Kothari, 1985; Sharma, 2000). The living conditions in Dharavi are beyond belief. It is overcrowded, dirty, filthy, and polluted. There is an acute shortage of clean water, lack of adequate sanitation, and no proper sewage system to speak of. During the monsoons (mid-July to mid-September), it becomes a breeding ground for infectious diseases, which result in high levels

RT44367_C001.indd 8

7/13/05 9:33:04 AM

STRESS, ANXIETY AND COPING

9

of infant mortality. I worked in Dharavi for over a week, visiting families and observing children, before returning to London. Within three weeks of my return to London, I fell acutely ill and was admitted into the Central Middlesex Hospital in North West London. To get admitted into a hospital in England on the National Health Service, one normally needs a referral letter from one’s G.P. One then waits in the Accidents and Emergencies (A&E) area until examined by a doctor, and a final decision is made concerning the nature and the severity of the illness and whether or not the patient will be admitted. My wife handed our G.P.’s referral letter to the receptionist, and we waited, along with other patients, to be called. After a short wait, I was ushered into an anteroom in which there were several beds. Presently, a young nurse came up to me and said, “Please take your clothes off and lie in this bed.” Although her statement had been prefaced by the word please, it was clear from the firm and authoritative way in which she spoke that it was not meant to be taken as a request, but as a command. I was genuinely puzzled. Why did she want me to take my clothes off? I was also shocked by the intimacy that the command implied and resentful of the threat to my dignity. Moreover, I was angry that she could give me orders in such a peremptory manner. “Why do you want me to take my clothes off?” I inquired. Without answering my question, she walked away and summoned two other nurses who were standing nearby. “The patient refuses to take his clothes off!” she complained as the three of them turned to me. I tried defending myself by arguing that (a) I wasn’t a patient, and (b) I wasn’t refusing to take my clothes off. All I was seeking was an explanation. None of them responded to my defense. This incident, though trivial in itself, made me aware of three points. First, upon entering the hospital, my own individuality and independence had been taken away from me. Only a couple of hours earlier, outside the hospital environment, I had been a free agent. I had also entered a new environment that had its own rules and norms. It had its own culture. To stay in hospital, I was now expected to learn and follow the rules of the new culture, and to play the role of patient. I had to play this role not as I might have seen fit to define it, but as it was being defined for me. As a potential patient, still to be officially admitted, I had made my first mistake: In seeking a reasonable explanation of the nurse’s command to take my clothes off, I was accused of disobeying her orders. In that instant, it became clear to me that disobedience was seen as a serious misdemeanor. I took my clothes off and lay in bed. During the course of the day, several doctors came to see me, each carrying out their own specialist investigations. Several hours later, the Senior Registrar came and spoke with me. He said that he had examined all the reports of the investigations, explained that I was seriously ill, and that he had decided to admit me into hospital. I was very weak, dehydrated, and anemic. My breathing was shallow, and it appeared that my kidneys were not functioning adequately. Ignoring all the other symptoms, I focused on my kidneys. “What’s wrong with my kidneys?” I cried in alarm, my mind terrorized by visions of being strapped to a dialysis machine for the rest of my life. Although the Senior

RT44367_C001.indd 9

7/13/05 9:33:04 AM

10

P. LAUNGANI

Registrar reassured me that there was nothing to worry about, he informed me that I was critically ill and needed looking after. He also said he had decided to admit me to the ITU, where they could carry out all the required investigations and monitor my progress during the night. I stared at him in disbelief. I was aware that I was ill, but not that I was critically ill. In the last two weeks or so, I knew I had been running high temperatures. I was also finding it increasingly difficult to climb stairs, my joints ached, my body was racked by pain, and I was unable to eat and swallow. Breathing was becoming an acute problem, and I was even having serious difficulties in squeezing toothpaste out of a tube. But in my own misplaced optimism, I had disregarded the severity of these symptoms. I had convinced myself that I was passing through a temporary phase, a bad patch so to speak, which would disappear after appropriate medication and rest. But the word ITU struck terror in my heart. It aroused in me a feeling of dread and foreboding. From the depths of my unconscious arose terrifying images, which until then had remained concealed. I saw myself being strapped to all sorts of complicated machinery, my nose and throat invaded by tubes attached to respirators, wires attached to my chest, my arms bruised and punctured by drips of saline, albumin, and blood, my mouth compressed by oxygen masks. I saw the nurses keeping a close vigil during the night, watching the monitors in case my heart skipped a beat or showed an erratic rhythm, or in case my blood pressure rose or dropped dangerously, or my breathing slowed . . . or stopped altogether. As though sensing my fears (for I had suddenly become very silent), the doctor tried to reassure me. “There is nothing to be frightened of. The ITU is one of the safest places to be in a hospital.” For a moment, I thought he was being flippant. Even in my confused state, I knew that one was taken into the ITU only when critically ill. The ITU could hardly be described as a safe place for me to be in, but I was too weak and tired and drowsy to argue the point. The incident with the nurse in the A&E had chastised me into meek submission. Transfer to St. Mary’s Hospital A day or two later I was told that I had developed an acute renal crisis. I did not know what was meant by the word crisis and asked for an explanation. No one bothered to explain. In my own fevered imagination, I reinterpreted this to mean renal failure. Once again, the specter of being attached to a dialysis machine for the rest of my life leered menacingly at me. Luckily, I was given no time to elaborate on these terrifying images. For even before the news could sink in fully, and before I could fully comprehend the implications of acute renal failure, I was informed that the Central Middlesex Hospital did not have a specialist renal unit and that they were transferring me to St. Mary’s Hospital, which they did. I was also assured I would be in better hands there. It was clear that I had had no say whatever in the decision concerning the transfer. Because the hospital ambulance drivers were on a national strike, I was transported to St. Mary’s Hospital by two young police officers in a police van and was admitted. Despite the medical emergency, for that was how

RT44367_C001.indd 10

7/13/05 9:33:05 AM

STRESS, ANXIETY AND COPING

11

I construed my own condition, I was disappointed that the police officers refused to turn on the siren when driving me to the hospital, where I was immediately assigned into the renal ward. The nurses on the ward made me as comfortable as they could. While I lay in bed, hovering between sleep and wakefulness, fresh IV lines were opened and, from time to time, the nurses came and collected blood samples. Apart from these intrusions, I was left completely alone. The following morning, I was seen by the Renal Consultant, accompanied by his Senior Registrar and several other junior doctors, who throughout the interchange remained attentively silent. The Consultant explained that the report from the previous hospital suggested renal complications. I asked him what that meant. He took my question quite seriously and explained that kidneys were meant to perform several functions. They acted as filtering mechanisms; they produced urine; they maintained balances such as fluid, electrolyte, and inorganic chemical balances; and so on. My kidneys were not performing any of those functions. Once again, the terror of dialysis returned, and I asked the Consultant if that meant I might be kept on a dialysis machine. He said he very much hoped that it wouldn’t come to that, and promised to send his Senior Registrar to see me later. I looked at my wife, seeking reassurance from her calm face. At some point, I fell asleep. Late in the evening, the Senior Registrar came to see me. He said he had examined all the reports of the blood tests that had been done and had come to the conclusion that the next logical step was to perform a kidney biopsy, which he would do the following morning. He handed me a printed sheet of paper and asked me to read and sign it. It was a voluntary consent form. In an abstract sense, I had the option of saying yes or no; but in a very real sense, I had little or no choice. Therein lay an interesting paradox. If I refused to sign, I was paradoxically preventing further examinations and was potentially endangering my own life. Yet if I signed, I was giving the doctor the right to inflict pain upon me and carry out investigations that could lead to potentially harmful unintended consequences. In signing the form, I was also absolving him from any blame in the unlikely event of the biopsy leading, at best, to unknown and unintended consequences, or at worst, to a fatal conclusion. I signed the consent form. A few days later, the Senior Registrar informed me that the biopsy had revealed nothing of any significance. As a matter of fact, my kidneys had stopped “misbehaving” and, for no known reasons, my renal functions had returned to normal. From this it was clear that the problem lay elsewhere. Fresh investigations started all over again, and I was subjected to a variety of tests. Notwithstanding the scores on these tests, the biopsies that were previously done on me, and the dozens of specialist consultants who came to visit me every day, the nature of my illness remained undiagnosed. Because I had only just returned from India, there was a suspicion that I may have picked up a tropical disease. As a result, I was examined by groups of specialists in tropical medicine, who carried out their investigations and left. However, none of them were able to arrive at an accurate diagnosis. My undiagnosed illness remained a mystery. As the days went by, my condition

RT44367_C001.indd 11

7/13/05 9:33:05 AM

12

P. LAUNGANI

worsened. I found it increasingly difficult to bear the stress of not knowing what unidentified illness or disease I might be dying of. I wanted to be certain. I wanted to know. For I knew that in certain knowledge there was also hope. Sadly, my illness, whatever it was, remained elusive. Yet, even in my frightened, delirious, and confused state, I noticed that the doctors were doing their utmost to arrive at an accurate diagnosis. They were attempting to do this by a process of elimination. “If not this, then that, if not A, then B.” It was obvious that their investigations were not guided by any theoretical formulations. For them, it was a tedious empirical process. For me, it was a nightmare that involved my being taken in a wheelchair, and when my condition worsened, on a bed trolley, to various units and departments of the hospital. Each visit tired me intensely. I felt increasingly exhausted and drained of all energy. As the search for an accurate diagnosis continued, my condition deteriorated. I developed frightening complications, and became paralyzed. Much to my horror, I found that I could not move my body, nor could I exercise any control over my limbs. I could neither raise my hands, nor turn in bed, nor raise myself to a sitting position. I found it impossible to swallow any fluids, and developed severe respiratory difficulties. I could barely raise my voice above a whisper. The doctors obviously were baffled by my illness. One or two of them saw my mysterious illness as an intellectual challenge, thus separating the patient from his illness. In their naive enthusiasm, they even told me so. Given the critical state I was in, I could not bring myself to share their misplaced enthusiasm. Arriving at a Correct Diagnosis Strangely enough, it was the paralysis, combined with respiratory difficulties, that eventually helped the medical staff to narrow down the diagnosis to a neurological or muscular dysfunction. To confirm this diagnosis, I was subjected to two further tests: a myograph and a muscle biopsy. The test itself, which was quite painful, consisted of my being given a series of sharp electric shocks that were administered to different parts of my body. The purpose of the test was to measure my muscular and neurological responses. The results of the myograph revealed that my neurological functions were unaffected. My muscles, however, were very severely damaged. The doctors looking after me were almost certain as to the precise nature of my illness. They were of the opinion that I was suffering from a disease known as polymyositis. But because there were several parallel diseases with similar symptoms, they recommended a muscle biopsy to be absolutely certain of their diagnosis. A decision was made to send me to National Hospital for the muscle biopsy. In my state of total exhaustion, I was extremely unhappy about the impending visit to another hospital. Was it not possible for the biopsy to be done where I was, I inquired. I was told that the National Hospital had all the facilities for such types of investigations. Under normal circumstances, the entire trip to and from National Hospital would have taken less than three to four hours. However, due to the ambulance drivers being on strike, it took nearly 20 hours

RT44367_C001.indd 12

7/13/05 9:33:05 AM

STRESS, ANXIETY AND COPING

13

before I returned to the “safety” of my own bed in the ward to which I had become accustomed. Retrospectively, I am now convinced that had my wife not accompanied me and prevented my spirits from flagging, I would have succumbed to the ordeal that this journey imposed upon me. The entire round-trip, to National Hospital and back, left me feeling totally exhausted. Transfer to Intensive Care Upon returning to my hospital, I was informed that I was going to be transferred to the ITU and was moved there the following day. While all the investigations were being carried out, I was given no medication and no food. As a result I had become extremely weak, frail, and emaciated, and I had lost almost 45 pounds during the three weeks that I had been in hospital. I wanted to know why I was being denied food and medication. The Senior Registrar, with whom by now I had become quite friendly, explained that an accurate diagnosis required them to carry out pure, uncontaminated investigations, and that any intake of food or medication would contaminate their observations. But now that the diagnosis was confirmed, there was no reason to delay treatment. I was told that I had acquired an extremely rare disease called polymyositis, which, as I subsequently learned, occurs in the ratio of only 1 of 250,000 people in any given population. As noted by Hochberg, Feldman, & Stevens (1986, p. 168) and confirmed by other investigators (see DeVere & Bradley, 1975; Ogle, 1980), polymyositis “is a disorder of unknown etiology characterized by an inflammatory myopathy, involving striated skeletal muscles and less commonly in the cardiac muscle.” Although the cause of polymyositis is unknown, the underlying problem is an autoimmune disorder in which the immune system attacks the muscles in the body. Individual muscle fibers become inflamed and are then destroyed by the immune system. According to Mitchell (1992), “The muscles around the shoulder girdle tend to be worst affected, and the polymyositis sufferer develops tenderness, pain, and weakness in these muscles” (p. 198). Other muscles may also be damaged by polymyositis, such as those concerned with breathing and swallowing, making it necessary for the patient to be intubated and, if necessary, ventilated on a breathing machine. I should point out that my understanding of polymyositis was gradually acquired. The knowledge that I gained came by probing, asking questions, seeking explanations, and “not letting the doctors go” until even the most uncommunicative of them had answered some of the questions to my satisfaction. I felt, though, that the information should have been offered to me as a matter of course, even as a right. To have been given information only because of my continued insistence was not right. There was a moral problem there. It needed to be discussed, which, of course, I did. But the doctors argued that, on the occasions in which they had tried to give me information, I was too weak, delirious, and unwell to respond rationally to their explanations. There the matter rested. Most of the scientific information related to polymyositis was acquired later—much, much, later—after I had

RT44367_C001.indd 13

7/13/05 9:33:06 AM

14

P. LAUNGANI

been discharged from hospital. Enforced convalescence gave me the leisure to pursue my interests in my medical problems. I lay in the ITU wondering about the critical nature of my illness and all of the other matters related to its etiology, diagnosis, treatment, and prognosis. By now, I was also completely paralyzsed. Because no one seemed to know a great deal about my illness, its etiology, or its prognosis, I became very agitated and frightened, and turned to my wife for support. I never knew when—or even if—my wife ever went home. Whenever I opened my eyes, I saw her sitting by my bedside. She usually came alone, but was accompanied on a few occasions by my relatives, who had come from different parts of the world. (For the benefit of the reader, it should be noted that in keeping with ancient Indian [Hindu] tradition, ours is a family-centered and community-oriented society.) Over the years, the members of our large, extended family have spread extensively and settled in different parts of the world. For a detailed account of the structure and the psychological functions served by an extended family network in Indian society, the reader is referred to Kakar (1981), Laungani (1995, 1996, 2001), and Roland (1988). Upon hearing of my “serious illness”—as they put it—my relatives found it necessary to put aside their private concerns and fly vast distances to be by my side. They arranged among themselves to come in turns. I found the presence of my relatives, particularly my sisters and nephews, who came from distances as far away as Hong Kong, Manila, Bombay, and Athens, to be very, very comforting. My friends, too, in turn came to visit me. One of the persons who visited me in the ITU was my publisher, David Whiting of Whiting & Birch in London. I could see that he was shocked by what he saw. I was like an intricate piece of machinery, wired to a variety of monitors, which bleeped and whizzed and recorded the necessary cardiovascular and other bodily functions. He saw tubes going into my mouth, tubes running through my nostrils, drips connected to my arms, ankles, and wrists, an oxygen mask suspended over my face. David sat in a chair, looked at me intently, and said rather abruptly: “Forget about the book on India which you are planning to write. Why don’t you instead write a book based on your experiences of being seriously ill, eh?” “Not many psychologists,” he added, “would have such a unique opportunity and be in such a privileged position as you are.” He stared at me for a moment and continued: “Look. Finish the book before you die. Your wife will become a rich widow. I’ll make some money out of it. And you will be posthumously famous! What have you got to lose?” Only my life, I said to myself, smiling at his gallows humor. However, in listening to his response, I continued to be intrigued by what he said: “Not many psychologists had the ‘good’ fortune to contract such a rare and crippling disease,” and “You could obviously combine both roles: the investigator and the patient. Look at the profound psychological insights you would offer!” The opportunity was too precious to let slip, for I could not only write as an “outsider” and as a research psychologist but could also write as an “insider” and as a patient. I was not sure about the profound psychological insights he spoke of. What insights, let alone their profundity, could I possibly offer? Nonetheless, I was

RT44367_C001.indd 14

7/13/05 9:33:06 AM

STRESS, ANXIETY AND COPING

15

fascinated by the idea of writing a book. Without fully understanding what I was committing myself to, I agreed. Yet I had not the least idea what a book of this nature might contain. I convinced myself that the invitation to write the book was but a ruse—a clever attempt by Whiting to lift me out of my stress and depression. I did not believe that he would hold me to my word were I unwilling or unable to write it. However, it was nice to pretend that I had been commissioned to write a book from my “deathbed,” so to speak. This lent a different and new dimension to my illness. It allowed me to see my own illness more objectively, and with it, the environment of the ITU around me. For instance, I could now see the funny side of the morning routine, which started with the “whistle-stop” visits of the Consultant in charge of the ITU, accompanied by her junior doctors. I noticed that they usually stood a few feet away from the bed of each patient, glanced through the patient’s case notes, spoke among themselves, gave instructions to the nurse looking after that particular patient, and moved on to the next patient . . . until they had raced through the last of the seven patients in the ITU. What surprised me was that they hardly ever spoke to any of the patients they had come to visit. Most of the patients, needless to say, were in a comatose state. So in a sense, it did not matter if they did or did not speak with them. However, what surprised me even more was that their behavior toward me was no different. They stood by my bedside, whispered among themselves, spoke with the nurse, and almost immediately rushed off to another bed. This routine continued for a few days. I was distressed and angry at being ignored by them and felt they saw me with unseeing eyes. To them, I had ceased being a person with his own past history and individuality. I had been transformed into a critically ill patient with whom they saw no need to enter into any communication. This upset me considerably. I did not wish to be ignored. One morning, as they all stood by my bed, talking among themselves, I pointed out to them quite firmly, and as loudly as I could manage, that I would appreciate it if they would address their comments to me! Their reaction was one of astonishment, followed by embarrassment. Mine, of course, was of pure joy, for I had scored a point. It puzzled me that in spite of all my medical notes, which they thumbed through twice a day, they assumed, falsely as it turned out, that I, too, was in a comatose state and, therefore, did not merit the courtesy of any conversation. Whether this pattern of behavior on the part of the medical staff was a function of my being in the ITU, or whether such behavior is fairly common across wards and across hospitals, I could not tell. If they took offense at my criticism, they did not show it. It did, however, have the desired effect of changing their behavior toward me. During their subsequent visits, they were friendly, solicitous, and concerned in their inquiries after my health. One or two of them even began to call me by my first name. When my disease had been diagnosed correctly, I was immediately put on a course of treatment, which consisted of extremely high doses of steroids (prednisolone), 120 mg per day, all administered intravenously. Doubtless, there were other drugs too, but the steroids stand out sharply in my mind because I was actually asked by the Consultant if I would let them administer

RT44367_C001.indd 15

7/13/05 9:33:06 AM

16

P. LAUNGANI

such high doses. He explained that my condition was quite desperate. Because I had no understanding of what the consequences of such high doses of steroids might be upon my body (and upon my mind), I replied: “Desperate diseases require desperate remedies.” Now that the disease had been diagnosed and identified, it was necessary to put me on a regular diet. Being unable to swallow, I was fed by means of a nasogastric tube. A concoction of fluids, rich in proteins and high in calories (2,000+), flowed into my stomach 24 hours a day. Within a few days of being in the ITU, I began to realize that it all seemed quite hopeless. I did not appear to be responding to treatment. My lungs were drowning in sputum; my breathing was extremely shallow; my electrolytes were out of balance; my potassium and creatinine levels were abnormal; I was unable to eat, drink, or sleep; and my body was racked by unbearable pain. I had also lost over 40 to 45 pounds in weight, was totally paralyzed, and the steroids did not appear to be having any positive effect on my damaged muscles. My situation seemed quite hopeless. I had the uncanny feeling that the consultants, too, felt the same way. Although they did not actually say so, I could tell from their faces and the way in which they spoke with my wife that my prognosis seemed rather poor. Fear of Pain and Death For the first time in my life I had the feeling that I was going to die. I could not work out in my mind what dying meant. The only thought that went through my mind—again and again and again—was that if I died I would simply cease to be, and my body would rot and decay. To me, death did not mean a period of long sleep, nor the possibility of an afterlife, or going into heaven, or heaven forbid, into hell. As a Hindu, I might have taken spiritual sustenance from the notion of karma, with its unshakable belief in an unending cycle of births and rebirths. But, no matter how hard I tried, I could not bring myself to believe in it. To me, death at that moment of my life meant total extinction. Although it was important for me to believe in some form of spiritual or religious “force” that would give meaning, if not poignancy, to my existence, I found myself unable to do so. It is not as though I am unacquainted with death. Coming as I do from a large extended family in India, death is not an uncommon event. I have seen uncles, aunts, and children die. In my own family, I have witnessed the death of my sister, my brother, and my mother and father. I have seen their bodies grow cold and stiff. I have washed, anointed, and dressed their bodies and, as mourner and pallbearer, have carried the bier to the crematorium, where amid the religious chants of the mourners, I have seen bodies burn. Over the years I have grown used to death. In a sense I have been close to death; yet in another sense, I have remained apart. Although I have been deeply affected by the deaths of those who have been dear to me, in the final analysis, it is they who have died. Not I. Their deaths had not made it any easier for me to come to terms with the possibility of my own death. Luckily, I was too tired and in considerable pain to contemplate the consequences of my impending mortality.

RT44367_C001.indd 16

7/13/05 9:33:07 AM

STRESS, ANXIETY AND COPING

17

A few days later, the consultants in whose care I was informed me that the steroids, despite the high doses, did not appear to be working as well as they had expected. Therefore, they had decided that, in addition to the steroids, I would also be given another form of treatment, which was described to me as plasma exchange, or plasmapheresis. I knew nothing about plasma exchange. Ignorance promoted its own terrors. Although one of the doctors explained the entire process to me in some detail, I still felt extremely nervous and anxious about the treatment. I was to be given 10 plasma exchanges, one every other day. Each treatment would consist of three liters of plasma being given to me in six bottles of a half liter each. It took about three hours for the six bottles of plasma to enter my nervous system. I received the first treatment the following day. An arterial line was opened and a tube was connected to my neck. Although the treatment itself was not painful, I found it difficult to get over my primitive fear of plasma exchange. I dreaded receiving it and was relieved when it was over. I was as terrified of it on the tenth and final exchange as I had been on the first. Added to my fears was the uncertainty about when the nurse from the renal ward, who was specially trained to give the plasma exchange, would come into the ITU to administer the treatment. Even after the 10 plasma exchanges, there was no dramatic improvement in my condition. I was still racked by agonizing pain, my paralysis made it impossible for me to move my body, and I had severe respiratory difficulties. The physiotherapists came more than twice a day to help me bring up the sputum that clogged my lungs. The oxygen mask was still one of the permanent features of the wires and lines to which I was attached, I could not swallow, and it was also impossible to discharge the normal bodily functions without assistance from the nursing staff. But the doctors were convinced that the plasma exchange had had some beneficial effects. I was declared out of danger. Transfer from the ITU A few days after the plasma exchanges were completed, I was transferred from the ITU to a high dependency ward. Because I was still extremely weak and frail, I was kept in a private room. It was a great relief and welcome luxury to have a room to myself. I no longer felt like a specimen, continuously under examination. I loved and cherished the privacy that the room offered. After hours and hours of boredom in the ITU—the boredom sometimes was harder to bear than pain—it was wonderful being able to turn my head and look out the window to see the clouds racing through the sky, and to hear the screeches of seagulls and ducks that lived in the canal below and the sounds of aircraft as they came in to land at Heathrow. Hitherto, my world had shrunk to a bed and the central control station of the ITU, with only its array of telephones and television monitors. While listening to the sounds outside the window, my world expanded before my disbelieving eyes. As the days went by, my swallowing returned and I was gradually able to eat tiny morsels of soft food. Because I had lost a lot of weight, I was kept

RT44367_C001.indd 17

7/13/05 9:33:07 AM

18

P. LAUNGANI

on a high-protein diet and was allowed to eat what I liked. I ate chocolate ice cream and puddings. I started to regain partial movement in my hands and could wiggle my fingers. It was a pleasure to move my hands, which until then had remained inert by my side. With the help of the physiotherapists, I learned to raise my arms above my head, stretch them backwards and forwards, hold a cup of tea in one hand, and slide down in bed from a sitting position to a reclining position. But by no strength of effort or will could I raise myself from a reclining to a sitting position. Much to my humiliation, I still needed the services of two nurses, three to four times during the night, to be turned from side to side. I knew I had a long, long way to go before I’d learn to raise myself out of bed, stand up, take a few steps, and climb a flight of stairs. Needless to say, I was totally dejected and disheartened at my lack of progress. The doctors congratulated me on my “miraculous” recovery, but warned me not to read too much into it. They pointed out that my body had taken a severe battering, and that my muscles were extremely weak and fragile and were unlikely to heal quickly. The Consultant who came to see me was of the opinion that it would take a very long time before I was able to walk. In the meantime, I would have to settle for a wheelchair. Although he took care not to specify what he meant by a “long time,” I could tell from the hints he dropped that he was thinking in terms of two to three years, and not two to three months as I, in my insane enthusiasm, had hoped. I was devastated! I’d rather die than live a living death, an albatross around my wife’s neck. I was unwilling to believe that I would have to remain in a wheelchair for a few years, before I started walking, if at all. I told him so. But because that wasn’t his problem—and it wasn’t—he said nothing and walked away. My wife came to see me in the afternoon. By looking at her face, I could tell that she knew. We sat looking at one another, each unwilling to articulate the “death sentence” that had been passed upon me. “I think it is time for me to go,” I said to her—or words to that effect. “Doctors have been known to be wrong,” she said. “It’s your muscles, you know. You never know what physiotherapists might be able to do.” I was being handed a chocolate-coated palliative. I had never felt more alone and more sorry for myself than at that moment. She tried to nurture, bolster, raise, and kick-start my shattered confidence, by fair means and foul. I felt battered, worn out, unable to argue, unable to talk back, but she was determined not to give in, nor was she going to let me give up. The fact that I was a victim of my own duplicity did little to alleviate my depression. I could not countenance the idea of being in a wheelchair. I had no wish to be a permanent liability to my wife. I knew I would not allow my life to come to such a pass, and as the days went by, my conviction that the doctors were wrong in their long-term prognostications gained strength. There was no doubt that their judgments were based on their vast clinical experience and expertise, and that they were being nothing if not candid. However, I refused to be convinced and felt that they did not have a firm basis for making these judgments. First, they had admitted that polymyositis was an extremely rare disease. Given this rarity, it was unlikely that they would have

RT44367_C001.indd 18

7/13/05 9:33:07 AM

STRESS, ANXIETY AND COPING

19

had a great deal of experience in understanding the course of the illness. Second, as physicians, they were either unaware of, or unwilling to consider seriously, the influence of psychological factors on recovery rates. Third, none of the doctors were prepared to concede that personality characteristics could also influence recovery rates. No psychological considerations had been taken into account in making their bleak assessments. Were I to accept their clinical assessments, I knew that sooner or later I would prove them right, thus lending credibility to the notion of the selffulfilling prophecy. Therefore, I found it easier not to believe in their pessimistic foreboding and chose instead to invest faith in my own ability to make a speedy recovery. Right at the outset, I refused the wheelchair and requested that the nursing staff remove it from the room. That put the onus of recovery firmly on my shoulders, and I was determined to walk and return home as soon as possible. The idea that I might not ever again be able to walk never entered my consciousness. Help came from the nurses and the physiotherapists. Seeing me as a keen “disciple,” the physiotherapists and the nurses worked with me every day. Within four to five weeks, the prognostications of the doctors turned out to be wrong. It took about four weeks before I was able to stand and take a few clumsy steps. It wasn’t what might be called a proper walk. I walked awkwardly, with a shuffling gait, dragging one foot behind the other, always afraid that I would topple over, which, of course, I did on several occasions. Even to me such a walk seemed pitiable, but I was proud of my achievement. Every day I walked a few yards more. Although each effort was extremely exhausting, it was exhaustion of triumph and not of despair. Two weeks after I took my first step, I was discharged from hospital. It was wonderful coming home! THE NATURE OF STRESS IN A LIFE-THREATENING ILLNESS I have described briefly and sequentially the experiences of my four-month stay in St. Mary’s Hospital in London. St. Mary’s is an old, prestigious teaching hospital. As was explained earlier, my disease had not rendered me speechless, nor had it affected my ability to think, observe, and put my observations into words, and subsequently on paper. Because I did not have the strength to hold a tape recorder in my hands, I dictated my observations to my wife, who wrote them down in longhand. Later on, after I had recovered some strength, I was able to use a tape recorder. Toward the end of my stay in hospital, when I was able to sit in an armchair for an hour or so every day, I used a computer. Conceptualizing Stress Selye (1956, 1974) adopted a physiological approach to conceptualizing stress. Based on his research, he formulated the General Adaptation Syndrome (GAS), a sequence of physiological responses that occur when humans and animals experience stress. He also observed that the GAS response occurs in

RT44367_C001.indd 19

7/13/05 9:33:08 AM

20

P. LAUNGANI

three stages, which he identified as alarm, resistance, and the stage of exhaustion. These formulations have subsequently stimulated sustained research on the nature of stress in America, Britain, and in other parts of the Western world. Although Selye’s (1956, 1974) model served as an important guide to research, it soon outgrew its usefulness. There were serious limitations to his GAS model. In particular, psychologists and behavioral scientists questioned Selye’s exclusive emphasis on physiological processes, finding his explanations restrictive, reductionist, and, in some sense, tautological. They could not reconcile their own psychological constructions with the neurophysiological conclusions offered by Selye and others. Consequently, the concept of stress has been re-examined from a more psychological perspective, stimulating a broader conceptual framework, and leading to definitions of stress as a psychophysiological process (e.g., Cox, 1978; Fisher, 1986; Lazarus, 1990). Recent conceptual definitions of stress have focused on (a) stimulus factors, viz., stress is seen as a negative force impinging upon an individual; (b) response factors, viz., an individual’s emotional, behavioral, or physiological response to external or internal environmental events; and (c) interactive processes, viz., the individual’s perceptual and cognitive appraisal of internal or external environments (Cox, 1978, 1983; Fisher, 1986). These definitions have also brought to light the sharp differences that exist between physiological and psychological conceptualizations of stress (Mason, 1975). Although remarkable progress has been achieved in the field of stress in the last two decades, with new and exciting methodological and treatment implications (e.g., Lazarus, 1990), there are strong voices of dissent. For example, Cooper and Payne (1988) argue that the field of stress lacks an integrative framework for explaining the majority of research findings from a logical theoretical perspective. Recognizing that individual variations in perceptions of stress make it impossible to make unequivocal statements, let us now examine those factors that seem most likely to lead to stress in a patient suffering from a life-threatening illness. Based largely on the author’s personal observations of his extremely stressful hospital experiences, I have identified three major conditions under which a patient is likely to experience stress, which may be classified as follows: (1) extrapersonal factors, (2) interpersonal factors, and (3) intrapersonal factors. Before examining each of these factors in depth, it should be emphasized that the three categories are by no means mutually exclusive. They are all interrelated. Stress experienced under one condition may “overflow” into the others, and, conversely, a successful strategy aimed at managing stress in one condition may lead to successful coping strategies in the other categories. Each of the three factors will now be carefully examined. 1. Extrapersonal Factors Extrapersonal factors relate to those conditions, events, physical situations, structures, and so on that are independent of and beyond the control of the patient, and directly related to the very fact of being in a hospital. There are two interrelated problems that a patient experiences when first entering

RT44367_C001.indd 20

7/13/05 9:33:08 AM

STRESS, ANXIETY AND COPING

21

hospital. The first is concerned with familiar experiences that become strange for the patient. The second is the patient’s experience of strangeness itself. The familiar becomes strange. Soon after admission, patients lose their familiar ways of life. Even the most simple, mundane, and ordinary actions acquire a strange quality, which the patient is unable to explain in a satisfactory manner. For instance, the patient learns that decisions that would normally have been exercised virtually as independent choices are now made by the hospital staff. Even the simplest decisions—the type of clothes that the patient might wish to wear, using the toilet, having a bath or shower, the kind of food that the patient might wish to eat, making or receiving a telephone call, wishing to read or sleep or watch television, turning out the lights, sleeping with the curtains drawn—are now essentially made by hospital staff. Patients soon learn that they have very little control over most daily activities that, prior to hospital admission, were taken for granted. It is not because the patient is incapable or unwilling to make decisions; the patient simply is not allowed to make them. Decisions that were previously taken for granted, without much forethought, now acquire a different meaning. Thus, the patient’s world, which was once familiar and ordinary, suddenly becomes strange and bewildering, causing severe stress. It is, of course, possible for patients to question the norms, routines, and rituals of the ward into which they have been admitted, and even to rebel against the strangeness that threatens one’s sense of identity and loss of dignity. However, based on my own observations, such acts of rebellion are extremely rare. Most patients submit meekly to the norms of the ward. Occasionally, when patients question the norms or seek rational explanations, they are usually silenced with statements such as “these are the rules,” or put off with excuses such as “I’m sorry, but I’m busy right now.” To the author, it appeared that many of the rules were quite arbitrary, and formulated more for the convenience and smooth running of the hospital ward than for the care of the patient. Patients were expected to follow these rules unquestioningly. Let me offer a brief example. Being conscious of my need for privacy, I requested that the nurses on my ward draw the curtains round my bed. On most occasions my request went unheeded; occasionally a compliant nurse, after some persuasion, yielded. When I demanded an explanation from the nurses who refused, they often said, “If we let every patient do what they wanted, we’d never get through the day,” or words to that effect. The fact that one was expected to follow the norms of the ward and to experience the indignity of having one’s reasonable requests turned down without any rational explanation caused severe stress in many patients, including the author. To overcome stress, the patient has to learn to accept, at least at a surface level, the “strangeness” of the hospital setting. In essence, the patient needs to engage in a cognitive switch by turning the strangeness of the hospital environment into one of familiarity. This is achieved when the patient becomes “institutionalized,” or loses the incentive to question the arbitrary nature of the rules. In so doing, the norms of the hospital, however arbitrary and irrational they might seem, are internalized. Patients must “rearrange” their lives around

RT44367_C001.indd 21

7/13/05 9:33:09 AM

22

P. LAUNGANI

the routines and the rituals of the hospital and the particular ward to which they are assigned. The experience of strangeness. Hospitals are imposing structures. A patient who has never been in a hospital before may find its atmosphere quite intimidating. Upon admission, one enters into a strange and unpredictable environment: a “buzzing, blooming confusion.” In addition to the frenetic activity, the patient’s senses are continuously assailed by sounds, smells, and sights that are unfamiliar, bewildering, and at times, extremely frightening. Patients do not have a choice regarding the ward into which they are admitted, and virtually all decisions are made by the doctors and the nursing staff. Although unavoidable and integral to the smooth running of a hospital, these factors are nonetheless stressful to patients, especially for those who have never been admitted into hospital, as has been confirmed in a number of studies (Johnston, 1982; Pitts, 1991; Taylor, 1979; van der Ploeg, 1988; Volicier & Bohannon, 1975). Luckily, most patients are either too ill, or too preoccupied with their own illnesses, to notice the strangeness and the seeming chaos around them. It is, indeed, a strange experience for a patient in a ward with other hospitalized patients to hear the noises, cries, and moans from the other patients. It is also a strange experience to see nurses and doctors rushing about, consultants accompanied by the junior doctors doing the ward rounds, visitors streaming in, wheelchairs and trolleys being pushed, and so on, and occasionally to see a patient dying on the ward. In order to cope with these strange sources of stress, patients need to create new cognitive categories or structures that allow them to accommodate these unfamiliar experiences. They need to learn to perceive the bewildering and disturbing experiences as though they were a regular part of the hospital environment, which, indeed, they are. It is only then that the strange can become familiar and the patient may be able to overcome, at least to some extent, the stress of being in a hospital. 2. Interpersonal Relationships with Hospital Staff Interpersonal factors are concerned with the nature, quality, and types of relationships that patients form, or are unable to form, with the doctor, nurses, and other hospital staff. During the course of their hospital stay, patients come in contact with a variety of persons who, directly or indirectly, are involved in their care. They meet a wide range of personnel, including nurses, ward managers, physiotherapists, dietitians, technicians, secretarial staff, chaplains, hairdressers, catering and cleaning staff, porters, newspaper vendors, and the doctors and consultants of various specialties, particularly those in whose care the patient has been assigned. These contacts may be brief or even cursory, as when patients make a rapid recovery and are discharged from hospital in two or three days or sadly, succumb to their illness. But when the patient remains in hospital for several weeks or months, as was the case with the author, the patient comes into contact with a variety of persons. In the initial stages of their admittance into hospital, most patients are preoccupied with their own fears and anxieties. They may hardly notice the persons involved in their care and may not be aware of the sudden and

RT44367_C001.indd 22

7/13/05 9:33:09 AM

STRESS, ANXIETY AND COPING

23

dramatic changes that have occurred in their lives. But as the days go by, they become more aware of their altered surroundings and of the persons involved in their treatment and care. The daily routine of each hospital ward imposes itself upon the patient, who becomes part of the routine. In the process of adapting to the hospital setting, patients learn to distinguish between three groups of persons who are more or less directly involved with their care during their stay in hospital. The persons most important to them are the ward nurses, ward sister, the house manager, the team of doctors assigned to the ward, and above all, the consultant(s) in whose care they have been assigned. Of least importance are the domestic and cleaning staff, florists, hairdressers, and newspaper sellers making their daily rounds. In between these two groups are hospital staff whom the patients may come to regard as being of some importance even though they may not be directly involved in their care and treatment. This group consists of the porters who wheel the patient from one department to another when further investigations are required, clerical and administrative ward staff, technicians from other departments, and the auxiliary staff who serve breakfast, lunch, and dinner. How stress arises in relation to the three groups can be discussed by means of an example in which a patient may be sent to various departments and units for a variety of tests and investigations, for example, X-rays, MRI scans, and other internal examinations, such as bronchoscopies, endoscopies, gastroscopies, and biopsies. Under these circumstances, the patient is normally transported by the porters, and it is here that the patient is very likely to experience stress because the patient is not always given sufficient information by the doctors and nurses as to the nature of the examination, its duration, and the extent of discomfort or pain that will be experienced during such examinations. Unprepared for such an eventuality and not knowing what to expect, fear will feed on ignorance, and the patient might then turn to the porters for further information, clarification, and reassurance. The porters, of course, are unable to offer any reassurances or guidance, and they depart as soon as the patient has been deposited to the appropriate unit. Patients are often left wondering if the porter will remember to come back and transport them once again to the safety of their bed. The feeling of being abandoned in cold, drafty corridors may cause severe stress. On one or two occasions, this was precisely what happened to the author; he was left unattended for a couple of hours before the “mistake” was discovered and a porter was hastily summoned to bring him back to the ward. The major sources of stress experienced by the patients concerns their relationships with the nurses and the team of doctors who are directly involved in their care and treatment. From the patients’ perspective, contact with the consultants involved in their care and treatment, although minimal, is most important because they soon learn that all major decisions concerning their care, treatment, investigations, diet, discharge, and so on are normally made by the consultants. They are the powerful persons in the patient’s life. The junior doctors who are involved in the day-to-day care of the patient generally defer to the consultants when important decisions are to be made, and thus legitimize the power-based hierarchical structure within the hospital.

RT44367_C001.indd 23

7/13/05 9:33:09 AM

24

P. LAUNGANI

Consequently, if the consultant in charge of a particular patient is unwilling to answer questions and fails to allay the patient’s fears and anxieties and lend reassurance in a manner considered desirable, the patient is likely to experience quite severe stress. Several examples of such stressors that were experienced by the author, which were subsequently confirmed in discussions with other patients, verify this observation (Laungani, 1992). The day-to-day care of the patient is, of course, left to the nurses, the ward sister, and the junior staff doctors. For a variety of reasons, ranging from the trivial and ridiculous to the grave, patients may experience stress as in the following examples: The patient may be kept waiting for food or the food served may not be what was ordered (e.g., a Muslim patient, as a result of an oversight, might be served pork). Linguistic and cultural differences may prevent a patient from understanding and being understood by the nursing and the medical staff. For example, being ordered to obey commands or follow rules, which might seem trivial and irrational; being awakened and forced to eat breakfast when sleep is preferred; being asked to sleep when wishing to stay awake and read; being ignored and kept waiting. All of these problems and more are likely sources of stress. Although some of these strictures might seem trivial, they can become sources of acute stress to hospital patients. Let me illustrate these sources of hospital stress by citing an example based on my own experience. During the final days of my hospital stay, the muscles that facilitated swallowing and eating functions were gradually returning to normal, whereas the other muscles in my body were unresponsive to either prayer or effort. By no act of will could I raise myself in bed from a sleeping position to a sitting position; but when moved into a sitting position, I found that I could hold a cup of tea in my hands, which was for me a significant achievement. Every morning, one of the domestic staff would bring me my cup of tea, wake me up, place the cup by the bedside table, and leave. The tantalizing cup of tea remained beyond my reach. When I requested that the tea lady raise me into a sitting position, she claimed it was not her job to move patients. By the time one of the nurses came into my room, the tea had turned stone cold. Only occasionally, the visit of a nurse would coincide with the visit of the tea lady—and that was bliss! To an outsider, this might seem a trivial episode, hardly worth the bother, but to me the inability to drink my tea was a source of acute stress. Whenever I complained, I was told in tones of condescension that they had far more serious matters to attend to than to bother about my cup of tea. They claimed that I was being needlessly petulant, ungrateful, and small-minded. Could I not think of patients who were far more critically ill than I was? Their attempts to instill in me a sense of moral rectitude fell on deaf ears. Other than in an abstract sense, in my pain I wasn’t concerned about the fate of other patients on the ward. After several such stressful episodes, which left me feeling anxious, fretful, and demoralized, I devised a simple strategy that helped me to resolve the problem. With the help of an alarm clock, I trained myself to wake up half an hour before the expected arrival of the tea lady. It was then quite easy

RT44367_C001.indd 24

7/13/05 9:33:10 AM

STRESS, ANXIETY AND COPING

25

to summon the night nurse, who at that hour was in peaceful slumber, to come into my room and raise me into a sitting position. An hour later I was rewarded with a steaming cup of tea, which served as the delayed reinforcer for my behavioral response! 3. Intrapersonal Factors In addition to environmental and human factors that lead to stress, the most common sources of stress for hospital patients are those that lie buried within the patient’s psyche. It is important to emphasize that such fears should not be confused with Freudian notions of repression, despite their striking resemblance. Repressions, as Freud (1975) emphasized, tend to remain embedded or buried in one’s unconscious. From time to time, such repressions may find an outlet and come rushing into conscious awareness, manifesting themselves in a bewildering variety of symptoms, which the afflicted person is neither able to cope with nor comprehend. Although some of the primeval stresses and anxieties experienced by patients may indeed be explained in terms of Freudian theory, there are a host of other fears of which patients are aware and may even be able to explain satisfactorily to themselves. Of all these fears, the most startling fear experienced by most patients who are critically ill is related to death. The idea of death, which may have seemed rather remote and unreal prior to the patient’s illness, now stands out as a stark reality. Given the reminders of death in a hospital setting, it is difficult, if not impossible, to employ the subtle stratagems that patients may have used in the past to deny death. In the hospital, death cannot be forgotten or put aside. It remains a constant threat. The thought of dying, what it means to die, and what might happen when one dies arouses a variety of distressing feelings and emotions in the patient. Fear of death is perhaps the greatest source of stress experienced by patients suffering from a life-threatening illness. When I had made sufficient recovery and was able to talk to other patients on the ward, I found that some had overcome their initial reticence and were eager to voice their fears related to dying. However, the manner in which they approached the theme of death varied from one person to another, ranging from obsessive concerns to oblique thoughts. Their attempts at coping with the stress caused by the fear of death lay, to a large extent, in seeking reassurances from the doctors and nursing staff; from family members, relatives, and friends who came to visit them; and occasionally even from me. In my own case, I was not immune from such stresses. On one occasion, when I was in an extremely critical condition in the Intensive Care Unit, I became convinced that I was going to die and wouldn’t survive the night. I spoke of my impending fears to the nurse who was assigned to look after me; I was inconsolable. At first she tried to make light of my fears. When that failed, she resorted to ridicule, which only made matters worse, and I criticized her for patronizing me and not taking my forebodings seriously. Eventually, she succeeded in soothing my anxieties by indicating that my disturbed and

RT44367_C001.indd 25

7/13/05 9:33:10 AM

26

P. LAUNGANI

frightened mental state seemed plausible. She convinced me that the high-tech Intensive Care Unit was the safest place for a patient to be, and that a patient could be kept alive there almost indefinitely. At a rational level I knew that her arguments were mistaken because many patients died in the ITU. Having lain in the ITU for almost two months, I had witnessed patients dying and their bodies being removed to the mortuary. But in my emotionally overwrought state, I found that the boundaries between reality and make-believe were blurred and was prepared to be convinced by her soothing arguments. I finally drifted off into a fitful sleep. The second major stressor experienced by hospital patients is concerned with their own assessment of their condition and the chances of their making a successful recovery. Some patients, particularly those who had double and triple coronary bypass operations, were acutely aware of their condition, and they knew they were critically ill. But they were uncertain of both their shortand long-term prognosis and turned to the medical staff, who could offer only vague assurances. In such situations the patients experienced extremely severe stress. Those patients with whom I was able to speak were dissatisfied with vague explanations and found it difficult to tolerate such cognitive ambiguities. They wanted to know, as precisely as possible, what their chances were of making a reasonable recovery. I was not able to find out how those who knew that their chances of making a reasonable recovery were poor were able to cope with this devastating news. One or two of them died on the ward, and a few of them were discharged from hospital. CONCLUSION Stress is a common experience in both humans and animals, which occurs when coping resources are overstrained. Although there are obvious personality differences in people regarding their reaction to an infinite variety of stressors, in a life-threatening illness the patient is subjected to many strange and unfamiliar stressors and, consequently, is more susceptible to severe stress than in most other situations. Patients are prevented from managing and coping with stress in such a situation by a number of factors, which include: 1. 2. 3. 4. 5. 6. 7. 8.

The strange environment into which the patient is confined Unfamiliarity with the rules, rituals, and the culture of the hospital Distress caused by the illness Pain and suffering Being away from one’s home Uncertainties related to recovery and prognosis Fear of being permanently disabled Fear of dying.

Each of the above stressors makes heavy demands on the patient’s coping resources and may have long-term deleterious effects on the recovery rates of patients suffering from life-threatening illnesses. To ensure that patients

RT44367_C001.indd 26

7/13/05 9:33:10 AM

STRESS, ANXIETY AND COPING

27

understand the types of stressors they are likely to experience in hospital settings and the positive coping strategies they can initiate to reduce stress levels will require long-term educational programs. In addition to helping hospitalized patients learn to take the initiatives required to manage their stress levels, it will also be essential to re-educate the hospital staff. It will be essential to help them to understand and accept the vital role played by psychological processes in the management of stress during the treatment and recovery from life-threatening illnesses. Only when hospital staff and patients work together toward a common goal, viz., the successful management of stress, can the much-desired progress be achieved.

REFERENCES Alarcon, G. A. (1992, January–April). The value and treatment of subjectivity in science, psychology, and psychotherapy. Clínica y Análisis Grupal, 14, pp. 49–82. Bauer, P. F. (1984). Personal reflection on participant observation as a methodology in the social sciences. Pastoral Psychology, 32, pp. 140–145. Bettelheim, B. (1970). The informed heart. London: Paladin. Bickhard, M. H. (1992). Myths of science: Misconceptions of science in contemporary psychology. Theory and Psychology, 2, pp. 321–337. Boring, E. G. (1957). A history of experimental psychology (2nd ed.). New York: Appleton-Century-Crofts. Brush, S. G. (1976). Fact and fantasy in the history of science. In M. H. Marx & F. E. Goodson (Eds.), Theories in contemporary psychology (pp. 66–86). New York: Macmillan. Cooper, C. L., & Payne, R. (Eds.) (1988). Causes, coping, and consequences of stress. Chichester, UK: John Wiley. Cox, T. (1978). Stress. London: Macmillan. Cox, T. (1983). The psychological and physiological response to stress. In A. Gale & J. A. Edwards (Eds.), Physiological correlates of human behaviour, Vol. 1: Basic issues London: Academic Press. DeVere, R., & Bradley, W. G. (1975). Polymyositis: Its presentation, morbidity, and mortality. Brain, 98, pp. 637–666. Festinger, L., Reicken, H. W., & Schachter, S. (1956). When prophecy fails. Minneapolis: University of Minnesota Press. Feyerabend, P. K. (1975). Against method. London: NLB. Feyerabend, P. K. (1978). Science in a free society. London: NLB. Fisher, S. (1986). Stress and strategy. London: Lawrence Erlbaum Associates. Freud, S. (1975). New introductory lectures on psychoanalysis. London: Pelican. Gillies, D. (1993). Philosophy of science in the 20th century: Four central themes. Oxford: Blackwell. Heisenberg, W. (1930). The physical principles of the quantum theory. Chicago: University of Chicago Press.

RT44367_C001.indd 27

7/13/05 9:33:11 AM

28

P. LAUNGANI

Hochberg, M. C., Feldman, D., & Stevens, M. B. (1986). Adult onset polymyositis/dermatomyositis: An analysis of clinical and laboratory features and survival in 76 patients with a review of the literature. Seminars in Arthritis and Rheumatism, 15, pp. 168–178. Howe, R. B. (1991). Reassessing introspection: A reply to Natsoulas, Lyons, and Eri Crutcher. New Ideas in Psychology, 9, pp. 383–394. Jain, U., & Misra, G. (1991). Reflections of psychology’s scientific concerns. Indian Journal of Current Psychological Research, 6, pp. 83–92. Johnston, M. (1982). Recognition of patients’ worries by nurses and by other patients. British Journal of Clinical Psychology, 21, pp. 255–261. Judd, C. M., Smith, E. R., & Kidder, L. H. (1991). Research methods in social relations. New York: Harcourt Brace Jovanovich. Kakar, S. (1981). The inner world: A psycho-analytic study of childhood and society in India. Delhi: Oxford University Press. Kant, I. (1781/1787). Critique of pure reason [English translation by Peter G. Lucas]. Manchester: Manchester University Press, 1959. Kothari, G. (1985). An investigation of infant feeding patterns in the major urban slum of Dharavi, Bombay: Final report. Bombay, India: Dharavi Project. Kovac, D. (1991). Sciences of man and psychology: A challenge to integration of science. Studia Psychologica, 33, pp. 113–117. Kuhn, T. S. (1962). The structure of scientific revolutions. Chicago: University of Chicago Press. Lakatos, I. (1970). Falsification and the methodology of scientific research programmes. In I. Lakatos & A. Musgrave (Eds.), Criticism and the growth of knowledge (pp. 91–196). London: Cambridge University Press. Lakatos, I., & Musgrave, A. (Eds.) (1970). Criticism and the growth of knowledge. London: Cambridge University Press. Laplane, D. (1992). Use of introspection in scientific psychological research, (R. Giggins and M. Laplane-Gibbins, Trans.). Behavioral Neurology, 5, pp. 199–203. Latanè, B., & Darley, J. M. (1968). Group inhibition of bystander intervention in emergencies. Journal of Personality and Social Psychology, 10, pp. 215–221. Laungani, P. (1992). It shouldn’t happen to a patient. London: Whiting & Birch. Laungani, P. (1995). Stress in Eastern and Western cultures. In C. D. Spielberger, G. Sarason, J. M. T. Brebner, E. Greenglass, P. Laungani, & A. M. O’Roark (Eds.), Stress and emotion: Anxiety, anger, and curiosity (Vol. 15, pp. 265–280). Washington, DC: Taylor & Francis. Laungani, P. (1997). Death in a Hindu family. In C. M. Parkes, P. Laungani, & B. Young (Eds.), Death and bereavement across cultures (pp. 52–72). London: Routledge. Laungani, P. (2001). Hindu deaths in Britain: Part 2. International Journal of Health Promotion and Education, 39, pp. 114–120. Lazarus, R. (1990). Theory-based stress measurement. Psychological Inquiry, 1, pp. 3–13. Mason, J. W. (1975). A historical view of the stress field. Journal of Human Stress, 1, pp. 22–36.

RT44367_C001.indd 28

7/13/05 9:33:11 AM

STRESS, ANXIETY AND COPING

29

Mathews, A., & Ridgeway, V. (1984). Psychological preparation for surgery. In A. Steptoe & A. Matthews (Eds.), Health care and human behavior. London: Academic Press. McGuigan, F. J. (1993). Experimental psychology: Methods of research (6th ed.). Englewood Cliffs, NJ: Prentice Hall. Meyer, M.A. (1992). How to apply the anthropological technique of participant observation to knowledge acquisition for expert systems. IEEE Transaction Systems, Man, and Cybernetics, 22, pp. 983–991. Milgram, S. (1992). The individual in a social world: Essays and experiments (2nd ed.), J. Sabatini & M. Silver (Eds.). New York: McGraw-Hill. Mitchell, D. (1992). Polymyositis and the immune system. In P. Laungani, It shouldn’t happen to a patient (pp. 195–199). London: Whiting & Birch. Morstyn, R. (1992). The quantum of meaning: An approach to the paradox of self-observation. Australian and New Zealand Journal of Psychiatry, 26, pp. 287–294. Ogle, S. (1980). Retrospective study of polymyositis in Auckland over 10 years. New Zealand Medical Journal, 92, pp. 433–435. Orne, M. T. (1962). On the social psychology of the psychological experiment: With particular reference to demand characteristics and their implications. American Psychologist, 17, 776–783. Pitts, M. (1991). The experience of treatment. In M. Pitts & K. Phillips (Eds.), The psychology of health: An introduction (pp. 64–74). London: Routledge. Popper, K. R. (1963). Conjectures and refutations. London: Routledge & Kegan Paul. Popper, K. R. (1968). The logic of scientific discovery. London: Hutchinson. Popper, K. R. (1972). Objective knowledge: An evolutionary approach. Oxford: Clarendon. Roland, A. (1988). In search of self in India and Japan. Toward a cross-cultural psychology. Princeton, NJ: Princeton University Press. Rosenhan, D. (1973). On being sane in insane places. Science, 79, pp. 250–258. Rosenthal, R. (1964). Effects of the experimenter on the results of psychological research. In B. A. Maher (Ed.), Progress in experimental personality research (Vol. 1, pp. 79–114). New York: Academic Press. Rosenthal, R., & Rosnow, R. L. (1975). The volunteer subject. New York: Wiley Interscience. Rosenthal, R., & Rosnow, R. L. (Eds.). (1969). Artifact in behavioral research. New York: Academic Press. Ryback, D. (1974). A return to introspectionism? American Psychologist, 29, pp. 362. Sakamoto, H. (1984). Consciousness and introspection: From a philosophical viewpoint [Special issue: Consciousness.] Japanese Psychological Review, 27, pp. 83–105. Scheffler, I. (1967) . Science and subjectivity. Indianapolis: Bobbs-Merrill. Selye, H. (1956). The stress of life. New York: McGraw-Hill. Selye, H. (1974). Stress without distress. Philadelphia: Lippincott.

RT44367_C001.indd 29

7/13/05 9:33:11 AM

30

P. LAUNGANI

Sharma, K. (2000). Rediscovering Dharavi. New Delhi: Penguin. Skinner, B. F. (1953). Science and human behavior. New York: Macmillan. Taylor, S. E. (1979). Hospital patient behavior: Reactance, helplessness, or control? Journal of Social Issues, 35, pp. 156–184. van der Ploeg, H. M. (1988). Stressful medical events: A survey of patients’ perceptions. In S. Maes, C. D. Spielberger, P. B. Defares, & I. G. Sarason (Eds.), Topics in health psychology. New York: John Wiley. Veatch, H. B. (1972). Science and humanism. In M. H. Marx & F. E. Goodson (Eds.), Theories in contemporary psychology (2nd ed., pp. 61–65). New York: Macmillan. Volicier, B. J., & Bohannon, M. W. (1975). A hospital rating scale. Nursing Research, 24, pp. 352–359. Wilson-Barnett, J. (1992). Psychological reactions to medical procedures. Psychotherapy and Psychosomatics, 57, pp. 118–127. Zimbardo, P. G. (1971). The psychological power and pathology of imprisonment. Statement prepared for the U.S. House of Representatives Committee on the Judiciary. Subcommittee No. 3: Hearings on Prison Reform. San Francisco, October 25.

RT44367_C001.indd 30

7/13/05 9:33:12 AM

2 Anxiety, Anger, and Moral Judgment Angela M.B. Biaggio Universidade Rio Grande do Sul, Brazil Abstract This chapter reviews research on anxiety, anger, and moral judgment, on which I have worked for more than a decade. It also notes how participation in the Interamerican Society of Psychology (Sociedad Interamericana de Psicologia: SIP) has focused my attention on specific areas of research. The intertwining of theoretical hypotheses, methodological approaches, and research findings with reminiscences about my involvement in SIP are described. These topics are presented in the spirit of Albert Bandura, who cogently observed that the casual encounters of personal life often have significant impact on one’s professional career. The main results that are reported include the discussion of how moral dilemmas arouse cognitive conflict and state anxiety, how these processes are involved in moral development, and gender differences in the relation of trait anxiety to moral judgment. The development of a Portuguese adaptation of Spielberger’s State-Trait Anger Expression Inventory (STAXI) is also described. Correlations between the Portuguese STAXI scale scores and extrapunitiveness and impunitiveness, as measured by Rosenzweig’s Picture Frustration Test, provide evidence of the validity of this instrument.

ANXIETY, ANGER, AND MORAL JUDGMENT1 My career in university teaching and research was greatly influenced by Professor Julian C. Stanley, now at Johns Hopkins University. At the 1962 Interamerican Congress in Mar del Plata (Argentina), Dr. Stanley expressed interest in stimulating the development of psychology in Latin America. Toward this goal, he offered to accept promising students for graduate work at the Madison campus of the University of Wisconsin, where he was teaching at that time. Professor Antonius Benko, S.J., an active SIP member and head of the Institute of Psychology at the Pontificia Universidade Católica do Rio de Janeiro, where I was an undergraduate student, approached Professor Stanley to discuss the possibility for me and two other Brazilian students to obtain graduate training at Wisconsin.

1 This chapter is based on the author’s presidential address, which was delivered at the XXV Interamerican Congress of Psychology in Santiago, Chile.

31

RT44367_C002.indd 31

7/13/05 9:35:18 AM

32

A.M.B. BIAGGIO

Under the expert guidance of Professor Stanley, I discovered the mysteries of statistics and psychometrics, and developed a taste for measurement issues and the construction and validation of psychological instruments. My first personal contact with the Interamerican Society of Psychology (Sociedad Interamericana de Psicologia: SIP) occurred in 1964, when Professor Stanley challenged his graduate students to submit “papers” for the IX Interamerican Congress, to be held in Miami. As beginning graduate students in a foreign country who were unfamiliar with the academic jargon, we three Brazilian students looked at each other, not daring to admit that we were not sure what a paper was. With Professor Stanley’s consultation and support, I presented a paper at the 1964 SIP Congress in Miami, based on my master’s thesis-to-be, in which I investigated the validity of the Scholastic Aptitude Test as a predictor of college grades in institutions with predominantly black or white students. I dealt with correlations, regression analyses, beta weights, ANOVAs, and corrections for restriction of range. In analyzing my data at a time when there were no personal computers, nor canned programs like SPSS, I chose to use huge and noisy manual calculators, rather than getting into the labyrinth of IBM programming. It was a great honor to have this paper accepted by Professor Wayne Holtzman, chair of the SIP Scientific Program Committee, and a thrill to participate in my first Congress. It was especially exciting to discover that those big names in psychology were real people, not just authors of books! During the next couple of years I was completely absorbed in obtaining a doctorate. No longer under the protective wings of Dr. Stanley, who had moved to Stanford, I entered the Wisconsin doctoral program under the supervision of Professor Robert Grinder, later SIP Executive Secretary for North America. Consistent with Professor Grinder’s orientation, I developed an interest in the field of moral development, which was my primary area of research. My first paper on moral development was presented in a symposium organized by Dr. Gordon Finley, who invited me to participate in the 1974 SIP Congress in Bogotá. Thus, my research and academic career received continuing significant impetus through interactions generated by the Interamerican Society. Between 1965 and 1973, I had little contact with SIP, being involved with my doctoral program and beginning a career in Brazil as a university faculty member. In 1973, I was invited to participate in a symposium on anxiety at the XII Interamerican Congress of Psychology in São Paulo by Professor C. D. Spielberger, whose theory of state and trait anxiety had already captured my attention (Spielberger, 1966, 1972). Thanks to this contact with Professor Spielberger, who at that time was SIP Vice President for North America, I became involved in anxiety research and spent a month at his Center for Research in Behavioral Medicine and Health Psychology at the University of South Florida. During the 1970s, I translated, validated, and standardized a Brazilian Portuguese adaptation of the State-Trait Anxiety Inventory (STAI; Spielberger, Gorsuch, & Lushene, 1970) and conducted a number of research studies on anxiety, using this theoretical framework (Biaggio, 1990). When I began work on the Portuguese adaptation of Spielberger’s anxiety measures, the STAI had already been translated and adapted into more than a dozen languages.

RT44367_C002.indd 32

7/13/05 9:35:18 AM

ANXIETY, ANGER, AND MORAL JUDGMENT

33

Portuguese translations of the adult and children’s forms of the STAI were adapted from the published English-language originals. More recently, I have worked on adapting and validating Spielberger’s (1988) State-Trait Anger Expression Inventory (STAXI) for Brazil and other Portuguese-speaking populations (Biaggio, 1995). Research studies using the Brazilian adaptations of the STAI and STAXI have been reported at several SIP Congresses and in the test manuals for these instruments (Biaggio, 1995; Biaggio, Natalicio, & Spielberger, 1979). The Portuguese adaptations of Spielberger’s anxiety scales have been used in numerous studies to investigate problems of diverse theoretical interest. Data on the reliability and validity of the Portuguese STAI, along with norms for Brazilian children and adults, are reported in the test manual (Biaggio et al., 1979). Our earliest investigations focused on obtaining evidence of the construct validity of the Portuguese STAI State (S-Anxiety) and Trait (T-Anxiety) scales, which were administered under both stressful and nonstressful conditions. For example, in one study, these scales were given to college students immediately before a statistics examination and subsequently readministered during a regular class period. The pre-examination S-Anxiety mean score of 45.7 was more than 10 points higher than the mean of 35.3 obtained in the nonstressful classroom condition (t = 3.32, p < .005). In contrast, the T-Anxiety means of 37.3 and 37.1 were almost identical in both the stressful and nonstressful conditions, demonstrating that scores on the Portuguese S-Anxiety scale increased in response to stress, whereas scores on the T-Anxiety scale were relatively impervious to situational stress. Similar findings have been reported in investigations of examination stress with the English STAI (Spielberger et al., 1970). Further evidence of the construct validity of the Portuguese STAI was demonstrated in a study by Kacelnik, Oliveira, and Farias (1975), who tested 30 eighth-grade students in a cognitive dissonance experiment. To induce dissonance, the students were required to write arguments in favor of having to take an unexpected and very difficult examination under low-reward conditions. Portuguese STAI S-Anxiety and T-Anxiety scores obtained under these conditions were then compared with scores on these scales when they were given in a neutral condition. The S-Anxiety scores were substantially higher after dissonance induction (t = 6.78; p < .001), whereas T-Anxiety scores were essentially the same in both the dissonance and neutral conditions. Additional studies with the Portuguese STAI have reported relationships between anxiety and other constructs such as self-esteem, locus of control, and moral judgment. THEORY AND RESEARCH ON ANXIETY AND MORAL DEVELOPMENT Studies of the relationship between anxiety and moral judgment, a major focus of my research, were guided by Kohlberg’s (1963, 1981, 1984) theory of moral development, one of the most important psychological theories in this field, which continues to have important implications for philosophy and edu-

RT44367_C002.indd 33

7/13/05 9:35:18 AM

34

A.M.B. BIAGGIO

cation. According to Kohlberg, maturity of moral judgment evolves through a sequence of six stages, which are presented in Appendix A. As can be noted, Kohlberg’s stages of moral development are similar in several respects to the stages of cognitive development identified by Piaget (1948). It may be helpful to call attention to those unique aspects of Kohlberg’s theory that have had great appeal in Brazil, as well as significant influence in other developing countries. The theory postulates a universal sequence of moral development in which Stages 5 and 6, the highest levels of development, comprise what Kohlberg called “postconventional thinking.” Contrary to most social and psychological explanations, which consider internalization of societal values as the goal of moral development, Kohlberg assumed that moral maturity was attained when an individual becomes capable of understanding that justice is not equal to law, and that existing laws that are morally unjust must be changed. A major assumption in Kohlberg’s theory is that individuals are potentially capable of transcending the values of the culture in which they were socialized. This key point provides common ground with sociological theories relating to the transformation of society. In order to clarify how Kohlberg’s theory differs from other social and psychological explanations of moral development, several of these theories will be briefly reviewed for comparative purposes. Durkheim (1973) believed that morality was inextricably linked to the acceptance of a social contract. In the context of this contract, it is the group that was valued. Because acceptance and assimilation of group values and norms was essential for the functioning of society, children born into a group come to accept these norms as the basis for their moral values. It should be noted that this view is similar in a number of respects to the Freudian conception of the origins of the superego. For example, in order to resolve the so-called Oedipus complex (love for the other-sex parent), a child identifies with the same-sex parent and internalizes basic incest prohibitions. Through this identification, the child incorporates societal values as if they had not been externally imposed and then comes to accept these values, independently of external sanctions. Early-learning theorists equated conscience or morality with resistance to extinction. For example, a child who is punished for certain unacceptable behaviors refrains from performing them even when punishment is unlikely. Conversely, good behavior that has been frequently reinforced will be sustained in the absence of positive reinforcement. The technical term for this phenomenon is resistance to extinction. More recently, social learning theorists have added cognitive dimensions to learning and internalization processes, such as expectancies, incentive value, and hypothesis testing. According to Bandura’s (1978) concept of reciprocal determinism, the internalization of societal values is considered to be fundamental in the development of conscience. Similarly, in Piaget’s constructivism, one may also detect an important role for human agency in the development of morality. According to Piaget, conscience evolves from an early stage of unilateral respect for the law, which is seen as sacred and immutable, to a later stage in which conscience functions autonomously. However, in this progression, Piaget gives greater emphasis to the role of peer interaction. Consequently, even the highest levels of morality do not appear to go beyond democratic cooperation and internalization of societal rules.

RT44367_C002.indd 34

7/13/05 9:35:19 AM

ANXIETY, ANGER, AND MORAL JUDGMENT

35

Although inspired by Piaget’s cognitive-developmental model, with its emphasis on peer interaction, Kohlberg’s postconventional thinking leaves much room for conscious freedom of choice in individual development and the possibility for transformation. It is important to note that Kohlberg (1971) believed that Piaget and Freud failed to distinguish among different law-andorder authority-obeying perspectives. According to Kohlberg, sociologists “have more adequately described the rule-maintaining side of Stage 4 morality and are correct in seeing it as the normal adult morality of any society” (p. 199). Normal is used here in the sense of most frequent. While Stage 4 in Kohlberg’s theory represents conventional morality and maintenance of the status quo, individuals at postconventional levels are able to change laws through due constitutional procedures (Stage 5), or by civil disobedience, that is, standing up for one’s own principles (Stage 6), as did Gandhi, Martin Luther King, Nelson Mandela, and Giordano Bruno in their political, religious, and scientific leadership. There can be no transformation if individual members of a society do not reach postconventional thinking. Boyd and Levine (in Wren, 1990) provide examples of earlier formulations of Stage 6, which include notions of “. . . a Kantian sense of categorical oughtness: A sense of the intrinsic moral value of certain norms such as saving a human life and keeping promises,” and considering “moral law as higher than the legal law, when the two are perceived to be in conflict” (p. 151). It is the revolutionary nature of Stage 6 thinking that seems to make this stage so appealing to Brazilian developmental psychologists. Moral education programs in Brazil have not focused on children, adolescents, or delinquents, as do most moral education programs. Rather, the focus has been on working with young adults and college students for the purpose of promoting societal change by fostering postconventional thinking. In keeping with my long-standing preoccupation with theoretical integration, I have endeavored to relate moral development to the constructs of anxiety and anger as emotional states and personality traits. Kohlberg’s theory of moral judgment assumes that cognitive conflict, a concept similar to Piaget’s notion of disequilibrium, accounts for progression to higher stages of moral maturity. According to Kohlberg, it is exposure to moral dilemmas one stage above one’s own that stimulates upward change. Although widely accepted, Kohlberg’s theory has not escaped criticism. For example, in finding that moral judgment correlated positively with guilt feelings, the author (1967) argued that the correlations between moral judgment and moral behavior are mediated by personality variables such as guilt and anxiety. Relation of Moral Judgment to State and Trait Anxiety The relation between maturity of moral judgment and state and trait anxiety was examined in a study of university students in which the following hypotheses were tested: (a) Participation in a moral judgment discussion group will increase S-Anxiety, and (b) T-Anxiety will correlate negatively with maturity of moral judgment. The rationale for the first hypothesis was that cognitive conflict is threatening and, therefore, likely to evoke an emotional reaction.

RT44367_C002.indd 35

7/13/05 9:35:19 AM

36

A.M.B. BIAGGIO

The second hypothesis was based on evidence that T-Anxiety is a nonadaptive, disruptive trait that would be expected to correlate negatively with desirable characteristics such as intelligence and moral judgment. The study participants were randomly assigned to experimental and control conditions. To ensure a manageable number of students for the moral dilemma discussions, the experimental group was further divided into two subgroups. In assigning students to these subgroups, moral judgment pretest scores were taken into account to ensure a wide variance within each group that would foster different stage thinking in the discussion sessions. One of Blatt and Kohlberg’s (1975) dilemmas was used as the basis for discussion in the experimental group. Students in the control group engaged in a placebo activity that consisted of responding to a self-concept scale. Following these discussions, the Portuguese STAI was administered to all study participants. Consistent with the first hypothesis, the mean S-Anxiety score for the experimental group was found to be significantly higher than the corresponding mean for the control group. In addition, gains in moral maturity attributable to the discussion were negatively correlated with S-Anxiety for males (r = –.78, p < .001). Although the corresponding correlation for females was also negative, it was not statistically significant. The theoretical relevance of this confirmation of the first hypothesis provides evidence of a relationship between the cognitive and emotional aspects of morality, but does not support or justify criticism of Kohlberg (1963) for ignoring how emotions are related to morality. In essence, our findings indicated that cognitive conflict, as conceptualized by Kohlberg, was related to S-Anxiety as conceptualized and measured by Spielberger (1966, 1972; Spielberger et al., 1970). The second hypothesis was tested by correlating T-Anxiety scores with scores on the Defining Issues Test. As predicted, these correlations were negative, but significant only for the males (r = –.69, p < .001). Thus, although S-Anxiety was consistently higher during the discussions of moral dilemmas than when no such discussions occurred, the predicted negative correlation between T-Anxiety and moral maturity was found only for males. It is interesting to note that a positive correlation was found between T-Anxiety and moral maturity for females in a study that did not include male participants. These results were consistent with Gilligan’s (1982) findings of gender differences in moral judgment and support her criticism of Kohlberg for failing to take gender specificity into account. According to Gilligan, moral judgments for men entail abstract notions of universal moral principles of justice, whereas for women higher stages of maturity may link morality to care and responsibility, which might also explain why mature women tend to be higher in T-Anxiety. MEASURING THE EXPERIENCE, EXPRESSION, AND CONTROL OF ANGER As the focus of Professor Spielberger’s research moved from anxiety to anger, I decided to develop a Portuguese adaptation of his State-Trait Anger Expression Inventory. Anger as an emotional state, S-Anger, is characterized by

RT44367_C002.indd 36

7/13/05 9:35:19 AM

ANXIETY, ANGER, AND MORAL JUDGMENT

37

subjective feelings that can vary in intensity from mild irritation or annoyance to intense fury and rage, accompanied by muscular tension and arousal of the autonomic nervous system (Spielberger, 1988; Spielberger, Jacobs, Russell, & Crane, 1985). The intensity of S-Anger varies as a function of perceived injustice, being attacked, unfair treatment by others, or frustrations resulting from obstacles to goal-directed behavior. T-Anger is defined in terms of individual differences in the disposition to perceive a variety of situations as unpleasant or frustrating. Individuals higher in T-Anger experience S-Anger more frequently, and with greater intensity, than individuals who are low in T-Anger. Anger expression is conceptualized by Spielberger (1988) as consisting of three main components: (a) anger turned inward (Anger-In), that is, the suppression of angry feelings; (b) the expression of anger toward other persons or objects in the environment (Anger-Out); and (c) how often a person attempts to control the expression of anger (Anger-Control). The expression and control scales constructed to assess each of these anger components were combined with the State-Trait Anger Scale to form the 44-item State-Trait Anger Expression Inventory (STAXI: Spielberger, 1988), which consists of five scales and two subscales that evaluate the intensity of the experience of anger, or individual differences in the frequency that anger is experienced, expressed, or controlled. The STAXI manual (Spielberger, 1988) provides information on the construction and psychometric properties of this inventory and includes detailed data on the reliability, validity, and standardization of the scales and subscales. The names and the number of items comprising each STAXI scale and subscale are described below: State-Anger: A 10-item scale that measures the intensity of angry feelings at a particular moment. Trait-Anger: A 10-item scale that measures individual differences in the disposition to experience angry feelings. The T-Anger scale has two four-item subscales: the Angry-Temperament subscale, which measures a general tendency to experience and express unprovoked anger, and the Angry-Reaction subscale, which measures differences in the disposition to express anger when criticized or treated unfairly by others. Anger-In: This eight-item scale measures the frequency that angry feelings are suppressed and kept inside. Anger-Out: An eight-item scale that measures the frequency that an individual expresses anger toward other people or objects in the environment. Anger-Control: An eight-item scale that measures how often an individual tries to control the expression of anger. Construction and Validation of the Portuguese STAXI The first step in adapting the Portuguese STAXI was translation of the individual items by persons with expertise in both Portuguese and English. The Portuguese and English forms were then administered individually, or in small groups, to 30 bilingual subjects who were English teachers or advanced college students with English as their major and translation/interpretation as their special field of study (Machado & Biaggio, 1990). To avoid possible

RT44367_C002.indd 37

7/13/05 9:35:20 AM

38

A.M.B. BIAGGIO

Table 2.1 Correlations for Bilingual Subjects between the Seven Portuguese and English STAXI Scales and Subscales, and Alpha Coefficients for the Portuguese Scales STAXI Scales

STAXI Scale Correlations

Alpha Coefficients

State Anger Trait Anger Temperament Reaction Anger-In Anger-Out Anger-Control

.94 .82 .88 .79 .82 .74 .80

.88 .79 .77 .69 .74 .70 .80

effects of order of presentation, half of the respondents were given the English form first, and the other half responded first to the Portuguese form. The interval between the administration of the two forms was three days. Adequacy of the translation. The correlations between the English and Portuguese STAXI scales and subscales are reported in Table 2.1. All of these correlations were quite large, ranging between .74 and .94 (mdn r = .82), and were highly significant (p < .0001). The magnitude of each correlation indicated that similar results were obtained for the bilingual subjects in the present study when responding to the Portuguese or English STAXI, demonstrating that the Portuguese translation was essentially equivalent to the original English test. Reliability. The internal consistency reliability for the Portuguese STAXI scales and subscales was evaluated by computing Cronbach alpha coefficients. The alpha values, reported in Table 2.1, varied between .69 and .88, indicating satisfactory reliability, especially for the very brief T-Anger subscales, which contained only four items. The internal consistency reliability of the Portuguese STAXI was further evaluated by computing item-remainder correlations between scores for each item with the corrected score for the total scale or subscale to which the item was assigned. These correlations, which are reported in Table 2.2, were also considered quite good; correlations of .30 or higher are generally acceptable for this type of analysis. Of the 52 item-remainder correlations, 48 were .30 or higher, and 43 were significant at the .01 level. Thus, the results obtained for the bilingual Brazilian sample provide evidence of a high level of internal consistency reliability for the Portuguese STAXI scales and subscales, and essential equivalence with the original English form. Validity. The validity of the Portuguese STAXI was examined by computing correlations between the seven scales and subscales and scores on the Rosenzweig (1978) Picture Frustration Test (PFT), a projective test in which respondents are confronted with a frustrating situation and instructed to report how they would react to this situation (see Crusius, Mattos, & Biaggio, 1990). According to the PFT manual, “There is frustration whenever the organism faces an obstacle or impairment more or less insurmountable, in a way that interferes with satisfaction of any vital need” (Rosenzweig, 1978, p. 7). The

RT44367_C002.indd 38

7/13/05 9:35:20 AM

39

ANXIETY, ANGER, AND MORAL JUDGMENT Table 2.2 Correlations between the 44 Individual STAXI Items with the Score for the Particular Portuguese STAXI Scale and Subscale to Which the Item Was Assigned State Item 1 Item 2 Item 3 Item 4 Item 5 Item 6 Item 7 Item 8 Item 9 Item 10

.83** .70** .91** .87** .21 .91** .26 .36 .90** .83**

Trait Item 11 Item 12 Item 13 Item 14 Item 15 Item 16 Item 17 Item 18 Item 19 Item 20

Anger-In Item 23 Item 25 Item 26 Item 30 Item 33 Item 36 Item 37 Item 41

.74** .57** .49** .66** .77** .50** .54** .55**

.59** .68** .72** .58** .63** .56** .59** .70** .29 .43*

Temperament Item 11 Item 12 Item 13 Item 16

.69** .83** .86** .69**

Item 14 Item 15 Item 18 Item 20

Reaction .69** .77** .82** .61**

Anger-Out Item 22 Item 27 Item 29 Item 32 Item 34 Item 39 Item 42 Item 43

.45* .64** .28 .74** .60** .74** .56** .58**

Control Item 21 Item 24 Item 28 Item 31 Item 35 Item 38 Item 40 Item 44

.87** .81** .69** .81** .45* .33 .76** .76**

*p < .05; **p < .01.

organism’s reaction to the stimulus situation that constitutes this obstacle is expressed as an increase in tension. Responses to the PFT are analyzed and classified according to how aggression is expressed. A response is considered (a) extrapunitive (E) if aggression is directed outward; (b) intrapunitive (I) if aggression is directed toward the individual; and (c) impunitive (M) if aggression is avoided. The type of reaction that a respondent reports is also analyzed in terms of (1) obstacle dominance; (2) ego defense; or (3) need persistence. Thus, when direction of aggression and type of reaction are both considered, there are six response possibilities. Given the conceptualization of the STAXI scales, the hypotheses of this study were formulated in regard to the direction of the expression of aggression. The Portuguese STAXI and the PFT were administered during a regular class period to 37 college students (25 females, 12 males). The significant correlations that were found for the total sample are reported in Table 2.3. These results provide evidence for the validity of the Anger-Out and T-Anger Reaction scales, which correlated positively with extrapunitiveness, and the Anger-In scale, which correlated negatively with extrapunitiveness and positively with impunitive responses. For the female study participants, the Anger-Out and T-Anger scales and the Angry-Reaction subscale correlated positively with extrapunitiveness,

RT44367_C002.indd 39

7/13/05 9:35:21 AM

40

A.M.B. BIAGGIO

Table 2.3 Significant Correlations for the Combined Sample of Males and Females between the Portuguese STAXI Scales and Subscales and the Rosenzweig Picture Frustration Test STAXI and Rosenzweig Ranges

Correlations

Anger-Out and extrapunitive responses Anger-Reaction and extrapunitive responses Anger-In and extrapunitive responses Anger-In and use of impunitive responses

(r = .38, p < .05) (r = .33, p < .05) (r = –.39, p < .05) (r = .48, p < .01)

and Anger-In correlated positively with impunitiveness. For the males, Anger-In correlated negatively with extrapunitiveness, and T-Anger correlated positively with intrapunitiveness, which was opposite to the corresponding correlation for the females. The gender differences in the correlations of T-Anger with extrapunitiveness and intrapunitiveness lend support to the hypothesis that women are more apt to express their angry feelings, whereas men are more likely to suppress them, thus increasing their risk for coronary heart disease. Anxiety, Anger, and Risk-Taking Behavior Lummertz & Biaggio (1989) investigated the relationships between anxiety and anger and risk-taking behavior. According to Farley’s (1981, 1986) theory of risk-taking, which is based on Hebb’s (1949) early work on the psychobiology of arousal, there are intrinsic individual differences in levels of optimal stimulation. In searching for external stimulation, people with low intrinsic arousal engage in risk-taking behaviors to increase arousal, whereas those with high arousal levels attempt to reduce external stimulation. According to Farley, thrill-seekers have a Type-T personality, and he further distinguishes between two types, T+ and T–. Type-T+ persons search for socially acceptable stimulation and become creative; they are inventors, artists, alpinists, and F-1 racers. In contrast, Type-T– individuals are attracted to unacceptable stimulation and may become delinquents, criminals, alcoholics, or dangerous drivers. These characteristics have been evaluated by physiological measures and self-report psychological tests that distinguish between physical and mental risks. The relations between Type-T physical and mental reactions and scores on the STAI and STAXI scales were evaluated by Lummertz and Biaggio (1989) for 51 university students (30 females, 21 males), who responded to these self-report measures during a regular class period. For the total group, a negative correlation of –.40 (p < .05) was found between T-Anxiety and the tendency to search for “mental risks.” For males, a strong positive correlation of .77 (p < .001) was observed between the tendency to assume physical risks and scores on the STAXI T-Angry Temperament subscale, and a negative correlation of –.42 (p < .05) was found between physical risks and STAI S-Anxiety scores. For the females, the negative correlation between mental risks and Anger-In (r = –.47, p < .05) was significant. The positive correlation

RT44367_C002.indd 40

7/13/05 9:35:22 AM

ANXIETY, ANGER, AND MORAL JUDGMENT

41

between physical risks and angry temperament suggested that individuals who frequently experience anger without provocation engage in risk-taking behaviors like driving erratically, participating in dangerous sports, or indulging in delinquent behavior. Females who turn anger inward have a tendency to avoid mental risks, a pattern opposite to that found for males. Thus, Anger-In may inhibit performance and competitiveness for women with high levels of this type of anger reaction. Gurski, Giongo, Muller, and Biaggio (1990) investigated the relationships between scores on the STAXI and STAI scales for a sample of 51 college students (21 males, 30 females). Based on Spielberger’s (1988) research, they expected to find positive correlations between S-Anxiety and S-Anger, and between T-Anxiety and T-Anger, its Temperament and Reaction subscales, and Anger-In. For the total group, these predicted correlations were all confirmed: S-Anxiety and S-Anger (r = .60, p < .001); and T-Anxiety with T-Anger (r = .52, p < .001), Angry-Temperament (r = .42, p < .01), and Angry-Reaction (r = .47, p < .001); and between T-Anxiety and Anger-In (r = .49, p < .001). For the males, there was also confirmation of the hypothesized relationships between T-Anxiety and T-Anger (r = .60, p < .01) and Angry-Temperament (r = .65, p < .01). For the females, the predicted relations between S-Anxiety and S-Anger (r = .66, p < .001), and between T-Anxiety and T-Anger (r = .46, p < .01), Angry-Reaction (r = .45, p < .01), and Anger-In (r = .60, p < .001), were also confirmed. The magnitude of the significant positive correlation between state and trait anxiety (r = 50, p < .001) was similar to the value reported in the STAI manual. A number of nonpredicted significant correlations also emerged between S-Anxiety and T-Anger (r = .58, p < .001), Angry-Temperament (r = .65, p < .001), Angry-Reaction (r = .39, p < .01), and Anger-Expression (r = .35, p < .001); and between T-Anxiety and Anger-Expression (r = .54, p < .001). For the females, positive correlations were found between S-Anxiety and T-Anger, Angry-Temperament, and Angry-Reaction (r = .55, .46, and .44, respectively), all significant at the .01 level; and between T-Anxiety and T-Anger (r = .46, p < .01), Anger-In (r = .60, p < .001), and Anger-Expression (r = .57, p < .001). The large number of significant correlations suggested that chronically anxious persons tend to score higher on trait anger because they are more likely to have an angry temperament and to react with anger when criticized or treated unfairly. The obtained pattern of correlations suggested that anxiety was consistently related to state and trait anger and anger expression, but was not related to how individuals endeavor to control their angry feelings. It is interesting to note that the only difference between males and females in this study was in the relationship between trait anxiety and Anger-In. The chronically anxious women tended to turn their anger inward, where it was transformed into guilt or depression, but this relationship was not significant for the men. However, the fact that the male sample was substantially smaller may have contributed to the nonsignificant correlations for males. Muller, Gurski, Giongo, and Biaggio (1990) administered the STAXI to 47 Brazilian subjects to investigate relations between the experience, expression, and control of anger and socioeconomic level. The study participants

RT44367_C002.indd 41

7/13/05 9:35:22 AM

42

A.M.B. BIAGGIO

were classified as higher or lower in socioeconomic status, according to the scale used by the Brazilian Association of Institutes of Marketing Research (ABIPEME). Negative correlations were found between social class and Anger-Expression (r = –.35, p < .01) and Angry-Reaction (r = –.34, p < .01), indicating that persons from the lower socioeconomic classes experienced and expressed anger more frequently than those persons from the higher classes. A significant positive correlation between anger expression and gender (r = .47, p < .001) suggested that men expressed their anger more frequently and with greater intensity. Biaggio, Simões, and Gurski (1990) studied the anger reactions of patients with motor impediments who were undergoing physiotherapy. They hypothesized that these patients would differ from a control group in both Anger-In and Anger-Out because they were unable to express their anger through physical means such as gestures, pounding on the table, slamming doors, and similar expressive reactions. However, the results were only significant for the Angry-Reaction subscale. The controls had higher scores than the physiotherapy patients, indicating that they reported experiencing anger more frequently than the patients when they were treated badly. It may be that the patients’ physical limitations were so frustrating that threats to the ego were not as important. Gurski et al. (1990) examined the anger profiles of 38 adult hypertensive patients (12 male, 26 female) who were being treated at the University Hospital Outpatient Clinic in Porto Alegre, Brazil. The most interesting finding in this study was a large negative correlation between Anger-In and diastolic blood pressure (r = –.73, p < .001). The hypertensive men and women scored above the median in state and trait anger, Angry-Temperament, and Angry-Reaction; the hypertensive males scored below the median on Anger-Out. The results of this study and those previously reported contribute to the construct validity of the Portuguese STAXI and are described in greater detail in the Annals of SBPC and in the STAXI manual (Spielberger, 1988). Finally, in keeping with our concern with the integration of anger research and moral judgment, a study currently in progress investigates the relationship between anger and the maturity of moral judgment. Because most modern societies consider anger expression to be an unacceptable form of behavior, it is expected that persons who are more morally mature will score lower on most of the STAXI scales, with the possible exception of Anger-In and Anger-Control.

REFERENCES Bandura, A. (1978). The psychology of chance encounters and life paths. American Psychologist, 37, pp. 747–755. Biaggio, A. (1967). Relationships among behavioral, cognitive, and affective aspects of children’s conscience. Unpublished doctoral dissertation, University of Wisconsin.

RT44367_C002.indd 42

7/13/05 9:35:22 AM

ANXIETY, ANGER, AND MORAL JUDGMENT

43

Biaggio, A. (1990). A decade of research on state-trait anxiety in Brazil. In C. D. Spielberger & R. Diaz-Guerrero (Eds.), Cross-cultural anxiety (Vol. 4, pp. 107–119). New York: Hemisphere Publishing. Biaggio, A. (1995). Adaptação do STAXI para o portugués. Revista Iberoamericana de Avaliação e Psicodiagnóstico, 1, pp. 15–35. Biaggio, A., Natalicio, L., & Spielberger, C. D. (1979). Manual do Inventário de Ansiedade traço-Estado. Rio de Janeiro: CEPA. Biaggio, A., Simões, S. C., & Gurski, R. (1990). Expressão da raiva em pacientes portadores de impedimentos motores nos membros superiores o inferiores. Porto Alegre, Brazil: Sociedade Brasileira para o Progresso da Ciência. Blatt, M., & Kohlberg, L. (1975). The effects of classroom moral discussion upon children’s level of moral judgment. Journal of Moral Education, 4, pp. 129–161. Crusius, T., Mattos, F. B., & Biaggio, A. (1990). Correlações entre o Inventário de Expressão de Raiva Estado-Traço de Spielberger e o Teste de Frustração de Rozensweig. Encontro Anual de Psicologia. Ribeirão Preto, São Paulo. Durkheim, E. (1973). Sociology and philosophy. Glencoe, IL: Free Press. [Originally published 1900] Farley, F. H. (1981). Basic process individual differences: A biologically based theory of individualization for cognitive, affective and creative outcomes. In F. H. Farley & N. J. Gordon (Eds.), Psychology and education: The state of the union. Berkeley, CA: McCutchan. Farley, F. H. (1986). The Big T in personality. Psychology Today, 20, pp. 44–52. Gilligan, C. (1982). In a different voice. Cambridge, MA: Harvard University Press. Gurski, R., Giongo, A. L., Muller, M. L. P., & Biaggio, A. (1990). O perfil do hipertenso em relação a seu estado-traço de raiva estado-traço de ansiedade [Anxiety and anger in hypertensive subjects]. Encontro Anual de Psicologia. Ribeirão Preto, São Paulo. Hebb, D. O. (1949). The organization of behavior. New York: Wiley. Kacelnik, C., Oliveira, E. S., & Farias, M. E. (1975). Relationships between cognitive dissonance and anxiety. Unpublished senior research paper, Pontificia Universidade Católica do Rio de Janeiro. Kohlberg, J. (1963). The development of children’s orientation toward a moral order: I. Sequence in the development of moral thought. Vita Humana, 6, pp. 11–33. Kohlberg, J. (1971). From is to ought: How to commit the naturalistic fallacy and get away with it in the study of moral development. In T. S. Mischel (Ed.), Genetic epistemology. New York: Academic Press. Kohlberg, L. (1981). Essays on moral development (Vol. 1). San Francisco: Harper & Row. Kohlberg, L. (1984). Essays on moral development (Vol. 2). San Francisco: Harper and Row. Lummertz, J., & Biaggio, A. (1989). Relaçães entre o comportamento de ariscarse, ansiedade e raiva. [Relationships between risk-taking behavior and anxiety]. XXII Congreso Interamericano de Psicologia Buenos Aires.

RT44367_C002.indd 43

7/13/05 9:35:23 AM

44

A.M.B. BIAGGIO

Machado, S., & Biaggio, A. (1990). Adaptação do STAXI de Spielberger para o Brasil: Estudos preliminaries com sujeitos bilingües. Porto Alegre, Brazil: Sociedade Brasileira para o Progresso da Ciência. Muller, M. L. P., Gurski, R., Giongo, A. L., & Biaggio, A. (1990). A experiencia e a expressão de raiva e sua relação com diferentes niveis sociais em ambos os sexos [Anxiety and anger in different social classes, by sex]. Encontro Anual de Psicologia. Ribeirão Preto, São Paulo. Piaget, J. (1948). The moral judgment of the child. New York: Free Press. [Originally published as Le jugement moral chez l’enfant. Paris: Z. Alcan, 1932] Rosenzweig, S. (1978). The PF Study: Basic Manual. St. Louis: Rana House. Spielberger, C. D. (1966). Theory and research on anxiety. In C. D. Spielberger (Ed.), Anxiety and behavior (pp. 3–20). New York: Academic Press. Spielberger, C. D. (1972). Anxiety as an emotional state. In C. D. Spielberger (Ed.), Anxiety: Current trends in theory and research (Vol. 1, pp. 23–29). New York: Academic Press. Spielberger, C. D. (1988). Manual for the State-Trait Anger Expression Inventory. Odessa, FL: Psychological Assessment Resources. Spielberger, C. D., Gorsuch, R. L., & Lushene, R. E. (1970). Manual for the State-Trait Anxiety Inventory. Palo Alto, CA: Consulting Psychologists Press. Spielberger, C. D., Jacobs, G. A., Russell, S. F., & Crane, R. S. (1985). Assessment of anger: The State-Trait Anger scale. In J. N. Burcher & C. D. Spielberger (Eds.), Advances in personality assessment (Vol. 2). Hillsdale, NJ: Erlbaum. Wren, T. (Ed.). (1990). The moral domain. London: MIT Press. APPENDIX A KOHLBERG’S STAGES OF MORAL DEVELOPMENT I.

Preconventional Level Stage 1. Morality of authority and punishment Stage 2. Instrumental relativistic hedonism II. Conventional Level Stage 3. Morality of “good boy” interpersonal relationships Stage 4. Morality of “law and order” III. Postconventional Level Stage 5. Morality of democratic contract Stage 6. Morality of individual principles of conscience

RT44367_C002.indd 44

7/13/05 9:35:23 AM

3 Trait-State Anxiety, Worry, and Emotionality in Athletic Competition Frontisek Man University of South Bohemia, Ceske Budejovice, Czech Republic F. Táborský Charles University, Prague, Czech Republic Charles D. Spielberger University of South Florida, Tampa, FL Abstract A fundamental tenet of Spielberger’s Trait-State Anxiety Theory is that persons high in trait anxiety (T-Anxiety) perceive stressful situations in which they are evaluated as more threatening, and respond to such situations with greater elevations in state anxiety (S-Anxiety), than persons low in T-Anxiety. Members of the Czechoslovakian National Female Handball Team (N = 24) completed the T-Anxiety scale of the State-Trait Anxiety Inventory (STAI) two days before a practice game. They also responded to the STAI S-Anxiety scale and the Competitive State Anxiety Inventory (CSAI) immediately before and after the practice game and a competitive game that was played 1 week later. The Cognitive Interference Questionnaire (CIQ) was administered immediately following both games. The Anxiety scales all correlated negatively with game performance, but only the CSAI correlations were significant. Although the high T-Anxiety players showed a greater increase in S-Anxiety and task-irrelevant cognitive responses than those low in T-Anxiety, this predicted interaction effect was significant only for irrelevant cognitions in the competitive game. The findings were discussed in the context of theories of anxiety, motivation, and action control.

TRAIT-STATE ANXIETY, WORRY, AND EMOTIONALITY IN ATHLETIC COMPETITION Emotional reactions to stress can be observed in almost all spheres of human activity and are invariably present in competitive sports, especially in championship competition. When the pressures of competition are very high, the performance of some athletes is facilitated, while others tend to perform more poorly. Moreover, players who exhibit an outstanding performance in 45

RT44367_C003.indd 45

7/13/05 9:28:54 AM

46

F. MAN ET AL.

training and practice games may perform poorly in subsequent national and international competition. A number of general theoretical approaches have been proposed for conceptualizing the impact of stress and anxiety on athletic competition (e.g., Froehlich, 1983; Krohne, 1980; Laux, 1983; Rand, Lens, & Decock, 1991; Schwenkmezger & Laux, 1986). Although a comprehensive review of these theories is beyond the scope of this chapter, two necessary prerequisites are required to understand and alleviate the negative effects of high levels of stress and anxiety on performance in athletic competition: (1) an adequate theoretical framework for examining the complex relationships among these variables, and (2) reliable and valid diagnostic methods for assessing stressor events and emotional reactions. The conceptual framework for defining the major variables in the two studies reported in this chapter was provided by Trait-State Anxiety Theory (Spielberger, 1966, 1972), which has influenced the research of a number of sport psychologists who have investigated the effects of anxiety on performance (e.g., Hackfort & Schwenkmezger, 1980, 1989; Hackfort & Spielberger, 1989; Hanin, 1983, 1986, 1989; Klavora, 1977; Krohne, Schumacher, & Neumann, 1989; Nideffer, 1989; Schwenkmezger, 1980, 1985; Spielberger, 1983, 1986). Trait-State Anxiety Theory has also guided the development of psychometric measures of the intensity of anxiety as an emotional state (S-Anxiety) and individual differences in anxiety proneness as a personality trait (T-Anxiety). The introduction of worry and emotionality by Liebert and Morris (1967) as key concepts in their two-component model of test anxiety has stimulated the development of diagnostic instruments for assessing these components (Deffenbacher, 1980; Spielberger, 1980). Davidson and Schwartz (1976) and Borkovec (1976) have also conceptualized anxiety as comprised of two primary components, which they labeled cognitive and somatic anxiety. In a review of research on these and other components of anxiety, Morris, Davis, and Hutchings (1981) concluded that cognitive and somatic anxiety are essentially the same as Liebert and Morris’s (1967) concepts of worry and emotionality. Consistent with Liebert & Morris’s (1967) two-component model, Spielberger conceptualized test anxiety as a situation-specific anxiety trait, with worry and emotionality as correlated components (Spielberger, Anton, & Bedell, 1976), and theorized that the emotionality experienced in test situations corresponded with state anxiety (S-Anxiety), as measured by the State-Trait Anxiety Inventory (Spielberger, 1983; Spielberger, Gorsuch, & Lushene, 1970). Trait-State Anxiety Theory predicts that persons high in the emotionality component of test anxiety are more likely to experience greater increases in S-Anxiety and more worry cognitions in examination situations than individuals who are low in test anxiety (Spielberger et al., 1976; Spielberger, Gonzalez, & Fletcher, 1979; Spielberger, Gonzalez, Taylor, Algaze, & Anton, 1978; Spielberger & Vagg, 1995). The hypothesized interactive effects on S-Anxiety of situational stress and individual differences in T-Anxiety would seem to apply equally well

RT44367_C003.indd 46

7/13/05 9:28:54 AM

TRAIT-STATE ANXIETY, WORRY, AND EMOTIONALITY

47

to competitive sports situations. Martens (1977) and his coworkers defined competitive anxiety as a situation-specific personality trait and developed the Sport Competition Anxiety Test (SCAT) to measure individual differences in this trait. They have also used a brief version of the STAI S-Anxiety scale to investigate the emotional components of S-Anxiety in sports situations. However, no systematic effort has been made to explore the relation between SCAT scores with STAI T-Anxiety and the worry cognitions experienced in athletic competition, nor has the impact of competition-related worry on sports performance been investigated. Cognitive theories of anxiety (e.g., Bandura, 1977; Ellis, 1969; Meichenbaum, 1977; Wine, 1971, 1980, 1982) implicitly conceptualize self-confidence and anxiety as opposite poles of a worry continuum; that is, self-confidence is viewed as the absence of cognitive anxiety. In developing the sport-specific Competitive State Anxiety Inventory-2 (CSAI-2), Martens, Burton, Vealey, Bump, and Smith (1990) included a 27-item Self-Confidence subscale with three nine-item subscales: Cognitive Anxiety, Somatic Anxiety, and Self-Confidence. Subjects respond to each CSAI-2 item by rating themselves on a four-point scale, with choices ranging from 1 (Not at all) to 4 (Very much so). The reliability, concurrent validity, factor structure, and general psychometric properties of the CSAI are very good (cf. Martens et al., 1990; Gould, Petlichkoff, & Weinberg, 1984). In a meta-analysis of the relationship between anxiety and sport performance, Kleine (1990) analyzed the results from 77 independent samples, comprising a total of 3,539 subjects. He examined effect size as a function of anxiety measure, type of sport, gender, and the time at which the anxiety measures were given. The most important findings in these analyses were (a) anxiety and sport performance were consistently negatively related, (b) effect size was significantly greater for males than for females, (c) sport-specific tests of state or trait anxiety did not account for a larger proportion of the variance than general anxiety measures, (d) significantly larger effects were found for worry than for emotionality, and (e) the shorter the interval between competition and the measurement of anxiety, the stronger the relationship. The largest effects were obtained when the anxiety measure was administered immediately following the competitive performance. The findings in two studies of the effects of state and trait anxiety on the performance of female handball players are reported in this chapter. STUDY 1 The main goal of Study 1 was to endeavor to replicate the results of Schwenkmezger & Laux (1986), who found that high-anxious players have more task-unrelated (irrelevant) cognitions than low-anxious players, as predicted by Spielberger’s Trait-State Model of Anxiety. A second goal was to compare the scores on Spielberger’s STAI S-Anxiety scale with CSAI sport-specific anxiety scores in low- and high-stress situations. On the basis of Spielberger’s Trait-State Anxiety Theory and Liebert and Morris’s (1967) two-component model of test anxiety, it was hypothesized that

RT44367_C003.indd 47

7/13/05 9:28:54 AM

48

F. MAN ET AL.

1. Players with high T-Anxiety would have more task-irrelevant responses than those with low T-Anxiety in a high-stress selection game than in a low-stress practice game. 2. A similar person-by-situation interaction was predicted for the intensity of emotional reactions, as measured by the STAI S-Anxiety scale and the three CSAI subscales. Method and Procedure Subjects and instruments. The participants were 24 female handball players between 18 and 27 years of age, who were members of the newly formed Czech Republic National Team. All participants were attending their first training camp for 14 days. Each subject completed the following three questionnaires: 1. State-Trait Anxiety Inventory (Czech version; Muellner, Ruisel, & Farkas, 1980; Spielberger, Gorsuch, & Lushene, 1970). Due to limited time, the STAI S-Anxiety scale was selectively reduced by eliminating items that did not appear to be directly relevant to the anxiety experienced in a competitive sport. The resulting 12-item S-Anxiety short form consisted of 7 anxiety-present and 5 anxiety-absent items. 2. Competitive State Anxiety Inventory-2 (Czech version; Man, 1988; Martens et al., 1990). This inventory was also completed prior to any competition and included the following subscales: Cognitive Anxiety, Somatic Anxiety, and Self-Confidence. 3. Cognitive Interference Questionnaire (CIQ) (Man, 1987). This self-report scale was developed for assessing task-irrelevant, sport-specific cognitions by modifying items from Schwenkmezger and Laux’s (1986) adaptation of Sarason’s (1978) Cognitive Interference Questionnaire, which was translated and adapted by Man (1987). Responses to the cognitive interference items were rated on a five-point scale, from 1 (Never) to 5 (Very often). Procedure. Each player participated in two games that were assumed to evoke different levels of stress: a practice (training) game of slight importance (low-stress condition), and a highly competitive game (high-stress condition) that took place one week later. Prior to the competitive game, the players were informed that a bad performance would result in their elimination from the Czech National Team. The players were previously aware of this fact because, at the training camp, they had been informed that more players were invited than could participate in the World Championship games in Denmark in December 1989. This reality resulted in very high levels of stress for most of the players. The STAI T-Anxiety scale was administered on the second day of the training camp, one week before the first game. Immediately prior to each game, the players were given the short form of the STAI S-Anxiety scale to assess precompetition S-Anxiety, followed by the Martens et al. (1990) CSAI-2 to assess precompetition cognitive and somatic anxiety, and self-con-

RT44367_C003.indd 48

7/13/05 9:28:55 AM

TRAIT-STATE ANXIETY, WORRY, AND EMOTIONALITY

Figure 3.1

49

Effects of Stress on S-Anxiety Players Who Were High or Low on T-Anxiety.

fidence. Immediately after each game, the players reported how often during the game they had experienced the various task-irrelevant sport-specific cognitions described by the CIQ items. Results and Discussion The data were analyzed in 2 × 2 ANOVAs in which T-Anxiety (high vs. low) and stress level (low vs. high) were the independent variables. Based on the median of their T-Anxiety scores, the players were divided into high (N = 10) and low (N = 11) T-Anxiety groups. The dependent variables in these analyses were S-Anxiety, cognitive and somatic state anxiety, self-confidence, and task-irrelevant responses, with repeated measures of each of these variables in the high- and low-stress conditions. The effects on S-Anxiety of the low and high levels of stress associated with practice and competition games can be noted in Figure 3.1. Similar increases were also found for scores on the CSAI-2 Cognitive and Somatic Anxiety scales, whereas scores on the CSAI-2 Self-Confidence scale were lower in the high-stress competitive game.

RT44367_C003.indd 49

7/13/05 9:28:56 AM

50

Figure 3.2

F. MAN ET AL.

Effects of Stress on Task-Irrelevant Cognitions for Players Who Were High or Low on T-Anxiety.

Significant main effects of T-Anxiety were also found for S-Anxiety [F (1,19) = 10.64, p < .01] and all three CSAI-2 subscales: Cognitive Anxiety [F (1,19) = 10.02, p < .01]; Somatic Anxiety [F (1,19) = 5.87, p < .05]; and Self-Confidence [F (1,19) = 9.07, p < .001]. The T-Anxiety main effect for S-Anxiety can be noted in Figure 3.1, in which the mean S-Anxiety scores for the high T-Anxiety players were consistently higher than those of players with low T-Anxiety. Although the differential increase in the intensity of S-Anxiety for the high and low T-Anxiety subjects was consistent with the prediction that the high T-Anxiety players would show a greater increase from the low- to the high-stress condition, the T-Anxiety by Stress Level interaction was not significant. The T-Anxiety by Stress Level interaction effect was significant for the measure of Task-Irrelevant Cognitive Responses [F (1,19) = 5.75; p < .05], as may be noted in Figure 3.2. The players with high T-Anxiety responded to the high-stress condition with a much larger increase in the frequency of task-irrelevant cognitions than the low T-Anxiety subjects, whose irrelevant

RT44367_C003.indd 50

7/13/05 9:28:57 AM

TRAIT-STATE ANXIETY, WORRY, AND EMOTIONALITY

51

cognition scores were more similar in the practice and competitive games. Thus, the hypothesis was confirmed that players with high STAI T-Anxiety would experience a greater increase in task-irrelevant worry responses than the low T-Anxiety players in the high-stress competitive game as compared to the low-stress practice game. In contrast, the hypotheses concerning the differential effects of high and low T-Anxiety on the STAI S-Anxiety and CSAI-2 scale scores in the practice and competitive games were not confirmed. The findings in Study 1 were consistent with the results of Schwenkmezger and Laux (1986), who also found that the mean S-Anxiety scores for high T-Anxiety subjects were not very different in practice and competitive games, which supported their interpretation that high T-Anxiety subjects perceived practice games as highly threatening. However, when the results reported in Figure 3.1 for the present study were compared with the findings of Schwenkmezger and Laux (1986), there were important differences that appeared to be attributable to the greater importance of the selection game for the Czechoslovakian sample. At the time the present experiment was conducted, it was very difficult to go to Western countries due to administrative problems with the official authorities. Consequently, to be a member of a National Team offered many positive opportunities. Nowadays, the situation in the Czech Republic has become much more open to the world. In contrast to Schwenkmezger and Laux’s (1986) findings, the Czech players in the present study reported fewer task-irrelevant cognitive responses than the German players, and the high T-Anxiety players in the low-stress condition reported fewer task-irrelevant responses than the low T-Anxiety players. The number of task-irrelevant responses of the high T-Anxiety Czech players in the high-stress competitive game corresponded approximately to those of the low T-Anxiety German players in the low-stress practice game. However, the results of the Czechoslovak and German studies seemed to agree for the high-stress condition. The findings in Study 1 that scores on the CSAI-2 Cognitive and Somatic Anxiety subscales increased from the low-stress practice game to the high-stress competitive game were consistent with our predictions. However, the changes in S-Anxiety were not significantly related to the T-Anxiety level of the participants and thus failed to support the prediction from Trait-State Anxiety Theory that high T-Anxiety players would show a greater increase in S-Anxiety. Previous research in Czechoslovakia with the CSAI-2 measures had focused on the effects of anxiety on player performance (Gould et al., 1984; Gould, Petlichkoff, Simons, & Vevera, 1987; Lander & Boutcher, 1986; McAuley, 1985), rather than on differential changes in S-Anxiety for players with high and low T-Anxiety. However, the results of a dissertation study by Scalan (described by Martens, 1977, pp. 57–60), were consistent with the hypotheses derived from Spielberger’s (1983, 1986) Trait-State Anxiety Theory. STUDY 2 The goals of Study 2 were to investigate the relationship of player performance with the measures of anxiety, self-confidence, and task-irrelevant cognitions that were described in Study 1. These relationships were evaluated

RT44367_C003.indd 51

7/13/05 9:28:58 AM

52

F. MAN ET AL.

for both the success and the failure-related performance of players in the lowstress practice and high-stress competitive games. Method and Procedure Subjects and instruments. The game performance of the same subjects who participated in Study 1 was further evaluated for the low- and high-stress conditions of the practice and competitive games. Analysis of critical incidents in player performance was guided by Flanagan’s (1954) Method of Quantitative Evaluation (MQE), as elaborated by the second author (1976, 1989). These evaluations were based on direct observations of the action taking place in the game, augmented by auxiliary written records. Critical incidents (cases) were identified that were considered to have had a substantial positive or negative impact on the game performance of players as related to their potential for scoring or obtaining a goal. The choice of critical cases was made from an empirical perspective on the basis of content criteria or face validity. By identifying specific critical incidents apart from other aspects of game performance, hypotheses were derived that could be tested by using the MQE. Because it was not possible to repeat game conditions, the reliability and validity of the MQE method was assessed on the basis of the agreement of the results of independent observers. The criterion for evaluating and comparing each player’s performance was determined by a group of experts. The validity of this method was evaluated by calculating approximate correlation coefficients between experts. In applying the MQE method, 15 positive and 15 negative critical cases were observed for field players, and 15 positive and 14 negative cases were evaluated for goalies. We attempted to define each critical case according to its game utility value, using a 10-point rating scale with scores ranging from +5 points to –5 points. The calculation of the total score for a player’s game performance was based on adding the sum of positive points plus half minutes in the game, and subtracting the sum for negative points. Results and Discussion The correlations between the anxiety and self-confidence measures and the critical incidents relating to a player’s success or failure in the lowstress practice game are reported in Table 3.1. The highest correlations with player performance, as evaluated by the MQE, were obtained for the Cognitive Anxiety scale (CSAI-cog), which correlated negatively (r = –.60) with Successful Shootings (p < .01) and positively with both the number of Negative Assistances (r = .55) and Other Negative Activities (r = .50). The CSAI-2 Somatic Anxiety scale also correlated negatively with Successful Shootings (r = –.45) and positively with Other Negative Activities (r = .42). In addition, the CSAI Self-Confidence scale correlated positively and significantly with Successful Shootings (r = .44, p < .05), and the correlation of –.40 between Cognitive Anxiety scores and Total Game Performance approached statistical significance (p < .08). It should also be noted that the CSAI-2 was administered

RT44367_C003.indd 52

7/13/05 9:28:59 AM

53

TRAIT-STATE ANXIETY, WORRY, AND EMOTIONALITY Table 3.1 Correlations between the Anxiety Measures and Activities Related to Player Performance in the Low-Stress Situation (Practice Game)

Successful Shootings Unsuccessful Shootings Positive Assistances Negative Assistances Other Positive Activities Other Negative Activities Total Perform. (sum of points)

CSAIsom

CSAIsc

TaskIrrelevant Responses

T-Anx.

S-Anx.

CSAIcog

–.34

–.36

–.60**

–.45*

.44*

.26

.13

–.14

.27

.11

.10

–.06

–.09

.04

.06

–.04

.40

–.16

–.30

–.11

.17

.28

–.08

–.24

.14

.41

.03

.004

.02

.08

.07

–.08

.02

.03

.09

.13

.55** –.24 .50* –.40

.42* –.22

Note: *p < .05; **p < .01.

immediately before the practice game, indicating that self-confidence and low anxiety prior to this game facilitated performance. The findings with the CSAI-2 corroborated the results of Martens et al. (1990) in demonstrating the adverse effects of cognitive anxiety on game performance under the low-stress practice game conditions. Although the negative correlations of scores on Spielberger’s STAI S-Anxiety and T-Anxiety scales with Successful Shootings were also consistent with our predictions, these correlations were not statistically significant. However, it should be noted that the positive correlation (r = .41) between task-irrelevant responses and total game performance in the low-stress practice game, which approached significance (p < .07), was opposite to our expectation. As previously noted, task-irrelevant cognitive responses were measured as a postcompetitive variable. The results for the same players and variables under the high-stress conditions of the competitive game are reported in Table 3.2. In contrast to the low-stress practice game, the only significant finding was a negative correlation of –.43 (p < .05) between task-irrelevant responses and total game performance. The correlation of .41 between the CSAI Cognitive Anxiety Scale and Unsuccessful Shootings also approached significance (p < .07). These findings may be interpreted as indicating that thoughts of doing poorly interfered with performance in the competitive game. Man and Hrabal (1989) have discussed a number of theoretical perspectives relating to anxiety, self-concept, and ability as correlates of action control and performance. Among these theories, the concept of flow might help us

RT44367_C003.indd 53

7/13/05 9:28:59 AM

54

F. MAN ET AL.

Table 3.2 Correlations between the Anxiety Variables and Activities Related to Player Performance in the High-Stress Situation (Selection Game)

Successful Shootings Unsuccessful Shootings Positive Assistances Negative Assistances Other Positive Activities Other Negative Activities Total Perform. (sum of points)

CSAIsom

CSAIsc

Task Irrelevant Responses

T-Anx.

S-Anx.

CSAIcog

–.30

–.08

.02

.05

–.14

–.13

–.09

.21

.41

.16

–.007

–.02

.15

.09

–.19

.05

–.06

–.32

.24

.09

–.03

.37

–.09

.28

–.19

–.03

–.05

.01

–.17

.35

.09

.16

.33

.20

–.02

.09

–.22

–.09

–.09

–.13

–.13

–.43*

Note: *p < .05; **p < .01.

to interpret the results of our research. According to Csikszentmihalyi (1975), under low-stress conditions, players should be more able to become totally involved when a game starts and can forget personal problems and more readily focus their attention on the immediate situation. They also feel competent and in control, and have a sense of harmony with their surroundings. Under these circumstances, a player can concentrate on the demands of the game and temporarily exclude irrelevant thoughts and feelings from consciousness (cf. Beckmann, 1987; Csikszentmihalyi, 1985). Consistent with differences in their theoretical views, Csikszentmihalyi’s Experience-Sampling Form (Csikszentmihalyi & Larson, 1987) and Sarason’s (1978) Cognitive Interference Questionnaire measure different dimensions of cognitive anxiety that might facilitate or interfere with athletic performance. Kuhl’s concepts of action versus state orientation seem to encompass both of these theoretical orientations, especially as related to highly centered sport-related activities (cf. Kuhl, 1983, 1985; Stiensmeier-Pelster, 1988; Sonstroem & Bernardo, 1982). In keeping with Beckmann’s (1987) elaboration of Kuhl’s theory as related to sport settings, the concerns of Czech players regarding the negative social consequences of whether or not they can qualify to be a member of a National Team would contribute to generating task-irrelevant worry cognitions. Control of cognitive activity is also influenced by high-level metaprocesses (cf. Bandura, 1977) that require substantial amounts of processing capacity (cf. Man, Stuchlik, & Hagtvet, 1992). Consequently, cognitive control processes that interfere with action-centered sport activities and the positive experience of

RT44367_C003.indd 54

7/13/05 9:29:00 AM

TRAIT-STATE ANXIETY, WORRY, AND EMOTIONALITY

55

flow were likely to have an adverse impact on the performance of prospective members of the Czechoslovakian National Female Handball Team, as observed in the studies reported in this chapter. Once the members of a National Team have been selected and team participation is stabilized, the experience of worry and other negative cognitions should be substantially reduced, as was found with the Czech National Ice Hockey Team. SUMMARY AND CONCLUSIONS The central tenet of Spielberger’s Trait-State Anxiety Theory (1966, 1972), which provided the conceptual framework for this research, is that persons high in T-Anxiety tend to perceive threats to self-esteem as more stressful than persons who are low in T-Anxiety and are therefore more likely to respond to such situations with greater elevations in S-Anxiety. Two studies designed to test these predictions were conducted with members of the Czechoslovakian National Female Handball Team (N = 24), who participated in both a lowstress practice game and a high-stress competitive game. The players who participated in the studies reported in this chapter completed Spielberger’s (1980) STAI T-Anxiety scale two days before the practice game. Immediately before this game, they also responded to a short form of the STAI S-Anxiety scale and the first author’s (1988) adaptation of Martens et al.’s (1990) Competitive State Anxiety Inventory, which included Cognitive and Somatic Anxiety and Self-Confidence subscales. Following the practice game, they responded to an adaptation of Sarason’s (1978) Cognitive Interference Questionnaire. Similar measures were also obtained a week later for the same study participants, before and after they played in a competitive game. As predicted, the results of the first study indicated that the players with high T-Anxiety reported significantly more cognitive interference responses than those with low T-Anxiety. However, the results failed to confirm the expected findings that the high T-Anxiety players would experience greater increases in the STAI and CSAI-2 state anxiety measures. Although only one of the predicted interactions was statistically significant, the trends in the mean scores for the study participants were generally consistent with the findings reported by Schwenkmezger and Laux (1986). In the second study, the relationship of scores on the STAI and CSAI-2 scales with the players’ performance in the low-stress (practice) and highstress (competitive) games was investigated. Quantitative evaluation of game performance was based on Flanagan’s (1954) critical incident technique. In the low-stress condition, a number of significant correlations were found between scores on the anxiety scales and Successful Shootings, Negative Assistances, and Other Negative Activities, which were generally consistent with these hypothesized relationships. A significant negative correlation was also found between Task-Irrelevant Responses and Total Performance in the high-stress condition, indicating that only these responses had a decisive explanatory power in the competitive game. The results were discussed in a broader context of theories of anxiety, motivation (flow-experience), and action control.

RT44367_C003.indd 55

7/13/05 9:29:00 AM

56

F. MAN ET AL.

REFERENCES Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavior change. Psychological Review, 84, pp. 191–215. Beckmann, J. (1987). Hochleistung als folge missgluecketer selbstregulation. [Top performance as a consequence of wrong regulation.] In J. P. Janssen, W. Schlicht, & H. Strang (Eds.), Handlungskontrolle und soziale prozesse in sport (pp. 52–63). Cologne: Bps-Verlag. Borkovec, T. D. (1976). Psychological and cognitive processes in the regulation of anxiety. In G. Schwartz & D. Shapiro (Eds.), Consciousness and selfregulation: Advances in research (Vol. 1, pp. 261–312). New York: Plenum. Csikszentmihalyi, M. (1975). Beyond boredom and anxiety. San Francisco: Jossey-Bass. Csikszentmihalyi, M. (1985). Emergent motivation and the evaluation of the self. In D. A. Kleiber & L. Maehr (Eds.), Advances in motivation and achievement (Vol. 4, pp. 93–119). Greenwich, CT: JAI Press. Csikszentmihalyi, M., & Larson, R. (1987). Validity and reliability of the Experience-Sampling Method. Journal of Nervous and Mental Disease, 175, pp. 526–536. Davidson, R. J., & Schwartz, C. E. (1976). The psychobiology of relaxation and related states: A multiprocess theory. In G. Mostofsly (Ed.), Behavioral control and modification of physiological activity (pp. 399–442). Englewood Cliffs, NJ: Prentice Hall. Deffenbacher, J. L. (1980). Worry and emotionality in test anxiety. In I. G. Sarason (Ed.), Test anxiety: Theory, research, and applications. Hillsdale, NJ: Erlbaum. Ellis, A. (1969). Reason and emotion in psychotherapy. New York: Lyle Stuart. Flanagan, J. C. (1954). The critical incident technique. Psychological Bulletin, 51, pp. 327–358. Froehlich, W. D. (1983). Perspektiven der angstforschung. [Perspectives of anxiety research.] In H. Thomae (Ed.), Psychologie der Motive, 4th ser., Vol. 2, Enzyklopädie der Psychologie (pp. 111–320). Göttingen: Hogrefe. Gould, D., Petlichkoff, L., Simons, J., & Vevera, M. (1987). The relationship between Competitive State Anxiety Inventory-2 subscale scores and pistol shooting performance. Journal of Sport Psychology, 9, pp. 33–42. Gould, D., Petlichkoff, L., & Weinberg, R. S. (1984). Antecedents of temporal changes in and relationship between CSAI-2 subcomponents. Journal of Sport Psychology, 6, pp. 289–304. Hackfort, D., & Schwenkmezger, P. (1980). Angst und angstkontrolle im sport. [Anxiety and anxiety control in sport.] In Sport relevante ansätze theoretischer und empirischer angst forschung. Cologne: Bps-Verlag. Hackfort, D., & Schwenkmezger, P. (1989). Measuring anxiety in sport: Possibilities, problems, and perspectives. In D. Hackfort & C. D. Spielberger (Eds.), Anxiety in sports: An international perspective (pp. 55–76). Washington, DC: Hemisphere.

RT44367_C003.indd 56

7/13/05 9:29:00 AM

TRAIT-STATE ANXIETY, WORRY, AND EMOTIONALITY

57

Hackfort, D., & Spielberger, C. D. (1989). Sport-related anxiety: Current trends in theory and research. In D. Hackfort & C. D. Spielberger (Eds.), Anxiety in sports: An international perspective (pp. 261–268). Washington, DC: Hemisphere. Hanin, Y. L. (1983). STAI in sport: Problems and perspectives. In E. Apitzsch (Ed.), Anxiety in sport (pp. 129–141). Magglingen, Switzerland: ETS (FEPSAC). Hanin, Y. L. (1986). State-trait anxiety research on sports in USSR. In C. D. Spielberger & R. Diaz-Guerrero (Eds.), Cross-cultural anxiety (Vol. 3, pp. 45–64). New York: Hemisphere. Hanin, Y. L. (1989). Interpersonal and intragroup anxiety in sport. In D. Hackfort & C. D. Spielberger (Eds.), Anxiety in sports: An international perspective (pp. 19–28). Washington, DC: Hemisphere. Klavora, P. (1977). An attempt to derive inverted U curves on the relationship between anxiety and athletic performance. In D. M. Landers & R. W. Christina (Eds.), Psychology of motor behavior and sport (pp. 369–377). Champaign, IL: Human Kinetics. Kleine, D. (1990). Anxiety and sport performance: A meta-analysis. Anxiety Research, 2, pp. 113–131. Krohne, H. W. (1980). Angsttheorie: Vom mechanistischen zum kognitiven ansatz. [Anxiety theories: From mechanistic to cognitive approach.] Psychologische Rundschau, 31, pp. 12–29. Krohne, H. W., Schumacher, A., & Neumann, R. (1989). Personality determinants of actual competence expectancies and state anxiety. In R. Schwarzer, H. M. van der Ploeg, & C. D. Spielberger (Eds.), Advances in test anxiety research (Vol. 6, pp. 27–36). Lisse, Netherlands: Swets & Zeitlinger. Kuhl, J. (1983). Motivations, konflikt und handlungskontrolle [Motivation, conflict and action control]. Berlin, Heidelberg, New York: Springer. Kuhl, J. (1985). Volitional mediators of cognition-behavior consistency: Selfregulatory processes and action versus state orientation. In J. Kuhl & J. Beckmann (Eds.), Action control: From cognition to behavior (pp. 101–128). Berlin, Heidelberg, New York: Springer. Lander, D. M., & Boutcher, S. H. (1986). Arousal-performance relationships. In J. M. Williams (Ed.), Applied sport psychology: Personal growth to peak performance (pp. 163–184). Palo Alto, CA: Mayfield. Laux, L. (1983). Psychologische stresskonzeptionen. [Psychological stress conceptions.] In H. Thomae (Ed.), Theorien und formen der motivation, 4th ser., Vol. 1, Enzyklopädie der Psychologie (pp. 453–535). Göttingen: Hogrefe. Liebert, R. M., & Morris, L. W. (1967). Cognitive and emotional components of test anxiety. Psychological Reports, 20, pp. 975–978. Man, F. (1987). A short manual of the Czech version of the sport adaptation of Sarason’s Cognitive Interference Questionnaire. Unpublished report, University of South Bohemia [in Czech]. Man, F. (1988). A preliminary short manual for the CSAI-2 Czech version of the CSAI-2. Unpublished report, University of South Bohemia [in Czech].

RT44367_C003.indd 57

7/13/05 9:29:01 AM

58

F. MAN ET AL.

Man, F., & Hrabal, V. I. (1989). Self-concept of ability, social consequences, anxiety, and attribution as correlates of action control. In F. Halisch & J. van den Bercken (Eds.), International perspectives on achievement and task motivation (pp. 309–316). Amsterdam: Swets & Zeitlinger. Man, F., Stuchlik, L., & Hagtvet, K. A. (1992). The effects of cognitive interference and experimentally manipulated failure on information processing in a complex strategic game. In K. A. Hagtvet & T. Backer Johnsen (Eds.), Advances in test anxiety research (Vol. 7, pp. 332–343). Lisse, The Netherlands: Swets & Zeitlinger. Martens, R. (1977). Sport Competition Anxiety Test. Champaign, IL: Human Kinetics. Martens, R., Burton, D., Vealey, R. S., Bump, L. A., & Smith, D. E. (1990). Competitive State Anxiety Inventory-2. Unpublished manuscript, University of Illinois at Urbana-Champaign. McAuley, E. (1985). State anxiety: Antecedents or results of sport performance. Journal of Sport Behavior, 8, pp. 71–77. Meichenbaum, D. (1977). Cognitive-behavior modification: An integrated approach. New York: Plenum. Morris, L., Davis, D., & Hutchings, C. (1981). Cognitive and emotional components of anxiety: Literature review and revised worry-emotionality scale. Journal of Educational Psychology, 73, pp. 541–555. Muellner, J., Ruisel, I., & Farkas, G. (1980). Czech form of Spielberger’s STAI. Bratislava: Psychodiagnostika. Nideffer, R. M. (1989). Anxiety, attention, and performance in sports. Theoretical and practical considerations. In D. Hackfort & C. D. Spielberger (Eds.), Anxiety in sports: An international perspective (pp. 117–136). Washington, DC: Hemisphere. Rand, P., Lens, W., & Decock, B. (1991). Negative motivation is half the story. Unpublished report, University of Oslo, Norway. Sarason, I. G. (1978). The Test Anxiety Scale: Concept and research. In C. D. Spielberger & I. G. Sarason (Eds.), Stress and anxiety (Vol. 5, pp. 193–216). Washington, DC: Hemisphere. Schwenkmezger, P. (1980). Untersuchungen zur koegnitiven angsttheorie im sportmotorischen bereich (state-trait anxiety). [Research on cognitive theories of anxiety in sport-settings.] Zeitschrift für Experimentalle und Angewandte Psychologie, 27, pp. 607–630. Schwenkmezger, P. (1985). Modelle der eigenschafts und zustandangst. [Model on trait and state anxiety]. Göttingen: Hogrefe. Schwenkmezger, P., & Laux, L. (1986). Trait anxiety, worry, and emotionality in athletic competition. In C. D. Spielberger & R. Diaz-Guerrero (Eds.), Cross-cultural anxiety (Vol. 3, pp. 65–77). New York: Hemisphere. Sonstroem, R. J., & Bernardo, P. (1982). Intraindividual pregame state anxiety and basketball performance: An examination of the inverted U curve. Journal of Sport Psychology, 4, pp. 235–245. Spielberger, C. D. (1966). Theory and research in anxiety. In C. D. Spielberger, (Ed.), Anxiety and behavior (pp. 3–20). New York: Academic Press.

RT44367_C003.indd 58

7/13/05 9:29:01 AM

TRAIT-STATE ANXIETY, WORRY, AND EMOTIONALITY

59

Spielberger, C. D. (1972). Anxiety as an emotional state. In C. D. Spielberger (Ed.), Anxiety: Current trends in theory and research (Vol. 1, pp. 23–49). New York: Academic Press. Spielberger, C. D. (1980). Test Anxiety Inventory: Preliminary professional manual. Palo Alto, CA: Consulting Psychologists Press. Spielberger, C. D. (1983). State-trait anxiety and sports psychology. In E. Apitzsch (Ed.), Anxiety in sport (pp. 74–85). Magglingen, Switzerland: FEPSAC. Spielberger, C. D. (1987). Stress, emotions, and health. In W. P. Morgan & S. E. Goldston (Eds.), Exercise and mental health (pp. 11–16). New York: Hemisphere/Harper & Row. Spielberger, C. D., Anton, W. D., & Bedell, J. (1976). The nature and treatment of test anxiety. In M. Zuckerman & C. D. Spielberger (Eds.), Emotions and anxiety: New concepts, methods, and applications (pp. 315–345). Hillsdale, NJ: Erlbaum. Spielberger, C. D., Gonzalez, H. P., & Fletcher, T. (1979). Test anxiety reduction, learning strategies, and academic performance. In H. F. O’Neil, Jr., & C. D. Spielberger (Eds.), Cognitive and affective learning strategies (pp. 111–131). New York: Academic Press. Spielberger, C. D., Gonzalez, H. P., Taylor, C. J., Algaze, B., & Anton, W. D. (1978). Examination stress and test anxiety. In C. D. Spielberger & I. G. Sarason (Eds.), Stress and anxiety (Vol. 5, pp. 167–191). New York: Hemisphere/Wiley. Spielberger, C. D., Gorsuch, R. L., & Lushene, R. E. (1970). Manual for the State-Trait Anxiety Inventory. Palo Alto, CA: Consulting Psychologists Press. Spielberger, C. D., & Vagg, P. R. (1995). Test anxiety: Theory, assessment and treatment. Washington, DC: Taylor & Francis. Stiensmeier-Pelster, J. (1988). Erlernte hilflosigkeit, handlungskontrolle und leistung. [Learned helplessness, action control and performance]. Berlin, Heidelberg, New York: Springer. Táborský, F. (1976). Hodnoceni vykonu hrace v utkani hazene. [Evaluation of the handball player during competition.] In F. Táborský (Ed.), Otazky soucasne vykonnosti ve sportovnich hrach. Prague: Handball Association. Táborský, F. (1989). Metody kvantitativniho hodnoceni vykonu. [Method of quantitative evaluation of performance]. Prague: Handball Association (Monograph). Wine, J. D. (1971). Test anxiety and direction of attention. Psychological Bulletin, 76, pp. 92–104. Wine, J. D. (1980). Cognitive-attentional theory of test anxiety. In I. G. Sarason (Ed.), Post anxiety: Theory, research, and applications (pp. 319–385). Hillsdale, NJ: Erlbaum. Wine, J. D. (1982). Evaluation anxiety: A cognitive attentional construct. In H. V. Krohne & L. Laux, (Eds.), Achievement, stress and anxiety. Washington, DC: Hemisphere.

RT44367_C003.indd 59

7/13/05 9:29:01 AM

RT44367_C003.indd 60

7/13/05 9:29:01 AM

4 Evaluation of Test Anxiety Using Pictorial Evaluation of Test Reactions (PETER) Joseph Toubiana Zahar Clinic, Petach Tikva, Israel Abstract This study reports cross-cultural research that compares the test anxiety levels of adolescents from eight cultures: Argentina, Czech Republic, Ireland, Israel, Israeli Arabs, South Africa, Turkey, and the United States. The Pictorial Evaluation of TEst Reactions-Trait Questionnaire (PETER-TQ), a self-report questionnaire developed by the author to evaluate test anxiety in students of different ages, and Spielberger’s Test Anxiety Inventory (TAI) were administered to 1,510 students (737 females, 773 males), ages 14–17. Participants first responded to the TAI and then completed the PETER-TQ. The PETER-TQ scores of students in all eight cultures were quite similar, with high internal consistency in all cultures (alpha coefficients of .84 or higher). Although the scores of females were higher than those of males, these gender differences were significant in only two cultures. Correlations of the PETER-TQ with the TAI were moderately high, ranging from .45 to .70; the overall correlation between the two measures was .61. The findings of the present study provide evidence that pictorial measures can be successfully employed to assess individual differences in test anxiety in different cultures.

EVALUATION OF TEST ANXIETY USING PICTORIAL EVALUATION OF TEST REACTIONS (PETER) Test anxiety reveals itself in a variety of physiological, behavioral, and cognitive manifestations, and has been widely studied for the past 40 years as a major problem in technological cultures. The development of self-report questionnaires has stimulated cross-cultural research to evaluate levels of test anxiety reported by students from different countries. Most of these studies have used Spielberger’s (1980) Test Anxiety Inventory (TAI) or Sarason’s (1984) Reaction to Tests (RTT) to measure test anxiety as a situation-specific trait. For example: Schwarzer & Kim (1984) found that students from Korea reported higher levels of test anxiety, as measured by the TAI, than German, Dutch, Hindi, Hungarian, and U.S. students, and El-Zahhar & Hocevar (1991) found higher levels of test anxiety among students in Egypt and Brazil, as compared to U.S. students. In a study of four Asian and five Euro-American cultures, females scored consistently higher in test anxiety than males (Sharma & Sud, 1990). 61

RT44367_C004.indd 61

8/15/05 6:58:44 AM

62

J. TOUBIANA

The use of verbal self-report questionnaires in cross-cultural research may have significant limitations that result from the translation of a test from one language to another. Sperber, Devellis, and Boehlecke (1994, p. 504) noted that “most researchers have used direct translation methods despite their shortcomings for cross-cultural comparisons,” and pointed out that faulty test translations often lead to erroneous conclusions about differences between cultures. Moreover, translation and evaluation methods designed to address these problems, such as performing a back-translation (Brislin, Lonner, & Thorndike, 1973) or using bilingual subjects (Prince & Mamboour, 1967), may also have validity problems (Hulin, 1987; Yang & Bond, 1980). Hocevar and El-Zahhar (1985) suggested a nine-step procedure for analyzing crosscultural translations based on Joreskog & Sorbom’s (1983) LISREL model. In applying this model, they noted: “The first five steps deal with whether a given measure remains valid in the factorial sense after it is translated into another language . . . the remaining four steps deal with cross-cultural differences” (p. 212). The overwhelming majority of instruments used to evaluate test anxiety are based on self-report questionnaires (e.g., Sarason, 1984; Spielberger, 1972, 1980), which assume that respondents are capable of coping with the verbal text. Most of these measures generate information about situation-specific anxiety reactions to stressful situations. A possible solution for coping with the limitations and biases of verbal self-report measures of test anxiety is the use of pictures to elicit the thoughts and feelings that are experienced during tests. As the famous saying goes, “One picture is worth a thousand words,” but it might be added that “one picture may sometimes be worth many translated items in a psychometric scale.” Meaningful visual stimuli, such as pictures, cartoons, or caricatures, can be used as a universal language to bridge between cultures. Klagsburn and Bowlby (1976) developed an instrument based on Hansburg’s Separation Anxiety Questionnaire (1972) that used pictures to evaluate the reactions of young children to separation from their parents. Halpern (1985, 1990) adapted the Pictorial Experimental Paradigm (PEP) from Spielberger’s StateTrait Anxiety Inventory for Children (STAIC; Spielberger, Edwards, Lushene, Montuori, & Platzek, 1973) to assess developmental anxiety in young children. The PEP uses pictures of separation and death anxiety situations to complement verbal stimuli. According to Halpern (1990, p. 273), “Such concrete aids lend existential reality to the situations for which the children’s reactions are assessed, and thus avoid an exclusive reliance on words as a stimulus, such as found in the STAIC.” Markham and Wang (1996) contend that “there is strong evidence for the universality of recognition of at least six emotions” (cited by Ekman, Friesen, & Ellsworth, 1982; Ekman et al., 1987). In their cross-cultural research on the facial expression of emotions, they found that most Grade 3 students correctly recognized different emotions when presented to them via pictures. In his comprehensive review of research on facial expressions, Izard (1994) also concluded that innate and universal facial expressions were associated with basic emotions (e.g., happiness, joy, fear, and so on), which develop

RT44367_C004.indd 62

8/15/05 6:58:45 AM

EVALUATION OF TEST ANXIETY USING PETER

63

in infancy and early childhood before language comprehension. Therefore, it seems reasonable to assume that pictures that reflect emotional reactions to testing situations shared by students around the world can be understood by students from different cultures. Most test anxiety questionnaires focus on individual differences in behavioral manifestations of anxiety in test situations (e.g., Spielberger, 1972). However, being tested can also be viewed as a dynamic process which includes a range of possible experiential reactions. This concept of testing led to the development of the Pictorial Evaluation of TEst Reactions State Questionnaire (SQ) to evaluate the phenomenological experience of state anxiety in test situations. The PETER-SQ uses only pictures to evoke a variety of verbal reactions related to testing situations, such as blanking out, having disturbing thoughts, or behavioral responses that reflect active coping or competitiveness. Respondents to the PETER-SQ report, on a five-point Lickert scale, the extent to which each picture fits what they experienced during the test just completed. The primary goal of the present study was to compare test anxiety in different cultures, using a new trait measure, the PETER-Trait Questionnaire (TQ), to examine cross-cultural and gender differences between these cultures. A second major goal was to evaluate the concurrent validity of the PETER-TQ in different cultures by correlating it with Spielberger’s (1980) TAI. METHOD Subjects and Instruments The subjects in this study were 1,510 students (737 females, 773 males) from eight different countries. The number of male and female participants from each of the eight countries, and the age range and the means and standard deviations for age, are reported in Table 4.1. The students ranged in age from 14 to 19 years. Most of the participants came from middle-class socioeconomic backgrounds and resided in large cities (e.g., New York, Buenos Aires, Tel-Aviv). The PETER-TQ and the TAI were administered in grouptesting sessions to study participants in all eight countries. Pictorial Evaluation of TEst Reactions-Trait Questionnaire. The PETER-TQ contains brief instructions to be followed in responding to the 20 pictures. Subjects respond to each picture by rating, on a five-point Lickert scale, the extent that the picture reflects the respondent’s usual experience in testing situations. A rating of 1 indicates that the picture does not, at all, reflect the subject’s experience. A rating of 5 means that the picture accurately reflects the subject’s experience. The respondent’s PETER-TQ score is the average of the sum of the scores for all 20 pictures. The mean PETER-TQ scores for the female and male students from each of the eight countries who participated in this study are reported in Figure 4.1. Test Anxiety Inventory (TAI). Spielberger’s (1980) TAI was used as the criterion measure to validate the PETER-TQ. The TAI is a self-report questionnaire containing 20 items that are each rated on a four-point frequency scale,

RT44367_C004.indd 63

8/15/05 6:58:46 AM

64

J. TOUBIANA

Table 4.1 Gender and Age Distributions of Study Participants for Eight Countries Country

N

Female

Male

Age Range

M

SD

Argentina Czech Republic Ireland Israel Israeli Arabs South Africa Turkey United States

179 199 181 184 173 195 200 199

103 100 87 54 112 100 91 90

76 99 94 130 61 95 109 109

12–19 13–18 12–19 15–18 13–18 13–19 14–19 13–18

14.26 15.02 14.66 15.35 15.17 17.10 16.10 14.33

2.01 .77 1.98 .83 1.36 1.07 .90 1.86

1,510

737

773

12–19

—

—

Total

60

53.41

51.52

50.90

50.84

50.04

49.12

47.72

MEAN

50

47.39

40

30

20

S. AF

CZE

ARG

IRL

U.S.

ISR

ISR.A

TUR

COUNTRY PETER-TQ’s mean scores in each culture

Figure 4.1. PETER-TQ mean scores for males and females in each culture.

with scores ranging from 1 (Almost never) to 4 (Almost always). Scores on the first TAI item (“I feel confident and relaxed while taking tests”) that reflect a low level of test anxiety are reversed (e.g., an item rating of 4 is scored as 1). Items 2 through 20 are direct scored. The TAI test anxiety score is based on the sum of the responses to all 20 items, with a score range of 20 to 80. The TAI contains two factorially derived subscales, Worry and Emotionality, that evaluate the major components of test anxiety that were identified in previous research (Liebert & Morris, 1967; Morris & Liebert, 1970). The TAI Worry subscale assesses negative thoughts relating to the possibility of failing a test, and adverse reactions that may result from poor test performance (e.g., “During tests I find myself thinking about the consequences of failing”). The

RT44367_C004.indd 64

8/15/05 6:58:47 AM

EVALUATION OF TEST ANXIETY USING PETER

65

Emotionality subscale assesses anxious feelings and tension associated with arousal of the autonomic nervous system, which may be reflected in bodily symptoms (e.g., “During tests I feel very tense”). Widely used in numerous studies to evaluate individual differences in test anxiety, the TAI has been translated and adapted in more than 20 different languages, including Italian (Comunian, 1985), Dutch (van der Ploeg, 1983), Hungarian (Sipos, Sipos, & Spielberger, 1985), Hebrew (Zeidner, Nevo, & Lipschitz, 1988), Arabic (Ahlawat, 1989), Nigerian (Kalu, 1987), Czech (Man & Hosek, 1989), and Chinese (Rocklin & Ren-Min, 1989). The TAI has a high internal consistency reliability (alpha = .90 or higher) and strong predictive validity as indicated by negative correlations with academic performance and achievement (Comunian, 1985; Man & Hosek, 1989; van der Ploeg, 1983; Sipos et al., 1985). The validity of the TAI as an outcome measure has also been demonstrated in “before” and “after” research projects, in which the effectiveness of different therapeutic approaches, such as systematic desensitization and cognitive behavior modification, have been evaluated (Spielberger & Vagg, 1987; Zeidner, Klingman, & Papko, 1988). Previous Research with the PETER-TQ The original version of the PETER-TQ contained 14 pictorial items, which were designed to present different aspects of test reactions that were varied in intensity. The pictures were drawn by Yaakov Farkash (“Ze’ev”), a journalist associated with two Israeli daily newspapers, the Maariv and the Ha’aretz, who received the prestigious Israel Prize in 1993 for his life work as a cartoonist. The pictorial items were presented to two groups of judges (approximately 15 per group) that included educational psychologists and 11th grade students. The judges were asked to rate, on a five-point Lickert scale, the amount of test anxiety associated with each picture. High agreement between the groups of judges provided the basis for selecting those pictures that reflected different test anxiety reactions. In a study by Zeidner, Nevo, & Lipschitz (1988), the PETER-TQ correlated substantially with the Hebrew version of Spielberger’s TAI. The author (1990) administered the PETER-SQ and PETER-TQ to 94 eighth graders, ages 14 and 15. The internal consistency alpha reliability of the PETER-SQ was .82; the alpha for the PETER-TQ was .86. The predictive validity of the PETER-SQ in this study was reflected in a moderate negative correlation of r = –.36 with anagram performance. The long-term stability of the PETER-TQ was demonstrated in a study of a large sample of 7th to 11th graders (N = 350), who were assessed on two occasions, one year apart (Toubiana, Milgram, & Shmueli, 1993). Table 4.2 presents the Pearson product-moment correlations of the PETER-TQ with the Learning Anxiety Scale (LAS), the TAI, and measures of self-control, learning skills, and procrastination (Toubiana et al., 1993). The moderate correlations between the PETER-TQ and the LAS and the TAI, which were similar for boys and girls, provide further evidence of the concurrent validity of the PETER-TQ as a measure of test anxiety. In a study of high school students suffering from test anxiety (N = 20), a correlation of .58 was found between

RT44367_C004.indd 65

8/15/05 6:58:48 AM

66

J. TOUBIANA

Table 4.2 Pearson Correlations between the PETER-TQ and Various Personality Measures Measure

Boys (N = 144)

Girls (N = 178)

LAS –

Learning Anxiety Scale (Toubiana et al., 1993)

.51**

.45**

TAI –

Test Anxiety Inventory (Spielberger, 1980) adapted to Hebrew (Zeidner et al., 1988)

.51**

.48**

SCS –

Self-Control Schedule (Rosenbaum, 1980)

.25**

.10

LSQ –

Learning Skills Questionnaire (Binyamini, 1986)

PLA –

Procrastination of Learning Assignments (Toubiana et al., 1993)

–.11

–.10

.04

.15

**p < .01.

scores on the PETER-TQ and the TAI (Toubiana, 1994). Barda (1990) reported a similar correlation of .54 between the PETER-TQ and TAI scores for 11th graders (N = 30). PROCEDURES For the research reported in this chapter, the original version of the PETER-TQ was extended by adding six additional pictorial items to the first version of the PETER-TQ, which consisted mainly of items measuring anxiety reactions. Of the 20 pictures in the extended PETER-TQ, 17 were related to the test anxiety spectrum; the other 3 pictures reflected active coping reactions during a test situation. Thus, in addition to measuring test anxiety, the extended form of the PETER-TQ also evaluates active coping with tests. The researchers in each country introduced themselves to the samples of about 200 subjects and requested their consent to participate in the study. An “explanatory sheet,” in which the research was described as a cross-cultural study of reactions to tests, was read to the students, who were asked to do their best in responding as sincerely as possible to the test materials. They were also informed that their responses would be anonymous and remain confidential. Each student was then handed the PETER-TQ, with the TAI questionnaire attached on top, and asked to record their age, grade, and gender on the test form. Similar procedures were carried out in each culture. The entire testing process took about 30 to 40 minutes to complete, and usually went smoothly. Most students had no problem in understanding and responding to the PETER-TQ pictures and completing the TAI. The reactions to the pictures

RT44367_C004.indd 66

8/15/05 6:58:49 AM

67

EVALUATION OF TEST ANXIETY USING PETER Table 4.3 PETER-TQ Alpha Coefficients for Eight Countries and Pearson Correlations of the PETER-TQ with the TAI for Each Country Country

Alpha

Pearson Correlations

Argentina Czech Republic Ireland Israel Israeli Arabs South Africa Turkey United States

.82 .71 .90 .85 .78 .84 .83 .83

.61 .46 .70 .65 .48 .49 .73 .71

Total

.84

.61

were generally positive and were “friendly and humanistic,” as was observed by the test administrators in the eight countries. Other researchers have also reported that their students responded to the PETER-TQ with full cooperation; many students commented on the “user-friendliness” of this measure. RESULTS Cronbach alpha coefficients for the PETER-TQ that were calculated for each culture are reported in Table 4.3. Pearson correlation coefficients between the PETER-TQ and Spielberger’s (1980) TAI are also reported in this table. The internal consistency of the PETER-TQ was reasonably high, with alphas ranging between .71 and .90 for the eight countries (median alpha = .83). However, because pictures 12 and 16 were found to have relatively low correlations with the PETER-TQ total score, these pictures were excluded from the analysis of the internal consistencies. The PETER-TQ also correlated substantially with the TAI for all countries, with correlations ranging from r = .46 in the Czech sample to r = .73 in Turkey (median r = .63). Cultural and Sex Differences in Test Anxiety The test anxiety means for females and males in each of the eight countries are reported in Figure 4.2. A culture-by-gender analysis of variance was calculated to evaluate differences in PETER-TQ scores within and between cultures. The main effect for the overall difference between cultures was highly significant [F(7,1450) = 8.09; p < .001]. A significant gender main effect [F(1,1456) = 29.56; p < .001] and an interaction of culture with gender [F(7,1450) = 3.45; p < .001] were also found. The PETER-TQ scores of the South African students were significantly higher than those of the students from Turkey and Israel. The Czech Republic

RT44367_C004.indd 67

8/15/05 6:58:49 AM

68

J. TOUBIANA FEMALE

70 60

54.01 52.78

52.56

50.36 49.58

48.72

50

MEAN

55.57

52.50 50.47

MALE

46.46

50.42 48.69 48.58

45.93

48.05 46.84

40 30 20 S. AF

CZE

ARG

IRL

U.S.

ISR

ISR.A

TUR

COUNTRY Sex differences in PETER-TQ

Figure 4.2. Gender differences in PETER-TQ scores for the eight countries.

students also had higher PETER-TQ scores than students in the Turkish sample. In general, females reported higher levels of test anxiety than males, but differences between the sexes were culturally dependent, as indicated by the significant interaction effect. The PETER-TQ scores of females from Ireland [F(1,180) = 37.29; p < .001] and Argentina [F(1,178) = 5.29; p < .05] were significantly higher than the scores of males from these countries. South African males had higher PETER-TQ scores than males from Ireland, Israel, and Turkey, and the Czech Republic males had higher scores than Ireland males. Females from Ireland and South Africa were higher in test anxiety than Turkish females, whereas U.S. females were significantly lower in test anxiety than Irish females. Discussion The Cronbach alpha for the PETER-TQ seemed to be quite reasonable for pictorial items that, in many cases, might be understood differently even among students from the same culture. In comparison to the internal consistency reliability of the TAI, the pictorial scale was also quite reliable and was consistent across the different cultures that took part in the research project. This consistency was reflected in the small differences between cultures of the alpha coefficients, which ranged from .71 to .90, and were .82 or higher in six of the eight cultures. The correlations of the PETER-TQ with the TAI were moderate to high and statistically significant, ranging from .46 to .73 (see Table 4.3). The overall correlation between these measures across all cultures was .61, validating the PETER-TQ as a reliable measure for evaluating test anxiety among students from different cultures. However, in several cultures (e.g., Czech Republic, Israeli Arabs, South Africa), the correlations with the TAI were lower in comparison

RT44367_C004.indd 68

8/15/05 6:58:51 AM

EVALUATION OF TEST ANXIETY USING PETER

69

with the other cultures, which could be attributed to the students’ difficulty to fully comprehend translated instructions or the meaning of the TAI items and PETER-TQ pictures. In regard to the Israeli Arabs, a Hebrew version of the PETER-TQ, rather than an Arabic translation, was administered. In this case, difficulty in understanding the Hebrew instructions for the TAI and the PETER-TQ might have contributed to the lower correlation, as compared to the correlation obtained for their Jewish Israeli peers. The relatively low correlations between the PETER-TQ and TAI for the South African and Czech Republic students could be attributed to difficulties in fully understanding the meaning of some of the pictures. The overall correlation of .61 between the PETER-TQ and the TAI for the total sample provides strong evidence of concurrent validity of the PETER-TQ as a cross-culturally valid measure for assessing test anxiety. Interestingly, the highest correlations between the two scales of approximately r = .73 was found for all of the English-speaking culture, except South Africa, where English was not the mother tongue. These findings suggested that the Englishspeaking cultures might share common visual perceptions of stress reactions and the experience of calmness. In the U.S. sample, which was comprised of four different ethnic groups (Black, Hispanic, Caucasians, and Jewish students), no differences were found in the reactions to the PETER-TQ pictures, which were sketched as human outlines on white paper. The overall findings regarding cross-cultural reactions to the PETER-TQ support previous research that has examined the universality of understanding facial expressions (Ekman et al., 1982; Ekman et al., 1987; Izard, 1994). The findings of the present study also demonstrated that pictorial/descriptive reactions to stressful test situations are manifested not only in facial expressions, but by the entire body and also by behavior, and that these reactions were shared by the different cultures. Because this study concentrated primarily on Western-oriented cultures, it will be interesting to discover whether the PETER-TQ is valid in Far Eastern cultures (e.g., China, Japan, South Korea). Our future research will compare reactions to the PETER-TQ in Western and Far Eastern cultures. Findings from a study of U.S. high school seniors from Minnesota (Toubiana & Adler, 1997) evaluated the PETER-TQ for assessing general anxiety, depression, and coping by correlating it with other scales that are valid measures of these variables. A different line of research with the PETER-TQ is to investigate the clinical use of this measure when the pictures are individually presented by a clinician to a respondent, who is asked to choose the pictures in a certain way (Toubiana, 1990) and to explain why they were chosen. It will be interesting to discover which themes were dominant in motivating subjects’ reactions across cultures and between cultures, and whether the content analyses of the pictured themes was shared by different cultures. Differences in Anxiety across Cultures Higher levels of test anxiety were found for students from South Africa, the Czech Republic, and Argentina than for students from Turkey and Israel (see Figure 4.2). A possible explanation is that these societies are undergoing

RT44367_C004.indd 69

8/15/05 6:58:51 AM

70

J. TOUBIANA

rapid and crucial economic and social changes that may cause increased stress because of parental expectations for higher achievement, leading to increased “ego-threat” (namely, test anxiety). The relatively high level of test anxiety among the Czech Republic students corresponded to the results of a study by Comunian and Schuller (1995), in which university students in Italy and Slovakia were compared, using El-Zahhar’s (1986) Anxiety and Arousability Inventory (AAI). The Czech students (both sexes) scored significantly higher on the AAI than the Italian students. The relatively low level of anxiety among the Turkish students in the present study was also consistent with results obtained by Oner and Kaymak (1987), using their Turkish adaptation of the TAI. The mean score for Turkish bilingual high school students was relatively low, as compared to other cultures. Gender Differences As hypothesized, females reported higher levels of test anxiety than the males, which was consistent with previous findings for general trait anxiety as measured by Spielberger’s STAI (e.g., Bowler, Rauch, & Schwarzer, 1985). Females from six different ethnic groups (Blacks, Asians, Filipinos, Hispanics, Caucasians, and Indo-Chinese), with an age range from 14 to 18 years, reported higher levels of general trait anxiety (Spielberger, Gorsuch, & Lushene, 1970). Research on test anxiety has revealed a similar trend (e.g., Arnbt, Guly, & McManus, 1986; Comunian & Schuller, 1995; El-Zahhar, 1986; Hembree, 1988; Sharma & Sud, 1990; Sowa & La Fleur, 1986; van der Ploeg, 1982; Zeidner & Safir, 1989). Females generally report higher levels of test anxiety than males. Sharma and Sud (1990) compared test anxiety in four Asian countries (India, Jordan, China, and Korea) with five Euro-American countries (Hungary, Turkey, Italy, Germany, and America), and concluded that greater conflict for females in role expectations was the major cause for their higher anxiety. The results of the present study are clearly consistent with previous findings. The means and SDs of the PETER-TQ for the male and female students from each of the eight countries are reported in Table 4.4. Of the cultures studied in the present research, despite the higher anxiety of the females, only Ireland and Argentina reported significant gender differences. The difference for Ireland was most significant and could be attributed to Ireland’s Catholic society, which upholds distinct sex roles for females, who are allowed to express their anxiety more openly. Whelan, who collected the Irish sample for the present study, explained this gap as resulting from the females’ higher achievement motivation, which was reflected in their outperforming their brothers on State exams (Whelan & Toubiana, 1997). The girls drew on their anxiety in a positive way, and used it “. . . as support for action where appropriate action was called for.” Hannan, Smyth, McCullagh, O’Leary, and McMahon (1996) also observed that the stress levels among girls in Irish schools was significantly and consistently higher than for boys.

RT44367_C004.indd 70

8/15/05 6:58:52 AM

71

EVALUATION OF TEST ANXIETY USING PETER

Table 4.4 Means (M) and Standard Deviations (SD) of the PETER-TQ for Male and Female Study Participants from Eight Countries Males

Females

Total

Country

M

SD

M

SD

M

SD

Argentina Czech Republic Ireland Israel Israeli Arabs South Africa Turkey United States

48.72 50.47 46.46 48.58 45.93 52.78 46.84 49.58

10.75 7.76 10.42 9.75 10.00 10.57 8.80 10.59

52.50 52.56 55.57 50.42 48.69 54.01 48.05 50.36

10.93 7.49 9.60 9.32 10.15 8.74 9.77 11.08

50.90 51.52 50.84 49.12 47.72 53.41 47.39 50.04

10.99 7.68 11.00 9.63 10.15 9.67 9.25 10.86

In a recent study of Irish female students who responded to the PETER-TQ (Whelan, personal communication, 1997), no differences in test anxiety were found for 103 junior-high and high school students from single-sex schools, and 114 same-grade students from coeducational schools. This finding was consistent with previous research (Hannan et al., 1996), indicating that “controlling for the social and ability background of pupils, being in a coeducational school was found to reduce stress levels . . . but did not have the same ameliorative effect for girls” (pp. 193–194). CONCLUSIONS In the present study, the PETER-TQ proved to be a cross-culturally reliable and valid measure for evaluating test anxiety. Differences were found within and between cultures in test anxiety levels. Gender differences were also found within and between cultures. The main implications of the findings in the present study suggested that pictorial stimuli can successfully replace verbal items in self-report questionnaires for assessing test anxiety in different cultures. The PETER-TQ can aid researchers and clinicians in overcoming biases resulting from the process of translating and adapting psychological measures. Furthermore, the PETER-TQ can also facilitate the evaluation of test anxiety and coping for students with reading difficulties. Additional cross-cultural research on the validity of the PETER-TQ needs to be carried out in order to evaluate the suitability for using this measure in clinical settings. Such research would also strengthen the “culture freeness” of the PETER-TQ by identifying pictures that might be irrelevant to a particular culture. Further studies might also focus on the relations between the PETER-TQ and other personality variables (e.g., general trait anxiety, social anxiety, hardiness, locus of control, and introversion/extroversion).

RT44367_C004.indd 71

8/15/05 6:58:53 AM

72

J. TOUBIANA

REFERENCES Ahlawat, K. S. (1989). Psychometric properties of the Yarmouk Test Anxiety Inventory (Y-TAI). In R. Schwarzer, H. M. van der Ploeg, & C. D. Spielberger (Eds.), Advances in test anxiety research (Vol. 6, pp. 263–287). Lisse, The Netherlands: Swets & Zeitlinger. Arnbt, C. B., Guly, U. M., & McManus, I. C. (1986). Preclinical anxiety: The stress associated with vivo vole examinations. Medical Education, 20, pp. 274–280. Barda, A. (1990). Test anxiety [In Hebrew]. Paper in lieu of matriculation examination. Ahad Haam High School, Petach Tikva, Israel. Bowler, R., Rauch, S., & Schwarzer, R. (1985). Anxiety in multiethnic high schools. In H. M. van der Ploeg, R. Schwarzer, & C. D. Spielberger (Eds.), Advances in test anxiety research (Vol. 4, pp. 202–213). Lisse, The Netherlands: Swets & Zeitlinger. Brislin, R., Lonner, W., & Thorndike, R. (1973). Questionnaire wording and translation. In Cross-cultural research methods (pp. 32–58). New York: Wiley. Comunian, A. L. (1985). The development and validation of the Italian form of the Test Anxiety Inventory. In H. M. van der Ploeg, R. Schwarzer, & C. D. Spielberger (Eds.), Advances in test anxiety research (Vol. 4, pp. 215–229). Lisse, The Netherlands: Swets & Zeitlinger. Comunian, A. L, & Schuller, I. S. (1995). Cross-cultural comparison of the Anxiety and Arousability Inventory (AAI). Studia Psychologica, 37, pp. 11–19. Ekman, P., Friesen, W. V., & Ellsworth, P. (1982). What are the similarities and differences in facial behaviour across cultures? In P. Ekman (Ed)., Emotion in the human face (2nd ed., pp. 128–144). Cambridge: Cambridge University Press. Ekman, P., Friesen, W. V., O’Sullivan, M., Diacoayanni-Tarlatzis, I., Krause, R., Pitcairn, T., Scherer, K., Chan, A., Heider, K., Ayan LeCompte, W., Ricci-Bitti, P. E., & Tomita, M. (1987). Universals and cultural differences in the judgments of facial expressions of emotion. Journal of Personality and Social Psychology, 53, pp. 712–717. El-Zahhar, N. (1986). An investigation of the multidimensional nature of trait anxiety. Journal of Educational Sciences, 2, pp. 67–78. El-Zahhar, N., & Hocevar, D. (1991). Cultural and sex differences in test anxiety, trait anxiety, and arousability: Egypt, Brazil, and the USA. Journal of Cross-Cultural Psychology, 22, pp. 238–249. Halpern, E. (1985). Separation and death anxiety differentiated with pictorial paradigms in 4-year-olds. In H. M. van der Ploeg, R. Schwarzer, & C. D. Spielberger (Eds.), Advances in test anxiety research (Vol. 4). Lisse, The Netherlands: Swets & Zetlinger. Halpern, E. (1990). A pictorial experimental paradigm in the assessment of developmental anxieties. In P. J. D. Drenth, J. A. Sergeant, & R. J. Takens (Eds.), European perspectives in psychology (Vol. 2). New York: John Wiley & Sons.

RT44367_C004.indd 72

8/15/05 6:58:54 AM

EVALUATION OF TEST ANXIETY USING PETER

73

Hannan, D. F., Smyth, E., McCullagh, J., O’Leary, R., & McMahon, D. (1996). Pupil stress levels. In Coeducation and gender equality: Exam performance, stress and personal development. Dublin: Oak Tree Press. Hansburg, H. G. (1972). Adolescent separation anxiety: A method for the study of adolescent separation problems. Springfield, IL: C. C. Thomas. Hembree, R. (1988). Correlates, causes, effects, and treatment of test anxiety. Review of Educational Research, 58, pp. 47–77. Hocevar, D., & El-Zahhar, N. E. (1985). Test anxiety in the USA and Egypt: A paradigm for investigating psychometric characteristics across cultures. In H. M. van der Ploeg, R. Schwarzer, & C. D. Spielberger (Eds.), Advances in test anxiety research (Vol. 4, pp. 202–213). Lisse, The Netherlands: Swets & Zeitlinger. Hulin, C. L. (1987). A psychometric theory of evaluation of items and scale translations: Fidelity across languages. Journal of Cross-Cultural Research, 18, pp. 115–142. Izard, C. E. (1994). Innate and universal facial expressions: Evidence from developmental and cross-cultural research. Psychological Bulletin, 115, pp. 288–299. Joreskog, K., & Sorbom, D. (1983). LISREL V: Analysis of linear structural relationships by the method of maximum likelihood. Chicago: International Educational Services. Kalu, W. (1987, June). Test anxiety in Nigerian undergraduate students. Paper presented at the Eighth International Conference of the Society for Anxiety Research, Bergen, Norway. Klagsburn, M., & Bowlby, J. (1976). Responses to separation from parents: A clinical test for young children. Projective Psychology, 21, pp. 7–8. Liebert, R. M., & Morris, L. W. (1967). Cognitive and emotional components of test anxiety: A distinction and some initial data. Psychological Reports, 20, pp. 975–978. Man, F., & Hosek, V. (1989). The development and validation of the Czech form of the Test Anxiety Inventory. In R. Schwarzer, H. M. van der Ploeg, & C. D. Spielberger (Eds.), Advances in test anxiety research (Vol. 6, pp. 233–243). Lisse, The Netherlands: Swets & Zeitlinger. Markham, R., & Wang, L. (1996). Recognition of emotion by Chinese and Australian children. Journal of Cross-Cultural Psychology, 27, pp. 616–643. Morris, L. W., & Liebert, R. M. (1970). Relationship of cognitive and emotional components of test anxiety of physiological arousal and academic performance. Journal of Consulting and Clinical Psychology, 35, pp. 332–337. Oner, N., & Kaymak, D. A. (1987). The transliteral equivalence and the reliability of the Turkish TAI. In R. Schwarzer, H. M. van der Ploeg, & C. D. Spielberger (Eds.), Advances in test anxiety research (Vol. 5, pp. 227–239). Lisse, The Netherlands: Swets & Zeitlinger. Prince, R., & Mamboour, W. (1967). A technique for improving linguistic equivalence in cross-cultural surveys. International Journal of Social Psychiatry, 13, pp. 229–237.

RT44367_C004.indd 73

8/15/05 6:58:55 AM

74

J. TOUBIANA

Rocklin, T., & Ren-Min, Y. (1989). Development and adaptation of the Chinese Test Anxiety Inventory: A research note. In R. Schwarzer, H. M. van der Ploeg, & C. D. Spielberger (Eds.), Advances in test anxiety research (Vol. 6, pp. 245–252). Lisse, The Netherlands: Swets & Zeitlinger. Rosenbaum, M. (1980). A schedule for assessing self-control behavior: Preliminary findings. Behavior Therapy, 11, 109–121. Sarason, I. G. (1984). Stress, anxiety, and cognitive interference: Reaction to tests. Journal of Personality and Social Psychology, 46, pp. 920–938. Schwarzer, C, & Kim, M. J. (1984). Adaptation of the Korean form of the Test Anxiety Inventory: A research note. In H. M. van der Ploeg, R. Schwarzer, & C. D. Spielberger (Eds.), Advances in test anxiety research (Vol. 3, pp. 277–282). Lisse, The Netherlands/Hillsdale, NJ: Swets & Zeitlinger/ Erlbaum. Sharma, S., & Sud, A. (1987, July–December). Examination stress and test anxiety: A cross-cultural perspective. Psychology and Developing Societies 2, pp. 183–201. Sipos, K., Sipos, M., & Spielberger, C. D. (1985). The development and validation of the Hungarian form of the Test Anxiety Inventory. In H. M. van der Ploeg, R. Schwarzer, & C. D. Spielberger (Eds.), Advances in test anxiety research (Vol. 4, pp. 215–220). Lisse, The Netherlands: Swets & Zeitlinger. Sowa, C. J., & La Fleur, N. K. (1986). Gender differences within test anxiety. Journal of Instructional Psychology, 13, pp. 75–80. Sperber, A. D., Devellis, R. F., & Boehlecke, B. (1994). Cross-cultural translation methodology and validation. Journal of Cross-Cultural Psychology, 25, pp. 501–524. Spielberger, C. D. (1972). Anxiety as an emotional state. In C. D. Spielberger (Ed.), Anxiety: Current trends in theory and research (Vol. 1, pp. 23–49). New York: Academic Press.. Spielberger, C. D. (1980). Test Anxiety Inventory: Preliminary professional manual. Palo Alto, CA: Consulting Psychologists Press. Spielberger, C. D., Edwards, C. F., Lushene, R. E., Montuori, K., & Platzek, D. (1973). Preliminary test manual for the State-Trait Anxiety Inventory for Children. Palo Alto, CA: Consulting Psychologists Press. Spielberger, C. D., Gorsuch, R., & Lushene, R. (1970). Manual for the State-Trait Anxiety Inventory. Palo Alto, CA: Consulting Psychologists Press. Spielberger, C. D., & Vagg, P. R. (1987). The treatment of test anxiety: A transactional process model. In R. Schwarzer, H. M. van der Ploeg, & C. D. Spielberger (Eds.), Advances in test anxiety research (Vol. 5, pp. 179–186). Lisse, The Netherlands: Swets & Zeitlinger. Toubiana, J. (1990). Personality variables and processing the experience of anxiety among the test-anxious [In Hebrew]. Doctoral dissertation, Bar-Ilan University, Ramat Gan, Israel. Toubiana, J. (1994). A phenomenological approach to the assessment of test reactions. Unpublished manuscript, Bar-Ilan University, Ramat Gan, Israel. Toubiana, J., & Adler, L. (1997). Validation of PETER-TQ as a measure of general anxiety and depression among schoolchildren. Paper in preparation.

RT44367_C004.indd 74

8/15/05 6:58:55 AM

EVALUATION OF TEST ANXIETY USING PETER

75

Toubiana, J., Milgram, N., & Shmueli, A. (1993). The reciprocal relationship between the development of test anxiety and academic procrastination and the development of coping styles among students: A follow-up study [In Hebrew]. Postdoctoral technical report, Department of Psychology, Tel Aviv University, Israel. van der Ploeg, H. M. (1982). The relationship of worry and emotionality to performance in Dutch schoolchildren. In R. Schwarzer, H. M. van der Ploeg, & C. D. Spielberger (Eds.), Advances in test anxiety research (Vol. 1, pp. 55–68). Lisse, The Netherlands/Hillsdale, NJ: Swets & Zeitlinger/Erlbaum. van der Ploeg, H. M. (1983). The validation of the Dutch form of the test anxiety inventory. In R. Schwarzer, H. M. van der Ploeg, & C. D. Spielberger (Eds.), Advances in test anxiety research (Vol. 2, pp. 191–202). Lisse, The Netherlands/Hillsdale, NJ: Swets & Zeitlinger/Erlbaum. Whelan, A., & Toubiana, J. (1997). Irish female students in coeducation schools versus single schools: A comparison study of reported test anxiety levels, using the TAI and PETER-TQ. Unpublished manuscript. Yang, K. S., & Bond, M. H. (1980). Ethnic affirmation by Chinese bilinguals. Journal of Cross-Cultural Psychology, 11, pp. 411–425. Zeidner, M., Klingman, A., & Papko, O. (1988). Enhancing students’ test coping skills: Report of a psychological health education program. Journal of Educational Psychology, 80, pp. 95–101. Zeidner, M., Nevo, B., & Lipschitz, H. (1988). TAS (Test Anxiety Scale): Tester’s manual [in Hebrew]. Haifa, Israel: Haifa University. Zeidner, M., & Safir, M. (1989). Sex, ethnic, and social differences in test anxiety among Israeli adolescents. Journal of Genetic Psychology, 150, pp. 175–185. ACKNOWLEDGMENTS The author wishes to convey his appreciation and gratitude to the following colleagues who most helpfully collected data for the cross-cultural research in their respective countries (listed in alphabetical order): Dr. Ferda Aysan, Psychology Department, University of Izmir, Turkey Dr. Bonnie Bienstock, New York School District, New York City Dr. Milada Krejci, Pedagogical Faculty, Jihocesska Univerzita, Czech Republic Dr. George Savag, Academic Development Department, The Cape Technion, Cape Town, South Africa Mrs. Selma Sporn, M.A., Educational Psychologist, San Miguel, Argentina Ms. Ann Whelan, M.A., Complete Testing and Assessment Service, Dublin, Ireland Thanks also to Mr. Dov Har-Eben, M.A., of the Psychology Department at Bar-Ilan University, Ramat Gan, Israel, for his assistance in the data analysis and helpful comments on the early version of the manuscript.

RT44367_C004.indd 75

8/15/05 6:58:56 AM

RT44367_C004.indd 76

8/15/05 6:58:57 AM

5 Culture and Gender Factors in the Structure of the Test Anxiety Inventory: A Meta-Analysis Brad J. Bushman University of Michigan, Ann Arbor, MI Peter R. Vagg and Charles D. Spielberger University of South Florida, Tampa, FL Abstract This study reports a meta-analysis of the factor structure of Spielberger’s (1980) Test Anxiety Inventory (TAI), taking culture and gender into account. The 20-item TAI is comprised of factorially derived eight-item subscales for measuring worry and emotionality. Kaiser’s method for evaluating factor similarity was used to compare the TAI factor structures for 37 samples obtained by researchers from 11 different countries with the factor structure of males and females in the original English-language TAI. The effect size metric was the cosine of the angle between matching factors. For the 29 samples in which two distinctive factors were identified, the mean factor cosine was .984, indicating a very high degree of correspondence of the Worry and Emotionality factors obtained by other researchers with the factors for males and females in the original English-language TAI. These two factors were also found to be invariant when the 29 samples were divided into Euro-American and Asian cultural groups. Worry and Emotionality factors similar to those in the English-language TAI were also found in the 8 additional samples in which more than two factors were identified. The results suggest that the factor structure of the TAI is essentially invariant with respect to culture and gender.

CULTURE AND GENDER FACTORS IN THE STRUCTURE OF THE TEST ANXIETY INVENTORY: A META-ANALYSIS Conceptual ambiguity and empirical confusion regarding the nature and measurement of anxiety have resulted from the failure to distinguish between anxiety as an emotional state and individual differences in anxiety as a personality trait. The distinction between states and traits is an old one. In the first century BC, for example, Cicero (106–43 BC) distinguished between anxiety and anger as emotional states, and anxiousness and irascibility as 77

RT44367_C005.indd 77

8/15/05 6:58:02 AM

78

B.J. BUSHMAN ET AL.

personality traits. Spielberger (1966) suggested that emotional states and personality traits were analogous to the concepts of kinetic and potential energy in physics. Like kinetic energy, psychological states describe the expression of energy here and now. Personality traits are like potential energy in that they refer to latent dispositions to respond with certain types of reactions under appropriate circumstances. Psychological states are comprised of the cognitive, emotional, and physiological reactions that occur at a specific moment, and that may fluctuate over time as circumstances change. In contrast to the transitory nature of psychological states, personality traits reflect relatively stable individual differences among people in the tendency or disposition to respond in specifiable ways to particular circumstances or situations. State Anxiety (S-Anxiety) is defined as a transitory emotional reaction “consisting of unpleasant, consciously perceived feelings of tension and apprehension, with associated activation or arousal of the autonomic nervous system” (Spielberger, 1972, p. 29). In keeping with the definition of personality traits, Trait Anxiety (T-Anxiety) is defined in terms of “relatively stable individual differences in anxiety proneness, i.e., differences in the disposition to perceive a wide range of stimulus situations as dangerous or threatening, and in the frequency that a person responds to such situations,” with an increase in the intensity of state anxiety (Spielberger, 1972, p. 39). Trait Anxiety may influence the frequency of occurrence and the level of intensity of State Anxiety in two ways. First, individuals high in T-Anxiety experience more intense elevations of S-Anxiety in a wider range of situations than do people low in this trait, especially in evaluative interpersonal situations that involve threats to self-esteem. Second, people high in T-Anxiety generally experience higher levels of S-Anxiety in neutral or nonstressful situations than those low in T-Anxiety, because worry cognitions stored in memory may become conscious and stimulate thoughts that evoke feelings of apprehension (Spielberger & Vagg, 1995). Another source of confusion in anxiety research is that the terms stress and anxiety are often used interchangeably. In keeping with Lazarus’s (1966) analysis of stress as a complex psychological process, Spielberger (1979) has proposed the following terminological conventions for distinguishing between stress and anxiety: The term stressor . . . describes situations or stimuli that are objectively characterized by some degree of physical or psychological danger. The term threat refers to the individual’s perception or appraisal of that situation or stimulus as potentially dangerous or harmful. People who see a stressful situation as threatening will experience an anxiety reaction. . . . The overall process is referred to as stress. (p. 17)

Test Anxiety, Worry-Emotionality, and Academic Achievement The conceptual distinctions among stress, threat, and anxiety, and between State and Trait Anxiety, are now generally accepted. Within this conceptual framework, examination stress refers to the objective stimulus properties of

RT44367_C005.indd 78

8/15/05 6:58:03 AM

CULTURE AND GENDER FACTORS

79

test situations that are perceived as stressful by most persons, and test anxiety can be defined in terms of “. . . individual differences in anxiety proneness in examination situations” (Spielberger, Gonzalez, Taylor, Algaze, & Anton, 1978, p. 174). During examinations, high test-anxious individuals experience “greater activation or arousal of the autonomic nervous system, and more self-centered worry cognitions and test-irrelevant thoughts that interfere with attention and performance” than do low test-anxious individuals (Spielberger & Vagg, 1995, p. 7). Because high test-anxious individuals respond to examination stress with more intense and more frequent elevations in State Anxiety, test anxiety may be conceptualized as a situation-specific anxiety trait (Spielberger et al., 1978; Spielberger & Vagg, 1995). The construct of test anxiety is ubiquitous and prevalent across nations and cultures (Sharma & Sud, 1990). Test anxiety has been studied extensively because of its relationship to academic achievement and its association with a wide array of unfavorable outcomes, such as poor cognitive performance and painful psychological distress (Hembree, 1988; Seipp, 1991; Sharma & Rao, 1984). In his comprehensive review of theory and research on examination stress and test anxiety, Zeidner (1997) concluded that test anxiety is: . . . one of the key villains in the ongoing drama surrounding psycho-educational testing. Consequently, test anxiety may limit educational or vocational development, as test scores and grades influence entrance to many educational and vocational training programs in modern society. (p. 12)

To explain the adverse effects of test anxiety on academic achievement, Wine (1971, 1980) hypothesized that evaluative situations affect the attention of persons who are high or low in test anxiety in opposite ways. She theorized that high test-anxious individuals focus attention inward, on their thoughts and feelings, whereas low test-anxious individuals focus their attention more directly on the task at hand. For the high test-anxious individual, feeling nervous, experiencing physiological arousal, and worrying about the consequences of failure compete with, and often displace, the cognitive processes needed to comprehend and respond successfully to examination questions. This self-preoccupation diverts the attention of the test-anxious person and thereby contributes to performance decrements. Wine’s (1971, 1980) attentional theory of test anxiety has received substantial empirical support from studies in the United States (e.g., Deffenbacher, 1978, 1980) and in India (Sharma & Sud, 1989, 1990). Liebert and Morris (1967) postulated that test anxiety consists of two major components, worry and emotionality, which they regarded as qualitatively different responses. The intensity of the emotionality elicited by the cues associated with an examination situation (e.g., classroom setting, instructor passing out tests) is highest just prior to and at the beginning of an exam, and tends to dissipate during the testing period as attention is focused on examination questions. Worry responses involve complex cognitive processes that result from current concerns about possible failure and its consequences, as well as previous experiences involving negative feedback and evaluation. The worry cognitions of test-anxious persons that are prevalent at the beginning

RT44367_C005.indd 79

8/15/05 6:58:04 AM

80

B.J. BUSHMAN ET AL.

of an examination may intensify during the testing period, unless cues of possible failure are reduced or removed. In most studies of academic performance, the worry component of test anxiety has been found to be negatively related to achievement, whereas the emotionality component is essentially unrelated to academic performance (see Hembree, 1988; Seipp, 1991). Such evidence demonstrates the critical need to distinguish, conceptually and empirically, between worry and emotionality as the major components of test anxiety. This conceptual distinction has also stimulated the development of objective psychometric measures of test anxiety, with subscales that assess worry and emotionality as its major components. The Test Anxiety Inventory Although there are several good instruments for assessing test anxiety as a global construct, such as I. G. Sarason’s (1978) Test Anxiety Scale (TAS), most of these measures do not provide separate subscales for assessing worry and emotionality. The Test Anxiety Inventory (TAI) was constructed by Spielberger (1980) and his colleagues (Spielberger et al., 1978) to measure stable individual differences in the tendency to experience emotional arousal and worry cognitions in examination situations. The TAI consists of 20 item statements, with instructions essentially the same as those used to measure trait anxiety with the State-Trait Anxiety Inventory (Spielberger, 1983). In responding to each TAI item, subjects rate themselves on a four-point frequency scale, with response choices ranging from 1 (Almost never) to 4 (Almost always). Although developed to assess test anxiety in high-school and college students, the TAI has also proved useful with elementary and junior high-school students. In addition to a Test Anxiety score based on all 20 items, the TAI yields scores on factorially derived eight-item subscales for assessing worry (W) and emotionality (E). The TAI W subscale measures individual differences in how often worry cognitions, such as negative expectations about test performance and concerns about the consequences of failure, are experienced in test situations (e.g., “During tests I find myself thinking about the consequences of failing”). The TAI E subscale assesses individual differences in the unpleasant feelings and autonomic arousal that are experienced during examinations (e.g., “I feel my heart beating very fast during an important test”). Correlations between the TAI W and E subscales vary between .64 and .71 (Spielberger et al., 1978). Alpha coefficients for TAI Total scores are very high, ranging between .92 and .96 (Spielberger, 1980). The median alpha coefficients for the eight-item W and E subscales for large normative samples of high-school and college students were .88 and .90, respectively. The test-retest stability coefficients for TAI Total scores were .81 and .80, respectively, for high-school and college students over time periods of three to four weeks (Spielberger, 1980). The TAI is currently the most widely used measure of test anxiety in the United States and has been translated and adapted in more than 20 languages, including Arabic (Ahlawat, 1989; El-Zahhar & Hocevar, 1991;

RT44367_C005.indd 80

8/15/05 6:58:04 AM

CULTURE AND GENDER FACTORS

81

Hocevar & El-Zahhar, 1985), Chinese (Rocklin & Ren-Min, 1989), Czech (Man, Blahus, & Spielberger, 1990; Man, Budéjovice, & Hosek, 1989), Dutch (van der Ploeg, 1984), German (Hodapp, Laux, & Spielberger, 1982), Hebrew (Zeidner, Nevo, & Lipschitz, 1988), Hindi (Sharma, Sud, & Spielberger, 1983), Hungarian (Sipos, Sipos, & Spielberger, 1984), Italian (Comunian, 1985, 1986), Japanese (Iwawaki & Araki, 1989), Korean (Schwarzer & Kim, 1984), Norwegian (Hagtvet, 1984), Portuguese (Hocevar & El-Zahhar, 1988), Spanish (Bauermeister, Collazo, & Spielberger, 1983), and Turkish (Öner & Kaymak, 1987). The Worry and Emotionality factors identified in adapting the TAI for assessing test anxiety in other languages and cultures were quite similar to the W and E factors in the original English-language TAI. However, the equivalence of the W and E factors obtained in these diverse cultures with the English-language TAI factor structure has not been empirically verified. Virtually all societies administer tests. Even in primitive societies, tests of courage and strength are sometimes used as rites of passage. Can we then conclude that the anxiety experienced in testing situations is similar for all people? Does everyone experience emotional arousal and worry cognitions in testing situations? Is the experience of worry and emotionality similar for people from different cultures? The primary goal of the present study was to collect and analyze empirical evidence of the correspondence of the Worry and Emotionality factors identified in adaptations of the TAI in other languages and cultures with the factors reported for the normative samples of the original English-language TAI. A second goal was to compare the factor structure of the TAI in composite samples of male and female Euro-Americans and Asians with the factor structure of the original English-language TAI. METHOD AND PROCEDURES In the present study, the factor structures for males and females of the original English-language TAI (Spielberger, 1980) were compared with those obtained by other researchers in the United States and in other countries. Confirmatory factor analysis is the best procedure for comparing factor structures from different samples (Jöreskog, 1971). Previous confirmatory factor analytic studies of the TAI have consistently demonstrated that a two-factor structure clearly encompasses the variations in the loadings of the TAI Worry and Emotionality items (Everson, Millsap, & Rodriguez, 1991; Gierl & Rogers, 1996; Hocevar & El Zahhar, 1985; Schwarzer, 1984; Ware, Galassi, & Dew, 1990), and that an oblique factor structure fits the data better than an orthogonal structure (Everson et al., 1991; Gierl & Rogers, 1996; Ware et al., 1990). Previous research has also shown that the factor structure of the TAI is largely invariant with respect to gender (Benson, Moulin-Julian, Schwarzer, Seipp, & El-Zahhar, 1992; Benson & Tippets, 1990; Everson et al., 1991; Gierl & Rogers, 1996; Hedl, 1984; Ware et al., 1990) and culture (Hocevar & El-Zahhar, 1985, 1988). However, a major limitation of confirmatory factor analysis is that it requires data that are usually not readily available, such as the covariance

RT44367_C005.indd 81

8/15/05 6:58:05 AM

82

B.J. BUSHMAN ET AL.

or correlation matrices for different samples. Consequently, the present study used Kaiser’s method for comparing factor structures (Kaiser, Hunka, & Bianchini, 1971), which only requires the factor loadings for each sample, and these are generally reported in research publications. Kaiser’s procedure, described in Appendix A, has been used to compare the factor structures of a number of personality questionnaires, including the Eysenck Personality Questionnaire (Barrett & Eysenck, 1984; Eysenck, Barrett, & Eysenck, 1985; Vagg & Hammond, 1976), the Buss-Durkee Hostility Inventory (Bushman, Cooper, & Lemke, 1991), the Bem Sex-Role Inventory (Thompson, 1989), and the Reducer-Augmenter Scale (Kohn, Hunt, Cowles, & Davis, 1986). Kaiser’s procedure compares two factor structures by calculating the cosine of the angle between matching factors. According to Kaiser et al. (1971), a factor cosine can be interpreted in essentially the same way as a correlation coefficient. The factors in two samples are identical if the cosine between them equals 1.00 (i.e., the angle between them equals 0°); the factors are orthogonal if the cosine between them equals 0 (i.e., the angle between them equals 90°). When two samples both contain two orthogonal factors, there is only one factor cosine, which indicates the degree of match between the factor structures of the samples. When two samples contain a different number of factors, the number of factor cosines equals the product of the number of factors in the two samples. For example, if Sample 1 contains two factors and Sample 2 contains three factors, there are six (2 × 3) factor cosines, i.e., the two factors in Sample 1 are compared to each of the three factors in Sample 2. Cosines of angles between factors in different studies are similar to correlation coefficients, having a range of 0 to 1.00, but they do not measure the covariation of the same subjects on the measures under study. Because the sampling distribution of cosines is unknown, one cannot determine whether the relation between the two factors in the two samples is statistically significant. Because factor cosines refer to geometric relations between factors, Vagg and Hammond (1976) suggest using the following geometric criteria to describe the relation between factors: identical or invariant (±5º; cosine = .985), close similarity (±10º; cosine = .940), and similar (±15º; cosine = .866). Literature Search Procedures and Criteria for Inclusion of Studies In the present study, both formal and informal channels were used to search the literature. First, the contents and reference lists of Advances in Test Anxiety Research (Vols. 1–7) and Cross-Cultural Anxiety (Vols. 1–4) were reviewed; 12 factor-analytic studies were obtained from these sources. Second, the following four on-line computer databases were searched for the years 1978 (the year Spielberger and his colleagues developed the TAI) through 1996: Psychological Abstracts, Dissertation Abstracts International, Educational Resources Information Center (ERIC), and the Social Science Citation Index. All studies with descriptors that included “Test Anxiety Inventory” or “TAI” in the title or abstract, or that cited the TAI Manual (Spielberger, 1980), were retrieved. Eleven factor-analytic studies were obtained from these four on-line computer databases. Finally, reprints or preprints of current papers were requested

RT44367_C005.indd 82

8/15/05 6:58:06 AM

CULTURE AND GENDER FACTORS

83

from colleagues who had previously published research on the TAI or on test anxiety. Four unpublished studies were obtained from those who responded. Of the 27 TAI factor-analytic studies that were located, 15 met the inclusion criteria recommended by Thompson (1989). First, the factor analyses in these studies had to include all 20 TAI items, and only these items. Authors of studies that factored the TAI items along with items from other scales were asked to supply item loadings when only the TAI items were factored. Second, only studies using principal-components factor analysis with orthogonal (e.g., varimax) rotation methods were considered. Authors who had used other methods (e.g., confirmatory factor analysis) were requested to reanalyze their TAI data, using principal-components analyses with varimax rotation. Third, factor loadings for all 20 TAI items had to be reported; those authors who had not reported nonsalient item loadings were asked to supply the missing loadings. Fifteen of the 27 studies met all three criteria. These studies included 37 samples from 11 different countries; the total number of subjects in these samples was 13,042. RESULTS The initial step in the data analysis was to compare the factor structure of the males with that of the females in the original English-language TAI (Spielberger, 1980). Using Kaiser et al.’s (1971) method, the mean cosine of the 20 individual TAI item cosines obtained in this analysis was .996, demonstrating that the relations among the item variables in the original English-language sample were essentially identical for males and females. The cosine of the angle between the corresponding Worry and Emotionality factor axes for the male and female samples of the original English-language TAI was .999, indicating that the factor structures in these samples were virtually invariant. The second step in the data analysis was to compare the factor structures for the 29 samples in which two factors were identified, with the corresponding factors for males and females in the original English-language TAI. Factor cosines for each of the 29 samples are reported in Table 5.1, along with the mean item cosines (shown in parentheses). The mean item cosine across all 29 samples, weighted by sample size, was .938, which is sufficiently large to allow a meaningful interpretation of the factor cosine (Gorsuch, 1983). The weighted mean factor cosine across all samples was .984, indicating a very high degree of correspondence between factors obtained by other researchers, with the Worry and Emotionality factors identified for males and females in the original English-language TAI (Spielberger, 1980). Using the conventional values proposed by Vagg and Hammond (1976) for interpreting factor cosines, a mean factor cosine of .984 indicates that the Worry and Emotionality factors obtained in the meta-analysis of the 29 samples were almost identical (invariant) with those of the original TAI. Comparisons of the factor structures for the eight samples in which more than two factors were identified with the corresponding factors for males and females in the original English-language TAI are reported in Table 5.2. Except

RT44367_C005.indd 83

8/15/05 6:58:07 AM

84

B.J. BUSHMAN ET AL.

Table 5.1 Factor Cosines between the Spielberger (1980) Male and Female Structures and Other Structures with Two Factors Spielberger (1980) samples Country

Source

Brazil

Hocevar & El-Zahhar (1988)

China

Rocklin & Ren-Min (1989)

Czech

Man et al. (1989)

Egypt

Benson et al. (1992) Hocevar & El-Zahhar (1988)

Germany

Benson et al. (1992)

Italy

Comunian (1986)

Puerto Rico

Bauermeister et al. (1983)

Saudi Arabia Hocevar & El-Zahhar (1988) South Korea C. Schwarzer & Kim (1984) USA McAuliffe (1985) Benson et al. (1992) Hedl (1984) Hocevar & El-Zahhar (1988)

Gender Male Female Male Female Male Female Male Female Male Female Male Female Male Female Male Female Male Female Male Female Male Bothc Bothd Male Female Male Female Male Female

Age

n

Child Child Botha Bothb Child Child Adult Adult Child Child Adult Adult Adult Adult Child Child Child Child Adult Adult Child Child Adult Adult Adult Adult Adult Child Child

109 125 267 268 73 81 83 146 114 163 40 121 100 100 100 100 100 100 305 300 318 670 138 431 677 219 324 73 68

Male 0.986 0.998 0.994 0.858 1.000 0.990 0.924 0.887 0.985 1.000 0.986 0.946 0.978 1.000 0.983 0.995 0.991 0.997 1.000 0.994 0.962 0.994 0.996 0.998 0.997 0.998 1.000 0.996 1.000

(0.912) (0.938) (0.874) (0.821) (0.861) (0.863) (0.876) (0.886) (0.937) (0.972) (0.948) (0.928) (0.898) (0.860) (0.935) (0.925) (0.900) (0.969) (0.974) (0.981) (0.877) (0.943) (0.984) (0.978) (0.987) (0.947) (0.991) (0.938) (0.947)

Female 0.982 0.999 0.982 0.998 0.999

(0.921) (0.940) (0.863) (0.844) (0.865)

0.931 0.893 0.988 1.000 0.989 0.954 0.974 1.000 0.980 0.997 0.717 0.999 1.000 0.992 0.994 0.993 0.998 0.999 0.997 0.997 0.999 0.995 1.000

(0.892) (0.856) (0.941) (0.967) (0.943) (0.921) (0.892) (0.885) (0.934) (0.928) (0.801) (0.965) (0.978) (0.981) (0.883) (0.947) (0.990) (0.979) (0.986) (0.941) (0.995) (0.952) (0.951)

Note: Values in parentheses are the mean cosines for the 20 TAI items; no solution could be obtained for the Man et al. Female versus Spielberger female comparison. a = 100 high-school students and 167 college students. b = 105 high-school students and 163 college students. c = 367 males and 389 females. d = Mostly females.

for Everson et al. (1991), for whom the sample sizes were very small, the Emotionality or Worry factors of the original English-language TAI always matched very closely with at least one factor in each of the other samples. For the two largest samples, each with more than 2,000 subjects (Mueller, 1990), the match for both the Worry and Emotionality factors with the original English-language TAI factors was almost identical.

RT44367_C005.indd 84

8/15/05 6:58:08 AM

85

CULTURE AND GENDER FACTORS Table 5.2 Comparisons between the Spielberger (1980) Male and Female Structures and Other Structures with More Than Two Factors

Spielberger (1980) Samples

Source

EigenValue

Factor

835

6.320 1.216 1.065 1.026

I II III IV

Adult

609

7.119 1.319 1.096

I II III

Adult

62

7.102 1.602 1.524 1.358 1.181 1.100

I II III IV V VI

Adult

116

8.393 1.359 1.275 1.083

I II III IV

Adult

2173

8.316 1.371 1.056

I II III

Female

Adult

2786

9.407 1.337 0.992

I II III

Male

Adult

274

9.070 1.336 1.093 1.023

I II III IV

Female

Adult

307

9.694 1.776 1.027

I II III

Gender

Öner & Male Kaymak (1987) Female

Everson Male et al. (1991)

Female

Mueller Male (1990)

Piatka (1987)

Males Factor

Age

n

Adult

E

Females Factor W

–0.165 0.980 0.715 0.152 0.657 0.102 0.172 –0.079 (0.804) –0.074 0.957 0.479 0.277 0.877 –0.071 (0.847) 0.737 0.159 0.492 0.169 0.311 0.186 –0.106 0.592 0.132 0.173 –0.299 0.729 (0.749) 0.896 0.069 –0.115 0.801 0.050 0.594 0.425 0.003 (0.879) 0.990 0.071 –0.061 0.994 –0.125 0.082 (0.978) 0.997 0.057 –0.049 0.991 –0.062 0.122 (0.962) 0.996 0.038 –0.005 0.800 –0.038 0.583 –0.087 0.141 (0.925) 0.999 0.041 –0.040 0.998 –0.028 0.049 (0.962)

E

W

–0.131 0.967 0.818 0.039 0.521 0.227 0.205 –0.114 (0.794) 0.052 0.849 0.255 0.494 0.968 –0.176 (0.846) 0.760 0.122 0.527 0.122 0.271 0.224 –0.137 0.637 0.072 0.248 –0.219 0.673 (0.751) 0.941 0.001 –0.068 0.775 0.031 0.624 0.331 0.099 (0.876) 0.991 0.047 –0.035 0.994 –0.127 0.095 (0.978) 0.997 0.034 –0.022 0.988 –0.079 0.149 (0.981) 0.996 0.015 –0.019 0.833 0.024 0.533 –0.084 0.147 (0.928) 0.999 0.018 –0.015 0.997 –0.047 0.076 (0.965)

Note: Values in parentheses are the mean cosines for the 20 TAI items. For the Spielberger et al. (1978) samples E = Emotionality and W = Worry. Adult = at least 18 years old; child = less than 18 years old. The Öner & Kaymak (1987) samples are from Turkey; the Everson et al. (1991), Mueller (1990), and Piatka (1987) samples are from the USA.

RT44367_C005.indd 85

8/15/05 6:58:09 AM

86

B.J. BUSHMAN ET AL.

Table 5.3 Mean Factor Cosines for the TAI Worry and Emotionality Factors as a Function of Culture and Gender1 Culture Original TAI

Gender

Asian

Euro-American

Male

Female

Male Sample Mean n Samples

.960 2,029 8

.994 3,604 21

.987 2,332 14

.974 2,573 13

Female Sample Mean n Samples

.983 2,029 8

.987 3,603 20

.978 2,332 14

.990 2,492 12

1

The mean factor cosines are based on the 29 samples reported in Table 5.1. The EuroAmerican category included samples from Brazil, the former Czechoslovakia, Germany, Italy, Puerto Rico, and the United States. The Asian category included samples from China, Egypt, Korea, and Saudi Arabia. In the culture and gender analyses for females, the Czech female sample was excluded because the factor cosine was not available. In the gender analysis, the two samples that included both males and females were excluded.

Invariance of the Factor Structure of the TAI across Culture and Gender Weighted analyses of variance (ANOVAs) of the factor cosines for the 29 samples reported in Table 5.1 were used to examine the stability of the factors across culture and gender. Because of the small number of samples, only main effects were included in the model. For the gender variable, the samples were assigned to male and female groups. For the culture variable, the samples were assigned to two culture groups, Euro-American and Asian, as defined by Sharma and Sud (1990). The Euro-American category included samples from Brazil, the former Czechoslovakia, Germany, Italy, Puerto Rico, and the United States. The Asian category included samples from China, Egypt, Korea, and Saudi Arabia. The mean factor cosines obtained in the analyses for culture and gender, as compared with the factors for males and females in the original Englishlanguage TAI, are reported in Table 5.3. The values of these cosines were uniformly high across both gender and culture. All eight cosines substantially exceeded Vagg & Hammond’s (1976) criteria for close similarity (.940), and four of the eight cosines satisfied the criteria for identical factors (.985). As might be expected, the cosines for the Euro-Americans were somewhat higher than those of the Asians, and the cosines for same-gender samples were slightly higher than those for the comparisons of males with females. However, no significant main effects were found for either culture or gender. In essence, the findings of this study demonstrate a close correspondence between the factors obtained for Euro-American and Asian males and females and the corresponding Worry and Emotionality factors of the original English-language TAI (Spielberger, 1980).

RT44367_C005.indd 86

8/15/05 6:58:10 AM

CULTURE AND GENDER FACTORS

87

DISCUSSION AND CONCLUSIONS Kaiser et al.’s (1971) method was used in the present study to compare the factors obtained by researchers from a number of different countries and cultures with the Emotionality and Worry factors identified by Spielberger (1980) for the normative samples of male and female subjects in the original English-language TAI. For the 29 samples in those TAI factor structure studies in which Worry and Emotionality factors were identified, the mean factor cosine was .984. These findings indicated very close similarity, approaching invariance, for the Worry and Emotionality factors identified in these studies as related to the corresponding factors in the original English-language TAI. The factor cosines for the corresponding Worry and Emotionality factors were more stable for those samples that were comprised of larger numbers of subjects, as may be noted in Tables 5.1 and 5.2, especially for those studies that reported more than two factors. For example, in the Mueller (1990) study, in which the samples were very large, the Emotionality and Worry factors both matched very closely with the corresponding factors in the English-language TAI, as can be noted in Table 5.2. In those studies with small samples, in which more than two factors were extracted, the smaller cosines appear to reflect a fracturing of the W and E factors due to overextraction. Most of the subjects comprising the 37 samples in this meta-analytic study were high-school or college students. However, several samples consisted of elementary-school students or adults who were employed as teachers. Given the heterogeneous characteristics of the 37 samples in terms of age and culture, the stability of the TAI Worry and Emotionality factors across both gender and culture was truly remarkable. Moreover, the TAI factor structures for EuroAmerican and Asian males and females from these widely differing cultures were also nearly identical. Thus, the findings of this study provide impressive evidence of the stability, across culture and gender, of Worry and Emotionality as major components of test anxiety as a situation-specific personality trait.

REFERENCES Ahlawat, K. S. (1989). Psychometric properties of the Yarmouk Test Anxiety Inventory Y-TAI. In R. Schwarzer, H. M. van der Ploeg, & C. D. Spielberger (Eds.), Advances in test anxiety research (Vol. 6, pp. 263–278). Lisse, The Netherlands: Swets & Zeitlinger. Barrett, P., & Eysenck, S. (1984). The assessment of personality factors across 25 countries. Personality and Individual Differences, 5, pp. 615–632. Bauermeister, J. J., Collazo, J. A., & Spielberger, C. D. (1983). The construction and validation of the Spanish form of the Test Anxiety Inventory: Inventario de Auto-Evaluación sobre Exámenes (IDASE) In C. D. Spielberger & R. Diaz-Guerrero (Eds.), Cross-cultural anxiety (Vol. 2, pp. 67–85). New York: Hemisphere/McGraw-Hill.

RT44367_C005.indd 87

8/15/05 6:58:11 AM

88

B.J. BUSHMAN ET AL.

Benson, J., Moulin-Julian, M., Schwarzer, C., Seipp, B., & El-Zahhar, N. (1992). Cross-validation of a revised test anxiety scale using multinational samples. In K. Hagtvet (Ed.), Advances in Test Anxiety Research (Vol. 7, pp. 62–102). Lisse, The Netherlands: Swets & Zeitlinger. Benson, J., & Tippets, E. (1990). Confirmatory factor analysis of the Test Anxiety Inventory. In C. D. Spielberger, R. Diaz-Guerrero, & J. Strelau (Eds.), Cross-cultural anxiety (Vol. 4, pp. 149–156). New York: Hemisphere/ McGraw-Hill. Bushman, B. J., Cooper, H. M., & Lemke, K. M. (1991). Meta-analysis of factor analyses: An illustration using the Buss-Durkee Hostility Inventory. Personality and Social Psychology Bulletin, 17, pp. 344–349. Comunian, A. L. (1985). The development and validation of the Italian form of the Test Anxiety Inventory. In H. M. van der Ploeg, R. Schwarzer, & C. D. Spielberger (Eds.), Advances in test anxiety research (Vol. 4, pp. 215–220). Lisse, The Netherlands: Swets & Zeitlinger. Comunian, A. L. (1986). A factor analytic study of the Italian Test Anxiety Inventory. University of Padua, Institute of Psychology, Report 120. Deffenbacher, J. L. (1978). Worry, emotionality, and task-generated interference in test anxiety: An empirical test of attentional theory. Journal of Educational Psychology, 70, pp. 248–256. Deffenbacher, J. L. (1980). Worry and emotionality in test anxiety. In I. G. Sarason (Ed.), Test anxiety: Theory, research, and applications (pp. 111–128). Hillsdale, NJ: Erlbaum. El-Zahhar, N. E., & Hocevar, D. (1991). Cultural and sexual differences in test anxiety, trait anxiety, and arousability. Journal of Cross-Cultural Psychology, 22, pp. 238–249. Everson, H. T., Millsap, R. E., & Rodriguez, C. M. (1991). Isolating gender differences in test anxiety: A confirmatory factor analysis of the Test Anxiety Inventory. Educational and Psychological Measurement, 51, pp. 243–251. Eysenck, H. J., Barrett, P., & Eysenck, S. B. G. (1985). Indices of factor comparison for homologous and nonhomologous personality scales in 24 different countries. Personality and Individual Differences, 6, pp. 503–504. Gierl, M. J., & Rogers, W. T. (1996). A confirmatory factor analysis of the Test Anxiety Inventory using Canadian high school students. Educational and Psychological Measurement, 56, pp. 315–324. Gorsuch, R. L. (1983). Factor analysis. Hillsdale, NJ: Erlbaum. Grossman, S. I. (1987). Elementary linear algebra (3rd ed.). Belmont, CA: Wadsworth. Hagtvet, K. A. (1984). A Norwegian adaptation of the Test Anxiety Inventory: A first try-out. International Review of Applied Psychology, 33, pp. 257–265. Hedl, J. J., Jr. (1984). A factor analytic study of the Test Anxiety Inventory. International Review of Applied Psychology, 33, pp. 267–283. Hembree, R. (1988). Correlates, causes, effects, and treatment of test anxiety. Review of Educational Research, 58, pp. 47–77.

RT44367_C005.indd 88

8/15/05 6:58:11 AM

CULTURE AND GENDER FACTORS

89

Hocevar, D., & El-Zahhar, N. E. (1985). Test anxiety in the USA and Egypt: A paradigm for investigating psychometric characteristics across cultures. In H. M. van der Ploeg, R. Schwarzer, & C. D. Spielberger (Eds.), Advances in test anxiety research (Vol. 4, pp. 203–213). Lisse, The Netherlands: Swets & Zeitlinger. Hocevar, D., & El-Zahhar, N. E. (1988). Arousability, trait anxiety, and the worry and emotionality components of test anxiety. Anxiety Research, 1, pp. 99–113. Hodapp, V., Laux, L., & Spielberger, C. D. (1982). Theorie und messung der emotionalen und kognitiven komponente der prüfungsangst. Zeitschrift für Differentielle und Diagnostiche Psychologie, 2, pp. 169–184. Iwawaki, S., & Araki, N. (1989). The development and validation of the Japanese form of the Test Anxiety Inventory. Paper presented at the 10th International Conference of the Society for Test Anxiety Research, Amsterdam, The Netherlands, 30 June–2 July. Jöreskog, K. G. (1971). Simultaneous factor analysis in several populations. Psychometrika, 36, pp. 409–426. Kaiser, H. F., Hunka, S., & Bianchini, J. C. (1971). Relating factors between studies based upon different individuals. Multivariate Behavioral Research, 6, pp. 409–422. Kohn, P. M., Hunt, R. W., Cowles, M. P., & Davis, C. (1986). Factor structure and construct validity of the Vando Reducer-Augmenter Scale. Personality and Individual Differences, 7, pp. 57–64. Lazarus, R. S. (1966). Psycholgical stress and coping process. New York: McGraw-Hill. Liebert, R. M., & Morris, L. W. (1967). Cognitive and emotional components of test anxiety: A distinction and some initial data. Psychological Reports, 20, pp. 975–980. Man, F., Blahus, P., & Spielberger, C. D. (1990). The relationship of test anxiety to intelligence and academic performance. In C. D. Spielberger & R. Diaz-Guerrero (Eds.), Cross-cultural anxiety (Vol. 4, pp. 183–191). New York: Hemisphere/McGraw Hill. Man, F., Budéjovice, C., & Hosek, V. (1989). The development and validation of the Czech form of the Test Anxiety Inventory. In R. Schwarzer, H. M. van der Ploeg, & C. D. Spielberger (Eds.), Advances in test anxiety research (Vol. 6, pp. 233–243). Lisse, The Netherlands: Swets & Zeitlinger. McAuliffee, E. A. (1985). An investigation of preservice elementary teachers mathematics anxiety and its relationship to their state-trait anxiety, attitudes towards mathematics, attitudes toward reading, test anxiety, and selected organic variables. Dissertation Abstracts International, 46/06A, 1501. (University Microfilms No. AAD85–16064.) Mueller, J. (1990). A factor analytic study of the Test Anxiety Inventory. Unpublished manuscript, University of Missouri-Columbia. Öner, N., & Kaymak, D. A. (1987). The transliteral equivalence and the reliability of the Turkish TAI. In R. Schwarzer, H. M. van der Ploeg, & C. D. Spielberger (Eds.), Advances in test anxiety research (Vol. 4, pp. 227–239). Lisse, The Netherlands: Swets & Zeitlinger.

RT44367_C005.indd 89

8/15/05 6:58:12 AM

90

B.J. BUSHMAN ET AL.

Piatka, A. (1987). The conceptual and methodological distinction between state and trait anxiety: Construct validation of the State-Trait Anxiety Inventory. Dissertation Abstracts International, 49, 05B. (University Microfilms No. AAD87–29279.) Rocklin, T., & Ren-Min, Y. (1989). Development and adaptation of the Chinese Test Anxiety Inventory: A research note. In R. Schwarzer, H. M. van der Ploeg, & C. D. Spielberger (Eds.), Advances in test anxiety research (Vol. 6, pp. 245–252). Lisse, The Netherlands: Swets & Zeitlinger. Sarason, I. G. (1978). The test anxiety scale: Concept and research. In C. D. Spielberger & I. G. Sarason (Eds.), Stress and anxiety (Vol. 5, pp. 193–216). Washington, D.C.: Hemisphere. Schwarzer, C., & Kim, M. (1984). Worry and emotionality as separate components in test anxiety. International Review of Applied Psychology, 33, pp. 205–220. Schwarzer, R. (1984). Adaptation of the Korean form of the Test Anxiety Inventory: A research note. In H. M. van der Ploeg, R. Schwarzer, & C. D. Spielberger (Eds.), Advances in test anxiety research (Vol. 3, pp. 277–285). Lisse, The Netherlands: Swets & Zeitlinger. Seipp, B. (1991). Anxiety and academic performances: A meta-analysis of findings. Anxiety Research, 4, pp. 27–41. Sharma, S., & Rao, U. (1984). Test anxiety research in India: A review. In H. M. van der Ploeg, R. Schwarzer, & C. D. Spielberger (Eds.), Advances in test anxiety research (Vol. 4, pp. 255–264). Lisse, The Netherlands: Swets & Zeitlinger. Sharma, S., & Sud, S. (1989). Cognitive attentional theory of test anxiety: An empirical test in Indian settings. In R. Schwarzer, H. M. van der Ploeg, & C. D. Spielberger (Eds.), Advances in test anxiety research. (Vol. 6, pp. 53–60). Lisse. The Netherlands: Swets & Zeitlinger. Sharma, S., & Sud, A. (1990). Examination stress and test anxiety: A crosscultural perspective. Psychology and Developing Societies, 2, pp. 183–201. Sharma, S., Sud, A., & Spielberger, C. D. (1983). Development of the Hindi form of the Test Anxiety Inventory. In H. M. van der Ploeg, R. Schwarzer, & C. D. Spielberger (Eds.), Advances in test anxiety research (Vol. 2, pp. 183–189). Lisse, The Netherlands: Swets & Zeitlinger. Sipos, K., Sipos, M., & Spielberger, C. D. (1984). First results with the Hungarian Test Anxiety Inventory. In C. D. Spielberger & R. Diaz-Guerrero (Eds.), Crosscultural anxiety (Vol. 3, pp. 37–44). New York: Hemisphere/McGraw-Hill. Spielberger, C. D. (1966). Theory and research on anxiety. In C. D. Spielberger (Ed.), Anxiety and behavior (pp. 3–20). New York: Academic Press. Spielberger, C. D. (1972). Anxiety as an emotional state. In C. D. Spielberger (Ed.), Anxiety: Current trends in theory and research (Vol. 1, pp. 23–49). New York: Academic Press. Spielberger, C. D. (1979). Understanding stress and anxiety. New York: Harper & Row. Spielberger, C. D. (1980). Manual for the Test Anxiety Inventory. Palo Alto, CA: Consulting Psychologists Press. Spielberger, C. D. (1983). Manual for the State-Trait Anxiety Inventory (Form Y). Palo Alto, CA: Consulting Psychologists Press.

RT44367_C005.indd 90

8/15/05 6:58:13 AM

CULTURE AND GENDER FACTORS

91

Spielberger, C. D., Gonzales, H. P., Taylor, C. J., Algaze, B., & Anton, W. D. (1978). Examination stress and test anxiety. In C. D. Spielberger & I. G. Sarason (Eds.), Stress and anxiety (Vol. 5, pp. 167–191). New York: Hemisphere/Wiley. Spielberger, C. D., & Vagg, P. R. (1995). Test anxiety: Theory, assessment, and treatment. Washington, DC: Taylor & Francis. Thompson, B. (1989). Meta-analysis of factor structure studies: A case study example with Bem’s androgyny measure. Journal of Experimental Education, 57, pp. 187–197. Vagg, P. R., & Hammond, S. B. (1976). The number and kind of invariant (Q) factors: A partial replication of Eysenck and Eysenck. British Journal of Social and Clinical Psychology, 15, pp. 121–129. van der Ploeg, H. M. (1984). The development and validation of the Dutch form of the Test Anxiety Inventory. International Review of Applied Psychology, 33, pp. 243–255. Veldman, D. J. (1967). FORTRAN programming for the behavioral sciences. New York: Holt, Rinehart, and Winston. Ware, W. B., Galassi, J. P., & Dew, K. M. H. (1990). The Test Anxiety Inventory: A confirmatory factor analysis. Anxiety Research, 3, pp. 205–212. Wine, J. D. (1971). Test anxiety and direction of attention. Psychological Bulletin, 76, pp. 92–106. Wine, J. D. (1980). Cognitive-attentional theory of test anxiety. In I. G. Sarason (Ed.), Test anxiety: Theory, research, and applications (pp. 349–385). Hillsdale, N.J.: Erlbaum. Zeidner, M. (1997). Test anxiety: The state of the art. New York: Plenum Press. Zeidner, M., Nevo, B., & Lipschitz, H. (1988). The Hebrew version of the Test-Anxiety Inventory. Unpublished paper, University of Haifa, Israel. APPENDIX A A GEOMETRICAL DESCRIPTION OF KAISER’S METHOD FOR COMPARING FACTOR STRUCTURES In this description of Kaiser’s method (Kaiser et al., 1971), variables (e.g., questionnaire items) and factors are represented as vectors (for a good elementary text on linear algebra see Grossman, 1987). A vector is a quantity that has both a magnitude (length) and direction. For example, in the two-dimensional Cartesian coordinate system, the vector a is given by (a1, a2), where a1 and a2 are the x- and y-components, respectively. If the vector a is placed at the origin, the terminal point of a coincides with a point in the xy-plane with coordinates (a1, a2), as shown in Figure 5.1. The length of a is given by ||a|| = √a12 + a22. The direction of a nonzero vector a is defined to be the angle θ, measured in radians, that the vector makes with the positive x-axis. The sine, cosine, and tangent trigonometric functions are defined as sin θ = a2 / ||a||, cos θ = a1 / ||a||, and tan θ = a2 /a2||a||, respectively. Thus, one equation for the direction of a is θ = arctan (a1/a2), where arctan(.) is the inverse tangent function.

RT44367_C005.indd 91

8/15/05 6:58:14 AM

92

B.J. BUSHMAN ET AL.

y a = (a1, a2) 2

|= |a

2,+ a1 √

a2

a2

θ 0 Figure 5.1

a1

x

Vector a in the two-dimensional Cartesian coordinate system.

Kaiser’s method compares factor structures from two samples of subjects. In this description, it is assumed that the two factor analyses were performed on the same set of variables and that the factor vectors are orthogonal.1 Kaiser’s method involves three steps. First, the variable vectors from both samples are transformed to unit vectors, each with length 1.0. To obtain a unit vector u in the same direction as vector a, one simply divides the a by its length, u = a/||a||. The variable vectors are transformed to unit vectors to ensure that each variable makes an equal contribution in determining the amount of rotation necessary to align the factor axes (see Step 3).2 Second, in the case where the two samples contain the same number of factors, the variable vectors from one of the samples are located in vector space by means of their factor loadings. The dimension of the space is determined by the number of factors (e.g., two-dimensional space for two factors, three-dimensional space for three factors). The variable vectors from the other sample are then projected onto this same space. For example, suppose that one wants to project vector b = (b2, b2) onto vector a. The projection of b on a is given by projab = ||b||(cos θ)u, where u is a unit vector in the same direction as a. The vector projab is in the same direction as a when cos θ > 0 (i.e., θ < 90°) and is in the opposite direction as a when cos θ < 0. In the case where the two samples do not contain the same number of factors, the variable vectors from the sample containing the smaller number of factors are projected onto the vector space of the sample containing the larger number of 1 Kaiser’s procedure can be used when two factor analyses include different variables, although some difficulties arise (see Kaiser et al., 1971). Kaiser’s procedure can also be used to compare correlated factor vectors. 2 Instead of allowing all variables to make an equal contribution in determining the solution, one could allow the variables to influence the solution in proportion to their commonalities so that variables having little common variance have little influence. In this case, the raw loadings can be used rather than the normalized ones.

RT44367_C005.indd 92

8/15/05 6:58:15 AM

93

CULTURE AND GENDER FACTORS

factors. In the common space, all vectors have the same origin and the same factor orientation (i.e., they are parallel). Third, the factors are rigidly rotated until a maximum degree of overlap between matching variables is obtained. The two sets of variable vectors are rotated simultaneously to that position where they are pairwise as close as possible. That is, the factor axes are rotated until the sum of products of loadings on corresponding factors is maximized. The rotation yields a unique solution. After the solution is obtained, the cosines of angles between matching variables are measured. For example, suppose that one wants to calculate the cosine of the angle between unit vectors u and v. One possible equation is cosθ = u•v, where u•v is the scalar product defined as u•v = u1v1 + u2v2. Two variable vectors are identical if the cosine of the angle between them equals 1.0 (i.e., the angle between them equals 0º), and orthogonal if the cosine of the angle between them equals 0 (i.e., the angle between them equals 90°). Gorsuch (1983) also recommends calculating the mean of the cosines between matching variable vectors. If the mean variable vector is low, “it is not possible to relate factors at all because the relationships among the variables are different” (p. 284). Consider the following hypothetical example. Suppose that a five-item questionnaire is given to two samples of subjects. The data from both samples are analyzed by principal-components factor analysis, and the factors are orthogonally rotated to a simpler structure using the varimax criterion. Also suppose that the number of factors obtained for both samples is two. Thus, in each sample, the variation in the five observed questionnaire items can be adequately explained in terms of two unobserved factors. To compare the factor structures from the two samples, one first transforms the original vectors to unit vectors. The factor 1.0

0

1

RT44367_C005.indd 93

2

4

-1.0 -1.0 Figure 5.2

F1

5

3

0

F2

1.0

Hypothetical data for Sample 1. F1 = factor 1 axis, F2 = factor 2 axis. The five questionnaire items are numbered 1 through 5.

8/15/05 6:58:17 AM

94

B.J. BUSHMAN ET AL.

1.0

F1

4

1

3

F2

0 2

-1.0 -1.0 Figure 5.3

5

1.0

0

Hypothetical data for Sample 2. F1 = factor 1 axis, F2 = factor 2 axis.

F1 F2

1.0

0

1 1

2

a1

a2 a5

4 4

2

a4

5 5

a3

-1.0 -1.0 Figure 5.4

3

F2 F1

β

0

3

1.0

β = angle between the factor axes for the two samples, α1 through α5 = angles.

structure for Sample 1 is shown in Figure 5.2. In Sample 1, Items 1, 2, and 3 load on factor 1, whereas Items 4 and 5 load on factor 2. The factor structure for Sample 2 is shown in Figure 5.3. In Sample 2, Items 4 and 5 load on factor 1, whereas Items 1, 2, and 3 load on factor 2.

RT44367_C005.indd 94

8/15/05 6:58:19 AM

CULTURE AND GENDER FACTORS

95

In Figure 5.4, the vectors for Sample 2 were projected onto the vectors for Sample 1. The common vector space is two-dimensional because both samples have two factors. Note that the vectors have a common origin. The factor axes were rotated to maximize the degree of overlap between matching items. The angle between the factors for the two samples is given by β, and the angles between matching items are given by α1, α2, …, α5. Veldman (1967) has written a FORTRAN computer program to carry out the steps described above. Although Kaiser’s method can be used to compare oblique factor structures, Veldman’s program only handles orthogonal factor structures. The output from the program includes the cosines of angles between matching variables and the cosines of angles between factors. An updated version of Veldman’s computer program can be obtained from the first author via E-mail ([email protected]).

RT44367_C005.indd 95

8/15/05 6:58:19 AM

RT44367_C005.indd 96

8/15/05 6:58:20 AM

6 Vigilant and Avoidant Coping: Theory and Measurement1 Heinz Walter Krohne and Boris Egloff Johannes-Gutenberg-Universität, Mainz, Germany Abstract This chapter describes the construction and validation of the Mainz Coping Inventory (MCI), an instrument for the assessment of individual differences in two central coping dimensions: vigilance and cognitive avoidance. Devised as a stimulusresponse inventory, the MCI originated from the model of coping modes developed by the first author (1993b) for describing and explaining individual differences in behavioral regulation under stressful conditions. After outlining the theoretical model, the construction and validation of the MCI is presented on the basis of a sequence of five analyses. The appraised aversiveness, controllability, and predictability of the different scenarios that were devised for assessing coping strategies are reported in Analysis 1. Analyses 2 and 3 present the statistical and psychometric properties of the scales and the internal factor structure of the MCI. Analyses 4 and 5 examine the relationships between the MCI scales and various measures of coping and related personality dimensions.

VIGILANT AND AVOIDANT COPING: THEORY AND MEASUREMENT Research on the processes by which individuals cope with stressful, and especially ego-threatening, situations has grown substantially over the past three decades (cf. Krohne, 2003; Lazarus, 1991; Moos, 1986; Zeidner & Endler, 1996). Many personality-oriented approaches in this field have established two constructs central to an understanding of cognitive responses to threat: vigilance, that is, the orientation toward the threat-related aspects of an aversive situation, and cognitive avoidance, that is, averting attention away from the threatening information (cf. Janis, 1983; Krohne, 1978, 1989, 1993a, 1993b, 1996b; Roth & Cohen, 1986; Suls & Fletcher, 1985).

The research reported in this chapter was partly supported by a joint grant from the German Academic Exchange Service and the American Council of Learned Societies to Heinz Walter Krohne. We would like to thank Lawrence R. Burns, Henry C. Ellis, Larry J. Varner, and Gerdi Weidner for making data collected with the MCI available to us.

1

97

RT44367_C006.indd 97

7/13/05 9:33:21 AM

98

H.J. KROHNE AND B. EGLOFF

Theoretical approaches that correspond with these conceptions are repression-sensitization (Bell & Byrne, 1978; Byrne, 1964), monitoring-blunting (Miller, 1980, 1987, 1996), and attention-rejection (Mullen & Suls, 1982). In regard to the relationship between the constructs of vigilance and cognitive avoidance, Byrne (1964) specifies a unidimensional, bipolar structure; Miller (1987) and Mullen & Suls (1982) leave this question open; while the first author (1989, 1993b) explicitly postulates the independent functioning of these dimensions. An advantage of the first author’s two-dimensional conceptualization is that one can distinguish between individuals who employ an increased amount of both vigilance and cognitive avoidance when confronted with stressors, and those who do not engage in either of these forms of coping. Although a multitude of instruments for measuring coping have been developed (for an overview, see Schwarzer & Schwarzer, 1996), many of these approaches suffer from a variety of weaknesses (cf. Krohne, 1996a). These include poor reliability of subscales, inappropriate use of factor-analytic techniques to determine simple structure, failure to explicitly distinguish between assessing habitual (trait) coping and measuring actual (state) coping behaviors, and a general failure to relate the devised scales to an elaborated theoretical basis. This chapter describes the construction and validation of an instrument for the assessment of individual differences in two central coping dimensions: vigilance and cognitive avoidance. The Mainz Coping Inventory (MCI; cf. Krohne, 1989, 2003; Krohne et al., 2000; Krohne, Schmukle, Burns, Egloff, & Spielberger, 2001) was developed as a stimulus-response (S-R) inventory based on the “model of coping modes” developed by the first author (1993b) for describing and explaining individual differences in behavioral regulation under stressful conditions. After outlining the theoretical model that guided the development of the MCI, the construction and validation of this inventory will be presented in the context of a sequence of five analyses. Analysis 1 reports the appraised aversiveness, controllability, and predictability of different scenarios devised for assessing coping strategies. The importance of taking personal appraisals of potentially stressful situations into account when constructing a stimulus-response instrument for measuring habitual (trait) coping is explained. Analyses 2 and 3 present the statistical and psychometric properties and the internal factorial structure of the MCI scale. Analyses 4 and 5 examine construct validity, as indicated by the relationships between the MCI scales and other measures of coping and related personality dimensions. MODEL OF COPING MODES The first author’s (1993b) model of coping modes specifies two basic dimensions of cognitive orientation toward an aversive situation: vigilance and cognitive avoidance. Vigilance refers to the intensified search for and processing of stress-related information. In contrast, cognitive avoidance is related to a class of coping strategies that aim to avert threat-relevant cues (cf. Krohne, 1989, 1996b, 2003; Krohne, Hock, & Kohlmann, 1992). Both dimensions describe actual stress-related operations as well as individual differences

RT44367_C006.indd 98

7/13/05 9:33:21 AM

99

VIGILANT AND AVOIDANT COPING Threat Situation

Immediate Feeling State

Anxiety/Fear Reaction

Coping Behavior

Expectation of an Uncontrollable Increase in Emotionality –Fear of Fear–

Danger Stimuli

Perception of Somatic Arousal

Cognitive Avoidance –Fear Control–

Situational Ambiguity

Experience of Uncertainty

Vigilance –Danger Control– Apprehension –Fear of Danger–

Figure 6.1

Hypothetical relationships between the components of an aversive situation, immediate feeling states, anxiety/fear reactions, and coping behavior. (The dashed arrow specifies a further possible relationship not explicitly considered in the model of coping modes.)

in the disposition to employ one or the other group of coping strategies. Concerning the dispositional (trait) level, it was posited that the frequency of employing vigilance and cognitive avoidance would vary independently. This model extends beyond similar approaches in coping research, for example, the concepts of “repression-sensitization” (Bell & Byrne, 1978; Byrne, 1964) and “monitoring-blunting” (Miller, 1987; Miller, Combs, & Kruus, 1993), in relating the descriptive constructs of vigilance and cognitive avoidance to an explicative basis (cf. Krohne, 1993b). According to other theoretical approaches and the findings in a number of empirical studies (cf. Epstein, 1972), most threat situations can be characterized by two general aspects: the actual presence of danger stimuli and a high degree of situational ambiguity (cf. Figure 6.1). Ambiguity in a danger situation is an important factor in triggering anxiety because it counteracts the immediate employment of overt reactions that are aimed at removing the danger. Corresponding to these two general aspects of a threatening situation, it is postulated that two reactions will be triggered in persons who are confronted with a frightening situation: the perception of somatic arousal caused by the presence of aversive stimuli and the experience of uncertainty as related to ambiguity. These two reactions are viewed as elementary feeling states that can link with more complex anxiety/fear reactions. When confronted with aversive stimuli, the perception of somatic arousal can lead to the expectation of a further increase in emotionality, which might not be controllable. This type of reaction has been designated as fear of fear (cf. Mineka, 1985; Reiss & McNally, 1985). In contrast, the experience of uncertainty may lead to apprehension about being surprised by negative developments in a stressful situation, a tendency that has been called fear of

RT44367_C006.indd 99

7/13/05 9:33:22 AM

100

H.J. KROHNE AND B. EGLOFF

danger (cf. Breznitz, 1984). In addition, it was postulated that the perception of intense somatic arousal associated with the anticipation of danger will initiate the tendency to avoid threat-related cues. This type of coping behavior has been called fear control (Leventhal, 1970; Wilson, 1985). In contrast, the intensified experience of uncertainty releases behavioral impulses characteristic of vigilance (danger control). In these cases, we are referring to cognitive processes instead of overt reactions like fight or flight, which are probably not possible because of situational ambiguity. Cognitive avoidance can be more precisely defined as a class of coping strategies that aim at shielding the organism from stimuli that induce arousal. This type of avoidance behavior, which is called arousal-motivated, might reduce an existing emotional state that is experienced as too intense or prevent an impending strong and possibly uncontrollable increase in arousal. Vigilance, in contrast, designates a class of coping strategies that are employed in order to reduce or prevent a further increase in uncertainty; that is, it designates uncertainty-motivated coping. Thus, individuals strive to construct a cognitive schema for understanding the aversive situation and its eventual course in order to prevent the possibility of being negatively surprised. Habitual preferences for cognitive avoidance or vigilance are explained by the constructs of intolerance of emotional arousal, which leads to arousal-motivated coping, and intolerance of uncertainty, leading to uncertainty-motivated coping. Individuals vary in the extent to which they are able to tolerate higher states of emotional arousal or uncertainty (cf. Roth & Cohen, 1986; Rothbart & Mellinger, 1972). Persons with a high level of intolerance of emotional arousal are likely to employ cognitive-avoidant coping. Intolerance of uncertainty, on the other hand, would be associated with vigilant coping strategies (cf. Figure 6.2). The specific configuration of a person’s scores on both dimensions of dispositional coping (e.g., high avoidance, low vigilance) is called coping mode. Four coping modes are distinguishable: consistent vigilance, consistent cognitive avoidance, fluctuating coping, and situation-related coping. High intolerance of uncertainty, along with low intolerance of arousal (sensitization), is associated with consistent vigilance (high vigilance, low cognitive avoidance). Persons with this coping mode characteristically fail to organize their coping behavior in a manner that takes advantage of the coping-related aspects inherent in different threatening situations, for example, according to the degree to which different situations can be influenced. Instead, they consistently attend to threat-relevant information. In threatening situations, their prime concern is to construct a cognitive schema of the impending danger and, hence, to reduce uncertainty. In contrast, persons with a high intolerance of arousal and a low intolerance of uncertainty (repression) employ consistent cognitive avoidance (low vigilance, high cognitive avoidance). These individuals exhibit a highly generalized tendency to disregard, both mentally and instrumentally, the aversive characteristics of a situation. Their main problem is to defend themselves against the emotional arousal triggered by cues prior to confrontation with an aversive event, rather than developing valid expectations concerning the further development of a threatening situation.

RT44367_C006.indd 100

7/13/05 9:33:23 AM

101

Figure 6.2

Intolerance of Uncertainty –Uncertainty-Motivated Coping–

Vigilance

high

Sensitization

High Anxiety

low

high

Low Anxiety

Repression low

high Consistent Vigilance low

SituationRelated Coping

Fluctuating Coping high

Consistent Cognitive Avoidance low

Cogitive Avoidance

Intolerance of Emotional Arousal –Arousal-Motivated Coping–

VIGILANT AND AVOIDANT COPING

Model of coping modes.

People who are highly intolerant of both arousal and uncertainty are classified under the mode of fluctuating coping, that is, high vigilance and avoidance, and are characterized by a strong propensity for deploying vigilant as well as avoidant strategies. Thus, they are threatened by a lack of valid expectations and the emotional arousal triggered by danger-related cues. Because they cannot defend themselves against both types of threat at the same time (e.g., attend to threat-related information while disregarding this information), they exhibit unstable or fluctuating coping behavior. For example, they may first employ attentional diversion, but then feel that actually monitoring the situation would be better. This fluctuating or erratic coping behavior is typical of anxious individuals. However, one should not interpret Figure 6.2 as implying a one-to-one correspondence. Although most anxious individuals tend toward the coping pattern of high vigilance and high avoidance, not everyone with this pattern can be designated as anxious. In contrast, individuals who are characterized by low intolerance can withstand both uncertainty and arousal relatively well. Because of their tolerance, they may pursue a particular coping strategy long enough to determine whether or not it is effective. Their coping behavior is characterized by a marked orientation toward whatever situational requirements prevail at any given time. Persons with this mode (called situation-related coping) do not feel especially hampered by either uncertainty or emotional arousal. Consequently, they use cognitive avoidance only in threatening situations that cannot be influenced by overt behavior, and vigilance in situations in which controllability could be increased by carefully monitoring certain threat-related aspects. This type of coping is considered to be typical of low-anxious, nondefensive individuals. It should be noted, however, that not all individuals who are low on both types of coping dimensions can be designated as nondefensive. A low tendency to employ vigilant or avoidant coping behavior could also indicate lack of sensitivity to uncertainty and emotional arousal, and perhaps a general deficit in coping resources. This specific configuration seems to be typical of psychopathic persons, who are described as being hyporesponsive to future dangers. Aversive cues are not appropriately appraised and, hence, do not initiate coping behavior

RT44367_C006.indd 101

7/13/05 9:33:24 AM

102

H.J. KROHNE AND B. EGLOFF

that, in turn, may be reinforced after its successful execution (cf. Gray, 1985; Hare, 1978; Patrick, Cuthbert, & Lang, 1994). Empirical studies provide support for the validity of the model of coping modes. For example, Hock, Krohne, and Kaiser (1996) examined the association of vigilance and cognitive avoidance in processing ambiguous stimuli with a threatening or a neutral meaning. Vigilant individuals were characterized by processing activities that favored the intake and storage of threatening stimuli, rather than giving thought to the neutral meaning of ambiguous stimuli (see also Hock & Krohne, 2004). The authors (1996) investigated the influence of coping variables on behaviors that have been described as “repressive emotional discreteness” (cf. Hansen & Hansen, 1988; Orbach & Mikulincer, 1996). This term refers to findings in research that repressors and nonrepressors do not differ in reporting the dominant emotion elicited in a stressful situation or by a specific memory. However, repressors report less intense nondominant emotions than nonrepressors. Using an experimental design in which emotions were measured as they were actually felt, the authors (1996) demonstrated that repressors, compared to nonrepressors, reported the same amount of guilt as the dominant emotion after failure, but showed less fear, sadness, and hostility. No differences were observed for positive emotions after success, indicating that repressive discreteness is restricted to negative emotions (see also Krohne, Pieper, Knoll, & Breimer, 2002). Kohlmann, Weidner, and Messina (1996) found support for the dissociation of verbal-autonomic responses in persons characterized by an avoidant coping style (cf. Asendorpf & Scherer, 1983; Kohlmann, 1993; Weinberger, Schwartz, & Davidson, 1979). Persons high in avoidance (as compared to low avoiders) showed stronger increases in systolic blood pressure in relation to an increase in self-reported anxiety during speech preparation and delivery. Krohne, Slangen, and Kleemann (1996) investigated the influence of vigilance and cognitive avoidance on state anxiety and indicators of intra- and postoperative adjustment in patients undergoing elective maxillofacial surgery under general anesthesia. Their results demonstrated the important role of dispositional coping variables in predicting preoperative anxiety and adaptation. For example, patients characterized by high vigilance reported higher levels of affective anxiety symptoms and, especially, cognitive anxiety. In comparison to high-vigilant patients, individuals low in vigilance had a higher risk of manifesting an unstable hemodynamic course during surgery. On the other hand, female patients with high dispositional vigilance needed and received significantly higher doses of narcotic agents at anesthesia induction than lowvigilant women (cf. also de Bruin, Schäfer, Krohne, & Dreyer, 2001). MEASUREMENT OF VIGILANCE AND COGNITIVE AVOIDANCE Design of the Mainz Coping Inventory Following the conceptual approach of stimulus-response inventories in anxiety research, the MCI was constructed to assess the dimensions of vigilance

RT44367_C006.indd 102

7/13/05 9:33:25 AM

103

VIGILANT AND AVOIDANT COPING MCI-E 1. Speech. Imagine that you have to make a speech to a group of people (i.e. participants of a course or seminar, parents at a meeting at their children’s school) in about one hour. 2. Exam. Imagine that you will have an important examination the next morning. 3. Job interview. Imagine that you applied for a job and that, in a few minutes, your application interview will start. 4. Mistake on the job. Imagine that you have made a mistake on the job which shouldn’t have happened and that you are to have a talk with your boss.

Figure 6.3

MCI-P 1. Dentist. Imagine that you haven’t been to the dentist for quite a long time. You are now sitting in his waiting room because you are having problems with your teeth. 2. Group of people. Imagine that you are walking alone through town in the late evening. A group of people, who look suspicious, approach you from out of a side street. 3. Inexperienced driver. Imagine that you are a front seat passenger next to an obviously inexperienced driver. Road conditions are poor due to snow and ice. 4. Turbulent flight. Imagine that you are sitting in an airplane. The flight has been turbulent for quite a while now, the “No Smoking” and “Fasten Seat Belts” signs have lit up.

Eight scenarios of the MCI.

and cognitive avoidance. This inventory contains descriptions of potentially anxiety-evoking situations (cf. Figure 6.3). Depending on the purpose of an investigation (such as coping with aversive medical procedures or with an important examination), only situations that involved ego-threats (subtest MCI-E) or physical dangers (subtest MCI-P) are presented. Descriptions of five vigilance and five cognitive avoidance reactions were assigned to each scenario. Examples of vigilant and avoidant responses to the “Exam” scenario from the subtest MCI-E are presented in Figure 6.4. Subjects were instructed to indicate which of the listed coping reactions they would employ in a given situation. The answers were summed separately for the vigilance and cognitive avoidance items across the four situations for each subtest, thus yielding four dispositional coping subtest scores: vigilance in response to ego-threat (VIG-E) or physical threat (VIG-P), and cognitive avoidance as assessed by the CAV-E and CAV-P subtests. In addition, total VIG-T and CAV-T scores can be calculated. The design of the MCI is similar to the one employed by Miller (1987, 1996) and her coworkers in the Monitor-Blunter Style Scale (MBSS), which focuses on assessing the cognitive coping styles of monitoring and blunting. However, in contrast to the MCI, the MBSS items describe only fictitious uncontrollable situations. Within this design, only the manifestation of monitoring behaviors, which are more likely to be unadaptive and, consequently, not demanded in uncontrollable situations, can be related to habitual tendencies (traits). However, for blunting, an adaptive form of coping in uncontrollable situations (cf. Krohne, 1986; Lazarus, 1983), it remains open as to whether such behaviors can be related to situational demands, or to habitual factors employed by individuals who generally use blunting as a form of coping. The MCI tries to overcome this problem by systematically varying the degree of controllability and predictability of its fictitious threat scenarios, thus creating the basis for interpreting individual differences in the endorsement of MCI reactions in terms of habitual coping.

RT44367_C006.indd 103

7/13/05 9:33:27 AM

104

H.J. KROHNE AND B. EGLOFF Imagine that you will have an important examination the next morning. In this situation...

Figure 6.4

True

False

1. Imagine that I could be surprised by unexpected questions....................................................................................O

O

2. I try not to think about the examination and do something else.............................................................................................O

O

3. I remember previous examinations........................................... O

O

4. I stay calmer than most people I know......................................O

O

5. I tell myself that the examination will probably proceed in a fair way............................................................................... O

O

6. I review the catalogue listing the questions that will be asked during the examination....................................................O

O

7. I tell myself: “I will pass this examination at any rate.”............O

O

8. I once again ask my friends who have already taken the examination which questions they were asked..........................O

O

9. I tell myself: “I’ve coped with situations that were far more difficult.”...........................................................................O

O

10. I think about what I can do if I have difficulties with some questions....................................................................................O

O

Coping reactions for the threat scenario “Exam.” Items 1, 3, 6, 8, and 10 indicate vigilant coping reactions; items 2, 4, 5, 7, and 9 indicate cognitive avoidant coping reactions.

Participants Data were obtained from three independent samples of undergraduate students. Sample 1 consisted of 227 students from the University of New Mexico. Samples 2 and 3 consisted, respectively, of 169 and 325 participants from the State University of New York at Stony Brook. The number of males and females and the mean age of the participants in each sample are reported in Table 6.1. The total combined sample of 721 students completed the questionnaires described below in group sessions as part of their course requirements. Analysis 1: Appraisal of the MCI Situations Situations presented as stimuli in an inventory intended to measure habitual coping have to meet a number of prerequisites. They should, first, induce a

RT44367_C006.indd 104

7/13/05 9:33:29 AM

105

VIGILANT AND AVOIDANT COPING Table 6.1 Description of the Samples Gender

Sample 1 Sample 2 Sample 3 Total Sample

N

Male

Female

Age

227 169 325 721

57 45 132 234

170 124 193 487

21.8 21.3 20.1 20.9

certain degree of threat and, hence, require coping efforts. However, in order to allow for variability in coping reactions, the experienced threat should not be too severe. Second, the scenarios should vary with regard to the configuration of appraised controllability and predictability. Finally, the vast majority of respondents should have previously experienced similar stressors. Using the responses provided by the 227 participants of Sample 1, we first checked whether the four fictitious situations of each part were appraised as being aversive (threat-inducing) at all. In addition, we examined whether the scenarios met the criteria for variable predictability and controllability. Finally, we asked the participants to indicate whether they had already experienced a similar situation. Although responses to the appraised aversiveness, predictability, and controllability were made on a five-point scale (varying from “Not at all” to “Very”), experience was indicated by means of a “Yes-No” format. With the exception of the scenarios “Job application” (subtest MCI-E) and “Dentist” (MCI-P), mean aversiveness ratings of the situations were above the median point of 3 (see Figure 6.5). Although these results demonstrate that the vast majority of MCI scenarios are indeed experienced as threat-inducing, subsequent multivariate analyses of variance (MANOVAs) with repeated measures revealed differences among these situations. These analyses were carried out separately for the situations of both subtests and yielded significant results for MCI-E, Wilks’s Λ = .52, F(3, 224) = 69.79, p < .001, and MCI-P, Λ = .55, F(3, 224) = 61.59, p < .001. Analogous MANOVAs were conducted for the predictability and controllability appraisals, indicating that the scenarios of both MCI parts varied considerably on these dimensions [all Fs(3, 224) > 50, p < .001]. Further analyses indicated that the vast majority of participants had already experienced the situations presented in the MCI. These results demonstrate that we were able to meet the prerequisites for the construction of stimuli in an S-R inventory of habitual (trait) coping. The situations are experienced as moderately threat-inducing. They exhibit very different patterns of controllability and predictability. For example, although the situation “Group of people” is appraised as comparatively controllable and less predictable, the reverse is true for the scenario “Dentist.” In contrast, the situation “Turbulent flight” yields very low ratings for both dimensions (cf. Figure 6.5). Finally, the vast majority of participants reported that they had previously experienced similar situations.

RT44367_C006.indd 105

7/13/05 9:33:29 AM

106

H.J. KROHNE AND B. EGLOFF

MCI-E 5 4 3 2 1 0

ch

ee

am

n

n

io

at

in

i

bl

Pu

p cs

p

Ex

ob

ap

io at

lic

b

jo

ak ist

J

Aversiveness

n

eo

e th

M

Predictability

Controllability

MCI-P 5 4 3 2 1 0 t

tis

en

D

p ou

of

r

G

Aversiveness

Figure 6.5

t gh

er iv

e

pl

o pe

nc

rie

e

p ex

ed

dr

In

Predictability

l bu

li tf

en

r Tu

Controllability

Appraisals of the MCI scenarios.

Although these response patterns were very similar for males and females, significant gender differences could be observed on a number of variables. In a first step, we analyzed the scenario appraisals summed up separately for each threat type (ego vs. physical threat) and for the complete inventory. Concerning appraised aversiveness, women as compared to men gave higher ratings in general (i.e., for the complete inventory as well as for each threat

RT44367_C006.indd 106

7/13/05 9:33:31 AM

107

VIGILANT AND AVOIDANT COPING Table 6.2 Means, Standard Deviations, and Reliabilities of the MCI Subscales Total Sample M SD

Scale VIG-E CAV-E VIG-P CAV-P VIG-T CAV-T

14.74 11.26 12.97 10.68 27.70 21.94

3.43 3.65 3.64 3.56 6.31 6.18

M

Men

14.05 12.16 12.03 11.38 26.06 23.54

Women SD

SD

M

3.56 3.44 3.62 3.36 6.37 5.67

15.07* 10.83* 13.42* 10.34* 28.49* 21.17*

3.33 3.68 3.56 3.60 6.13 6.27

Total Sample α rretest .73 .73 .73 .68 .83 .80

.70 .81 .71 .78 .76 .84

Note: N(total sample) = 721, N(men) = 234, N(women) = 487, N(retest) = 145 (subsample of Sample 2; time span: 2 weeks). α = Cronbach’s α. *p < .001 (two-tailed t-test for gender differences).

type, ps < .001). There was a tendency for men to report a higher level of predictability of the ego-threat scenarios, t(225) = 1.72, p = .087, whereas no gender differences were observed regarding the appraised controllability of the scenarios. In a second step, we conducted t-tests separately for each scenario. Female participants gave higher aversiveness ratings for the scenarios “Job interview” (p < .001), “Mistake on the job” (p < .001), “Group of people” (p < .001), “Inexperienced driver” (p < .01), and “Turbulent flight” (p < .05). For males the situation “Mistake on the job” was more predictable (p < .05), and the scenarios “Group of people” (p < .01) and “Turbulent flight” (p < .05) more controllable than for females. In addition, the scenario “Speech” was more predictable (p < .01) and controllable (p < .05) for males than for females. Analysis 2: Statistical and Psychometric Properties of the MCI The means, standard deviations, and internal consistency reliability for the MCI total scores and the four subscales are reported in Table 6.2. Cronbach alpha coefficients varied between α = .68 for CAV-P and .83 for VIG-T. The retest reliabilities for a time span of two weeks varied between rtt = .70 for VIG-E and .84 for CAV-T. Although the psychometric properties of the scales were similar for both males and females, the female participants scored higher on the vigilance and lower on the cognitive avoidance scales (all differences: p < .001). Analysis 3: Internal Structure of the MCI Correlations between the MCI scales and subscales are reported in Table 6.3 for the total sample of 721 students. As predicted from the coping modes model, vigilance and cognitive avoidance were only moderately associated, with correlations of r = –.15 for MCI-E, r = –.26 for MCI-P, and r = –.19 for the total MCI scores. On the other hand, the corresponding VIG and CAV scales were strongly correlated (VIG: r = .59; CAV: r = .47). This pattern

RT44367_C006.indd 107

7/13/05 9:33:31 AM

108

H.J. KROHNE AND B. EGLOFF

Table 6.3 Correlations of the MCI Subscales Scale

CAV-E

VIG-P

CAV-P

VIG-T

CAV-T

VIG-E CAV-E VIG-P CAV-P VIG-T

–.15

.59 –.14

–.01 .47 –.26

.88 –.16 .90 –.16

–.09 .86 –.23 .85 –.19

Note: N = 721. Convergent validity coefficients are shown in italics, discriminant validity coefficients are shown in boldface.

of results provided strong support for the relatively independent functioning of vigilance and avoidance, and for the discriminant validity of the VIG and CAV scales. The correlations among the corresponding MCI scales also demonstrated that the variables for both parts of this inventory were related to the same trait constructs, but were not identical. In order to further verify the dimensionality of the MCI, the scenario scores were factored, with subsequent varimax rotation. Ratings of the five vigilance and five cognitive avoidance strategies for each of the eight scenarios were summed separately, forming 16 coping variables, which were subjected to principal-components factor analyses (PCA). The following criteria were applied in regard to the number of factors to be extracted and the extent to which these factors were considered meaningful (cf. Rost & Schermer, 1989): Eigenvalues greater than 1, accounting for at least 3% of the total variance, and defined by at least three marker variables. The marker variables for a factor had to meet the criteria of a significant absolute factor loading (a > .40) and of relative unidimensionality; that is, the share of both highest loadings on the communality should differ by at least 25% ([a12 – a22]/h2 > .25). The PCA produced five eigenvalues greater than 1 (3.64, 2.56, 1.28, 1.15, 1.01). Based on this solution, the rotated factor matrices with two, three, four, and five components were inspected. As can be noted, the first two eigenvalues were distinctly larger than the remaining ones, with the slope (scree) leveling off after the third eigenvalue. Because the solutions with three or more components produced factors with two or fewer marker variables, the two-factor solution, which accounted for 38.8% of the total variance, was considered most meaningful. This solution is reported in Table 6.4, in which a clear separation of the two factors can be noted for the eight vigilance and eight avoidance variables. Analysis 4: Relationships with Other Coping Variables (Construct Validity) The construct validity of the MCI was evaluated by examining the relationships between the VIG and CAV scales of this instrument with other coping measures. The following instruments were administered: the COPE Inventory (Carver, Scheier, & Weintraub, 1989), the Social Desirability Scale (Crowne & Marlowe, 1964), and the Social Problem-Solving Inventory (SPSI; D’Zurilla

RT44367_C006.indd 108

7/13/05 9:33:32 AM

109

VIGILANT AND AVOIDANT COPING

Table 6.4 Results of a Principal-Components Analysis with Varimax-Rotation of the Vigilance and Cognitive Avoidance Scores Summed up for Each Situation

VIG-E1 VIG-E2 VIG-E3 VIG-E4 VIG-P1 VIG-P2 VIG-P3 VIG-P4 CAV-E1 CAV-E2 CAV-E3 CAV-E4 CAV-P1 CAV-P2 CAV-P3 CAV-P4 Explained variance

F1

F2

H2

.64 .65 .70 .67 .56 .60 .62 .60 –.08 –.06 –.17 .01 –.06 .08 .05 –.21

.06 .13 –.05 –.10 –.20 –.02 –.07 –.13 .61 .65 .64 .72 .63 .41 .53 .48

.41 .44 .50 .46 .35 .36 .39 .37 .38 .42 .44 .52 .40 .18 .28 .28

20.6%

18.2%

Note: N = 721. Marker variables are shown in boldface. VIG-E1 = vigilance score in ego threat scenario 1; CAV-P4 = cognitive avoidance score in physical threat scenario 4. H2 = communality.

& Nezu, 1990). The analyses of the responses to the SPSI were based on scales derived from the five-factor model proposed by Maydeu-Olivares and D’Zurilla (1996). In previous research that compared the coping reactions of men and women, complex relationships have been observed with criterion variables, such as health status, anxiety, and depression (Miller & Kirsch, 1987; Weidner & Collins, 1993; see also Analyses 1 and 2). These findings emphasize the importance of including both men and women in studies of coping so that gender differences can be taken into account. Following this strategy, the correlations of the MCI variables with other coping and criterion variables were evaluated separately for men and women in Analysis 4 and Analysis 5. Table 6.5 presents the correlations of the four MCI subscales and the total scores with the different variables from the other coping tests. In interpreting these differences, it is important to consider the unequal male and female sample sizes. In general, the coefficients exhibit a theoretically meaningful pattern. As hypothesized by Carver et al. (1989), vigilance is linked with planning (thinking about how to cope with a stressor), restraint coping (holding oneself back and not acting prematurely), and seeking support for instrumental reasons (seeking advice, assistance, or information). Not predicted by Carver et al. are the positive correlations of vigilance with active coping (taking active measures against a stressor or its effects) and seeking emotional support (moral support, sympathy,

RT44367_C006.indd 109

7/13/05 9:33:33 AM

110

H.J. KROHNE AND B. EGLOFF

Table 6.5 Correlations of the MCI for Women (First Row) and Men (Second Row) with other Coping Variables Coping Scales

VIG-E

VIG-P

VIG-T

CAV-E CAV-P CAV-T

.24** .25*

.15 .21

.22** .25*

–.03 .31*

.13 .20

Planning

.22** .25*

.17* .22

.22** .26*

–.03 .26*

–.03 .13

–.03 .23

Suppression of Competing Activities

.17* .14

.18* .18

.20* .17

–.09 .27*

–.02 .13

–.07 .24*

Restraint Coping

.28*** .16

.22** .16

.29*** .02 .18 –.03

.08 .09

.06 .03

Seeking Social Support: Instrumental

.23** .29*

.15 .25*

.22* .29*

–.11 .11

.01 .15

–.06 .15

Seeking Social Support: Emotional

.27** .35**

.11 .29*

.21* .35**

.03 .09

.04 –.01

.04 .05

Positive Reinterpretation & Growth

.10 .31**

.13 .28*

.13 .32**

.22** .35**

.01 .21

.14 .33**

Acceptance

.02 .17

.07 .07

.05 .13

.18* .15

.09 .21

.16 .20

Turning to Religion

.20* .26*

.25** .15

.26** .22

.16 .12

.08 .06

.14 .10

Focus on & Venting of Emotions

.30*** .06

.30*** .01

.34*** –.13 .04 –.11

–.09 .03

–.13 –.06

Denial

.01 .26*

.08 .18

.05 .24*

–.14 –.05

–.01 .06

–.09 .00

.03 .04

.02 .01

–.21* –.10

.04 .04

–.10 –.04

.23** .19

.16 .11

–.18* –.12

–.03 –.05

–.12 –.10

.00 .27*

–.04 .04 –.33** –.26*

COPE Inventorya Active Coping

Behavioral Disengagement

.00 –.02

Mental Disengagement

.04 .01

Alcohol-Drug Disengagement

.07 .25*

–.06 .24*

–.01 –.31*

.01 –.14

Social Desirability Scaleb

RT44367_C006.indd 110

.00 –.24

.16* .14

.03 –.04

.05 .30*

.00 –.35** .12 .05

7/13/05 9:33:34 AM

111

VIGILANT AND AVOIDANT COPING Table 6.5 (Continued) Coping Scales SPSI-Ra Positive Problem Orientation

VIG-E

VIG-P

–.11 .05

–.11 .11

VIG-T –.13 .09

CAV-E .37*** .54***

Negative Problem Orientation

.16 .14

.21* .12

.22** .14

–.34*** –.43***

Rational Problem Solving

.12 .21

.09 .29*

.12 .27*

.01 .36**

CAV-P CAV-T .19* .29*

.33*** .50***

–.15 –.29*** –.26* –.41*** .01 .21

.01 .33**

Impulsivity/Carelessness

–.06 –.08

.02 –.15

–.02 –.13

–.06 –.10

–.03 –.03

–.05 –.08

Avoidance Style

–.01 .10

.06 .01

.03 .06

–.14 –.19

.01 –.04

–.08 –.14

Note: The first row of correlations is for females, the second row is for males. a Based on a subsample of Sample 3 (N = 205: 137 women, 68 men). b Based on Sample 1 (N = 227: 170 women, 57 men). *p < .05. **p < .01. ***p < .001 (two-tailed). VIG = Vigilance, CAV = Cognitive avoidance: -E = Ego, -P = Physical, -T = Total. SPSI-R = Social Problem Solving Inventory-Revised.

or understanding). However, if one considers the very high associations between planning and active coping as well as between the two support-seeking variables as reported by Carver et al. (1989, Table 3; r = .67 and .69, respectively), this pattern of correlations is not surprising. Turning to religion is a coping strategy that may be important to only some people. These individuals might employ it under stress for widely varying reasons (for example, emotional support or positive reinterpretation). Not surprisingly, no clear pattern of associations with CAV or VIG emerged. Positive reinterpretation and growth is an emotion-focused way of coping that relies primarily on a positive reappraisal of the aversive situation (cf. Lazarus, 1991). Consequently, this tendency should positively correlate with cognitive avoidance, a hypothesis confirmed by our data. However, the positive association of this coping scale with vigilance (observed in the male sample) comes unexpectedly. This result might again be mediated by the positive correlations of this scale with active coping and planning (cf. Carver et al., 1989, Table 3). The modest coefficients observed for the COPE subscale Denial clearly disconfirm our theoretical expectations. However, in our opinion the items of this scale have an ambiguous relationship with the coping dimensions cognitive avoidance and vigilance as conceptualized in the model of coping modes. Wording of these items contains elements of orientation toward as well as away from a stressor (e.g. “I pretend that it hasn’t really happened”). A clear avoidance item would

RT44367_C006.indd 111

7/13/05 9:33:35 AM

112

H.J. KROHNE AND B. EGLOFF

be, for example, “Very rarely have bad things happened to me.” Consequently, Carver et al. (1989, Table 4) report zero correlations of their denial scale with the blunting and monitoring dimensions of the MBSS. Marked gender differences emerged for the COPE scales focus on and venting of emotions (to focus on whatever distress one is experiencing and to ventilate those feelings) and alcohol-drug disengagement (a one-item scale: “I drink alcohol or take drugs, in order to think about it less”; cf. Carver et al., 1989, Table 1). The first tendency clearly involves vigilant strategies, which are perhaps more preferably employed by vigilant women. Vigilant men, on the other hand, might be more inclined to rely on externally oriented vigilant strategies, as the comparatively high correlations of vigilance and seeking instrumental support in the male sample indicate. The finding that coping with distress by drinking alcohol or taking drugs is significantly related to all MCI variables is theoretically meaningful: High avoiders deny relying on this (socially disapproved) way of coping, while vigilant individuals (who are more distressed by aversive situations) admit employing these strategies. However, this pattern only holds for male participants. Unfortunately, Carver et al. do not report gender differences in the means and standard deviations of this scale. We suppose that both parameters are considerably smaller for females, thus explaining the zero correlations for this subsample. Correlations with the Marlowe-Crowne Scale (which measures the tendency to portray oneself in a highly favorable light or to defensively avoid social disapproval) are small but yield the expected pattern, with a positive relationship to cognitive avoidance and a negative association with vigilance in situations that represent a threat to the self (CAV-E). The SPSI-R scale positive problem orientation (an optimistic view concerning the outcome of a problem) is positively correlated with avoidance, while negative problem orientation (anxiety and pessimism) yields negative correlations with cognitive avoidance and positive associations with vigilance. For male participants, rational problem solving is positively related to vigilant and cognitive avoidant coping. This finding might be a product of the male stereotype as effective or adaptive problem solver: The more men cope (either vigilantly or avoidantly) with a problem situation, the more they view themselves as rational problem solvers. The SPSI-R subscale avoidance style does not exhibit the expected association with cognitive avoidance. However, we have severe doubts whether the name avoidance style is justified, because the main elements of this dimension are procrastination, passivity, and dependency. None of these items address the central aspect of cognitive avoidance: turning attention away from threat-relevant cues. Analysis 5: Relationships with Various Criterion Variables A second analysis collected further information about the validity of the MCI scales by assessing relationships between these dimensions and measures of trait affectivity (the Positive and Negative Affect Schedule, PANAS; Watson, Clark, & Tellegen, 1988; and the EASI-III Temperament Survey; Buss & Plomin, 1975), trait anxiety (the State-Trait Anxiety Inventory, STAI; Spielberger, 1983),

RT44367_C006.indd 112

7/13/05 9:33:35 AM

VIGILANT AND AVOIDANT COPING

113

depression (the Beck Depression Inventory, BDI; Beck, Ward, Mendelson, Mock, & Erbaugh, 1961), symptom reporting (the Pennebaker Inventory of Limbic Languidness, PILL; Pennebaker, 1982), self-esteem (Rosenberg, 1965), and appraisal of the threat scenarios. Appraisal scores were gained by summing up the aversiveness, predictability, and controllability ratings for the eight threat scenarios described above (cf. Analysis 1). In addition, the NEO Five Factors Inventory (Costa & McCrae, 1989) was administered. As expected from the model of coping modes, cognitive avoidance, especially as related to ego-threatening stressors, correlates positively with positive affectivity and negatively with negative affectivity, while vigilance is only associated with negative affectivity (cf. Table 6.6). A similar pattern of associations exists for the subscale fear of the EASI-III survey. Correlations for negative affectivity and fear are generally somewhat stronger for females than for males. Trait anxiety and symptom reporting are positively associated with vigilance and negatively with cognitive avoidance (with a smaller coefficient for CAV-P). Depression correlates highly with both vigilance scales and (only for females) with cognitive avoidance. Interestingly, only the tendency to employ cognitive avoidance in ego-threatening situations is substantially related to self-esteem. As expected from the definition of the coping dimensions vigilance and cognitive avoidance, the aversiveness appraisal is positively related to VIG and negatively to CAV. Similarly meaningful is the finding that an increase in vigilance is associated with a decrease in appraised predictability of threatening situations. In contrast, appraised controllability is unrelated to both coping variables. Of the five NEO dimensions, neuroticism shows the expected strong correlations with the MCI scales. For males, extraversion yields a positive association with cognitive avoidance. For females, on the other hand, avoidance correlates positively with both agreeableness and conscientiousness. DISCUSSION In this chapter, a personality-oriented theory that focused on two central coping constructs, vigilance and cognitive avoidance, was presented, and the construction of the MCI, an instrument for assessing these characteristics, was described. The first author’s (1989, 1993b) model of coping modes specifies vigilance and cognitive avoidance as the central dimensions of a cognitive orientation toward coping with aversive (especially threatening) situations. Vigilance is characterized by an intensified search for and processing of aversive information; cognitive avoidance refers to a class of coping strategies that aim at averting threat-related cues. Consistent with Carver et al.’s (1989) view that coping has a motivational basis, vigilance was defined as uncertainty-motivated behavior. Individuals who frequently employ vigilant strategies are especially influenced by the ambiguity inherent in most threatening situations (“intolerance of uncertainty”). Consequently, vigilant coping is employed in order to reduce or prevent a further increase in uncertainty.

RT44367_C006.indd 113

7/13/05 9:33:36 AM

114

H.J. KROHNE AND B. EGLOFF

Table 6.6 Correlations for Women and Men of the MCI Scales with Selected Criterion Measures SCALE Variable PANASa Positive Affectivity Negative Affectivity EASI-IIIa Fear Anger STAIb Trait anxiety BDIb Depression PILLc Symptom Reporting ROSENBERG Scalea Self-esteem SCENARIO Appraisalsb Aversiveness

VIG-E

VIG-P

VIG-T

CAV-E

CAV-P

CAV-T

.03 .04

.03 .09

.03 .07

.20* .43***

–.06 .09

.32 .32**

.15 .10

.17* .16

.18* .15

–.45*** –.23

.25** .08

.31*** –.33*** .05 –.34**

.29*** –.01 .20* –.05

.15 –.06

.20* –.06

–.10 –.24*

–.25** –.02

–.41*** –.16

–.16 –.12

–.29*** –.28*

–.10 –.16

–.12 –.23

.29*** .35**

.28*** .39**

.32*** –.35*** .41** –.30*

–.14 –.20

–.29*** –.28*

.24** .21

.24** .29*

.27*** –.24** .28* –.17

–.17* –.15

–.24** –.18

.22* .31*

.26** .46**

.27** .46**

–.22* –.13

–.25** –.11

–.16 .08 .40*** .37**

–.17* .18 .50*** .48***

–.19* .14

–.24* –.05 .40*** .49***

.50*** –.33*** .48*** –.36**

.05 .11

.26** .37***

–.30*** –.37*** –.27* –.36**

Predictability

–.30*** –.23** –.26* –.41**

–.29** –.37**

.17* .20

–.02 .29*

.09 .28*

Controllability

–.11 –.13

–.11 –.18

.11 .00

.09 .12

.12 .07

NEO-FFIc Neuroticism Extraversion Openness

RT44367_C006.indd 114

.25** .33*

–.10 –.19 .35*** .55***

.34*** –.43*** –.33*** –.43*** .53*** –.39* –.22 –.36*

.10 .10

–.05 .12

.02 .13

.13 .39*

–.20* –.39*

.01 –.14

–.09 –.31*

.07 .29

.16 .32* .14 –.12

.17 .43** .12 .08

7/13/05 9:33:37 AM

115

VIGILANT AND AVOIDANT COPING Table 6.6 (Continued) SCALE Variable

VIG-E

VIG-P

VIG-T

CAV-E

Agreeableness

–.10 –.09

–.16 –.01

–.15 –.06

.12 .09

Conscientiousness

–.03 –.03

–.03 –.15

–.03 –.11

.43*** .08

CAV-P

CAV-T

.30** .01

.24* .06

.27** –.11

.39*** –.03

Note: The first row of correlation coefficients is for females, the second is for the males. a Based on a subsample of Sample 3 (N = 205: 137 women, 68 men). b Based on Sample 1 (N = 227: 170 women, 57 men). c Based on a subsample of Sample 2 (N = 152: 109 women, 43 men). *p < .05. **p < .01. ***p < .001 (two-tailed). VIG = Vigilance, CAV = Cognitive avoidance: -E = Ego, -P = Physical, -T = Total. PANAS = Positive and Negative Affect Schedule; EASI-III Temperament Survey; STAI = State Trait Anxiety Inventory; BDI = Beck Depression Inventory; PILL = Pennebaker Inventory of Limbic Languidness; NEO-FFI = NEO Five Factors Inventory.

Individuals who frequently employ cognitive avoidance are especially sensitive to the emotional arousal triggered by danger signals (“intolerance of emotional arousal”). Cognitive avoidant coping strategies that aim at shielding the organism from danger are arousal motivated. Although vigilance and cognitive avoidance are core constructs, they do not cover the full range of coping possibilities associated with cognitive reactions to stress. Consequently, the model of coping modes that guided the construction of the MCI can be conceptualized as a partial-area theory (cf. Krohne, 1993b) that concentrates on two specific classes of cognitive coping strategies. Research strategies that aim at describing and measuring different forms of coping may be classified as either micro- or macroanalytic (cf. Krohne, 1996b). Microanalytic strategies, such as those related to the COPE Inventory (Carver et al., 1989; see also Carver & Scheier, 1993), examine a large number of specific unidimensional coping strategies. Macroanalytic analysis, which operates at a higher level of aggregation and focuses on more fundamental constructs, guided the construction of the Mainz Coping Inventory (Krohne et al., 2000). The two-dimensional macroanalytic conception of the MCI has a major advantage over unidimensional instruments. It can distinguish between individuals who employ an increased amount of both vigilance and avoidance, and those who do not consistently use either of these coping forms. In unidimensional approaches, these groups would be located in the middle range of the data distribution. The MCI also meets two other important prerequisites for the construction of dispositional coping inventories. Different threat scenarios are presented as stimuli that vary according to the degree of controllability and predictability as reported by test respondents. Furthermore, ratings of these scenarios also differentiate between coping in ego- and physically threatening situations, thus creating two subtests, the MCI-E and the MCI-P.

RT44367_C006.indd 115

7/13/05 9:33:38 AM

116

H.J. KROHNE AND B. EGLOFF

In addition to describing the design and construction of the MCI, the results of five sets of analyses of the MCI data were reported. Analysis 1 presented findings related to the appraised aversiveness, controllability, and predictability of different MCI threat scenarios. Analysis 2 and Analysis 3 presented statistical data on the psychometric properties of the MCI and the factor structure of this inventory. Analysis 4 and Analysis 5 further examined the construct validity of the MCI as indicated by correlations of vigilance and cognitive avoidance with a number of criterion measures. The MCI scenarios were found to be appropriately threat inducing and were reported as previously experienced by most of the study participants. The reliabilities of the MCI scales were also quite satisfactory, and the internal structure of the MCI was consistent with hypotheses derived from the model of coping modes, as indicated by the results of factor analyses that provided clear evidence for the postulated two-dimensional structure of this inventory. Although the measures of vigilance and cognitive avoidance varied independently, the corresponding VIG and CAV scales for assessing ego and physical threat were strongly correlated, thus confirming the postulated habitual nature of vigilant and avoidant coping. On the other hand, the associations between both parts are not such as to make a separate assessment of coping with both types of threat meaningless. Furthermore, results of the factor analysis provided evidence for the postulated two-dimensional structure of the MCI. An inspection of the association of the MCI scales with emotionality, anxiety, depression, and other coping measures generally confirmed the concurrent and discriminant validity of this inventory. It is especially noteworthy that none of the correlations with the validity criterion measure reached a level that would indicate lack of discriminant validity of the MCI scales. This was most relevant in regard to the correlations of vigilance with neuroticism and measures of negative affectivity, and between cognitive avoidance and social desirability, indicating that these constructs were related but separable. The observed gender differences call for further systematic analyses. For example, it seems necessary to register the gender-specific intentions underlying the employment of certain coping strategies. The same strategy (e.g., seeking support) can yield different relationships with other variables, such as vigilance or cognitive avoidance, depending on the intention behind it. Seeking social support, for example, may be employed in order to receive more information about a threatening event (clearly a vigilant strategy) or to divert one’s attention away from the stressor (an avoidant strategy). This idea, however, calls for adoption of a broader conceptualization of gender that includes socially determined aspects such as life and gender roles (Weidner & Collins, 1993). In conclusion, the results of research with the MCI reported in this chapter regarding the construction of threatening scenarios, the reliability of the scales, and the internal structure of the MCI provide evidence of the construct validity and strong psychometric properties of this inventory. The findings in the studies of concurrent and construct validity and discrimination also demonstrate that the Mainz Coping Inventory is a useful instrument for assessing vigilance and cognitive avoidance as major coping dimensions.

RT44367_C006.indd 116

7/13/05 9:33:38 AM

VIGILANT AND AVOIDANT COPING

117

REFERENCES Asendorpf, J. B., & Scherer, K. R. (1983). The discrepant repressor: Differentiation between low anxiety, high anxiety, and repression of anxiety by autonomic-facial-verbal patterns of behavior. Journal of Personality and Social Psychology, 45, pp. 1334–1346. Beck, A. T., Ward, C. H., Mendelson, M., Mock, J., & Erbaugh, J. (1961). An inventory for measuring depression. Archives of General Psychiatry, 4, pp. 561–571. Bell, P. A., & Byrne, D. (1978). Repression-sensitization. In H. London & J. E. Exner (Eds.), Dimensions of personality (pp. 449–485). New York: Wiley. Breznitz, S. (1984). Cry wolf: The psychology of false alarms. Hillsdale, NJ: Erlbaum. Buss, A. H., & Plomin, R. (1975). A temperament theory of personality development. New York: Wiley. Byrne, D. (1964). Repression-sensitization as a dimension of personality. In B. A. Maher (Ed.), Progress in experimental personality research (Vol. 1, pp. 169–220). New York: Academic Press. Carver, C. S., & Scheier, M. F. (1993). Vigilant and avoidant coping in two patient samples. In H. W. Krohne (Ed.), Attention and avoidance: Strategies in coping with aversiveness (pp. 295–319). Seattle and Toronto: Hogrefe & Huber. Carver, C. S., Scheier, M. F., & Weintraub, J. G. (1989). Assessing coping strategies: A theoretically based approach. Journal of Personality and Social Psychology, 56, pp. 267–283. Costa, P. T., & McCrae, R. R. (1989). NEO PI/FFI manual supplement. Odessa, FL: Psychological Assessment Resources. Crowne, D. P., & Marlowe, D. (1964). The approval motive. New York: Wiley. de Bruin, J. T., Schäfer, M. K., Krohne, H. W., & Dreyer, A. (2001). Preoperative anxiety, coping, and intraoperative adjustment: Are there mediating effects of stress-induced analgesia? Psychology and Health, 16, pp. 253–271. D’Zurilla, T. J., & Nezu, A. M. (1990). Development and preliminary evaluation of the Social Problem-Solving Inventory. Psychological Assessment, 2, pp. 156–163. Egloff, B., & Krohne, H. W. (1996). Repressive emotional discreteness after failure. Journal of Personality and Social Psychology, 70, pp. 1318–1326. Epstein, S. (1972). The nature of anxiety with emphasis upon its relationship to expectancy. In C. D. Spielberger (Ed.), Anxiety: Current trends in theory and research (Vol. 2, pp. 291–337). New York: Academic Press. Gray, J. A. (1985). Issues in the neuropsychology of anxiety. In A. H. Tuma & J. D. Maser (Eds.), Anxiety and the anxiety disorders (pp. 5–25). Hillsdale, NJ: Erlbaum. Hansen, R. H., & Hansen, C. H. (1988). Repression of emotionally tagged memories: The architecture of less complex emotions. Journal of Personality and Social Psychology, 55, pp. 811–818. Hare, R. D. (1978). Electrodermal and cardiovascular correlates of psychopathy. In R. D. Hare & D. Schalling (Eds.), Psychopathic behavior: Approaches to research (pp. 107–143). New York: Wiley.

RT44367_C006.indd 117

7/13/05 9:33:40 AM

118

H.J. KROHNE AND B. EGLOFF

Hock, M., & Krohne, H. W. (2004). Coping with threat and memory for ambiguous information: Testing the repressive discontinuity hypothesis. Emotion, 4, pp. 65–86. Hock, M., Krohne, H. W., & Kaiser, J. (1996). Coping dispositions and the processing of ambiguous stimuli. Journal of Personality and Social Psychology, 70, pp. 1052–1066. Janis, I. L. (1983). Stress inoculation in health care: Theory and research. In D. Meichenbaum & M. Jaremko (Eds.), Stress reduction and prevention (pp. 67–99). New York: Plenum. Kohlmann, C. W. (1993). Development of the repression-sensitization construct: With special reference to the discrepancy between subjective and physiological stress reactions. In U. Hentschel, G. J. W. Smith, W. Ehlers, & J. G. Draguns (Eds.), The concept of defense mechanisms in contemporary psychology. Theoretical, research, and clinical perspectives (pp. 184–204). New York: Springer-Verlag. Kohlmann, C. W., Weidner, G., & Messina, C. (1996). Avoidant coping style and verbal-cardiovascular response dissociation. Psychology and Health, 11, pp. 371–384. Krohne, H. W. (1978). Individual differences in coping with stress and anxiety. In C. D. Spielberger & I. G. Sarason (Eds.), Stress and anxiety (Vol. 5, pp. 233–260). Washington, DC: Hemisphere. Krohne, H. W. (1986). Coping with stress: Dispositions, strategies, and the problem of measurement. In M. H. Appley & R. Trumbull (Eds.), Dynamics of stress: Physiological, psychological, and social perspectives (pp. 207–232). New York: Plenum. Krohne, H. W. (1989). The concept of coping modes: Relating cognitive person variables to actual coping behavior. Advances in Behaviour Research and Therapy, 11, pp. 235–248. Krohne, H. W. (Ed.). (1993a). Attention and avoidance: Strategies in coping with aversiveness. Seattle and Toronto: Hogrefe & Huber. Krohne, H. W. (1993b). Vigilance and cognitive avoidance as concepts in coping research. In H. W. Krohne (Ed.), Attention and avoidance: Strategies in coping with aversiveness (pp. 19–50). Seattle and Toronto: Hogrefe & Huber. Krohne, H. W. (1996a). Angst und angstbewältigung [Anxiety and coping]. Stuttgart, Germany: Kohlhammer. Krohne, H. W. (1996b). Individual differences in coping. In M. Zeidner & N. S. Endler (Eds.), Handbook of coping: Theory, research, applications (pp. 381–409). New York: Wiley. Krohne, H. W. (2003). Individual differences in emotional reactions and coping. In R. J. Davidson, K. R. Scherer, & H. Goldsmith (Eds.), Handbook of affective sciences (pp. 698–725). New York: Oxford University Press. Krohne, H. W., Egloff, B., Varner, L. J., Burns, L. R., Weidner, G., & Ellis, H. C. (2000). The assessment of dispositional vigilance and cognitive avoidance: Factorial structure, psychometric properties, and validity of the Mainz Coping Inventory (MCI). Cognitive Therapy and Research, 24, pp. 297–311.

RT44367_C006.indd 118

7/13/05 9:33:40 AM

VIGILANT AND AVOIDANT COPING

119

Krohne, H. W., Hock, M., & Kohlmann, C. W. (1992). Coping dispositions, uncertainty, and emotional arousal. In K. T. Strongman (Ed.), International review of studies on emotion (Vol. 2, pp. 73–95). Chichester, UK: Wiley. Krohne, H. W., Pieper, M., Knoll, N., & Breimer, N. (2002). The cognitive regulation of emotions: The role of success versus failure experience and coping dispositions. Cognition and Emotion, 16, pp. 217–243. Krohne, H. W., Schmukle, S. C., Burns, L. R., Egloff, B., & Spielberger, C. D. (2001). The measurement of coping in achievement situations: An international comparison. Personality and Individual Differences, 30, pp. 1225–1243. Krohne, H. W., Slangen, K., & Kleemann, P. P. (1996). Coping variables as predictors of perioperative emotional states and adjustment. Psychology and Health, 11, pp. 315–330. Lazarus, R. S. (1983). The costs and benefits of denial. In S. Breznitz (Ed.), The denial of stress (pp. 1–30). New York: International Universities Press. Lazarus, R. S. (1991). Emotion and adaptation. New York: Oxford University Press. Leventhal, H. (1970). Findings and theory in the study of fear communications. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 5, pp. 119–186). New York: Academic Press. Maydeu-Olivares, A., & D’Zurilla, T. J. (1996). A factor-analytic study of the Social Problem-Solving Inventory: An integration of theory and data. Cognitive Therapy and Research, 20, pp. 115–133. Miller, S. M. (1980). When is a little information a dangerous thing? Coping with stressful life events by monitoring versus blunting. In S. Levine & H. Ursin (Eds.), Health and coping (pp. 145–169). New York: Plenum. Miller, S. M. (1987). Monitoring and blunting: Validation of a questionnaire to assess styles of information seeking under threat. Journal of Personality and Social Psychology, 52, pp. 345–353. Miller, S. M. (1996). Monitoring and blunting of threatening information: Cognitive interference and facilitation in the coping process. In I. G. Sarason, G. R. Pierce, & B. R. Sarason (Eds.), Cognitive interference: Theories, methods, and findings (pp. 175–190). Mahwah, NJ: Erlbaum. Miller, S. M., Combs, C., & Kruus, L. (1993). Tuning in and tuning out: Confronting the effects of confrontation. In H. W. Krohne (Ed.), Attention and avoidance: Strategies in coping with aversiveness (pp. 51–69). Seattle and Toronto: Hogrefe & Huber. Miller, S. M., & Kirsch, N. (1987). Sex differences in cognitive coping with stress. In R. C. Barnett, L. Biener, & G. K. Baruch (Eds.), Gender and stress (pp. 278–307). New York: Free Press. Mineka, S. (1985). Animal models of anxiety-based disorders: Their usefulness and limitation. In A. H. Tuma & J. D. Maser (Eds.), Anxiety and the anxiety disorders (pp. 199–244). Hillsdale, NJ: Erlbaum. Moos, R. E. (Ed.). (1986). Coping with life crises: An integrated approach. New York: Plenum. Mullen, B., & Suls, J. (1982). The effectiveness of attention and rejection as coping styles: A meta-analysis of temporal differences. Journal of Psychosomatic Research, 26, pp. 43–49.

RT44367_C006.indd 119

7/13/05 9:33:40 AM

120

H.J. KROHNE AND B. EGLOFF

Orbach, I., & Mikulincer, M. (1996). Repressive coping style and the architecture of emotion-related memories: Is their relationship that simple? British Journal of Social Psychology, 35, pp. 459–471. Patrick, C. J., Cuthbert, B. N., & Lang, P. J. (1994). Emotion in the criminal psychopath: Fear image processing. Journal of Abnormal Psychology, 103, pp. 523–534. Pennebaker, J. W. (1982). The psychology of physical symptoms. New York: Springer-Verlag. Reiss, S., & McNally, R. J. (1985). Expectancy model of fear. In S. Reiss & R. R. Bootzin (Eds.), Theoretical issues in behavior therapy (pp. 107–121). New York: Academic Press. Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, NJ: Princeton University Press. Rost, D. H., & Schermer, F. J. (1989). Diagnostik des leistungsangsterlebens [Test anxiety measurement]. Diagnostica, 35, pp. 287–314. Roth, S., & Cohen, L. J. (1986). Approach, avoidance, and coping with stress. American Psychologist, 41, pp. 813–819. Rothbart, M., & Mellinger, M. (1972). Attention and responsivity to remote dangers: A laboratory simulation for assessing reactions to threatening events. Journal of Personality and Social Psychology, 24, pp. 132–142. Schwarzer, R., & Schwarzer, C. (1996). A critical survey of coping instruments. In M. Zeidner & N. S. Endler (Eds.), Handbook of coping: Theory, research, applications (pp. 107–132). New York: Wiley. Spielberger, C. D. (1983). Manual for the State-Trait Anxiety Inventory (Form Y). Palo Alto, CA: Consulting Psychologists Press. Suls, J., & Fletcher, B. (1985). The relative efficacy of avoidant and nonavoidant coping strategies: A meta-analysis. Health Psychology, 4, pp. 249–288. Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief measures of positive and negative affect: The PANAS scales. Journal of Personality and Social Psychology, 54, pp. 1063–1070. Weidner, G., & Collins, R. L. (1993). Gender, coping, and health. In H. W. Krohne (Ed.), Attention and avoidance. Strategies in coping with aversiveness (pp. 241–265). Seattle and Toronto: Hogrefe & Huber. Weinberger, D. A., Schwartz, G. E., & Davidson, R. J. (1979). Low-anxious, highanxious, and repressive coping-styles: Psychometric patterns and behavioral and physiological responses to stress. Journal of Abnormal Psychology, 88, pp. 369–380. Wilson, J. F. (1985). Stress, coping styles, and physiological arousal. In S. R. Burchfield (Ed.), Stress. Psychological and physiological interactions (pp. 263–281). Washington, DC: Hemisphere. Zeidner, M., & Endler, N. S. (Eds.). (1996). Handbook of coping: Theory, research, applications. New York: Wiley.

RT44367_C006.indd 120

7/13/05 9:33:41 AM

II EXPERIENCE, EXPRESSION, AND CONTROL OF ANGER

RT44367_S002.indd 121

7/13/05 9:28:43 AM

RT44367_S002.indd 122

7/13/05 9:28:43 AM

7 The Expression and Management of Anger and Hostility: Psychological Implications Esther R. Greenglass York University, Toronto, Ontario, Canada Abstract Research on the relationship among anger, hostility, and risk for coronary heart disease (CHD) is reviewed and analyzed from a multidimensional perspective. Psychological implications of the simultaneous experience and control of anger/hostility are examined in relation to risk factors for CHD. The review of research findings indicated the importance of incorporating variables associated with social and interpersonal context in models that examine relations between anger and hostility with CHD. The empirical data also suggested that the management of anger and its impact on the development of illness may be better understood by including gender-role variables and by examining coexisting personality traits and behavioral configurations within a social context.

THE EXPRESSION AND MANAGEMENT OF ANGER AND HOSTILITY: PSYCHOLOGICAL IMPLICATIONS Early research findings showed that the Type-A Behavior Pattern (TABP) predicted coronary heart disease (CHD) independently of traditional risk factors. Later studies established a relationship between anger/hostility, a component of the TABP, and the incidence and mortality of CHD (Barefoot, Dahlstrom, & Williams, 1983; Shekelle, Gale, Ostfeld, & Paul, 1983). However, more recent research has focused on a multidimensional approach to the psychological study of the coronary-prone individual. Simply experiencing anger or hostility may not be the only aspect of these emotional reactions that have disease-related consequences. How emotions are expressed, particularly negative emotions, may also play an important role as a precursor of CHD. Research evidence to date suggests that it is how anger and hostility are expressed, rather than the mere experience of these emotions, that is most relevant to coronary artery (CAD) and cardiovascular disease (CVD). Recent research suggests that an individual’s behavioral expression of anger, relative to the use of nonexpressive angry behavior, is related to cardiovascular risk. Engebretson and Stoney (1995) used an intraindividual approach in which scores on the Anger-In scale of Spielberger’s (1988) State-Trait Anger Expression Inventory (STAXI) were subtracted from STAXI Anger-Out scores 123

RT44367_C007.indd 123

7/13/05 9:29:28 AM

124

Cholesterol level

E.R. GREENGLASS High

Low High Nonexpression

Moderate Nonexpression

Moderate High Expression Expression

Anger expression Figure 7.1. Cholesterol levels as a function of anger expression. Adapted from Engebretson & Stoney (1996).

and were then examined in relation to cholesterol levels. Employing this intraindividual difference approach, they determined the relative dominance of an individual’s Anger-In and Anger-Out scores and their associations with cholesterol levels. The results showed a curvilinear relationship between anger expression and total cholesterol. It should be noted that high values of differences in anger expression scores in the Engebretson and Stoney (1995) study did not necessarily indicate that anger was more frequently expressed in aggressive behavior. Rather, the frequency that anger was expressed in aggressive behavior was relative to the frequency of holding anger in. Furthermore, individuals at both extremes of the anger expression distribution (i.e., almost always expressing anger outwardly or holding anger in) had significantly higher total cholesterol and lipid concentration levels relative to those in the middle range of anger expression scores, as may be noted in Figure 7.1. Engebretson and Stoney (1995) suggested that this middle range of scores indicated a more flexible use of the expression of anger-related behavior. Because extreme styles of anger expression or nonexpression are toxic to health, the message seems to be that individuals should be taught flexibility in how to manage their anger, perhaps even erring on the side of moderate nonexpression. Multiple Emotions and CHD Research findings have increasingly indicated the importance of the simultaneous consideration of the experience and expression of anger, hostility, and other emotions in order to understand the precursors of CHD. For example, in a study of hostility, anger expression, and the severity of coronary disease in patients with coronary atherosclerosis, Dembroski, MacDougal, Williams, Haney, and Blumenthal (1985) found that potential for hostility was only associated with heart disease for those hostile patients who were also high on the suppression of anger, as measured by the STAXI Anger-In (AX/In) scale.

RT44367_C007.indd 124

7/13/05 9:29:29 AM

THE EXPRESSION AND MANAGEMENT OF ANGER AND HOSTILITY

125

Julkunen, Salonen, Kaplan, and Salonen (1992), in a prospective study of middle-aged Finnish men, examined the role of different aspects of anger and hostility as predictors of the progression of carotid atherosclerosis (PCA), using ultrasonographic examination of the carotid artery during a two-year followup. Anger control was measured by the STAXI AX/Con scale, which assesses an individual’s efforts to control angry feelings by inhibiting the open expression of anger in hostile-aggressive behavior (Spielberger, 1988; Spielberger et al., 1985). Hostility was measured with a nine-item Cynical Distrust scale derived by Greenglass and Julkunen (1989) from the Cook-Medley Hostility (Ho) scale. Julkunen et al. (1992) found a fourfold greater acceleration in PCA for participants who were high on both anger-control and any one of the three anger/ hostility measures used in this study: Cynical Distrust, Impatience/Irritability, and the STAXI Anger-Out (AX/Out) scale. Moreover, contrary to expectation, anger control was more important than high scores on the three anger/hostility scales in predicting the progression of atherosclerosis. Thus, greater progression of atherosclerosis was more likely to occur for angry individuals who invested a great deal of energy in monitoring and preventing the expression of their anger, which was driving their intense feelings of cynicism, suspicion, and distrust. Interactive Approaches to the Study of CHD Research has increasingly focused on an interactive approach to the psychological study of CHD. The results of recent studies indicate that it is essential to examine the interaction of emotion and personality with other risk factors in predicting disease proneness, rather than limiting investigation to single traits. Multidimensional analysis of coronary-prone traits provides better predictions of CHD than focusing on single traits or syndromes, such as hostility or the TABP. For example, Fontana et al. (1989) found that hostile coronary patients reported high dependency needs, whereas hostile noncoronary patients did not. These findings suggested that the interaction between hostility and dependency contributed to the etiology of CHD and that emotions and personality characteristics that potentiate the effects of other CHD risk factors appear to be important components of the coronary-prone personality. The findings in a number of studies also provide evidence that anxiety and depression predict CHD (e.g., Booth-Kewley & Friedman, 1987; Carney et al., 1987; Coryell, Noyes, & Hause, 1986; Haines, Imeson, & Meade, 1987; Kawachi et al., 1994) and that anxiety may function as a mediator of the relationship between anger control and CHD. Consistent with this interpretation, in a study of male and female Canadian government supervisors, the author (1996) found that high anxiety was associated with higher scores on measures of Anger-In and Cynical Distrust. Anxiety was assessed with the State Anxiety Scale of Spielberger’s (1983) State-Trait Anxiety Inventory (STAI); Anger-In and Anger Control were measured with the eight-item STAXI AX/In and AX/Con scales (Spielberger, 1988). The Cynical Distrust scale, which was previously described (Greenglass & Julkunen, 1989), was used to measure hostility. Male and female respondents, rank-ordered according to their scores

RT44367_C007.indd 125

7/13/05 9:29:30 AM

126

E.R. GREENGLASS

on the Cynical Distrust and Anger-In scales, were assigned to four experimental groups on the basis of whether their scores on these measures were above or below the median score for the total sample. Study participants who scored high on both Cynical Distrust and Anger-In had significantly higher anxiety scores than their counterparts, whose scores were below the median on one or both of these scales. Similar results were found for both women and men. In the Greenglass and Julkunen (1989) study, individuals with high hostility scores who also made a strong effort to control or suppress their angry feelings were at greater risk for CHD. The increased risk associated with the combination of Cynical Distrust and Anger-In suggested that pathogenic effects resulted from the frequency that a person experienced anger, which was generally greater for individuals who were high in cynicism and had a strong tendency to suppress anger once it was aroused. This process also leads to higher anxiety, associated physiological arousal, and hormonal changes that contribute to the development of CHD. Previous research has also shown that hypertensive patients, as compared with normal controls, have higher levels of suppressed anger and higher anxiety and depression scores (DeQuattro et al., 1981). Thus, the emotional configuration of high Cynical Distrust, High Anger-In, and high anxiety appears to be a psychological precursor of both hypertension and CHD. Reactivity, CHD, and psychological factors. The “Reactivity Hypothesis” proposed by Williams, Barefoot, and Shekelle (1985) is perhaps the most widely cited mechanism for explaining the association between anger/hostility and CHD. According to Williams et al. (1985), hostility contributes to CHD by increasing physiological responses to potential stressors. In responding to environmental stress, individuals with high hostility scores experience greater cardiovascular reactivity and have higher levels of serum catecholamines than those low in hostility. Given this link among personality, excessive cardiovascular activation, and CHD, the final common path to coronary disease clearly involves excessive catecholaminergic reactivity (Obrist, 1981; Wright, Contrada, & Glass, 1985). Recent research findings have also shown that the expression of anger is related to heightened levels of systolic and diastolic blood pressure, and that heightened cardiovascular reactivity (CVR) is associated with CAD and CHD (Manuck, Kaplan, Clarkson, Adams, & Shively, 1992; Manuck, Muldoon, Kaplan, Adams, & Polefrone, 1989; Williams, 1989). Why does the expression and control of anger, and not simply the experience of intense angry feelings, produce high levels of CVR, which result in coronary disease? Angry people raise their voices, accelerate their speech rate, and often interrupt others, which further intensifies the subjective experience of anger and increases blood pressure, heart rate, and levels of cortisol and epinephrine (Siegman, 1994). Furthermore, as suggested by Tavris (1982), most persons who vent their anger get angrier, not less angry. It should be noted, however, that more needs to be learned about the relationship between anger and hostility and cardiovascular reactivity and other coronary disease risk factors before we can conclude that any personality trait, including hostility, increases the risk for CHD (Burns & Katkin, 1992).

RT44367_C007.indd 126

7/13/05 9:29:30 AM

THE EXPRESSION AND MANAGEMENT OF ANGER AND HOSTILITY

127

Hostility, interpersonal stressors, and reactivity. Research findings indicating that anger and hostility are associated with interpersonal conflict and provocation provide evidence that person-by-situation interactions and environmental conditions contribute to increased cardiovascular hyperreactivity. For example, Hardy and Smith (1988) measured systolic and diastolic blood pressure before and during role-play interactions that involved high or low levels of interpersonal conflict. During the high-conflict interactions, study participants with high hostility scores displayed larger increases in diastolic blood pressure than those with low hostility scores. In contrast, during the low-conflict interactions, scores on the hostility measure were not related to blood pressure reactivity. In a similar study, Powch and Houston (1996) examined the effects on cardiovascular reactivity of cynical hostility, aggressiveness, and Anger-In for college women under conditions of high and low interpersonal stress. High stress was created by having participants discuss an issue on which they held a strong opinion with a confederate who adamantly espoused an opposing view. Participants in the low-stress condition discussed an issue about which they had little concern with a confederate who expressed agreement. Only subjects in the high interpersonal stress condition who were high on cynical hostility showed greater systolic blood pressure reactivity than those with lower hostility scores. Thus, in a situation that involved interpersonal conflict and provocation, hostility was associated with greater cardiovascular reactivity. Suarez and Williams (1989) evaluated the effects of stress on the cardiovascular responses of men with high and low hostility scores while they worked on a word-identification task. Subjects in the stressful experimental condition were harassed by the experimenter; those in the control condition received neutral comments about their performance. Heightened cardiovascular reactivity was associated with hostility scores only for the subjects who were harassed. Similar findings were reported by Siegman, Anderson, Herbst, Boyle, and Wilkinson (1992) in a study of male undergraduates who worked on a serial subtraction test, with and without harassment and provocation. Significant positive correlations of anger/hostility scores with systolic and diastolic blood pressure reactivity were found only for participants in the stressful harassment condition. Engebretson, Matthews, and Scheier (1989) also found a significant positive relationship between the outward expression of anger and blood pressure and heart rate reactivity for participants who were harassed in a study in which subjects performed on a task with either a pleasant or harassing confederate. An interactive relationship between anger expression (Anger-In + Anger-Out) and cardiovascular reactivity was reported by Burns and Katkin (1991) for subjects who were harassed during a reaction-time task, but not for those who worked on this same task when social evaluation was emphasized. Participants in this study with high trait anger scores, as measured by Spielberger’s (1988) STAXI T-Anger Scale, reported that they more frequently experienced intense anger in response to a wider range of potential provocations than those subjects with low T-Anger scores (Deffenbacher et al., 1996; Spielberger, Jacobs, Russell, & Crane, 1983). They also reported that they had greater difficulty in

RT44367_C007.indd 127

7/13/05 9:29:30 AM

128

E.R. GREENGLASS

expressing their anger. Individuals who reported that they habitually expressed their anger also responded to harassment and provocation with heightened blood pressure reactivity. In summary, research findings have consistently indicated that confronting others, and experiencing and expressing or suppressing intense emotions, is maladaptive. The discharge of emotions in confrontive coping reactions has also been consistently associated with worsened emotional states (Folkman & Lazarus, 1988; Tavris, 1984). Research findings have further suggested that hostility and the expression of anger make a person feel worse and that persons with depressive symptoms are more likely to experience and express their emotions than those low in depression and to more often engage in confrontive coping behavior (Barnett & Gottlib, 1988; Folkman & Lazarus, 1986). The Social Context of Anger/Hostility and CHD The results previously reported highlight the importance of considering the social context of situations related to the expression of anger and hostility in evaluating the psychophysiological reactivity hypothesis. Regarding the psychological mediators of these effects, Averill (1982) argued that anger and hostility often represent attempts to control the actions of other persons, whereas Price (1982) suggested that hostility is a defense against the anticipated pain of being criticized for being “wrong.” There is also growing evidence that the social context in which anger is experienced can exacerbate the expression of anger and hostility, as well as buffer the individual against such patterns of response. Job stress and hostility. Stress in the workplace can be a major source of frustration and occupational stress. Spector (1987) reported significant positive correlations between excessive workload and work-related interpersonal conflict and frustration, anxiety, and health symptoms. Work stress can also exacerbate cynical distrust and hostility, which, in turn, can trigger higher levels of anxiety (Greenglass & Julkunen, 1989) and occupational stress (Fiksenbaum & Greenglass, 1999). Fiksenbaum (1997) administered the Job Stress Survey (JSS), a self-report questionnaire that assesses occupational stress (Spielberger, 1994; Spielberger & Vagg, 1999), to 640 men and women government employees. The JSS items inquire about the perceived intensity and frequency of occurrence of 30 generic sources of stress that are encountered by managerial, professional, and clerical personnel in a variety of work settings (Spielberger, 1994; Spielberger & Vagg, 1999; Turnage & Spielberger, 1991). Examples of JSS items are “frequent interruptions,” “meeting deadlines,” and “dealing with crisis situations.” Perceived Severity and Frequency scores are obtained for each item. Total Severity and Frequency scores, and a Job Stress Index based on all 30 items, are computed by combining the Severity and Frequency scores for the individual items. Fiksenbaum (1997) found significant positive relationships between the JSS Job Stress Index and Cynical Distrust scores for both males and females (Fiksenbaum & Greenglass, 1999). She also found significant positive correlations for both women and men between Cynical Distrust and Depersonalization,

RT44367_C007.indd 128

7/13/05 9:29:31 AM

THE EXPRESSION AND MANAGEMENT OF ANGER AND HOSTILITY

129

which were assessed with a five-item subscale from the Maslach Burnout Inventory (Maslach & Jackson, 1986), which measured the degree to which respondents treated their clients in a detached and impersonal manner. The more cynical and distrusting the respondents, the more they depersonalized their clients. This finding of a positive relation between cynical distrust and depersonalization was consistent with evidence that hostile individuals are more likely to expect mistreatment from other persons whom they consider as antagonistic, and to attribute hostile intent to them (Smith, 1994). Hence, depersonalization may be seen as a defensive coping technique whereby individuals distance themselves from anyone who is perceived as potentially dangerous or harmful. The transactional model. The results of Fiksenbaum’s (1997) study suggested that hostile individuals were more likely than their less hostile counterparts to interpret the work environment as stressful, threatening, and lacking in social support. Additional findings by Fiksenbaum and Greenglass (1999) showed that hostility and social support were inversely related and that the interaction of hostile persons with their work environment had negative physiological consequences. These findings were consistent with Smith and Pope’s (1990) transactional model, which was founded on the premise that personal reactions and characteristics of the social environment were reciprocally determined. Although hostile individuals respond with heightened reactivity to social stressors that are common to all people, they also react with anger to the additional stressors that they, themselves, have created (Smith, 1994). The cognitive consequences of hostility have important implications for the experience of stress. According to Smith’s (1994) transactional model, chronically angry and hostile persons contribute to shaping the negative characteristics of their social environment through their thoughts and actions. By mistrusting others and expecting mistreatment, and by attributing hostile intent to others, hostile persons are prone to express antagonistic and aggressive actions, which are likely to undermine social support. Thus, through their thoughts and actions, chronically hostile individuals provoke a nonsupportive environment that is likely to maintain and enhance their anger and hostility. Figure 7.2 presents a schematic representation of the transactional model, which is based on the assumption that personal reactions and the social environment are reciprocally determined. According to this model, the interrelationships between job stress, anxiety, hostility, and depersonalization contribute to an unsupportive work environment and diminished social support from supervisors and coworkers. Social Support and Health Social support plays an important role in how individuals deal with environmental stress and in mediating the relationship of stress with anger and health. Research on social support (e.g., Sarason, Sarason, Brock, & Pierce, 1996) has previously focused on three major areas: (1) how differences in interpersonal connectedness influence reactions to stressful situations, (2) identification of

RT44367_C007.indd 129

7/13/05 9:29:31 AM

130

E.R. GREENGLASS Job stress

Anxiety

Hostility

Depersonalization

Unsupportive environment

Figure 7.2.

Interrelationships between job stress, hostility, depersonalization and social support.

supportive components of the environment, and (3) how a person’s sense of being is supported. Research findings underline the critical role of social support in the maintenance of health and provide evidence that perceived support is positively linked with physical and mental health, and the ability to cope with stress (Sarason et al., 1996). Kaplan, Berkman, and Breslow (1983), who followed several thousand healthy California residents for a number of years, found that individuals with fewer social ties had higher death rates than those with close connections to others. Research findings also indicate that social support can influence well-being by serving as a buffer against stress and by directly improving feelings of well-being (Cohen & Wills, 1985; Greenglass, 1993; Hobfoll, 1986). According to the buffering hypothesis, although stress adversely affects some people, those with strong social support resources are more resistant to the deleterious effects of stressful events. Although the direct effects of social support are documented by negative correlations between measures of social support and stress, it should be noted that evidence for the buffering hypothesis is mixed (Himle, Jayaratne, & Thyness, 1991). Evidence of the positive effects of social support on problems created by occupational stress was reported by LaRocco, House, and French (1980), who found that support from supervisors and coworkers reduced the psychological strain related to workplace stressors. However, no buffering effects of social support were found in other studies (Himle, Jayaratne, & Thyness, 1989; Linn, Simeone, Ensel, & Kuo, 1979; Shinn, Rosario, March, & Chestnut, 1984). These contradictory results suggest that the buffering effects of social support in work settings may depend on the social context defining the occupational stressors and on the nature of the available social support.

RT44367_C007.indd 130

7/13/05 9:29:32 AM

THE EXPRESSION AND MANAGEMENT OF ANGER AND HOSTILITY

131

The findings of Greenglass et al. (1996) also indicated that practical and informational support within the work environment functioned as a buffer to experiencing depersonalization. Social support on the job that provides practical advice and information can assist workers by helping them to structure their workload so that it more closely matches their individual resources. In the absence of such support, job demands may exceed the individual’s resources, leading to frustration, anger and hostility, depersonalization, and increased levels of occupational stress. Beehr and McGrath (1992) have observed that most research on job stress and social support has benefited from the contextual realism of employing nonexperimental field methods. However, these investigators also point out that when research relies primarily on nonexperimental approaches, inferences regarding causality rest primarily on theory and reasoned argument, rather than on methodologically strong evidence. Because these observations are clearly relevant to the study of the effects of anger, hostility, and social support on CHD, research studies that will allow investigators to gauge more accurately the direction of causality are needed. Social support, anger/hostility, and buffering. Research findings indicate that strong social support is associated with lower hostility and anger (Smith & Pope, 1990; Thomas, 1989) and that social support can function as a buffer of anger in managerial women (Greenglass, 1987). According to Smith and Pope (1990), hostile individuals tend to evoke high levels of interpersonal conflict while, at the same time, undermining potential sources of social support. Barefoot et al. (1983) found that higher hostility scores were associated with a lower quality of social support. Managers with stronger family support were found by the author (1991) to have less Anger-In. Rather than creating an environment where sources of frustration and anger can be dealt with constructively, hostile individuals create nonsupportive situations that maintain their level of hostility and distrust. As a consequence of expressing their anger, hostile individuals also reduce the possibility of receiving helpful information, practical assistance, and emotional support from others who could potentially help them to deal more constructively with their frustration and anger. In a study of teacher burnout, Greenglass, Fiksenbaum, and Burke (1996) examined the buffering effects of social support on emotional exhaustion, depersonalization, and lack of accomplishment, as measured by three subscales from the Maslach Burnout Inventory (Maslach & Jackson, 1986). The effects of informational, practical, and emotional support from supervisors, coworkers, family, and friends were assessed by a modified version of the Caplan, Cobb, French, Van Harrison, and Pinneau (1975) scale. Greenglass et al. (1996) found that emotional support from coworkers, family, and friends, which involved listening to the teachers’ problems and boosting their morale, was an important buffer of depersonalization when stress was high. To the extent that the teachers perceived receiving emotional support from those close to them when there was a great deal of stress at work, they were less likely to depersonalize their students. Because depersonalization covaried significantly with hostility, emotional support may also have served as a buffer of hostility. These findings suggested that when teachers were able to turn to others and

RT44367_C007.indd 131

7/13/05 9:29:32 AM

132

E.R. GREENGLASS

received emotional “boosting” at times of high stress, they were less likely to act in a hostile, detached, and impersonal manner. Anger, Hostility, and Coping Understanding the effective management of anger and hostility, and how the experience, expression, and control of anger may be related to illness, requires a conceptual integration of the dynamics of anger expression with the process of coping. Research has identified two critical dimensions of the ways in which people cope that appear to be related to how they deal with anger and hostility: (1) the extent to which an individual engages in suppression versus the disclosure and expression of feelings and thoughts; and (2) how often an individual uses active or passive coping techniques. Suppression versus disclosure. A psychodynamic interpretation of the benefits of coping considers the disclosure of thoughts and feelings as a key aspect of the coping process. Pennebaker and Susman (1988) discuss the benefits that follow from the disclosure of past traumatic events. Coping that involves disclosure is considered beneficial because it eliminates the adverse effects of the behavioral inhibition that results from not disclosing information or feelings to others. The process of behavioral inhibition may lead to healthrelated problems (Pennebaker, 1989), as demonstrated by research in which thought suppression was found to generate negative affect and increased anxiety (Roemer & Borkovec, 1993). Similar findings have also been reported regarding the negative effects of anger suppression. For example, in a study that examined the relationship between the experience and expression of anger and other mental health variables, Kopper and Epperson (1996) found that anger suppression correlated positively with depression, resentment, suspiciousness, guilt, and conflict avoidance, and negatively with assertiveness and self-confidence. There is also evidence that high anger suppression is inversely related to social support. For example, the author (1996) found that individuals high in cynical distrust, who turned anger in to suppress and control their angry impulses, reported less family support and interpersonal trust than persons with lower scores on measures of Cynical Distrust and Anger-In. These findings suggest that hostile individuals who suppress and do not disclose their angry thoughts and feelings are suspicious of others and lower in interpersonal trust, and consequently receive less social support. In contrast to the negative effects of anger suppression, dispositional optimism and self-efficacy may mediate disclosure, and thus facilitate more effective dealing with angry feelings. Ausbrooks, Thomas, and Williams (1995) investigated the relationship of trait anger and modes of anger expression to dispositional optimism and self-efficacy in college students. Their results showed that dispositional optimism and self-efficacy correlated positively with a tendency to express anger through discussion and negatively with the tendency to suppress anger. These investigators concluded that, when anger is aroused, individuals with a strong sense of self-efficacy and dispositional optimism can handle their anger more effectively.

RT44367_C007.indd 132

7/13/05 9:29:32 AM

THE EXPRESSION AND MANAGEMENT OF ANGER AND HOSTILITY

133

Active versus passive coping. The distinction between suppression and disclosure overlaps with concepts of active versus passive coping. Research findings indicate that persons who rely on passive coping and cognitive avoidance tend to report greater depression and more physical symptoms, have drinking problems (Moos, 1988), and experience problematic interpersonal behavior, including hostility and aggression, and lack of empathy (Kobak & Sceery, 1989). Passive coping is also associated with less access to social support. According to Moos (1988), respondents with fewer social resources rely more heavily on various forms of passive coping, including cognitive and behavioral avoidance, resigned acceptance, and emotional discharge, which are all positively correlated with anxiety. These passive and avoidant coping strategies tend to be inversely correlated with active coping and planning, and other positive coping strategies (Carver, Scheier, & Weintraub, 1989). Communicative Expression of Anger and Hostility Research findings suggest that the use of disclosure and active coping techniques, including access to and the utilization of social support resources, leads to communicative and nonthreatening expression of anger and hostility (Stoney & Engebretson, 1994). Consistent with Thoits’s (1986) concept of social support as a source of coping assistance, research in Spain has shown that people who were able to access social support resources, especially from friends, relied more heavily on logical analyses and cognitive redefinition (Paez, Basabe, Valdoseda, Velasco, & Iraurgi, 1995). Anger-provoking situations were thus redefined or restructured so as to be less threatening, thereby leading to lower levels of anger. Research has also shown that individuals who displayed health benefits following disclosure subsequently demonstrated increased insight and cognitive restructuring as compared with those who did not experience improved health (Pennebaker, 1993). Social sharing and talking about feelings, including anger, might also contribute to solving problems and lessening emotional distress. Through discussion with others, anger and hostility may be somewhat diffused and reduced. Discussion also provides a forum whereby practical, informational, and emotional support can be provided by others. It is suggested here that such support can be helpful in redefining or restructuring situations as less threatening, thereby leading to lower levels of anger and hostility. When individuals are highly hostile, the probability is considerably lower that they would benefit from social support in the ways described above. Given the self-involvement of hostile persons, they most often pay little attention to what others are saying or doing (Williams & Williams, 1993). This increased focus on the self, with little attention given to information gathering or seeking advice, may further distance the hostile individual from other persons. Anger and hostility also repel family and friends who might engage in beneficial discussion and provide social support. The degree to which individuals are willing to disclose their thoughts and feelings may also function as a moderator of social support. To the extent that individuals can disclose their thoughts and feelings to others, there is a great-

RT44367_C007.indd 133

7/13/05 9:29:33 AM

134

E.R. GREENGLASS Anger experience Social support

Practical Information Emotional

Cognitive restructuring

Anger-in

Anger-out

Anger-communicate

Figure 7.3. Experience and expression of anger.

er opportunity for problem solving, gaining insight, cognitively restructuring their thoughts, and lessening emotional distress, which should also reduce the intensity of their angry feelings and the expression of anger. Figure 7.3 presents a schematic representation of the ways in which anger can be expressed within a social and cognitive context. Anger, Hostility, and Gender Differences Issues relating to the mental health implications of the experience, expression, suppression, and control of anger raise questions in regard to whether men and women differ in their experience of angry feelings and how their feelings are expressed in aggressive behavior. Research findings generally indicate that men and women experience similar levels of anger. For example, in research on individual differences in trait anger using Spielberger’s T-Anger Scale (Engebretson & Matthews, 1992; Girdler, Turner, Sherwood, & Light, 1990), males and females reported experiencing similar amounts of trait anger. In a large national survey, Barefoot et al. (1991) also found that males and females had similar trait anger scores across ages, ranging from 18 to 90 years. It is generally believed that women have greater difficulty than men in expressing their anger, thereby resulting in greater anger suppression in women. These gender differences are attributed to factors associated with the feminine role, which is seen as antithetical to the expression of anger or aggression. Although most research does not support this belief, middle-aged women in the Framingham Heart Study were more likely to inhibit anger than their male counterparts (Haynes & Feinleib, 1980). However, Greenglass and Julkunen (1989) and Thomas and Williams (1990) found no significant gender differences in the anger expression scores of college students. In a study of middle-aged men and women, no gender differences were found in either Anger-In or Anger-Out (Engebretson & Matthews, 1992).

RT44367_C007.indd 134

7/13/05 9:29:34 AM

THE EXPRESSION AND MANAGEMENT OF ANGER AND HOSTILITY

135

There is, however, evidence that males are more likely than females to express anger in aggressive behaviors. Eagly and Steffen (1986), in a metaanalysis review of the literature, found that males were more likely than females to engage in aggressive behaviors, but the magnitude of gender differences was associated with the characteristics of the aggression being studied. Interpretations regarding appropriate anger expression also tend to differ for women and men. Women who express anger without restraint are more likely to be labeled negatively than men who express anger in a similar manner (Forgays & Hatch, 1996). Aggression and competitiveness are approved and socially sanctioned as masculine coping styles (Eisler & Blalock, 1991), and masculinity has also been linked with aggressive acting-out behavior (Kopper & Epperson, 1996). Male socialization encourages men to engage in more aggressive activities in order to validate their masculinity, which may contribute to the higher male death rates in accidents of all kinds (Eisler & Blalock, 1991). The greater vulnerability of males for death by suicide or homicide may also be due, at least partially, to a greater predisposition for them to engage in aggressive behavior. Research indicates that suppression of anger in both men and women is associated with depression, dependency, guilt, and conflict avoidance, but anger suppression appears to be more strongly associated with CHD in women. Haynes, Feinleib, and Kannel (1980) found that suppressed hostility was a significant predictor of CHD incidence in white-collar men and working women. However, when the data were adjusted for standard CHD risk factors, the contribution of anger suppression to the prediction of CHD remained significant for women, but not for men. Thus, anger suppression was a CHD risk factor for women, but not for men. Research findings also suggest that gender mediates the expression of hostility. In a study by Greenglass and Julkunen (1989), cynical distrust correlated positively with Anger-In for women and with Anger-Out for their male counterparts. These findings suggested that, for men, cynicism and hostility were more likely to be expressed through Anger-Out, whereas women expressed anger and hostility primarily by turning anger in. This interpretation was supported by Kopper (1993), who found higher indirect hostility scores for women and higher scores for men on measures of assault and aggressiveness. Communicative expression and gender. Females appear to express their anger in a communicative fashion to a greater degree than males and are less likely to express anger in aggressive behavior (Stoney & Engebretson, 1994). The results of a survey by Averill (1982), based on a combined community and undergraduate sample, indicated that females reported more frequently than males that they would like to talk over an angering situation with the individual involved. In a midlife sample, Thomas (1989) also found that women expressed a stronger desire than men to discuss their anger. Similarly, Harburg, Blakelock, and Roeper (1979), in a survey of 744 adults, found that females reported significantly more reflective problem-solving responses than males to a hypothetical scenario involving an angry boss. Communicative expression of anger may be used as a positive coping strategy that helps individuals deal constructively with their angry feelings. Moreover, as

RT44367_C007.indd 135

7/13/05 9:29:34 AM

136

E.R. GREENGLASS

previously noted, anger and hostility may be somewhat diffused through discussion with others, and active coping strategies, including social support, generally facilitate the communicative nonthreatening expression of anger. Previous research findings also suggest that social support linked with coping can be helpful in redefining or restructuring stressful situations to make them less threatening and that women are better able to utilize social support from others in developing effective instrumental and preventive coping strategies. The author (1993) examined the role of support by supervisors, family, and friends in predicting the coping strategies used by men and women managers. The results of regression analyses indicated that supervisor support for women was positively and significantly related to preventive and instrumental coping, and that support by friends and family was also positively and significantly associated with the development of preventive coping. In contrast, only supervisor support predicted preventive coping in men. Additional findings from this study indicated that, for women, support from friends and family was negatively related to palliative coping, which included self-blame and wishful thinking. Thus, women were able to use social support to lessen their reliance on coping strategies associated with greater job-related anxiety, depression, and somatization (Greenglass, 1993). Research findings also suggest that women are more likely than men to employ communicative nonthreatening forms of anger expression because their social support is linked to these more effective forms of coping. The communicative, nonthreatening expression of anger by women is consistent with their more effective utilization of various forms of social support, including information, advice, and guidance. CONCLUSIONS Understanding the relationship between anger, hostility, and CHD requires a multidimensional approach that simultaneously considers more than one emotion. Although past research tended to focus on the relationship between a single emotion, such as hostility or anger, and the incidence and severity of CHD, more recent research has focused on the interaction of hostility, social support, and how anger is managed. Recent research findings have also indicated that it is essential to study interactions of personality and emotion with other risk factors in predicting disease proneness, rather than limiting investigation to single traits. Multidimensional analyses of emotions and other coronary-prone traits provide better predictors of CHD than focusing on single emotions, traits, or syndromes. Recent research findings also point to the importance of how the social context interacts with individual differences in emotions, personality, and other risk factors to trigger the heightened physiological reactivity associated with CHD. The critical role of social support is also increasingly recognized as an important contributor to the development of effective communicative ways of coping with anger. Moreover, research findings also point to the importance of taking into account the social context of situations that stimulate the experience, expression, and control of anger and hostility.

RT44367_C007.indd 136

7/13/05 9:29:34 AM

THE EXPRESSION AND MANAGEMENT OF ANGER AND HOSTILITY

137

Examination of gender-related effects in the prediction of CHD requires taking into account the multidimensionality of the experience, expression, and control of anger and hostility. Consistent differences have been observed in the ways women and men communicate their feelings and manage their anger that are linked to gender-related obligations and expectations, and social roles. Research also shows that women, more than men, are likely to employ communicative nonthreatening forms of anger expression because their social support is often linked to more effective methods of coping. These findings underline the importance of examining the social context in order to understand how the experience and control of anger are related to anger expression.

REFERENCES Ausbrooks, E. P., Thomas, S. P., & Williams, R. L. (1995). Relationships among self-efficacy, optimism, trait anger, and anger expression. Health Values: The Journal of Health Behavior, Education and Promotion, 19, pp. 46–54. Averill, J. R. (1982). Anger and aggression: An essay on emotion. New York: Springer-Verlag. Barefoot, J. C., Dahlstrom, W. G., & Williams, R. B., Jr. (1983). Hostility, CHD incidence, and total mortality. A 25-year follow-up study of 255 physicians. Psychosomatic Medicine, 45, pp. 59–63. Barefoot, J. C., Peterson, B. L., Dahlstrom, W. G., Siegler, I. C., Anderson, N. B., & Williams, R. B., Jr. (1991). Hostility patterns and health implications: Correlates of Cook-Medley Hostility Scale scores in a national survey. Health Psychology, 10, pp. 18–24. Barnett, P. A., & Gottlib, I. H. (1988). Psychosocial functioning and depression: Distinguishing among antecedents, concomitants, and consequences. Psychological Bulletin, 104, pp. 97–126. Beehr, T. A., & McGrath, J. E. (1992). Social support, occupational stress, and anxiety. Anxiety, Stress, and Coping, 5, pp. 7–19. Booth-Kewley, S., & Friedman, H. S. (1987). Psychological predictors of heart disease: A quantitative review. Psychological Bulletin, 101, pp. 343–362. Burns, J. W., & Katkin, E. S. (1991). Anger, hostility, anxiety, and cardiovascular reactivity to different stress conditions. Paper presented at the meeting of the Society for Behavioral Medicine, Washington, DC, March. Burns, J. W., & Katkin, E. S. (1992). Commentary to part two: Hostility and the coronary-prone personality. In H. S. Friedman (Ed.), Hostility, coping and health. Washington, DC: American Psychological Association. Caplan, R. D., Cobb, S., French, J. R. P., Jr., Van Harrison, R., & Pinneau, S. P. (1975). Job demands and worker health: Main effects and occupational differences. Washington, DC: U.S. Government Printing Office. Carney, R. M., Rich, M. W., teVelde, A., Saini, J., Clark, K., & Jaffe, A. S. (1987). Major depressive disorder in coronary artery disease. American Journal of Cardiology, 60, pp. 1273–1275.

RT44367_C007.indd 137

7/13/05 9:29:35 AM

138

E.R. GREENGLASS

Carver, C. S., Scheier, M. F., & Weintraub, J. K. (1989). Assessing coping strategies: A theoretically based approach. Journal of Personality and Social Psychology, 56, pp. 267–283. Cohen, S., & Wills, T. (1985). Stress, social support and the buffering hypothesis. Psychological Bulletin, 98, pp. 310–357. Coryell, W., Noyes, R., & Hause, J. D. (1986). Mortality among outpatients with anxiety disorders. American Journal of Psychiatry, 143, pp. 508–510. Deffenbacher, J. L., Oetting, E. R., Thwaites, G. A., Lynch, R. S., Baker, D. A., Stark, R. S., Thacker, S., & Eiswerth-Dox, L. (1996). State-trait anger theory and the utility of the Trait Anger Scale. Journal of Counseling Psychology, 43, pp. 131–148. Dembroski, T. M., MacDougal, J. M., Williams, R. B., Haney, T. L., & Blumenthal, J. A. (1985). Components of Type A, hostility, and Anger-In: Relationship to angiographic findings. Psychosomatic Medicine, 47, pp. 219–233. DeQuattro, V., Sullivan, P., Foti, A., Schoentgen, S., Kolloch, R., Verasales, G., & Levine, D. (1981). Central neurogenic mechanisms in hypertension and in postural hypotension. In G. Laragh, F. Buhler, & D. Seldin (Eds.), Frontiers in hypertension research. New York: Springer-Verlag. Eagly, A. H., & Steffen, V. J. (1986). Gender and aggressive behavior: A meta-analytic review of the social psychological literature. Psychological Bulletin, 100, pp. 309–330. Eisler, R. M., & Blalock, J. A. (1991). Masculine gender role stress: Implications for the assessment of men. Clinical Psychology Review, 11, pp. 45–60. Engebretson, T. O., & Matthews, K. A. (1992). Dimensions of hostility in men, women, and boys: Relationships to personality and cardiovascular responses to stress. Psychosomatic Medicine, 54, pp. 311–323. Engebretson, T. O., Matthews, K. A., & Scheier, M. F. (1989). Relations between anger expression and cardiovascular reactivity: Reconciling inconsistent findings through a matching hypothesis. Journal of Personality and Social Psychology, 57, pp. 513–521. Engebretson, T. O., & Stoney, C. M. (1995). Anger expression and lipid concentrations. International Journal of Behavioral Medicine, 2, pp. 281–298. Fiksenbaum, L. (1997). Social support and coping as moderators of workrelated stress and burnout. Unpublished master’s thesis, York University, Toronto, Canada. Fiksenbaum, L., & Greenglass, E. R. (1999). Job stress, hostility, and social support in government workers. Unpublished manuscript. Toronto, Canada: York University. Folkman, S., & Lazarus, R. S. (1986). Estrés y procesos cognitivos [Stress and cognitive processes]. Barcelona, Spain: Martinez Roca. Folkman, S., & Lazarus, R. S. (1988). The relationship between coping and emotion: Implications for theory and research. Social Science Medicine, 26, pp. 309–317. Fontana, A. F., Kerns, R. D., Blatt, S. J., Rosenberg, R. L., Burg, M. M., & Colonese, K. L. (1989). Cynical mistrust and the search for self-worth. Journal of Psychosomatic Research, 33, pp. 449–456.

RT44367_C007.indd 138

7/13/05 9:29:35 AM

THE EXPRESSION AND MANAGEMENT OF ANGER AND HOSTILITY

139

Forgays, D. K., & Hatch, H. (1996). The female experience and expression of anger. Paper presented at the European Health Psychology Society Conference, Dublin, Ireland, September. Girdler, S. S., Turner, J. R., Sherwood, A., & Light, K. C. (1990). Gender differences in blood pressure control during a variety of behavior stressors. Psychosomatic Medicine, 52, pp. 571–591. Greenglass, E. (1987). Anger in Type-A women: Implications for coronary heart disease. Personality and Individual Differences, 8, pp. 639–650. Greenglass, E. (1991). Burnout and gender: Theoretical and organizational implications. Invited paper for Canadian Psychology, 32, pp. 562–572. Greenglass, E. (1993). Social support and coping of employed women. In B. C. Long & S. E. Kahn (Eds.), Women, work and coping: A Multidisciplinary approach to workplace stress (pp. 215–239). Montreal: McGill-Queens University Press. Greenglass, E. (1996). Anger suppression, cynical distrust, and hostility: Implications for coronary heart disease. In C. D. Spielberger & I. G. Sarason (Eds.), Stress and emotion: Anxiety, anger, and curiosity (Vol. 16, pp. 205–225). Washington, DC: Taylor & Francis. Greenglass, E., Fiksenbaum, L., & Burke, R. J. (1996). Components of social support, buffering effects, and burnout: Implications for psychological functioning. Anxiety, Stress, and Coping, 9, pp. 185–197. Greenglass, E., & Julkunen, J. (1989). Construct validity and sex differences in Cook-Medley hostility. Personality and Individual Differences, 10, pp. 209–218. Haines, A. P., Imeson, J. D., & Meade, T. W. (1987). Phobic anxiety and ischaemic heart disease. British Medical Journal, 295, pp. 297–299. Harburg, E., Blakelock, E. H., & Roeper, P. J. (1979). Resentful and reflective coping with arbitrary authority and blood pressure: Detroit. Psychosomatic Medicine, 41, pp. 189–202. Hardy, J. D., & Smith, T. W. (1988). Cynical hostility and vulnerability to disease: Social support, life stress, and physiological response to conflict. Health Psychology, 7, pp. 447–459. Haynes, S. G., & Feinleib, M. (1980). Women, work, and coronary heart disease: Prospective findings from the Framingham heart study. American Journal of Public Health, 70, pp. 133–141. Haynes, S. G., Feinleib, M., & Kannel, W. B. (1980). The relationship of psychosocial factors to coronary heart disease in the Framingham Study. III. Eight-year incidence of coronary heart disease. American Journal of Epidemiology, 3, pp. 37–58. Himle, D. P., Jayaratne, S., & Thyness, P. A. (1989). The buffering effects of four types of supervisory support on work stress. Administration in Social Work, 13, pp. 19–34. Himle, D. P., Jayaratne, S., & Thyness, P. (1991). Buffering effects of four social support types on burnout among social workers. Social Work Research and Abstracts, 27, pp. 22–27. Hobfoll, S. E. (1986). Stress, social support, and women. Washington, DC: Hemisphere.

RT44367_C007.indd 139

7/13/05 9:29:35 AM

140

E.R. GREENGLASS

Julkunen, J., Salonen, R., Kaplan, G. A., & Salonen, J. T. (1992). Hostility and the progression of carotid atherosclerosis. Paper presented at the 25th International Congress of Psychology, Brussels, Belgium, July. Kaplan, G., Berkman, L. F., & Breslow, L. (1983). Health and Ways of Living: The Alameda County Study. New York: Oxford University Press. Kawachi, I., Colditz, G. A., Ascherio, A., Rimm, E. B., Giovannucci, E., Stampfer, M. J., & Willett, W. C. (1994). Prospective study of phobic anxiety and risk of coronary heart disease in men. Circulation, 89, 1992–1997. Kobak, R., & Sceery, A. (1989). Attachment in late adolescence: Working models, affect regulation, and perceptions of self and others. Child Development, 59, pp. 135–146. Kopper, B. A. (1993). Role of gender, sex role identity, and Type-A behavior in anger expression and mental health functioning. Journal of Counseling Psychology, 40, pp. 232–237. Kopper, B. A., & Epperson, D. L. (1996). The experience and expression of anger: Relationships with gender, gender role socialization, depression, and mental health functioning. Journal of Counseling Psychology, 43, pp. 158–165. LaRocco, J. M., House, J. S., & French, J. R. P., Jr. (1980). Social support, occupational stress, and health. Journal of Health and Social Behavior, 21, pp. 202–218. Linn, N. R. S., Simeone, R. L., Ensel, W. M., & Kuo, W. (1979). Social support, stressful life events, and illness: A model and an empirical test. Journal of Health and Social Behavior, 20, pp. 108–119. Manuck, S. B., Kaplan, J. R., Clarkson, T. B., Adams, M. R., & Shively, C. R. (1992). Behavioral influences on coronary artery disease: A nonhuman primate model. In H. S. Friedman (Ed.), Hostility, coping, and health. Washington, DC: American Psychological Association. Manuck, S. B., Muldoon, M. F., Kaplan, J. R., Adams, M. R., & Polefrone, J. M. (1989). Coronary artery atherosclerosis and cardiac response to stress in cynomolgus monkeys. In A. W. Siegman & T. M. Dembroski (Eds.), In search of coronary-prone behavior: Beyond Type-A (pp. 207–227). Hillsdale, NJ: Erlbaum. Maslach, C., & Jackson, S. E. (1986). Maslach Burnout Inventory Manual (2nd ed.). Palo Alto, CA: Consulting Psychologists Press. Moos, R. H. (1988). Life stressors and coping resources influence health and well-being. Psychological Assessment, 4, pp. 133–158. Obrist, P. A. (1981). Cardiovascular psychophysiology: A perspective. New York: Plenum Press. Paez, D., Basabe, N., Valdoseda, M., Velasco, C., & Iraurgi, I. (1995). Confrontation: Inhibition, alexithymia, and health. In J. W. Pennebaker (Ed.), Emotion, disclosure, and health (pp. 195–222). Washington, DC: American Psychological Association. Pennebaker, J. W. (1989). Stream of consciousness and stress: Levels of thinking. In J.S. Uleman and J. Bargh (Eds.), Unintended thought. New York: Guilford Press.

RT44367_C007.indd 140

7/13/05 9:29:36 AM

THE EXPRESSION AND MANAGEMENT OF ANGER AND HOSTILITY

141

Pennebaker, J. W. (1993). Putting stress into words: Health, linguistic, and therapeutic implications. Behaviour Research and Therapy, 31, pp. 539–548. Pennebaker, J. W., & Susman, J. R. (1988). Disclosure of traumas and psychosomatic processes. Social Science and Medicine, 26, pp. 327–332. Powch, I. G., & Houston, B. K. (1996). Hostility, Anger-In, and cardiovascular reactivity in white women. Health Psychology, 15, pp. 200–208. Price, V. A. (1982). Type-A behavior pattern: A model for research and practice. New York: Academic Press. Roemer, L., & Borkovec, T. D. (1993). Worry: Unwanted cognitive activity that controls unwanted somatic experience. In D. M. Wegner & J. Pennebaker (Eds.), Handbook of mental control (pp. 220–238). Englewood Cliffs, NJ: Prentice Hall. Sarason, I. G., Sarason, B. R., Brock, D. M., & Pierce, G. R. (1996). Social support: Current status, current issues. In C. D. Spielberger & I. G. Sarason (Eds.), Stress and emotion: Anxiety, anger, and curiosity (Vol. 16, pp. 3–27). Washington, DC: Taylor & Francis. Shekelle, R. B., Gale, M., Ostfeld, A. M., & Paul, O. (1983). Hostility, risk of coronary heart disease, and mortality. Psychosomatic Medicine, 45, pp. 109–114. Shinn, M., Rosario, M., March, H., & Chestnut, D. (1984). Coping with job stress and burnout in the human services. Journal of Personality and Social Psychology, 46, pp. 864–876. Siegman, A. W. (1994). Cardiovascular consequences of expressing and repressing anger. In A. W. Siegman & T. W. Smith (Eds.), Anger, hostility, and the heart (pp. 173–197). Hillsdale, NJ: Erlbaum. Siegman, A. W., Anderson, R., Herbst, J., Boyle, S., & Wilkinson, J. (1992). Dimensions of anger-hostility and cardiovascular reactivity in provoked and angered men. Journal of Behavioral Medicine, 15, pp. 257–272. Smith, T. W. (1994). Concepts and methods in the study of anger, hostility, and health. In A. W. Siegman & T. W. Smith (Eds.), Anger, hostility, and the heart (pp. 23–42). Hillsdale, NJ: Erlbaum. Smith, T. W., & Pope, M. K. (1990). Cynical hostility as a health risk: Current status and future directions. Journal of Social Behavior and Personality, 5, pp. 77–88. Spector, P. E. (1987). Interactive effects of perceived control and job stressors on affective reactions and health outcomes for clerical workers. Work and Stress, 1, pp. 155–162. Spielberger, C. D. (1983). Manual for the State-Trait Anxiety Inventory (Form Y). Palo Alto, CA: Consulting Psychologists Press. Spielberger, C. D. (1988). State-Trait Anger Expression Inventory (STAXI). Odessa, FL: Psychological Assessment Resources. Spielberger, C. D. (1994). Preliminary Manual for the Job Stress Survey (JSS). Tampa, FL: University of South Florida. Spielberger, C. D., Jacobs, G. A., Russell, S. F., & Crane, R. J. (1983). Assessment of anger: The State-Trait Anger Scale. In J. N. Butcher & C. D. Spielberger (Eds.), Advances in personality assessment (Vol. 2, pp. 159–187). Hillsdale, N.J.: Erlbaum.

RT44367_C007.indd 141

7/13/05 9:29:36 AM

142

E.R. GREENGLASS

Spielberger, C. D., Johnson, E. H., Russell, S., Crane, R., Jacobs, G., & Worden, T. (1985). The experience and expression of anger: Construction and validation of an anger expression scale. In M. A. Chesney & R. H. Rosenman (Eds.), Anger and hostility in cardiovascular and behavioral disorders (pp. 5–29). New York: Hemisphere/McGraw-Hill. Spielberger, C. D., & Vagg, P. R. (1999). Professional manual for the Job Stress Survey (JSS). Odessa, FL: Psychological Assessment Resources. Stoney, C. M., & Engebretson, T. O. (1994). Anger and hostility: Potential mediators of the gender difference in coronary heart disease. In A. W. Siegman & T. W. Smith (Eds.), Anger, hostility, and the heart (pp. 215–237). Hillsdale, NJ: Erlbaum. Suarez, E. C., & Williams, R. B. (1989). Situational determinants of cardiovascular and emotional reactivity in high and low hostile men. Psychosomatic Medicine, 51, pp. 404–418. Tavris, C. (1982). Anger, the misunderstood emotion. New York: Simon & Schuster. Tavris, C. (1984). On the wisdom of counting to ten. Personal and social dangers of anger expression. In P. Shaver (Ed.), Emotions, relationships and health. Beverly Hills, CA: Sage. Thoits, P. (1986). Social support as coping assistance. Journal of Consulting and Clinical Psychology, 54, pp. 416–423. Thomas, S. P. (1989). Gender differences in anger expression: Health implications. Research in Nursing and Health, 12, pp. 389–398. Thomas, S. P., & Williams, R. L. (1990). Gender differences in modes of anger expression: Health implications. Paper presented at the 11th Annual Meeting of the Society of Behavioral Medicine, Chicago, April. Turnage, J. J., & Spielberger, C. D. (1991). Job stress in managers, professionals, and clerical workers. Work and Stress, 5, pp. 165–176. Williams, R. B., Jr. (1989). Biological mechanisms mediating the relationship between behavior and coronary heart disease. In A. W. Siegman & T. M. Dembroski (Eds.), In search of coronary-prone behavior: Beyond Type-A. Hillsdale, NJ: Erlbaum. Williams, R. B., Jr., Barefoot, J. C., & Shekelle, R. B. (1985). The health consequences of hostility. In M. A. Chesney & R. H. Rosenman (Eds.), Anger and hostility in cardiovascular and behavioral disorders (pp. 173–185). Washington, DC: Hemisphere. Williams, R. B., Jr., & Williams, V. (1993). Anger kills: Seventeen strategies for controlling the hostility that can harm your health. New York: Random House. Wright, R. A., Contrada, R. J., & Glass, D. C. (1985). Psychophysiological correlates of Type-A behavior. In E. S. Katkin & S. B. Manuck (Eds.), Advances in behavioral medicine (Vol. 1, pp. 39–88). Greenwich, CT: JAI Press.

RT44367_C007.indd 142

7/13/05 9:29:36 AM

8 School Stress, Anger Expression, and Adolescent Problem Behaviors Chok C. Hiew and Glendon Rayworth University of New Brunswick, Fredericton, Canada Abstract The effects of school stress and the risk of school failure on anger, violence, substance abuse, and delinquent behavior were investigated with junior high school students. “At-Risk” students were compared with “Average” students and “High Achievers” on five types of problem behaviors. Significant positive correlations were found between trait anger and anger expression (Anger-Out) with measures of Violence, Drug Use, Property Damage, Theft, and Nonserious Delinquency; AngerControl correlated negatively with these problem behaviors. At-Risk students had significantly higher scores than the High Achievers on all of these problem behaviors and significantly higher scores than the Average group on Violence and Drug Use. The predictive power of the anger measures was also demonstrated by the significant correlations that were found with the problem behaviors for all three groups. The finding that the experience and expression of anger were better predictors of adolescent problem behaviors than teachers’ratings supports an anger-mediated model for helping students cope with school stress.

SCHOOL STRESS, ANGER EXPRESSION, AND ADOLESCENT PROBLEM BEHAVIORS The relationship between stress and the high risk of experiencing problem behaviors has been well documented in previous research. Vaux and Ruggerio (1983), for example, found that stressful life changes added significantly to age and socioeconomic status as predictors of a variety of forms of delinquent behavior, which included violence, theft, drug use, property damage, and school truancy. Stressful life changes have also been found to be strongly associated with adult criminal behavior (e.g., Levinson & Ramsey, 1979; Masuda, Cutler, Hein, & Holmes, 1978). Although these studies have succeeded in linking stress with high-risk behaviors, they have not examined how specific types of stress are related to different particular problem behaviors. Despite the centrality of school stress in the lives of adolescents, the effects of school stress on problem behaviors has not been given sufficient attention. The findings by Elias, Gara, and Ubriaco (1985) of a positive relationship between 143

RT44367_C008.indd 143

7/13/05 9:35:32 AM

144

C.C. HIEW AND G. RAYWORTH

school stress and self-reported substance abuse in high-school students demonstrates the importance of research in this area. D’aurora & Fimian (1988) reported three basic factors in school stress: (a) student distress, (b) academic problems, and (c) poor instructional relations. Unfortunately, there are relatively few available measures of school-related stress with adequate psychometric properties. Fanshawe and Burnett’s (1991) inventory is generally adequate, but lacks evidence supporting its criterion validity. In contrast to other adolescent stressors, the risk of school failure is a high stress event that is routinely measured for almost all school-age adolescents. Low achievement in school is clearly evident in test scores, falling behind the modal grade, and leaving school before graduation. For adolescents, school failure is often related to a cluster of problem behaviors such as drug use, delinquent activities, and early sexual intercourse (Dryfoos, 1990). The longterm consequences of school failure and dropout have been linked empirically to a variety of social consequences, such as welfare dependency, physical and mental health problems, and criminality (Berlin & Sum, 1988; Ginzberg, Berliner, & Ostow, 1988). Clearly, adolescent stress caused by school failure reduces the probability of developing into a healthy, well-functioning, and mature adult. School failure and problem behaviors have long been suspected to be closely related, as suggested in several theories that have examined the relationship between the two phenomena. Jessor and Jessor’s (1977) Problem Behavior theory recognized a strong relationship between school failure and delinquency, together with a host of other problem behaviors. Jessor and his associates (Donovan & Jessor, 1985; Donovan, Jessor, & Costa, 1991) also concluded that certain adolescents were prone to engage simultaneously in a variety of problem behaviors, depending on a number of factors such as personality characteristics and perceptions of the school environment. Cloward and Ohlin (1960) have observed that when failure to achieve valued goals, for example, successful school performance, was attributed to injustices within the social system, a sense of normlessness ensued, in which the alienated individual was motivated to engage in delinquent behavior. Elliott and Voss (1974) extended this model, suggesting that delinquent behavior was the primary variable that causatively preceded school failure. This conceptualization challenged traditional thinking, which assumed that dropping out of school resulted in increased delinquent behavior. Empirical evidence to support the Elliott and Voss theory is reflected in the delinquency rates for dropouts while they were in and out of school, indicating that they had higher delinquency rates than the graduates and that their highest delinquency rates occurred while they were still in school. It is important to note that Elliott and Voss’s (1974) delinquency rates were based on the number of police contacts. While this measure is meaningful, it fails to take into account delinquent behavior that was undetected by police. Thus, graduating students may actually have participated in a similar or even greater level of delinquent behavior than the dropouts, which was undetected. The design of the present study addresses this methodological problem by employing anonymous self-report measures of delinquent behavior and substance

RT44367_C008.indd 144

7/13/05 9:35:32 AM

SCHOOL STRESS AND ANGER EXPRESSION

145

abuse, which, theoretically, are sensitive to all delinquent behaviors committed by individuals in various performance groups. PROBLEM BEHAVIOR AND ANGER EXPRESSION Rhodes and Jason (1990) derived a social stress model from Albee’s (1979) landmark prevention equation on stress and psychopathology. They proposed that an adolescent’s level of risk for engaging in delinquent or socially unacceptable behaviors increased with the number of environmental and social stressors that were experienced and decreased with the quality of their attachments, coping skills, and resources. Agnew (1985) also theorized that stressful aversive environments had a direct impact on delinquent behaviors and aggression problems. He further suggested that problem behaviors were maladaptive coping responses, mediated by anger, that motivated efforts to escape from an aversive environment. Vaux and Ruggerio (1983) have clearly demonstrated that measures of stressful life changes were related to increased adolescent involvement in self-reported substance abuse and delinquent behavior, and suggested three mechanisms to account for these relationships: (1) stress as an impetus to anxiety, (2) stress as an impetus to anger and aggression, and (3) stress as an impetus to reduced social attentiveness. In research on school stress, Agnew (1985) found that when adolescents became angered or frustrated with school failure, they often manifested their angry feelings in delinquent activities and drug abuse. Thus, problem behaviors resulted from inappropriate coping responses to anxiety and anger, especially the excessive expression of uncontrolled anger. The mediating role of anger appears to motivate coping responses that are manifested in various forms of delinquent activities. There is, however, considerable confusion in regard to the meaning and measurement of anger in previous research, for which Spielberger’s (1980, 1988) conceptualization of anger as a psychological construct provides needed clarification (Spielberger, Krasner, & Solomon, 1988; Spielberger, Reheiser, & Sydeman, 1995). He points out that anger, hostility, and aggression are overlapping concepts that are referred to as the AHA! Syndrome (Spielberger et al., 1995) and theorizes that problematic aggressive acts and hostile behavioral manifestations are generally preceded by the emotional experience of intense anger. Aggressive behavioral responses are the operative affective component of the expression of anger in aggressive and hostile behavior. According to Spielberger et al. (1988), the way that people cope with anger can be classified according to three modes of anger expression: (1) Anger-Out; (2) Anger-In; and (3) Anger-Control. The outward expression of anger tends to manifest itself in aggressive behavior, whereas turning anger in results in the relatively enduring suppression of angry feelings that are not readily observable to others. Anger-Control refers to coping successfully with angry feelings so that these feelings are not expressed (“keeping the lid on”) or are quickly dissipated (calming down) without manifesting themselves internally

RT44367_C008.indd 145

7/13/05 9:35:32 AM

146

C.C. HIEW AND G. RAYWORTH

or externally. Spielberger’s (1980) State-Trait Anger Scale (STAS) also assesses the intensity of anger as an emotional state and individual differences in trait anger or anger proneness as relevant anger-related concepts. Evidence of the potential relevance of the AHA! Syndrome to school-related behavior was reported by Parker and Asher (1987), who observed higher ratings of aggressiveness among dropouts by both teachers and peers at the time the dropouts first entered the ninth grade. Rowe and Nesdale (1976) confirmed that anger does, indeed, significantly heighten aggression, and Roff (1986) found that adolescents at high risk for delinquency could be identified by their aggressive behavior. The intensity and frequency that adolescents experience the emotion of anger seems to mediate aggressive behavior and acts of violence. Proneness to anger has also been found to be related to various forms of drug use (Swaim, Oetting, Edwards, & Beauvais, 1989), whereas elevated scores on Spielberger’s Anger-Out measure were related to higher cigarette and marijuana use (Stoner, 1988; Witt, Kaelin & Stoner, 1988). Given these results, it would seem reasonable to expect that the excessive experience of anger, that is, high anger proneness or trait anger, and its outward expression might motivate problem behaviors and mediate the effects of stress on school failure. According to Cloward and Ohlin (1960), school dropouts have a history of experiencing and expressing a disproportionate level of frustration and anger when confronted with school failure, and Loeber and Dishion (1983) found that high ratings by students as being easily angered was a significant predictor of delinquency. However, as noted by Cowen and Work (1988), the fundamental challenge is to determine what complex processes and stress-resilient factors moderate stress and overall adolescent adjustment. Therefore, the inclusion of an anger control measure as a moderating factor, along with anger expression measures, will facilitate investigating the positive role of anger management in mediating problem behaviors that contribute to school failure. METHOD Study participants. The participants in this study were 108 male ninth-grade students enrolled in a public school, who were predominantly Anglo-Saxon Caucasians. Participation was limited to males because of the disproportionately high number of male public-school dropouts (Presseisen, 1988) and to avoid the potential introduction of different effects of gender on the various measures of delinquency (Elliott & Voss, 1974). Classroom teachers rated the study participants on the basis of their school performance as belonging to one of three groups: (1) High Achievers, (2) Average Students, and (3) At-Risk Students. The number of students rated as Average was 59. However, for statistical purposes, only 30 of the 59 Average subjects were randomly selected for the final data analysis in order to conform to the relatively smaller sample size of the other two groups. Of the 79 study participants who were included in the final sample, 24 were rated as High Achievers, 30 as Average, and 25 as At-Risk. Differences between the three groups for each dependent measure were evaluated in one-way ANOVAs.

RT44367_C008.indd 146

7/13/05 9:35:33 AM

SCHOOL STRESS AND ANGER EXPRESSION

147

Predictor and Criterion Measures Teacher ratings. Standardized rating criteria were provided for the teachers, who rated each student in accordance with one of the following classifications: High Achiever (rated 1), Average (rated 2), At-Risk for school failure (rated 3). The criteria for the High Achiever group included high GPA, excellent class attendance, and an above-average level of overall school performance. Criteria for the Average rating were average GPA and average classroom behavior and school performance. For the At-Risk classification, the criteria included low or failing GPA, and two or more of the following characteristics: chronic truancy, a record of suspension or expulsion, chronic disturbance in class, or excessive disciplinary problems involving school authorities. Because record of academic success proved to be the best predictor of final dropout status (Elliott & Voss, 1974), GPA level was the most discriminating criterion for distinguishing among the three groups. Trait Anger Scale. The Trait Anger Scale (Spielberger, Jacobs, Russell & Crane, 1983) measures individual differences in the frequency that anger is experienced over time. Persons who score high on Trait Anger are more likely to perceive a wider range of situations as anger provoking and are considered to have a higher degree of anger proneness. Subjects are instructed to indicate how often they generally feel or react in the manner described by each item (e.g., “I feel angry”) by rating themselves on the following four-point scale: 1 = Almost never, 2 = Sometimes, 3 = Often, and 4 = Almost always. Adequate psychometric properties of the Trait Anger scale have been reported (Spielberger et al., 1983). Anger Expression Scale. The Anger Expression (AX) Scale was developed by Spielberger (1988) to provide a more comprehensive measure of the fundamental properties of anger, above and beyond the intensity and frequency that anger was experienced. The AX scale examines how anger is expressed, with instructions for respondents to indicate how often they react or behave in the manner described by each item when they feel angry or furious. The AX scale includes the following three subscales: (1) Anger-Out (AX/Out); (2) Anger-In (AX/In); and (3) Anger-Control (AX/Con). The AX/Out scale measures how often an individual expresses anger outwardly in an aggressive manner; for example, “I do things like slam doors.” The AX/In scale assesses how frequently an individual suppresses or holds in anger; for example, “I boil inside, but I don’t show it.” The AX/Con scale evaluates how frequently individuals cope with anger by controlling its expression so that it doesn’t affect them in a negative manner; for example, “I control my temper.” Each AX scale item is rated on the same four-point scale that is used with the Trait Anger Scale, as previously described. Problem behaviors. The Delinquent Activities Questionnaire (DAC) is a modified version of the self-report scale used by Vaux and Ruggerio (1983) and Richards (1981), which consists of 32 items representing five problem areas. Respondents are instructed to indicate how often during the previous six months they participated in the specific problem activities described by each item, by rating themselves on the following four-point scale: 1 = Never,

RT44367_C008.indd 147

7/13/05 9:35:33 AM

148

C.C. HIEW AND G. RAYWORTH

2 = Rarely, 3 = Sometimes, and 4 = Often. The DAC problem behavior scales assess the following five types of delinquent behaviors: (1) Violence, (2) Property Damage, (3) Drug Use, (4) Theft, and (5) Nonserious Delinquency or school infractions. A number of studies have supported the DAC as a reliable and valid measure of delinquent behavior (e.g., Farrington & West, 1971; Tolan & Lorion, 1988). Procedure. The instructions for each measure were explained by the two researchers who administered the questionnaires. An informed consent was obtained from each participant prior to questionnaire completion. The consent form assured participants of anonymity and the confidentiality of their responses. The questionnaires were coded with numbers that corresponded to the teachers’ ratings of the school performance of each participant as a High Achiever, Average, or At-Risk student. RESULTS Differences in the performance of the High Achiever, Average, and At-Risk groups for each of the five types of problem behaviors as measured by the DAC were evaluated by one-way ANOVAs. Subsequent Scheffe multiple comparisons were conducted to detect significant differences between the High Achiever, Average, and At-Risk groups. The relationships between the various dependent measures were analyzed by computing Pearson product-moment correlations for all possible paired combinations of the dependent measures. The means and standard deviations of each DAC behavior measure for the three groups and the results of the Scheffe comparisons are reported in Table 8.1. These analyses revealed significant between-groups differences for all five problem behaviors. Specifically, the results of the ANOVAs indicated the following significant F tests of differences between the three groups for each of the five problem behaviors: Violence, F(2, 74) = 12.26, p < .01; Property Damage, F(2, 72) = 4.95, p < .01; Drug Use, F(2, 71) = 17.70, p < .0001; Theft, F(2, 74) = 6.69, p < .02; and Nonserious Delinquency, F(2, 72) = 13.05, p < .0001. The At-Risk students scored significantly higher than the High Achievers on all five problem behaviors, and higher than the Average group on Violence and Drug Use. The Average group also had somewhat higher scores than the High Achiever group, but these differences were significant only for Nonserious Delinquency. The Scheffe pairwise comparisons also yielded significantly higher (p < .05) Violence and Drug Use scores for the At-Risk group as compared to the Average group (t = 6.60, t = 4.23). The At-Risk and Average groups both scored significantly higher than the High Achiever group on self-reported delinquent activity (t = 6.76, t = 3.77). These significant findings may be interpreted as indicating that the teachers’ ratings of the risk of school failure accurately reflected not only the students’ poor GPA, but also their record of chronic truancy, school suspension, acting-out behavior, and disciplinary problems.

RT44367_C008.indd 148

7/13/05 9:35:33 AM

149

SCHOOL STRESS AND ANGER EXPRESSION Table 8.1 Level of Academic Performance and Types of Problem Behaviors in Grade 9 Male Adolescents Group I

Group II

Group III

High Achiever

Average

At-Risk

Scheffe Comparisons

Variable

M

SD

M

SD

M

SD

HA.A HA.AR A.AR

Violence Property Damage Drug Use Theft Nonserious Delinquency

9.83

3.17

10.27

2.42

13.70

3.34

0.44

3.87*

3.43*

6.75 7.35 6.00

2.13 1.94 1.06

8.03 9.72 7.17

2.81 4.58 1.91

9.45 13.95 8.26

3.69 4.02 3.02

1.28 2.37 1.17

2.70* 6.60* 2.26*

1.43 4.23* 1.09

16.92

3.46

20.69

4.58

23.68

5.34

3.77*

6.76*

2.99

Note: *p < .05; HA = High Achiever group; A = Average group; AR = At-Risk group.

Table 8.2 Pearson Product-Moment Correlations between the Trait Anger and Anger Expression Scales

1. Trait Anger (T-Anger) 2. Anger-Out (AX/Out) 3. Anger-In (AX/In) 4. Anger-Control (AX/Con)

2

3

4

.64*** —

.21 .22 —

–.41*** –.49*** –.15 —

Note: ***p < .0001.

Because no significant differences were found among the three groups for any of the anger measures, the group means and SDs for the anger measures are not reported. The product-moment correlations among the four anger measures, reported in Table 8.2, indicated that Trait Anger and AX/Out correlated highly with each other, and that both of these scales correlated negatively with the AX/Con measure. Consistent with previous research (Spielberger, 1988), the findings indicated that experiencing anger more frequently was related to the outward expression of anger and that greater anger control was associated with experiencing and expressing anger less often. The mediating effects of anger were evaluated by computing correlations between the four anger measures and the five problem behaviors that are reported in Table 8.3. The results revealed multiple significant relations between the anger measures and the various problem activities. Trait Anger correlated strongly and significantly with all of the problem behavior variables except Drug Use, and

RT44367_C008.indd 149

7/13/05 9:35:34 AM

150

C.C. HIEW AND G. RAYWORTH

Table 8.3 Pearson Product-Moment Correlations between Anger Variables and Problem Behaviors Violence

Property Damage

Drug Use

Theft

Nonserious Delinquency

.46*** .43*** .15 –.24*

.31** .35** .33** –.20

.18 .30** .16 –.24*

.25* .43*** .26* –.23*

.33** .44*** .18 –.25*

Trait Anger Anger-Out (AX/Out) Anger-In (AX/In) Anger-Control (AX/Con)

Note: *p < .05; **p < .01; ***p < .001.

Table 8.4 Pearson Product-Moment Correlations between Five Types of Problem Behaviors

1. Violence 2. Property Damage 3. Drug Use 4. Theft 5. Nonserious Delinquency

2

3

4

5

.38** —

.56*** .62*** —

.37** .72*** .65*** —

.62*** .70*** .74*** .70*** —

Note: **p < .01; ***p < .001.

AX/Out correlated positively and significantly with all five problem behaviors. The AX/Con scale also correlated significantly and negatively with all of the problem behaviors, except Property Damage. AX/In only correlated significantly with Property Damage and Theft. Overall, the strongest correlations (p < .001) were between AX/Out and Violence (r = .43), Theft (r = .43), and Nonserious Delinquency (r = .44), and between Trait Anger and Violence (r = .46). Correlations of the five types of problem behaviors with each other are reported in Table 8.4. The single most prevalent problem behavior was Nonserious Delinquency, which correlated .74 with drug use, and .62 or higher with the other three problem activities. Clearly, the students who were identified as having more school-related problematic behaviors (e.g., skipping school, suspension, being expelled, troubles with teacher, unruly conduct) also had more problems with Violence, Property Damage, and Theft and especially with Drug Abuse. Violence was also a significant problem behavior, with implications for physical aggression and substance abuse. DISCUSSION AND CONCLUSIONS The results of this study strongly support the hypothesis that students who are judged by their teachers to be at risk of failure are involved in more schoolrelated delinquency and disciplinary problems. They are also more likely to

RT44367_C008.indd 150

7/13/05 9:35:34 AM

SCHOOL STRESS AND ANGER EXPRESSION

151

participate in a greater number of high-risk activities, including violence, drug use, theft, and vandalism. These relationships were most pronounced in comparing the At-Risk students with the High Achievers and provide strong support for Jessor and Jessor’s (1977) Problem Behavior theory, which posits that a wide range of adolescent problem behaviors tend to be highly interrelated. Adolescents who engage in a particular problem behavior are disposed to engage in a wide range of other problem behaviors. The findings of significant group differences further suggested that failing at school was a strong predictor not only of school problems, but also the likelihood of being involved in wide range of antisocial activities (Elliott & Voss, 1974; Loeber & Dishion, 1983). A second major finding in this study was that the problem behaviors were influenced by anger in a similar manner for all three student groups, which was consistent with the anger-mediated coping model for predicting problem behaviors suggested by a number of researchers (Agnew, 1985; Barnett, Fagan, & Booker, 1991; Cloward & Ohlin, 1960; Rhodes & Jason, 1990; Vaux & Ruggerio, 1983). These findings also support Spielberger et al.’s (1995) conception of the AHA! Syndrome, which distinguishes between the experience of angry feelings and how anger is expressed in hostility and aggressive behavior. Adolescents who were failing in school expressed their feelings of anger and frustration in various forms of hostile and aggressive behaviors, which were mediated by different types of angerrelated problems. The two anger measures that were the best predictors of all five types of problem behaviors (Violence, Property Damage, Drug Use, Theft, and Nonserious Delinquency) were Trait Anger (anger proneness) and Anger-Out (AX/Out), the expression of anger in aggressive behavior. Anger-In (Ax/In) also predicted delinquency as related to Theft and Property Damage. It is especially important to note that the At-Risk, Average, and High Achievers did not differ on any of the self-reported anger measures. The findings for anger cut across all three groups of students and were better than teacher ratings in predicting a wider range of problem behaviors. The significant negative correlations of Anger-Control (AX-Con) with most of the problem behaviors suggested the potential value of incorporating anger management programs for dealing with the high incidence of school failure and subsequent dropout. In previous research, anger management programs have been found to be efficacious in the treatment of adolescent problem behaviors (e.g., Feindler, Marriott, & Iwata, 1984; Novaco, 1978) and have proved helpful in working with unemployed former dropouts (Hiew, 1995). Finally, on approaching the relation between anger and school adjustment from the other end, that is, assuming that the source of excessive anger results from frustration related to school stress and failure, the school curriculum needs to be adjusted not only to facilitate academic learning, but also to be more sensitive to individual students’ needs. There is a vital need to promote resilience and hardiness in adolescents through highlighting each student’s healthy self-perceptions, coping skills, and emotional commitment to educational success.

RT44367_C008.indd 151

7/13/05 9:35:35 AM

152

C.C. HIEW AND G. RAYWORTH

REFERENCES Agnew, R. (1985). A revised strain theory of delinquency. Social Forces, 64, pp. 51–167. Albee, G. (1979). Primary prevention. Canada’s Mental Health, 27, pp. 5–9. Barnett, O. W., Fagan, R. W., & Booker, J. M. (1991). Hostility and stress as mediators of aggression in violent men. Journal of Family Violence, 6, pp. 217–241. Berlin, G., & Sum, A. (1988). Toward a more perfect union: Basic skills, poor families, and our economic future. New York: Ford Foundation. Cloward, R., & Ohlin, L. (1960). Delinquency and opportunity. New York: Free Press. Cowen, E. L., & Work, W. C. (1988). Resilient children, psychological wellness, and primary prevention. American Journal of Community Psychology, 16, pp. 591–606. D’aurora, D. L., & Fimian, M. J. (1988). Dimensions of life and school stress experiences by young people. Psychology in the Schools, 25, pp. 44–53. Donovan, J. E., & Jessor, R. (1985). Structure of problem behavior in adolescence and young adulthood. Journal of Consulting and Clinical Behavior, 53, pp. 890–904. Donovan, J. E., Jessor, R., & Costa, F. M. (1991). Adolescent health behavior and conventionality-unconventionality: An extension of problem behavior theory. Health Psychology, 10, pp. 52–61. Dryfoos, J. G. (1990). Adolescents at-risk: Prevalence and prevention. New York: Oxford University Press. Elias, M. J., Gara, M., & Ubriaco, M. (1985). Sources of stress and support in children’s transition to middle school: An empirical analysis. Journal of Clinical Child Psychology, 14, pp. 112–118. Elliott, D. S., & Voss, H. L. (1974). Delinquency and dropout. Toronto: D. C. Heath and Company. Fanshawe, J. P., & Burnett, P. C. (1991). Assessing school-related stressors and coping mechanisms in adolescents. British Journal of Educational Psychology, 61, pp. 92–98. Farrington, D. P., & West, D. J. (1971). A comparison between early delinquents and young aggressives. British Journal of Criminology, 11, pp. 341–358. Feindler, E. L., Marriott, S. A., & Iwata, M. (1984). Group anger control training for junior high school delinquents. Cognitive Therapy and Research, 8, pp. 299–311. Ginzberg, E., Berliner, H., & Ostow, M. (1988). Young people at risk: Is prevention possible? New York: Ford Foundation. Hiew, C. C. (1995). Job stress in the shifting Canadian economy. In C. D. Spielberger & I. G. Sarason (Eds.), Stress and emotion: Anxiety, anger, and curiosity (Vol. 15, pp. 103–119). Washington, DC: Taylor & Francis. Jessor, R., & Jessor, S. (1977). Problem behavior and psychosocial development: A longitudinal study of youth. New York: Academic Press. Levinson, R. M., & Ramsey, G. (1979). Dangerousness, stress, and mental health evaluations. Journal of Health and Social Behavior, 20, pp. 178–187.

RT44367_C008.indd 152

7/13/05 9:35:35 AM

SCHOOL STRESS AND ANGER EXPRESSION

153

Loeber, R., & Dishion, T. (1983). Early predictors of male delinquency: A review. Psychological Bulletin, 94, pp. 68–99. Masuda, M., Cutler, D. L., Hein, L., & Holmes, T. H. (1978). Life events and prisoners. Archives of General Psychiatry, 35, pp. 197–203. Novaco, R. M. (1978). Anger and coping with stress. In J. Forey & D. Rathzen (Eds.), Cognitive behavior therapy (pp. 135–173). New York: Plenum. Parker, J. G., & Asher, S. R. (1987). Peer relations and later personal adjustment: Are low-accepted children at risk? Psychological Bulletin, 102, pp. 357–389. Presseisen, B. Z. (1988). At-risk students and thinking: Perspectives from research. Washington, DC: Research and National Education Association for Better Schools. Rhodes, J. E., & Jason, L. A. (1990). A social stress model of substance abuse. Journal of Consulting and Clinical Psychology, 58, pp. 395–501. Richards, P. (1981). Quantitative and qualitative sex differences in middleclass delinquency. Criminology, 18, pp. 453–470. Roff, J. D. (1986). Identification of boys at high risk for delinquency. Psychological Reports, 58, pp. 615–618. Rowe, B. G., & Nesdale, A. R. (1976). Emotional arousal and aggressive behavior. Psychological Bulletin, 83, pp. 851–863. Spielberger, C. D. (1980). Preliminary manual for the State-Trait Anger Scale (STAS). Tampa, FL: University of South Florida, Department of Psychology. Spielberger, C. D. (1988). Manual for the State-Trait Anger Expression Inventory (STAXI). Odessa, FL: Psychological Assessment Resources. Spielberger, C. D., Jacobs, G., Russell, S., & Crane, R. S. (1983). Assessment of anger: The state-trait anger scale. In J. N. Butcher & C. D. Spielberger (Eds.), Advances in personality assessment. (Vol. 2, pp. 159–187). Hillsdale, NJ: LEA. Spielberger, C. D., Krasner, S. S., & Solomon, E. D. (1988). The experience, expression, and control of anger. In M. P. Janisse (Ed.), Individual differences, stress, and health psychology (pp. 89–107). New York: Springer-Verlag. Spielberger, C. D., Reheiser, E. C., & Sydeman, S. J. (1995). Measuring the experience, expression, and control of anger. In H. Kassinove (Ed.), Anger disorders: Definitions, diagnosis, and treatment (pp. 49–67). Washington, DC: Taylor & Francis. Stoner, S. S. (1988). Undergraduate marijuana use and anger. Journal of Psychology, 122, pp. 343–347. Swaim, R. C., Oetting, E. R., Edwards, R. W., & Beauvais, F. (1989). Links from emotional distress to adolescent drug use: A path model. Journal of Consulting and Clinical Psychology, 57, pp. 227–231. Tolan, P. H., & Lorion, R. P. (1988). Multivariate approaches to the identification of delinquency proneness in adolescent males. American Journal of Community Psychology, 16, pp. 547–561. Vaux, A., & Ruggerio, M. (1983). Stressful life change and delinquent behavior. American Journal of Community Psychology, 11, pp. 169–183. Witt, E. M., Kaelin, J. A., & Stoner, S. B. (1988). Smoking behavior and anger. Psychological Reports, 63, pp. 117–118.

RT44367_C008.indd 153

7/13/05 9:35:35 AM

RT44367_C008.indd 154

7/13/05 9:35:35 AM

9 Experience of Anger and Anxiety Emotions during Examinations: Cross-Cultural Evidence Norbert K. Tanzer Alliant International University, San Diego, CA Charles D. Spielberger University of South Florida, Tampa, FL Abstract The literature abounds with studies on test anxiety, but there is little research on the experience of anger and other emotions in examinations and other achievement-related situations. This chapter reports the findings of a series of studies of lower secondary school and university students from Austria, Singapore, and the United States who were asked to report their experience of anger and anxiety during tests. Following the second author’s state-trait distinction, trait measures of examination anger and anxiety were administered during regular classroom periods, and state measures of these emotions were administered immediately after written tests. In a second series of studies, three intelligence/aptitude tests were administered to Austrian and U.S. university students, who were instructed to rate the intensity of their feelings, including anger and anxiety, at the beginning of the experiment and immediately after each intelligence test. In all 11 studies, anger and anxiety emerged as distinct but correlated factors. During examination situations, some subjects frequently experienced intense anger while others experienced more intense anger. Like test anxiety, “test anger” appears to hinder effective test taking behavior.

EXPERIENCE OF ANGER AND ANXIETY EMOTIONS DURING EXAMINATIONS: CROSS-CULTURAL EVIDENCE1 The literature abounds with studies on test anxiety, but there is relatively little research on the experience of anger and other emotions in examination situations (e.g., Elkind, 1986; Folkman & Lazarus, 1985; Pederson & Hollandsworth, 1987; The major part of the research reported in this paper was conducted during a one-year Erwin Schrödinger scholarship by the first author at the Center for Research in Behavioral Medicine and Health Psychology at the University of South Florida. The scholarship was awarded by the Austrian Science Foundation (FWF Project #J0805-SOZ).

1

155

RT44367_C009.indd 155

8/15/05 6:57:23 AM

156

N.K. TANZER AND C.D. SPIELBERGER

Pekrun, 1994, 1997; Pekrun & Frese, 1992; Smith & Ellsworth, 1985, 1987). Nevertheless, even these few studies clearly indicate that a variety of positive and negative emotions are frequently evoked in examination situations. Positive emotions include happiness, hope/challenge, joy, relief, and satisfaction, while negative emotions include anger, annoyance, anxiety, frustration, hopelessness/ sadness, and self-blame/guilt (cf. Pederson & Hollandsworth, 1987; Pekrun, 1994; Smith & Ellsworth, 1987). Consistent with results from laboratory studies (e.g., Polivy, 1981), a considerable degree of emotional blending was found in the emotional state of students during or after an examination (Folkman & Lazarus, 1985; Smith & Ellsworth, 1987). This co-occurrence of emotions is also reflected in the relatively high correlations that were found between emotions during examination situations (Pekrun, 1994; Smith and Ellsworth, 1987). However, with the exception of a study by Smith & Ellsworth (1987), the factorial validity of the scales used for examination situations still needs to be verified. Furthermore, the abovementioned studies used relatively small samples of university students, mainly from the United States. Although anger and anxiety are universal emotions, the way in which people cope with them is governed by cultural factors, such as the traditions and norms of the respective society (cf. Mesquita & Frijda, 1992). Considering the differences in tradition, culture, child rearing, and educational systems in Eastern and Western societies (e.g., Bond, 1991; Wu, 1991), it is plausible that how individuals experience, express, and cope with anger or anxiety may be different in non-Western societies. These patterns may also change during the developmental process (cf. Törestadt, 1990). Furthermore, most theories and research on test anxiety have been based on a global concept of test anxiety that does not take into account the possibility of substantial individual differences in anxiety aroused by different academic domains. In recent research (Tanzer, 1992; Tanzer, Hochegger, & Linhart, 1993; Tanzer, Hochegger, & Schwetz, 1994), a certain degree of domain specificity in test anxiety was found, but no studies on the domain specificity of other achievement-related emotions such as anger have been conducted. The main objective of the research reported in this chapter was to evaluate differences between anger and anxiety in examination situations across samples of different age groups from several countries with different cultural backgrounds or different educational systems. METHOD Instruments Development of the Examination Anger and Anxiety Scales (EAAS).2 A review of the literature of existing trait and state scales on anger or anxiety such as the second author’s (1973, 1979, 1980, 1983a, 1983b, 1988) STAI, The first author is indebted to Prof. James Dickinson, University of South Florida, and Prof. Kjell Haseth, University of Oslo, for their valuable comments during many hours of discussion and brainstorming.

2

RT44367_C009.indd 156

8/15/05 6:57:24 AM

EXPERIENCE OF ANGER AND ANXIETY EMOTIONS

157

STAXI, STPI, and TAI; Liebert and Morris’s (1967) Worry-Emotionality Scale (Morris, Davis, & Hutchings, 1981); and Sarason’s RTT (1984), together with a screening of various English-language thesauri, yielded a working collection of anger and anxiety items. Note that throughout this chapter, we will use the term anxiety as a synonym for the emotionality component in the WorryEmotionality conceptual distinction introduced by Liebert and Morris (1967). In a series of small pilot studies, the collection of anger and anxiety items was presented to teachers and students. Based on their feedback, the items were rephrased and shortened so as to remove any ambiguities, either due to item content or colloquial use of English. For example, TAI Emotionality Item 2 (“While taking examinations, I have an uneasy, upset feeling”) was split into two new items, with “uneasy” and “upset” as the key words. “Uneasy” turned out to be a clearer anxiety (emotionality) item, while “upset” repeatedly loaded on both the anxiety (emotionality) and anger factors, indicating that respondents may use this expression to reflect a nonspecific blend of negative emotions, rather than one specific emotion. Another criterion in the development of the final set of items for the Examination Anger and Anxiety Scales was that each item could be used with either trait or state instruction (see next section). For example, items that referred to “being in a state of turmoil” or “being furious” can be used as trait items but not as state items, because one cannot be in a “slight” state of turmoil or a “little bit” furious. The final criterion was the prevalence of the anger and anxiety symptoms. Infrequently experienced symptoms, such as “I get nausea,” were discarded. Although the findings reported in this chapter focus only on the experience of anger and anxiety in examination situations, EAAS items were also part of a larger questionnaire on cognitions and emotions. Trait and State Instructions Following the second author’s state-trait distinction, we differentiated, for both anger (e.g., 1979, 1988; Spielberger, Jacobs, Russell, & Crane, 1983; Spielberger, Krasner, & Solomon, 1988) and anxiety (e.g., 1972, 1983a, 1983b, 1985), between (a) states, that is, the intensity of angry/anxious feelings that a student experiences during a particular test, and (b) traits, the proneness or frequency that a student experiences these emotions during tests. Studies 1–4 and Study 10 used trait measures to assess examination anger and anxiety, while Studies 5–9 applied state measures. The trait measures were administered during lessons; the state measures were administered immediately after a written examination or intelligence/aptitude assessment. In Study 11, trait measures were administered to explore the experience of feelings of anger and anxiety during regular classroom lessons, rather than during examinations. In Trait Studies 1–4, the EAAS items were administered by using a global trait instruction (i.e., “How often do you have this thought or feeling during tests?”). Trait Study 10 focused on the specific academic domains of reading and math by using a domain-specific trait instruction (i.e., “How often do you have this thought or feeling during tests in Math?” and “How often do you have this thought or feeling during tests in Reading?”). In State Studies 6-9,

RT44367_C009.indd 157

8/15/05 6:57:24 AM

158

N.K. TANZER AND C.D. SPIELBERGER

the EAAS scales were administered immediately after an actual examination or intelligence/aptitude assessment (i.e., “How did you feel during the test you just completed?”), while Study 5 aimed at assessing recollected states (i.e., “How did you feel during the last test you attended?”). Rating Scale Formats In all studies with trait instructions (i.e., Studies 1–4 and 10–11), the same four-point frequency rating scale used with the second author’s anxiety and anger measures was used (1 = Hardly ever, 2 = Sometimes, 3 = Often, 4 = Almost always). In State Studies 6 and 8, in which the English EAAS version was used (cf. Appendix A), subjects rated themselves on a five-point intensity rating scale (0 = Not at all, 1 = Somewhat, 2 = Moderately, 3 = Quite a bit, 4 = Very much). In State Studies 7 and 9, which used the German EAAS version (cf. Appendix B), a seven-point graphical intensity rating scale was used (with verbal anchors for 1 = Not at all, 3 = Somewhat, 5 = Quite a bit, and 7 = Very much). In Study 5, which used the German version, the subjects had to recall (i.e., recollect) their state of anger and anxiety during their last examination by rating themselves on a four-point intensity rating scale (1 = Not at all, 2 = Somewhat, 3 = Quite a bit, 4 = Very much). For all studies using the German EAAS, the closest English translation of the German rating scales was as given above. Statistical Analyses In all studies, the Principal-Axis factor analysis solutions were subjected to oblique rotation (direct oblimin method with d = 0; cf. Kim & Mueller, 1978; SPSS, 1990). The number of factors to be extracted was based on the postulated theoretical factor structure; that is, a two-factor solution in Studies 1–9 and Study 11, and a four-factor solution in Study 10. In all studies, the Anger and Anxiety Scale scores were defined as the average rating of the items comprising each scale. Because means and standard deviations (SDs) depend on the particular rating scale used, they cannot be directly compared between studies that used different rating scale formats. In particular, this prohibits any comparisons of means and SDs between the English trait studies (i.e., Studies 1–3) and their German counterparts (i.e., Studies 4 and 10). Factor loadings, however, are invariant for any linear scale transformation and are thus not affected by different rating scale formats. Endorsement rates will be reported in addition to item means for each anger item. These endorsement rates were calculated for the trait studies as the cumulative (i.e., pooled) percentage of subjects with either an “Often” or “Almost always” rating for the respective anger symptom. Similarly, for all state studies, the endorsement rates were calculated as the cumulative percentage of reports with either a “Quite a bit” or “Very much” rating for the respective anger symptom. In Study 10, course specificity was investigated by calculating the percentage of variance between classes. Because “class affiliation” has to be treated as a random ANOVA factor, the percentage of between classes variance can be calculated as 100 times the intraclass-correlation (cf. Winer, 1971).

RT44367_C009.indd 158

8/15/05 6:57:25 AM

EXPERIENCE OF ANGER AND ANXIETY EMOTIONS

159

Unbiased estimates of intraclass correlations were obtained from multilevel theory (cf. Goldstein, 1987; Paterson & Goldstein, 1991; Raudenbush, Rowan, & Jin Kang, 1991) using the ML3 software (Prosser, Rasbash, & Goldstein, 1991). Subjects and Procedures Study 1. The English trait version of the EAAS was administered to a sample of undergraduates enrolled in several introductory psychology courses at a large state university as part of a questionnaire survey. A total of 368 students (36% males, 64% females; 74% Anglo-American, 9% African-American, 9% HispanicAmerican, 4% Asian-American, 3% Other, 1% No answer) completed the questionnaire on an anonymous, voluntary basis during a regular class session. Study 2. The English trait version of the EAAS was administered to seventh and eighth graders in several middle schools as part of a questionnaire survey.3 A sample of 158 students (49% boys, 51% girls; 76% Anglo-American, 8% Hispanic-American, 4% African-American, 2% Asian-American, 6% Other, and 4% No answer) completed the questionnaire on an anonymous and voluntary basis during a regular classroom session. Study 3. The English trait version of the EAAS was administered to seventh to ninth graders in a Singaporean all-boys secondary school as part of a questionnaire survey.4 A sample of 199 boys (88% Chinese, 4% Indian, 1% Malay, 7% No answer) completed the questionnaire on an anonymous and voluntary basis during a regular classroom session. Because English is the medium of instruction for all academic subjects beginning in the first year of primary school, the English comprehension level of Singaporean seventh graders is generally similar to that of native English speakers. Therefore, the original English EAAS version was used in this study to avoid any bias arising from translation (cf. Hambleton, 1991; Tanzer, Sim, and Marsh, 1992). However, in order to ensure the validity of using the English EAAS, an additional check of the comprehension of each item by the Singaporean students was conducted in six primary-school classes.5 Study 4. The German trait version of the EAAS (see Appendix B) was administered to a sample of Austrian university undergraduates enrolled in several introductory psychology courses as part of a questionnaire survey.6 This questionnaire was completed on an anonymous and voluntary basis during a regular psychology course session by 445 students (48% males, 52% females). Study 5. After first completing the trait version of the EAAS, the subjects who participated in Study 4 were asked to respond to the EAAS items a second time. However, unlike in Study 4, they were instructed to indicate, on a four-point intensity rating scale, how they felt while taking their last 3 We would like to thank Ray Gadd from the District School Board of Pasco County, Florida, for supervising the data collection. 4 We would like to thank Dr. Ong Teck Chin for conducting the questionnaire administration for this study. Data collection approved by the Singapore Ministry of Education (Approval Number ESS/94/1). 5 We would like to thank Ho Yeng Chow for conducting this comprehension survey. 6 We would like to thank Dr. Georg Gittler and Dr. Heinrich Hochegger for conducting the data collection in Vienna and in Graz, respectively.

RT44367_C009.indd 159

8/15/05 6:57:26 AM

160

N.K. TANZER AND C.D. SPIELBERGER

examination (i.e., “recollected state” instruction), which in most cases was at least several weeks prior to their participation in Study 5. Study 6. The English state version of the EAAS was administered on two occasions to 355 U.S. university undergraduates (31% males, 69% females) enrolled in three introductory psychology courses. This scale was given immediately after their first and second written tests as part of a questionnaire survey of students from a variety of majors who participated anonymously and voluntarily for partial course credit.7 A total of 312 state reports were collected after the first examination; 281 reports were obtained following the second examination. Study 7. The German state version of the EAAS was administered to a sample of 146 Austrian university students (33% males, 67% females) in an introductory experimental psychology course, immediately after their midterm and final written tests. Each student participated anonymously and voluntarily for partial course credit. A total of 123 state reports were collected after the first examination, 111 after the second examination, and 35 after the reexamination. Study 8. A sample of 126 U.S. university undergraduates (13% males, 87% females) participated anonymously and voluntarily in this study, and received partial course credit in their introductory psychology courses.8 The subjects had to work on three nonverbal intelligence/ability tests: (1) the “d2” (Brickenkamp, 1974), which is a Bourdon-like letter cancellation test measuring short-term concentration; (2) the “Viennese Matrices Test” (VMT; Forman & Piswanger, 1979), a Raven SPM-like reasoning test; and (3) the “Three-Dimensional Cube Comparison Test” (3DC; Gittler, 1990; Tanzer, Gittler, & Sim, 1994), which measures spatial ability. The test battery was administered in small groups (less than 8 subjects per group) in the following sequence: concentration test (d2), reasoning test (VMT), and spatial ability test (3 DC). The concentration test was given under speed instructions (i.e., timed). The reasoning test (18–30 minutes) and the spatial ability test (20–35 minutes) were given under self-paced power conditions. At the beginning of the experiment, which was used as the baseline, and after each of the three tests, the subjects responded to the English EAAS state version. Approximately 90 minutes was required to complete this experiment. All but one subject completed all four EAAS state scales. Study 9. The experimental design and procedures for Study 9 were identical to those of Study 8, except the language was German. Two Austrian samples, namely, 120 university psychology students (26% males, 74% females) and 70 military academy cadets (males only) participated in this study on an anonymous and voluntary basis.9 All 190 subjects completed the German EAAS state version four times. We would like to thank Lois Prettyman for screening and keying in the data and for her assistance in collecting the questionnaires during the second examination period. 8 The experimental sessions were conducted by Cynthia Miller, Lois Prettyman, Sabrina Robinson, Gloria Sanchez, and Ursula Taylor under the joint supervision of Catherine Q. E. Sim and the first author. 9 The experimental sessions were conducted by Doris Bäck, Margit Gorgi, Christine Hadler, Elke Hochsteiner, Dr. Ursula Horak, Karin Kotsis, and Peter Strempfl under the joint supervision of Dr. Georg Gittler and the first author. 7

RT44367_C009.indd 160

8/15/05 6:57:27 AM

EXPERIENCE OF ANGER AND ANXIETY EMOTIONS

161

Study 10. The German trait version of the EAAS was administered to a sample of 356 seventh and eighth graders (51% boys, 49% girls) enrolled in 15 Austrian lower secondary school classes (i.e., “gymnasium”) as part of a large-scale longitudinal study of the state and trait components of test anxiety.10 The students were instructed to report how often they experienced anger and anxiety symptoms during reading and math tests (i.e., domain-specific trait instructions). See Hochegger (1995) for a detailed description of the sample and the study procedures for Study 10. Study 11. The set of anger and anxiety items displayed in Table 9.1 was given as part of a questionnaire survey to seventh and eighth graders in several U.S. middle schools.11 Unlike the previous studies, the students were instructed to indicate how often they experienced the anger or anxiety symptom during regular classroom lessons in both math and reading (i.e., trait instructions: “How often do you have this thought or feeling during Math classes and during Reading classes?”). A total of 149 students completed the questionnaire on an anonymous and voluntary basis during a regular classroom session (48% boys, 52% girls; 76% Anglo-American, 12% Hispanic-American, 3% African-American, 1% Asian-American, 5% Other, 3% No answer). RESULTS Experience of Anger and Anxiety during Examinations: Results from Trait Studies Table 9.1 shows the means and standard deviations of the EAAS items for the three English trait studies. Although the means of the anger items were lower than those of the anxiety items for the U.S. university students in Study 1, the endorsement rates (as defined in the Method section) of 14% for “Annoyed,” “Irritated,” and “Fed up” indicated that one out of seven students gets angry frequently (i.e., “Often” or “Almost always”) during examinations. Although not part of the EAAS, the endorsement rates for “burned up” (7%) and “furious” (8%) indicated 1 of every 10 university students frequently experienced intense anger feelings. For the U.S. lower secondary school students (Study 2), the means of the anger items were considerably higher than those of the university students. A substantial percentage of the middle-school students reported that they were frequently annoyed (21%), irritated (22%), and fed up (16%) during tests. In the Singapore sample of lower secondary school students (Study 3), the means of the anger items were similar to those of the U.S. university students. It should, be noted, however, that this sample, unlike the two more representative U.S. samples, was drawn from only one Singaporean secondary school. 10 The questionnaire administration was conducted by Dr. Heinrich Hochegger as part of his doctoral dissertation project under the supervision of the first author. 11 We would like to thank Ray Gadd from the District School Board of Pasco County, Florida, for supervising the questionnaire administration.

RT44367_C009.indd 161

8/15/05 6:57:28 AM

162

N.K. TANZER AND C.D. SPIELBERGER

Table 9.1 Means and Standard Deviations for the English Trait Version of the Examination Anger and Anxiety Items University Students U.S.a

Lower Secondary School Students U.S.b

Singaporec

Mean

SD

Mean

SD

Mean

SD

Anxiety Items Uptight Uneasy Tense Nervous Jittery Getting butterflies Heart beats faster

2.25 1.91 2.12 2.12 1.79 2.13 1.94

.92 .76 .92 .95 .84 .89 .91

1.96 1.92 1.91 2.08 1.73 1.86 1.85

.92 .93 1.06 1.18 .99 .93 .99

2.24 2.05 2.20 2.25 1.83 1.97 2.35

.84 .80 .92 .93 .84 .79 .92

Anger Items Angry Annoyed Irritated Fed up Madd

1.49 1.69 1.63 1.64 1.34

.79 .80 .79 .80 .66

1.70 2.18 1.92 2.04 —

.97 1.04 .93 1.08 —

1.44 1.58 1.73 1.77 —

.71 .78 .80 .84 —

Note: For all items, a four-point frequency rating scale (1 = Hardly ever, 2 = Sometimes, 3 = Often, 4 = Almost always) was used. a Study 1 (N = 368), bStudy 2 (N = 158), c Study 3 (N = 199). dItem was not used in Studies 2 and 3.

Table 9.2 shows the oblimin-rotated factor patterns for the three English trait studies. In all three studies, the items split into two factors that can be clearly identified as Examination Trait Anxiety (Factor 1) and Examination Trait Anger (Factor 2). The two factors were strongly correlated; the factor intercorrelations ranged between .56 and .64. With only one exception, the factor patterns displayed a perfect simple-structure. All but one target loading was above .51, and all nontarget loadings were below .30. The single exception was the Emotionality item “Jittery,” which loaded in the Singaporean lower secondary school sample on Anger instead of Anxiety. Because “Jittery” was one of the items that caused substantial comprehension problems for the Singaporean primary-school students (see Method section), this deviation from the posited factor structure can be attributed to poor comprehension of “Jittery” in the Singaporean sample of seventh to ninth graders. The means, standard deviations, and factor loadings of the German EAAS trait version for the Austrian university students (Study 4) are reported in Table 9.3. With the exception of one item, “Annoyed,” with a weak loading on Factor 2, the German EAAS items had perfect simple-structure, with a correlation between factors (r = .36) that was considerably lower compared to English trait studies. Endorsement rates between 7% and 16% for the five anger items

RT44367_C009.indd 162

8/15/05 6:57:29 AM

163

EXPERIENCE OF ANGER AND ANXIETY EMOTIONS

Table 9.2 Factor Loadings for the English Trait Version of the Examination Anger and Anxiety Items University Students

Anxiety Items Uptight Uneasy Tense Nervous Jittery Getting butterflies Heart beats faster

U.S.a

U.S.b

Singaporec

Factor 1 Factor 2

Factor 1 Factor 2

Factor 1 Factor 2

.59 .56 .81 .73 .60 .64 .61

Anger Items Angry Annoyed Irritated Fed up Madd Eigenvaluese

Lower Secondary School Students

.51 .67 .70 .74 .47 .72 .78 .76 .76 .77 .60 .88

48%

12%

.59 .72 .76 .75 .53 .57 .56 .57 .71 .83 —

48%

11%

.46

.52 .61 .71 .69 — 40%

13%

Note. The Principal-Axis 2-factor solutions were oblimin rotated. Loadings less than | .30 | are omitted. aStudy 1 (factor correlation r = .61). bStudy 2 (factor correlation r = .64). c Study 3 (factor correlation r = .56). dItem was not used in Studies 2 and 3. eVariance explained by the first two unrotated Principal-Axis factors.

indicated that a substantial percentage of Austrian university students frequently (i.e., “Often” or “Almost always”) experienced angry feelings. Table 9.4 shows means and SDs of the English EAAS items for the U.S. university students during written examinations (Study 6) and during intelligence/aptitude assessments (Study 8). Consistent with the results of the trait studies, a substantial percentage of the students reported being considerably angry, annoyed, irritated, fed up, and mad during the written examinations; the pooled endorsement rates for “Quite a bit” and “Very much” ranged between 9% and 12%. Although not part of the EAAS, the endorsement rate for “Furious” (8%) indicated that 1 out of 10 university students experienced highly intense anger feelings. Every fifth student felt “Frustrated” (respective endorsement rate 20%) during the examination, but only 3% reported being very “Angry with the professor.” The means and standard deviations of the German EAAS items for the samples of Austrian university students for their recollection of their last examination (Study 5), during (i.e., immediately after) written examinations (Study 7), and during intelligence/aptitude assessments (Study 9) are reported

RT44367_C009.indd 163

8/15/05 6:57:30 AM

164

N.K. TANZER AND C.D. SPIELBERGER

Table 9.3 Means, Standard Deviations, and Factor Loadings for the German Trait Version of the Examination Anger and Anxiety Items (Austrian University Students) Mean

SD

Factor 1

Anxiety Itemsa Funny feeling in stomach Muscles being tense Heart beats faster Oppressed Tremble with fear Anxious feeling

2.64 1.54 2.09 2.02 1.78 1.97

.97 .80 .91 .94 .90 .94

.58 .64 .77 .69 .68 .72

Anger Itemsa Angry/irritated Annoyed Feel like shouting Angry/Furious Fed up

1.59 1.79 1.59 1.34 1.46

.85 .75 .86 .67 .74

Eigenvaluesb

Factor 2

.60 .36 .64 .87 .84 34%

15%

Note: Study 4 (N = 445). For all items, a four-point frequency rating scale (1 = Hardly ever, 2 = Sometimes, 3 = Often, 4 = Almost always) was used. The Principal-Axis two-factor solution was oblimin rotated (factor correlation r = .36). Loadings less than | .30 | are omitted. aClosest translation of the German items (see Appendix B). bVariance explained by the first two unrotated Principal-Axis factors.

in Table 9.5. The pooled endorsement rates for the two categories “Quite a bit” and “Very much” ranged between 10% and 17% for the five anger items, indicating once more that a substantial percentage of students experienced intense angry feelings during their midterm or final examinations. Tables 9.6 and 9.7 report the oblimin-rotated factor patterns for the two English and the three German EAAS state studies. With the exceptions of only two items (see Table 9.6, Study 8), the factor patterns in all studies displayed a perfect simple-structure with target loadings above .48 and nontarget loadings below .32. These two exceptions regarded the English EAAS anxiety items “Uneasy” and “Uptight,” which had substantial nontarget loadings on anger for the intelligence/aptitude assessments in the U.S. sample. For the state versions of both the English and the German EAAS, the anger and anxiety factors for examinations correlated more highly than for the intelligence/ aptitude assessments. Passing the course examinations was crucial for the students’ academic career, while taking the intelligence tests under the anonymous administration conditions had no consequences for their real life and was not threatening to their self-esteem (no individual feedback given). Consequently, the intelligence/aptitude assessment situation stimulated appraisal patterns that influenced the experience of different emotions (cf. Lazarus, 1991).

RT44367_C009.indd 164

8/15/05 6:57:30 AM

165

EXPERIENCE OF ANGER AND ANXIETY EMOTIONS Table 9.4 Means and Standard Deviations for the English State Version of the Examination Anger and Anxiety Items (U.S. University Students) After Written Examinationa

Anxiety Items Uptight Uneasy Tense Nervous Jittery Getting butterflies Heart beats faster Anger Items Angry Annoyed Irritated Fed up Mad

Intelligence Test Batteryb

Mean

SD

Mean

SD

1.24 .85 .93 .90 .79 .85 .63

1.20 1.08 1.19 1.15 1.12 1.01 1.01

.64 .65 1.03 .61 .56 .44 .68

.97 .94 1.10 .91 .85 .75 .99

.71 .75 .81 .67 .57

1.15 1.12 1.17 1.12 1.08

.30 .73 .78 .61 .30

.70 1.18 1.14 1.09 .80

Note: For all items, a five-point intensity rating scale (0 = Not at all, 1 = Somewhat, 2 = Moderately, 3 = Quite a bit, 4 = Very much) was used. aStudy 6 (N = 593 reports from 355 students). b Study 8 (N = 4 × 126 – 1 = 503 reports from 126 subjects.

Temporal and Situational Specificity of Anger and Anxiety during Examinations The test-retest correlations of state anxiety measures during the real examinations were r = .50 for the U.S. sample (Study 6, N = 184 students) and r = .70 in the Austrian sample of university students (Study 7, N = 50 students). The temporal stability of state anger during examinations was considerably lower, r = .31 for the U.S. sample and r = .30 for the Austrian sample. The means of the EAAS state anger and anxiety scores reported during the course examinations (“test occasions”) are reported in Tables 9.8 and 9.9 for the U.S. and the Austrian samples, respectively. The magnitude of the variation between the course/occasion means, as compared to the total variation, clearly demonstrated that the levels of state anger and state anxiety depended substantially on characteristics of the test situations that are common to all test takers, such as difficulty and fairness of the questions, and the behavior of the test administrator. In the U.S. sample, the largest difference between the anxiety and anger means for the six test occasions was about the size of one SD for the total sample of reports. In the Austrian sample, the largest difference between the anxiety means for the seven test occasions was almost one SD, while the largest difference between the anger means was 1.5 times one SD.

RT44367_C009.indd 165

8/15/05 6:57:32 AM

166

N.K. TANZER AND C.D. SPIELBERGER

Table 9.5 Means and Standard Deviations for the German State Version of the Examination Anger and Anxiety Items (Austrian University Students) Last Examinationa (Recollected State)

Immediately after Written Examinationb (Actual State)

Intelligence Test Batteryc (Actual State)

Meand

SD

Meane

SD

Meane

SD

Anxiety Items f Funny feeling in stomach Muscles being tense Heart beats faster Oppressed Tremble with fear Anxious feelingg

2.26 1.55 1.81 1.77 1.77 1.75

.91 .82 .97 .90 .89 .92

3.48 2.63 2.72 3.00 2.36 2.88

1.71 1.61 1.80 1.69 1.76 1.66

1.49 1.34 1.32 1.39 1.61 —

.93 .79 .75 .80 .97 —

Anger Items f Angry/irritated Annoyed Feel like shouting Angry/furious Fed up

1.34 1.46 1.37 1.29 1.31

.71 .78 .77 .68 .70

2.19 2.54 2.28 2.15 2.35

1.51 1.68 1.82 1.69 1.76

1.39 1.45 1.41 1.34 1.48

.98 1.04 1.11 .94 1.04

Note: aStudy 5 (N = 445 reports from 445 students). bStudy 7 (N = 269 reports from 146 students). cStudy 9 (N = 760 reports from 190 subjects). dA four-point intensity rating scale (with verbal anchors at 1 = Not at all, 2 = Somewhat, 3 = Quite a bit, 4 = Very much) was used. eA seven-point graphical intensity rating scale (1 = Not at all, 3 = Somewhat, 5 = Quite a bit, and 7 = Very much) was used. fClosest translation of the German items (see Appendix B). gNot used in Study 9.

For the intelligence/aptitude assessment studies, the EAAS mean state anger and state anxiety scores for the baseline, and for each of the three intelligence/aptitude tests, are reported in Table 9.10 for the U.S. sample (Study 8) and in Table 9.11 for the Austrian sample (Study 9). The three intelligence/aptitude tests (i.e., Bourdon test, reasoning test, and spatial ability test) evoked quite different emotional responses within and between the two samples. In both studies, mean state anger increased from the baseline level throughout the test sequence. Compared to the baseline, the mean state anxiety scores in the U.S. sample were equally high for all three intelligence tests. State anxiety in the Austrian sample increased steeply from the baseline level, reaching its maximum during the Bourdon test, and dropped back to the initial baseline level during the spatial ability test. Unlike most U.S. university students, the Austrian students had taken mandatory courses in technical drawing during their secondary school. Consequently, the spatial ability test materials, namely the drawings of cubes and the mental operations needed for solving the problems, were more familiar to them than to their U.S. counterparts.

RT44367_C009.indd 166

8/15/05 6:57:33 AM

167

EXPERIENCE OF ANGER AND ANXIETY EMOTIONS

Table 9.6 Factor Loadings for the English State Version of the Examination Anger and Anxiety Items (U.S. University Students) After Written Examinationa Factor 1 Anxiety Items Uptight Uneasy Tense Nervous Jittery Getting butterflies Heart beats faster

.70 .58 .69 .79 .69 .85 .68

Anger Items Angry Annoyed Irritated Fed up Mad Eigenvaluec

Factor 2

.32

Intelligence Test Batteryb Factor 1

Factor 2

.48 .55 .66 .82 .84 .74 .71

.43 .39

.79 .86 .75 .85 .88 66%

8%

.75 .89 .87 .83 .82 46%

23%

Note: The Principal-Axis two-factor solutions were oblimin rotated. Loadings less than .30 are omitted. aStudy 6 (factor correlation r = .78). bStudy 8 (factor correlation r = .29). c Variance explained by the first two unrotated Principal-Axis factors.

Academic Domain Specificity and Course (Teacher) Specificity of Trait Anger and Trait Anxiety during Examinations Table 9.12 reports the means and standard deviations of the German EAAS items during the reading and math tests for the Austrian lower secondary school students (Study 10). When compared to the reading tests, the math tests evoked more intense anxiety and, to a lesser degree, stronger anger reactions. Once again, a substantial percentage of the students frequently (i.e., “Often” or “Almost always”) experienced angry feelings; the endorsement rates for the five anger items were between 8% and 12% for the math tests, and were 6% to 7% for the reading tests. However, compared to the Austrian university students (see Table 9.3, Study 4), the Austrian lower secondary school students reported less anger and anxiety. The factor loadings of the four posited factors (i.e., Reading and Math test anxiety and test anger) are shown in Table 9.13. These loadings revealed a perfect simple-structure, with all target loadings higher than .45; all nontarget loadings were below .33. In terms of a multitrait-multimethod paradigm, the factor correlations were equally high for between-domain/within-emotions factors (r = .32 for anxiety; r = .45 for anger), and between-emotions/within-domain

RT44367_C009.indd 167

8/15/05 6:57:34 AM

168

N.K. TANZER AND C.D. SPIELBERGER

Table 9.7 Factor Loadings for the German State Version of the Examination Anger and Anxiety Items (Austrian University Students) Last Examinationa (Recollected State) Factor 1 Anxiety Itemsd Funny feeling in stomach Muscles being tense Heart beats faster Oppressed Tremble with fear Anxious feelinge

.69 .68 .84 .78 .76 .81

Anger Itemsd Angry/irritated Annoyed Feeling like shouting Angry/furious Fed up Eigenvaluesf

Factor 2

Immediately after Written Examinationb (Actual State) Factor 1 .68 .76 .85 .73 .85 .73

.76 .55 .79 .92 .90 46%

Factor 2

16%

Intelligence Test Batteryc (Actual State) Factor 1 .69 .65 .66 .80 .73 —

.78 .90 .79 .89 .93 51%

Factor 2

17%

.90 .92 .76 .92 .84 42%

21%

Note: The Principal-Axis two-factor solutions were oblimin rotated. Loadings less than .30 are omitted. aStudy 5 (N = 445 reports, factor correlation r = .46). bStudy 7 (N = 269 reports, factor correlation r = .48). cStudy 9 (N = 760 reports, factor correlation r = .26). dClosest translation of the German items (see Appendix B). eNot used in Study 9. fVariance explained by the first two unrotated Principal-Axis factors.

factors (r = .47 for Reading; r = .43 for Math). These findings demonstrate the emotion-specificity (anger vs. anxiety) and domain-specificity (Reading vs. Math) of examination stress. The intraclass correlations for all four scales were not significant (p > .10). The between-classes variance was less than 4% for Reading and Math anxiety, and less than 1% for Reading and Math anger, indicating that the teachers had no substantial impact on the long-term test anxiety and anger of their students. These results contrasted with the substantial course specificity of anger and anxiety that was found in the previous section. It should be noted, however, that trait measures were used in this study, whereas the results reported in the previous section were based on state measures. Trait Anger and Anxiety during Regular Classroom Lessons Table 9.14 shows the means, standard deviations, and factor loadings for the symptoms of anxiety and anger reported by a sample of U.S. lower secondary school students during regular classroom lessons (Study 11). Although

RT44367_C009.indd 168

8/15/05 6:57:35 AM

169

EXPERIENCE OF ANGER AND ANXIETY EMOTIONS Table 9.8 Means and Standard Deviations of State Anger and State Anxiety Experienced during Written Examinations (U.S. University Students) Examination Group (Course)

Anxiety

Anger

Mean

SD

Mean

SD

Reports

1st Examination Group (Course) a Group b Group c

.85 .72 1.18

.93 .79 .97

.31 .55 1.04

.79 .85 1.16

50 192 70

2nd Examination Group a Group b Group c

1.14 .64 1.49

.87 .83 1.13

1.00 .49 1.39

.93 .86 1.35

67 156 58

Total Sample of Reports

.88

.92

.70

1.00

593

Maximum Mean Difference

.85

1.08

Note: Study 6 (N = 593 reports from 355 students). For all items, a five-point intensity rating scale (0 = Not at all, 1 = Somewhat, 2 = Moderately, 3 = Quite a bit, 4 = Very much) was used.

each anger symptom was reported separately for the reading and math classes, the students’ Reading and Math ratings were averaged in order to calculate the “global” ratings (i.e., not domain-specific) displayed in Table 9.14. Similar to the results for the examination situations (see Table 9.1, Study 2), the means for the anger symptoms were as high as, if not higher than, the means for the anxiety symptoms. The factor loadings again showed almost perfect simple-structure; only one item (“Uneasy”) had a weak target loading. The two factors were highly intercorrelated (r = .58), as was found in the other English trait studies. DISCUSSION In the factor analyses of all 11 studies, anger and anxiety emerged as distinct emotions. Like anxiety, anger was frequently experienced in achievementrelated situations, and similar results were found across different cultures and educational systems (Austria, Singapore, United States), different age groups (lower secondary school students and university students), and achievementrelated situations (evaluation of course examinations and intelligence/aptitude assessments), and by learning in regular classroom periods in different academic domains such as math and reading. The anger and anxiety factors were also quite similar for the trait and state measures of these emotional reactions. The results of the pilot studies also demonstrated that some of the items that have not been previously included in test anxiety scales loaded on both anger and anxiety factors. These items seem to inquire about a blend of several negative

RT44367_C009.indd 169

8/15/05 6:57:36 AM

170

N.K. TANZER AND C.D. SPIELBERGER

Table 9.9 Means and Standard Deviations of State Anger and State Anxiety Experienced during Written Examinations (Austrian University Students) Examination Group (Course)

Anxiety

Anger

Mean

SD

Mean

SD

Reports

1st Examination Group (Course) a Group b Group c

2.35 2.59 2.39

1.11 1.12 1.10

1.40 1.72 1.63

.89 0.86 1.16

42 41 40

2nd Examination Group a Group b Group c

3.03 3.51 3.09

1.27 1.66 1.57

2.70 3.65 2.37

1.71 1.81 1.24

44 28 39

Reexamination

2.58

1.19

3.20

1.74

35

Total Sample of Reports

2.76

1.32

2.29

1.54

269

Maximum Mean Difference

1.16

2.25

Note: Study 7 (N = 269 reports from 146 students). For all items, a seven-point graphical intensity rating scale (with verbal anchors at 1 = Not at all, 3 = Somewhat, 5 = Quite a bit, and 7 = Very much) was used.

Table 9.10 Means and Standard Deviations of State Anger and State Anxiety Experienced during an Intelligence Test Battery (U.S. University Students) Anxiety Test

Anger

Mean

SD

Mean

SD

Reports

Baseline (Begin) Bourdon Test (Attention) Matrices Test (Reasoning) Cube Comparison Test (Spatial Ability)

.40 .79 .75 .75

.60 .72 .72 .73

.18 .34 .79 1.00

.50 .64 .90 1.09

126 126 126 125

Total Sample of Reports

.67

.71

.58

.88

503

Note: Study 8 (N = 503 reports from 126 subjects). For all items, a five-point intensity rating scale (0 = Not at all, 1 = Somewhat, 2 = Moderately, 3 = Quite a bit, 4 = Very much) was used.

emotions (e.g., “Upset” or “Oppressed”), or a state of nonspecific hyperarousal associated with both anger and anxiety, and other negative emotions (e.g., “Being in a state of turmoil,” “Feeling high-strung,” or “Feeling overexcited”). Consequently, in order to ensure factorial validity, items related to anger and

RT44367_C009.indd 170

8/15/05 6:57:37 AM

171

EXPERIENCE OF ANGER AND ANXIETY EMOTIONS Table 9.11 Means and Standard Deviations of State Anger and State Anxiety Experienced during an Intelligence Test Battery (Austrian University Students) Anxiety Test

Anger

Mean

SD

Mean

SD

Reports

Baseline (Begin) Bourdon Test (Attention) Matrices Test (Reasoning) Cube Comparison Test (Spatial Ability)

1.35 1.57 1.43 1.37

.56 .75 .68 .60

1.17 1.33 1.49 1.66

.47 .79 .97 1.20

190 190 190 190

Total Sample of Reports

1.43

.66

1.41

.91

760

Note: Study 9 (N = 760 reports from 190 subjects). For all items, a seven-point graphical intensity rating scale (with verbal anchors at 1 = Not at all, 3 = Somewhat, 5 = Quite a bit, and 7 = Very much) was used.

Table 9.12 Means and Standard Deviations for the Domain-Specific German Trait Version of the Examination Anger and Anxiety Items (Austrian Lower Secondary School Students) Reading

Mathematics

Mean

SD

Mean

SD

Mean Difference

Anxiety Items a Funny feeling in stomach Muscles being tense Heart beating faster Oppressed Tremble with fear Anxious feeling

1.51 1.26 1.43 1.38 1.42 1.34

.83 .64 .83 .73 .77 .72

1.95 1.36 1.68 1.56 1.65 1.70

.91 .65 .93 .82 .87 .91

.44 .10 .25 .18 .23 .36

Anger Items a Angry/irritated Annoyed Feel like shouting Angry/furious Fed up

1.24 1.32 1.30 1.28 1.30

.61 .68 .68 .68 .69

1.38 1.40 1.46 1.40 1.40

.69 .73 .84 .76 .73

.14 .08 .16 .12 .10

Note: Study 10 (N = 356). For all items, a four-point frequency rating scale (1 = Hardly ever, 2 = Sometimes, 3 = Often, 4 = Almost always) was used. aClosest translation of the German items (see Appendix B).

other negative emotions should routinely be included in the development of anxiety scales or the translation of such scales. Regarding the correlations between anger and anxiety, the picture is less clear. For a sample of U.S. university students, Smith and Ellsworth (1987) reported

RT44367_C009.indd 171

8/15/05 6:57:38 AM

172

N.K. TANZER AND C.D. SPIELBERGER

Table 9.13 Factor Loadings for the Domain-Specific German Trait Version of the Examination Anger and Anxiety items (Austrian Lower Secondary School Students) Anxiety Factor 1 Anxiety Items Reading: Funny feeling in stomach Muscles being tense Heart beats faster Oppressed Tremble with fear Anxious feeling

Factor 2

Factor 1

Factor 2

.71 .72 .65 .74 .86 .78

Mathematics: Funny feeling in stomach Muscles being tense Heart beats faster Oppressed Tremble with fear Anxious feeling Anger Items a Reading: Angry/irritated Annoyed Feel like shouting Angry/furious Fed up

Anger

.85 .60 .58 .45 .53 .70

.58 .53 .80 .82 .58

.33

Mathematics: Angry/irritated Annoyed Feel like shouting Angry/furious Fed up

.58 .94 .65 .76 .93 Factor Intercorrelations

Factor 1 Factor 2 Factor 3 Factor 4

— .32 .47 .20

.32 — .02 .43

.47 .02 — .45

.20 .43 .45 —

Eigenvalueb

36%

14%

11%

3%

Note: Study 10 (N = 356). The Principal-Axis four-factor solution was oblimin rotated. Loadings less than | .30 | are omitted. aClosest translation of the German items (see Appendix B). bVariance explained by the first four unrotated Principal-Axis factors.

RT44367_C009.indd 172

8/15/05 6:57:39 AM

173

EXPERIENCE OF ANGER AND ANXIETY EMOTIONS

Table 9.14 Means, Standard Deviations, and Factor Loadings of Anger and Anxiety Symptoms Experienced during Regular Classroom Lessons (U.S. Lower Secondary School Students) Factor Loadings Mean

SD

Factor 1

Factor 2

Anxiety Items Uptight Uneasy Anxious and tense Nervous Jittery Panicky Scared

1.77 1.83 1.85 1.87 1.83 1.70 1.44

.80 .75 .88 .86 .90 .85 .71

.81 .33 .69 .47 .58 .51 .57

.31

Anger Items Angry Annoyed Irritated Fed up

1.74 2.18 1.86 2.03

.82 .92 .85 .88

Eigenvaluea

.68 .65 .37 .56 42%

11%

Note: Study 11 (N = 149). For all items, a four-point frequency rating scale (1 = Hardly ever, 2 = Sometimes, 3 = Often, 4 = Almost always) was used. The Principal-Axis 2-factor solution was oblimin rotated (factor correlation r = .58). Loadings less than | .30 | are omitted. aVariance explained by the first two unrotated Principal-Axis factors.

correlations of .42 and .63 for state measures during pre-exam and postfeedback conditions, respectively. Likewise, for a sample of German university students, Pekrun (1997) reported correlations of .57 and .63 for state measures during examinations and learning situations, respectively. Although the correlations were similarly high for the English trait and state studies (between r = .56 and r = .78), we found considerably lower correlations in the German trait and state studies (between r = .36 and r = .48). As such, our findings are only partially in accord with the results previously reported in the literature. For the intelligence/aptitude assessment situations, which were less threatening to the students’ careers, the correlations of .26 and .29 for the state anger and anxiety measures were consistently lower than those for examinations. In summary, the results of the studies reported in this chapter showed that some students frequently experienced high levels of anger in achievementrelated situations, whereas other students with the same level of anxiety did not. Another interesting observation is that items addressing external attributions of blame (e.g., “The test questions were unfair”) load on anger, while items related to the internal attribution of blame (e.g., “I should have studied more”) do not. Although only speculation regarding the reasons for these individual differences can be made, including instruments in future studies to assess anger

RT44367_C009.indd 173

8/15/05 6:57:39 AM

174

N.K. TANZER AND C.D. SPIELBERGER

expression and control (e.g., Spielberger, 1988), and the causal attributions of success and failure (e.g., Weiner, 1986; Weiner, Russel, & Lerman, 1979), may help to clarify these questions.

REFERENCES Bond, M. (1991). Beyond the Chinese face. Hong Kong: Oxford University Press. Brickenkamp, R. (1974). d2 Aufmerksamkeits-Belastungs-Test. [Concentration test d2.] Göttingen, Germany: Hogrefe. Elkind, D. (1986). Stress and the middle grader. School Counselor, 33, pp. 196–206. Folkman, S., & Lazarus, R. S. (1985). If it changes it must be a process: Study of emotion and coping during three stages of a college examination. Journal of Personality and Social Psychology, 48, pp. 150–170. Forman, A. K., & Piswanger, K. (1979). Wiener Matrizen-Test. Ein Raschskalierter sprachfreier Intelligenztest. The Viennese Matrices Test: A Raschcalibrated nonverbal intelligence test. Weinheim, Germany. Beltz Test. Gittler, G. (1990). Dreidimensionaler Würfeltest 3DW. Ein Rasch-skalierter test zur messung des räumlichen vorstellungsvermögens. Theoretische grundlagen und manual [Three-dimensional cube comparison test 3DC. A Rasch-calibrated spatial ability test. Theoretical conception and test manual] Germany: Weinheim Beltz Test. Goldstein, H. (1987). Multilevel models in educational and social research. London & New York: Charles Griffin & Oxford University Press. Hambleton, R. K. (1991). Test translation for cross-cultural studies. Bulletin of the International Test Commission, 18 (1 & 2). Hochegger, H. (1995). State-trait komponenten von prüfungsangst. Eine längsschnittstudie an österreichischen AHS-unterstufen-schülern. [State-trait components of test anxiety: A longitudinal study on Austrian lower secondary school students]. Unpublished doctoral dissertation. University of Graz. Kim, J.-O., & Mueller, C. W. (1978). Factor analysis: Statistical methods and practical issues. Beverly Hills, CA: Sage. Lazarus, R. S. (1991). Emotion and adaption. New York/Oxford: Oxford University Press. Liebert, R. M., & Morris, L. W. (1967). Cognitive and emotional components of test anxiety: A distinction and some initial data. Psychological Reports, 20, pp. 975–978. Mesquita, B., & Frijda, N. H. (1992). Cultural variations in emotions: A review. Psychological Bulletin, 112, pp. 179–204. Morris, L. W., Davis, M. A., & Hutchings, C. H. (1981). Cognitive and emotional components of anxiety: Literature review and a revised Worry-Emotionality scale. Journal of Educational Psychology, 73, pp. 541–555. Paterson, L., & Goldstein, H. (1991). New statistical models for analysing social structures: An introduction to multilevel models. British Educational Research Journal, 17, pp. 387–393.

RT44367_C009.indd 174

8/15/05 6:57:40 AM

EXPERIENCE OF ANGER AND ANXIETY EMOTIONS

175

Pederson, D. J., & Hollandsworth, J. G. (1987). The effects of an annoying test administrator on effective and ineffective test-takers during an actual test. Educational and Psychological Research, 7, pp. 191–199. Pekrun, R. (1994). Academic emotions in students’ self-regulated learning and achievement. Paper presented at the Annual Meeting of AERA, New Orleans, April. Unpublished manuscript, University of Regensburg, Germany. Pekrun, R. (1997). Emotionen in Lern- und Leistungssituationen. [Emotions experienced in learning situations and examination situations]. Paper presented at the symposium “Motivation und Lernen,” Vienna, Austria. Pekrun, R., & Frese, M. (1992). Emotions in work and achievement. In C. L. Cooper & I. T. Robertson (Eds.), International Review of Industrial and Organizational Psychology (Vol. 7, pp. 153–200). Chichester, UK: Wiley. Polivy, J. (1981). On the induction of emotion in the laboratory: Discrete moods or multiple affect states? Journal of Personality and Social Psychology, 41, pp. 403–417. Prosser, R., Rasbash, J., & Goldstein, H. (1991). ML3: Software for three-level analysis. User’s guide. London: University of London. Raudenbush, S. W., Rowan, B., & Jin Kang, S. (1991). A multilevel, multivariate model for studying school climate with estimation via the EM algorithm and application to U.S. high-school data. Journal of Educational Statistics, 16, pp. 295–330. Sarason, I. G. (1984). Stress, anxiety, and cognitive interference: Reactions to tests. Journal of Personality and Social Psychology, 46, pp. 929–938. Smith, C. A., & Ellsworth, P. C. (1985). Patterns of cognitive appraisal in emotion. Journal of Personality and Social Psychology, 48, pp. 813–838. Smith, C. A., & Ellsworth, P. C. (1987). Patterns of appraisal and emotion related to taking an exam. Journal of Personality and Social Psychology, 52, pp. 475–488. Spielberger, C. D. (1972). Anxiety as an emotional state. In C. D. Spielberger (Ed.), Anxiety: Current trends in theory and research (Vol. 1, pp. 23–49). New York: Academic Press. Spielberger, C. D. (1973). State-Trait Inventory for Children (STAI C) manual. Palo Alto, CA: Consulting Psychologists Press. Spielberger, C. D. (1979). State-Trait Personality Inventory (STPI): Preliminary manual. Unpublished manuscript, Department of Psychology, University of South Florida, Tampa. Spielberger, C. D. (1980). Test Anxiety Inventory (TAI): Professional manual. Palo Alto, CA: Consulting Psychologists Press. Spielberger, C. D. (1983a). Manual for the State-Trait Anxiety Inventory (STAI: Rev. ed.). Palo Alto, CA: Consulting Psychologists Press. Spielberger, C. D. (1983b). State-Trait Anxiety Inventory: A comprehensive bibliography. Palo Alto, CA: Consulting Psychologists Press. Spielberger, C. D. (1985). Assessment of state and trait anxiety: Conceptual and methodological issues. The Southern Psychologist, 2, pp. 6–16. Spielberger, C. D. (1988). State-Trait Anger Expression Inventory (STAXI: Research Ed,): Professional manual. Odessa, FL: Psychological Assessment Resources.

RT44367_C009.indd 175

8/15/05 6:57:41 AM

176

N.K. TANZER AND C.D. SPIELBERGER

Spielberger, C. D., Jacobs, G., Russell, S., & Crane, R. J. (1983). Assessment of anger: The State-Trait Anger Scale. In J. N. Butcher & C. D. Spielberger (Eds.), Advances in personality assessment (Vol. 2, pp. 159–187). Hillsdale, NJ: LEA. Spielberger, C. D., Krasner, S. S., & Solomon, E. P. (1988). The experience, expression, and control of anger. In M. P. Janisse (Ed.), Health psychology: Individual differences and stress (pp. 89–108). New York: Springer. SPSS. (1990). SPSS reference guide. Chicago: SPSS. Tanzer, N. K. (1992). A component- and domain-specific concept of test anxiety: Theoretical considerations and first empirical support. (Research Report 1992/1). Graz, Austria: Department of Psychology, University of Graz. Tanzer, N. K., Gittler, G., & Sim, C. Q. E. (1994). A cross-cultural comparison of a Rasch-calibrated spatial ability test between Austrian and Singaporean adolescents. In A. M. Bouvy, F. J. R. van de Vijver, P. Bowski, & P. Schmitz (Eds.), Journeys into cross-cultural psychology (pp. 96–110). Lisse, The Netherlands: Swets & Zeitlinger. Tanzer, N. K., Hochegger, H., & Linhart, M. (1993). Gegenstandsspezifische Prüfungsangst sowie Kompetenz- und Unterrichtsklimawahrnehmung bei AHS-Unterstufenschülern. [Domain specificity of test anxiety, academic self-concept and instructional climate in Austrian lower secondary schools.] (Research Report 1993/1). Graz, Austria: Department of Psychology, University of Graz. Tanzer, N. K., Hochegger, H., & Schwetz, H. (1994, July). The impact of academic domain and test form on test anxiety. Paper presented at the 15th International Conference of the Stress and Anxiety Research Society (STAR), Madrid, Spain. Tanzer, N. K., Sim, C. Q. E., & Marsh, H. W. (1992). Using personality and attitude inventories over cultures. Theoretical considerations and empirical findings. Bulletin of the International Test Commission, 19, pp. 151–172. Törestad, B. (1990). What is anger provoking? Provokers and provoked in anger situations: Developmental trends and sex differences. Aggressive Behavior, 16, pp. 353–359. Weiner, B. (1986). An attributional theory of motivation and emotion. New York: Springer. Weiner, B., Russel, D., & Lerman, D. (1979). The cognition-emotion process in achievement-related contexts. Journal of Personality and Social Psychology, 37, pp. 1211–1220. Winer, B. J. (1971). Statistical principles in experimental design. New York: McGraw-Hill. Wu, D. Y. H. (1991). The construction of Chinese and non-Chinese identities. Daedalus. Journal of the American Academy of Arts and Sciences, 120, pp. 159–179.

RT44367_C009.indd 176

8/15/05 6:57:42 AM

EXPERIENCE OF ANGER AND ANXIETY EMOTIONS

177

APPENDIX A ITEMS OF THE ENGLISH VERSION A. Emotionality 1. I am uptight 2. I feel uneasy 3. I feel tense 4. I feel nervous 5. I feel jittery 6. I get butterflies 7. My heart beats faster B. Anger 1. I feel angry 2. I feel annoyed 3. I feel irritated 4. I am fed up 5. I am mad

APPENDIX B Items of the German Version A. Emotionality 1. Ich spüre ein komisches Gefühl im Magen 2. Ich bin am ganzen Körper verkrampft 3. Ich spüre mein Herz pochen 4. Ich habe ein beklemmendes Gefühl 5. Ich zittere vor Aufregung 6. Ich fühle mich ängstlich B. Anger 1. Ich bin wütend und gereizt 2. Ich bin verärgert 3. Ich möchte am liebsten laut schimpfen 4. Ich fühle mich richtig/zornig 5. Ich bin sauer und “angefressen”

RT44367_C009.indd 177

Closest English Translation 1. I have a funny feeling in the stomach 2. My muscles are tense 3. My heart beats faster 4. I feel oppressed 5. I tremble with fear 6. I feel anxious 1. 2. 3. 4. 5.

I feel angry/irritated I feel annoyed I feel like shouting I am angry/furious I am fed up

8/15/05 6:57:42 AM

RT44367_C009.indd 178

8/15/05 6:57:43 AM

10 Family Cohesiveness, Flexibility, and Maternal Anger: Boon or Detriment to Children’s Adjustment? Solly Dreman Ben Gurion University of the Negev, Israel1 Abstract This chapter examines relationships between family structure, parental anger, and children’s behavior problems. Divorced mothers were tested in a multiple respondent design, along with one of their adolescent children. The mothers and the children both rated the child’s behavior problems. When mothers rated these problems, the children of families with the highest level of cohesion and adaptability were found to have the fewest behavior problems, and those with the lowest levels had the most problems. When children performed the ratings, higher levels of family cohesion/adaptability were associated with more behavior problems. The daughters of mothers with higher levels of state anger had more behavior problems. In contrast, boys whose mothers were higher in state anger had fewer problems. Mothers with moderate levels of state anger rated their sons as having more behavior problems, while boys whose mothers were low in maternal anger tended to rate their own behavior problems as more severe. These findings were interpreted as reflecting the needs of adolescent children for clear role hierarchies, family stability, and parental assertiveness in order to promote optimal adjustment.

FAMILY COHESIVENESS, FLEXIBILITY, AND MATERNAL ANGER: BOON OR DETRIMENT TO CHILDREN’S ADJUSTMENT? This chapter examines relationships between family structure, as expressed in family cohesion and adaptability, and the behavior problems of children. Olson, Portner, and Lavee (1985) observed that the relation between cohesion or adaptability and adjustment was curvilinear; the highest and lowest levels of these variables were related to poorer adjustment in individuals and families. In a later formulation, Olson (1997) suggested that nonclinical families do not exhibit extreme ranges of cohesion and adaptability. Hence, there should be a positive linear relation between family cohesion and adjustment outcomes. Correspondence concerning this chapter should be addressed to Professor Solly Dreman, Department of Behavioural Sciences, Ben Gurion University of the Negev, P.O. Box 654, Beer Sheva 84105, Israel, E-mail: [email protected]. 1

179

RT44367_C010.indd 179

7/13/05 9:28:24 AM

180

S. DREMAN

In a study conducted in Israel, Dreman and Ronen-Eliav (1997) predicted that higher levels of family cohesion and adaptability perceived by divorced mothers would be related to fewer behavior problems in their children. This prediction was based on the assumption that divorced families are undergoing a situationally based life crisis and that higher levels of cohesion and adaptability should be positively related to adjustment in these families. The author (1995) also predicted that higher levels of maternal anger in divorced women would be related to more behavior problems in their children because previous research has shown that angry mothers tend to judge their children more negatively (Dix, Reinhold, & Zambarano, 1990), leading to negative parentchild interactions (Tsemach, 1996). Parental anger also contributes to both diminished parental cooperation and increased conflict, factors that have been found to have negative effects on children’s adjustment in both divorced and two-parent families. Another factor that might influence adjustment is imitation, as postulated in Bandura’s (1962) social learning theory, which predicts that angry parents who serve as negative role models directly affect their children’s behavior and adjustment. Therefore, very low levels of maternal anger might be related to fewer behavioral problems in children. However, low maternal anger might result in more behavior problems if associated with extreme suppression or repression of anger. In the present study, a multiple-respondent design was employed in which both mothers and their children rated the child’s behavior problems in order to obtain different perspectives of the child’s adjustment. The children’s ratings were considered essential in this study because Dreman and Aldor (1994) found that divorced mothers who reported high levels of family cohesion and adaptability and rated themselves as low in anger appeared to project these positive feelings in the ratings of their children by reporting fewer behavior problems. Obtaining children’s ratings controlled for possible respondent biases in their mothers’ ratings. Because research has shown that boys in postdivorce maternal custody displayed poorer adjustment than girls (Zaslow, 1988), children’s gender was also examined in the present study. A possible explanation of this finding is that divorced mothers have difficulty in establishing parental control and in disciplining their male children in the absence of a father figure (Hetherington, 1981; Hetherington, Cox, & Cox, 1985; Hetherington, Stanley-Hagan, & Anderson, 1989). METHOD The participants in this study were divorced mothers with the custody of their children, who were at least four years postdivorce. This group was chosen in order to measure the long-term effects of family structure and maternal anger on children’s behavior problems under relatively stable conditions, rather than during the crisis period immediately following divorce. Because the divorced mothers’ group approximated the normative population, the relations between

RT44367_C010.indd 180

7/13/05 9:28:24 AM

FAMILY COHESIVENESS, FLEXIBILITY, AND MATERNAL ANGER

181

cohesion, adaptability, and maternal anger with children’s behavior problems may be more generalizable to two-parent families. The study participants were randomly selected from an inclusive national computerized list of 14,000 divorced custodial mothers receiving child stipend payments from the National Insurance Institute in Israel. A random sample of 2,126 divorced mothers was contacted by a letter from the Institute, requesting their participation in a national study of coping and adjustment in divorced families. Because of the mobility of this population, many of the letters were not answered, nor returned. Although the positive response rate was only 18.5% of the letters that were delivered (393 mothers), this level of participation compares favorably with other family survey studies conducted in Israel (Hornick & Maier, 1989). A second letter was subsequently sent to mothers of the Jewish faith (85% of the divorcees) who both responded and agreed to participate in the study (337 mothers). Jewish mothers were selected because they constitute the major religious group in Israel and there are pronounced cultural differences in child-rearing practices between Jewish and other religious denominations. This second letter explained the purpose of the research and was accompanied by a demographic questionnaire designed to clarify the resident members and present living arrangements of the reconstituted family. Only divorced mothers who had at least one child between 6 and 18 years of age were included in the first part of the study. Remarried mothers, or those who had remarried and again divorced, were excluded. The research focused on single-parent custodial households resulting from a first divorce in which the mother was living without an adult partner. The final sample for the first part of the study consisted of 280 custodial mothers who agreed to participate, and one of their children who was between 11 and 18 years of age. In the second part of the study, the sample was further reduced in order to focus on the long-term effects of maternal anger. Only mothers who were four or more years past the divorce event, which is generally considered the period of stabilization of the divorce process (Herz Brown, 1988), were selected. In addition, only mothers who had a child between the ages of 11 and 18 were selected, and one child in this age range was selected from each household to fill out questionnaires (60 girls, 59 boys). Controls for age and birth order were also employed. Given the selection criteria and a 20% dropout, the sample for the second part of the study consisted of 119 custodial mothers. These mothers were similar in ethnicity, education, and number of children in custody, as compared with the general population of divorced Jewish mothers. However, the sample was positively skewed in terms of the age distribution of the mothers (mean age = 41.87 years) and their children (mean age = 14.92 years). The mean time since divorce for the participants in the second part of the study was eight years, five months (SD = three years, two months). Measures State-Trait Anger Expression Inventory (STAXI). The 44-item STAXI has been used extensively in studies of stress and emotion; good reliability and

RT44367_C010.indd 181

7/13/05 9:28:25 AM

182

S. DREMAN

validity have been consistently demonstrated for this inventory (Spielberger, 1988). Two STAXI scales measure the experience of anger: State Anger (S-Anger) and Trait Anger (T-Anger). Four scales assess anger expression and control: the Anger-Out (AX/Out), Anger-In (AX/In), Anger-Control (AX/Con), and Anger-Expression (AX) Index. In responding to the STAXI, the participants in this study were informed that the questionnaire inquired about their feelings, attitudes, and behavior. The Hebrew adaptation of the STAXI (Dreman, 1996) was used in this study. Only the STAXI S-Anger, AX/Out, and AX/In scales were administered. The S-Anger scale consists of 10 items (e.g., “I feel angry”) that ask respondents to rate how they feel, “right now,” on a four-point intensity scale, with response choices ranging from “Not at all” to “Very much so.” In responding to the AX/Out and AX/In scales, each comprised of eight items, subjects are asked to report, on a four-point frequency scale, how often they express anger, with response choices ranging from “Almost never “ to “Almost always.” Examples of AX/Out and AX/In items are “When I get mad, I say nasty things,” and “When I get angry, I keep things in.” Alpha coefficients for adult females reported in the American STAXI Test Manual (Spielberger, 1988) for the S-Anger, AX/Out, and AX/In scales were .91, .78, and .81, respectively. The STAXI S-Anger scale was chosen to measure the intensity of the experience of anger. In previous research with women in families of divorce (Dreman & Aldor, 1994), scores on this scale were related to children’s adjustment. For the present study, it was hypothesized that S-Anger was influenced by ongoing stressors (e.g., single parenting, diminished socioeconomic standing, work overload) that directly affected divorced mothers in the postdivorce situation. Because S-Anger is experienced as an internal state, it cannot be directly imitated by children, but might indirectly affect children’s adjustment if expressed in increased maternal assertiveness or hostility. The AX/Out and AX/In scales, which evaluate behavior styles that may be directly imitated by children, were used to evaluate the effects of maternal anger expression on children’s “externalizing” and “internalizing” of behavior problems (Achenbach & Edelbrock, 1983). Family Adaptability and Cohesion Evaluation Scales (FACES III). The FACES III is a well-established instrument in family research (Olson et al., 1985), consisting of 20 items that are each rated on a five-point Likert scale. The Hebrew version of FACES III (Teichman & Navon, 1990) was used in this study to evaluate the adaptability and cohesion of the Israeli participants. Examples of items measuring family cohesion and adaptability are “Family members ask each other for help,” and “Different people act as leaders in our family.” Each of these FACES III dimensions defines a continuum that consists of four levels, ranging from low to high levels of cohesion and adaptability. The degrees of cohesion, from the lowest to the highest level, are disengaged, separated, connected, and enmeshed. The four adaptability levels, from lowest to highest, are rigid, structured, flexible, and chaotic. The intermediate levels of adaptability and cohesion for each FACES III dimension were defined for the present study as being less than one standard deviation above or below the mean; the highest and lowest levels were defined

RT44367_C010.indd 182

7/13/05 9:28:25 AM

FAMILY COHESIVENESS, FLEXIBILITY, AND MATERNAL ANGER

183

by scale scores of more than one standard deviation above or below the mean. For the cohesion dimension, the frequency distribution for the divorced mothers was disengaged = 15, separated = 43, connected = 38, and enmeshed = 23. For adaptability, the distribution was rigid = 16, structured = 43, flexible = 43, and chaotic = 17. Child Behavior Checklist (CBCL). The Hebrew adaptation of the parent form of the CBCL (Achenbach & Edelbrock, 1983) developed by Zilber, Auerbach, and Lerner (1994) was used to assess the Israeli mother’s perceptions of her child’s behavior problems, “now or in the last 6 months.” The mothers responded to the 118 questions, rating each item on the following three-point checklist: 0: Not true, 1: Somewhat or sometimes true, 2: Very true or often true. Responses to the CBCL items yield a total behavior problem score based on responses for all 118 items and separate scores for externalizing and internalizing problem behaviors. The CBCL Total Behavior Problems scale (e.g., “Poor peer relations”) included three narrow-band subscales: Social problems, Thought problems, and Attention problems. The CBCL Externalizing (e.g., “Attacks people”) and Internalizing (e.g., “Likes to be alone”) scales assess two broad dimensions, which were identified in factor analysis of the items describing behavior problems in the development of the CBCL (Achenbach & Edelbrock, 1983). The Externalizing scale includes two narrow-band subscales that assess Delinquent and Aggressive behaviors. The Internalizing scale is comprised of three subscales: Withdrawn, Somatic Complaints, and Anxious or Depressed. When comparing scores on the CBCL scales, it should be noted that the score for the Total Behavior Problems scale is not identical to the sum of the scores of the Externalizing and the Internalizing scales. Youth Self-Report (YSR). The YSR was used to assess the children’s perceptions of their own behavior problems “now or in the last 6 months” (Achenbach & Edlelbrock, 1987). This scale was standardized on children 11 years of age and older, which was consistent with the age range for the children in the present study. The items comprising the YSR are identical to 113 of the 118 CBCL items and are rated on the same three-point scale, yielding a Total Behavior Problems score for all items combined and separate scores for the YSR Externalizing and Internalizing behavior scales. However, five CBCL items were not included in the YSR because of content that might arouse resistance or defensiveness for children who were asked these questions (e.g., “Plays openly with his or her sexual organs”). Given the unequal number of items in the YSR and the CBCL, in order to facilitate comparisons of the responses of the mothers and their children to these measures, the raw scores for each scale were converted into standard or z scores, with means of 0 and a standard deviation of 1.00. The scores for the total samples were used to compute the z scores in order to permit examination of possible gender differences for each scale. Because higher raw scores on the CBCL and the YSR indicate a larger number of behavior problems, higher, more positive z scores are indicative of children with more behavior problems, whereas lower and negative z scores indicated that the children were rated as having fewer behavior problems.

RT44367_C010.indd 183

7/13/05 9:28:26 AM

184

S. DREMAN

Procedure Because the FACES III, CBCL, and YSR were not introduced until the second part of this longitudinal study, only the data for the 119 mothers and their children who participated in both parts of the study will be reported. The participants in the second part of the study were examined, on average, 1.5 years after the first part was completed. Testing was done by mail and in-home visits by undergraduate students who were trained in interviewing and test administration. The FACES III was administered as part of a larger mail questionnaire; the STAXI, CBCL, and YSR were administered individually by the research assistants at the family’s home after the mothers had completed the mail-in questionnaire. The mothers were also given an opportunity to participate in group meetings, conducted in different parts of the country, in which they could discuss their research experience and divorce-related problems. RESULTS Relation between Cohesion, Adaptability, and Behavior Problems The data were evaluated in a number of ANOVAs of responses to the Family Adaptability and Cohesion Evaluation scales. In examining the relations between family structure and children’s behavior problems, gender differences were taken into account in evaluating the response for each of the three behavior problems scales. No main effects or interactions were found for cohesion, adaptability, or gender in the analyses of the children’s behavior problems. However, visual inspections of the mothers’ CBCL scores and the children’s responses to the YSR revealed a consistent tendency for mothers to report differently from their children regarding the extent of the child’s behavior problems. In responding to the CBCL, mothers, as predicted, reported the fewest behavior problems at the highest (enmeshed) level of cohesion and the most problems at the lowest (disengaged) level. They also reported, as expected, the fewest behavior problems at the highest level of adaptability (chaos) and the most problems at the lowest (rigid) adaptability level. In contrast, children tended to report more behavior problems at the highest levels of both cohesion and adaptability, and fewer problems at the lowest cohesion and adaptability levels. Given these observations of interactions between cohesion and adaptability with respondent sources (parent vs. child), the data were collapsed over gender, and 4 × 2 ANOVAs (cohesion or adaptability levels by respondent source) were computed for each of the three behavior problems scales based on the responses of the mothers and their children to the CBCL and the YSR. Cohesion and Children’s Behavior Problems The means and standard deviations for the mothers’ and children’s z scores on Total Behavior Problems, and for the Externalizing and Internalizing scales, of responses to the CBCL and YSR are reported in Table 10.1 for each of

RT44367_C010.indd 184

7/13/05 9:28:26 AM

FAMILY COHESIVENESS, FLEXIBILITY, AND MATERNAL ANGER

185

Table 10.1 Behavior Problems Means and Standard Deviations (z scores) Reported by Divorced Mothers and Their Children as a Function of Family Cohesion and Respondent Source Panel A: Total Behavior Problem Scores

CBCL YSR

Mean SD Mean SD N

Disengaged

Separated

Connected

Enmeshed

.12 .78 –.26 1.13 13

–.07 1.19 .12 .87 34

.03 .74 –.19 1.17 21

–.43 .59 .56 .80 15

Panel B: Externalizing Behavior Problem Scores Disengaged CBCL YSR

Mean SD Mean SD N

.47 .54 .22 .91 10

Separated

Connected

Enmeshed

–.04 1.08 .03 .99 30

.02 .84 –.25 1.05 17

–.52 .61 .32 .88 12

Panel C: Internalizing Behavior Problem Scores

CBCL YSR

Mean SD Mean SD N

Disengaged

Separated

Connected

Enmeshed

.40 .85 .29 1.06 10

–.02 .89 .10 .88 29

.01 .79 –.17 1.03 18

–.29 .75 .71 .99 11

the four cohesion levels. A significant cohesion by respondent interaction was found for the responses of mothers and children to the Total Behavior Problems scale, F(3,79) = 4.00, p < .01. As predicted, mothers with the highest scores on cohesion, that is, enmeshment, reported the fewest behavior problems, and those with the lowest scores, that is, disengagement, reported the most behavior problems (planned contrast of enmeshed vs. disengaged, t(111) = 1.876, p < .06). In contrast, children of mothers perceiving enmeshment reported the most total behavior problems in their YSR responses, whereas those whose mothers perceived disengagement reported the fewest problems (planned contrast, enmeshed vs. disengaged, t(81) = –2.187, p < .03). Thus, the reports of mothers and their children were diametrically opposed at the extreme levels of cohesion. A significant interaction was also found between cohesion and motherchild responses for Internalizing Behavior Problems, F(3,64) = 2.65, p < .05. Mothers with high cohesion scores, that is, enmeshment, reported the fewest

RT44367_C010.indd 185

7/13/05 9:28:27 AM

186

S. DREMAN

behavior problems, whereas those with low cohesion scores, that is, disengaged, reported the most problems, t(93) = 2.067, p < .04. In contrast, in their responses to the YSR, mother’s enmeshment was related to the most behavior problems reported by children, and disengagement was related to fewer behavior problems. Although these differences were not statistically significant, t(77) = –.86, p < .39, a curvilinear relation was evidenced by extreme levels of cohesion being more related to behavior problems than intermediate levels (planned contrast, enmeshed and disengaged vs. separated and connected, t(77) = –2.23, p < .03). Children in connected families reported fewer behavior problems than those in enmeshed families, as indicated by significantly lower scores on the Total and Internal Behavior Problems scales, respectively, t(81) = –.2057, and p < .03; t(77) = –2.516, p < .02. Children of disengaged mothers reported relatively high rates of behavior problems on both the Externalizing and Internalizing scales, while also reporting a smaller number of problems on the Total Behavior Problem scale. This appears to be a contradiction because scores on the Total Behavior Problem scale were based on the sum of the scores on the Externalizing and Internalizing scales. It should be noted, however, that the Total Behavior Problem scores included three additional subscales, which might account for the discrepancy between Total Problem scores and scores on the Externalizing and Internalizing scales. Adaptability and Children’s Behavior Problems A significant interaction between adaptability and respondent source was found for scores on the Total Behavior Problems measure, F(3,79) = 3.43, p < .02, which was based on the CBCL and YSR scores of mothers and their children. As predicted, mothers with the highest CBCL adaptability scores, that is, chaotic (mean = –.43), reported the fewest behavior problems, and those with the lowest scores, that is, rigid (mean = .30), reported the most behavior problems, t(81) = –2.102, p < .04, as can be noted in Table 10.2. In contrast, in their responses to the YSR, children whose mothers perceived chaotic family patterns reported the most total behavior problems (mean = –.65), whereas those whose mothers perceived a rigid family pattern reported having the fewest problems (mean = .20); these differences were significant, t(81) = –2.10, p < .04. Thus, the reports of mothers at the extreme levels of adaptability were very different from those of their children. Consistent with Olson’s revised model of family structure, divorced mothers who indicated that their families were highly cohesive and adaptable reported fewer behavior problems for their children than mothers who rated their families as low on these dimensions. In contrast, children reported more behavior problems when family cohesion and adaptability levels were high and reported fewer problems at the lowest levels of these family structure scales. These trends were particularly strong on the cohesion dimension for both Total and Internalizing Behavior Problems and for Total Behavior Problems on the Adaptability dimension. It is also interesting to note that the number of significant interactions for the children was greater for the cohesion scale than for the adaptability scale.

RT44367_C010.indd 186

7/13/05 9:28:27 AM

187

FAMILY COHESIVENESS, FLEXIBILITY, AND MATERNAL ANGER

Table 10.2 Behavior Problems Means and Standard Deviations (z scores) Reported by Divorced Mothers and Their Children as a Function of Family Adaptability and Respondent Source Panel A: Total Behavior Problem Scores

CBCL YSR

Mean SD Mean SD N

Rigid

Structured

Flexible

Chaotic

.30 .94 –.20 .94 11

.02 1.20 –.10 1.09 27

–.18 .73 .12 .93 36

–.43 .75 .65 1.02 9

Panel B: Externalizing Behavior Problem Scores

CBCL YSR

Mean SD Mean SD N

Rigid

Structured

Flexible

Chaotic

.28 .64 –.25 1.03 9

.23 1.23 .23 1.21 21

–.17 .75 –.06 .78 31

–.56 .49 .21 .97 8

Panel C: Internalizing Behavior Problem Scores

CBCL YSR

Mean SD Mean SD N

Rigid

Structured

.52 1.27 .15 1.15 10

–.23 .67 .02 1.00 19

Flexible

Chaotic

.05 .71 .10 .88 29

–.17 .88 .59 1.13 10

Custodial Mothers’ Anger and Children’s Behavior Problems The relationships for boys and girls between maternal anger and mothers and children’s ratings of Externalizing and Internalizing Behavior Problems were evaluated in 3 × 2 ANCOVAs (mother’s anger level × child’s gender) for each of the three behavior problems scales. Mothers with anger scores one-half standard deviation or more above or below the mean were classified as “high anger” or “low anger”; those within one-half SD of the mean were considered as “medium anger.” For scores on the Total, Externalizing, and Internalizing Behavior Problem scales, two ANCOVAs were performed for each anger measure, one for the mothers’ responses to the CBCL, and one for the children’s responses to the YSR. The mothers’ age, education, and income, which have been found to be related to their children’s postdivorce adjustment, were used as covariates in these analyses.

RT44367_C010.indd 187

7/13/05 9:28:27 AM

188

S. DREMAN

Anger Expression and Children’s Behavior Problems. The ANCOVA analyses revealed no significant main or interaction effects for gender and maternal anger expression (Anger-Out, Anger-In) for the three behavior problems scales, contradicting our predictions. However, as expected, there was a trend for higher levels of maternal Anger-Out to be associated with higher ratings for Total and Externalizing Behavior Problems for both boys and girls, regardless of whether the child’s behavior problems were rated by the mothers or their children. Anger Experience (S-Anger) and Children’s Behavior Problems. The results of the ANCOVA analyses of the relations between the S-Anger level of the custodial mothers and the behavior problems of boys and girls, as measured by their responses to the CBCL and the YSR, are presented in Table 10.3. In these analyses, the hypothesis that high maternal anger would be related to higher ratings of children’s behavior problems was examined. For the mothers’ CBCL scores, the interaction between maternal S-Anger level and gender for the Total Behavioral Problems measure was significant, F(2,109) = 5.378, p < .006. As predicted, maternal S-Anger was positively associated with mothers’ ratings of their daughters’ behavior problems. A planned contrast showed that the difference in the number of children’s behavior problems reported by mothers with high and low levels of maternal anger was highly significant, t(57) = –3.975, p <. 0001. However, in contrast to what was predicted, mothers with medium S-Anger scores reported their sons as having the most behavioral problems. A planned contrast also showed that the sons of mothers with medium anger reported significantly more behavior problems than the sons of the combined group of mothers with low and high S-Anger scores, t(52) = 2.197, p < .03. The finding that mothers high in state anger reported that their boys had the fewest behavior problems was clearly contrary to our prediction. The interaction of maternal anger by gender for responses to the Youth Self-Report measure approached significance, F(2,79) = 2.437, p < .09. As predicted, maternal anger was positively associated with daughters’ reports of Total Behavioral Problems, but this difference in the behavioral problems of daughters did not reach statistical significance. Unexpectedly, boys whose mothers had low S-Anger reported the most behavior problems, whereas boys of mothers with medium maternal anger reported the fewest behavior problems. A planned contrast also showed that the sons of mothers with low S-Anger reported significantly more behavior problems than the sons of mothers with medium maternal anger, t(37) = 1.995, p < .05. In the analyses of Externalizing Behavior Problems, a significant state anger by gender interaction was found for mothers’ responses to the CBCL, F(2,100) = 3.723, p < .03. As predicted, maternal S-Anger was positively associated with behavior problems in daughters; a planned contrast showed that mothers with high maternal anger reported that their daughters had significantly more behavior problems, t(54) = –2.209, p < .03. However, mothers with medium state anger reported that their sons had the most behavior problems, while those with high or low maternal anger levels reported that their sons had fewer problems (planned contrast for mothers with medium vs. combined high and low state anger, t(46) = –3.208, p < .002).

RT44367_C010.indd 188

7/13/05 9:28:28 AM

FAMILY COHESIVENESS, FLEXIBILITY, AND MATERNAL ANGER

189

Table 10.3 The Relation between the State Anger of Custodial Mothers and the Mean z Scores of Boys and Girls (N = 119) for Total, Externalizing, and Internalizing Behavior Problems as Measured by the Child Behavior Check List and the Youth Self Reports Child Behavior Check List Low

Med

High

Row Mean

Total Behavior Problems Boys Girls CM*

–.217 –.339 –.280

.313 .045 .183

–.277 –.302 –.290

.723 –.028 .325

–.323 .965 .438

–.061 .052

–.287 –.136 –.214

.068 .102 .089

Med

High

Row Mean

–.268 .392 .122

–.255 .937 .474

.237 –.220 .003

–.402 .104 –.110

–.139 .363 .112

–.023 .004

Externalizing Behavior Problems .051 .057

Internalizing Behavior Problems Boys Girls CM

Low

Total Behavior Problems

Externalizing Behavior Problems Boys Girls CM

Youth Self Reports

.313 –.250 .039

–.520 .178 –.095

–.394 .400 1.05

–.062 .039

Internalizing Behavior Problems –.193 .168

–.233 .031 –.101

–.344 .351 .030

.182 .467 .162

–.257 .231

* CM = Column mean. Standard deviations ranged from .581 to 1.691, most between .750 and 1.100, thus not violating the assumptions of the analysis.

In children’s responses to the Youth Self-Report, a significant state anger by gender interaction was also found for Externalizing problems, F(2,72) = 4.280, p < .02. As predicted, daughters of mothers with high maternal anger reported significantly more behavior problems than those whose mothers were low in maternal anger, t(58) = 2.555, p < .01. However, contrary to our prediction, boys whose mothers had medium levels of maternal anger reported the fewest behavior problems, whereas those whose mothers had low maternal anger reported the most behavior problems. A planned contrast also showed that the sons of mothers with low levels of S-Anger reported significantly more behavior problems than those whose mothers had medium levels of maternal anger, t(34) = 2.596, p < .04. Consistent with the hypotheses of this study, the state anger of divorced mothers was associated with more Total, Externalizing, and Internalizing Behavioral Problems for girls, as indicated by the responses of both mothers and their children. However, in contrast to our predictions, mothers with medium levels of maternal anger reported the most behavior problems for their sons. Also unexpected were the findings that the sons of mothers with

RT44367_C010.indd 189

7/13/05 9:28:28 AM

190

S. DREMAN

medium maternal anger reported the fewest behavior problems and those whose mothers had the lowest level of maternal anger reported the most Total and Externalizing behavior problems. Similar trends to those found for Externalizing Behavior Problems were also evidenced for Internalizing Behavior Problems, but the results of the analyses did not reach statistical significance. The only exception was in the YSR responses; boys of mothers with the highest level of maternal anger reported the most Internalizing Behavior Problems, and those whose mothers were at the lowest level of maternal anger reported the fewest problems. Internalizing Behavior Problems was also the only measure for which a main effect of gender was found. In responding to the YSR, girls reported significantly more Internalizing Behavior Problems than boys, F(1,75) = 4.753, p < .03. DISCUSSION AND CONCLUSIONS The findings of the studies reported in this chapter (Dreman, 1995; Dreman & Ronen-Eliav, 1997) confirmed, in part, the predictions regarding the effects of maternal anger and family structure on the behavior problems of children. It was predicted that high levels of maternal anger, as compared with medium maternal anger, would result in children having more behavioral problems, and that high levels of family cohesion and adaptability would be related to fewer behavior problems in children. Although mothers did report that higher levels of cohesion and adaptability were related to fewer behavior problems, children reported the most behavior problems at the highest levels of cohesion and adaptability, and fewer problems at the lowest levels (Dreman & Ronen-Eliav, 1997). High maternal anger was positively related to more behavior problems for daughters, as predicted, but was related to fewer behavior problems for their sons. Family Structure and Adjustment. To understand these findings, it is useful to consider how different observers perceive similar environments. For example, Bernard (1972) reported that husbands and wives render very different reports of their marital situation. Similar biases may occur when parents and children rate family structure in relation to children’s behavior problems. These biases could reflect different locations in the family structure hierarchy (e.g., superordinate vs. subordinate), different frames of reference for adolescents and their middle-aged parents, and culturally linked biases. For example, adolescents may experience clear parent-child role boundaries and stability, when expressed in lower levels of family cohesion and adaptability, as security in the face of profound psychological, physiological, and social changes that they are undergoing at this developmental stage. This may be especially true for adolescent children of divorced parents who are faced with the unsettling effects of the divorce process. On the other hand, a divorced parent who has undergone considerable emotional turmoil in the divorce process might perceive high levels of family cohesion and adaptability as promoting better adjustment in their children. These observations may help to explain why parents report fewer behavior problems in highly cohesive and adaptable families, while their adolescent children report expe-

RT44367_C010.indd 190

7/13/05 9:28:29 AM

FAMILY COHESIVENESS, FLEXIBILITY, AND MATERNAL ANGER

191

riencing more problems in the identical structural contexts, as did the sons of divorced mothers in this study. It should be noted that children in connected families generally report fewer behavior problems than those in enmeshed families. Dreman and Ronen-Eliav (1997) hypothesized that the better adjustment of children in connected families reflects not only parental warmth, but also clear family role hierarchies and controls, which lead to better adjustment in their adolescent children, irrespective of gender. Supporting this is a recent meta-analysis of 47 studies, which found that a factor, called parental “acceptance-responsibility,” explained the largest part of the variance between parental behavior and diminished behavior problems in their children (Rothbaum & Weisz, 1994). This factor incorporated positive parental control in addition to emotional closeness. Anger and Adjustment. Maternal anger was related to more behavior problems for daughters, as predicted, but was related to fewer behavior problems in sons. To understand these findings, it is instructive to consider research that shows that boys in maternal custody display poorer adjustment (Camara & Resnick, 1989). Single-parent custodial mothers appear to have difficulty in establishing parental controls and discipline without a father figure present, especially in the case of their adolescent sons (Hetherington, 1981; Hetherington et al., 1989). Hence, boys in maternal custody may be deprived of the disciplinary measures and control usually provided by their fathers, which may result in more behavior problems. The findings in the author’s (1995) study provide evidence of the importance of maternal discipline and control. A factor analysis in this study identified an action potential factor, represented by STAXI S-Anger items, such as “I feel like banging on the table” and “I feel like hitting someone,” which contrasted with an angry feelings factor, represented by items such as “I feel irritated” and “I feel angry.” The children of mothers with high scores on the action potential factor had fewer behavior problems as measured by both the CBCL and the YSR. These findings were interpreted as indicating that action potential represented internal energy, which was expressed by custodial mothers in increased assertiveness and control, rather than overtly expressed anger, which might explain the lower incidence of behavior problems in their boys in the absence of a father figure in the home. Behavior problems of both boys and girls tended to increase as a function of maternal Anger-Out scores. Although these trends were not statistically significant, the fact that the behavior problems of children increased somewhat at higher Anger-Out levels may be attributable to the fact that the items included in the STAXI AX/Out scale (e.g., “When I am angry, I slam doors”) reflect explicit behaviors, demonstrating impulsiveness and loss of control. Such behaviors serve as visible negative role models that can be imitated by children and thereby evoke behavior problems irrespective of gender. In contrast, state anger items represent only angry feelings that may not be expressed. The differential gender effect of maternal S-Anger may be explained by the ability of daughters to empathize with their mothers’ experience of internal distress, while boys are more influenced by their mothers’ external assertiveness, which, unlike Anger-Out, is not an aggressive act.

RT44367_C010.indd 191

7/13/05 9:28:29 AM

192

S. DREMAN

Family Structure, Anger, and Gender Differences. No interaction was found between gender and either cohesion or adaptability; both boys and girls reported that they had more behavior problems when their families were perceived as displaying higher cohesion or adaptability. This was in contrast to their mothers who, as predicted, reported the lowest incidence of behavior problems at the highest levels of cohesion and adaptability. The finding that gender interacted with maternal state anger, resulting in better adjustment in boys than girls, may be attributable to the fact that the action potential factor was related to assertiveness in the divorced custodial mothers. Given the absence of a father figure in the household, such maternal assertiveness may enhance the mothers’ effectiveness in disciplining their adolescent sons. In contrast to previous findings that boys displayed poorer adjustment in maternal custody situations (e.g., Zaslow, 1988), few main effects of gender were found in the two studies that were reported in this chapter. A possible explanation is that gender differences among children in divorce tend to become less pronounced as adolescence approaches (Hess & Camara, 1979). A possible reason is that girls display more behavior problems in the heterosexual realm as they approach puberty. Theory, Research, and Intervention. Contemporary theories of stress contend that cognitive appraisals of crisis events and a person’s coping resources mediate between environmental stressors and adjustment (e.g., Lazarus & Folkman, 1984; McCubbin & Patterson, 1983; Olson, 1993, 1997). Common to these theories is their failure to take into account the mediating role that emotions, such as anger or anxiety, may play in explaining adjustment. Recent research by Cummings and associates provides convincing evidence that parental emotions are important in explaining children’s adjustment to stressful events and have a causal role in organizing and directing children’s reactions to these stressors (Cummings, 1995; Cummings & Davies, 1994, 1995; El-Sheikh & Cummings, 1995). These findings also highlight the role of emotionality of children in coping with marital conflict (e.g., Fincham, 1994; Grych & Fincham, 1993; Osborne & Fincham, 1996). High levels of state anger may be evoked in divorced mothers by the ongoing stressors that they face in the postdivorce situation, such as work overload, single parenting, and diminished socioeconomic standing, which could affect both appraisal and coping efforts. To illustrate this point, an angry mother may make negative cognitive appraisals of her children (see Dix et al., 1990) that could affect her coping behaviors and ultimately affect her children’s adjustment. Such observations emphasize the importance of emotions as intervening variables in understanding the effects of parental stress management in children’s adjustment. Although the results of the studies reported in this chapter indicate that the experience of anger may be adaptive in certain family contexts, such conclusions should also be treated with caution, because anger might affect adjustment through intervening coping variables such as assertiveness. The theoretical framework for this research should also be considered with caution, given the correlational nature of the divorce studies that were reported (Dreman, 1995). Conclusions regarding causality should be held in abeyance until longitudinal

RT44367_C010.indd 192

7/13/05 9:28:29 AM

FAMILY COHESIVENESS, FLEXIBILITY, AND MATERNAL ANGER

193

studies that systematically investigate ongoing relationships between stress, emotions, cognitive assessment, coping, and adjustment are conducted. The author and associates plan to develop a “maternal assertiveness” scale to investigate whether state anger is positively correlated with assertiveness in mothers. If higher correlations are obtained between assertiveness and state anger than with cohesion or adaptability, this might explain the differential gender effects that were found with maternal anxiety. Such research would have important implications in providing guidelines for parents on how to more effectively modulate angry behavior. Conclusions drawn from the studies cited in this chapter may also be limited by the sample considerations, given that the participants were single-parent mothers of the Jewish faith who were more than eight years past the divorce event. In order to determine the more general implications of our findings, future studies must systematically investigate other populations such as two-parent families, different ethnic/religious groups, and divorced families at different periods of time since the divorce. Preadolescent children should also be investigated to determine if family structure and anger affect their behavior problems in the same way as was found for the adolescents in the present study. In regard to divorce, several lines of research support the notion that marital discord and its emotional residues, rather than divorce per se, result in poor adjustment in children (Emery, 1982; Rutter, 1979). Moreover, children are not only affected by anger and aggression directed at them, but are also responsive to the background climate or environment of anger, including parental conflict (Cummings, Zahn-Waxler, & Radke-Yarrow, 1984). This suggests that anger may affect children’s adjustment in a similar fashion regardless of differing family structures such as in single- or two-parent families. Process considerations, such as different time periods in the divorce process or different stages of the family life cycle, should also be taken into consideration when considering the effects of anger or family structure on children’s adjustment. In a study by Abelsohn and Saayman (1991), which investigated the relation between cohesion/adaptability and behavior problems, divorced mothers in highly adaptable (chaotic) families rated their children as having significantly more behavior problems 10 months after the divorce, whereas family cohesion was found to have no significant effects. However, Dreman and Ronen-Eliav (1997) found that family cohesion accounted for more of the total variance for reported behavior problems than did adaptability. In the initial stages of the divorce process, the situation is very stressful and family roles are in a state of flux. More structure and greater stability, represented by lower adaptability scores, may be needed to help the family cope. For longterm adjustment, the findings of our study suggested that family cohesion plays a more important role than adaptability in children’s behavior problems, as perceived by both parents and children. Future cross-cultural studies are also necessary because anger may have different meanings in different cultural contexts. For example, behaviors viewed as aggressive and hostile in some cultures may be considered as assertive and positive behavior in others. Regarding family structure, Shamgar-Handleman and Bar Yosef (1991) point out that Israeli society is characterized by “familism”

RT44367_C010.indd 193

7/13/05 9:28:30 AM

194

S. DREMAN

and that more value is attached to family-related subjects such as cohesiveness and child-centered concerns. The relative influence of family cohesion and adaptability in less familistic societies should be investigated in order to determine if these structural dimensions affect children’s adjustment in a different fashion than indicated in the findings of the present study. A unique characteristic of the present study was that mothers rendered selfreports of anger, cohesion, and adaptability, whereas both mothers and their children independently reported on the children’s behavior problems. This research paradigm overcomes the respondent bias involved in studies using only parental respondents for reporting on family structure or parenting in relation to their children’s adjustment. There was evidence in this study that mothers with more negative emotions rated their children’s behavior more negatively (Dix et al., 1990; Emery, 1982), indicating that the parental ratings of their children’s adjustment were influenced by the emotional state of the parent. Such respondent bias may have been reflected in our finding that divorced mothers rated their sons’ behavior problems as more severe when they had medium versus low levels of maternal state anger, while their sons did the opposite. The angrier mothers may have projected their own emotional distress in their ratings of their children’s behavior problems. Similarly, mothers who rated their families as low in cohesion or adaptability may have projected their accompanying negative emotional states in their assessments of their children’s adjustment. This could explain the higher ratings of behavior problems by mothers in disengaged and rigid families compared with those in enmeshed or chaotic families. These findings clearly indicate that children’s self-reports may provide a different perspective of their adjustment, which should be taken into consideration in family research.

REFERENCES Abelsohn, D., & Saayman, G. (1991). Adolescent adjustment to parental divorce: An investigation from the perspective of basic dimensions of structural family therapy. Family Process, 30, pp. 177–191. Achenbach, T. M., & Edelbrock, C. (1983). Manual for child behavior checklist and revised child behavior profile. Burlington, VT: University of Vermont. Achenbach, T. M., & Edelbrock, C. (1987). Manual for youth self-report and profile. Burlington, VT: University of Vermont. Bandura, A. (1962). Social learning through imitation. In M. R. Jones (Ed.), Nebraska symposium on motivation (Vol. 10, pp. 211–274). Lincoln: University of Nebraska Press. Bernard, J. (1972). The future of marriage. New York: Bantam. Camara, K. A., & Resnick, G. (1989). Styles of conflict resolution and cooperation between divorced parents: Effects on child behavior and adjustment. American Journal of Orthopsychiatry, 59, pp. 560–575. Cummings, E. M. (1995). Security, emotionality, and parental depression: A commentary. Developmental Psychology, 31, pp. 425–427.

RT44367_C010.indd 194

7/13/05 9:28:30 AM

FAMILY COHESIVENESS, FLEXIBILITY, AND MATERNAL ANGER

195

Cummings, E. M., & Davies, P. (1994). Children and marital conflict. New York: Guilford Press. Cummings, E. M., & Davies, P. (1995). Children’s emotions as organizers of their reactions to interadult anger: A functionalist perspective. Developmental Psychology, 31, pp. 677–684. Cummings, E. M., Zahn-Waxler, C., & Radke-Yarrow, M. (1984). Developmental changes in children’s reactions to anger in the home. Journal of Child Psychology and Psychiatry, 25, pp. 63–74. Dix, T., Reinhold, D. P., & Zambarano, R. J. (1990). Mothers’ judgments in moments of anger. Merrill-Palmer Quarterly, 36, pp. 465–486. Dreman, S. (1995). The experience and expression of anger in divorced mothers: Effects on postdivorce adjustment in children. In C. D. Spielberger & I. Sarason (Eds.), Stress and emotion: Anger, anxiety and curiosity (Vol. 15, pp. 75–89). New York: Hemisphere/Taylor & Francis. Dreman, S. (1996). The experience and expression of anger in a stressful society: Standardization of the STAXI in Israel [in Hebrew]. Unpublished manuscript, Ben Gurion University, Beer Sheva, Israel. Dreman, S., & Aldor, R. (1994). Work or marriage? Competence in custodial mothers in the stabilization phase of the divorce process. Journal of Divorce & Remarriage, 22, pp. 3–22. Dreman, S., & Ronen-Eliav, H. (1997). The relation of divorced mothers’ perceptions of family cohesion and adaptability to behavior problems in children: A cross-respondent analysis. Journal of Marriage and the Family, 59, pp. 324–331. El-Sheikh, M., & Cummings, E. M. (1995). Children’s responses to angry adult behavior as a function of experimentally manipulated exposure to resolved and unresolved conflict. Social Development, 4, pp. 75–91. Emery, R. E. (1982). Interparental conflict and the children of discord and divorce. Psychological Bulletin, 92, pp. 310–330. Fincham, F. D. (1994). Understanding the association between marital conflict and child adjustment: Overview. Journal of Family Psychology, 8, pp. 123–127. Grych, J. H., & Fincham, F. D. (1993). Children’s appraisal of marital conflict. Child Development, 64, pp. 215–230. Herz Brown, F. (1988). The postdivorce family. In B. Carter & M. McGoldrick (Eds.), The changing family life cycle: A framework for family therapy (2nd ed., pp. 371–398). New York: Gardner Press. Hess, R. D., & Camara, K. A. (1979). Postdivorce family relationships as mediating factors in the consequences of divorce for children. Journal of Social Issues, 35, 79–96. Hetherington, E. M. (1981). Children and divorce. In R. W. Henderson (Ed.), Parent-child interaction: Theory, research, and prospects (pp. 33–58). New York: Academic Press. Hetherington, E. M., Cox, M., & Cox, R. (1985). Long-term effects of divorce and remarriage on the adjustment of children. Journal of the American Academy of Child Psychiatry, 24, pp. 518–530. Hetherington, E. M., Stanley-Hagan, M., & Anderson, E. R. (1989). Marital transitions: A child’s perspective. The American Psychologist, 44, pp. 303–312.

RT44367_C010.indd 195

7/13/05 9:28:30 AM

196

S. DREMAN

Hornick, Y., & Maier, N. (1989). Nee-tuach shal niv-dakim sha’lo anu b-seker doar [An analysis of nonrespondents in a mail survey]. Megamot, 32, pp. 386–400. Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal and coping. New York: Springer. McCubbin, H. I., & Patterson, J. M. (1983). The family stress process: The double ABCX model of adjustment and adaptation. In H. I. McCubbin, M. B. Sussman, & J. M. Patterson (Eds.), Social stress and the family: Advances and developments in family stress theory and research (pp. 7–37). New York: Haworth Press. Olson, D. H. (1993). Circumplex model of marital and family systems: Assessing family functioning. In F. Walsh (Ed.), Normal family processes (2nd ed., pp. 104–137). New York: Guilford Press. Olson, D. H. (1997). Family stress and coping: A multisystem perspective. In S. Dreman (Ed.), The family on the threshold of the 21st century: Trends and implications. Mahwah, NJ: Erlbaum. Olson, D. H., Portner, J., & Lavee, Y. (1985). Family adaptability and cohesion evaluation scales (FACES III). St. Paul: University of Minnesota, Family Social Science. Osborne, L. N., & Fincham, F. D. (1996). Marital conflict, parent-child relationships, and child adjustment: Does gender matter? Merrill-Palmer Quarterly, 42, pp. 48–75. Rothbaum, F., & Weisz, J. R. (1994). Parental caregiving and child externalizing behavior in nonclinical samples: A meta-analysis. Psychological Bulletin, 116, pp. 55–74. Rutter, M. (1979). Maternal deprivation, 1972–1978: New findings, new concepts, new approaches. Child Development, 50, pp. 283–305. Shamgar-Handleman, L., & Bar Yosef, R. (1991). Mishpachot b’yeesrael [Families in Israel]. Jerusalem: Academon-The Hebrew University of Jerusalem. Spielberger, C. D. (1988). Manual for the State-Trait Anger Expression Inventory (STAXI). Odessa, FL: Psychological Assessment Resources. Teichman, Y., & Navon, S. (1990). Ha-arachah mishpachtit: Ha-model hacircumplexi [The circumplex model: Translation and standardization of FACES III to Hebrew]. Psychologia, 2, pp. 36–46. Tsemach, S. (1996). The relationship between state anger of divorced women and different areas of functioning after divorce. Unpublished thesis, Ben Gurion University, Beer Sheva, Israel. Zaslow, M. J. (1988). Sex differences in children’s response to divorce: 1. Research methodology and postdivorce family forms. American Journal of Orthopsychiatry, 58, pp. 355–378. Zilber, N., Auerbach, J., & Lerner, Y. (1994). Israeli norms for the Achenbach Child Behavior Checklist: Comparison of clinically referred and nonreferred children. Israel Journal of Psychiatry and Related Sciences, 31, pp. 5–12.

RT44367_C010.indd 196

7/13/05 9:28:31 AM

11 Hindi Anger Expression Scale: Construction and Validation Sagar Sharma and Archna Krishna Himachal Pradesh University, Shimla, India Charles D. Spielberger University of South Florida, Tampa, FL Abstract Construction and validation of the Hindi Anger Expression (AX) Scale followed the same general approach that had proved successful in developing the Hindi adaptations of the third author’s State-Trait Anxiety Inventory and Test Anxiety Inventory. To establish the equivalence of the Hindi AX Scale with the original English scale, both forms were administered to 160 bilingual university students (80 males, 80 females). Comparable means, standard deviations, and internal consistency reliability coefficients, and very high correlations for both males and females between scores on the AX Scale and all three subscales, documented the equivalence of the Hindi and English forms. Evidence of the construct validity and predictive power of the Hindi AX Scale was reported based on the results of eight studies conducted in India with samples of graduate students, electrical engineers, nurses, and patients suffering from essential hypertension, bronchial asthma, and chronic peptic ulcer.

THE HINDI ANGER EXPRESSION SCALE: CONSTRUCTION AND VALIDATION The 24-item Anger Expression (AX) Scale provides relatively brief, reliable, and valid self-report measures for assessing the expression and control of anger when a person feels angry or furious (Spielberger, 1988; Spielberger, Johnson, Russell, Crane, Jacobs, & Worden, 1985). Three major components of anger expression—suppressed anger (AX/In), anger directed toward other persons or objects in the environment (AX/Out), and anger control (AX/Con)—are assessed by the eight-item AX subscales. The AX/In and AX/Out subscales were initially developed by Spielberger et al. (1985) based on 1,114 U.S. high-school students (634 males, 480 females) and subsequently standardized on large samples of university students and working adults (Spielberger, 1988). Extensive psychometric analyses of responses to the AX Scale items, including separate factor analyses for the males and 197

RT44367_C011.indd 197

7/13/05 9:34:49 AM

198

S. SHARMA ET AL.

females, indicated that the AX/In and AX/Out subscales assess two orthogonal (independent) anger expression dimensions. The original 20-item AX Scale yielded a total Anger Expression (AX/EX) score based on the sum of the ratings for all 20 items. Scores on the eight-item AX/ In and AX/Out subscales were comprised of items with dominant salient loadings in factor analyses on either Anger-In or Anger-Out and essentially zero loadings on the second factor (Spielberger et al., 1985). Four items that were included in the AX/EX Scale were not scored in terms of the direction of anger expression. However, when subsequent research demonstrated the need to develop an AngerControl subscale (Krasner, 1986; Spielberger, 1988), three of the four previously unscored items were included in this new eight-item AX/Con subscale. Further studies of the factor structure of the 24-item AX Scale have confirmed that the Anger-Control items defined a third factor that was relatively independent of Anger-In and Anger-Out (Spielberger, Krasner, & Solomon, 1988). A major reason for constructing the AX Scale was to investigate how anger expression contributed to the etiology of hypertension, cardiovascular disease, and other psychosomatic disorders. Although the construct validity of the English form of the AX Scale has been verified in a number of studies (Spielberger, 1988), meaningful cross-cultural research requires demonstration of the reliability, construct validity, and equivalence of this scale in different languages and cultures, and with diverse national, ethnic, and socioeconomic groups. The three factors identified in factor studies of the 24 AX Scale items are operationally defined by the AX/In, AX/Out, and AX/Con subscales, each consisting of eight items (see Spielberger, 1988; Spielberger, Reheiser, & Sydeman, 1995). Strong evidence of the relative independence and internal consistency of these factors has been demonstrated cross-culturally in the construction and validation of adaptations of the AX Scale in Germany (Schwenkmezger, Hodapp, & Spielberger, 1992), Italy (Spielberger & Comunian, 1992), The Netherlands (Mook, van der Ploeg, & Kleijn, 1990), Norway (Håseth, 1996), and Singapore (Tanzer, Sim, & Spielberger, 1996). In applying tests developed in the United States to other cultures, especially in Eastern countries such as India (Sinha & Holtzman, 1984), the crosscultural equivalence and utility of the adapted scale must be empirically verified. The major goals of this study were to (1) describe the development of a Hindi adaptation of the Anger Expression Scale that can be utilized in all Hindi-speaking states of India, (2) present data that demonstrate the psychometric equivalence of the Hindi AX Scale with the original English form, and (3) report findings with the Hindi AX Scale based on a series of studies conducted in India that provide evidence of the usefulness of this measure in behavioral medicine and in clinical and health psychology research. ADAPTATION AND VALIDATION OF THE HINDI AX SCALE The construction of the Hindi AX Scale followed the same general approach that proved successful in developing the Hindi forms of the State-Trait Anxiety Inventory (Spielberger & Sharma, 1976; Spielberger, Sharma, & Singh, 1973)

RT44367_C011.indd 198

7/13/05 9:34:49 AM

HINDI ANGER EXPRESSION SCALE

199

and the Test Anxiety Inventory (Sharma, Sud, & Spielberger, 1983). The initial translation of the AX Scale into Hindi was carried out with the assistance of five instructors at the Himachal Pradesh University who were knowledgeable about the substantive fields in which the scale was to be used. In the initial phase of test development, the instructors were asked to construct simple Hindi translations of each of the 24 English AX Scale items and to provide at least three translations for each item. However, rather than developing literal item translations, they were instructed to give special attention to psychological connotation and grammatical syntax in order to retain the essential meaning of the item and to keep in mind, as the framework for item construction, the definitions of the Anger-In, Anger-Out, and Anger-Control concepts that guided the construction of the English form of the scale. In developing the Hindi translations for the 24 English AX Scale items, problems arose in adapting two Anger-In items that were expressed in English idioms: “I pout or sulk”; “I boil inside, but I don’t show it.” In constructing adaptations of the former item, the feeling connotation was emphasized by describing facial expressions in the Hindi language that reflected the experience of these feelings. Corresponding Hindi idioms were also identified for describing “boiling inside.” The resulting preliminary form of the Hindi AX Scale was comprised of a total of 72 items, each written in simple Hindi language. In the second phase of developing the Hindi AX Scale, interviews were conducted with six instructors who taught courses in English, Hindi, and Sanskrit languages at the Himachal Pradesh University and had not previously participated in the translation and adaptation of the English items. Each Hindi item was carefully examined, along with its English equivalent, to ascertain which of the Hindi translations was conceptually most like the original English item. On the basis of discussions with these language experts, a number of items were modified or replaced. The final phase of the adaptation process involved obtaining evaluations of each Hindi AX Scale item from a diverse group of 20 professionals, which included psychologists, psychiatrists, and language teachers at universities in Shimla, Chandigarh, and Lucknow, the capital cities of three North Indian states. After first clarifying the conceptual definitions of Anger-In, Anger-Out, and Anger-Control for these professionals, they were asked to review and rate the adequacy of the three alternatives for each English item, using the following three-point rating scale: G = Good; S = Satisfactory; U = Unsatisfactory. The raters were also encouraged to provide suggestions for making the Hindi items linguistically more acceptable to the average person. Based on these ratings and suggestions, the final set of 24 items was selected for the Hindi Anger Expression Scale. Psychometric Properties and Equivalence of the Hindi and English AX Scales Empirical demonstration of the cross-language equivalence of psychological scales, subscales, and individual test items is an essential step in the development of psychometric measures for cross-cultural research. For this

RT44367_C011.indd 199

7/13/05 9:34:50 AM

200

S. SHARMA ET AL.

Table 11.1 Means, Standard Deviations, and Alpha Coefficients for the Hindi and English Forms of the Anger Expression Scale for Bilingual Male and Female University Students Males (N = 80)

Females (N = 80)

Scale

Hindi

English

Hindi

English

AX/EX M SD Alpha

28.42 11.63 .96

28.32 10.88 .96

30.80 11.44 .89

30.30 11.36 .91

AX/In M SD Alpha

15.62 4.55 .92

15.50 4.57 .90

17.60 5.02 .88

17.29 4.73 .85

AX/Out M SD Alpha

15.67 4.63 .92

15.30 4.69 .93

16.46 3.29 .62

16.45 3.48 .70

AX/Con M SD Alpha

18.92 2.95 .82

18.77 3.01 .88

19.54 3.53 .82

19.36 3.52 .88

purpose, it is often convenient to utilize bilingual subjects who have reading proficiency in both languages (Spielberger & Sharma, 1976). To establish the equivalence of the Hindi AX Scale with the standard English scale, both forms were administered to 160 bilingual students (80 males, 80 females) from Lucknow University, ranging in age from 19 to 23 years. The students were randomly assigned to one of two presentation sequences, English-Hindi or Hindi-English, with 80 students (40 males, 40 females) in each sequence. The means and standard deviations for the Experimental Hindi and the English AX/EX, AX/In, AX/Out, and AX/Con scales are reported in Table 11.1, separately for the male and female bilingual students. As can be noted, the mean scores within gender for all four Hindi and English AX subscales were very similar. However, the scores of the females were consistently higher than those of the males for both language forms. Although the gender differences were not significant for the AX/Out, AX/Con, or AX/EX Scales, the females had significantly higher AX/In scores on both the Hindi (CR = 2.91, p < .01) and English (CR = 2.45, p < .05) forms, indicating greater suppression of their angry feelings. The manner in which anger is expressed depends on a person’s cultural history and the norms, customs, and cultural traditions of the society in which the individual is reared (cf. Mesquita & Frijda, 1992; Törestad, 1990). Social

RT44367_C011.indd 200

7/13/05 9:34:50 AM

201

HINDI ANGER EXPRESSION SCALE Table 11.2 Correlations between the Hindi and English Forms of the Anger Expression Scales for Male and Female University Students AX Scales AX/EX AX/In AX/Out AX/Con

Males

Females

.88 .88 .87 .88

.85 .74 .90 .88

customs tend to result in gender biases in regard to expressing anger and to reinforce gender-specific expectations. In India, and perhaps throughout Asia, female stereotypes continue to be associated with social norms such as gentleness, shyness, and the restraint and control of emotions. Although changes in gender roles are increasingly accepted in pursuing careers, Indian women are still expected to adhere to standards of dignity and composure much more than men. Sociocultural expectations that Indian females assimilate from their early childhood experiences undoubtedly contribute to the observed tendencies to suppress and direct anger inward. The equivalence of the Hindi and English AX Scales was evaluated for the bilingual Indian university students by computing product-moment correlations between scores on the AX/EX, AX/In, AX/Out, and AX/Con subscales for these measures, which are reported separately for males and females in Table 11.2. Except for the correlation of .74 between the AX/In subscales for females, all of these correlations were .85 or higher, providing strong evidence of the equivalence of the Hindi and English AX scales. The lower correlation for the AX/In subscales, though nevertheless quite adequate, may reflect the operation of a male bias in adapting the English AX/In items, which contributed to the Hindi AX/In items being more consistent with the values of Hindi males in regard to the inward expression of anger. Internal Consistency Reliability of the Hindi AX Scale Alpha coefficients for the Hindi AX Scales are also reported in Table 11.1. For scores on the 24-item Hindi AX/EX Scale, the alphas were .96 for males and .91 for females. The alpha coefficients for the eight-item Hindi and English AX/In and AX/Con subscales were also quite high for both sexes, ranging from .82 to .92. The alphas for the Hindi and English AX/Out scales were very high for the males (.92 and .93), but much lower for females (.62 and .70). Here again, the adaptation of the English AX/Out items may have been influenced by a male bias in defining the Anger-Out concept and in constructing the items for this subscale. In general, the internal consistency of the Hindi and English AX/EX, AX/ In, and AX/Con Scales was very satisfactory, indicating that the items comprising these scales are quite homogeneous for bilingual Indian university students. Indeed, the alphas for both the Hindi and English scales were somewhat higher than the values reported by the third author (1988) for large samples of U.S. students and similar to those observed by Tanzer et al. (1996) for

RT44367_C011.indd 201

7/13/05 9:34:51 AM

202

S. SHARMA ET AL.

Chinese young adults in Singapore. Given the somewhat lower alphas of the females for the Hindi and English AX/Out subscales, indicating marginal internal consistency, scores on this subscale need to be interpreted with caution for Indian females. Item-remainder (IR) correlations were computed for each item of the four AX scales. For the Hindi AX/EX Scale, the IRs ranged from .48 to .80 for males (median IR = .64) and from .28 to .81 for females (median IR = .62). Except for two items for females (“I control my behavior” and “If someone annoys me, I am apt to tell him or her how I feel”), the IR correlations of the 24 items with the AX/EX Scale scores were .40 or higher for both sexes, attesting to the overall homogeneity of the items comprising the Hindi AX Index Scale. The Hindi AX/In items were more homogeneous for the females than the males, as indicated by higher IR correlations, ranging from .58 to .87 (median IR = .80). The AX/In IRs for the males were .45 or higher (median IR = .60), except for “pout or sulk” (r = .27); in contrast, for females the IR for “pout or sulk” was .78. To “pout or sulk” is unacceptable for Indian males because such responses are considered to be characteristic of women. The item-remainder correlations for the Hindi AX/Out subscale were much higher for males than for females, ranging from .66 to .78 (median IR = .68); the IRs for females ranged from .24 to .60 (median IR = .42). Indeed, for all eight AX/Out items, the IRs for the females were markedly lower than those for the males. Thus, although the mean AX/Out scores for males and females did not differ, the males were more consistent than the females in responding to the AX/Out items. In India, the expression of anger is socially unacceptable for females, who are much less likely to engage in aggressive behavior than males when angry feelings are experienced. The IRs for the Hindi AX/Con subscale were .48 or higher for both males (median IR = .53) and females (median IR = .69), except for one item, “I control my behavior,” for which the IRs were .35 for males and .25 for females. Although this item may need revision or replacement, it is interesting to note that it deals with controlling the expression of anger in aggressive behavior and may be misinterpreted by Indian people, who tend to be more concerned with the internal control of anger than are persons from Western countries. The internal consistency of the English AX/EX, AX/In, AX/Out, and AX/Con scales was also evaluated by computing alpha coefficients and item-remainder correlations. In general, the psychometric properties of the English and Hindi AX scales were quite comparable. Thus, evidence of the equivalence of the Hindi and English forms of the Anger Expression Scale, as documented by comparable means, standard deviations, and internal consistency reliability, indicated that both forms of these scales can be used to assess the expression and control of anger in bilingual Indian university students. RESEARCH WITH THE HINDI ANGER EXPRESSION SCALE During the brief period of its availability, the Hindi AX Scale has been employed in India to investigate (1) gender differences as measured by the three

RT44367_C011.indd 202

7/13/05 9:34:51 AM

HINDI ANGER EXPRESSION SCALE

203

components of the AX Scale; (2) relationships of anger expression to Type-A behavior, organizational stress, and coping strategies; and (3) how the components of anger expression differ for patients with essential hypertension, bronchial asthma, and peptic ulcer. In regard to gender differences, greater suppression of angry feelings was observed in female university students than in their male counterparts, as was previously noted. Similar findings have also been reported for students from Aligarh Muslim University (Iqbal & Ahmad, 1993). Sharma and Acharya (1989) investigated the relation of anger expression to different strategies for dealing with organizational stress for a sample of 150 male electrical engineers. They found that engineers who internalized their anger, as indicated by higher AX/In scores, made greater use of avoidancecoping, whereas those with higher AX/Out scores were more likely to use approach-coping. The avoidance-coping group also reported more suppressed anger than the approach-coping group and less external expression and control of their angry feelings. Sharma, Sood, and Spielberger (1999) observed evidence of an association of anger expression with occupational stress and Type-A behavior in a sample of 160 registered nurses working at four Indian state-administered hospitals. Occupational stress correlated positively and significantly (p < .001) with AX/In (r = .33), AX/Out (r = .34), and AX/EX (r = .33) scores. Type-A behavior, as measured by the Jenkins Activity Survey (Jenkins, Zyzanski, & Roseman, 1979), also correlated positively and significantly (p < .01) with AX/In (r = .27), AX/ Out (r = .26), and AX/EX (r = .30) scores. In addition, the Type-A nurses had significantly higher (p < .001) AX/In, AX/Out, and AX/EX scores than their Type-B counterparts. It is interesting to note that nurses in India are highly stressed and that Type-A nurses, in particular, tend to cope with their angry feelings not only by suppression, but also by expressing their anger in aggressive behavior. The authors (1996) administered the Hindi AX Scale to middle-aged male and female patients with essential hypertension who were undergoing treatment in an outpatient clinic of a general hospital. The hypertensives had significantly higher AX/In scores and significantly lower AX/Out scores than a control group of patients with normal blood pressure (BP). In addition, the correlations of the AX/In scores of the hypertensives with BP were positive, whereas the correlations between AX/Out scores and BP were negative; these correlations were highly significant (p < .001) for both sexes. No significant relations were found between the AX/Con scores and BP for the hypertensives, nor between any of the anger measures and BP for the normotensive control group. The second author (1992) administered the Hindi AX Scale to a large sample of hypertensive and normotensive outpatients in the city of Kanpur. She found that scores on the AX/In subscale were the best predictors of systolic BP for the hypertensives, followed by AX/Out scores. These findings were consistent with the results of Sharma et al. (1996) and Ghosh and Sharma (1998), providing additional evidence of the contribution of suppressed anger to the etiology of essential hypertension. Anger expression was also investigated by Ghosh (1992) in a study of married, middle-aged bronchial asthma and peptic ulcer patients who were being

RT44367_C011.indd 203

7/13/05 9:34:51 AM

204

S. SHARMA ET AL.

treated in the outpatient clinic of a general hospital in Shimla. Compared with a control group of male surgical and orthopedic patients, the asthma and ulcer patients had significantly higher AX/In scores and lower AX/Out scores than the controls, indicating that they tend to cope with their angry feelings by suppressing them and by inhibiting aggressive behavior directed toward other people or objects in the environment. The asthma and ulcer patients also reported less control of the outward expression of anger, as reflected in lower AX/Con scores. Similar results were reported by Sharma, Ghosh, and Spielberger (1995) for chronic male gastric ulcer patients. In summary, the results of the studies of anger expression and control in patients with asthma and peptic ulcers demonstrated that anger contributed to nonvascular psychosomatic disorders as well as to hypertension and heart disease. When considered along with the psychometric data previously described, these findings also demonstrated the cross-language equivalence of the Hindi and English AX scales and the utility of the Hindi AX Scale in psychosomatic research. It is hoped that the availability of the Hindi AX Scale will serve to stimulate and facilitate research that investigates the effects of the expression and control of anger on other psychosomatic disorders in the Indian sociocultural context.

REFERENCES Ghosh, S. N. (1992). Life stress, anxiety, and anger in patients of bronchial asthma and peptic ulcer: An empirical investigation. Unpublished doctoral thesis, Himachal Pradesh University, Shimla, India. Ghosh, S. N., & Sharma, S. (1998). Trait anxiety and anger expression in patients with essential hypertension. Journal of the Indian Academy of Applied Psychology, 224, pp. 9–14. Håseth, K. J. (1996). The Norwegian adaptation of the State-Trait Anger Expression Inventory. In C. D. Spielberger & I. G. Sarason (Eds.), Stress and emotions: Anxiety, anger, and curiosity (Vol. 16, pp. 83–106). Washington, DC: Taylor & Francis. Iqbal, N., & Ahmad, H. (1993). Gender differences and birth order in expressed and suppressed anger. Journal of Personality and Clinical Studies, 9, pp. 73–77. Jenkins, C. D., Zyzanski, S. J., & Rosenman, R. H. (1979). Jenkins Activity Survey: Form B. New York: The Psychological Corporation. Krasner, S. S. (1986). Anger, anger control, and the coronary prone behavior pattern. Unpublished master’s thesis, University of South Florida, Tampa, FL. Krishna, A. (1992). The role of anger and anxiety in essential hypertension. Unpublished doctoral thesis, Himachal Pradesh University, Shimla, India. Mesquita, R., & Frijda, N. H. (1992). Cultural variations in emotions: A review. Psychological Bulletin, 112, pp. 179–204. Mook, J., van der Ploeg, H. N., & Kleijn, W. C. (1990). Anxiety, anger, and depression: Relationship at the trait level. Anxiety Research, 3, pp. 17–31.

RT44367_C011.indd 204

7/13/05 9:34:52 AM

HINDI ANGER EXPRESSION SCALE

205

Schwenkmezger, P., Hodapp, V., & Spielberger, C. D. (1992). Das State-Trait Arger-Ausdrucks-Inventar (STAXI): Handbuch [Test manual of the German version of the State-Trait Anger Expression Inventory]. Verlag Hans Hube Bern, Switzerland. Sharma, S., & Acharya, T. (1989). Coping strategies and anger expression. Journal of Personality and Clinical Studies, 5, pp. 15–18. Sharma, S., Ghosh, S. N., & Spielberger, C. D. (1995). Anxiety, anger expression, and chronic gastric ulcer. Psychological Studies, 40, pp. 187–191. Sharma, S., Krishna, A., & Spielberger, C. D. (1996). Anger and anxiety in hypertensive patients in India. In C. D. Spielberger & I. G. Sarason (Eds.), Stress and emotions: Anxiety, anger, and curiosity (Vol. 16, pp. 261–268). Washington, DC: Taylor & Francis. Sharma, S., Sood, N., & Spielberger, C. D. (1999). Occupational stress, anxiety, anger, and Type A Behaviour: A study of nurses from India. In D. M. Pestonjee, U. Pareek, & R. Agrawol (Eds.), Studies in stress and its management (pp. 111–120). New Delhi: Oxford & IBH Publishing Co. Sharma, S., Sud, A., & Spielberger, C. D. (1983). Development of the Hindi form of the Test Anxiety Inventory. In H. M. van der Ploeg, R. Schwarzer, & C. D. Spielberger (Eds.), Advances in test anxiety research (Vol. 2, pp. 183–190). Lisse, The Netherlands: Swets & Zeitlinger. Sinha, D., & Holtzman, W. H. (1984). The impact of psychology on Third World development. International Journal of Psychology (Special Issue), 19, pp. 3–192. Spielberger, C. D. (1988). Professional manual for the State-Trait Anger Expression Inventory (STAXI) (Research Ed.). Odessa, FL: Psychological Assessment Resources. Spielberger, C. D. & Comunian, A. L. (1992). STAXI: State-Trait Anger Expression Inventory: Versione adattamento italiano. Manuale [Test manual of the Italian version of the State-Trait Anger Expression Inventory]. Florence, Italy: Organizzazioni Speciali. Spielberger, C. D., Johnson, E. G., Russell, S. F., Crane, R. S., Jacobs, G. A., & Worden, T. J. (1985). The experience and expression of anger. In M. A. Chesney & R. H. Rosenman (Eds.), Anger and hostility in cardiovascular and behavioral disorders (pp. 5–30). New York: Hemisphere/McGraw-Hill. Spielberger, C. D., Krasner, S. S., & Solomon, E. P. (1988). The experience, expression, and control of anger. In M. P. Janisse (Ed.), Health psychology: Individual differences and stress (pp. 89–108). New York: Springer-Verlag. Spielberger, C. D., Reheiser, E. C., & Sydeman, S. J. (1995). Measuring the experience, expression, and control of anger. In H. Kassinove (Ed.), Anger disorders: Definitions, diagnosis, and treatment (pp. 49–72). Washington, DC: Taylor & Francis. Spielberger, C. D., & Sharma, S. (1976). Cross-cultural measurement of anxiety. In C. D. Spielberger & R. Diaz-Guerrero (Eds.), Cross-cultural anxiety (pp. 13–25). Washington, DC: Hemisphere/Wiley. Spielberger, C. D., Sharma, S., & Singh, M. (1973). Development of the Hindi edition of the State-Trait Anxiety Inventory. Indian Journal of Psychology, 48, pp. 11–20.

RT44367_C011.indd 205

7/13/05 9:34:52 AM

206

S. SHARMA ET AL.

Tanzer, N. K., Sim, C. Q. E., & Spielberger, C. D. (1996). Experience, expression, and control of anger in a Chinese society: The case of Singapore. In C. D. Spielberger & I. G. Sarason (Eds.), Stress and emotions: Anxiety, anger, and curiosity (Vol. 16, pp. 51–65). Washington, DC: Taylor & Francis. Törestad, B. (1990). What is anger provoking? Provokers and provoked in anger situations: Developmental trends and sex differences. Aggressive Behavior, 16, pp. 353–359.

RT44367_C011.indd 206

7/13/05 9:34:52 AM

III CROSS-CULTURAL RESEARCH ON OCCUPATIONAL STRESS

RT44367_S003.indd 207

7/13/05 9:34:26 AM

RT44367_S003.indd 208

7/13/05 9:34:26 AM

12 The German Adaptation of the Job Stress Survey: A Multistudy Validation in Different Occupational Settings Volker Hodapp J.-W. Goethe-Universität, Frankfurt, Germany Norbert K. Tanzer Alliant International University, San Diego, CA Christian Korunka University of Vienna, Austria Ehrentraud R. Maier and Iris A. Pestemer Karl-Franzens-Universität, Graz, Austria Abstract The Job Stress Survey (JSS) is a self-report inventory designed to measure the perceived severity and frequency of occurrence of 30 job-related stressors that are encountered by female and male employees in a broad range of occupational settings. Two occupational stress factors, Job Pressure (e.g., meeting deadlines) and Lack of Organizational Support (e.g., fellow workers not doing their job), have been identified and found to be stable across gender and organizational levels. This chapter describes the development of a German adaptation of the JSS (JSS-G) and reports the results of validation studies for four occupational groups: primary-school teachers; secondaryschool teachers; medical and nonmedical personnel associated with a large hospital; and bank employees. In all four studies, the two-factor structure of the original English JSS was replicated in the German adaptation, demonstrating cross-cultural construct validity for the JSS-G. Intercorrelations between the JSS-G and an organizational climate questionnaire, with scales that assessed staff cohesion, supervisory support, job satisfaction, and burnout, provided evidence of the convergent and discriminant validity of the German JSS.

209

RT44367_C012.indd 209

7/13/05 9:31:45 AM

210

V. HODAPP ET AL.

THE GERMAN ADAPTATION OF THE JOB STRESS SURVEY: A MULTISTUDY VALIDATION IN DIFFERENT OCCUPATIONAL SETTINGS Stress in the workplace has become an increasing problem in our society. It is well established that job-related stress affects productivity, absenteeism, worker turnover, and employee health and well-being (Cooper & Payne, 1988; Hinton & Hodapp, 1997; Karasek & Theorell, 1990; Keita & Sauter, 1992; Matteson & Ivancevich, 1987; Schnall & Landsbergis, 1994; Spielberger, 1994; Spielberger & Vagg, 1999). Absenteeism, turnover, and medical problems also have substantial direct costs for employers. Other factors, such as reduced productivity and diminished customer services, are hidden costs that often result from distressed and exhausted employees (Karasek & Theorell, 1990). Moreover, many employees with stress-related workplace problems expect compensation, including disability payments and early retirement benefits. Given the serious consequences of job-related stress for both employees and organizations, it is increasingly important to understand and evaluate the adverse effects of stress at work. Unfortunately, the tremendous increase in research on occupational stress has not contributed to clarifying the findings of these studies because of the wide range of differences in research methodology and ambiguity in the conceptual definition of job stress (Spielberger & Reheiser, 1994; Weyer, 1982). Although some researchers have focused on antecedent conditions and the duties and responsibilities of a particular job, others have been concerned primarily with the consequences of environmental and job-related pressures. Moreover, the specific job pressures and work environment variables that differentially influence stress levels for men and women working at higher and lower organizational levels tend to be neglected. This chapter describes the construction of a German adaptation (JSS-G) of Spielberger’s (1994) Job Stress Survey (JSS) and evaluates the factor structure of this measure. The results of four JSS-G validation studies carried out with employees from different occupational groups are reported. These groups included teachers from primary and secondary schools (studies 1 and 2); medical, nursing, administrative, and technical personnel from a large psychiatric hospital (Study 3); and employees from commercial banks (Study 4). Because student misbehavior has been identified as a major source of stress for teachers in a number of studies (cf. Barth, 1992; Boyle, Borg, Falzon, & Baglioni, Jr., 1995; Brissie, Hoover-Dempsey, & Bassler, 1988; Burke, 1992; Cooper, 1995; Friedman, 1991; Harris, Halpin, & Halpin, 1985; Hart, 1994; Hart, Wearing, & Conn, 1995; Litt & Turk, 1985; Pithers & Fogarty, 1995; Tellenback, Brenner, & Löfgren, 1983), several items pertaining to student misconduct and teaching-related stress were included in Studies 1 and 2. THE ORIGINAL ENGLISH-LANGUAGE JOB STRESS SURVEY The Job Stress Survey is a self-report inventory designed to assess the impact of sources of stress in the workplace (Spielberger, 1994; Spielberger &

RT44367_C012.indd 210

7/13/05 9:31:45 AM

THE GERMAN ADAPTATION OF THE JOB STRESS SURVEY

211

Vagg, 1999). It consists of 30 statements that describe job-related events identified as stressful by employees in a variety of occupations (“job stressors”). Each of the 30 job stressor events is rated twice: Respondents first rate the perceived severity of the stress evoked by each JSS stressor event (“Severity” ratings); the second rating is for the frequency that each stressor was experienced during the last six months (“Frequency” ratings). The perceived severity of each JSS item is rated on a nine-point intensity scale. Extreme ratings (1 and 9) are respectively labeled “Low” and “High”; the midpoint of the Severity scale, with a score of 5, is labeled “Moderate.” Respondents are instructed to base their ratings on personal experience, taking into account the amount of time and energy required to adjust to each stressor event as compared to the “Assignment of disagreeable duties.” This “standard stressor,” with a preassigned severity rating of 5, was rated as moderately stressful in previous research by employees working at different organizational levels in diverse work settings. Respondents also rated, on a 10-point scale, the approximate number of days each stressor was experienced during the last six months. Ratings of 0 indicate that the event did not occur at all; a rating of 9+ indicates that the stressor event was experienced on nine or more days. The sum of the individual Severity and Frequency ratings of all 30 JSS items yields overall stress Severity and Frequency scores. For each JSS stressor event, a Stress Index score is calculated by multiplying the Severity rating by the Frequency rating. The sum of the cross-products of these ratings for the 30 JSS items (i.e., the sum of the Item Index scores) gives an overall Job Stress Index score, which provides an estimate of the total impact of occupational stress experienced by a particular worker. Factor analyses of the 30 JSS Severity and Frequency items have consistently identified Job Pressure (JP) and Lack of Organizational Support (LS) as the major components of occupational stress (cf. Spielberger & Reheiser, 1994). These components are assessed with factorially derived 10-item JSS Job Pressure and Lack of Organizational Support subscales. DEVELOPMENT OF THE GERMAN ADAPTATION OF THE JOB STRESS SURVEY It is widely accepted that the adaptation of a psychological test involves more than simply translating each test item into another language. The appropriate adaptation of a psychological test requires a balanced consideration of psychological, linguistic, and cultural factors (cf. Bracken & Barona, 1991; Brislin, 1980, 1986; Geisinger, 1994; Hambleton, 1994; Poortinga, 1989; Van de Vijver & Hambleton, 1996; Van de Vijver & Leung, 1997a, 1997b; Van de Vijver & Tanzer, 1997). In keeping with previous research, the German adaptation of the Job Stress Survey was developed in a series of studies (Hodapp, Tanzer, & Spielberger, 1996). Based on the empirical results and the experience gained from the preceding studies, the JSS-G was further modified for administration in each subsequent study. The four different versions of the JSS-G and the results of the studies based on each of these versions will be described.

RT44367_C012.indd 211

7/13/05 9:31:46 AM

212

V. HODAPP ET AL.

Version 1. This first version of the JSS-G consisted of 28 items that were translated from the original English-language JSS by the second author for a study of Austrian teachers (cf. Study 1). Item 7 (“Dealing with crisis situations”) and Item 12 (“Periods of inactivity”) were not included. The JSS items were administered as part of a larger questionnaire (cf. Tanzer & Schwetz, 1996a), which included job stressors specific to the teaching profession and items related to student misconduct that were adapted from the Teacher Stress Survey (Grier, 1981). In rating stress severity, the respondents were instructed to assign a numerical value, between 0 (“Not stressful at all”) and 100 (“Highly stressful”), with a rating of 50 to indicate an average or moderate amount of stress. The frequency that each stressor was experienced was rated on a five-point scale, ranging from 0 (“Never”) to 4 (“All the time”). In order to evaluate the appropriateness of the English anchor item, “Assignment of disagreeable duties,” as a standard stressor that evoked an average amount of stress, the rating procedure was also applied to this item. Version 2. Minor linguistic revisions based on the empirical results of Study 1 were made in the version of the JSS-G used in Study 2. The format for the Severity scale in Version 2 also differed from Version 1. Instead of using a numerical rating scale with values ranging from 0 to 100, respondents rated the severity of each stressor on a five-point scale, with ratings ranging from 0 (“No stress at all”) to 4 (“Very high degree of stress”). The format for the Frequency scale and the comparison of the perceived severity of each stressor event with the standard anchor item were the same as in Version 1. In addition, translations of Items 7 and 12 from the English JSS were included in Version 2. Version 3. In developing Version 3 of the JSS-G, the form used in Version 2 was compared with an independent German translation of the original English JSS developed by the first author. The content and the connotative meanings of the 30 German JSS items were also back-translated, and each item was discussed with the author of the original English version. In addition, two alternative German translations of Item 26 (“Meeting deadlines”) were also included in Version 3, resulting in a 31-item form of the JSS-G. The procedures for rating perceived Severity and Frequency in Version 3 were modified to make them more similar to the original English JSS. The format for the perceived severity ratings was changed from a five-point scale to a nine-point scale; scores of 1 to 3 indicate a “Low degree of stress,” scores of 4 to 6 indicate an average or moderate amount of stress, and ratings of 7 to 9 indicate a “High degree of stress.” As in the original English JSS, “Assignment of disagreeable duties” was used as the standard anchor for the Severity ratings, with an assigned value of 5 on the 9-point Severity scale. The procedures for the Frequency ratings in Version 3 were also more similar to those used with the English JSS. Frequency of stressor experience was rated on a nine-point scale, with scores ranging from 1 (“Never”) to 9+ (“All the time”). Version 4. As compared to Version 3, only relatively minor changes were made in the instructions for Version 4 and in reformulating the wording of the individual items. Based on the results of the three previous studies, the most important difference between Version 4 and the earlier versions was that

RT44367_C012.indd 212

7/13/05 9:31:46 AM

THE GERMAN ADAPTATION OF THE JOB STRESS SURVEY

213

“Assignment of disagreeable duties” was no longer used as the anchor for rating the Severity items. STUDY 1: PRIMARY-SCHOOL TEACHERS Participants and Procedure A sample of 175 Austrian teachers (74% female, 26% male) who were teaching in primary or vocational schools in both rural and urban areas participated in this study on a voluntary and anonymous basis (cf. Tanzer & Schwetz, 1996a). They responded to a 76-item questionnaire on job-related stressors (cf. Version 1) that was administered at a meeting of a local teachers association or during a continuing-education workshop. Each item was rated for both perceived severity and frequency of occurrence. The abbreviated content for the 28 JSS stress items is given in Table 12.1 (Items 1–30; Items 7 and 12 were not included). The four items related to teacher-specific job pressures (Items 41–44) and the six items on teaching stress and student misconduct (Items 51 to 56) are also listed in Table 12.1. After completing the questionnaire, the participants were asked to rate the overall severity and frequency of their experience of job-related stress (“Overall Job-Stress Rating”) by responding to a single question: “How much and how often do you feel stressed in your job?” Results In all four studies, separate principal-axis exploratory factor analyses (EFAs), with subsequent oblimin-rotation (δ = 0), were carried out for the Severity and Frequency ratings and for the Stress Index scores (i.e., Frequency × Severity), which provided an estimate of the overall Stress Impact of the JSS-G stressor events. The EFAs were based on Pearson product-moment correlation matrices and were carried out with a Windows 6.0 version of the commercially available SPSS program (Norusis, 1993). The number of factors that were extracted in these analyses was based on eigenvalues greater than 1.00 and scree tests, the criteria most commonly used in EFA (cf. Bernstein, 1988). Factor structure. The Severity and Frequency ratings and Stress Index Impact scores based on the 28 JSS-G items were factored together with similar data based on the four teacher-specific job stress and six teacher stress/student misconduct items. For all three sets of scores, the scree plots indicated that the three-factor solutions provided the best fit, accounting for 36% of the total variance for the Stress Index scores, and 39% and 34% of the respective variance for the Severity and Frequency ratings. The oblimin-rotated factor loadings for the Stress Index (Impact), Frequency, and Severity scores are reported in Table 12.1. On the basis of the loadings of the original 28 JSS-G items (1–30), the first factor identified in all three analyses was Lack of Organizational Support, one of the posited job stress dimensions assessed by a JSS subscale. Eight of the 10 JSS-G items adapted from the original JSS Lack of Support subscale

RT44367_C012.indd 213

7/13/05 9:31:46 AM

214

V. HODAPP ET AL.

Table 12.1 Factor Loadings for the Stress Index, Frequency, and Severity Items in Study 1 Factor 1 Scale/Items Lack of Organizational Support 05. Colleagues not doing their job 06. Inadequate supervisory support 08. No recognition for good work 13. Difficulties with supervisor 14. Negative attitudes toward school 17. Insult from colleagues, etc. 18. No participation in policy-making 20. Competition for advancement 21. Poor or inadequate supervision 29. Unmotivated colleagues 30. Conflicts with others Job Pressure 01. Disagreeable duties 02. Working overtime 04. Unfamiliar duties 09. Tasks not in job description 10. Inadequate equipment 11. Increased responsibility 15. Insufficient personnel 16. Making on-the-spot decisions 19. Inadequate salary 23. Frequent interruptions 24. Boring/demanding activities change 25. Excessive paperwork 26. Meeting deadlines 27. Insufficient personal time 28. Covering work for others 41. Red tape by school board 42. Preparing lessons 43. Correcting assignments 44. Grading Teaching Stress 22. Noisy work area 51. Cheeky students 52. Students with behavior disorders 53. Disciplinary problems 54. Unmotivated students 55. Covering child-rearing tasks 56. Teaching in noisy classrooms Nontarget Item 03. Lack of advancement prospects

Factor 2

Im

Fr

Se

67 65 55 53 62 57 34 40 58 58 69

62 71 52 56 54 68

72 62 41 60 71 62 35 48 61 52 71

53 53 62 78

32 39 37

Im

Fr

Factor 3 Se

Im

Fr

Se

43 66 75 61 55 47 73

54 72 66 71 53 39 67

49 72 74 57 59 37 78

42

32

34

60 47 56 56 31 70 40 52 37 42 44 60 50 40 58 47 42 50 63

48 46 45 51 65 38 63 41 46 64 60 48 54 49 61 54

59 65 46 45 38 65 31 35 47 40 44 51 48 49 48 38 53 54 66

32

Note. JSS Items 7 and 12 were not used in this study. Decimal points and item loadings of less than .30 are omitted from the table. Im = Impact (i.e., Frequency × Severity) scores; Fr = Frequency ratings; Se = Severity ratings.

RT44367_C012.indd 214

7/13/05 9:31:47 AM

215

THE GERMAN ADAPTATION OF THE JOB STRESS SURVEY

Table 12.2 Intercorrelations between Lack of Organizational Support, Job Pressure, Teaching-Related Stress, and Overall Stress Rating for the Impact, Frequency, and Severity Scores (Study 1) Impact Scale

Frequency

Severity

1

2

3

1

2

3

1

2

3

Impact 1. Lack of Support 2. Job Pressure 3. Teaching Stress

— .45 .20

.45 — .32

.20 .32 —

.89 .39 .20

.36 .77 .25

.12 .21 .90

.80 .46 .17

.39 .90 .30

.20 .35 .87

Frequency 4. Lack of Support 5. Job Pressure 6. Teaching Stress

.89 .36 .12

.39 .77 .21

.20 .25 .90

— .40 .18

.40 — .26

.18 .26 —

.68 .34 .06

.31 .62 .14

.13 .26 .72

Severity 7. Lack of Support 8. Job Pressure 9. Teaching Stress

.80 .39 .20

.46 .90 .35

.17 .30 .87

.68 .31 .13

.34 .62 .26

.06 .14 .72

— .57 .34

.57 — .43

.34 .43 —

Overall Stress Rating 10. Impact 11. Frequency 12. Severity

.28 .23 .31

.51 .46 .50

.49 .50 .50

.24 .23 .25

.37 .33 .36

.39 .42 .39

.30 .24 .33

.52 .44 .53

.47 .45 .52

loaded on this factor. Nine of the 10 JSS-G items adapted from the original JSS Job Pressure subscale had dominant salient loadings on Factor 2, clearly a Job Pressure factor. Two LS items (10, “Inadequate Equipment,” and 19 “Inadequate Salary”) also had small dominant loadings on Factor 2, reflecting job pressures experienced by most public-school teachers. JP Item 3 (“No advancement prospects”) had a moderate nontarget loading. All four of the teacher-specific job pressure items (41–44) also loaded on Factor 2, the Job Pressure factor. The third factor was defined almost entirely by the dominant salient loadings of all six items that were designed to assess teacher stress/student misconduct (51–56). For example, Item 52 (“Students with behavior disorders11”) had the highest loadings on Factor 3 for both the JSS-G Stress Impact and Severity scores. Item 22 (“Noisy work area”) also had a moderate loading on Factor 3, in keeping with the typical work environment of the teaching profession. Convergent and discriminant validity. Correlations of the JSS Lack of Support, Job Pressure, and Teacher-Related Stress scales with the overall Stress Index (Impact), Frequency, and Severity scores based on all 28 JSS items are reported in Table 12.2. As expected, the Lack of Organizational Support and Job Pressure subscales correlated substantially with all three

RT44367_C012.indd 215

7/13/05 9:31:48 AM

216

V. HODAPP ET AL.

Figure 12.1. Distribution of the Severity ratings for the JSS-stressor “Assignment of disagreeable duties” on the 100-point rating scale used in Study 1.

JSS scale scores, providing evidence of convergent validity. However, the correlations of the overall Stress Impact scores were consistently higher with Job Pressure (.33 to .53) and Teacher-Related Stress (.39 to .52) than with Lack of Organizational Support (.23 to .33). Severity ratings for the standard stressor. In contrast to the original English JSS, in which “Assignment of disagreeable duties” (Item 1) is used as an anchor, with an assigned Severity rating of 5, the German translation of this item in Study 1 was freely rated by all participants. The distribution of the Severity ratings for this item are reported in Figure 12.1. The broad range of these ratings, along with a high proportion of 0 ratings, casts doubt on the appropriateness of the translation of this anchor item as a standard stressor that evoked an average amount of stress for a majority of the study participants. Consequently, a different rating format was used in Study 2 to investigate the average level of perceived severity for “Assignment of disagreeable duties” as the standard for the JSS-G. STUDY 2: SECONDARY-SCHOOL TEACHERS Participants and Procedure The participants in this study were 106 teachers (86% female, 14% male) from 11 primary schools and 131 teachers (60% female, 40% male) from eight secondary vocational schools in a rural Austrian school district. The

RT44367_C012.indd 216

7/13/05 9:31:48 AM

THE GERMAN ADAPTATION OF THE JOB STRESS SURVEY

217

teachers were informed that participation in the study was voluntary and anonymous. Those who agreed to participate responded to two questionnaires, which were given to them in their schools, and were instructed to return the completed questionnaires to the researchers in sealed envelopes (cf. Tanzer & Schwetz, 1996a). The first questionnaire included the 30 JSS-G items (cf. Version 2) and the same four teacher-specific job pressure and six teacher stress/student misconduct items as in Study 1. The second questionnaire was a teacher version of the 36-item Organizational Climate Questionnaire (OCQ; Tanzer & Schwetz, 1996b; see also Oswald, Pfeifer, Ritter-Berlach, & Tanzer, 1989), which consisted of four scales for assessing Staff Cohesion (9 items), Supervisory Support (11 items), Job Satisfaction (8 items), and Burnout (8 items). Each of the 36 OCQ items comprising these scales was rated on a four-point Agree-Disagree scale. The participants were also asked to rate the overall severity and frequency of their job-related stress experience, as in Study 1 (“Overall Job-Stress Rating”). Results Factor structure. The scree plot for the Severity ratings in Study 2 indicated that a two-factor solution, which explained 36% of the total variance, was optimal. Because all 30 JSS-G items loaded on the first factor and the six teaching stress/student misconduct items (51–56) loaded on the second factor, the detailed results of this analysis are not reported. In the factor analyses of the Stress Index Impact and Frequency scores, the scree plots indicated that the three-factor solutions provided the best fit, accounting for 40% of the total variance for the Impact scores and 38% for the Frequency scores. The results of the oblimin-rotated factor loadings for these three-factor solutions are reported in Table 12.3. The results of the factor analysis for the Impact and Frequency scores were similar to the findings in Study 1. However, a larger number of items had dual or nontarget loadings. Of the 14 items with dominant loadings on Factor 1 (Organizational Support), the Impact scores for four of these items (5, 17, 29, 30) had salient secondary loadings on Factor 2, which was defined primarily by Job Pressure items. The Impact score for Item 9 (“Tasks not in job description”) had an expected dominant loading on Factor 2 (Job Pressure), but an even stronger nontarget salient loading on Factor 1 (Organizational Support). Furthermore, six Job Pressure items (1, 2, 4, 7, 11, 16) had secondary salient loadings on the Organizational Support factor, and Items 19 and 24 did not have salient loadings on either factor. Convergent and discriminant validity. Table 12.4 shows the intercorrelations of the four OCQ scales with the Impact (Severity × Frequency) scores of the JSS-G Lack of Organizational Support, Job Pressure, and TeachingStress scales and the Overall Stress rating. The Overall Stress rating correlated more highly with Job Pressure (r = .48) and Teaching-Related Stress (r = .57) than with Lack of Organizational Support (r = .26). Two of the OCQ scales, Supervisory Support and Staff Cohesion, were more highly correlated with Lack of Organizational Support (rs = –.68 and –.60) than with either Job

RT44367_C012.indd 217

7/13/05 9:31:48 AM

218

V. HODAPP ET AL.

Table 12.3 Factor Loadings for the Impact and Frequency Scores of the Job Stressors (Study 2) Scale/Items Lack of Organizational Support 03. No advancement prospects 05. Colleagues not doing their job 06. Inadequate supervisory support 08. No recognition for good work 10. Inadequate equipment 13. Difficulties with principal 14. Negative attitudes toward school 15. Insufficient personnel 17. Insult from colleagues, etc. 18. No policy-making decision participation 20. Advancement competition 21. Poor supervision 29. Unmotivated colleagues 30. Conflicts with others

Factor 1

Factor 2

Factor 3

Im

Fr

Im

Im

Fr

41 44 84 74 37 69 74 42 42 60 52 74 52 50

61 82 70 46 77 70 61 61 65 56 79 65 68

42

40 33 33 32 43 44

76 55 48 40 69 62 35 60 77 42 40 37

55 74 82 72 51 41 76

50 76 67 78 51 44 72

Job Pressure 01. Disagreeable duties 02. Working overtime 04. Unfamiliar duties 07. Dealing with crisis situations 11. Increased duties 16. Making on-the-spot decisions 23. Frequent interruptions 25. Excessive paperwork 26. Meeting deadlines 41. Red tape by school board 42. Preparing lessons 43. Correcting assignments 44. Grading

40

34 44

Teaching 22. Noisy work area 51. Cheeky students 52. Students with behavior disorders 53. Disciplinary problems 54. Unmotivated students 55. Covering child-rearing tasks 56. Teaching in noisy classrooms Nontarget/Dual-Loading Items 09. Tasks not in job description 12. Periods of inactivity 19. Inadequate salary 24. Boring/demanding activity changes 27. Insufficient personal time 28. Covering work for others

31

56

30 48

Fr

35 33 44 39 60 68 49 33 45 37

59 34

55

Note: For the Severity ratings, the scree test indicated a two-factor solution. Decimal points and item loadings of less than .25 are omitted from the table. Im = Impact (i.e., Frequency × Severity) scores; Fr = Frequency ratings.

RT44367_C012.indd 218

7/13/05 9:31:49 AM

219

THE GERMAN ADAPTATION OF THE JOB STRESS SURVEY Table 12.4 Intercorrelations between the Organizational Climate Scales and the Impact Scores of the Job Stress Scales (Study 2) Job Stress Scale

1

Job Stressa 1. Organizational Support 2. Job Pressure 3. Teaching Stress 4. Overall Stressb

2

3

Climate 4

5

JS/BO

6

7

8

Study 2 (N = 285) — .49 .28 .26

.49 — .34 .48

.28 .34 — .57

.26 .48 .57 —

–.68 –.23 –.13 –.10

–.60 –.28 –.23 –.26

–.33 –.20 –.38 –.30

.19 .50 .50 .66

Organizational Climate 5. Supervisory Support 6. Staff Cohesion

–.68 –.60

–.23 –.28

–.13 –.23

–.10 –.26

— .62

.62 —

.36 .34

–.10 –.11

Job Satisfaction/Burnout 7. Job Satisfaction 8. Burnout

–.33 .19

–.20 .50

–.38 .50

–.30 .66

.36 –.10

.34 –.11

— –.44

–.44 —

a b

Impact scores. Overall Stress Rating.

Pressure (rs = –.23 and –.28) or Teaching-Related Stress (rs = –.13 and –.23). In contrast, the OCG Burnout scale correlated more highly with Job Pressure (r = .50), Teaching-Related Stress (r = .50), and Overall Stress (r = .66) than with Organizational Support (r = .19). Severity ratings of the standard stressor. Figure 12.2 shows the distribution of the Severity ratings for the German translation of the standard stressor, “Assignment of disagreeable duties,” which was rated on a five-point scale, with scores ranging from 0 (“No stress at all”) to 4 (“Very high degree of stress”). Although a majority of the participants rated this item in the average range of stress (M = 1.80), the broad dispersion of the ratings (SD = 0.99) and the high percentage of extreme ratings (13%) indicated that disagreeable duties had a different stress impact for different people. STUDY 3: HOSPITAL PERSONNEL Subjects and Procedure A set of questionnaires was given to 1,306 employees who were associated with a large Viennese psychiatric hospital. Of the 574 employees (62% female, 38% male) who responded (44% return rate), 94 (16%) were medical doctors, 324 (56%) were nurses, 116 (20%) were administrative staff, and 40 (7%) were technical personnel. Response rates were very similar for all groups of employees. In addition to Version 3 of the JSS-G, the participants completed a

RT44367_C012.indd 219

7/13/05 9:31:50 AM

220

V. HODAPP ET AL.

Figure 12.2. Distribution of the Severity ratings for the JSS-stressor “Assignment of disagreeable duties” on the five-point rating scale used in Study 2.

somatic complaints questionnaire and several other scales. Further information regarding these measures can be found in Jung (1997). Results Factor structure. The scree plots in the factor analyses of the JSS-G Impact, Severity, and Frequency items indicated that two-factor principal-axis solutions provided the best fit in all three analyses. These factors explained 40.6% of the total variance of the Impact scores, 41.7% for the Frequency ratings, and 40.0% for the Severity ratings. The factor loadings for the 20 items that most clearly defined each of the two factors are reported in Table 12.5. Based on the content of the items, these factors were labeled Lack of Organizational Support and Job Pressure. From the original 10-item JSS subscales, 7 Organizational Support and 8 Job Pressure items loaded on their respective posited factor. As in the two previous studies, the two factors were correlated positively for all three types of scores (rs between .47 and .52). Group and gender differences. The means, standard deviations, and Cronbach alpha coefficients for the JSS-G Impact, Frequency, and Severity scores are reported in Table 12.6 for females and males, and for the combined sample. Although the females tended to have higher JSS-G scores than the males, these differences were not significant (all ps > .05). For both gender groups and for the combined sample, the alpha coefficients were all .92 or greater, providing strong evidence of the internal consistency reliability of the JSS-G Stress Impact, Severity, and Frequency scales.

RT44367_C012.indd 220

7/13/05 9:31:51 AM

221

THE GERMAN ADAPTATION OF THE JOB STRESS SURVEY Table 12.5 Factor Loadings for the Impact, Frequency, and Severity Scores of the Job Stressors (Study 3) Factor 1 JSS-G Scales and Items Lack of Organizational Support 05. Colleagues not doing their job 06. Inadequate supervisory support 08. No recognition for good work 10. Inadequate equipment 12. Periods of inactivity 13. Difficulties with principal 14. Negative attitudes toward organization 18. No policy-making decision participation 21. Poor supervision 29. Unmotivated colleagues

Factor 2

Im

Fr

Se

52 74 59 45 57 73 52 79 40 72

45 79 69 41 50 80 53 74 56 67

64 77 60 54 54 65 52 83 46 73

Job Pressure 04. Unfamiliar duties 07. Dealing with crisis situations 11. Increased duties 16. Making on-the-spot decisions 23. Frequent interruptions 24. Boring/demanding activity changes 25. Excessive paperwork 27. Insufficient personal time 28. Covering work for others 31. Time pressure Factor Correlation

Im

34

63 68 79 73 68 43 63 59 58 86 47

52

Fr

Se

35

54 64 78 71 70 49 71 62 65 90

75 66 75 76 55 41 54 58 46 82

51

Note: Decimal points and item loadings of less than .25 are omitted from the table. Im = Impact (i.e., Frequency × Severity) scores, Fr = Frequency ratings, Se = Severity ratings.

To evaluate occupational group differences, the JSS-G Impact, Frequency, and Severity scores of medical doctors, nurses, administrative staff, and technical personnel were compared in 2 by 2 analyses of variance (ANOVAs) with Gender and Occupation as the independent variables. Because very few females were represented among the technical personnel, this group was combined with the administrative staff in the ANOVA analyses. The means and standard deviations for the resulting three occupational groups are reported in Table 12.7. No significant gender differences or Gender by Occupation interactions were found (all ps > .05). The F-statistics for the main effects of Occupation, which are reported in Table 12.7, indicated substantial group differences. The doctors had the highest stress scores on the JSS-G Lack of Organizational Support and Job Pressure subscales, followed by the nurses; these differences were more pronounced for Lack of Organizational Support.

RT44367_C012.indd 221

7/13/05 9:31:51 AM

222

V. HODAPP ET AL.

Table 12.6 Means, Standard Deviations, and Alpha Coefficients for Impact, Frequency, and Severity Scores of the JSS-G Total Scale (Study 3) JSS-G Total Scores

Female

Male

Total

t

22.52 11.18 246 .92

21.42 11.14 165 .92

22.08 11.17 411 .92

.98

Frequency M SD n α

122.43 40.58 283 .93

121.21 40.16 180 .93

121.96 40.38 463 .93

.32

Severity M SD n α

138.83 39.79 289 .92

133.95 39.46 190 .92

136.81 39.69 480 .92

1.32

Impact M SD n α

The results of Study 3 were generally similar to the data reported for the original JSS (Spielberger & Vagg, 1999). Severity ratings of the standard stressor. In Study 3, “Assignment of disagreeable duties” was used as the anchor item, as with the original English-language JSS. The participants were instructed to base their Severity ratings on the amount of time and energy required to adjust to each stressor and to compare this with the anchor item. However, a number of subjects did not comply with this instruction and explicitly cancelled the preassigned Severity rating for the standard stressor by marking a different rating category for this item. Consequently, in Study 4, the standard anchor stressor was converted back to a free-rating item. STUDY 4: BANK EMPLOYEES Participants and Procedure The participants were 287 bank employees (51% women, 49% men) from 27 branches of nine Austrian banks. The wide range of job profiles included employees with administrative, customer service, and clerical duties. After an interval of four to five months, 213 of the employees were retested. On both the initial and retest occasions, the questionnaires were given at their place of employment to the participants, who were informed of the purpose of the study and that their participation was voluntary and anonymous. The

RT44367_C012.indd 222

7/13/05 9:31:51 AM

223

THE GERMAN ADAPTATION OF THE JOB STRESS SURVEY Table 12.7 Means, Standard Deviations, and Alpha Coefficients for the Impact, Frequency, and Severity Scores of the Job Stressors for Different Occupational Groups (Study 3) OCCUPATIONAL GROUPS

JSS-G Subscales

Doctors Female

Male

Nurses

Administrative and Technicala

Female

Male

Female

Male

Fb

Lack of Organizational Support Impact M SD n

23.89 12.51 64

27.34 16.08 25

22.66 13.53 194

23.25 12.76 93

16.92 13.15 48

15.78 12.80 70

14.00***

Frequency M SD n

39.89 15.17 65

44.84 18.62 25

39.27 15.06 201

41.36 15.11 96

33.02 15.60 54

31.78 13.92 77

15.04***

Severity M SD n

52.38 15.92 66

52.63 16.26 27

48.33 16.29 204

48.92 14.94 100

40.86 17.30 57

37.56 16.50 75

21.45***

Impact M SD n

23.65 12.51 64

24.57 16.08 25

21.95 13.53 194

24.22 12.76 93

22.45 13.15 48

17.93 12.80 70

2.93

Frequency M SD n

44.89 16.69 64

46.56 13.56 25

43.61 16.27 200

47.72 14.18 98

41.85 17.40 55

36.84 15.57 75

7.87***

Severity M SD n

46.18 16.43 66

47.92 13.39 25

44.01 14.62 197

45.89 13.67 102

43.44 17.19 57

39.08 15.43 78

4.44**

Job Pressure

Note: aGroups 3 (administrative staff) and 4 (technical personnel) were combined for the ANOVA analyses. b F-ratios for main effect of Occupation. **p < .01; ***p < .001.

RT44367_C012.indd 223

7/13/05 9:31:52 AM

224

V. HODAPP ET AL.

completed questionnaires were returned to the researchers in sealed envelopes (cf. Maier, Pestemer, & Tanzer, 1997). The test and retest questionnaires of 157 participants were matched according to their bank affiliation, gender, age, and an anonymous code. Instruments. Version 4 of the JSS-G was used in this study. In addition to responding to the 31 JSS-G items, the participants were asked to describe up to three additional stressful job-related events that they had experienced during the last six months. A second instrument, a version of the Organizational Climate Questionnaire that was specially designed for bankers, was also used (cf. Maier et al., 1997). The 32-item banker’s OCQ consisted of four scales for assessing Staff Cohesion (8 items), Supervisory Support (11 items), Job Satisfaction (7 items), and Burnout (6 items). As in Study 2, the participants rated each OCQ item on a four-point Agree-Disagree rating scale. Results Factor structure. The scree tests in the factor analyses of the JSS-G Stress Impact, Severity, and Frequency scores indicated that two factors could be extracted which explained 30% of the total variance of the Stress Impact scores, 32% of the Frequency ratings, and 33% of the Severity ratings. The oblimin-rotated factor loadings in these analyses, which are reported in Table 12.8, identified Lack of Organizational Support and Job Pressure factors that were similar to those found in the three previous studies. All 10 items that were adapted from the original English-language JSS Lack of Organizational Support scale loaded on the first factor. Nine of the 10 JSS-G Job Pressure items had dominant salient loadings on Factor 2. Similar to the findings in Study 2, Items 24 (“Frequent changes from boring to demanding activities”) and 27 (“Insufficient personal time”) did not have salient loadings on either factor. Convergent and discriminant validity. The correlations between the six JSS-G subscales are reported in Table 12.9. As in the previous studies, the JSS-G Lack of Organizational Support and Job Pressure subscales were substantially correlated (rs between .47 and .52). The pattern of intercorrelations between the Stress Impact (Severity × Frequency) scores for the two JSS-G subscales and the four OCQ scales are reported in Table 12.10 and closely resembled the findings in Study 2. The OCQ Supervisory Support and Staff Cohesion Organizational Climate scales correlated more highly with Lack of Organizational Support (rs = –.60 and –.54) than with Job Pressure (both rs = –.27). Similar to the findings in Study 2, the OCQ Burnout scale correlated more highly with Job Pressure (r = .51) than with Lack of Organizational Support (r = .38), and Job Satisfaction correlated more negatively with Lack of Organizational Support (r = –.42) than with Job Pressure (r = –.30). Severity ratings of the standard stressor. Figure 12.3 shows the distribution of the Severity ratings for the German translation of “Assignment of disagreeable duties.” Despite the average mean severity rating of 4.80 on the nine-point scale, the broad dispersion of the item ratings (SD = 1.89) and the substantial percentage of ratings for both “Low-stress” and “High-stress” indicated that the

RT44367_C012.indd 224

7/13/05 9:31:52 AM

225

THE GERMAN ADAPTATION OF THE JOB STRESS SURVEY Table 12.8 Factor Loadings for the Impact, Frequency, and Severity Scores of the Job Stressors (Study 4) Factor 1 JSS-G Scales and Items Lack of Organizational Support 03. No advancement prospects 05. Colleagues not doing their job 06. Inadequate supervisory support 08. No recognition for good work 09. Tasks not in job description 10. Inadequate equipment 13. Difficulties with supervisor 14. Negative attitude toward organization 17. Insult from colleagues, etc. 18. No policy-making participation 19. Inadequate salary 20. Advancement competition 21. Poor supervision 28. Covering work for others 29. Unmotivated colleagues 30. Conflicts with other departments

Factor 2

Im

Fr

Se

54 59 66 62 46 38 46 54 33 65 60 53 37 50 52 37

49 59 69 67 55 38 57 62 44 64 62 57 49 57 54 41

57 53 66 56 30 61 56 44 44 61 72 60 46 43 61 48

Job Pressure 01. Disagreeable duties 02. Working overtime 04. Unfamiliar duties 07. Dealing with crisis situations 11. Increased duties 16. Making on-the-spot decisions 23. Frequent interruptions 25. Excessive paperwork 26. Meeting deadlines 31. Time pressure

31

Nontarget/Dual Loading Items 12. Periods of inactivity 15. Insufficient personnel 22. Noisy work areas 24. Boring/demanding activity changes 27. Insufficient personal time

41

39 37 47

39

Im

Fr

30

Se

31

31

52 46 41 55 53 54 49 45 78 76

41 56 45 33 51 56 50 75 76

56 34 52 56 65 54 39 31 74 76

30

36

32

33

.

Factor Correlations

41

36

39

Note: Decimal points and item loadings of less than .25 are omitted from the table. Im = Impact (i.e., Frequency × Severity) scores; Fr = Frequency ratings; Se = Severity ratings.

RT44367_C012.indd 225

7/13/05 9:31:53 AM

226

V. HODAPP ET AL.

Table 12.9 Intercorrelations between JSS-G subscales for the Impact Frequency and Severity Scores (Study 4) Impact

Frequency

Severity

Scale

1

2

3

4

5

6

Impact 1. Organizational Support 2. Job Pressure

— .47

.47 —

94 .46

.48 .86

.78 .43

.35 .83

Frequency 3. Organizational Support 4. Job Pressure

.94 .48

.46 .86

— .52

.52 —

.59 .39

.27 .53

Severity 5. Organizational Support 6. Job Pressure

.78 .35

.34 .83

.59 .27

.39 .53

— .51

.51 —

Table 12.10 Intercorrelations between the Four Organizational Climate Scales and the Impact Scores for the Two JSS-G Subscales (Study 4) Job Stress

Climate

JS/BO

Scale

1

2

3

4

5

Job Stressa 1. Organizational Support 2. Job Pressure

— .47

.47 —

–.60 –.27

–.54 –.27

–.42 –.30

.38 .51

Organizational Climate 3. Supervisory Support 4 Staff Cohesion

–.60 –.54

.27 –.27

— .58

.58 —

.49 .50

–.32 –.38

Job Sat./Burnout 5. Job Satisfaction 6. Burnout

–.42 .38

–.30 .51

.49 –.32

.50 –.38

— –.38

–.38 —

a

6

Impact scores.

standard stressor item did not evoke the same amount of stress for all people, an important requirement for it to be used as a valid anchor item. DISCUSSION AND CONCLUSIONS The construction, development, and validation of a German adaptation of the Job Stress Survey (Spielberger, 1994; Spielberger & Vagg, 1999) was described in this chapter. In the translation of the English-language JSS

RT44367_C012.indd 226

7/13/05 9:31:54 AM

THE GERMAN ADAPTATION OF THE JOB STRESS SURVEY

227

Figure 12.3. Distribution of the Severity ratings for the JSS-stressor “Assignment of disagreeable duties” on the nine-point scale used in Study 4.

items, psychological, linguistic, and cultural factors were taken into account. The German adaptation of the Job Stress Survey was developed in a series of empirical studies (Hodapp et al., 1996). Based on the administration experience and results from preceding studies, the earlier versions of the JSS-G were further modified in each subsequent study. During this process, it became evident that the German translation of the anchor item “Assignment of disagreeable duties,” with a preassigned average Severity rating in the original English JSS, consistently received controversial ratings. The German translation of the original standard stressor clearly did not evoke the same amount of stress for all people, a requirement that was considered important for its use as a valid anchor item. The principal-axis factor analyses in all four validation studies that were carried out with employees from different occupational groups, namely, teachers (Studies 1 and 2), medical and nonmedical staff associated with a psychiatric hospital (Study 3), and bank employees (Study 4), confirmed the construct validity of the German adaptation of the Job Stress Survey. In all four studies, Job Pressure and Lack of Organizational Support consistently emerged as separate factors, confirming the cross-cultural validity of these two components of occupational stress. In the four studies, almost all of the items from the original 10-item Englishlanguage JSS subscales had dominant salient loadings on the appropriate Lack of Organizational Support and Job Pressure factors. However, a few items loaded on different factors in the four studies. In particular, Items 9 (Tasks not in job description), 10 (Inadequate equipment), 19 (Inadequate salary), and

RT44367_C012.indd 227

7/13/05 9:31:54 AM

228

V. HODAPP ET AL.

28 (Covering work for others) loaded on Job Pressure in Studies 1 and 3, and on Lack of Organizational Support in Study 4. Most of these inconsistencies can probably be explained by differences in the organizational and supervisory structure of bureaucratically regulated (“red tape”) Austrian state-owned schools (Studies 1 and 2) and hospitals (Study 3) staffed predominantly by tenured personnel, as compared to commercially oriented private enterprises such as banks (Study 4). Additional research is needed to identify a core set of JSS-G items that have invariant loading patterns across a wide range of different occupational settings. Evidence of the convergent and discriminant validity of the JSS-G Organizational Support and Job Pressure subscales was demonstrated in the pattern of correlations between these subscales and the four scales of the Organizational Climate Questionnaire that were reported in Studies 2 and 4. In both studies, the OCQ Supervisory Support and Staff Cohesion scales correlated more highly with Organizational Support than with Job Pressure, and the OCQ Burnout scale correlated more highly with Job Pressure than with Organizational Support. In addition, the correlations between the two JSS-G subscales and an overall stress rating (“How much and how often do you feel stressed in your job?”) in Studies 1 and 2 further confirmed the convergent and discriminant validity of the JSS-G. In both studies, the correlation of the overall stress rating was consistently higher with Job Pressure than with Organizational Support. Finally, in a quasiexperimental intervention study (Tanzer, Pestemer, Maier, & Kogler, 1998) that was not reported in this chapter, 90 Austrian customs officers who received a three-day stress reduction program were compared with the Austrian bank employees described in Study 4 who served as an untreated control group (cf. Maier et al., 1997). The preliminary results of this intervention study gave further support to the validity of the German adaptation of the Job Stress Survey.

REFERENCES Barth, A. R. (1992). Burnout bei Lehrern [Burnout in teachers]. Göttingen, Germany: Hogrefe. Bernstein, I. H. (1988). Applied multivariate analysis. New York: Springer. Boyle, G. J., Borg, M. L., Falzon, J. M., & Baglioni, A. J., Jr. (1995). A structural model of the dimensions of teacher stress. British Journal of Educational Psychology, 65, pp. 49–67. Bracken, B. A., & Barona, A. (1991). State of the art procedures for translating, validating and using psychoeducational tests in cross-cultural assessment. School Psychology International, 12, pp. 119–132. Brislin, R. W. (1980). Translation and content analysis of oral and written material. In H. C. Triandis & J. W. Berry (Eds.), Handbook of cross-cultural psychology (Vol. 1, pp. 389–444). Boston: Allyn & Bacon.

RT44367_C012.indd 228

7/13/05 9:31:55 AM

THE GERMAN ADAPTATION OF THE JOB STRESS SURVEY

229

Brislin, R. W. (1986). The wording and translation of research instruments. In W. J. Lonner & J. W. Berry (Eds.), Field methods in cross-cultural research (pp. 137–164). Newbury Park, CA: Sage. Brissie, J. S., Hoover-Dempsey, K. V., & Bassler, O. C. (1988). Individual, situational contributors to teacher burnout. Journal of Education Research, 82, pp. 106–112. Burke, R. J. (Guest Ed.). (1992). Occupational stress, psychological burnout, and anxiety. Anxiety, Stress, and Coping [special issue], 5, no. 1. Cooper, C. L., & Payne, R. (Eds.). (1988). Causes, coping, and consequences of stress at work. Chichester, UK: Wiley. Cooper, G. L. (1995). Life at the chalkface: Identifying and measuring teacher stress. British Journal of Educational Psychology, 65, pp. 69–71. Friedman, I. A. (1991). High- and low-burnout schools: School culture aspects of teacher burnout. Journal of Educational Research, 84, pp. 325–333. Geisinger, K. F. (1994). Cross-cultural normative assessment: Translation and adaptation issues influencing the normative interpretation of assessment instruments. Psychological Assessment, 6, pp. 304–312. Grier, K. S. (1981). A comparison of job stress in law enforcement and teaching. Doctoral dissertation, University of South Florida, Tampa. Dissertation Abstracts International, 43, 870B. Hambleton, R. K. (1994). Guidelines for adapting educational and psychological tests: A progress report. European Journal of Psychological Assessment (Bulletin of the International Test Commission), 10, pp. 229–244. Harris, K. R., Halpin, G., & Halpin, G. (1985). Teacher characteristics and stress. Journal of Educational Research, 78, pp. 346–350. Hart, P. M. (1994). Teacher quality of work life: Integrating work experiences, psychological distress, and morale. Journal of Occupational and Organizational Psychology, 67, pp. 109–132. Hart, P. M., Wearing, A. J., & Conn, M. (1995). Conventional wisdom is a poor predictor of the relationship between discipline policy, student misbehaviour, and teacher stress. British Journal of Educational Psychology, 65, pp. 26–48. Hinton, J. W., & Hodapp, V. (1997). The psychophysiology of work stress. Introductory comments on an invited set of papers from members of the European research group MATCH. Journal of Psychophysiology, 11, pp. 195–199. Hodapp, V., Tanzer, N. K., & Spielberger, C. D. (1996). Der Berufs-StreßFragebogen [The German adaptation of the Job Stress Survey]. Unpublished test, University of Düsseldorf, Germany, and University of Graz, Austria. Jung, B. (1997). Evaluierung des Job Stress Survey [Validation of the Job Stress Survey]. Unpublished master´s thesis, University of Vienna, Austria. Karasek, R. A., & Theorell, T. (1990). Healthy work. New York: Basic Books. Keita, G. P., & Sauter, S. L. (Eds.). (1992). Work and well-being: An agenda for the 1990s. Washington, DC: American Psychological Association. Litt, M. D., & Turk, D. C. (1985). Sources of stress and dissatisfaction in experienced high school teachers. Journal of Educational Research, 78, pp. 178–185.

RT44367_C012.indd 229

7/13/05 9:31:55 AM

230

V. HODAPP ET AL.

Maier, E. R., Pestemer, I. A., & Tanzer, N. K. (1997, July). Validation of a German adaptation of the Job Stress Survey in a sample of Austrian bank employees. Paper presented at the 18th International Conference of the Stress and Anxiety Research Society, Düsseldorf, Germany. Matteson, M. T., & Ivancevich, J. M. (Eds.). (1987). Controlling work stress: Effective human resource and management strategies. San Francisco: Jossey-Bass. Norusis, M. J. (1993). SPSS for Windows: Professional Statistics Release 6.0. Chicago: SPSS. Oswald, F., Pfeifer, B., Ritter-Berlach, G., & Tanzer, N. K. (1989). Schulklima. Die wirkungen der persönlichen beziehungen in der schule [School climates: The impact of interpersonal relationships]. Vienna, Austria: Wiener Universitätsverlag. Pithers, R. T., & Fogarty, G. J. (1995). Occupational stress among vocational teachers. British Journal of Educational Psychology, 65, pp. 3–14. Poortinga, Y. H. (1989). Equivalence of cross-cultural data: An overview of basic issues. International Journal of Psychology, 24, pp. 737–756. Schnall, P. L., & Landsbergis, P. A. (1994). Job strain and cardiovascular disease. Annual Review of Public Health, 15, pp. 381–411. Spielberger, C. D. (1994). Preliminary manual for the Job Stress Survey (JSS). Tampa: University of South Florida, Psychology Department. Spielberger, C. D., & Reheiser, E. C. (1994). Job stress in university, corporate, and military personnel. International Journal of Stress Management, 1, pp. 19–31. Spielberger, C. D., & Vagg, P. R. (1999). Professional manual for the Job Stress Survey (JSS). Odessa, FL: Psychological Resources Assessment. Tanzer, N. K., Pestemer, I. A., Maier, E. R., & Kogler, A. (1998, August). Assessment of job stress in Austrian bank employees and customs officers. In C. D. Spielberger (Chair), Stress, Emotions, and Health. Invited Symposium conducted at the 24th International Congress of Applied Psychology, San Francisco. Tanzer, N. K., & Schwetz, H. (1996a). Was streßt: Ergebnisse zweier umfragen an steirischen volks- und hauptschullehrern [Job Stressors: Results of two surveys on Austrian primary and vocational school teachers]. Unser Weg, 51, pp. 31–35. Tanzer, N. K., & Schwetz, H. (1996b). Arbeits- und beziehungskulturen an schulen und wohlbefinden von lehrern: Eine empirische untersuchung an 47 steirischen und burgenländischen hauptschulen [Organizational and social climates, and emotional well-being in teachers: A survey in 47 Austrian vocational schools]. Unser Weg, 51, pp. 12–30. Tellenback, S., Brenner, S.-O., & Löfgren, H. (1983). Teacher stress: Exploratory model building. Journal of Occupational Psychology, 56, pp. 19–33. Van de Vijver, F. J. R., & Hambleton, R. K. (1996). Translating tests: Some practical guidelines. European Psychologist, 1, pp. 89–99. Van de Vijver, F. J. R., & Leung, K. (1997a). Methods and data analysis of comparative research. In J. W. Berry, Y. H. Poortinga, & J. Pandey (Eds.), Handbook of cross-cultural psychology (2nd ed., Vol. 1, pp. 257–300). Chicago: Allyn & Bacon.

RT44367_C012.indd 230

7/13/05 9:31:55 AM

THE GERMAN ADAPTATION OF THE JOB STRESS SURVEY

231

Van de Vijver, F. J. R., & Leung, K. (1997b). Methods and data analysis for cross-cultural research. Newbury Park, CA: Sage. Van de Vijver, F. J. R., & Tanzer, N. K. (1997). Bias and equivalence in crosscultural assessment: An overview. European Review of Applied Psychology, 47, 263–279. Weyer, G. (1982). Self-report measures in research on job-related stress. In H. W. Krohne & L. Laux (Eds.), Achievement, stress, and anxiety (pp. 333–344). Washington, DC: Hemisphere. AUTHORS’ NOTES AND ACKNOWLEDGMENTS Correspondence concerning this chapter should be sent to Volker Hodapp, Institute of Psychology, University of Frankfurt, P.O. 11 19 32, D-60054 Frankfurt, Germany (E-mail: [email protected]). For their assistance in data collection, we wish to thank Herbert Schwetz and Oliver Heimrath (Studies 1 and 2) and Barbara Jung (Study 3). We would also like to thank Catherine Q. E. Sim for her valuable comments on an earlier version of this chapter. Study 1 was designed and partly conducted during a one-year ErwinSchrödinger Scholarship awarded to the second author by the Austrian Science Foundation (FWF Project #J0805-SOZ) that supported his work at the University of South Florida Center for Research in Behavioral Medicine and Health Psychology, directed by Charles D. Spielberger.

RT44367_C012.indd 231

7/13/05 9:31:56 AM

RT44367_C012.indd 232

7/13/05 9:31:56 AM

13 Spanish Adaptation of the Job Stress Survey: All Dressed Up and Ready to Go to Work1 María Jesús González and José M. Prieto Complutense University, Madrid, Spain Abstract The construction and psychometric properties of the Spanish adaptation of the Job Stress Survey (JSS-SV) are described in this chapter, along with detailed information on how the Spanish JSS was developed and validated. The JSS-SV was administered to 450 health, clerical, and maintenance workers, along with a General Reactions Survey. The internal consistency reliability of the JSS-SV for the 294 respondents was demonstrated in alpha coefficients of .93 for the Severity Scale, .92 for the Frequency Scale, and .91 for the overall Job Stress Index. Factor analyses confirmed the existence of two factors that were similar to the Lack of Organizational Support and Job Pressure factors of the American JSS. A third factor, which was unique to the JSS-SV, consisted of items related to Job Annoyance. The meaningfulness of the three JSS-SV factors and the concurrent validity of the scales derived from them was reflected in a strong correspondence of the individual items in the Spanish and English forms. Highly significant correlations of the JSS-SV with the General Reactions Survey were also found, demonstrating the congruence between both instruments.

SPANISH ADAPTATION OF THE JOB STRESS SURVEY: ALL DRESSED UP AND READY TO GO TO WORK This chapter reports the work carried out in adapting the Job Stress Survey (JSS) for the Spanish population. The general aim was to construct a reliable and valid Spanish version of the JSS (JSS-SV), which was achieved by pursuing the attainment of three specific objectives: (a) to develop a comprehensible, flexible, and applicable measure of occupational stress; (b) to calculate the reliability of the JSS-SV by determining the internal consistency of this measure; and (c) to evaluate the construct and concurrent validity of the JSS-SV. The first aim required the collaboration of a group of experts. The second involved the calculation of alpha (α) coefficients to assess internal consistency. The third required a factor analysis of the JSS-SV items to determine its construct equivalence validity with the American JSS. Correlations with Correspondence concerning this chapter should be addressed to the authors, c/o Faculty of Psychology, Department of Individual Differences and Work Psychology, Somosaguas, 28223 Madrid, Spain; Fax 34-91-3943189; E-mail [email protected] and [email protected]. 1

233

RT44367_C013.indd 233

7/13/05 9:32:44 AM

M.J. GONZ´ALEZ AND J. PRIETO

234

other measures were also computed to estimate the concurrent validity of the Spanish JSS as compared with the English form developed in the U.S. with the American form. CONSTRUCTION AND DEVELOPMENT OF THE JSS-SV The original JSS was constructed to evaluate occupational stress affecting employees in a variety of workplaces (Spielberger, 1994; Spielberger & Vagg, 1999). This was achieved by obtaining ratings of the perceived severity of 30 specific job situations for generating stress in workers who were exposed to them and estimates of how often each of these stressor situations was experienced during the preceding six months (Spielberger & Reheiser, 1994a, 1994b; Turnage & Spielberger, 1991). The overall stress level for each source of occupational stress was determined by combining the ratings of perceived severity and frequency of occurrence. In the original American version of the JSS (Spielberger, 1994; Spielberger & Vagg, 1999), workers were instructed to evaluate, on a scale from 1 to 9, the level of stress produced by the circumstances or situations described by each item. The sum of these ratings defined the worker’s score on the JSS Severity (S) scale. In the second part of the JSS, workers were asked to indicate, on a 10-point scale with ratings of 0 to 9+ (9 or more days), how often during the past six months they personally experienced each of the 30 JSS stressor situations. These scores defined the Frequency (F) Scale. A third part was added to the Spanish JSS in which the respondents were asked (a) to describe additional situations or circumstances in their organization that gave rise to tensions during the last six months, (b) to specify what level of tension was experienced, and (c) to indicate how often these situations or circumstances had occurred. In constructing the items for the Spanish JSS, a group of experts2 made individual and joint studies of the American JSS. Each item was analyzed with two aims in mind: (a) to determine, in the experts’ opinion, whether the content of the item was transferable from the American to the Spanish version; and (b) to ensure that the items in the Spanish adaptation were clear, brief, flexible, and relevant to occupational stress for Spanish workers. In order to achieve the first of these aims, which pertained to content validity, the experts introduced four new items, which were numbered from 31 to 34. The first additional item, which referred to long working days, was included because it was considered desirable to elaborate on the meaning of the original JSS item that inquired about working overtime. In Spain, “overtime” leads to tension because (a) the worker is given little notice, (b) social problems or family conflicts may occur, and (c) compensatory elements such as payment for additional work may not be forthcoming. Excessively long work days may also contribute to problems similar to “overtime,” such as (d) increased tiredness, (e) whether or not the overtime is foreseeable, and (f) active or passive acceptance. At times of economic crises, payments for overtime or for long 2

Our gratitude is expressed to Carmen Cid and Manuel Gancedo.

RT44367_C013.indd 234

7/13/05 9:32:44 AM

SPANISH ADAPTATION OF THE JOB STRESS SURVEY

235

working days are often curtailed, which contributes to these stressors being regarded as occupational hazards by many workers. The second item that was added to the JSS-SV refers to unsuitable temperatures in the workplace. In the American JSS, noise was included as a stressor, but not temperature, which has some of the same negative effects (Poulton, 1978). In the first European survey of the work environment (1991–1992), the European Foundation for the Improvement of Living and Working Conditions (EFILWC; 1992) reported that 43% of the workers who were surveyed reported exposure to heat or cold for at least 25% of their working time, 36% for at least half of this time, and 23% for all of their working time. In a 1993 survey of working conditions carried out by the Instituto Nacional de Seguridad e Higiene en el Trabajo (INSHT; 1995), the results indicated that 25% of the respondents considered the temperature in their work environment to be unsuitable. These findings appear to confirm a strong link between job stress and unsuitable temperature. The third and fourth items added to the JSS-SV refer to role ambiguity and role conflict. These items pertain to lack of clarity in marking the limits of job competence and working conditions that give rise to contradictory demands. The experts who discussed the two new items agreed on their being added to the JSS-SV because problems relating to role ambiguity and conflict are frequently encountered in Spanish work settings. The importance of these problems for workers’ health has been pointed out by a number of investigators since first described by Kahn, Wolfe, Quinn, Snoek, and Rosenthal (1964). Previous reviews by Fisher and Gittelson (1983), Jackson and Schuler (1985), Peiró (1992), and Van Sell, Brief, and Schuler (1981) have also emphasized the importance of role ambiguity and conflict. The final pool of 34 JSS-SV items was back-translated into English and reviewed by Professor Spielberger, the author of the American JSS (Spielberger, 1994; Spielberger & Vagg, 1999). This preliminary version of the JSS-SV was given to a pilot sample of Spanish subjects with different levels of education, ranging from basic studies to university level. The responses of the pilot sample demonstrated that the instructions were clear and comprehensible, and that the JSS-SV items were adequate. The JSS-SV has essentially the same Severity and Frequency Scales as the American version and also includes a third part devoted to asking open questions, which was of great interest because it identified (a) sources of stress that were relatively unique to a particular organization, (b) the importance that workers attached to these stressors, and (c) how often each stressor was experienced. Such information added important organizational conditions to the database of situations evaluated by the JSS-SV. Thus, the JSS-SV provides a detailed picture of organizational reality and facilitates identifying problems that can be addressed by specific preventive programs. Choice of Complementary Tests To establish the concurrent validity of the JSS-SV, other indicators of stress symptoms and psychological equilibrium were needed. In applying the

RT44367_C013.indd 235

7/13/05 9:32:45 AM

236

M.J. GONZ´ALEZ AND J. PRIETO

old principle of “the shorter the better,” the choice was made to administer (a) 12 items from the Norfolk (1989) questionnaire, which referred to the individual worker’s perception of the state of stress; (b) 12 items referring to wellbeing, as tested in work environments by Banks et al. (1980); and (c) Goldberg and Blackwell’s (1980) General Reactions Survey (GRS), which will be described in a later section on evaluating the concurrent validity of the JSS-SV. Selecting and Obtaining the Sample The following steps were taken prior to administering the JSS-SV to a representative sample of Spanish workers: (a) profiles were established for the subjects to whom the survey would be given; (b) the makeup of teams to be trained to interview and hand out the surveys was determined; (c) key persons were contacted in the different organizations or firms where the tests were to be administered, who could act as presenters with their workmates and inform subjects of the goals of the study; and (d) general rules for responding to the JSS-SV items were developed. Given the interest of subsequently working with employees on tasks with health and safety risks in selected professional environments, two types of worker profiles were established. Profile A applied to groups of health workers, which included different types of professionals in contact with the sick who worked in state or private hospitals, or in public health care. The Profile A workers were doctors, nurses, psychologists, social workers, and therapists. Profile B applied to workers carrying out administrative, office (clerical), or maintenance work for the Civil Service or for private-sector firms. Once the profiles had been decided, contact was made with Profile A people in five public and two private hospitals and in four health centers. Contact with Profile B individuals was made in 5 public agencies and 10 private firms. The decision to collect data with a selected sample was a consequence of the difficulties involved in obtaining a truly random sample. The directors of human resources and personnel managers in the different organizations whom we contacted were not enthusiastic about recruiting random groups. Application of the Tests All subjects collaborated voluntarily in responding to the test instruments. They were informed of the goals of the study and given general rules for responding to the survey. It was also emphasized that anonymity was guaranteed. Both the JSS-SV and the General Reactions Survey were self-administered as stapled copies, with no identification code. The subjects deposited the completed surveys in a conveniently located container after responding to them. Description of the Sample In initial interviews with each subject, they were asked for their cooperation, given the questionnaires, and urged to respond to all items. Emphasis was given to rating the level of stress caused by each JSS-SV stressor condition

RT44367_C013.indd 236

7/13/05 9:32:45 AM

SPANISH ADAPTATION OF THE JOB STRESS SURVEY

237

by comparing it with the tension normally produced by “taking on unpleasant matters.” A total of 450 questionnaires3 were handed out. Of the 225 surveys given to the health workers, 120 full replies were collected. Of the 225 surveys handed out to administrative, clerical, and maintenance workers, 174 were completed and returned. Of the total sample of 294 subjects who participated in the study, 58% were women and 42% were men; 57% were married, 38% unmarried, and 5% were widowers, divorced, or separated. Approximately 40% of the participants were health workers and 60% were administrative or clerical employees. In Spain, females comprise the majority of the workers in both health and office settings (INSHT, 1995). The study participants were asked to indicate their age, using one of the four age ranges suggested by Cooper, Sloan, and Williams (1988). No participant was younger than 21 years, 54% were between 21 and 36 years, 42% were between 36 and 55, and 4% were over 55. In regard to level of education, 7% of the sample had completed only primary studies, 15% had received job-training education, 24% had studied in high schools, 34% had medium grade certificates, and 20% had higher education qualifications. In terms of length of service with the firm by whom they were employed, the mean was 12.5 years, the median was 11 years, and the mode was 6 years. Mean length of time on their present job was 7.5 years, the median was 5 years, and the mode was 1 year. RESULTS AND DISCUSSION Psychometric Properties of the JSS-SV All statistical analyses were made using the BMDP and SPSS statistical packets. The statistics for the JSS-SV Severity (S), Frequency (F), and Index (JS-I) scales were calculated after first checking the normality of the distributions and determining their suitability for multivariate analysis. The Severity score, which indicates the total amount of stress perceived by the respondent and the subject’s vulnerability to job stress, is calculated by means of the formula S = Σ S1–34. The Frequency score, which indicates how often the subjects are exposed to the various sources of job stress, is calculated by F = Σ F1–34. The Job Stress Index indicates the global level of stress-producing pressure the subjects are experiencing, which depends on both the vulnerability perceived by the subjects when facing stress-producing circumstances and how often they have been exposed to them. The JSS-SV Index scores were obtained by the same procedure used in computing JSS Index scores with the American version. The Index score for each item is determined by multiplying the value of Scale S by that of Scale F and summing the Index scores for all 34 items. The JSS-SV Stress Index is then calculated as follows: JS–I = Σ (Si × Fi /10). 3 Our gratitude is expressed to Carmen Cid and Manuel Gancedo as well as students who assisted in the data collection in the research team headed by Dr. González.

RT44367_C013.indd 237

7/13/05 9:32:45 AM

M.J. GONZ´ALEZ AND J. PRIETO

238

Table 13.1 Means, Standard Deviations, and Alpha Coefficients for the JSS-SV Severity and Frequency Scales and Index Scores, and Alpha Coefficients for U.S. Samples Scale

Means

Standard Deviations

JSS-SV Alphas

USA Sample Alpha Range

Severity Frequency JS-Index

178.04 124.02 66.83

41.53 61.05 40.51

.93 .92 .91

.89–.93 .89–.92 .87–.92

The means and standard deviations for the JSS-SV Severity and Frequency Scales, and the JS-I Index, are reported in Table 13.1. Cronbach’s alpha coefficients were computed to evaluate the internal consistency of the JSS-SV scales and Index scores for the total sample and also separately for men and women. The alpha values obtained for the three JSS-SV scales for the total sample, reported in Table 13.1, were .93 for Scale S, .92 for Scale F, and .91 for the JS-I. The range of alpha values for a number of different U.S. samples, which are also reported in Table 13.1, serve to compare the alpha values obtained for the Spanish sample. It can be noted that the JSS-SV α values were at the upper limit of the range of alphas obtained for heterogeneous samples of U.S. subjects. Given the magnitude of the α values obtained for the Spanish sample (.91 to .93), which were comparable to those for the American JSS, the internal consistency of the three JSS-SV scales was considered to have been adequately established. However, when comparing the alpha values obtained for the JSS-SV with the corresponding range of values obtained for the American samples, it must be kept in mind that the Spanish sample was comprised of different percentages of men and women, and that the U.S. samples included subsamples of managers, professionals, and clerical and maintenance workers. Validity of the JSS-SV The classic statement that “A test is valid if it achieves what it sets out to do” (Yela, 1987) was the objective goal of our work. To evaluate this goal, we examined the JSS-SV for three types of validity: content, construct, and criterion. The first aim was to establish content validity, using the procedures recommended by Crocker and Algina (1986) for defining the area to which the test applies. The most suitable items were selected by a group of experts in the field of occupational stress, who established a structured framework for identifying items corresponding to those in the original JSS that were considered to be equally appropriate for Spanish workers. Although this aspect of test construction had been followed in selecting items for the original JSS (Spielberger, 1994; Spielberger & Vagg, 1999), it was repeated to evaluate and ensure the content validity of the items for the Spanish version. García-Cueto (1993) divided the content validity of test items into two subcategories: face validity and sampling validity. It was decided that a group of

RT44367_C013.indd 238

7/13/05 9:32:46 AM

SPANISH ADAPTATION OF THE JOB STRESS SURVEY

239

experts should study the content validity of the individual JSS-SV items by analyzing both the appropriateness of the content of each item and the degree to which the pool of items covered the broad area of job stress. The procedure in this part of the process was based entirely on the subjective judgments of the experts; no congruence index via statistical procedures was obtained. By the agreement and consensus among the experts, the decision was made to include four new items in the Spanish JSS-SV, which were numbered 31 to 34. The content validity of the 34 JSS-SV items was considered by the experts to be sufficiently demonstrated. Construct validity was established through factor analysis, using the principal components method with varimax rotation. Communality was estimated by multiple squared correlations, which were computed using the BMDP statistical package. The number of meaningful factors was established by first taking into account the magnitude of the eigenvalues and by following the steps of Cattell’s (1978) scree test. This procedure was followed in the factor analyses of the JSS-SV Severity and Frequency items and the JS-I item index scores. Factor Structure of the Spanish Job Stress Survey The factor structure for the JSS-SV S, F, and JS-I scales are shown in Table 13.2. In the trifactor solution, which was considered the most suitable for all three scales, the order of listing the first, second, and third factors was based on the percentage of variance associated with each factor, which was similar for all three JSS-SV scales. In Table 13.2, the first three columns report the factor loadings for the Severity Scale, which assesses the total amount of perceived stress for the 33 stressor situations described by the JSS-SV items, as compared to item 1 (Disagreeable duties), the standard stressor that was assigned a median value of 5. The similarity in the pattern of factor loadings for the JSS-SV Severity and Frequency Scales and the Job Stress Index provides strong evidence that essentially the same three groups of stress-generating situations were assessed by all three JSS-SV measures. The content of each JSS-SV item with a dominant loading on one of the three factors is reported in Table 13.3, in which the items are grouped together on the basis of the consistency of their loadings on the three JSS-SV factors. The items are listed as a function of the descending magnitude of their Severity loadings on each factor. The content of the items with dominant loadings on the first factor highlights Lack of organizational support, especially from supervisors. The stressor items with dominant loadings on Factor 2 are related to the requirements of the job, which lead to feelings of Job pressure resulting from increased responsibility, making critical decisions, assignment of new duties, and working overtime. The content of the items with dominant loadings on Factor 3 describe work-related situations that are irksome or unpleasant, giving rise to Job annoyance because of stressful working conditions that involve doing the work of unmotivated colleagues, inadequate salary, uncomfortable temperatures, noisy work environment, poor equipment, and lack of personal time.

RT44367_C013.indd 239

7/13/05 9:32:46 AM

M.J. GONZ´ALEZ AND J. PRIETO

240

Table 13.2 Factor Structure for the Spanish JSS-SV Severity and Frequency Scales and Job Stress Index Severity Scale Items

F-1

6 13 8 30 21 5 14 18 17 29 33 34

74 73 64 63 62 57 56 54 53 49 43 44

11 16 4 7 2 24 26 31 25 1

40

**

F-2

43

44 43 74 70 65 60 50 50 48 42 36 **

28 32 19 10 27 22 23 9 3 12 15 20

47 36

VP

39%

F-3

31%

51

Frequency Scale F-1 79 81 68 48 74 48 55 64 43 48 52 48

45

53 **

F-2

F-3

37 43 39

Job Stress Index F-1 80 75 61 47 69 47 55 61 41 40 52 54

65 65 62 40 73 52 58 59 57 42

65 59 57 55 55 50 47 41

35

36

37

43

35

36

43

30%

42%

36%

45 51 54

F-2

38 40 40 62 60 50 37 64 53 57 55 51 47

42 74 43 52

36

75 46

F-3

38 49

52

44 75 58 54 72 42

42 22%

40%

35%

25%

Note: Only loadings above .35 are reported. Each loading has been multiplied by 100. **Item 1, with a fixed value of 5, served as the standard against which the perceived severity of the other items were compared and was not rated for the Severity Scale.

RT44367_C013.indd 240

7/13/05 9:32:47 AM

241

SPANISH ADAPTATION OF THE JOB STRESS SURVEY

Table 13.3 Content of the JSS-SV Items with Dominant Significant Loadings on the Three Factors Item

Factor 1

6

Lack of support by supervisor Difficulties with supervisor Lack of recognition Conflicts between departments Poor supervision Unreliable fellow workers Negative attitudes Lack of participation Personal insults Unmotivated partners Dubious self-competence Contradictory demands

13 8 30 21 5 14 18 17 29 33 34 11 16 4 7 2 24 26 31 25 1 28 32 19 10 27 22 23 9

Factor 2

Factor 3

Increased responsibility On-the-spot decisions New duties Crisis situations Working overtime Frequent changes Deadlines Lengthy working days Backlog of paperwork Disagreeable duties Doing the work of another Inadequate heating Inadequate salary Poor equipment Lack of personal time Noisy environment Regular interruptions Tasks not in job description

Do the three sets of stress-generating situations identified as factors by the participants in this study represent the organizational reality of the work setting? As can be noted in Tables 13.2 and 13.3, the items grouped together for each factor were essentially the same for the JSS-SV Severity

RT44367_C013.indd 241

7/13/05 9:32:48 AM

242

M.J. GONZ´ALEZ AND J. PRIETO

and Frequency Scales and Job Stress Index. The global trifactor structure of the JSS-SV is also reasonably consistent with general theory in the field of occupational stress. Further analyses of the content of the individual items with dominant loadings on the three factors suggested the following interpretations of each factor: • Factor 1 can be understood as a relational factor, reflecting lack of support at work that was analogous to the Lack of organizational support factor in the American JSS. This factor explained 38.8% of the variance for Scale S, 41.6% for Scale F, and 39.6% for the JS-I. The strongest loadings on Factor 1 were found for items describing stressful human relationships that were organizationally determined. Somewhat lower loadings were found for task-stressors that interacted with human relationships. Among the former were (a) difficulties with supervisors, (b) lack of recognition at work, and (c) conflicts with other departments. The latter included (d) feeling offended by others, (e) poorly motivated colleagues, and (f) situations that imposed contrary work demands. • Factor 2 is clearly similar to the Job pressure factor in the American JSS. The items with strong loadings on this factor included (a) increased responsibilities, (b) making important on-the-spot decisions, (c) coping with new duties; (d) meeting deadlines, and (e) working overtime. Long working days, dealing with excessive paperwork, and being assigned disagreeable duties also contributed to experiencing job pressures stemming from the nature of the work. The Job pressure factor explained 30.8% of the variance on Scale S, 35.85% on Scale F, and 34.85% for the JS-I. • Factor 3 highlights the job annoyance and discomfort caused by a wide range of working conditions related to the physical atmosphere in work settings, such as an uncomfortable temperature, high noise level, and poor equipment. The psychosocial atmosphere of the work setting (e.g., frequent interruptions, lack of resources, inadequate salary, and being required to perform tasks not in the job description) also contributed to the Job Annoyance factor. Such conditions generally cause increased irritability and anger, which reduce the employee’s ability to adapt to social interactions at work. The respective variances explained by Factor 3 for the JSS-SV Severity and Frequency Scales and JS-I were 30.4%, 22.6%, and 25.6%. Concurrent Validity of the Spanish JSS In order to evaluate the concurrent validity of the JSS-SV, a General Reactions Survey was constructed and administered to study participants. The GRS was made up of two different components that assessed psychological well-being and stress-related symptoms. The 12 items comprising the psychological well-being component were drawn from Banks et al.’s (1980) 30-item General Health Questionnaire. The stress symptoms component was based on 12 of the 13 items formulated by Norfolk (1989). The psychological well-being items were placed at the beginning of the GRS, followed by the stress-related symptom items.

RT44367_C013.indd 242

7/13/05 9:32:48 AM

243

SPANISH ADAPTATION OF THE JOB STRESS SURVEY

Table 13.4 Factor Structure for the G and N Components of the General Reactions Surveys for the Spanish Study Participants I/G Survey Items

Factor 1

5 2 6 9 10 1 7 11 12 4 8 3

74 74 73 71 61 46 49 40 50

VP

61%

I/N Survey

Factor 2

Items

Factor 1 86 73 64 61 57 54 43

51 79 65 60

1 2 5 10 13 3 4 9 7 8 11 6

39%

60%

Factor 2

43 64 57 56 45 43 40%

Note: Only loadings above .35 are reported. Each loading is multiplied by 100.

Each GRS item was question based, with four possible responses for each question; respondents underlined the answer they considered most appropriate for each item. For items posing negative consequences, the possible responses were “Not at all,” “No more than usual,” “More than usual,” and “Much more than usual.” For items that described positive outcomes, the response alternatives were “More than usual,” “Same as usual,” “Less than usual,” and “Much less than usual.” Responses to the GRS were available for 213 study participants. For the statistical analyses of responses to the GRS, a global index was obtained for each part of this survey. The name I/G was given to the psychological well-being component, which was based on the Goldberg method (Goldberg & Blackwell, 1970); the acronym for the stress symptoms component derived from the Norfolk (1989) questionnaire was I/N. The procedure for calculating these indices was similar for the two measures: I/G = Σ G1–12 and I/N = Σ N1–12. The reliability of the I/G and I/N measures was evaluated in the same manner as for the JSS-SV. Cronbach internal consistency alpha coefficients obtained for the two GRS Index scores were α = .81 for the I/G psychological well-being Index and α = .80 for the I/N measure of stressrelated symptoms. The content and construct validity of the I/G and I/N measures were evaluated by factor analyses, using the same procedure that was used in computing the factor analyses for the JSS-SV. The two-factor solutions reported in Table 13.4 for the GRS I/G and I/N measures were considered to provide the most consistent and meaningful solutions for both measures. In the analysis of the I/G severity items, the following items had the highest loadings on the first

RT44367_C013.indd 243

7/13/05 9:32:49 AM

244

M.J. GONZ´ALEZ AND J. PRIETO

factor: (a) feeling overwhelmed and tense, (b) loss of sleep through worries, (c) perception that one's own difficulties cannot be overcome, (d) feeling unhappy or depressed, (e) loss of self-confidence, (f) inability to concentrate, (g) difficulty enjoying day-to-day activities; and (h) harboring doubts about one's own worth. Given the content of these items, this factor was considered to provide a global measure of negative affect, which included both anxiety and depression. The items with the strongest loadings on the second factor obtained in the analysis of the I/G survey were (a) feeling unable to make decisions, (b) avoiding facing up to one's own problems, (c) not managing to play a useful role in life, and (d) considering oneself unhappy. The content of these items suggested a substratum of depressive symptoms. Although items related to feeling unhappy or depressed and the inability to enjoy day-to-day activities had salient loadings on both factors, the loadings on the first factor were stronger, suggesting that this was a negative affect factor, defined by items that reflected both anxiety and depression. The results of the factor analyses of the I/N survey items are also reported in Table 13.4. The first I/N factor was defined by hyperactivation items that described inability to relax, difficulty in "switching off," difficulty falling asleep and interruptions during sleep, physical tiredness and mental fatigue, and symptoms characteristic of vegetative overactivation. The second factor identified for the I/N survey was defined by items indicating hyperactive social behavior. Items with dominant salient loadings on this factor described impatience, always being in a hurry, frustration resulting from being held up or having to wait, feeling frustrated or annoyed by other persons who were slow in completing a task, and using tranquilizers, tobacco, and alcohol. An item with content referring to irritability loaded equally on both I/N factors, which suggested that underlying anger and anxiety contributed to the hyperactive feelings and behaviors described by most of the I/N items. After establishing that the internal consistency of the I/G and I/N measures were within an acceptable range and determining the factor structure for these questionnaires, correlations between the JSS-SV Severity, Frequency, and Index scores and the GRS measures were computed to evaluate concurrent validity. However, before commenting on the obtained matrix of correlations, which is reported in Table 13.5, the criteria employed in selecting the items for the GRS and JSS-SV should be noted. The GRS I/G and I/N scales were comprised of items that described global indicators of psychological wellbeing and stress-related symptoms. The JSS-SV items assess the perceived severity of sources of occupational stress from the perspective of respondents and how often they were exposed to them. Because the JSS Index takes into account both the perceived severity of the work-related stressors described by the JSS items and how often each stressor was experienced, scores on this measure can be considered as a global indicator of the level of stress experienced by a particular individual. Most theories of occupational stress assume that stress reactions are influenced by three major characteristics of the stressful situations that are

RT44367_C013.indd 244

7/13/05 9:32:49 AM

245

SPANISH ADAPTATION OF THE JOB STRESS SURVEY

Table 13.5 Correlations of the JSS-SV Index, Severity, and Frequency Scores with the GRS I/G and I/N Total Scores GRS Indexes and JSS-SV Scales GRS I/G GRS I/N JSS-SV S JSS-SV F

Spanish Job Stress Survey JS-Index

Severity

Frequency

24 31 63 93

20 24

20 26

51

Note: The correlation coefficients have been multiplied by 100, and all were significant at p < .001.

encountered in the workplace: (a) the duration of exposure to a stressor, (b) the acute or chronic nature of the exposure, and (c) how often the exposure occurs. Stress reactions may also be manifested in a variety of behaviors or symptoms and have a wide range of effects on health and feelings of psychological wellbeing. Moreover, a stress reaction (understood as a response) is not simply the direct result of exposure to a particular source of stress, but results from a person's interaction with the stressor, which may be influenced by individual differences in personality and other work-related variables that function as modulators of stress. In the present context, it is not possible to analyze what these modulating variables might be and how they influence stress reactions. However, it is important to note that there may be a large number of potential modulating variables and that their often powerful direct and indirect effects can markedly influence (a) the perception of stress-producing situations, (b) specific stress responses and their manifestations, and (c) the effects that sustained reactions to particular sources of stress may have on the person. The feedback produced by such effects may also influence the interactions between the person and a particular source of stress, and the environment in which a person works. The effects of occupational stress on emotions, behavior, and the work environment do not necessarily occur at the same time. Most stress researchers differentiate between (a) symptoms of stress as an immediate or, in some cases, a delayed reaction to contact with the stress source, and (b) the effects of stress as a transactional process that continues to occur over an extended period of time (Folkman & Lazarus, 1984). The importance of the time difference between the initial contact with a stress source and its effects on behavior and health was strongly emphasized by Holmes and Rahe (1967) and is now generally accepted. The instruments used in the present study evaluated five different aspects of stress as a transactional process, that is, the immediate impact of the stressor, the perceived severity and frequency of occurrence of stress reactions, and the effects of the stress process on the psychological well-being and the symptoms manifested by a worker. The correlations obtained between the JSS-SV Severity, Frequency, and Index scores and the GRS I/G and I/N Survey measures reported

RT44367_C013.indd 245

7/13/05 9:32:50 AM

246

M.J. GONZ´ALEZ AND J. PRIETO

in Table 13.5 are relatively small, but all of these correlations were positive and highly significant (p < .001). Although from a strictly psychometric viewpoint, the values of these correlations do not reach the level required for a clear and solid demonstration of the concurrent validity of the JSS-SV, the findings are generally consistent with those of other studies of occupational stress in which the correlations obtained among variables that assessed reactions to stressful conditions were low to moderate. The correlation of .51 between the JSS-SV Severity and Frequency scores reported in Table 13.5 indicated a moderate relation between the perceived severity and frequency of occurrence of the JSS-SV stressors. As may also be noted in Table 13.5, the JSS-SV Index correlated .93 with the Frequency scores and only .63 with the Severity scores, suggesting that how often the sources of stress assessed by the JSS-SV were encountered had a substantially greater impact than the perceived severity of the stressors on the total amount of stress experienced by the respondents. The correlation of .55 between the GRS I/G and I/N measures, which is not reported in Table 13.5, indicated that positive feelings of psychological well-being were characteristic of persons who experienced fewer stress-related symptoms, as would be expected. Comparison of the Factor Structures of the Spanish JSS-SV and the American JSS The items with dominant salient loadings on the three factors for the Spanish JSS-SV are compared in Table 13.6 with the items that defined the Job pressure and Lack of organizational support factors in the American JSS (Spielberger, 1994; Spielberger & Vagg, 1999). To facilitate these comparisons, the items from the American JSS Frequency Scale that are numbered from 31 to 60 on the JSS test form were renumbered 1 to 30 so that the corresponding items in the Frequency Scales of the American and Spanish forms of the JSS would have the same numbers in both surveys. It should be noted that items numbered 31 to 34 were unique to the Spanish JSS-SV. As may be noted in Table 13.6, 8 of the 10 Severity and Frequency items that had dominant salient loadings on the first factor in the Spanish JSS-SV were essentially the same as those that were included in the Lack of organizational support scale of the American JSS. The items with dominant loadings on the second factor obtained in the analysis of the Spanish JSS-SV were also quite similar to those in the Job Pressure scale of the American JSS. Seven of the 10 Job Pressure items in the American JSS had dominant salient loadings on the second factor of the Spanish JSS-SV. Of the three remaining American job pressure items, Item 23 loaded on the Job annoyance factor, and Items 9 and 27 were among the six items that did not significantly load on any of the three factors in the JSS-SV. A more detailed analysis of the findings reported in Table 13.6 may help to clarify the similarities and differences between the JSS-SV and the American JSS. The factor loadings of five Lack of support items did not coincide in the Spanish and American forms:

RT44367_C013.indd 246

7/13/05 9:32:50 AM

SPANISH ADAPTATION OF THE JOB STRESS SURVEY

247

Table 13.6 Comparison of the JSS Factor Structure for the U.S. and Spanish Samples JSS Factors Lack of Support Severity Items Frequency Items Job Pressure Severity Items Frequency Items Job Annoyance Severity Items Frequency Items

Sample

Items on U.S. and Spanish JSS Scales

U.S. Spain

3, 5, 6, 8, 13, 14, 18, 19, 21, 29 5, 6, 8, 13, 14, 17, 18, 21, 29, 30, 33, 34

U.S. Spain

3, 5, 6, 8, 13, 14, 18, 19, 21, 29 5, 6, 8, 13, 14, 17, 18, 21, 29, 30, 33, 34

U.S. Spain

2, 4, 7, 9, 11, 16, 23, 25, 26, 27 2, 4, 7, 11, 16, 24, 25, 26, 31, 34

U.S. Spain

2, 4, 7, 9, 11, 16, 23, 25, 26, 27 2, 4, 7, 11, 16, 24, 25, 26, 31, 34

Spain Spain

10, 19, 22, 23, 28, 29, 32 10, 19, 22, 23, 28, 29, 32

• Item 3 (Opportunities for advancement) and Item 19 (Inadequate salary), which were included in the Lack of support subscale in the American JSS, did not have dominant loadings on Factor 1 for the Spanish JSS-SV. However, Item 19 loaded on the Job annoyance factor in the Spanish sample. In Spain, most payroll issues are determined by generally applicable collective agreements that affect every employee’s wages for a given job category; employee-company agreements rarely exist. Pay adjustments are also generally bargained in relation to the cost-of-living index, without taking productivity into consideration. • Item 17 (Personal insults) and Item 30 (Conflicts between departments), which were not included in the Lack of organizational support scale of the American JSS, loaded on Factor 1 for the Spanish JSS-SV. • Item 29 (Unmotivated partners) was included in the Lack of organizational support scales for both the American JSS and Spanish JSS-SV, and also loaded on the Job annoyance factor of the Spanish JSS-SV. In Spain, the presence of unmotivated colleagues indicates a lack of mutual assistance and is a source of perturbation in the job. The following four Job Pressure items did not coincide in the Job Pressure scales of the Spanish JSS-SV and the American JSS. Item 9 (Tasks not in job description), Item 23 (Interruptions), and Item 27 (Very short breaks), which were included in the American JSS Job Pressure scale, did not load on Factor 2, the Job Pressure factor of the Spanish JSS-SV. However, Item 23 had a dominant salient loading on the Spanish JSS-SV Job Annoyance factor. Item 24

RT44367_C013.indd 247

7/13/05 9:32:51 AM

248

M.J. GONZ´ALEZ AND J. PRIETO

(Frequent changes in task requirements), which loaded on the Job pressure factor of the Spanish JSS-SV, was not included in either the Job Pressure or Lack of organizational support scale of the American JSS. In the factor analyses of the Spanish JSS-SV items, Item 10 (Poor equipment), Item 22 (Noisy environment), and Item 28 (Doing the work of another), which were not included in either the Lack of organizational support or Job pressure scale of the American JSS, had unique dominant salient loadings on the JSS-SV Job annoyance factor. Item 19 (Inadequate salary), which was included in the American JSS Lack of organizational support scale, also had a dominant salient loading on Factor 3, the Spanish JSS-SV Job annoyance factor. Item 29 (Unmotivated partners), which was included in the American JSS Lack of organizational support scale, loaded on both the Lack of organizational support and Job annoyance factors of the Spanish JSS-SV. All four of the new experimental items that were added to the Spanish JSS-SV loaded on one of the three factors that were identified for this measure. Item 31 (Length of working days) and Item 34 (Contradictory demands) both loaded on the JSS-SV Job pressure factor, and Item 34 also loaded on the Lack of organizational support factor. Item 32 (Inadequate heating) loaded on the JSS-SV Job annoyance factor, and Item 33 (Dubious competence) loaded on the Lack of organizational support factor. The results of the factor analyses of the JSS-SV items provide strong evidence that Job annoyance, in addition to Lack of organizational support and Job pressure, should be considered a major category of occupational stress. In summary, of the 30 American JSS items that were translated and adapted for the Spanish JSS-SV, 10 of these items loaded on the Spanish JSS-SV Lack of organizational support factor, 8 on the Job pressure factor, and 6 on the Job annoyance factor. Thus, 24 of the 30 American JSS items contributed to the three-factor structure that was identified for the Spanish JSS-SV, including four items that were not in the American JSS Lack of organizational support or Job pressure scales. Of the 34 items comprising the Spanish JSS-SV, 27 items had dominant salient loadings on one of the three factors, and all four of the new experimental items that were added to the JSS-SV had salient loadings on one or more of the three JSS-SV factors. CONCLUSIONS AND SUGGESTIONS FOR FURTHER RESEARCH The findings in the analyses of the data obtained with the JSS-SV indicated that this instrument was reliable and had good construct validity when compared with the factor structure of the American JSS. The concurrent validity findings were also significant, but the magnitude of the correlations obtained between the JSS-SV and the reference criteria for psychological well-being and stress-related symptoms was only moderate. It should be noted, however, that the quality of the concurrent validity findings might be due to limitations in the criterion measures that were developed for this study. Clearly, more detailed research on the validity of the JSS-SV is needed. However, based on the research reported in this chapter, it seems reasonable to conclude that:

RT44367_C013.indd 248

7/13/05 9:32:51 AM

SPANISH ADAPTATION OF THE JOB STRESS SURVEY

249

1. The JSS-SV is a versatile instrument that can be used in both general and individual applications. The instructions and the design of the JSS-SV facilitate responding to each individual item and enhance the potential for using this measure in a self-administered organizational system. 2. Using separate measures in the JSS-SV to assess the perceived severity of potentially stress-generating situations (Severity scale), the frequency that these situations are experienced (Frequency scale), and the joint effects of both (JS-I scale) facilitates obtaining a global view of organizational reality and the tension produced by work-related stressors in individuals and groups of employees. 3. The JSS facilitates discovering and evaluating stress in the workplace and obtaining knowledge on which to base interventions that enable stress to be reduced, both at general organizational and individual levels. When providing psychological counseling, or introducing cognitive-behavioral techniques for tackling stress, it is essential to (a) discover the degree of perceived severity that employees attribute to work situations that generate stress, (b) compare these personal perceptions with those of an appropriate reference group, and (c) evaluate perceptions of specific sources of job stress for different occupations and organizational groups. 4. The relatively brief time required to administer the JSS, as compared to other job stress evaluation procedures, along with the interest aroused among respondents to the survey items, makes the JSS a very useful tool for evaluating personal stress levels. The JSS can also be used in the selection process and in follow-up and control evaluations of persons in the organizational context. 5. The JSS-SV has identified Job Annoyance as a third factor present in stressful situations and transactions within the workplace. It should be noted, however, that Job annoyance is not a brand-new factor, because it may be traced back to research within the context of occupational health psychology. For example, the nexus between noise and worker annoyance was highlighted by McLean and Tamopolsky (1977). Within a hospital milieu, Haas (1964) indicated that when group members perceive job roles divergently, irritation and annoyance follow. Injustice experienced by medical staff generated not only moral judgments but also emotions dominated by annoyance rather than anger, as noticed by Wosinska (1987). Bycio (1992) examined the relationship between individual employee performance and absenteeism in a meta-analytical study of 46 articles and identified supervisors’ annoyance as an important mechanism in ratings of frequently absent employees as poor performers. Finally, two suggestions based on research with the JSS-SV should be noted that are regarded as having important implications for future research: 1. Retaining the third part of the survey that was used with the JSS-SV, in which the subjects were presented with open questions about their own and their organizational sources of stress, facilitates analyzing the importance of

RT44367_C013.indd 249

7/13/05 9:32:51 AM

250

M.J. GONZ´ALEZ AND J. PRIETO

particular stress sources and the frequency of contact with them. Because the types of open-ended questions included in Part 3 of the JSS-SV are difficult to deal with and are generally regarded as not very effective, such questions are often included only in the experimental phase of test construction and are subsequently eliminated. However, in the context of a specific organization, the responses of employees to open-ended questions may provide information on general organizational problems that are not addressed by the JSS. They may also help to identify specific problems that are unique to particular departments or services within the organization. 2. The possibility of providing a frequency response category to identify stressors that are encountered daily, or almost daily, should be considered. The frequency response of 9+ days of exposure over a six-month period may not be sufficient to identify “daily” exposure to a specific stress source. A detailed analysis of the differences produced by daily exposure, as compared to frequent exposure, would provide information to facilitate (a) establishing categories of stressors, as proposed by Elliot and Eisdorfer (1982), (b) identifying normal and abnormal reactions to different sources of work-related stress (Dohrenwend, Dohrenwend, Dodson & Shrout, 1984), (c) a detailed analysis of the effects of chronic daily stress sources (Dohrenwend & Shrout, 1985), and (d) analysis of the emotional responses to daily sources of job stress as an indicator of a worker’s vulnerability (Folkman & Lazarus, 1984). Information on the time pattern for specific sources of job stress in regard to their intermittent or continuous action would facilitate efforts to reduce or eliminate daily producers of job stress. However, daily stressors are sometimes considered trivial and are accepted as habitual and unmodifiable annoyances in organizational life. Efforts directed toward increasing well-being by reducing stress in the workplace have been emphasized in the publications of Keita and Sauter (1992) and Quick, Murphy, and Hurrell (1992). If primary interventions in the workplace carried out by an organization are effective in reducing the perceived severity and frequency of occurrence of sources of job stress, this would demonstrate that the intervention was successful (MacLennan, 1992). Worker-oriented interventions that provide assistance for employees exposed to occupational stress can be implemented by developing training programs within an organizational framework (Martin, 1992). In both types of interventions, the data collected with the JSS-SV should prove useful in designing and evaluating these interventions.

REFERENCES American Psychological Association (APA). (1974). Standards for educational and psychological tests. Washington, DC: APA.

RT44367_C013.indd 250

7/13/05 9:32:52 AM

SPANISH ADAPTATION OF THE JOB STRESS SURVEY

251

Banks, M. H., Clegg, C. W., Jackson, P. R., Kemp, N. J., Stafford, E. M., & Wall, T. D. (1980). The use of the General Health Questionnaire as an indicator of mental health in occupational studies. Journal of Occupational Psychology, 53, pp. 187–194. Bycio, P. (1992). Job performance and absenteeism: A review and meta-analysis. Human Relations, 45, pp. 193–220. Cattell, R. B. (1978). The scientific use of factor analysis. New York: Plenum Press. Cooper, C. L., Sloan, S. J., & Williams, S. (1988). Occupational stress indicator management guide. Windsor, UK: Nelson NFER. Crocker, L. J., & Algina, J. (1986). Introduction to classical and modern test theory. New York: Holt, Rinehart, and Winston. Dohrenwend, B. S., Dohrenwend, B. P., Dodson, M., & Shrout, P. (1984). Symptoms, hassles, social supports, and life events: Problems of confounded measures. Journal of Abnormal Psychology, 93, pp. 222–230. Dohrenwend, B. S., & Shrout, P. (1985). Hassles in the conceptualization and measurement of life stress variables. American Psychologist, 40, pp. 780–785. Elliot, G. P., & Eisdorfer, C. (Eds.) (1982). Stress and human health. New York: Springer-Verlag. European Foundation for the Improvement of Living and Working Conditions (EFILWC) (1992). First European survey on the work environment. Luxembourg: Office of Official Publications of the European Communities. Fisher, C. D., & Gittelson, R. (1983). A meta-analysis of correlates of role conflict and ambiguity. Journal of Applied Psychology, 68, pp. 320–333. Folkman, S., & Lazarus, R. S. (1984). 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 Psychology, 48, pp. 150–170. García-Cueto, E. (1993). Introducción a la psicometría. Madrid: S. XXI de España. Goldberg, D. P., & Blackwell, B. (1970). Psychiatric illness in general practice: A detailed study using new method of case identification. British Medical Journal, 2, pp. 439–443. Haas, J. E. (1964). Role conception and group consensus: A study of disharmony in hospital work groups. Columbus: Ohio State University Press. Holmes, T. H., & Rahe, R. J. (1967). The social readjustment rating scale: A cross-cultural study of Western Europeans and Americans. Journal of Psychosomatic Research, 14, pp. 391–400. Instituto Nacional de Seguridad e Higiene en el Trabajo (INSHT) [National Institute of Safety and Hygiene at Work]. (1995). Encuesta nacional de condiciones de trabajo, 1993. Salud y Trabajo, pp. 107–108, I–XXXII. Jackson, S. E., & Schuler, R. S. (1985). A meta-analysis and conceptual critique of research on role ambiguity and role conflict in work settings. Organizational Behavior and Human Decision Processes, 36, pp. 16–78. Kahn, R. L., Wolfe, D. M., Quinn, R. P., Snoek, J. D., & Rosenthal, R. A. (1964). Organizational stress: Studies in role conflict and ambiguity. New York: John Wiley & Sons.

RT44367_C013.indd 251

7/13/05 9:32:52 AM

252

M.J. GONZ´ALEZ AND J. PRIETO

Keita, G. P., & Sauter, S. L. (Eds.). (1992). Work and well-being: An agenda for the 1990s. Washington, DC: American Psychological Association. MacLennan, B. W. (1992). Stress reduction: An organizational alternative to individual stress management. In J. C. Quick, L. R. Murphy, & J. J. Hurrell, Jr. (Eds.), Stress and well-being at work: Assessments and interventions for occupational mental health. Washington, DC: APA. Martin. E. V. (1992). Designing stress training. In J. C. Quick, L. R. Murphy, & J. J. Hurrell, Jr. (Eds.), Stress and well-being at work: Assessments and interventions for occupational mental health. Washington, DC: APA. McLean, E. K., & Tamopolsky, A. (1977). Noise, discomfort, and mental health: A review of the sociomedical implications of disturbance by noise. Psychological Medicine, 7, pp. 19–62. Norfolk, D. (1989). El estrés del ejecutivo. Bilbao, Spain: Deusto. Peiró, J. M. (1992). Desencadenantes del estrés laboral. Madrid: Eudema. Poulton, E. C. (1978). Blue collar stressors. In C. L. Cooper, & R. Payne (Eds.), Stress at work. Chichester, UK: John Wiley & Sons. Quick, J. C., Murphy, L. R., & Hurrell, J. J., Jr. (Eds.). (1992). Stress and well-being at work: Assessments and interventions for occupational mental health. Washington, DC: APA. Spielberger, C. D. (1994). Preliminary Manual for the Job Stress Survey (JSS). Tampa: University of South Florida. Spielberger, C. D., & Reheiser E. C. (1994a). Job stress in university, corporate, and military personnel. International Journal of Stress Management, 1, pp. 19–31. Spielberger, C. D., & Reheiser, E. C. (1994b). The Job Stress Survey: Measuring gender differences in occupational stress. Journal of Social Behavior and Personality, 9, pp. 199–218. Spielberger, C. D., & Vagg, P. R. (1999). Professional manual for the Job Stress Survey (JSS). Odessa, FL: Psychological Resources Assessment. Turnage, J. J., & Spielberger, C. D. (1991). Job stress in managers, professionals, and clerical workers. Work & Stress, 5, pp. 165–176. Van Sell, M., Brief, A. P., & Schuler, R. S. (1981). Role conflict and role ambiguity: Integration of the literature and directions for future research. Human Relations, 34, pp. 43–71. Wosinska, W. (1987) Stress in the hospital setting and injustice experienced by medical staff. Polish Psychological Bulletin, 18, pp. 177–186. Yela, M. (1987). Apuntes de psicología matemática (Vol. 2). Madrid: Facultad de Psicología, U.C.M.

RT44367_C013.indd 252

7/13/05 9:32:52 AM

14 Occupational Level and Gender Differences in Job Stress in Venezuelan Employees Lya Feldman, Grace Chacón-Puignau, Nuri Bagés, and María Gabriela Pérez Universidad Simón Bolívar, Caracas, Venezuela Abstract The goals of this study were to develop and validate a Spanish adaptation of Spielberger’s Job Stress Survey (JSS) and to evaluate organizational level and gender differences for Venezuelan employees. The JSS was administered, along with a Social Support Survey (SSS) and an Automatic Thinking (AT) questionnaire, to 168 employees of an insurance company and a major university. The participants belonged to four occupational groups: managers, clerical personnel, technicians, and blue-collar workers. The Spanish JSS had good reliability and construct validity. Stressors related to work overload and coping were perceived as more severe and occurring more frequently. In general, men rated the 30 JSS occupational stressors as more severe than women and reported experiencing them more frequently. Men in blue-collar and technical jobs reported Lack of support and Environmental problems at work as more stressful than women at the same organizational levels. Blue-collar men also reported more Job overload than blue-collar women, whereas white-collar women reported greater Job overload than their male counterparts.

OCCUPATIONAL LEVEL AND GENDER DIFFERENCES IN JOB STRESS IN VENEZUELAN EMPLOYEES Stress constitutes an extremely important psychosocial factor that has been studied extensively in the past 25 years, especially in relation to the work environment. It is now well established that occupational stress is strongly related to productivity and job satisfaction, absenteeism at work, physical and mental health problems, and feelings of well-being (Karasek & Theorell, 1990; Matteson & Ivancevich, 1982; Pereyra, 2004; Sutherland & Cooper, 1990). Organizational researchers have also consistently found that stress at work contributes to negative job attitudes and problematic behaviors (Caplan, Cobb, French, Harrison, & Pinneau, 1975; Cooper, 2002; Margolis, Kroes, & Quinn; 1974), and nonwork factors such as family strain (Barling, 1990; Bartolome & Evans, 1979; Burke, Weir, & Dowors, 1980; Chan, Lai, Ko, & Boey, 2000). 253

RT44367_C014.indd 253

7/13/05 9:50:08 AM

254

L. FELDMAN ET AL.

According to Jick and Payne (1980), many researchers in the medical sciences have observed that stress-related problems are reaching epidemic proportions. For example, a study by Northwestern National Life (1991, 1992) found that the proportion of workers who reported “feeling highly stressed” more than doubled from 1985 to 1990, and those who reported “having multiple stressrelated illnesses” increased from 13% to 25%. Moreover, 69% of the 600 workers surveyed in this study indicated that their productivity was reduced because of high stress levels, and one in three reported that “job stress is the single greatest stress in their lives” (Spielberger & Reheiser, 1994, p. 1). Although research findings have firmly established the importance of occupational stress, the definition of job stress is still unclear (Briner, 1993; Spielberger & Reheiser, 1994). It is essential to clearly define the concept of occupational stress and to develop reliable and valid measuring instruments that can be used across investigations and cultures to make meaningful comparisons. The Job Stress Survey (JSS), which is now widely used in the United States (Spielberger, 1994; Spielberger & Reheiser, 1994; Spielberger & Vagg, 1999), focuses on assessing generic sources of occupational stress that are encountered by men and women in a wide variety of work settings. The occupational stressors described in the content of the JSS items appear to be adequate for evaluating job stress in various cultures; a Spanish version of this inventory was recently translated and adapted by González and Prieto (Chapter 13, this volume). The number of women entering the labor force has increased substantially in recent years, which has stimulated research on gender and employment in the workplace. Important differences in the reactions of men and women to various sources of occupational stress have also been reported, creating a new and challenging topic for investigation. The work environment differs substantially for men and women, who often differ in their responses to job demands (Feldman, Payne, & Viva, 2002; Karasek & Theorell, 1990; La Croix & Haynes, 1987; Miller, 1980). There are very few published studies on work stress in Venezuela. In two investigations of high-level managers (Bagés, Feldman, & Chacón, 1995; Feldman, Bagés, & Chacón, 1991), it was found that occupational stress influenced their personality characteristics and that a very high proportion of these managers (85%) were Type-A individuals. The results of these studies also indicated that high-level married managers reported less stress than lowerlevel single managers and that managers who exhibited positive thinking had more effective coping styles than those with negative thinking. The main purpose of the research that will be reported in this chapter was to evaluate and validate the utility of a Spanish adaptation of the Job Stress Survey (Spielberger, 1994; Spielberger & Vagg, 1999) for use in a Latin American culture, as was done in Spain by González and Preito; Chapter 13, this volume). A second goal was to examine the effects of differences in gender and organizational level on occupational stress as measured by the JSS and to investigate the relation between the JSS scale and subscale scores with measures of social support and automatic thinking.

RT44367_C014.indd 254

7/13/05 9:50:08 AM

OCCUPATIONAL LEVEL AND GENDER DIFFERENCES

255

METHOD Participants A total of 231 insurance company and university employees were recruited to participate in this study. Of this number, 168 employees (105 women, 63 men) returned the questionnaires (72.72% response rate). The age of the participants varied between 18 and 54 years, with a mean age of 29.39 (SD = 7.96); the men (mean = 30.20, SD = 8.05) and women (mean = 29.08, SD = 7.89) did not differ significantly in age (t = 0.88; NS). The distribution of participants by organizational level was managers and directors (22.1%), clerical (38.7%), technical (27.6%), and blue-collar (11.7%). The number of participants at each organizational level was unequally distributed by gender (chi-sq = 22.49; gl = 3; p < .001). There were more men (61.1%) than women in the management positions. In the clerical and technical staff positions, there were more women than men (84.1% and 57.8%, respectively). Men comprised 47.4% of the blue-collar group, as compared to 52.6% for the blue-collar women. There were also gender differences in the mean age of the personnel at higher and lower organizational levels. As expected, the managerial positions were occupied by older men and women (38.1 and 38.5 years, respectively), as compared to the younger participants at the lower organizational levels. Instruments Job Stress Survey. The JSS assesses the perceived severity and frequency of occurrence of job-related stress in a variety of occupations. The individual JSS items describe 30 potentially stressful work-related situations. Each stressor event is first rated in terms of its perceived severity (intensity) on a nine-point scale, from low to high stress. The frequency of experiencing each stressor event during the past six months is then rated on a 10-point scale, from 0 to 9+ days. Developed in the mid-1980s by Spielberger (1994) and his colleagues (Spielberger & Vagg, 1999; Spielberger, Vagg, & Wasala, 2003), the JSS has been translated and adapted in a number of languages. Social Support Scale (SSS). The SSS was derived from the Social Support Questionnaire (Dunn, Puttallaz, Sheppard, & Lindstrom, 1987), which was translated and adapted in Spanish by Bagés et al. (1995). This measure consists of 18 items, each rated on a four-point scale. The Cronbach alpha internal consistency reliability coefficient was .74 for the 18-item SSS. Factor analysis of the SSS items identified six dimensions of social support. Three of these dimensions were related to sources of support (fathers and siblings, partners and colleagues at work). Two SSS dimensions were concerned with general support and institutional and group support (including religious). The items comprising the sixth factor referred to the respondents’ perception of the availability of social support. Automatic Thinking (AT) Questionnaire. The AT Questionnaire was derived from a measure developed by Ingram & Wisnicki (1988) to evaluate thinking

RT44367_C014.indd 255

7/13/05 9:50:08 AM

256

L. FELDMAN ET AL.

styles. This questionnaire is comprised of six positive and five negative thoughtrelated items, with two subscales for which the split-half Cronbach alpha reliability coefficients were .72 and .75. The presence of distinctive positive and negative thought processes was confirmed by a principal-components factor analysis. Procedure Permission was obtained from an insurance company to administer the questionnaires to their employees. The package of three questionnaires was personally delivered to the insurance company participants by their supervisors, who asked them to respond to the questionnaires and return them within 48 hours. If the questionnaires were not returned to the supervisors, they were collected by a research assistant. To ensure adequate comprehension of the instructions and item content by the blue-collar workers, the questionnaires were verbally administered to them in an interview. University personnel received the questionnaires through their internal mail system and were instructed to return them within two days, via internal mail. RESULTS Reliability and Factor Analyses of the Spanish JSS Severity and Frequency Scales Table 14.1 reports the Cronbach alpha and Guttman split-half internal consistency reliability coefficients for the JSS Severity and Frequency scales. As may be noted, both scales were found to be highly reliable. The results of the principal-components factor analysis of the Frequency items, with varimax rotation, in which five factors were identified, was considered to provide the most meaningful solution. Item loadings for each of the five factors are reported in Appendix A, Table 14.A1. The first factor in this analysis, which explained 36.3% of the variance, was related to Lack of support by supervisors and organizational policies. The second factor, which was related to Recognition and Advancement, explained 7.4% of the variance. Factors 3, 4, and 5 were related to overload, inadequate personnel and equipment, and coping and crisis management, explaining, respectively, 5.8%, 5.2%, and 4.6% of the variance. In the factor analysis of the JSS Severity items, the four-factor principal components solution with varimax rotation was most consistent with the eigenvalues criterion and was also considered to be theoretically most meaningful. The dominant loadings of the Severity items on each of the four factors are reported in Appendix A, Table 14.A2. Similar to the findings for the Frequency items, the first large factor, which explained 30.6% of the variance, was related to Lack of support and Job dissatisfaction. The second factor, explaining 9.4% of the variance, was related to Overload and Competition for advancement. The third and fourth factors, which explained 6.5% and 4.9% of the variance respectively, were related to Crisis management and Environmental conditions.

RT44367_C014.indd 256

7/13/05 9:50:09 AM

257

OCCUPATIONAL LEVEL AND GENDER DIFFERENCES Table 14.1 Reliability of the JSS Severity and Frequency Scales JSS Scales

Number of Items

Cronbach Alpha

Guttman Split-half

Severity

30

.879

.859

Frequency

30

.936

.911

Table 14.2 Descriptive Statistics for the Spanish JSS Scales Variable

Mean

Standard Deviation

Possible Range

Observed Range

Number of situations Frequency Severity Total Job Stress Survey

18.8 84.0 156.5 18.2

10.8 54.6 39.7 11.6

0–30 0–270 0–270 0–81

0–30 0–270 47–250 0–55

Descriptive statistics for the Spanish JSS are reported in Table 14.2. The respondents reported that they had experienced approximately two thirds of the 30 stressor situations (mean = 18.8) during the past six months. The stressor items reported as occurring most frequently were Excessive paperwork (89.1%), Meeting deadlines (87.0%), Making critical on-the-spot decisions (86.4%), Assignment of new or unfamiliar duties (85.4%), Performing tasks not in job description (84.2%), Dealing with crisis situations (84.2%), and Inadequate salary (84.0%). The mean perceived severity levels for the 30 JSS stressor events were also relatively high. The following individual stressors were rated as most severe Inadequate salary (mean = 6.59), Problems with supervisor (mean = 5.61), Lack of opportunity for advancement (mean = 5.91), Dealing with crisis situations (mean = 5.79), Lack of recognition for good work (mean = 5.85), Inadequate supervision (mean = 5.84), and Frequent interruptions (mean = 5.83). Differences in Gender and Organizational Level To evaluate the effects of gender and organizational level for the Spanish JSS Frequency items, the number of stressor events experienced by the study participants during the past six months and the frequency of their occurrence were evaluated in 2 × 2 analyses of variance (ANOVAs). The main effect for gender was significant for both the number of stressor events experienced by the study participants [F(1,155) = 9.44; p < .003] and the frequency of occurrence of these stressors [F(1,153) = 5.95; p < .016]. On both measures, men had higher scores than women. The mean number of stressor situations reported as experienced by men and women were 21.85 and 16.65, respectively; the mean Frequency scores were 100.08 for men and 75.75 for women.

RT44367_C014.indd 257

7/13/05 9:50:09 AM

258

L. FELDMAN ET AL.

Table 14.3 Descriptive Statistics for the Frequency Scores of the Five Spanish JSS Stress Factors Factors

Mean

Standard Deviation

Possible Range

Observed Range

Lack of Support Recognition/Advancement Overload Inadequate Personnel/Equipment Coping with Work

19.12 20.25 34.72 10.17 14.84

15.66 13.86 19.55 7.21 9.03

0–72 0–54 0–81 0–27 0–36

0–72 0–54 0–81 0–27 0–36

Table 14.4 Descriptive Statistics for the Spanish JSS Stress Severity Factors

Factors

Mean

Standard Deviation

Possible Range

Observed Range

Lack of Support Overload and Competition Coping Management Environmental Conditions

77.18 25.86 36.36 21.70

21.06 7.47 10.04 6.89

0–126 0–45 0–63 0–36

25–126 5–45 10–62 4–36

The gender by organizational level interaction effect was not significant for either the total number of stressor events or their frequency, and no significant differences were found for organizational level. However, it is interesting to note that 85% of the men rated eight of the JSS situations as highly stressful, whereas only one stressor, Meeting deadlines, was rated as very stressful by 85% of the women. Most of the situations that men rated as highly stressful were related to job overload. The means, SDs, and range of scores for subscales based on the five JSS Frequency factors are reported in Table 14.3. Stressful situations related to Overload were reported as occurring most frequently. The mean Frequency scores for Lack of support and Recognition/Advancement factors were similar and were next highest in frequency. Inadequate personnel and equipment was rated as least frequent. When the Frequency scores for each of the five factors were evaluated as a function of gender and organizational level, the gender difference for the Lack of support factor was the only significant finding; men had higher frequency scores (mean = 20.84) than women (mean = 17.21). No differences in organizational level were found for any of the five factors, and there were no significant interactions. The means, SDs, and range of scores for the four Spanish JSS Stress Severity factors are reported in Table 14.4. Lack of support was rated as a much more severe source of stress than any of the other factors. In the 2 by 2 ANOVA for the Severity ratings, the gender by organizational level interaction was statistically significant [F(3,155) = 3.23; p < .02]. Separate analyses for each gender

RT44367_C014.indd 258

7/13/05 9:50:10 AM

259

OCCUPATIONAL LEVEL AND GENDER DIFFERENCES Table 14.5 Mean Job Stress Severity Scores as a Function of Gender and Organizational Level Organizational Level Males Females

Blue-Collar

Clerical

Technical

White-Collar

168.6* 128.3a

148.4 154.3

171.1* 148.8

157.3 175.5b

*Statistically significant differences (p < .05) between males and females. a and b indicate significant differences at the p < .05 level.

indicated that the difference in the mean Severity scores for women at higher and lower organizational levels approached significance [F(3,99) = 2.39; p < .07], whereas this difference was not significant for men [F(3,59) = 1.80; NS]. In separate analyses of the organizational level subgroups, the white-collar women reported a significantly higher level (p < .05) of Stress Severity (mean = 175.5) than the blue-collar women (mean = 128.3). The mean Severity scores as a function of gender and organizational level are reported in Table 14.5. Blue-collar men reported significantly higher overall levels of job stress severity than their female counterparts, which resulted from their higher scores on the Lack of support, Overload, and Environmental factors. Male technical workers also had significantly higher overall mean Severity scores than women, due primarily to their higher scores on the Lack of support and Environmental factors. The white-collar women reported higher Stress Severity scores than blue-collar women, especially on the Overload factor. No significant gender differences were found for the clerical personnel. The effects of gender and organizational level on JSS Severity scores were further evaluated in 2 × 2 ANOVAs that were computed for each of the four Spanish JSS Severity factors. Significant interaction effects were found for Lack of support [F(3,120) = 3.73; p < .01], Overload [F(3,137) = 5.68; p < .001], and Environmental conditions [F(3,143) = 4.92; p < .01]. No differences in gender or organizational level were found for the Coping factor. In separate one-way ANOVAs for each gender, significant organizational level differences were found on Overload for both women [F(3,84) = 5.24; p < .01] and men [F(3,53) = 2.83; p < .05]. Clerical and white-collar women had higher mean Overload Severity scores than blue-collar women; blue-collar men had higher Overload scores than white-collar men. Differences in organizational level for women were also found for the Lack of support [F(3,70) = 3.71; p < .01] and Environmental conditions factors [F(3,91) = 4.14; p < .07]. Clerical and white-collar women had substantially higher Severity scores than blue-collar women on these factors. In summary, the results of the analyses of the ratings of the Severity factors indicated that the clerical and white-collar women perceived Lack of support, Overload, and Environmental conditions as considerably more stressful than the blue-collar women, as can be noted in Table 14.6. For men, the only statistically significant difference indicated that the blue-collar men perceived Overload as more stressful than the white-collar men. The blue-collar and technical men

RT44367_C014.indd 259

7/13/05 9:50:10 AM

260

L. FELDMAN ET AL.

Table 14.6 Mean Job Stress Perceived Severity Scores for Male and Female Blue-Collar, Clerical, Technical, and White-Collar Workers Factors

Blue-Collar

Clerical

Technical

White-Collar

Lack of Support Males Females

75.75* 54.66a

69.66 82.22b

83.38* 71.21

77.31 86.3b

Overload and Competition Males Females

30.00*b 17.14a

26.40 27.97b

26.63 23.12

22.70*a 27.07b

Environment Conditions Males Females

26.12* 15.33a

19.77 22.06b

24.42* 18.87

22.45 24.15b

*Statistically significant differences (p < .05) between males and females. a and b indicate significant differences at the p < .05 level.

Table 14.7 Descriptive Statistics for the Social Support Survey (SSS) and the Automatic Thinking Questionnaire (AT) Variables

Mean

Standard Deviation

Possible Range

Observed Range

Social Support Survey (SSS) Parents and siblings Partner Work setting General support Groups and institutions Isolation perception

8.55 6.75 5.08 5.26 5.23 3.98

2.89 2.42 1.11 1.97 2.13 1.75

0–12 0–9 0–6 0–9 0–9 0–9

0–12 0–9 0–6 0–9 0–9 0–9

10.64 4.63

2.62 3.00

0–15 0–18

2–15 0–18

Automatic Thinking(AT) Positive Negative

also had significantly higher perceived Severity scores for the Lack of support and Environmental conditions factor than women at comparable organizational levels. The blue-collar men also reported significantly higher levels of job stress than the blue-collar women, due primarily to their higher Overload scores. In contrast, the white-collar women reported higher levels of perceived severity due to Overload than white-collar men. Social Support and Automatic Thinking The means and standard deviations for the Social Support Survey and Automatic Thinking Questionnaire are reported in Table 14.7. When the

RT44367_C014.indd 260

7/13/05 9:50:11 AM

OCCUPATIONAL LEVEL AND GENDER DIFFERENCES

261

scores for the subscales of these measures were compared, taking gender and organizational level into account, the only significant difference was the main effect of organizational level for the SSS General Social Support scale [F(3,132) = 3.01; p < .03]. The managerial personnel had substantially higher mean scores on the SSS General Social Support scale (mean = 6.1) than the technical (4.7), clerical (4.9), or blue-collar (4.7) workers. To further evaluate the relationships between the Spanish JSS scales and scores on the SSS and AT measures, Pearson product-moment correlations were computed, separately for males and females, between each SSS and AT scale and the four JSS measures that were based on all 30 items (number of stressor situations, and JSS Frequency, Severity, and Total scores). The only significant finding for women was a positive correlation, r = .33 (p < .001), between scores on the JSS Severity and SSS Isolation Perception scales. For men, the AT Negative Thoughts scale correlated significantly with the number of JSS stressor events (r = .345, p < .01) and Frequency scores (r = .32, p < .01). DISCUSSION The Spanish adaptation of the JSS developed in Venezuela showed good internal consistency reliability and construct validity. The content of the JSS items was also considered adequate for measuring stress in the workplace in Venezuela and sufficiently comprehensive to provide evidence of cross-cultural validity. Consistent with research with the American JSS, the findings of the present study indicated that gender and organizational level should be taken into account in research on occupational stress. However, factor analyses of the Spanish JSS items identified some important differences in the major dimensions of occupational stress for Venezuelan workers, as compared with the American form, and suggested that three or four Job pressure factors should be separately considered. Two major sources of occupational stress have been consistently identified in factor analyses of the American JSS: Lack of Organizational support and Job pressure (Spielberger, 1994; Spielberger & Vagg, 1999; Spielberger, Vagg, & Wasala, 2003). In the factor analyses of the Severity ratings in the present study, four factors were identified. The largest of these factors was clearly related to Lack of support, as measured by the American JSS, but three smaller factors related to Job pressure were also identified: Overload and competition, Environmental conditions, and Coping management. The Overload and Coping factors identified with the Spanish JSS were defined by excessive job demands and ways for dealing with them. The Environmental conditions factor was defined by physical conditions in the workplace. Most of the job stress conditions that were reported as more severe, or as occurring more frequently, were associated with the Overload and Coping factors. Karasek and Theorell (1990) have emphasized that high job demands (overload) and lack of control over one’s work are critical occupational stress factors that contribute to coronary heart disease. In the present study,

RT44367_C014.indd 261

7/13/05 9:50:11 AM

262

L. FELDMAN ET AL.

work overload (i.e., having too much to do) was identified as one of the major sources of occupational stress. Overload has been reported as the particular aspect of work stress that is most directly related to job dissatisfaction (MacLennan, 1992) and behavioral malfunctions (French & Caplan, 1972). In a study of 1,500 employees, work overload was associated with specific stress-related symptoms such as lowered self-esteem, low work motivation, and escapist drinking (Margolis et al., 1974). Work overload was also found to be a major stressor for public-school administrators (Borg & Riding, 1993) and among the lower ranks of British police officers (Cooper, Davidson, & Robinson, 1982). Gender and Organizational Level Differences in Occupational Stress The results of the present study indicated that men and women differed in their perceptions of sources of stress in the workplace. Men reported that the JSS stressor events were experienced more frequently and perceived them to be more severe than women. The major stressors for men were related to work overload, which might reflect differences in gender-related work assignments because men are often given tasks with greater responsibilities. It is also interesting to note that 85% of men reported experiencing eight or more JSS stressor situations as highly stressful, while only one JSS stressor situation (Meeting deadlines) was reported as highly stressful by women, a finding that deserves special consideration in future research. The results in the present study differed in several important respects from the findings reported by other investigators and especially those studies in which no gender differences were found (e.g., Richard & Krieshok, 1989). In previous research, men reported experiencing stressful situations more often than women, whereas women had higher perceived Severity scores for a large number of situations (Spielberger & Reheiser, 1994). These inconsistencies related to gender differences may be due to cultural differences, the specific occupations that have been investigated, and the procedures and measures that were used to assess occupational stress. LaCroix & Haynes (1987) have emphasized the importance of gender differences in response to job demands. Although both sexes may be required to cope with the same tasks at work, women develop different coping strategies, possibly due to their need to maintain multiple roles. Work stressors should be compared with home and family stressors, as well as with the different coping strategies that are used by men and women. Traditionally, the main concerns of men are related to job responsibilities, while women must more often deal with home-related tasks (Barnett & Marshall, 1991; Biernat & Wortman, 1991; Di Benedetto & Kehr, 1990; Feldman et al., 2002; Mederer, 1993). Gender differences in perceptions of the same environment, together with different cognitive and behavioral coping strategies, seem to contribute to observed differences of men and women in coping with work stressors. Clearly, research on gender differences in occupational stress needs further attention (Nelson, Burke, & Michie, 2002).

RT44367_C014.indd 262

7/13/05 9:50:12 AM

OCCUPATIONAL LEVEL AND GENDER DIFFERENCES

263

In the present study, organizational level seemed to affect women more than men. For example, women white-collar workers reported higher perceived Stress Severity than women blue-collar workers, whereas no differences in perceived Severity were found for men at higher and lower organizational levels. However, differences in organizational level were found for men when specific job stress dimensions were separately analyzed. For example, male blue-collar workers reported more Overload than white-collar men, whereas women white-collar and clerical workers complained more about Lack of support, Overload, and Environmental conditions than blue-collar women. Technical and blue-collar men also reported Lack of support and Environmental problems more often than women at the same organizational levels, and bluecollar men reported experiencing Overload more frequently than blue-collar women. In contrast, white-collar women reported experiencing Overload more often than their blue-collar counterparts. The results of the present study appear to confirm the impression that men at lower organizational levels experience work demands and Lack of support more often than women, while women at higher organizational levels experience more stress due to Overload. However, the specific nature of the tasks and work assignments at the various organizational levels appears to be the critical factor to consider, rather than the differences in organizational level. Further research is needed in which the reactions of men and women to specific tasks at different organizational levels can be investigated. The demand-control latitude (Karasek, 1998; Karasek & Theorell, 1990) for men and women must also be evaluated for different organizational levels. Moreover, as noted by Spielberger and Reheiser (1994), total JSS scores may not accurately reflect gender and organizational level differences. Therefore, it will be important to evaluate the effects of gender-specific aspects of occupational stress, in terms of both Frequency and Severity, for the pressures associated with a particular job. Social support has been systematically considered to buffer the effects of stress on physical and mental health (Chesney & Darbes, 1998; Cohen & Willis, 1985; Cutrona & Russell, 1990; Orth-Gomer, Rosengren, & Wilhelmsen, 1993). In the present study, job stress items relating to poor supervision were reported as the most severe sources of stress for both men and women. Furthermore, stress appeared to lead to different coping styles for men and women. Men reported more negative thinking as the frequency of stress situations increased, whereas women tended to isolate themselves when the severity of stress was high. In summary, the results of this study contribute to understanding differences between men and women in the perceived severity and frequency of occurrence of sources of stress in the workplace. Evaluating gender differences and making cross-cultural comparisons are viable through the use of instruments such as the Spanish Job Stress Scale, which was adapted from the American JSS. Future studies should include job role descriptions, along with analyses of differences in gender and organizational level, when comparing occupational stress in different cultures.

RT44367_C014.indd 263

7/13/05 9:50:12 AM

264

L. FELDMAN ET AL.

REFERENCES Bagés, N., Feldman, L., & Chacón. G. (1995). Patrón de conducta tipo A y reactividad cardiovascular en gerentes [Type-A behavior and cardiovascular reactivity in managers]. Salud Pública de México, 37, pp. 47–56. Barling, J. (1990). Employment, stress, and family functioning. New York: John Wiley & Sons. Barnett, R. C., & Marshall, N. L. (1991). The relationship of women’s work and family roles and their subjective well-being and psychological distress. In M. Frankenhaeuser, U. Lundberg, & M. Chesney (Eds.), Women, work, and health: Stress and opportunities (pp. 111–136). New York: Plenum Press. Bartolome, F., & Evans, P. A. (1979, Spring). Professional lives versus private lives: Shifting patterns of managerial commitment. Organizational Dynamics, 7, pp. 3–29. Biernat, M., & Wortman, C. B. (1991). Sharing of home responsibilities between professionally employed women and their husbands. Journal of Personality and Social Psychology, 60, pp. 844–860. Borg, M. G., & Riding, R. J. (1993). Occupational stress and job satisfaction among school administrators. Journal of Educational Administration, 31, pp. 4–21. Briner, R. B. (1993, April). Bad theory and bad practice in occupational stress. The British Psychological Society, 19, pp. 8–13. Burke, R. J, Weir, T., & Dowors, R. E. (1980). Work demands on administrative and spouse well-being. Human Relations, 33, pp. 253–298. Caplan, R., Cobb, S., French, J. R., Jr., Harrison, R. V., & Pinneau, S. (1975). Job demands and worker health: Main effects and occupational differences (pp. 75–160). Cincinnati, OH: NIOSH. Chan, K. B., Lai, G., Ko, Y. C., & Boey, K. W. (2000). Work stress among six professional groups: The Singapore experience. Social Science & Medicine, 50, pp. 1415–1432. Chesney, M., & Darbes, L. (1998). Social support and heart disease in women: Implications for interventions. In K. Orth-Gomer, M. A. Chesney, & N. K. Wenger (Eds.), Women, stress, and heart disease (pp. 165–182). Mahwah, NJ: Erlbaum. Cohen, S., & Willis, T. A. (1985). Stress, social support, and the buffering hypothesis. Psychology Bulletin, 98, pp. 310–352. Cooper, C. (2002). The changing psychological contract at work: Revisiting the job demands-control model (work psychology). Occupational and Environmental Medicine, 59, p. 355. Cooper, C. L., Davidson, M. J., & Robinson, P. (1982). Stress in the police service. Journal of Occupational Medicine, 24, pp. 30–36. Cutrona, C. E., & Russell, D. W. (1990). In B. R. Sarason, I. G. Sarason, & G. C. Pierce (Eds.), Social support: An interactional view (pp. 319–366). Oxford, UK: John Wiley & Sons. Di Benedetto, B., & Kehr, T. C. (1990). Gender and adult role commitment of women and men in job-family trade-off context. Journal of Counseling Psychology, 37, pp. 41–48.

RT44367_C014.indd 264

7/13/05 9:50:12 AM

OCCUPATIONAL LEVEL AND GENDER DIFFERENCES

265

Dunn, S., Puttallaz, M., Sheppard, B., & Lindstrom, R. (1987). Social support and adjustment in gifted adolescents. Journal of Educational Psychology, 79, pp. 467–473. Feldman, L., Bagés, N., & Chacón, G. (1991). Stress and the protector factors in the managerial environment. Presented at II Congreso de Asociaciones de Terapía del Comportamiento de Países de Lengua Latina “ Latini Dies.” Barcelona: Sitges. Feldman, L., Payne, S., & Viva, E. (2002). Gender roles, role quality, and health in Venezuelan working women. In S. P. Wamala & J. Lynch (Eds.), Gender and social inequalities in health: A public health issue. Lund, Sweden: Studentlitteratur. French, J., & Caplan, R. (1972). Organizational stress and individual strain. In A. J. Marrow (Ed.), The failure of success (pp. 31–66). New York: Amacon. Ingram, R. E., & Wisnicki, K. S. (1988). Assessment of positive automatic cognition. Journal of Consulting and Clinical Psychology, 56, pp. 898–902. Jick, T. D., & Payne, R. (1980). Stress at work. The Organizational Behavior Teaching Journal, 5, pp. 50–56. Karasek, R. A., Jr. (1998). Demand/Control Model: A social, emotional and physiological approach to stress risk and active behaviour development. In J. M. Stellman, L. McCann, L. Warshaw, & C. Dufresne (Eds.), Encyclopedia of occupational health and safety (pp. 34.6–34.14). Geneva: International Labour Office. Karasek, R. A., Jr., & Theorell, T. (1990). Healthy work: Stress, productivity, and the reconstruction of working life. New York: Basic Books. La Croix, A. Z., & Haynes, S. (1987). Gender differences in the health effects of workplace roles. In R. C. Barnett & L. Biener (Eds.), Gender and stress (pp. 96–121). New York: The Free Press. MacLennan, B. W. (1992). Stressor reduction: An organizational alternative to individual stress management. In J. C. Quick, L. R. Murphy, & J. J. Hurrell, Jr. (Eds.), Stress & well-being at work (pp. 79–95). Washington, DC: American Psychological Association. Margolis, B. L., Kroes, H. W., & Quinn, R. P. (1974). Job stress: An unlisted occupational hazard. Journal of Occupational Medicine, 16, pp. 654–661. Matteson, M. T., & Ivancevich, J. M. (1982). Managing job stress and health: The intelligent person’s guide. New York: The Free Press. Mederer, H. (1993). Division of labor in two-earner homes: Task accomplishment versus household management as a critical variable in perceptions about family work. Journal of Marriage and the Family, 55, pp. 133–145. Miller, J. (1980). Individual and occupational determinants of job satisfaction, a focus on gender differences. Sociology of Work and Occupations, 7, pp. 337–366. Nelson, D. L., Burke, .R. J., & Michie, S. (2002). New directions for studying gender, work stress, and health. In D. L. Nelson & R. J. Burke (Eds.), Gender, work stress, and health, (pp. 229–242). Washington, DC: American Psychological Association. Northwestern National Life. (1991). Employee burnout: America’s newest epidemic. Minneapolis: Northwestern National Life Insurance Company.

RT44367_C014.indd 265

7/13/05 9:50:13 AM

266

L. FELDMAN ET AL.

Northwestern National Life. (1992). Employee burnout: Causes and cures. Minneapolis: Northwestern National Life Insurance Company. Orth-Gomer, R., Rosengren, A., & Wilhelmsen, L. (1993). Lack of social support and incidence of coronary heart disease in middle-aged Swedish men. Psychology Medicine, 55, pp. 37–43. Pereyra, M. (2004). Estrés y salud [Stress and health]. In L. A. Oblitas (Ed.), Psicología de la salud y calidad de vida. Polanco, Mexico: Thompson Learning. Richard, G. V., & Krieshok, T. S. (1989). Occupational stress, strain, and coping in university faculty. Journal of Vocational Behavior, 34, pp. 117–132. Spielberger, C. D. (1994). Preliminary Manual for the Job Stress Survey (JSS). Tampa: University of South Florida, Center for Research in Behavioral Medicine and Health Psychology. Spielberger, C. D., & Reheiser, E. C. (1994). The Job Stress Survey: Measuring gender differences in occupational stress. Journal of Social Behavior and Personality, 9, pp. 199–218. Spielberger, C. D., & Vagg, P. R. (1999). Professional manual for the Job Stress Survey (JSS). Odessa, FL: Psychological Assessment Resources. Spielberger, C. D., Vagg, P. R., Wasala, C. F. (2003). Occupational stress: Job pressures and lack of support. In J. C. Quick & L. E. Tetrick (Eds.), Handbook of occupational health physiology, Part II, Chapter 9 (pp. 185–200). Washington, D. C. American Psychological Association. Sutherland, V. J. & Cooper, C. L. (1990). Understanding stress: A psychological perspective for health professionals. In D. Marcer (Ed.), Psychology and Health Series 5 (pp. 273–301). New York, London: Chapman and Hall. APPENDIX A Table 14.A-1 Construct Validity of Job Stress Frequency Factors in the JSS PrincipalComponents Analysis with Varimax Rotation and Kaiser Normalization Factor 1: Lack of Support Item #

Item

Loading

43 51 60 42 36 47 48 34

Difficulty in getting along with supervisor Poor or inadequate supervision Conflicts with other departments Periods of inactivity Inadequate support by supervisor Personal insult from customer/consumer/colleague Lack of participation in policy-making decisions Assignment of new or unfamiliar duties

.697 .686 .667 .639 .605 .582 .507 .429

Note: Eigenvalue 10.88, 36.3% of variance explained.

RT44367_C014.indd 266

7/13/05 9:50:13 AM

OCCUPATIONAL LEVEL AND GENDER DIFFERENCES

267

Factor 2: Recognition and Advancement Item # 50 38 33 49 59 44

Item Competition for advancement Lack of recognition for good work Lack of opportunity for advancement Inadequate salary Poorly motivated coworkers Experiencing negative attitudes toward the organization

Loading .743 .698 .604 .558 .534 .460

Note: Eigenvalue 2.20, 7.4% of variance explained.

Factor 3: Overload Item # 55 53 56 57 58 32 39 52 54

Item Excessive paperwork Frequent interruptions Meeting deadlines Insufficient personal time (coffee breaks, lunch) Covering work for another employee Working overtime Performing tasks not included in job description Noisy work area Frequent changes from boring to demanding activities

Loading .786 .709 .587 .525 .519 .480 .477 .449 .378

Note: Eigenvalue 1.75, 5.8% of variance explained.

Factor 4: Inadequate Personnel and Equipment Item # 40 45 35

Item Inadequate or poor quality of the equipment Insufficient personnel to handle an assignment adequately Fellow workers not doing their job

Loading .627 .622 .601

Note: Eigenvalue 1.56, 5.2% of variance explained.

Factor 5: Coping and Crisis Management Item # 37 31 41 46

Item Dealing with crisis situations Assignment of disagreeable duties Assignment of increased responsibility Making critical on-the-spot decisions

Loading .804 .765 .677 .671

Note: Eigenvalue 1.38, 4.6% of variance explained.

RT44367_C014.indd 267

7/13/05 9:50:14 AM

268

L. FELDMAN ET AL.

Table 14.A-2 Construct Validity of Job Stress Severity Factors in the PrincipalComponents Analysis with Varimax Rotation and Kaiser Normalization Factor 1: Lack of Support Item # 14 6 21 8 12 15 5 18 30 3 19 29 10 17

Item Experiencing negative attitudes toward the organization Inadequate support by supervisor Poor or inadequate supervision Lack of recognition for good work Periods of inactivity Insufficient personnel to handle an assignment adequately Fellow workers not doing their job Lack of participation in policy-making decisions Conflicts with other departments Lack of opportunity for advancement Inadequate salary Poorly motivated coworkers Inadequate or poor quality equipment Personal insult from customer/consumer/colleague

Loading .791 .641 .722 .677 .659 .647 .641 .635 .605 .588 .580 .550 .548 .422

Note: Eigenvalue 9.17, 30.6% of variance explained

Factor 2: Overload Item # 20 26 25 24 9

Item Competition for advancement Meeting deadlines Excessive paperwork Frequent changes from boring to demanding activities Performing tasks not in job description

Loading .751 .700 .603 .590 .580

Note: Eigenvalue 2.82, 9.4% of variance explained.

Factor 3: Coping with Crisis Management Item # 7 11 16 1 4 28 2

Item Dealing with crisis situations Assignment of increased responsibility Making critical on-the-spot decisions Assignment of disagreeable duties Assignment of new or unfamiliar duties Covering work for another employee Working overtime

Loading .655 .644 .620 .551 .543 .513 .295

Note: Eigenvalue 1.95, 6.5% of variance explained.

RT44367_C014.indd 268

7/13/05 9:50:15 AM

OCCUPATIONAL LEVEL AND GENDER DIFFERENCES

269

Factor 4: Environment Conditions Item # 27 23 22 13

Item Insufficient personal time (coffee breaks, lunch) Frequent interruptions Noisy work area Difficulty getting along with supervisor

Loading .647 .623 .549 .356

Note: Eigenvalue 1.48, 4.9% of variance explained.

RT44367_C014.indd 269

7/13/05 9:50:15 AM

RT44367_C014.indd 270

7/13/05 9:50:15 AM

15 Job Stress and Gender: A Comparison of Government and Corporate Employees Lisa Fiksenbaum and Esther R. Greenglass York University, Toronto, Ontario, Canada Abstract Stressful events such as downsizing, layoffs, and mergers have contributed to occupational stress becoming a fact of life for almost everyone in contemporary society. Common sources of job stress experienced by 199 male and female Canadian government workers involved in tax collection and the enforcement of tax laws were evaluated in this study, using Spielberger’s (1994) Job Stress Survey (JSS). The mean JSS Severity and Frequency Scale scores of the Canadian workers were higher than those of a normative group of 394 U.S. corporate employees; the Canadian employees also had higher scores on the JSS Lack of Organizational Support subscales. Analyses of responses to the individual JSS items identified the following specific stressors as most stressful for the Canadian workers: Less attractive pay scales, Lack of opportunity for advancement, Exclusion from policy-making decisions, Inadequate or poor quality equipment, and Noisy work areas. Canadian women rated Insufficient personnel to adequately handle an assignment, Making critical on-the-spot decisions, Noisy work areas, and Personal insult from customer/colleague as more stressful than Canadian men. Conflicts with other departments was the only JSS item that was rated as more stressful by the Canadian men.

JOB STRESS AND GENDER: A COMPARISON OF GOVERNMENT AND CORPORATE EMPLOYEES During the last quarter century, an extraordinary amount of research on the topic of stress has been conducted across a wide spectrum of disciplines (e.g., psychology, sociology, public health, epidemiology, management, and engineering). Vingerhoets and Marcelissen (1988) identified nearly 10,000 stress-related articles that were published between 1976 and 1985, and this research has continued to accelerate. Yet, remarkably, researchers investigating the effects of stress on health and behavior have not reached a consensus on the conceptual meaning of stress, much less formulated prescriptions for attenuating its deleterious effects. It is quite clear, however, that stress is, indeed, a very broad concept that encompasses a conglomerate of variables relating to both stressors and strains. The essential components of the stress 271

RT44367_C015.indd 271

7/13/05 9:35:00 AM

272

L. FIKSENBAUM AND E.R. GREENGLASS

process include the environmental stimuli that impinge on a person and the emotional reactions and behavioral responses that are evoked, which are often highly idiosyncratic. The research of Hans Selye in the 1930s and 1940s marked the beginning of the systematic study of stress. Selye (1993) defined stress as the nonspecific response of the body to any demand made upon it. The General Adaptation Syndrome (GAS) was identified by Selye as a stereotyped reaction to highly stressful circumstances, which involved a three-stage process. In the first stage, the organism responds to stressful environmental stimuli with an alarm reaction, in which intense emotions are mobilized to handle the demands produced by the environment. A resistance stage followed, with a return to homeostatic balance if the environmental demand ceased. If the demand did not attenuate and the organism continued to experience stress, the final stage of exhaustion occurred; death resulted if the stress was prolonged and severe. Although Selye’s research has been highly influential in interpreting the nature of stress in humans, his empirical work was largely based on the physiological reactions produced in animals by physical stressors. As research in this field progressed, it became apparent that the physiological mechanisms in stress reactions presented only one facet of the concept of stress as a complex process. Consequently, in recent years, physiological response models of stress have been superseded by interactional models that emphasize the importance of psychological and social factors in the stress process. Among these, Lazarus’ (1966) Transactional Process model, which examines stress as a unique interaction between a person and the environment (Lazarus & Folkman, 1984), has gained widespread acceptance among researchers. The basic tenet of the Transactional Process model is that an environmental event is stressful to the extent that it is appraised by a person as threatening or disturbing, taking into acount the individual’s coping resources. Because physical and psychological stressors evoke stress reactions only if they are perceived as threatening, a particular stimulus or event may evoke stress responses in one individual but not in another. Moreover, people differ not only in how they appraise a stressor, but also in their resources and capabilities for dealing with the threat. Although this subjectivity has contributed to a lack of clarity in the concept of stress, recognizing the critical importance of the perception and appraisal of a stressor has furthered our understanding of the nature of stress as a process that involves stressors and strain. Stress and Health Despite the lack of consensus in regard to the conceptual meaning of stress, the findings in this large and expanding literature suggest that stress reactions contribute to a variety of disorders and illnesses. In the stress research literature, the relationship between stress and coronary heart disease has received the greatest attention. Epidemiological studies have established stress as a risk factor for cardiovascular disease, along with age, elevated blood pressure and hypertension, obesity, physical inactivity, cigarette smoking, the Type-A Behavior Pattern, and genetic influences (Houston, Smith, & Cates,

RT44367_C015.indd 272

7/13/05 9:35:00 AM

JOB STRESS AND GENDER

273

1989; Schnall, Deveraux, Pickering, & Schwartz, 1992; Strube, 1991). In Canada, cardiovascular disease and stroke are the major causes of death, disability, and illness, accounting for more than 76,000 deaths in 1992 (Heart & Stroke Foundation of Canada, 1995). Lack of social support, problematic interpersonal relations, and anger/hostility, a component of the Type-A construct, have also been identified as major stressrelated factors that contribute to the development of coronary heart disease (Chesney, 1988). A combination of these factors, along with high scores on a measure of hostility, was found to be associated with greater susceptibility to illness and death (Smith & Pope, 1990). In addition to the detrimental effects on health, the enormous costs of stress-related problems to the economy have been recognized. For example, the Heart & Stroke Foundation of Canada (1995) estimated that $17 billion was spent in Canada alone for medical care, drugs, disability pensions, and lost wages due to heart attacks, stroke, and other cardiovascular disorders. Stress may also induce personality changes that contribute to pathological behaviors. Individuals who lack adaptive coping techniques may react to stressful situations with alcohol and drug abuse, smoking, and poor diet, and these behaviors may, in turn, render the individual susceptible to mental disorders such as chronic anxiety and depression. In addition, stress-induced behaviors may exacerbate reactions to stressful situations that perpetuate a vicious cycle for “stressed-out” individuals, who become a substantial burden for society, both economically and socially. Clearly, the consequences of stress for the individual, and for society as a whole, are far-reaching and warrant comprehensive analysis. Occupational Stress: Its Sources and Consequences Work occupies a major part of most people’s lives. Consequently, the work setting provides an important site for examining the impact of stress on individuals and the efficacy of various techniques used to attenuate their stress reactions. For many employees, stress in the workplace is steadily increasing for a variety of reasons, which include changing technology, downsizing of business organizations, and increased competition. In fact, stress at work is so commonplace in contemporary society that it is accepted as an inevitable characteristic of modern life. Given the growing concern about the detrimental effects of occupational stress, numerous studies have been undertaken to help delineate its causes and consequences (e.g., Beehr, 1995; Cooper & Payne, 1988; Cooper & Smith, 1985; Perrewé, 1991). Although the sources of occupational stress are ubiquitous, evidence from a growing body of literature suggests that five major job stress categories can be identified. Working conditions intrinsic to the job (e.g., noisy work areas, excessive paperwork, time pressures) constitute a major source of occupational stress. Other sources of occupational stress, such as difficulty in getting along with a supervisor and covering work for other employees, stem from the employee’s role within the organization. In addition, career development obstacles (e.g., lack of opportunity for advancement) and the

RT44367_C015.indd 273

7/13/05 9:35:00 AM

274

L. FIKSENBAUM AND E.R. GREENGLASS

structure and climate of an organization (e.g., office politics, participation in decision making, work/family conflict) contribute to occupational stress. Regardless of its particular source, stress in organizational life has the potential for eliciting reactions that are often maladaptive. Included among these negative reactions are job dissatisfaction (Bhagat, Allie, & Ford, 1991; Chen & Spector, 1992; Landsbergis, 1988; McLaney & Murrell, 1988), poor physical and mental health, psychosomatic symptoms, and heart disease (Ganster, Fusilier, & Mayes, 1986; Greenglass, 1988; Jayaratne & Chess, 1984; Kaufman & Beehr, 1986; Smith & Pope, 1990; Sutherland & Cooper, 1990). Occupational stress can also be very costly for organizations. Poor performance, low morale, absenteeism, and job turnover are frequently observed in stressed-out employees (Beehr & Newman, 1978; Leiter, 1991; Shirom, 1989). Given the complexity of stress-related problems, it is difficult to arrive at an accurate determination of these costs. However, in an Ontario provincial government survey, chief executives estimated that occupational stress contributed to more than 26% of employee sick time, which translated into 11.6 million lost days of work at a cost of $1.2 billion annually (Matthews, 1992). Because occupational stress has been found to result in numerous deleterious consequences for both individuals and organizations, it is important to focus on the antecedents of job-related stressors, along with their effects on performance and health outcomes. It is also important to develop sound instruments for measuring the specific sources of stress that are experienced by men and women working at different occupational levels. Unfortunately, most job stress measures confound the severity of stressful events with the frequency of their occurrence. In evaluating stress in the workplace, both of these major dimensions of job-related stress require careful evaluation. The Job Stress Survey (JSS) was developed by Spielberger (1994) and his colleagues (Spielberger & Reheiser, 1994a, 1994b; Spielberger & Vagg, 1999; Turnage & Spielberger, 1991) as a generic measure of occupational stress to facilitate comparison of the stress levels of employees working at different occupational levels in a wide variety of work environments. The JSS assesses the perceived Severity (intensity) and Frequency of occurrence of 30 general sources of occupational stress; self-ratings of both perceived Severity and Frequency are obtained for each of these stressor events. In addition to providing comprehensive measures of Severity and Frequency based on all 30 items, the JSS also assesses two major components of occupational stress: Job pressure and Lack of organizational support. The information provided by the JSS facilitates making changes in the work setting to alleviate job stress and the development of effective employee assistance programs as interventions that help workers to deal more effectively with job stress and improve their performance (O’Roark, 1995). Most research on occupational stress has focused on managers and professionals, such as lawyers, doctors, nurses, and teachers. In contrast, the effects of job stress on government employees and the problems they encounter has received relatively little attention. Given the current social, economic, and political climate, the demands on government workers are both substantial and increasing, whereas the available resources to help them meet these demands

RT44367_C015.indd 274

7/13/05 9:35:01 AM

JOB STRESS AND GENDER

275

are limited. The social and economic realities for many government workers include salary freezes, lack of mobility, limited opportunity for advancement, job uncertainty, downsizing, and technological advances. When coupled with the problems of a bureaucratic system, such shortcomings often result in a severely stressful work environment. Gender Differences in Occupational Stress The changing role of women in contemporary society has stimulated a substantial amount of research on understanding the stressors and strains experienced by working women. While paid employment may have beneficial effects for women, including positive health outcomes (Repetti, Matthews, & Waldron, 1989; Verbrugge, 1989), researchers have observed that working women experience unique stressors because of their relatively lower status in a patriarchal work society. These stressors included sex discrimination, threats of violence, social isolation, and work/family conflict (Greenglass, 1985). Female and male managers also differ in the kinds of work-related situations and events they perceive as stressful (McDonald & Korabik, 1991). In a study of 524 clerical and managerial personnel, Hendrix, Cantrell, and Steel (1988) found that women exhibited significantly higher levels of both occupational and life stress than their male counterparts. Age also significantly influenced the stress experienced by female executives as a consequence of work/family conflict, societal expectations, and the behavioral norms that women commonly face when they occupy a combination of roles (Beena & Poduval, 1992). On the basis of a comprehensive review of research on gender differences in job stress, Nelson and Quick (1985) concluded that employed women experienced greater stress than either unemployed women or men because of the unique stressors that they typically encountered. Although a number of studies have reported gender differences in occupational stress, DiSalvo, Lubbers, Rossi, and Lewis (1995) found that men and women did not differ in their perceptions of work-related stressors. Similarly, Martocchio and O’Leary (1989) found no gender differences in a metaanalysis of occupational stress studies and little support for the notion that stress was experienced differently by women and men in various work settings. Despite being subjected to more stressors than their male counterparts, Korabik, McDonald, and Rosin (1993) concluded that women managers did not experience any more severe stress reactions than men, nor more negative consequences of job-related stress. Jick and Mitz (1985) attributed the inconsistent findings in research on gender and occupational stress to variations in the definition of stress, methodological problems, and sampling differences. They concluded that stress was experienced differently by men and women, which resulted in depression and other emotional problems for women, whereas men were more likely to experience coronary heart disease, cirrhosis of the liver, and other physical disorders. Jick and Mitz (1985) also noted that investigators often failed to take into account the patriarchal structure of the workforce that relegates women to clerical and service jobs, whereas men are more likely to have executive/managerial positions.

RT44367_C015.indd 275

7/13/05 9:35:01 AM

276

L. FIKSENBAUM AND E.R. GREENGLASS

In order to understand gender differences in research on occupational stress, it is essential to take into account the fact that men outnumber women in supervisory/managerial positions, whereas women tend to be clustered in lower-status clerical jobs. When comparing male and female perceptions of job-related stress, it is also important to recognize that women who work outside the home must add this vocational role to their other duties and responsibilities. Consequently, women are expected to fulfill the demands associated with both work and family duties. It is also evident that global measures of occupational stress tend to mask gender differences, which may be found at a more molecular level of analysis (Spielberger & Vagg, 1999). The research reported in this chapter has focused on three main objectives. First, the structure and content of job stress among Canadian government workers responsible for the collection of taxes and the enforcement of the Income Tax Act were examined. Second, using the Job Stress Survey (Spielberger, 1994; Spielberger & Vagg, 1999) to assess occupational stress, two distinct samples, government workers and corporate employees, were compared. Finally, gender differences in job stress were examined within each of these samples. METHOD AND PROCEDURE Participants The participants in this study were 198 government workers (96 men, 102 women) who were employed in two Canadian District Income Tax offices (Fiksenbaum, 1997). The average age of these employees was 37.5 years (SD = 9.14), with ages ranging from 22 to 63 years. All respondents were employed full time and worked an average of 37.72 hours per week (SD = 1.72). The average tenure of employment was 7.01 years (SD = 6.53). The demographic characteristics of the Canadian government workers are described in greater detail by the first author (1997). The responsibilities of the government employees who participated in this study were defined in accordance with the four organizational levels established by the Programme Administrator (PM). At the first or basic PM level, workers respond to inquiries from the public regarding income taxes and child tax benefits, and are responsible for collecting personal and corporate tax arrears and negotiating payment arrangements. Workers at the next higher level, PM2, have limited supervisory responsibilities and are responsible for interviewing clients regarding their financial situations, reviewing records for compliance with the tax laws, and monitoring the enforcement of these laws. At the PM3 level, the employees provide technical advice and counseling for their coworkers. At PM4, the highest level, the employees’ responsibilities include supervising and directing team members, ensuring that standards are maintained, and resolving difficult cases with clients. When organizational levels were examined in relation to gender, two general trends emerged. For the lowest PM level, where respondents had the most contact with the public, 60% were women and 40% were men. When the three higher PM levels

RT44367_C015.indd 276

7/13/05 9:35:01 AM

277

JOB STRESS AND GENDER

involving supervisory activities were combined, there were more men (56%) than women (44%) at these levels. The Canadian government workers who participated in the present study were compared with a normative sample of U.S. workers employed in business and university settings (Spielberger, 1994). The U.S. normative sample consisted of 450 managerial, professional, and clerical employees (332 males, 118 females). Similar to the participants in the present study, a substantially higher proportion of the males were managers or professionals, whereas a majority of the females were employed in clerical positions Measures The Job Stress Survey, a self-report measure of occupational stress (Spielberger, 1994; Spielberger & Vagg, 1999), was designed to address the shortcomings that have been noted in previously published measures, such as the failure to take into account the severity (intensity) and frequency of occurrence of sources of stress in the workplace. The 30 generic JSS items were adapted from two earlier measures: the Police Stress Survey (Spielberger, Westberry, Grier, & Greenfield, 1981) and the Teacher Stress Survey (Grier, 1982). Examples of the JSS stressor items are Excessive paperwork, Frequent interruptions, Meeting deadlines, Poor or inadequate supervision, and Fellow workers not doing their job. Research with the JSS has demonstrated that this inventory is applicable for use with employees working at different organizational levels (i.e., managers, professionals, clerical workers) in a wide variety of occupational settings. In responding to the JSS, employees rate the perceived Severity and the frequency of occurrence of working conditions that are likely to have an adverse effect on their psychological well-being. The perceived Severity of each of the 60 JSS stressor events is rated on a nine-point scale, as compared to a standard stressor event, “Assignment of disagreeable duties,” which was consistently given a stress Severity rating of 5 in previous research (Spielberger & Reheiser, 1994b). The frequency of each JSS stressor event is rated on a 10-point scale, from 0 to 9+, to indicate the approximate number of days on which the stressor was experienced during the preceding six months. Severity and Frequency ratings for the 30 items are summed to calculate overall Severity and Frequency scores. The Job Stress Index, which indicates the overall level of occupational stress experienced by a worker, is computed by summing the cross-products of the Severity and Frequency ratings for each of the 30 JSS items and dividing these totals by 30 to determine the average stress score for all 30 items. Factor analyses of the 30 JSS items based on large samples of managerial, professional, and clerical employees have consistently identified two major sources of occupational stress: Job pressure and Lack of organizational support (Spielberger & Reheiser, 1994b). Examples of Job pressure items are Assignment of increased responsibility, Making critical on-the-spot decisions, and Meeting deadlines. Examples of Lack of support items include Difficulty getting along with supervisor, Lack of recognition for good work, and

RT44367_C015.indd 277

7/13/05 9:35:01 AM

278

L. FIKSENBAUM AND E.R. GREENGLASS

Poorly motivated coworkers. Job pressure and Lack of organizational support scores are computed by summing the ratings for the 10 JSS Severity and Frequency items that are included in these subscales. High scores on the Job pressure subscale signify greater perceived occupational stress resulting from the requirements of the job; high scores on the Lack of organizational support subscale indicate less support from supervisors, coworkers, and organizational policies and procedures. Procedures The JSS was administered in research packets that were distributed to the Canadian government workers by a representative of the Government Human Resources Department. Each packet contained a cover letter that (1) explained the nature and importance of the study, (2) informed the respondent that participation was voluntary, and (3) assured the anonymity and confidentiality of the respondent. The packet also contained a letter from the Director of the worker’s District Office, encouraging the employee to respond to the questionnaire. After completing the questionnaire, the respondents were instructed to place it in an addressed envelope and deposit this envelope in a sealed box located in the Resource Center of their District Office. A total of 640 questionnaires were distributed. Three employees were dropped from the study because they failed to respond to three or more questionnaire items. The final sample of 198 respondents represented approximately 32% of those invited to participate in the study. RESULTS The means, standard deviations, and alpha coefficients for the JSS Severity, Frequency, and Index scores of the male and female Canadian government workers (Fiksenbaum, 1997) are compared in Table 15.1 with a normative sample of U.S. corporate employees (Spielberger & Reheiser, 1994a). Alpha coefficients of .88 or higher for the Canadian workers were comparable to those of the U.S. employees, indicating a high level of internal consistency for the JSS Severity and Frequency scales based on responses to all 30 stressor events. The mean JSS Severity and Frequency scores of the Canadian government workers were higher than the corresponding scores of the U.S. corporate sample, especially the Frequency ratings of the Canadian workers, which were substantially higher than those of the U.S. employees. Thus, the Canadian workers reported that they perceived the JSS stressor events as somewhat more stressful than the U.S. employees, and that they experienced stress in the workplace much more frequently. The means and standard deviations for the 10-item JSS Job pressure and organizational support Severity and Frequency subscales of the Canadian government workers and U.S. corporate employees are compared in Table 15.2. The Canadian workers reported perceiving more severe occupational stress

RT44367_C015.indd 278

7/13/05 9:35:02 AM

279

JOB STRESS AND GENDER

Table 15.1 Means and Standard Deviations of the JSS Total Severity and Frequency Scales for Canadian Government Workers and U.S. Corporate Employees Canadian Workers

U.S. Employees

Male (N = 96)

Female (N = 102)

Male (N = 280)

Female (N = 114)

Severity Mean SD Alpha

155.14 33.18 .90

162.78 33.56 .91

152.34 25.64 .88

151.35 34.80 .90

Frequency Mean SD Alpha

130.50 51.04 .90

128.54 47.39 .88

115.96 46.47 .88

102.96 50.15 .89

JSS Scale

Table 15.2 Means and Standard Deviations for the JSS Job Pressure and Lack of Organizational Support Severity and Frequency Subscales for the Canadian Government Workers and U.S. Corporate Employees Government Workers (N = 198)

Corporate Employees (N = 394)

Mean

SD

Mean

SD

Job Pressure Severity Frequency

52.82 41.90

12.46 18.72

45.00 46.24

11.69 20.50

Lack of Support1 Severity Frequency

55.35 44.52

12.18 17.43

55.53 31.62

12.93 19.56

JSS Scale

1

Higher scores indicate less organizational support.

and experiencing stress much more frequently in the workplace than the U.S. corporate employees, as indicated by their higher Job pressure Severity scores and substantially higher Lack of organizational support Frequency scores. In contrast, the U.S. employees reported experiencing job pressures more frequently than the Canadian workers. There was no difference in the Lack of organizational support Severity subscale scores, which were quite similar for both samples. Although the analyses of the JSS scale and subscale scores indicated that the Canadian government workers experienced greater overall work stress than the U.S. corporate employees, the major difference between the two groups was that the government workers reported

RT44367_C015.indd 279

7/13/05 9:35:02 AM

280

L. FIKSENBAUM AND E.R. GREENGLASS

experiencing Lack of organizational support much more frequently than the corporate employees. Gender differences in the JSS scales and subscales were evaluated by t-tests of the Severity, Frequency, and Index scores of the male and female Canadian government workers. No significant gender differences were found in the JSS scales based on all 30 items, nor for the 10-item JSS Job pressure and Lack of organizational support subscales. Similar results were reported by Spielberger and Reheiser (1994b), who found no gender differences in the JSS Index scale and subscale scores for university and corporate employees. These results may be interpreted as indicating that the major dimensions of job stress measured by the JSS scales and subscales are too global to detect gender differences that are evident at the item level (Spielberger & Vagg, 1999). Also, as noted by Spielberger and Reheiser (1994b), when scores for specific job stressors are summed to compute scale and subscale scores, opposite differences between men and women on some of the items tended to cancel out. However, numerous group and gender differences in specific JSS stressor events have been identified in previous research that focused on the individual JSS items (Spielberger & Vagg, 1999). Group and Gender Differences in Ratings of Specific JSS Stressor Events Group and gender differences in the work environments of the Canadian government workers and U.S. corporate employees were examined in separate analyses of the Severity and Frequency ratings for each of the 30 JSS stressor events. The means and standard deviations of the Severity scores of male and female Canadian and U.S. workers for 29 of the 30 JSS items are reported in Table 15.3. Item 1, Assignment of disagreeable duties, was not rated for Severity because this stressor event, which was assigned a score of 5.0, served as the standard against which the perceived severity of the other 29 items was compared. Significant gender differences in the Severity item scores of men and women were found for 14 items in the U.S. sample and for 5 items in the Canadian sample. Women had higher scores than the men in 13 of these 19 comparisons. In same-gender analyses across the two samples of the mean scores of men and women for the 29 Severity items, the Canadian workers had higher scores in 48 of these 58 comparisons (Fiksenbaum, 1997). The results of similar analyses computed for the 30 JSS Frequency items are reported in Table 15.4. Only two significant gender differences were found for the Canadian government workers. Men reported higher Frequency scores for Poorly motivated coworkers, whereas women scored higher on Insufficient personnel to adequately handle an assignment. For the U.S. employees, significant gender differences were found for 18 of the 30 Frequency items. Men scored higher on 10 of these items; women had higher Frequency scores on 8 items. In the analyses of the same-gender scores of men and women across the two samples, the Canadian government workers had higher mean Frequency scores than the U.S. corporate employees in 45 of these 60 comparisons

RT44367_C015.indd 280

7/13/05 9:35:03 AM

281

JOB STRESS AND GENDER

(Fiksenbaum, 1997). Thus, the Canadian government workers reported that they experienced 75% of the JSS stressor events more frequently than the U.S. employees. As previously noted, most of the sources of job stress assessed by the JSS were also perceived as more severely stressful by the Canadian government workers than the U.S. corporate employees. Examples of the stressors perceived by the Canadian workers as both significantly more severe and experienced more often included Inadequate salary, Lack of opportunity for advancement, Noisy work areas, Inadequate or poor quality equipment, and Periods of inactivity. DISCUSSION The main goals of the research reported in this chapter were to evaluate the occupational stress of Canadian government workers as measured by the Job Stress Survey scales and subscales (Spielberger, & Vagg, 1999), and to compare the job stress of Canadian government workers with U.S. corporate employees. The findings of higher mean Severity, Frequency, and Index scale and subscale scores for the government workers indicated that they experienced substantially more stress in their work environment than the U.S. corporate employees. These findings also suggested that the government workers whose jobs required frequent contact with the public or extensive supervisory responsibilities experienced more work-related stress than the heterogeneous sample of U.S. corporate workers, which included managerial, professional, and clerical employees. The higher JSS scale and subscale scores of the Canadian government workers indicated that they experienced greater overall work stress than the U.S. corporate employees. In the comparisons of the scores on the individual JSS Severity items, the male and female government workers had higher mean scores than the U.S. corporate employees on 48 of the 58 items (excluding Item 1, the standard against which the perceived Severity of the other items was compared), and on 45 of the 60 Frequency items. The Canadian government workers were more likely to report that their work efforts were unappreciated, as reflected in their perception of being inadequately paid, unrecognized for their work, and lacking opportunities for advancement. They also reported that their working conditions were less satisfactory (e.g., greater noise, inadequate equipment, personal insults from others, and periods of inactivity). In addition to indicating that they experienced greater job pressures than the corporate employees, the government workers also reported that supervisory support was less adequate and that their coworkers were not doing their jobs. Consequently, it was not surprising that the Canadian government workers experienced stronger negative attitudes toward their organization than the corporate employees. Given the greater stress experienced by the Canadian govenrment workers, especially their feelings of lack of organizational support, these workers might also be expected to be more prone to illness, as was found in a previous study by the second author (1996) on relationships between emotional factors, coping, and illness.

RT44367_C015.indd 281

7/13/05 9:35:03 AM

282

L. FIKSENBAUM AND E.R. GREENGLASS

Table 15.3 Means and Standard Deviations of the JSS Severity Items for Male and Female Canadian Government Workers and U.S. Corporate Employees, and Statistical Tests of Gender Differences Canadian Workers JSS Items

Male

Female

M

5.00

5.00

M SD

3.38 2.09

3.73 1.97

3. Lack of opportunity for advancement

M SD

6.79 2.13

4. Assignment of new or unfamiliar duties

M SD

5. Fellow workers not doing their job

U.S. Employees Male

Female

5.00

5.00

1.20

3.85 2.09

3.83 2.19

.03

7.06 2.02

.91

6.01 2.25

6.22 2.25

4.02*

5.04 1.82

5.01 1.97

.12

4.30 1.97

4.28 1.97

.05

M SD

5.91 2.18

6.40 1.93

1.70

5.63 2.12

5.74 2.29

1.13

6. Inadequate support by supervisor

M SD

5.90 2.43

6.24 2.48

.98

5.78 2.29

5.69 2.64

.61

7. Dealing with crisis situations

M SD

5.80 1.86

5.87 1.86

.27

5.43 2.04

5.41 2.15

.02

8. Lack of recognition for good work

M SD

6.03 2.23

6.39 2.05

1.18

5.55 2.25

5.67 2.18

1.42

9. Performing tasks not in job description

M SD

4.51 1.96

4.70 2.01

.66

3.80 2.09

4.21 2.20

16.21***

10. Inadequate or poor quality equipment

M SD

5.31 1.83

5.47 2.10

.56

5.08 2.21

4.98 2.31

.88

11. Assignment of increased responsibility 12. Periods of inactivity

M SD

5.01 1.86

5.14 2.09

.45

4.30 1.94

4.44 2.06

2.82

M SD

4.00 2.44

4.08 2.76

.24

3.83 2.19

3.88 2.30

.16

M 13. Difficulty getting along with supervisor SD

4.41 2.63

5.04 2.64

1.67

5.21 2.63

5.07 2.89

1.14

14. Experiencing negative M attitudes toward the SD organization 15. Insufficient personnel M to adequately handle SD an assignment

5.83 2.14

5.78 2.18

.16

5.34 2.21

4.95 2.36

13.10***

5.44 2.36

6.34 1.82

3.04**

5.86 2.10

5.62 2.24

5.63*

1. Assignment of disagreeable duties 2. Working overtime

RT44367_C015.indd 282

t

F

7/13/05 9:35:04 AM

283

JOB STRESS AND GENDER Table 15.3 (Continued) Canadian Workers JSS Items

U.S. Employees

Male

Female

t

Male

Female

F

M SD

4.75 1.82

5.29 1.85

2.08*

4.36 1.97

4.68 2.00

12.14***

17. Personal insults from M customer/colleague SD

5.40 2.47

6.06 2.22

1.99*

5.28 2.26

5.32 2.35

.09

18 Lack of participation in policy-making decisions 19. Inadequate salary

M SD

5.29 2.28

5.52 2.21

.71

5.51 2.08

5.12 2.18

15.52***

M SD

7.10 2.03

7.31 1.74

.78

6.16 2.18

6.68 2.14

25.99***

20. Competition for advancement

M SD

6.26 2.07

6.58 1.96

1.11

4.59 2.08

4.97 2.24

13.23***

21. Poor or inadequate supervision

M SD

5.04 2.09

5.50 2.32

1.40

4.82 2.38

4.54 2.48

5.99**

22. Noisy work area

M SD

4.44 2.39

5.25 2.21

2.47*

4.28 2.26

4.60 2.45

8.54**

23. Frequent interruptions

M SD

5.13 2.32

5.57 2.18

1.38

5.08 2.11

5.44 2.33

11.43***

24. Frequent changes M from boring to SD demanding activities 25. Excessive paperwork M SD

4.52 1.95

4.58 1.95

.21

3.86 1.91

4.04 2.08

3.44

5.22 2.37

5.43 1.82

.71

5.07 2.15

4.91 2.27

2.44

26. Meeting deadlines

M SD

4.48 1.96

4.84 1.83

1.33

4.95 2.07

4.85 2.17

.95

27. Insufficient personal time

M SD

3.19 2.07

3.74 2.06

1.86

3.45 2.05

4.03 2.33

30.93***

28. Covering work for another employee

M SD

4.70 2.37

4.83 2.52

.41

4.12 2.13

4.48 2.28

12.12***

29. Poorly motivated coworkers

M SD

6.29 2.01

5.78 2.18

1.70

5.14 2.22

5.21 2.35

.38

30. Conflicts with other departments

M SD

5.14 2.06

4.48 2.29

2.12*

4.83 2.18

4.44 2.31

13.16***

16. Making critical onthe-spot decisions

***p < .001; **p < .01; *p < .05.

RT44367_C015.indd 283

7/13/05 9:35:05 AM

284

L. FIKSENBAUM AND E.R. GREENGLASS

Table 15.4 Means and Standard Deviations of the JSS Frequency Items for Male and Female Canadian Government Workers and U.S. Corporate Employees, and Statistical Tests of Gender Differences Canadian Workers JSS Items

U.S. Employees

Male

Female

t

Male

Female

F

1. Assignment of disagreeable duties

M SD

3.86 3.51

3.12 3.12

1.55

3.01 2.90

2.80 2.96

2. Working overtime

M SD

2.89 3.51

3.11 3.82

.43

5.14 3.93

3.61 3.81

3. Lack of opportunity for advancement

M SD

5.69 3.97

5.52 3.91

.31

3.05 3.79

3.31 3.80

1.95

4. Assignment of new or unfamiliar duties

M SD

4.07 3.25

3.87 3.28

.43

3.11 3.12

3.38 3.13

3.41

5. Fellow workers not doing their job

M SD

5.65 3.39

6.14 3.24

1.04

4.21 3.53

4.16 3.55

.09

6. Inadequate support by supervisor

M SD

3.88 3.51

3.29 3.58

1.15

2.46 3.18

2.47 3.25

.01

7. Dealing with crisis situations

M SD

4.14 3.17

3.59 2.94

1.24

4.72 3.43

4.16 3.29

12.27***

8. Lack of recognition for good work

M SD

5.05 3.62

5.15 3.48

.19

3.57 3.52

3.16 3.47

6.02**

9. Performing tasks not in job description

M SD

4.69 3.56

5.02 3.54

.66

4.08 3.72

4.12 3.62

.06

10. Inadequate or poor quality equipment

M SD

4.46 3.29

4.47 3.48

.03

3.87 3.68

3.60 3.66

2.40

11. Assignment of increased responsibility 12. Periods of inactivity

M SD

4.60 3.50

4.87 3.55

.54

3.65 3.22

4.18 3.32

11.55***

M SD

2.92 3.32

3.06 3.54

.29

1.56 2.59

1.85 2.70

5.24*

M 13. Difficulty getting along with supervisor SD

1.35 2.42

1.18 2.06

.65

1.43 2.51

1.41 2.52

.03

14. Experiencing negative M attitudes toward the SD organization 15. Insufficient personnel M to adequately handle SD an assignment

6.01 3.44

5.88 3.05

.28

4.29 3.47

3.49 3.47

23.81***

4.34 3.62

5.42 3.51

–2.13*

4.70 3.63

4.07 3.69

13.40***

RT44367_C015.indd 284

2.21 70.28***

7/13/05 9:35:06 AM

285

JOB STRESS AND GENDER Table 15.4 (Continued) Canadian Workers JSS Items

U.S. Employees

Male

Female

t

Male

Female

F

M SD

4.84 3.23

4.30 3.23

1.11

4.78 3.45

3.79 3.21

39.52***

17. Personal insults from M customer/colleague SD

3.29 3.42

3.81 3.29

1.09

1.83 2.54

2.18 2.83

7.12**

18 Lack of participation in policy-making decisions 19. Inadequate salary

M SD

5.25 3.78

4.48 3.64

1.46

3.89 3.41

3.39 3.48

9.19**

M SD

7.18 3.17

7.01 3.13

.37

4.53 4.07

5.27 4.02

14.69**

20. Competition for advancement

M SD

3.03 3.74

2.25 3.20

1.58

2.10 3.13

2.00 3.05

.38

21. Poor or inadequate supervision

M SD

3.38 3.47

3.02 3.44

.73

2.39 3.23

2.08 3.13

4.11*

22. Noisy work area

M SD

4.05 3.51

4.61 3.44

1.12

3.52 3.72

3.98 3.81

6.63*

23. Frequent interruptions

M SD

6.13 3.22

6.51 3.15

.85

6.28 3.45

6.52 3.28

2.37

24. Frequent changes M from boring to SD demanding activities 25. Excessive paperwork M SD

3.98 3.47

3.95 3.23

.06

3.54 3.61

3.99 3.57

7.01**

5.59 3.51

5.46 3.38

.26

4.97 3.65

5.02 3.58

.08

26. Meeting deadlines

M SD

4.56 3.37

4.72 3.22

.34

5.87 3.40

5.49 3.43

5.29*

27. Insufficient personal time

M SD

2.08 3.17

2.20 3.12

.28

2.38 3.43

3.04 3.66

16.05***

28. Covering work for another employee

M SD

4.49 3.34

4.58 3.27

.19

3.61 3.38

4.11 3.48

9.44**

29. Poorly motivated coworkers

M SD

6.50 3.12

5.57 3.32

2.03*

3.69 3.48

3.88 3.59

1.28

30. Conflicts with other departments

M SD

3.03 2.83

2.73 2.92

.75

3.05 3.21

1.98 2.81

16. Making critical onthe-spot decisions

57.61***

***p < .001; **p < .01; *p < .05.

RT44367_C015.indd 285

7/13/05 9:35:07 AM

286

L. FIKSENBAUM AND E.R. GREENGLASS

Table 15.5 Gender Differences for JSS Item Severity and Frequency: Summary of Results Severity Government Workers JSS Item

Corporate Employees

Male

Female

Male

Female

Insufficient personnel to adequately handle an assignment

M SD

5.44 2.36

6.34 1.82

5.86 2.10

5.62 2.24

Making critical on-the-spot decisions

M SD

4.75 1.82

5.29 1.85

4.36 1.97

4.68 2.00

Noisy work areas

M SD

4.44 2.39

5.25 2.21

4.28 2.26

4.60 2.45

Conflicts with other departments

M SD

5.14 2.06

4.48 2.29

4.83 2.18

4.44 2.31

Personal insult from customer/colleague

M SD

5.40 2.47

6.06 2.22

5.28 2.26

5.32 2.35

Frequency Government Workers JSS Item

Corporate Employees

Male

Female

Male

Female

Insufficient personnel to adequately handle an assignment

M SD

4.34 3.62

5.42 3.51

4.70 3.63

4.07 3.69

Poorly motivated coworkers

M SD

6.50 3.12

5.57 3.32

3.69 3.48

3.88 3.59

Job Stress and Gender Differences In her thesis research, the first author (1997) found no significant gender differences in the scores of Canadian government workers on the JSS Severity, Frequency, and Index scales, nor in their scores on the JSS Job pressure and Lack of organizational support subscales. Spielberger and Reheiser (1994b) reported a similar failure to find gender differences in their evaluation of male and female U.S. corporate employees. (See Table 15.5.) In both of these studies, the more global measures of job stress, which are assessed by the JSS scales and subscales,

RT44367_C015.indd 286

7/13/05 9:35:07 AM

JOB STRESS AND GENDER

287

masked gender-related effects that were identified in tests of the perceived severity of the specific sources of stress evaluated by the individual JSS items and in the frequency that these stressors were experienced by men and women. In both the first author’s (1997) evaluation of Canadian government workers and Spielberger and Reheiser’s (1994b) study of U.S. corporate employees, men rated Conflicts with other departments as a more severe stressor than women, which might be expected from a gender-role perspective. When masculine and feminine gender roles are compared, a differential emphasis on power is clearly evident. Given that the masculine gender role is more likely to emphasize independent achievement, dominance, and personal power (Blalock & Eisler, 1991; Greenglass, 1982), it is not surprising that the male government and corporate workers rated Conflicts with other departments as more severe than their female counterparts. In their analyses of gender differences for U.S. corporate workers, Spielberger and Reheiser (1994b) found that the specific sources of stress rated by women as more severe included Inadequate salary, Lack of opportunity for advancement, Having to cover work for another employee, and Performing tasks not in their job description. Stressors rated as more severe by men than women included Insufficient personnel to handle an assignment, Exclusion from policy decisions, and Negativity toward the organization. These gender differences in the corporate sample may have been due, in part, to differences in organizational level associated with gender, rather than gender alone. Although the number of men and women in Spielberger and Reheiser’s (1994b) sample was almost equal, nearly twice as many men were employed at higher managerial and professional organizational levels, and more than twice as many women were working at lower clerical levels. Thus, men in the U.S. corporate sample may have reported experiencing more stress in areas relating to crisis management because they were more likely to be employed in managerial and professional positions. Greater stress reported by women relating to periods of inactivity, frequent changes from boring to demanding activities, and workload issues may have been a function of their relatively low-paid positions as clerical workers, with little opportunity for advancement. Although the first author’s (1997) research did not identify any gender-related differences in scores on the JSS scales and subscales, evidence that job stress was experienced differently by men and women was found when the ratings of specific JSS items were examined. Examination of gender-related differences in occupational stress requires consideration of the organizational level of the employee (e.g., managerial, clerical) and the social context within which men and women work. It is also important to take contextual work-related factors into account when examining gender-related differences in stress on the job. When factors associated with organizational level are controlled, genderrelated differences in job stress are less likely to be found. In a study of male and female first-line government supervisors who were closely matched on management responsibilities, number of people supervised, and years on the job, the second author (1988) found no significant gender differences in job stress. As previously noted, the first author (1997) found relatively few gender-related differences in the mean scores for individual JSS items

RT44367_C015.indd 287

7/13/05 9:35:08 AM

288

L. FIKSENBAUM AND E.R. GREENGLASS

in her sample of male and female Canadian government workers. However, the number of men and women with supervisory responsibilities in her study was nearly equal, and two thirds of the women had many diverse duties that involved having extensive contact with the public, collecting tax arrears, and negotiating payment. Thus, it seems apparent that gender-related differences in job stress are less likely to be found when women and men have similar work and organizational responsibilities. In summary, the findings reported in this chapter demonstrated that examination of responses to each JSS item identified gender differences in occupational stress that were not evident in the mean scores for the JSS scales and subscales. Analyses of the responses to individual JSS items also identified important differences in the pattern of stressors experienced by government and corporate employees. The finding that the Canadian government workers experienced greater job stress than U.S. corporate employees was not surprising, given the pressures unique to their work environment, which included Inadequate pay, Lack of opportunity for advancement, Lack of recognition for good work, and Poor working conditions (e.g., Noisy work areas, Poor or inadequate quality equipment). In addition, deteriorating working conditions and staff shortages in Canada due to recent widespread restructuring and downsizing to cut costs and reduce deficits also contributed to the greater stress and less social support experienced by the government employees, which placed them at a greater risk for illness. Job stress not only is detrimental to workers’ performance on the job, but may also have adverse effects on their psychological and physical health. Although it is important for individuals to learn how to react to and cope with stress in the workplace, microlevel interventions with individual employees are generally inadequate solutions for reducing occupational stress, which often requires institutional changes. Management can effectively address workplace stress by providing more social support for employees, modifying overly restrictive organizational policies, improving opportunities for advancement, and assisting individuals in their coping efforts. It is essential for organizations to recognize the critical role they can play in helping to alleviate or minimize job stress, thus improving the productivity and health of their employees.

REFERENCES Beehr, T. A. (1995). Psychological stress in the workplace. New York: Routledge. Beehr, T A., & Newman, J. E. (1978). Job stress, employee health, and organizational effectiveness: A facet analysis, model, and literature review. Personnel Psychology, 31, pp. 665–669. Beena, C., & Poduval, P. R. (1992). Gender differences in work stress of executives. Psychological Studies, 37, pp. 109–113. Bhagat, R. S., Allie, S. M., & Ford, D. L. (1991). Organizational stress, personal life stress, and symptoms of life strains: An inquiry into the moderating role of styles of coping. Journal of Social Behavior and Personality, 6, pp. 163–184.

RT44367_C015.indd 288

7/13/05 9:35:08 AM

JOB STRESS AND GENDER

289

Blalock, J. A., & Eisler, R. M. (1991). Masculine gender role stress: Implications for the assessment of men. Clinical Psychology Review, 11, pp. 45–60. Chen, P. Y., & Spector, P. E. (1992). Relationships of work stressors with aggression, withdrawal, theft, and substance abuse. Journal of Occupational Psychology, 65, pp. 177–184. Chesney, M. A. (1988). The evolution of coronary-prone behavior. Annals of Behavioral Medicine, 10, pp. 43–45. Cooper, C. L., & Payne, R. (Eds.). (1988). Causes, coping, and consequences of stress at work. New York: Wiley. Cooper, C. L., & Smith, M. J. (Eds.). (1985). Job stress and blue-collar work. New York: Wiley. DiSalvo, V., Lubbers, C., Rossi, A. M., & Lewis, J. (1995). Unstructured perceptions of work-related stress: An exploratory qualitative study. In R. Crandall & P. L. Perrewé (Eds.), Occupational stress: A handbook (pp. 23–28). Washington, DC: Taylor & Francis. Fiksenbaum, L. (1997). Social support and coping as moderators of work-related stress and burnout. Unpublished master’s thesis, York University, Toronto. Ganster, D. C., Fusilier, M. R., & Mayes, B. T. (1986). Role of social support in the experience of stress at work. Journal of Applied Psychology, 71, pp. 102–110. Greenglass, E. R. (1982). A world of difference: Gender roles in perspective. Toronto: Wiley. Greenglass, E. R. (1985). Psychological implications of sex bias in the workplace. Academic Psychology Bulletin, 7, pp. 227–240. Greenglass, E. R. (1988). Type-A behaviour and coping strategies in female and male supervisors. Applied Psychology: An International Review, 37, pp. 271–288. Greenglass, E. R. (1996). Anger suppression, cynical distrust, and hostility: Implications for coronary heart disease. In C. D. Spielberger & I. G. Sarason (Eds.), Stress and emotion: Anxiety, anger, and curiousity (Vol. 16, pp. 205–225). Washington, DC: Taylor & Francis. Grier, K. S. (1982). A comparison of job stress in law enforcement and teaching. Doctoral dissertation, University of South Florida, 1981. Dissertation Abstracts International, 43, p. 870B. Heart & Stroke Foundation of Canada (1995). Heart disease and stroke in Canada. Ottawa: Heart & Stroke Foundation of Canada.. Hendrix, W. H., Cantrell, R. S., & Steel, R. P. (1988). Effects of social support on the stress-burnout relationship. Journal of Business and Psychology, 3, pp. 67–73. Houston, B. K., Smith, M. A., & Cates, D. S. (1989). Hostility patterns and cardiovascular reactivity to stress. Psychophysiology, 26, pp. 337–342. Jayaratne, S., & Chess, W. A. (1984). The effects of emotional support on perceived job stress and strain. Journal of Applied Behavioral Science, 20, pp. 141–153. Jick, T. D., & Mitz, L. E. (1985). Sex differences in work stress. Academy of Management Review, 10, pp. 408–420.

RT44367_C015.indd 289

7/13/05 9:35:09 AM

290

L. FIKSENBAUM AND E.R. GREENGLASS

Kaufman, G. M., & Beehr, T. A. (1986). Interactions between job stressors and social support: Some counterintuitive results. Journal of Applied Psychology, 71, pp. 522–526. Korabik, K., McDonald, L. M., & Rosin, H. M. (1993). Stress, coping, and social support among women managers. In B. C. Long & S. E. Kahn (Eds.), Women, work, and coping (pp. 133–153). Montreal: McGill Queen’s University Press. Landsbergis, P. A. (1988). Occupational stress among health care workers: A test of the job demands-control model. Journal of Organizational Behavior, 9, pp. 217–239. Lazarus, R. S. (1966). Psychological stress and the coping process. New York: McGraw-Hill. Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal, and coping. New York: Springer. Leiter, M. P. (1991). The dream denied: Professional burnout and the constraints of service organizations. Canadian Psychology, 32, pp. 547–558. Martocchio, J. J., & O’Leary, A. M. (1989). Sex differences in occupational stress: A meta-analytic review. Journal of Applied Psychology, 74, pp. 495–501. Matthews, J. (1992). Stress Inc. CA Magazine, 125, pp. 34–37. McDonald, L. M., & Korabik, K. (1991). Sources of stress and ways of coping among male and female managers. Journal of Social Behavior and Personality, 6, pp. 185–198. McLaney, M. A., & Murrell, J. J. (1988). Control, stress, and job satisfaction in Canadian nurses. Work and Stress, 2, pp. 217–224. Nelson, D. L., & Quick, J. C. (1985). Professional women: Are distress and disease inevitable? Academy of Management Review, 10, pp. 206–213. O’Roark, A. M. (1995). Occupational stress and informed interventions. In C. D. Spielberger & I. G. Sarason (Eds.), Stress and emotion: Anxiety, anger and curiosity (Vol. 15, pp. 121–136). Washington, DC: Taylor & Francis. Perrewé, P. L. (Ed.). (1991). Handbook on job stress [Special issue]. Journal of Social Behavior and Personality, 6, pp. 1–466. Repetti, R. L., Matthews, K. A., & Waldron, I. (1989). Employment and women’s health: Effects of paid employment on women’s mental and physical health. American Psychologist, 44, pp. 1394–1401. Schnall, P. L., Deveraux, R. B., Pickering, T. G., & Schwartz, J. E. (1992). The relationship between “job strain,” workplace diastolic blood pressure, and left ventricular mass index: A correction. Journal of the American Medical Association, 267, p. 1209. Selye, H. (1993). History of the stress concept. In L. Goldberger & S. Breznitz (eds.), Handbook of stress: Theoretical and clinical aspects (2nd ed., pp. 7–17). New York: Free Press. Shirom, A. (1989). Burnout in organizations. In C. L. Cooper & I. Robertson (Eds.), International review of industrial and organizational psychology. Chichester, UK: Wiley. Smith, T. W., & Pope, M. K. (1990). Cynical hostility as a health risk: Current status and future directions. Journal of Social Behavior and Personality, 5, pp. 77–88.

RT44367_C015.indd 290

7/13/05 9:35:09 AM

JOB STRESS AND GENDER

291

Spielberger, C. D. (1994). Preliminary manual for the Job Stress Survey (JSS). Tampa: University of South Florida. Spielberger, C. D. & Reheiser, E. C. (1994a). Job stress in university, corporate, and military personnel. International Journal of Stress Management, 1, pp. 19–31. Spielberger, C. D., & Reheiser, E. C. (1994b). The Job Stress Survey: Measuring gender differences in occupational stress. Journal of Social Behavior and Personality, 9, pp. 199–218. Spielberger, C. D., & Vagg, P. R. (1999). Professional Manual for the Job Stress Survey (JSS). Odessa, FL: Psychological Assessment Resources. Spielberger, C. D., Westberry, L. G., Grier, K. S., & Greenfield, G. (1981). The Police Stress Survey: Sources of stress in law enforcement (Human Resources Institute Monograph, series 3, no. 6. Tampa: University of South Florida, College of Social Behavioral Sciences. Strube, M. J. (Ed.), (1991). Type-A behavior. Newbury Park, CA: Sage. Sutherland, V. J., & Cooper, C. L. (1990). Understanding stress: A psychological perspective for health professionals. London: Chapman & Hall. Turnage, J. J., & Spielberger, C. D. (1991). Job stress in managers, professionals, and clerical workers. Work and Stress, 5, pp. 165–176. Verbrugge, L. M. (1989). The twain meet: Empirical explanations of sex differences in health and mortality. Journal of Health and Social Behavior, 30, pp. 282–304. Vingerhoets, A. J., & Marcelissen, F. H. (1988). Stress research: Its present status and issues for future developments. Social Science and Medicine, 26, pp. 279–291.

RT44367_C015.indd 291

7/13/05 9:35:10 AM

RT44367_C015.indd 292

7/13/05 9:35:10 AM

16 Norwegian Adaptation of the Job Stress Survey Kjell Håseth University of Oslo, Norway Abstract This chapter reviews theoretical and methodological issues relating to the measurement of occupational stress and describes the procedures used in developing the Norwegian adaptation of Spielberger’s Job Stress Survey (JSS). A preliminary form of the Norwegian JSS (JSS-N) was administered to 626 workers from seven different occupational groups, varying in size from 50 to 127 subjects. The results of factor analyses showed that the factor structure of the JSS-N was similar to that of the original American JSS and that Job pressure and Lack of organizational support were the main factors. In the seven groups of Norwegian workers who were studied, substantial differences in job stress levels were found in both the perceived Severity and Frequency of occurrence of occupational stress. Police officers, nurses, and teachers reported the highest overall stress levels; bank clerks and firemen reported the lowest levels. Norwegian females perceived most of the workplace stressors assessed by the JSS-N as more stressful to their well-being than the Norwegian males.

NORWEGIAN ADAPTATION OF THE JOB STRESS SURVEY In recent years, stress in the workplace has been recognized as a major problem affecting productivity, absenteeism, and job turnover, as well as job satisfaction, psychological well-being, and workers’ health. Stress-related costs are mounting to alarming heights in many industrialized countries (Dollard & Winefield, 1996). Consequently, the prevention of further escalation of these problems is now being given a great deal of attention by management and by labor organizations. The growing public awareness and concerns about work-related stress have led to a dramatic increase in efforts to study the causes, scope, and remedies for occupational stress. In reviewing these developments, Spielberger and Vagg (1999) determined the number of studies listed in PsycLit with titles that included “job stress,” “work stress,” or “occupational stress.” From 1971 to 1992, there was more than a 20-fold increase in the number of studies of workplace-related stress in which these keywords were used. The total number of “family stress” studies also increased during this time period, but to a much lesser extent. 293

RT44367_C016.indd 293

7/13/05 9:34:34 AM

294

K. HÅSETH

Theoretical Views of Occupational Stress The Person-Environment (PE) Fit theory (French & Caplan, 1972; French, Caplan, & Harrison, 1982) has provided the most widely influential theoretical framework for research within the area of occupational stress (Chemers, Hayes, Rhodewalt, & Wysocki, 1985). This theory holds that the particular characteristics of a job may constitute a threat to an individual worker because there is a mismatch between the worker’s abilities and job demands. Stress resulting from a poor person-environment fit produces strain in the individual. If this strain persists over a long period of time, the detrimental effects include emotional distress and stress-related illnesses. In studies guided by the Person-Environment Fit theory, sources of occupational stress and stressful work-related events were investigated to determine their role in producing emotional distress and illnesses among workers in a number of occupations. The typical methodological approach was to measure cross-sectionally how stressor variables were related to adverse outcomes at the present time, or retrospectively, during the last six months. Person-Environment Fit concepts, such as “role ambiguity” and “role conflict” (Fisher & Gitelson, 1983; Jackson & Schuler, 1985), are widely used in the job stress literature. These concepts portray sources of stress at an intermediate level of abstraction in relation to the actual situation. In terms of role demands or expectations, role ambiguity was more applicable to managers and executives, whereas employees in less-responsible positions tended to perceive more role conflict. However, Jackson and Schuler (1985) were not able to verify that role concepts could differentiate between persons at different organizational levels, nor did they find that role ambiguity or conflict influenced job performance or worker satisfaction. Studies within the Person-Environment Fit theoretical framework have also investigated other job conditions leading to occupational stress, such as heavy workload, interpersonal conflict, worker dissatisfaction, and health-related symptoms (Ganster, Fusilier, & Mayes, 1972; Spector, Dwyer, & Jex, 1988). Interactions between person and environment were found to be better predictors of job stress than either of these categories alone (Chemers et al. 1985). However, Person-Environment Fit theory has recently come under severe criticism from psychologists advocating a more interactionist, person-oriented theoretical outlook (Lazarus, 1995; Perrez & Reicherts, 1992). Research based on Person-Environment Fit theory has also been criticized as static and as possibly limiting progress because the role of enduring personal characteristics and dynamic processes that mediate between goodness of fit and situational demands are neglected. Karasek (1979) has proposed a model of occupational stress that is similar in many respects to Person-Environment Fit theory. In his Demands/Control or “job latitude” model, Karasek theorizes that the relation between degree of control and job demands is the central determinant of work-related stress. According to Karesek, high control over one’s working conditions lowers the probability of stress even when job demands are high, whereas low control and high demands contribute to poorer productivity and health risks. Sauter

RT44367_C016.indd 294

7/13/05 9:34:34 AM

NORWEGIAN ADAPTATION OF THE JOB STRESS SURVEY

295

and Hurrell (1989) also noted the importance of job latitude as related to job demands and that lack of control inhibits learning and reduces motivation to change the situation in order to overcome stress. However, they recognized and identified problems in operationalizing job control as a scientific construct. Nevertheless, during the past 10 years, research has intensified in studies of individual differences in the experience of emotional control and its breakdown. Lazarus (1993) conceptualizes occupational stress as a dynamic process involving a transaction between a person and the environment, and highlights the importance of primary and secondary appraisal and coping. From this theoretical perspective, it is essential to study and understand individual patterns of stress and the negative emotions that are evoked over time and across situations, keeping in mind that appraisals of situations can change as a function of the individual’s experience with the stressors. Lazarus’ (1995) Transactional Process theoretical model has led to a number of empirical studies (e.g., Perrez & Reicherts, 1992; Schwartz & Stone, 1993). Clearly, his highly person-oriented interactional outlook appeals to many researchers within this field. However, the ideographic methodological implications of Lazarus’ theory have resulted in very few large-scale investigations that adhere to his model. Brief and George (1995) and Harris (1995) have strongly criticized the ideographic implications of Lazarus’ model. In contrast, Spielberger and Reheiser (1994) contend that the Transactional Process and Person-Environment Fit models are not incompatible viewpoints and that these theories are complementary rather than contradictory. Person-Environment Fit theory deals with how stressors produce job strain, which affects a variety of dependent variables. Lazarus’ model, which is broader in scope, is based on a more general theory of how individuals experience and deal with their emotions and on the long-term effects that emotions have on individual adjustment and reactions to the environment. More recently, Lazarus (1995) has proposed that the concept of stress can be dispensed with altogether and replaced by focusing on the person’s primary and secondary appraisals of specific stressors, or combinations of stressors, that evoke negative emotions. Similarly, Spielberger (1994) suggests that occupational stress can best be measured by asking workers about the intensity and frequency of their reactions to specific stressors and general working conditions. This approach is also advocated by other stress researchers (e.g., Cooper, 1981). The Job Stress Survey The Job Stress Survey (JSS) was constructed to measure occupational stress as conceptualized within Spielberger’s (1983) state-trait model of stress and emotional experience (Spielberger, 1994; Spielberger & Vagg, 1999). In describing the evocation of negative emotions, such as anxiety and anger, this model is similar to, but not identical with, Lazarus’ (1993) Transactional Process theory of stress and emotion. The main difference is that in developing the JSS, Spielberger (1994) identified a generic set of prototypical situations

RT44367_C016.indd 295

7/13/05 9:34:34 AM

296

K. HÅSETH

that evoked occupational stress, whereas Lazarus (1995) advocates a more comprehensive person-specific stress process that focuses on the appraisal of sources of stress in the workplace. The 30-item JSS was “designed to assess the perceived severity (intensity) and frequency of occurrence of working conditions that are likely to affect, adversely, the well-being of employees who are exposed to them” (Spielberger & Reheiser, 1994, p. 207). The individual JSS items describe work-related events and situations that are encountered in a variety of occupations (e.g., Working overtime, Excessive paperwork). Because the sources of stress described by each JSS item are found in many occupations, they represent a core set of stressors, which makes the JSS well suited to compare stress levels in a wide variety of work settings. In Part 1 of the JSS, the format for responding to the Severity Rating Scale is similar to the rating procedure that was used in the Holmes-Rahe (1967) Social Readjustment Rating Scale. Respondents are asked to rate the perceived severity of the 30 JSS stressors on a nine-point scale by comparing each individual work-related stressor event with a standard, Assignment of disagreeable duties, which was given a constant midpoint scale value of 5. In early research on occupational stress with the JSS, this item was assigned ratings in the middle of the range of stress severity by both police officers and teachers (Spielberger, 1994; Spielberger & Vagg, 1999). Examinees respond to Part 2 of the JSS Frequency Scale by indicating the number of days they experienced each stressor event in their workplace during the past six months, with a score range of 0 to 9+ days. This rating format corresponds with that of the STAI Trait Anxiety Scale (Spielberger, 1983), but it differs in specifying a particular time period. The JSS Stress Index, which assesses the overall level of occupational stress experienced by an individual, is based on the Index scores of all 30 items. This Index score for each item is computed by multiplying the item’s Severity score by its Frequency score to assess the amount of job stress experienced by the examinees in the specific situations described by the JSS items. The overall Stress Index score is then computed by summing the Index scores for all 30 items. Although the format for the JSS Severity rating scale corresponds with the procedure used to assess state anxiety with the STAI (Spielberger, 1983), the Severity scale should not be regarded as a measure of the intensity of occupational stress. Unlike measures of emotional states, the JSS Severity scale does not require examinees to report the intensity of their emotional experience in either actual or hypothetical situations. Rather, respondents are instructed to report their perceptions of the potential of each stressor event to activate negative feelings. In addition to the JSS Stress Index and Severity and Frequency scores that are based on all 30 JSS items, the JSS includes the 10-item Job pressure (JP) and Lack of organizational support (LS) subscales that were developed on the basis of factor analyses of the Severity and Frequency items (Spielberger, 1994; Spielberger & Vagg, 1999). For these subscales, JP and LS Stress Index, Severity, and Frequency scores can be computed in the same manner as for the 30-item JSS scales.

RT44367_C016.indd 296

7/13/05 9:34:34 AM

NORWEGIAN ADAPTATION OF THE JOB STRESS SURVEY

297

THE NORWEGIAN JOB STRESS SURVEY The adverse effects of stressful workplace conditions on physiological, psychological, and social variables have been demonstrated in several Norwegian investigations in which a variety of measures were utilized to assess worker characteristics, such as job satisfaction, health, social support, and interpersonal conflict (Mykletun, 1995; Ursin, Endresen, & Ursin, 1988; Westgaard & Jansen, 1992). The measures used in these studies were translations of wellknown international inventories or were constructed to test specific hypotheses put forward by the researchers. The Cooper (1981) Stress Check was an important early contribution to the assessment of occupational stress in Norway. This 22-item questionnaire was translated and used in a number of studies by Holger Ursin and his colleagues (Ursin et al., 1988) and in research by other Norwegian investigators (Mykletun, 1995; Westgaard & Jansen, 1992) in which its psychometric properties were reasonably well established. However, the Cooper Stress Check has been rarely used in international research, and the original English form of this measure was subsequently replaced by the Occupational Stress Inventory (OSI; Cooper, Sloan, & Williams, 1988). Consequently, the time in Norway seemed ripe for developing or adapting a job stress inventory for use in future stress audits of troubled organizations and to facilitate comparisons of the results of Norwegian investigations with job stress studies in other countries. Developing the Norwegian JSS The items from the American JSS were translated into Norwegian, using graduate students as an audit group. Care was taken to construct items that described, as clearly as possible, the underlying meaning of the concepts that guided the construction of the original JSS items, rather than literal verbal translations. It was our impression that items that described the JSS stressor events were easier to translate than those found in most self-report instruments that measure emotional experience, such as anger and anxiety. The reason for this seemed to be that the JSS items provide very brief descriptions (two to three words) of specific situations and happenings that are frequently encountered in a variety of workplaces. Two JSS items were given alternative translations to clarify their meaning in the Norwegian language. Item 13, Difficulty getting along with supervisor, was translated as “Vanskeligheter med å komme overens med overordnet,” an approximate verbatim translation of the original item, and as “Samarbeidsproblemer med overordnet” (“Difficulties working with a supervisor”). Item 17, Personal insult from customer/consumer/colleague, was split into two items, which were translated as “Ubehageligheter fra arbeidskollegaer” (“Personal insult from colleagues”) and “Ubehageligheter fra brukergruppe, rettet mot egen person” (“Personal insult from customer/consumer user group”). To gain insight into the relevance of the translations of the entire set of JSS items, the 32-item experimental Norwegian JSS (JSS-N) was administered to

RT44367_C016.indd 297

7/13/05 9:34:35 AM

298

K. HÅSETH

a convenience sample of 30 regularly employed professionals (10 females, mean age = 35.3; 20 males, mean age = 38.1). In responding to Part 1 of the experimental JSS-N, the subjects were instructed to report the severity of each stressor on a nine-point intensity scale without comparing the perceived severity to a standard item. This allowed us to check on whether the designated standard in the original JSS, Assignment of disagreeable duties, would be rated in the middle range of the intensity scale by Norwegian subjects. The midscale values assigned to the standard item by the Norwegian subjects indicated that this item could also serve as the standard for anchoring the Severity ratings in the Norwegian adaptation of the JSS. In responding to Part 1 of the revised 32-item experimental Norwegian JSS-N, the subjects were instructed to report the perceived severity of the JSS-N items in the same manner as these items were rated in the original American JSS, by comparing each item with the standard stressor. Each item was rated in Part 2 of the revised experimental JSS-N on the same 10-point Frequency scale that was used with the American JSS to indicate the number of days on which the stressor event occurred during the past six months, with a score range of 0 to 9+ days. To further establish the relevance of the Norwegian JSS stressor items, the subjects were interviewed to determine which stressors they had actually encountered in their daily work; most of them reported that they had experienced all 32 stressors. Two additional major sources of stress were also frequently mentioned by Norwegian workers that were not included in the American JSS: stressful events relating to role ambiguity and role conflict. One new item relevant to each of these areas was constructed and included in the preliminary Norwegian JSS-N, which then totaled 34 items. The new role ambiguity and role conflict items were “Usikkerhet med hensyn til hvilke oppgaver jeg har ansvaret for” (“Uncertainty about my responsibility”) and “Motstridende forventninger til arbeidsoppgaver jeg skal ivareta” (“Conflicting expectations regarding my work”). The preliminary 34-item JSS-N was administered to 626 subjects (299 females; 327 males) who were currently employed in a variety of occupations, along with a demographic questionnaire and several other psychological tests. Subjects responded anonymously to these questionnaires on a voluntary basis. The total sample consisted of 127 police officers; 50 firefighters; 50 bus drivers; 121 bank clerks; 73 teachers; 105 nurses; 17 librarians; 30 medical, legal, and mental health professionals; and 53 managers of government agencies. The mean age for the total group of subjects was 40.56 (SD = 9.25). There was no significant age difference between the females (M = 40.76, SD = 9.93) and males (M = 40.47, SD = 8.32). The Factor Structure and Psychometric Properties of the JSS-N The factor structure for the 34 Severity and Frequency items that were included in the preliminary JSS-N were evaluated in separate factor analyses for males and females, using principal-components factor analyses. The number of extracted factors was based on eigenvalues greater than 1.00 and Cattell’s (1966) scree test and breaks criterion. Two large factors and two or

RT44367_C016.indd 298

7/13/05 9:34:35 AM

299

NORWEGIAN ADAPTATION OF THE JOB STRESS SURVEY Table 16.1 Means, Standard Deviations, and Alpha Coefficients for the JSS-N Stress Index, Severity, and Frequency Scales for Males and Females, and F-Tests for Gender Differences JSS Stress Scales

Males

Females

F-Test (Sig.)

Stress Index JS-X Mean SD Alpha N

17.49 10.12 322 .90

17.86 9.85 289 .91

F = 0.21

Severity JS-S Mean SD Alpha N

4.52 1.11 326 .90

4.73 1.20 297 .91

F = 5.10*

Frequency JS-F Mean SD Alpha N

3.39 1.58 326 .91

3.32 1.55 289 .90

F = 0.33

Note: *p < .05.

three relatively weak factors were initially extracted, which were similar to the factors identified in the American JSS (Spielberger, 1994; Spielberger & Vagg, 1999). Taking into account the scree test and the breaks criterion, two strong factors were extracted, using varimax rotation; essentially the same two factors were found for both males and females. Based on the results of the factor analyses of the individual Severity and Frequency items, 4 of the 34 preliminary Norwegian JSS items with dual or nonsalient (p < .30) factor loadings were eliminated. The means, standard deviations, and alpha coefficients for the 30-item JSS-N Index, Severity, and Frequency scores for females and males in the total sample are reported in Table 16.1. The internal consistency reliability of the Norwegian JSS, as indicated by the alpha coefficients of .90 and .91 for both females and males, were similar to those reported by Spielberger and Vagg (1999). The mean JSS-N Severity (JS-S) score for the females was significantly higher than the corresponding score for males (F = 5.10, p < .05). Gender differences for the JSS-N Stress Index and Frequency scores were minimal. Although significant gender differences were found for 10 of the 30 JSS-N items, these results are not reported because the design of this study did not permit a clear differentiation between gender and occupational group differences. Significant gender differences were also found for U.S. workers on a number of individual items. The only gender difference that was found in the JSS scale scores of U.S. workers based on all 30 items was that women employed in private corporations scored higher on the Frequency scale than men (Spielberger & Reheiser, 1994).

RT44367_C016.indd 299

7/13/05 9:34:36 AM

300

K. HÅSETH

Table 16.2 Factor Structure for the 30 JSS-N Severity and Frequency Items for Males and Females Severity Support No.

Items

M

F

3 13 17 12 30 21 29 6 19 18 5 22 20 28 8 15 14 4 24 11 16 9 26 27 2 23 25 7 10 1

Difficult working with a supervisor Difficult getting along with a supervisor Personal insult from colleagues Personal insult from user group Conflicts with other departments Poor/inadequate supervision Poorly motivated coworkers Inadequate support from superior Uncertainty about responsibility Lack of participation about decisions Fellow workers not doing job Conflicting work tasks Competition for advancement Covering work for others Lack of recognition for good work Insufficient personnel to handle tasks Negative attitudes toward organization Assignment of new/unfamiliar duties Freq. changes boring to demanding tasks Increased responsibility Make on-the-spot decisions Perform tasks not in job description Meeting deadlines Insufficient personal time Working overtime Frequent interruptions Excessive paperwork Dealing with crisis situations Inadequate equipment Assignment of disagreeable duties

81 77 72 70 68 67 65 61 60 59 54 54 42 42 38 37 36

64 79 80 66 61 76 70 67 70 51 55 66 41 38 32 31 (25)

Eigenvalues % of variance

Frequency

Pressure M

38

32 68 67 66 63 62 62 58 54 49 44 44 36 —

7.98 8.59 2.91 27.5 29.6 10.1

F

34

(25) 45 63 68 52 56 49 58 53 56 61 (26) 49 —

Support M

F

83 84 60 48 37 76 45 84 33 50 31

87 86 52 (27) 43 83 46 83 51 53 40 53 50

56 67 65 44

Pressure M

F

46 40 52

32 62

63 45

45

62 79 46 70 74 49 44 65 73 76 (28) (29) 55

39 76 48 65 58 55 63 53 69 65 49 33 49

31

2.58 3.53 7.86 8.24 3.37 8.9 11.8 26.2 27.5 11.2

The factor structure of the item scores for the final set of 30 JSS-N Severity and Frequency items was evaluated in separate analyses for men and women, using principal-components factor analyses with varimax rotation. The results of these analyses are presented in Table 16.2. The same two Severity and Frequency factors were identified for both sexes. In keeping with Spielberger and Vagg (1999), these factors were named Lack of organizational support and

RT44367_C016.indd 300

7/13/05 9:34:37 AM

NORWEGIAN ADAPTATION OF THE JOB STRESS SURVEY

301

Job pressure. Given the consistency in the factor structure of the Norwegian JSS-N and the American JSS, and the content of the items with the strongest loadings on the two factors, the factor names were considered to be both descriptive and appropriate. Of the 29 JSS-N Severity items, 26 had strong loadings for both sexes on only one factor and no dual loadings, as may be noted in Table 16.2. Of these 26 items, 15 had salient loadings of .30 or higher on the Lack of organizational support factor; 11 items had salient loadings on the Job pressure factor. Only three of the 29 Severity items had any salient dual loadings. Of the 30 Frequency items, 23 had salient loadings on only one factor and no dual loadings, 11 on Lack of organizational support and 12 on Job pressure. The Severity and Frequency items, with consistent salient loadings on either the Support or the Pressure factor for both females and males, were selected for the JSS-N 10-item Lack of support and Job pressure subscales. The JSS-N Lack of Organizational Support and Job Pressure Subscales In order to further verify the factor structure of the JSS-N, the scores for the 20 Severity and 20 Frequency items selected for the Lack of organizational support and Job pressure subscales were factored in separate analyses for males and females, using principal-components factor analyses with varimax rotation. The results of these analyses for the two factors that were extracted are reported in Table 16.3, in which it can be noted that the Lack of organizational support factor was stronger for both sexes and that all 20 Severity and Frequency items had salient loadings for both males and females on the appropriate factor, with no dual salient loadings. Eight of the 10 JSS-N Lack of organizational support items and 9 of the 10 Job pressure items were essentially the same as the items comprising the corresponding American JSS subscales. The three Norwegian JSS-N subscale items that differed from the items in the American JSS subscales are marked with an asterisk in Table 16.3. The means, standard deviations, and alpha coefficients for the JSS-N Lack of organizational support and Job Pressure Indexes, Severity, and Frequency subscale scores for males and females are reported in Table 16.4. The alpha coefficients for the three Lack of organizational support subscales, LS-X, LS-S, and LS-F, varied from .82 to .89; the alphas for the Job pressure subscales were slightly lower, varying from .79 to .84. Although the alpha coefficients for the 10-item JSS-N subscales were slightly lower than those of .90 or higher for the JSS-N Index, Severity, and Frequency scales, which were based on all 30 items (see Table 16.1), the internal consistency reliabilities for the LS and JP subscales were considered satisfactory. Males had significantly higher scores than females on the LS-X Lack of Support Index subscale, whereas females had significantly higher scores than males on the JP-X Job Pressure Index subscale. Females also tended to perceive job pressures as more stressful than males, as indicated by their significantly higher scores on the Job pressure Severity (JP-S) subscale, whereas men had substantially higher Lack of organizational support Frequency (LS-F) scores than women. The two sexes did not differ in their LS-S Severity and

RT44367_C016.indd 301

7/13/05 9:34:37 AM

302

K. HÅSETH

Table 16.3 Factor Structure for Males and Females of the JSS-N Severity and Frequency Items That Are Included in the Lack of Organizational Support and Job Pressure Subscales Severity Support No.

Items

Lack of Organizational Support (LS-S) Scale 3 Difficult working with a supervisor 6 Inadequate support from superior 8 Lack of recognition for good work 13 Difficult getting along with supervisor 17 Personal insult from colleagues 18 Lack of participation in decisions 20* Competition for advancement 21 Poor/inadequate supervision 29 Poorly motivated coworkers 30* Conflicts with other departments Job Pressure (JP) Scale 2 Working overtime 4 Assignment of new/unfamiliar duties 9 Perform tasks not in job description 11 Increased responsibility 16 Make on the spot decisions 23* Frequent interruptions 24 Var. in boring/demanding tasks 25 Excessive paperwork 26 Meeting deadlines 27 Insufficient personal time Eigenvalues % of variance

M

F

85 65 45 83 68 58 43 72 62 58

88 72 42 82 78 57 41 81 70 64

Frequency

Pressure M

55 70 60 67 64 49 65 48 61 55

F

Support M

F

87 86 64 86 61 52 58 77 47 37

83 83 62 85 55 48 52 80 45 40

55 54 57 69 50 55 68 62 51 56

Pressure M

F

64 60 69 50 68 74 76 72 49 47

55 47 55 45 60 72 78 64 60 62

5.59 6.09 2.64 2.46 5.77 5.32 3.24 2.99 28.0 30.5 13.2 12.3 28.9 26.6 16.2 15.0

Note: *JSS-N items that are not included in the corresponding American JSS subscales.

JP-F Frequency subscale scores. Thus, the perceived Severity of Job pressure had a greater impact on the occupational stress experienced by women, whereas the Frequency that Lack of organizational support was encountered by men was more influential. Intercorrelations among the JSS Scales and Subscales for Males and Females The intercorrelations among the JSS-N scales and subscales are reported for females and males, respectively, in the upper and lower parts of Table 16.5. The largest correlations, .93 for females and .90 for males, were found between

RT44367_C016.indd 302

7/13/05 9:34:38 AM

303

NORWEGIAN ADAPTATION OF THE JOB STRESS SURVEY

Table 16.4 Means, Standard Deviations, and Alpha Coefficients for the JSS-N Stress Index, and Severity and Frequency Subscales for Males and Females and F-Tests for Gender Differences JSS Stress Scales

Males

Females

F-Test (Sig.)

Stress Index LS-X Mean SD Alpha JP-X Mean SD Alpha

13.78 12.90 .84 22.00 12.31 .82

11.22 11.44 .82 24.62 13.50 .81

F = 6.66**

Stress Severity LS-S Mean SD Alpha JP-S Mean SD Alpha

4.66 1.51 .86 4.29 1.27 .81

4.74 1.69 .89 4.57 1.34 .79

F = 0.39

Stress Frequency LS-F Mean SD Alpha JP-F Mean SD Alpha

2.40 1.84 .87 4.74 2.20 .84

1.96 1.73 .85 4.91 2.13 .83

F = 9.00**

F = 6.30**

F = 7.17**

F = 0.88

Note: *p < .05; **p < .01.

the JS-X Stress Index and JS-F Frequency scales, which were based on all 30 JSS items. These correlations were influenced primarily by the Frequency that Lack of organizational support was encountered in the workplace, as indicated by the very strong correlations of .96 between the LS-X and LS-F subscales for both sexes. Thus, the Frequency that the respondents reported experiencing Lack of organizational support had a greater impact on overall stress levels for both sexes than the perceived Severity of the stressor events. The moderate correlations between the JS-S Severity and JS-F Frequency scales, which were .45 for females and .48 for males, may be interpreted as providing evidence that these scales were measuring somewhat different, but related, sources of occupational stress. VALIDITY OF THE NORWEGIAN JSS-N The Norwegian JSS-N was administered to three occupational groups along with two other measures of occupational stress, the Norwegian translations

RT44367_C016.indd 303

7/13/05 9:34:38 AM

304

K. HÅSETH

Table 16.5 Intercorrelations between the JSS-N Scales and Subscales for Norwegian Females and Males Jobs Stress JS-X JS-X JS-S JS-F JP-X JP-S JP-F LS-X LS-S LS-F

70 90 82 58 65 82 57 78

Job Pressure

Lack of Support

JS-S

JS-F

JP-X

JP-S

JP-F

LS-X

LS-S

LS-F

61

93 45

82 55 75

60 75 40 74

71 38 81 84 39

75 42 71 32 25 27

46 88 37 29 41 26 49

75 32 76 34 23 31 96 37

48 60 75 32 57 85 47

75 30 82 72 46 74

68 19 44 36 39

19 35 37 29

30 29 30

63 96

53

Note: Correlations for the females are reported in the upper parts of the table. For males, the corresponding correlations are reported in the lower part of the table.

Table 16.6 Intercorrelations of the JSS-N Scales and Subscales with Other Measures of Occupational Stress for Police Officers, Bank Clerks, and Teachers JSS-N Job Stress Scale

Support

Pressure

Groups

Measure*

JS-X

JS-S

JS-F

LS-X

JP-X

Police Bank Clerks Teachers

CSC CSC MSC

.51 .68 .64

.45 .48 .42

.48 .54 .57

.50 .57 .48

.38 .55 .58

Note: *CSC = Cooper Stress Check; MSC = Mykletun Stress Check.

of the Cooper (1981) Stress Check (CSC-N) and the Mykletun (1995) Stress Check (MSC-N), which was adapted from the CSC to assess the job stress experienced by teachers. The JSS-N and CSC-N were administered to 127 police officers and 123 bank clerks. The JSS-N and the MSC-N were administered by Nilsen and Schøyen (1998) to 73 Norwegian teachers. The correlations of the CSC-N and MSC-N with the 30-item JS-X Stress Index, JS-S Severity, and JS-F Frequency scales, and the 10-item JP-X Job Pressure and LS-X Lack of Support Index subscales, are reported in Table 16.6. The largest correlations of the JSS-N scales and subscales with the CSC-N for both police officers and bank clerks were found with the JS-X Stress Index scale. The correlations of the CSC-N with the JS-F Frequency scale were slightly larger for both groups than those for the JS-S Severity scale. For police officers, the finding that CSC-N scores correlated more highly with the LS-X Lack of Support Index than with the JP-X Job Pressure Index suggested that Lack of support contributed more than Job pressure to overall stress levels. These findings were consistent with the results of the factor analyses

RT44367_C016.indd 304

7/13/05 9:34:39 AM

305

NORWEGIAN ADAPTATION OF THE JOB STRESS SURVEY Table 16.7 Intercorrelations of the JSS-N Scales and Lack of Support and Job Pressure Index Scores with Measures of Emotions, Personality, and Health JSS-N Job Stress Scales STPI Scales

N

JS-X

JS-S

JS-F

S-Anxiety S-Anger S-Depression S-Curiosity T-Anxiety T-Anger T-Depression T-Curiosity

433 53 53 53 433 53 53 53

.20*** .08 –.13 –.02 .12*** .08 .16* .20*

.06 .11 .12 –.07 .05 .17* .16* .07

.13** .12 –.11 –.04 .04 .08 .11 .21*

Health-Intensity Health-Frequency

290 290

.43*** .26***

.30** .14*

.36*** .21***

Lack of Support

Job Pressure

LS-X

JP-X

.17*** .24** –.09 .00 .09* .07 .24** .10

.15** –.09 –.04 –.09 .11** .10 .16* .15*

.39*** .34***

.29** .09

Note: *p < .05; **p < .01; ***p < .001.

of the CSC-N items reported by Nilsen and Schøyen (1998), which indicated that three of the four CSC factors (leadership problems, social interaction, career opportunities) were related to organizational support. The strongest correlation of the MSC-N scale for teachers was also found with the JS-X Stress Index. The correlations of the MSC-N with the JS-F Frequency scale and Job Pressure Index were larger than those found for this measure with the JS-S Severity scale and Lack of Support Index. These findings suggested that the Frequency that Job pressures were experienced by the teachers contributed more to their overall level of occupational stress than Lack of support. The moderate-to-strong positive correlations of the JSS-N scales with the CSC and MSC demonstrated reasonably good concurrent validity and identified possible differences in the effects of Lack of support and Job pressure on the stress levels of different occupational groups. To further investigate the relation between occupational stress and emotions, personality, and health, the JSS-N and the Norwegian adaptation of Spielberger’s (1979) State-Trait Personality Inventory (STPI-N; Håseth & Spielberger, 2000) were administered to 53 police officers. The JSS-N and the STPI-N State (S-Anxiety) and Trait (T-Anxiety) scales were also administered to a heterogeneous sample of 380 Norwegian workers. The correlations of the STPI-N State and Trait Anxiety, Anger, Depression, and Curiosity scales with the JSS-N Stress Index, Severity, and Frequency scales, and the Lack of Support and Job Pressure Index scores are reported in Table 16.7. In computing the correlations for the S-Anxiety and T-Anxiety scores, the data for the police officers were combined with those of the other workers; the results for the total group (N = 433) are reported in Table 16.7.

RT44367_C016.indd 305

7/13/05 9:34:39 AM

306

K. HÅSETH

Small but significant positive correlations were found for the STPI-N S-Anxiety scores with the 30-item JS-X Stress Index and JS-F Frequency scales, and the 10-item Lack of Support and Job Pressure Index scores. A significant positive correlation was also found for the S-Anger scale with the LS-X Lack of Support Index scores. These findings suggested that participants who were more anxious or more angry at the time they responded to the questionnaires tended to perceive Lack of support as somewhat more stressful and that higher S-Anxiety was also related to the more frequent experience of job pressures. No significant correlations were found for the STPI-N State Depression and State Curiosity scales with any of the JSS-N scales and subscales. The STPI-N T-Anxiety Scale correlated positively and significantly with the JS-X Stress Index and the Lack of Support and Job Pressure Index scales; the T-Anger scale correlated significantly with the JSS-N Severity Scale. Small but significant positive correlations were found between the T-Depression scale and all of the JSS-N scales except JS-F Frequency. Significant positive correlations were also found for the T-Curiosity scale with the JS-X Stress Index and JS-F Frequency scales, and with the JP-X Job Pressure Index. These findings suggested that feeling depressed was associated with perceived Lack of support and greater occupational stress, and that persons with high T-Curiosity may engage in activities that lead them to experience occupational stress more frequently in relation to Job pressures. In another study, the JSS-N and the Ursin et al., (1988) Health Inventory (UHI) were administered to 290 teachers, bank clerks, and nurses. The UHI inquires about the intensity (Health-Intensity) and frequency (Health-Frequency) of discomfort associated with 29 common physical symptoms during a 30-day period prior to taking the test. The correlations of all the JSS-N scales and subscales with Health-Intensity were positive and significant. Somewhat smaller significant positive correlations were also found for four of the five JSS-N scales with Health-Frequency. The consistent significant moderate positive correlations of the LS-X Lack of Support Index with both Health-Intensity and Health-Frequency suggested that Lack of organizational support may be the most important health-related source of stress measured by the JSS-N. Occupational Group Differences in JSS-N Scores Table 16.8 reports the results of comparisons of the mean JSS Index, Severity, and Frequency scores for seven groups of Norwegian workers. In one-way analyses of variance, highly significant F-tests of differences between these groups were found for the three JSS-N scales that are based on all 30 items: JS-X (F = 8.07; p < .001); JS-S (F = 6.71; p < .000); JS-F (F = 10.77; p < .001). Duncan Multiple Range tests of differences between the seven groups on each of the JSS-N scales are also reported in Table 16.8. Police officers had the highest mean JS-X Stress Index scores; the Stress Index scores of the firefighters were much lower than those of the other groups. Nurses, teachers, and police officers had the highest JS-S Severity scores; firefighters had the lowest scores on this measure. The police officers and managers had the highest scores on the JS-F Frequency scale, and here,

RT44367_C016.indd 306

7/13/05 9:34:40 AM

307

NORWEGIAN ADAPTATION OF THE JOB STRESS SURVEY Table 16.8 Means and Standard Deviations of the JSS Index, Severity, and Frequency Scores for Seven Groups of Norwegian Employees and Significant Differences between these Groups Index Group

Severity

Frequency

N

M

SD

Sig.

M

SD

Sig.

M

SD

Sig.

50 121 50 53

10.64 16.10 16.26 17.87

6.11 9.00 9.57 6.49

A B BC BC

4.06 4.54 4.44 4.46

1.05 1.14 1.10 0.91

A B AB AB

2.22 3.23 2.80 3.96

1.13 1.51 1.47 0.93

A B AB CD

73

18.08 10.12

BC

5.06

1.31

C

3.09

1.42

B

Nurses

105

19.26 11.99

C

4.59

1.31

B

3.44

1.84

BD

Police Officers

127

21.57 10.15

D

5.00

0.92

C

4.00

1.58

C

Firefighters Bank Clerks Bus Drivers Managers Teachers

Note: Means with the same letter are not significantly different.

again, the firefighters had the lowest scores. In the comparisons of the seven groups of employees, the most consistent findings were that police officers had high scores on all three JSS-N scales and that the scores of the firefighters were substantially lower than those of the other occupational groups. DISCUSSION The factor structure and psychometric properties of the Norwegian adaptation of the JSS closely resembled those of the original American inventory (Spielberger, 1994; Spielberger & Vagg, 1999). However, in adapting a test instrument for use in another language or culture, it is generally necessary to make a number of small changes to bring the adapted instrument in line with local traditions and language habits, and this was required with the Norwegian JSS-N. It should be noted, however, that the translation of the JSS was, in many ways, easier than the translation of items with more abstract and general content, such as the anxiety items in Spielberger’s (1983) State-Trait Anxiety Inventory. The JSS items describe specific situations and events that are frequently encountered in a wide variety of work settings in many different countries. Although care was taken to keep this same level of specificity in developing the Norwegian adaptation, two stressor items were constructed and added to the JSS-N on the basis of information obtained in qualitative analyses of the workplace stressors that were frequently reported in interviews with Norwegian workers. These items described role ambiguity and role conflict, the more general and abstract stressors associated with the traditional sources of occupational stress identified by researchers working within the PE Fit tradition (Fisher & Gitelson,

RT44367_C016.indd 307

7/13/05 9:34:40 AM

308

K. HÅSETH

1983; Jackson & Schuler, 1985). The role ambiguity and conflict items had high item-total correlations for both the JSS-N Severity and Frequency scales, and were retained in the Norwegian JSS-N. However, factor analyses showed that these items had unclear factor loadings and could not consistently be classified as sources of stress resulting from Lack of organizational support or Job pressure. The addition of new items to measures of occupational stress raises the question of how many specific sources of job stress can be identified and included in these measures. It is obviously not possible to include all stressproducing situations that are encountered in a variety of work settings. The factor analyses of the Norwegian JSS-N, reported in Tables 16.2 and 16.3, clearly identified two major sources of occupational stress, which were essentially the same as the Lack of organizational support and Job pressure factors that were found in factor analyses of the American JSS (Spielberger & Reheiser, 1994; Spielberger & Vagg, 1999). Given that the Norwegian and American studies covered a multitude of occupations, it seems reasonable to conclude that these two broad sources of occupational stress are represented in most workplaces that employ managerial, professional, clerical, and service personnel. Items that describe Job pressures and Organizational support are well represented in several well-known test instruments that were developed to measure occupational stress. In his Work Stress Inventory (WSI), Barone (1995) included numerous Lack of organizational support and Job pressure items to assess “organizational stress,” but did not differentiate between these concepts. Barone also identified Job risk as an important component of occupational stress, which he assessed with items that described physical threats to self and others. In Cooper and Marshall’s (1976, 1978) theoretical model of occupational stress, which has been supported by research with the Occupational Stress Indicator (Cooper et al., 1988), six sources of job stress were identified that are measured by OSI scales. These scales include stressors that were intrinsic to the (1) job, (2) managerial role, (3) relationships with others, (4) career and achievement, (5) organizational structure and climate, and (6) home/work interface. Job pressures are clearly intrinsic to a particular job; Lack of support is assessed by OSI scales 2, 3, 4, and 5. To evaluate the concurrent validity of the JSS-N in the present study, the Cooper Stress Check, a forerunner of the OSI constructed on the basis of Cooper’s theoretical model, was administered to Norwegian workers along with the JSS-N. The moderately high correlations that were found between the CSC and the JSS-N scales and subscales indicated a great deal of overlap between the two major sources of stress identified by the JSS and those assessed by the CSC. Job pressure and Organizational support as major sources of occupational stress have also been substantiated in studies using qualitative methods. DiSalvo, Lubbers, Rossi, and Lewis (1995) used open-ended reports from groups of professionals to evaluate their respondents’ awareness of conditions in their work settings that created stress in their everyday work. In using a Flanagan (1955) procedure to classify their subjects’ spontaneous responses, they identified two broad categories of occupational stress: “work content,”

RT44367_C016.indd 308

7/13/05 9:34:41 AM

NORWEGIAN ADAPTATION OF THE JOB STRESS SURVEY

309

which was similar to Job pressure, and “work context,” which was almost identical with Organizational support. Kasl (1998) has pointed out that the two basic sources of occupational stress that are measured by the JSS-N items are likely to be close to what workers themselves consider as phenomenologically important. However, further research is needed to determine if these generic phenomenologically important sources of stress can be found in a variety of occupations in different countries and cultures and to explore whether or not the two major types of occupational stress assessed by the JSS should be expanded to include additional sources of stress that are encountered by female and male workers in a wide range of work settings. REFERENCES Barone, D. F. (1995). Developing a transactional psychology of work stress. In R. Crandall & P. L. Perrewé (Eds.), Occupational stress: A handbook (pp. 29–37). New York: Taylor & Francis. Brief, A. P., & George, J. M. (1995). Psychological stress in the workplace: A brief comment on Lazarus’ outlook. In R. Crandall & P. L. Perrewé (Eds.), Occupational stress: A handbook (pp. 3–14). New York: Taylor & Francis. Cattell, R. B. (1966). The meaning and strategic use of factor analysis. In R. B. Cattell (Ed.), Handbook of multivariate experimental psychology. Skokie, IL: Rand McNally. Chemers, M. M., Hayes, R. B., Rhodewalt, F., & Wysocki, J. (1985). A personenvironment analysis of job stress: A contingency model explanation. Journal of Personality and Social Psychology, 49, pp. 628–635. Cooper, C. L. (1981). The stress check. Englewood Cliffs, NJ: Prentice Hall. Cooper, C. L., & Marshall, J. (1976). Occupational sources of stress. Journal of Occupational Psychology, 49, pp. 11–28. Cooper, C. L., & Marshall, J. (1978). Understanding executive stress. London: Macmillan. Cooper, C. L., Sloan, S. G., & Williams, S. (1988). The Occupational Stress Indicator: Management guide. Windsor, UK: NFER-Nelson. DiSalvo,V., Lubbers, C., Rossi, A. M., & Lewis, J. (1995). The impact of gender on work-related stress. In R. Crandall & P. L. Perrewé (Eds.), Occupational stress: A handbook (pp. 39–50). New York: Taylor & Francis. Dollard, M. F., & Winefield, A. H. (1996). Managing occupational stress: A national and international perspective. International Journal of Stress Management, 3, pp. 69–90. Fisher, C. D., & Gitelson, R. (1983). A meta-analysis of the correlates of role conflict and ambiguity. Journal of Applied Psychology, 68, pp. 320–333. Flanagan, J. C. (1955). The critical incident technique. Psychological Bulletin, 51, pp. 327–358. French, J. R. P., Jr., & Caplan, R. D. (1972). Occupational stress and individual strain. In A. J. Marrow (Ed.), The failure of success (pp. 30–66). New York: Amacom.

RT44367_C016.indd 309

7/13/05 9:34:41 AM

310

K. HÅSETH

French, J. R. P., Jr., Caplan, R. D., & Harrison, R. V. (1982). The mechanisms of job stress and strain. Chichester, UK: Wiley. Ganster, D. C., Fusilier, M. R., & Mayes, B. T. (1972). Role of social support in the experience of stress at work. Journal of Applied Psychology, 71, pp. 102–110. Harris, J. R. (1995). The utility of the transaction approach for occupational stress research. In R. Crandall & P. L. Perrewé (Eds.), Occupational stress: A handbook (pp. 21–28). New York: Taylor & Francis. Håseth, K., & Spielberger, C. D. (2000). Forelopig manual for Deb Horske Tilstand-Trekk Personlighets Inventorium [Preliminary manual for the Norwegian State-Trait Personality Inventory]. Unpublished manuscript, Department of Psychology, University of Oslo, Norway. Holmes, T. H., & Rahe, R. H. (1967). The Social Readjustment Rating Scale: A cross-cultural study of Western Europeans and Americans. Journal of Psychosomatic Research, 14, pp. 391–400. Jackson, S. E., & Schuler, R. S. (1985). A meta-analysis and conceptual critique of research on role ambiguity and role conflict in work settings. Organizational Behavior and Human Decision Process, 36, pp. 16–78. Karasek, R. A. (1979). Job demands, job decision latitude, and mental strain: Implications for job redesign. Administrative Science Quarterly, 24, pp. 285–308. Kasl, S. V. (1998). Measuring job stressors and studying the health impact of the work environment: An epidemiologic commentary. Journal of Occupational Health Psychology, 4, pp. 390–401. Lazarus, R. S. (1993). Why we should think of stress as a subset of emotion. In L. Goldberger & S. Breznitz (Eds.), Handbook of stress: Theoretical and clinical aspects (2nd ed., pp. 21–39). New York: The Free Press/Macmillan. Lazarus, R. S. (1995). Psychological stress in the workplace. In R. Crandall & P. L. Perrewé (Eds.), Occupational stress: A handbook (pp. 3–14). New York: Taylor & Francis. Mykletun, R. J. (1995). Spørsmål om arbeidsmiljø [Questions about work environment]. Stavanger: Rogalandsforskning & Norsk Hotellhøgskole. Nilsen, R. T., & Schøyen, T. T. (1998). Arbeidsmiljø i skolen. En empirisk undersøkelse av stress, personlighetstrekk og helse hos lærere [The working environment: An empirical investigation of stress, personality traits, and health among teachers]. Master’s thesis, Department of Psychology, University of Oslo. Perrez, M., & Reicherts, M. (1992). A situation-behavior approach to stress and coping. In M. Perrez & M. Reicherts (Eds.), Stress, coping, and health (pp. 17–36). Seattle: Hogrefe & Huber. Sauter, S. L., & Hurrell, J. J., Jr. (1989). Introduction. In S. L. Sauter, J. J. Hurrell, Jr., & C. L. Cooper (Eds.), Job control and worker health (pp. xii–xx). Chichester, UK: Wiley. Schwartz, J. E., & Stone, A. A. (1993). Coping with daily work problems: Contributions of problem content, appraisals, and personal factors. Work and Stress, 7, pp. 47–62.

RT44367_C016.indd 310

7/13/05 9:34:41 AM

NORWEGIAN ADAPTATION OF THE JOB STRESS SURVEY

311

Spector, P. E., Dwyer, D. J., & Jex, S. M. (1988). Relation of job stressors to affective, health, and performance outcomes: A comparison of multiple data sources. Journal of Applied Psychology, 73, pp. 11–19. Spielberger, C. D. (1979). Preliminary manual for the State-Trait Personality Inventory (STPI). Unpublished manuscript, University of South Florida, Psychology Department. Spielberger, C. D. (1983). Manual for the State-Trait Anxiety Inventory (rev. ed.). Palo Alto, CA: Consulting Psychologists Press. Spielberger, C. D. (1994). Preliminary manual for the Job Stress Survey (JSS). Tampa: University of South Florida, Psychology Department. Spielberger, C. D., & Reheiser, E. C. (1994). The Job Stress Survey: Measuring gender differences in occupational stress. Journal of Social Behavior and Stress, 9, pp. 199–218. Spielberger, C. D., & Vagg, P. R. (1999). Professional manual for the Job Stress Survey (JSS). Odessa, FL: Psychological Assessment Resources. Ursin, H., Endresen, I., & Ursin, G. (1988). Psychological factors and self-reports of muscle pain. European Journal Applied Physiology, 57, pp. 282–290. Westgaard, R. H., & Jansen, T. (1992). Sporreskjema om arbeidsmiljo [The questionnaire about work environment]. Unpublished manuscript, University of Trondheim, Norwegian Institute of Technology, Department of Organization and Work Science.

RT44367_C016.indd 311

7/13/05 9:34:41 AM

RT44367_C016.indd 312

7/13/05 9:34:42 AM

17 Calibrating Productivity in the Workplace: Assessing Stress Judiciously Ann M. O’Roark Executive Consultant, St. Petersburg, FL Abstract This chapter describes a consultant-client partnership approach to the diagnosis and treatment of workplace stress. Two examples, a special courts judge and a family courts judicial community, provide detailed descriptions of applications of the Job Stress Survey (JSS) as a benchmarking tool for calibrating optimal productivity and pinpointing the effects of over- and understimulation associated with specific workplace stressors. The case of the individual judge provides perspective for interpreting the more complex judicial community data and illustrates the wide-ranging idiosyncratic factors that influence individual employees within an occupational field. In the judicial/legal culture of the family courts community, making critical on-the-spot decisions and dealing with crisis situations were identified as the most potent stressors for judges who must provide timely, fair, and quality decisions in emotionally charged family litigations. Diversity related to professional and gender subgroups was reflected in the perceived severity and frequency of occurrence of the stressors and the responses to the individual stressor items by members of the family courts judicial community. These variations, together with findings from personality assessments and critical incident discussions, served as the foundation for calibration consultation intervention planning.

CALIBRATING PRODUCTIVITY IN THE WORKPLACE: ASSESSING STRESS JUDICIOUSLY Burnout, overload, recurring traumatic experiences, and insufficient “spice of life” (Selye, 1974) are stressor conditions known to detract from employee morale and workplace productivity (Spielberger & Vagg, 1999). However, in the explosion of professional and popular literature addressing occupational stress, little is said about the importance of striving for optimal-stressor conditions in the workplace. Consulting psychologists who advise organizational leaders are frequently expected to identify factors that generate excessive job stress and reduce employee productivity. A judicious response to this corporate expectation is to agree with the observations of executive realists who conclude that some stress is normal and can be beneficial to productivity. Using a calibration consultation (CC) approach, those stressors that are excessive and likely to inhibit productivity can be identified, and in follow-up discussions, 313

RT44367_C017.indd 313

7/13/05 9:32:25 AM

314

A.M. O’ROARK

action interventions likely to reduce effects of stress overloads and professional burnout can be selected and implemented. In addition, employees need stimulation and challenges to make work interesting, to be productive, and to spark entrepreneurial initiative (Bass, 1997). In most occupations, high performers also need to be free from hassles and trauma while working (Spielberger, Westberry, Grier, & Greenfield, 1981). Psychological assessment, which is analogous in some respects to engineering tools like thermostats and alarm indicators, relies on facts. Consultants carefully observe and assess workplace conditions, collecting specifics about “what is” and “what is desired” in order to “calibrate” the effectiveness of current organizational, work group, or individual functioning. Such information helps establish expectations, credibility, and acceptability for discussing improvements that are needed within an organization. A relatively stress-free environment is proven to be conducive to positive change and contributes to both productivity and learning (Combs, Miser, & Whitaker, 1999; Combs & Snygg, 1959). Building on these research foundations, and before leaping to intervention, calibration consultation begins with an action research/process consultation mode (Schein, 1999) that requires first getting acquainted with current workplace conditions and norms. Then, it is essential to partner with the leaders in the client organization to collect, organize, and analyze data that are accepted as descriptive of present status and the wanted future status. The objective of this joint analysis is to agree upon critical issues for intervention attention. Productivity improvements as a result of consultation activities are more likely to be lasting when the method of intervention guides clients toward self-responsibility, proactive motivation, and interest in ongoing activities that enhance performance. CALIBRATION PROCEDURES AND SELECTING A PIVOTAL INSTRUMENT The CC psychologist’s objective in consulting with an organization is to propose sound approaches for developing, maintaining, and revitalizing employee stamina and productivity. Intervention effectiveness depends upon accurate identification of primary sources of stress that are experienced by individuals and work groups. Generic remedies for stress on the job are unlikely to be relevant to all situations, and universal/“canned” remedies are notoriously inept. Thus, what works in a manufacturing plant may not be useful for managers, professionals, and staff personnel employed in specialized work communities where emphasis is placed on helping individuals to realize their potential for high achievement. In the initial action research phase, consulting psychologists seek to recognize and describe the organization’s hardy, optimally motivated workers. These employees generally prove to be long-term “assets” and are the organization’s most productive human resources. They make strong collaborators for the calibration and intervention activities. Psychologists want to discover, as quickly as possible, what is inhibiting workers’ motivation and interfering with performance. The “hardy” elephant in Figure 17.1 symbolizes optimal performance. The hypothetical

RT44367_C017.indd 314

7/13/05 9:32:25 AM

315

CALIBRATING PRODUCTIVITY IN THE WORKPLACE

Optimal Stress Maximal

Performance

Satisfaction Change

Progress Creativity

Rational problem solving

boredom

irrational problem solving

fatigue frustration

exhaustion

illness

dissatisfaction depression

Minimal

low selfesteem

Performance LOW Stress Performance impaired by Lack of stress

HIGH Stress Performance impaired by Excessive stress

Figure 17.1. Stress and the Performance Curve. The elephant represents the hardy performers, leaders of the workforce parade, who engage in rational problem solving leading to progress and innovation. Similar to the normal population bell curve, the optimal stress arch shows that too much or too little stress results in minimal or impaired performance. Irrational problem solving and illness are symptoms of excessive stimulation. Boredom and depression are symptoms of insufficient stimulation.

bell-curve points out some of the conditions that are likely to exist when personnel are performing at maximal productivity, as well as those experienced by talented contributors who “derail.” Excessive stress leads to irrational problem solving, low self-esteem, exhaustion, and physical illness. Too little stress shows itself in boredom, fatigue, frustration, and dissatisfaction. An optimal balance between too much and too little challenge is found in workers who are creative and actively engaged in rational problem solving. Optimal stressor conditions keep current projects moving forward at expected rates, foster enthusiasm and health among employees, and stimulate innovative as well as adaptive changes. Such conditions are also associated with greater productivity and job satisfaction. Individuals in the organization

RT44367_C017.indd 315

7/13/05 9:32:28 AM

316

A.M. O’ROARK

whose behavior is characterized by emotional balance, endurance, and inventiveness are said to be working in a “peak performance zone.” Calibration methodology. For CC approaches, one or more assessment instruments are used in conjunction with interviews, observational data, and group discussions for making recommendations that are given to organizational leaders, work groups, and individual employees. Formal assessment increases confidence in conclusions regarding interventions or follow-up activities. CC permits both external and internal “benchmarking” to determine when results are typical or atypical. External comparisons can be readily made with norm tables developed for instrument validation studies and internally with averages that emerge from results obtained from employees in client-organization assessments. Initial steps in providing calibration consultation include selecting instruments that most directly examine issues presented by the client. The major goal of the CC described in this chapter was to improve productivity and reduce workplace stress. The Job Stress Survey (JSS; Spielberger & Vagg, 1999) brings immediate attention to workplace stressors associated with job pressure and lack of organizational support. This instrument provides the pivotal information for developing strategies to improve productivity and to move toward optimal stress by increasing stimulation for the undermotivated and eliminating stressor overloads. Due to space limitations, it is not possible to present a detailed discussion of the various types of tests that can be incorporated in the CC process and how each contributes to the development of calibration consultation interventions. Where the presenting issue is communication, pivotal instruments have been the LIFO (Atkins, 1973/1999) or the MyersBriggs Type Indicator (MBTI; Myers & McCaulley, 1989/1995). For conflict resolution, the FIRO-B (Schnell & Hammer, 1993) and the State-Trait Personality Inventory (STPI; Spielberger, 1983) have proved useful. Problemsolving instruments are most appropriate for assessing decision-making and cognitive preferences (Harrison, Bramson, Bramson, & Parlette, 1992). Recent client awareness and interest in “emotional intelligence” and interpersonal skills suggests using instruments developed by Goldman (1995) that can contribute to team-building interventions. In CC with the judicial community, two instruments provided useful supplementary information: the Myers-Briggs Type Indicator and the State-Trait Personality Inventory. Discussion of an individual judge’s profile in the following section includes suggested action planning based on his extraverted personality characteristics that were identified by the MBTI. Results from the STPI, which identified persons with high levels of anger, were significant in recommending the conflict resolution intervention for judges and lawyers reported in the second section to follow. Meaningful interpretation of complex psychological information brings the power of individual differences in personality to bear on productivity problems. What does the JSS tell us? The JSS evaluates the perceived severity of commonly occurring workplace stressors and the frequency that these stressors are experienced. The overall JSS Stress Index (JS-I) is based on the Severity and Frequency ratings for each of the 30 JSS stressor events, which are listed in the JSS Profile Report for the special courts judge in Table 17.1.

RT44367_C017.indd 316

7/13/05 9:32:28 AM

317

CALIBRATING PRODUCTIVITY IN THE WORKPLACE Table 17.1 Job Stress Survey Report for Male Special Courts Judge, Age 46 JSS

Index

Severity

Frequency

Overall: Index Job Pressure: JP Lack of Support: LS

4.2 5.3 3.2

4.8 4.7 4.9

3.5 5.9 1.4

A. Perceived Severity 1–9 (high) 1A. 2A. 3A. 4A. 5A. 6A. 7A. 8A. 9A. 10A. 11A. 12A. 13A. 14A.

5 3 6 4 5 5 5 4 4 4 5 4 4 4

B. Six Months’ Frequency 0–9+ days 1B. 2B. 3B. 4B. 5B. 6B. 7B. 8B. 9B. 10B. 11B. 12B. 13B. 14B.

0 9 0 9 0 0 9 0 0 0 0 9 0 5

15A. 5

15B. 0

16A. 17A. 18A. 19A. 20A. 21A. 22A. 23A. 24A. 25A. 26A.

16B. 17B. 18B. 19B. 20B. 21B. 22B. 23B. 24B. 25B. 26B.

5 6 6 6 6 5 5 5 5 5 6

9 0 0 9 0 0 0 9 9 5 9

27A. 5

27B. 0

28A. 3 29A. 5 30A. 5

28B. 6 29B. 0 30B. 9

Severity 4.8

Frequency 3.5

Job Stress Survey Items LS scale = dark shade

JP scale = light shade

Assignment of disagreeable duties Working overtime Lack of opportunity for advancement Assignment of new or unfamiliar duties Fellow workers not doing their job Inadequate support by supervisor Dealing with crisis situations Lack of recognition for good work Performing tasks not in job description Inadequate or poor equipment Assignment of increased responsibility Periods of inactivity Difficulty getting along with supervisor Negative attitudes toward organization Insufficient personnel to handle assignments Making critical on-the-spot decisions Personal insult from customer/colleague Lack of participation in policy decisions Inadequate salary Competition for advancement Poor or inadequate supervision Noisy work area Frequent interruptions Frequent changes in work activities Excessive paperwork Meeting deadlines Insufficient personal time (e.g., coffee breaks) Covering work for another employee Poorly motivated coworkers Conflicts with other departments

C. Index A×B –0– 27 –0– 36 –0– –0– 45 –0– –0– –0– –0– 36 –0– 20

Ranking and Action Ideas

2

–0– 45 –0– –0– 54 –0– –0– –0– 45 45 25 54

2 1

2 2 1

–0– 18 –0– 45

2

Index 4.2

Note: Individual JSS reports are distributed at the beginning of a stress management consultation. This report is either given directly to the individual client for one-to-one discussion or sent in sealed envelopes addressed to each participant in a group workshop.

RT44367_C017.indd 317

7/13/05 9:32:29 AM

318

A.M. O’ROARK

Severity ratings for each JSS item can range from a very low score of 1 to a high score of 9, with mean item Severity scores falling between 4 and 5 in the normative samples (Spielberger & Vagg, 1999). Ratings of the Frequency of occurrence during the past 6 months for each JSS stressor event can range from 0 to 9-or-more (9+) days on a 10-point scale, with mean Frequency scores of 3.5 to 4.0 in the normative samples (Spielberger & Vagg, 1999). Clients ratings for the individual stressor items are also examined and are used, along with responses of fellow workers, to generate collective information for each stressor event. The 30 JSS Severity and Frequency scales include two embedded 10-item subscales that assess Job Pressure (JP) and Lack of Organizational Support (LS). Job Pressure refers to the structural and technical aspects of the employee’s workload. Lack of Organizational Support involves relationships among employees who are carrying out the work, as well as the stress that results from the policies and procedures of the organization. Scores for the two JSS subscales are computed in the same manner as the overall JSS Stress Index, Severity, and Frequency scores. The mean ratings for the JP and LS Severity and Frequency subscales range between 2 and 8. Interpreting high and low JSS ratings. Responses to the JSS provide information about generic sources of stress in the workplace. Each of the 30 JSS stressors was selected because it represented a source of stress encountered by men and women at different pay levels in a variety of occupations in business and industry, government, and educational settings. The simplicity of the JSS items makes it possible for respondents to associate a specific stressor with a particular experience in their own workplace. The CC focuses on those items with the highest and lowest scores in follow-up sessions with clients that elicit observations, comments, and personal examples. Items rated high in perceived Severity that are also frequently experienced, or that are confirmed by critical incident examples collected from respondents, are selected as stress management targets to be addressed in intervention action plans. Understimulation or denial? Very low stressor ratings are also considered in follow-up discussions and included in stress-management planning. When individuals report low Severity scores or infrequent exposure to a particular stressor, the consultant seeks to determine whether these ratings reflect a nonstimulating work environment or a denial process within the individual. If the employee is experiencing boredom or a lack of challenge, more optimal working conditions may be established through interventions that involve job enrichment, added assignments, or expanded objectives. The introduction of novelty can also reduce boredom and dissatisfaction, prevent depression, and improve performance. When denial is at work, the consultant strives to call attention to “realities” in the workplace by reviewing JSS normative data and benchmark ratings reported by peers. Clients who are reluctant or too embarrassed to admit that they experience occupational stress are generally more comfortable in discussing their job pressures after the consultant explains the protective value of emotions. Stress-related emotions, like anxiety and anger, which serve as survival messengers, signal that it is time for problem-solving action.

RT44367_C017.indd 318

7/13/05 9:32:29 AM

CALIBRATING PRODUCTIVITY IN THE WORKPLACE

319

The responses of a participant in an executive workshop provide an example that illustrates how personality and denial can influence JSS scores. The president of an entrepreneurial engineering firm reported extremely low Severity ratings, with an average score of less than 2 on the nine-point scale. His Frequency ratings also indicated that he had experienced most of the JSS stressor events on fewer than two days in the past six months. When the results for all workshop participants were posted anonymously for class discussion, participants asked who had the bottomof-the range scores. The high-energy president acknowledged that they were his. Members of the group then provided immediate feedback, describing his workshop behavior, which was distinguished by strong opinions and frequent differences with others. They reminded him how he talked about his work setting as being full of high risks, time pressures, and the turmoil of changing technologies. Classmates said they were convinced that he was a major source of stress for those working around him, just as he had been for them during the executive workshop. The president explained his low JSS ratings by saying that a strong leader does not talk about feeling stressed. He said that company loyalty demanded that he, as president, could not acknowledge imperfections in the organization, especially to “outsiders.” Later, in a private discussion with the instructor, the president indicated a growing awareness that it might be time for him to retire or to change organizations. He was able to look directly at how his behavior might be doing damage to others around him, as well as to himself. Consultant-client partnership. Another important consideration in working with executive and professional populations is that they often prefer that the wording of stressor items include discipline-specific terminology. Sometimes, requests are made to add to the 30 JSS stressor events. In the judicial community to be discussed later, several JSS items were reworded without changing the focus of the query. For example, Item 5, Fellow workers not doing their job, was modified to Others in the judicial/legal system not doing their job. Similarly, Item 30, Conflicts with other departments, was changed to Conflicts with other persons or agencies in the judicial/legal system. Formulating adaptations of JSS items provides the consultant with insight into client realities and facilitates establishing consultant-client partnership approaches to resolving organizational problems. Adjusting questions to capture nuances of professional work demonstrates the consultant’s respect for a client’s special circumstances and unique challenges. The consultant’s attitude of collaboration carries over into feedback sessions and builds the sense of selfresponsibility in clients, who must ultimately determine the implementation of plans for improvement. Finally, collaborative modification of the questionnaire items increases motivation for candid responding to the JSS. INDIVIDUAL CALIBRATION: A SPECIAL COURTS JUDGE The JSS report for a 46-year-old male special courts judge is presented in Table 17.1 in a format designed to be given directly to the client. This report can serve as a scoring sheet, allowing the consultant or the individual client to easily calculate and better understand the results of the assessment. During

RT44367_C017.indd 319

7/13/05 9:32:30 AM

320

A.M. O’ROARK

feedback, three types of information are discussed: (1) the general concept of workplace stressors, (2) the perceived Severity and Frequency of occurrence of the specific stressors experienced by the client as most stressful, and (3) scores on the JSS subscales that assess Job Pressures and Lack of organizational support as important domains of occupational stress. The JSS Job pressure subscale assesses stressful experiences that result from the nature of the work and the requirements of the job. The Lack of organizational support subscale measures the effects of stressors that involve relationships with supervisors, coworkers, and management policies and practices. In collaboration consultation, the consultant first provides feedback that engages the client in talking about earlier assessments or extenuating current circumstances: “What recent events or pending decisions might have influenced your responses to the JSS questions?”; “What other assessments have you taken?” The consultant also endeavors to identify those personality characteristics of the client that might help in interpreting the JSS results. Following this introductory discussion, the consultant guides the client through a review and discussion of the data contained in the JSS Report, beginning with a discussion of the Stress Index, Severity, and Frequency scores based on all 30 items. Explaining the profile and calculations. In the JSS report for the special courts judge presented in Table 17.1, the judge’s overall Stress Index, Severity, and Frequency scores were all lower than 5, the average level of job stress. These scores are reported in the top row of the JSS report and are also noted in the shaded area at the bottom of Columns A and B. To compute the average perceived Severity score, the Severity ratings for each JSS item listed in Column A are added, and the sum of these ratings is divided by the number of items. The average Frequency score is obtained in the same manner by adding ratings for each Frequency item in Column B and dividing by the total number of items. The overall Stress Index score is derived for the “quick hand scoring” method used in CC by simply adding the average Severity and Frequency scores and dividing this sum by two. The Stress Index scores for each item, which are listed in Column C, were obtained by multiplying the item Severity rating by its Frequency rating. Analyzing the scores. The judge’s relatively low Stress Index score of 4.2 resulted primarily from his very low 3.5 Frequency score. The Frequency of occurrence of each stressor was rated on a 10-point scale, with a score range of 0 to 9+ to indicate the number of days the stressor was experienced in the last six months. The scores for the JP and LS subscales indicated that the Frequency that Job pressures were experienced was the dominant stress-inducing factor. The judge’s slightly above average JP Index score of 5.3 was obtained by adding his Frequency (5.9) and Severity (4.7) scores, and dividing this total by 2. His low LS Stress Index score of 3.2 was due to his very low mean LS Frequency score of only 1.4. Clearly, for this judge, Job pressures had a much greater impact than did Lack of support, and the Frequency that Job pressures were experienced contributed most to his occupational stress. Five of his seven highest-rated JSS items were Job pressure stressors (7, 16, 23, 24, 26), with Item Index scores ranging from 45 to 54 (Column C).

RT44367_C017.indd 320

7/13/05 9:32:30 AM

CALIBRATING PRODUCTIVITY IN THE WORKPLACE

321

Calibrating Severity. In the next stage of the CC with the special courts judge, attention was directed to his Severity ratings for each JSS item (Column A). The perceived Severity of each stressor event was rated on a nine-point scale in terms of how stressful it was as compared to the standard stressor, Disagreeable duties, with an assigned midpoint Severity rating of 5. The judge rated only six items as greater than average in perceived Severity, as indicated by ratings of 6. The stressors with the highest Severity ratings were 3 (Lack of opportunity for advancement), 17 (Personal insults), 18 (Lack of participation in policy-making decisions), 19 (Inadequate salary), 20 (Competition for advancement), and 26 (Meeting deadlines). However, it is interesting to note that four of these stressors received 0 Frequency ratings; only Inadequate salary and Meeting deadlines were rated as occurring with a relatively high Frequency. As may be noted in Column B, the judge assigned Frequency ratings of 0 for 17 of the 30 JSS stressor events and 9+ ratings for 10 stressors. Sources of stress experienced with a high Frequency of occurrence generally indicate that these stressors are inherent in the job. An important reality check that a consultant must consider is whether the nature of a client’s work inherently includes frequent encounters with particular stressors. Some business and service organizations have work conditions that are recognized as involving a highly intense pace, such as stock exchanges and emergency rooms at hospitals. Likewise, dealing with criminals and antisocial incidents in courts of law can be considered as a stressful occupation in which particular Job pressures are expected to be frequently encountered. Negative consequences of high Frequency/low Severity items. In consulting with a client about achieving optimal stressor conditions, it is essential to recognize that even low Severity stressors, when frequently experienced, can impair performance and productivity, and cause wear and tear on the body. For example, police officers who reported relatively low Severity levels for paperwork duties acknowledged experiencing daily aggravations as a consequence of such chores, which contributed to excessive paperwork becoming highly stressful for them (Spielberger et al., 1981). We have observed that midlevel managers who rated paperwork as one of their primary stressors were often comparative newcomers to management. In feedback discussions, it was found that the Severity of the paperwork stressor often reflected a client’s lack of experience and efficiency, or a personal preference for the hands-on physical activity associated with “doing” a job. Once the client is familiar with the sources of stress described by the JSS items, and with the meaning of the scores reported in Columns A, B, and C of the JSS Profile Report, the major stressors are discussed in depth. These stressors are rank-ordered in Column C of the Profile Report to establish priorities for Critical incident discussion and Action plan attention. Examples describing particular encounters with the top stressors, referred to as critical incidents, are also solicited from the client. Once validated, the stressors with the highest scores are selected for immediate attention. In the case of ties, as were found in the results for the special courts judge, the items with the highest rank are selected for primary attention.

RT44367_C017.indd 321

7/13/05 9:32:31 AM

322

A.M. O’ROARK

For JSS items with Frequency ratings of 0, the Item Index score is also 0, indicating that the stressor was not experienced during the past six months, and therefore had little or no impact on the worker. However, it can be informative to note the absence of stress as well as the pressures of stress in the workplace. Although the primary objective of CC is to engage clients in thinking about how to address situations associated with excessive trauma and hassles, it is equally important to be alert to profiles indicating stressor levels that are very low and to direct the client’s attention to personal and organizational benefits that can be obtained by taking on fresh challenges or expanding personal responsibilities. Providing employees and professionals who participate in organizationwide surveys with a personal and confidential copy of their JSS Profile Report stimulates and reinforces self-responsibility. For the judge whose JSS scores are presented in Table 17.1, items 19 (Inadequate salary) and 26 (Meeting deadlines) were identified as his top stressors, each with Item Index scores of 54 and 9+ Frequency ratings. The judge’s overall Frequency Index of 3.5, which was relatively low, suggested that boredom might be detracting from peak performance. With no salary increments on the horizon and increasing pressure from deadlines, the judge’s motivation and productivity could suffer despite his high achievement drive and prestigious civic/community status. Formulating action plans. To help the judge sort out the implications of his JSS data, it was important to look at his stressors that were ranked 2 in his Profile Report (Column C), each with an Item Index score of 45. The overload potential for these five stressors (Dealing with crisis situations, Making on-the-spot decisions, Frequent interruptions, Frequent changes in work intensity, Conflicts with other departments) comes not from perceived Severity, but from their frequency of occurrence in the work environment. Frequently occurring stressors affect multiple people and may be indicative of dysfunction in the work group or organization. The stressors that dominated the judge’s profile were all JP items; none of the JSS LS items were rated as highly stressful. Although some judges consider Job pressures as problems to be handled by court administrators and management staff, this special courts judge determined that his organization would benefit if he expanded his zone of responsibility and offered to help solve problems associated with his five most highly rated Job pressures. He set a goal to initiate collaboration among peers in his work community to deal with frequently occurring stressors. He also saw an opportunity to develop “triage plans” that could lead to greater efficiency during times of crisis (Item 7) and to facilitate the preparation of checklists of criteria and priorities to use when making quick decisions (Item 16). In moving toward an optimal stress level, the special courts judge considered specific stressors that were prominent in his JSS scores and decided to take the initiative of applying his natural abilities more creatively. His personal development/action plan included addressing his most stressful administrative JP problems, which were Dealing with crisis situations, Making on-the-spot decisions, Frequent interruptions, Frequent changes in work intensity, and Meeting deadlines. Systemwide improvements could also emerge from giving attention to his top-ranked stressor, Inadequate pay. Discussions with the state

RT44367_C017.indd 322

7/13/05 9:32:31 AM

CALIBRATING PRODUCTIVITY IN THE WORKPLACE

323

government hierarchy and wage-scale policymakers for the family courts could lead to updating salary and compensation programs, not only for himself but also for many others. Discussing Frequent changes in work intensity offered the CC a window of opportunity for helping the judge to identify ways for making the best use of slow time, for example, triage planning and advocacy for more adequate judicial compensation policies. The remaining high-ranked stressor, Conflict with other departments, with a Stress Index score of 45, stimulated another interesting addition to the judge’s stress management plan. On the basis of a critical review of incidences of conflict among coworkers (Item 30) and knowing himself to be characterized by a strongly extraverted personality, as highlighted by results from the MBTI assessment, the judge decided to volunteer as an arbitrator for resolving specific interdepartmental conflicts. LARGE-SCALE CALIBRATION: A FAMILY COURTS AGENCY Given that stress is inherent in work, as in life, realistic executives and managers do not aim for a stress-free environment. Furthermore, challenging levels of stress infuse motivational value to job-related activities, providing what Hans Selye (1974) called the spice of life. Even with accurate survey data of occupational stress, a vision of optimal stress becomes unrealizable if managers and executives ordain one set of conditions for all. Controlling stress in the workplace needs to be tailored to the organization’s management philosophy, the nature of the work, and the personality characteristics of the involved individuals. Global one-size-fits-all approaches did not result in harmonious productivity in the 18th century for paternalistic cottage industries established by the English in their homes at the beginning of the Industrial Revolution. Nor did similar concepts work in the 19th century for U.S. steel companies, farming plantations, or coal-mining towns. During the early decades of the 20th century, residential compounds established by industries to alleviate their employees’ most pressing stressors were abandoned as unions sprang into power, voicing workers’ opinions and concerns. Professional associations also expanded to ensure quality standards for services and the well-being of practitioners. In the 21st century, more and more employees and their fellow workers prefer to speak for themselves. Assessing stress judiciously. The work of the CC with the employees of a family courts judicial community is reported here to illustrate how JSS results are collated, subgrouped, and analyzed in ways that avoid the one-size-fits-all fallacy. Partitioning the JSS data highlights differences in the stressor experiences of the several work and peer groups within an organization and helps the consultant to tailor stress management interventions not only for individuals, but also for unique segments of a workplace community. A senior judge in the family courts community spearheaded the development of the arrangements for the CC assessment as part of a state-funded program. The first phase of this project involved selecting and modifying, where appropriate, the assess-

RT44367_C017.indd 323

7/13/05 9:32:31 AM

324

A.M. O’ROARK

Table 17.2 Mean JSS Stress Index, Severity, and Frequency Scores for the Judicial Community and the Five Professional Work Groups, and the Principal Sources of Stress for the Total Community

Judicial Community (N = 43) Professional Work Groups 1. Lawyers (N = 3) 2. Court administrators (N = 4) 3. Judges (N = 9) 4. Administrative staff (N = 18) 5. Judicial assistants (N = 9) Principal Stressors 1. Frequent interruptions 2. Dealing with crisis situations 3. Making critical on-the-spot decisions 4. Excessive paperwork 5. Meeting deadlines 6. Fellow workers not doing their job

Stress Index

Severity

Frequency

4.42

4.66

4.17

4.94 4.62 4.61 4.15 3.77

5.05 4.90 4.66 4.33 4.38

4.83 4.33 4.57 3.97 3.16

6.18 6.16 5.84 5.67 5.55 5.55

5.25 5.84 4.61 5.09 5.05 5.73

7.11 6.48 7.07 6.25 6.05 5.36

Note: Professional groups are rank-ordered as a function of the mean JSS Stress Index scores for each group. The principal stressors are rank-ordered as a function of mean JSS Index scores for the Total Judicial Community.

ment instruments. The CC then partnered with leaders in the client organization to determine the best methods for collecting psychometric data and providing feedback of results to individuals, work groups, and group leaders. An assessment battery was administered to the individuals responsible for the activities and services provided by members of the family courts community, which included 43 respondents (25 women, 18 men) who were client-classified as belonging to two groups of legal professionals: (a) judges and the lawyers representing the clients appearing before the family courts, and (b) court administrative personnel, including diverse specialists such as administrators, judicial assistants, and support staff, consisting of counselors, family violence specialists, and clerks. The JSS results for these two major groups were then broken out for five work-related subgroups: judges, lawyers, administrators, judicial assistants, and support staff. Table 17.2 reports the results of the JSS assessment for the total judicial community. The overall mean Stress Index score of 4.42 reflects a somewhat below-average amount of stress. At first glance, this finding is puzzling because of the intense nature of the cases and issues that comprise the work of the family courts community. However, a stressor climate lower than that reported in many private and commercial organizations becomes less surprising when considered in the context of a tendency among professionals and senior management to understate stress conditions. Survey responses of persons motivated to be high achievers are influenced by internalized values, expectations, and attitudes, such as satisfaction associated with meaningful

RT44367_C017.indd 324

7/13/05 9:32:32 AM

CALIBRATING PRODUCTIVITY IN THE WORKPLACE

325

work and loyalty to the group or organization. Admirable as these characteristics may be, they cloud stress survey results if compared only to statistical tables for workplaces in general. It can be detrimental to efforts to increase productivity and job satisfaction if survey results are not interpreted in light of the norms for the worksite that influence the way respondents “use their numbers” when providing ratings or estimating severity. The mean perceived Severity and Frequency ratings of 4.66 and 4.17 for the total judicial community, which are reported in Table 17.2, were below the average normative midpoints for these scales. However, the underlying story of optimal stress cannot be known by examining only the overall mean Stress Index scale and subscale scores. CC consultants must develop recommendations to fit unique circumstances by taking time to understand the occupation-related attitudes and individual personality characteristics that influence the perceived Severity ratings and the environmental conditions that contribute to the frequency of occurrence of a particular stressor. Effective stress management strategies emerge from one-on-one discussions with individuals and groups of employees for those stressors with the highest ratings. Stressor Differences in the Five Judicial Community Work Groups To provide a more useful picture of the major sources of stress for the judicial community, workforce data for each of the five professional groups were prepared, as reported in Table 17.2. In addition to the overall mean Stress Index, Severity, and Frequency ratings for each group, the six principal stressors rated as most stressful by the entire family courts judicial community are also reported in Table 17.2, in which they are listed in the descending rank-order of their mean Stress Index ratings. The JSS Stress Index scores for each group were obtained in the same quick-score manner as previously described for determining the individual Item Index scores, by adding the mean Severity and Frequency scores for each stressor event and dividing this sum by 2. Lawyers reported highest stressor levels. The mean JSS Stress Index scores for three of the five family courts work groups were higher than the Stress Index score of 4.42 for the total judicial community. The lawyers are paid by their clients, who are plaintiffs or defendants in court cases, and reported higher mean scores across all JSS dimensions than the other four groups. They were also the only group whose mean Severity score of 5.05 was above the 5.0 midpoint rating. Thus, stress management is particularly important for this small, highly involved group of lawyers who voluntarily participated in the CC assessments. Their participation in the calibration project allowed them to become better acquainted with the judges and other court personnel with whom they worked. The CC process itself most likely facilitated reduction in the lawyers’ stress levels. Judges and court administrators reported similar stressor levels. The judges and the court administrators reported very similar mean Stress Index ratings of 4.61 and 4.62, with Severity scores that were slightly higher than their Frequency scores. The administrators, who served as chief executives for

RT44367_C017.indd 325

7/13/05 9:32:32 AM

326

A.M. O’ROARK

their respective agencies, reported higher Severity scores and lower Frequency scores than the judges. Although the judges reported encountering JSS stressor events more frequently than the administrators, they were apparently more sanguine about dealing with them, as indicated by their lower perceived Severity ratings for most stressors. Support staff and judicial assistants reported lower stressor levels. As may be noted in Table 17.2, there is a clear difference between the JSS scores of the top three professional groups and those of the last two groups, whose lower Stress Index scores resulted primarily from less frequent exposure to major stressor events. The judicial assistants, who each work as executive aide to one judge, reported experiencing JSS stressor events less often than any of the other groups. These midlevel administrative staff work in a sheltered situation as gatekeepers for a judge and are infrequently exposed to many of the major stressors. It is also possible that judges prefer to hire assistants with low-key Type-B personality characteristics and reaction patterns, and who typically report less stress than Type-A achievers. An alternative interpretation of low JSS scores for a particular group is that the respondents are not willing to admit experiencing stress. The reason for the lower stressor results can only be determined by follow-up interviews and discussion about hiring practices. Stressor item analysis is the foundation for intervention recommendations. Stressor salience for the five professional work groups can be clarified by contrasting the mean Stress Index, Severity, and Frequency scores of the six principal JSS stressors for the total judicial community, which are listed in Table 17.2. The top two stressors are Frequent interruptions and Dealing with crisis situations, with Index scores of 6.18 and 6.16. The mean Severity score of 5.84 for Dealing with crisis situations is higher than the Frequent interruptions mean Severity score of 5.25, whereas the Interruptions mean Frequency score of 7.11 is higher than the Crisis Severity score of 6.48. Although Dealing with crisis situations is clearly more traumatic for most people than Frequent interruptions, the very high frequency ratings for Interruptions gives this stressor added impact, as indicated by its slightly higher mean Stress Index score. Decision making. The stressor ranked third highest in Table 17.2, Making critical on-the-spot decisions, with a Frequency score of 7.07, was reported as experienced almost as often as Frequent interruptions by the total judicial community, but was rated as less than average in perceived Severity. Given the responsibility of the courts to render decisions, this finding is consistent with what would be expected. Family courts produce many decisions relating to child and spousal abuse, and divorce and custody settlements, which are ideally handed down in a fair, ethical, and timely fashion. Such decisions are based on points of law, judicial precedents, and procedural judgment calls by judges and attorneys. Emotional reactions and dramatic behaviors by defendants and prosecuting parties are also frequently encountered. Consequently, members of the judicial community are required to make numerous on-thespot decisions. Individuals seeking employment in court organizations are likely to be those who enjoy the adrenaline rush produced by these types of stressful circumstances.

RT44367_C017.indd 326

7/13/05 9:32:33 AM

327

CALIBRATING PRODUCTIVITY IN THE WORKPLACE

Paperwork and deadlines. The mean perceived Severity scores for the judicial community of 5.09 for Excessive paperwork and 5.05 for Meeting deadlines were close to the generic averages for most work groups, whereas the Frequency means for these stressors were considerably higher than found in other organizations. Indeed, five of the top six stressors showed substantially elevated frequencies, as may be noted in Table 17.2. Stressors that are frequently generated by a work environment are best addressed collaboratively where joint perspectives can be presented and mutually acceptable changes in work flow or employee assignments can be considered. Fellow workers. The impact of the sixth-ranked stressor, Fellow workers not doing their job, with a mean Stress Index score of 5.55, ranks second only to Dealing with crisis situations in perceived severity. Could this indicate that dealing with crises and the limitations of fellow workers are yoked stressors? Could the incompetence or lack of effort by fellow workers have a causal role in precipitating crisis events? Such questions are formulated by the consultant in order to probe deeper into organizational conditions. Recommendations for Reducing Job Stress in the Judicial Community The mean JSS Item Index scores of the six principal stressors for the five judicial community professional work groups are reported in Table 17.3, in which the ranks of the top three stressors for each group are also noted. Differences among the five groups served as the basis for tailoring follow-up interventions with each group or combinations of groups. Frequent interruptions headed the list of organizationwide stressors and was rated as one of the top two stressors for four of the five professional groups. However, this stressor was rated very low by the judges. Although it would not be appropriate to recommend that all members of the judicial community spend time Table 17.3 Mean JSS Item Index Scores of the Principal Stressors for the Five Professional Work Groups

JSS Stressors Items 1. Frequent interruptions 2. Dealing with crisis situations 3. Making critical on-the-spot decisions 4. Excessive paperwork 5. Meeting deadlines 6. Fellow workers not doing their job

Judges (N = 9)

Court Admin. (N = 9)

Judicial Assists. (N = 9)

Admin. Staff (N = 9)

Lawyers (N = 9)

3.34 6.22-2 7.22-1

7.50-2 7.50-2 6.50

6.45-2 6.50-1 6.11-3

5.86-1 5.59-2 4.97

7.76-2 7.36 4.98

5.88 6.00 6.11-3

7.50-2 7.63-1 5.00

5.50 4.62 4.45

4.22 4.92 5.09

7.50-3 8.36-1 7.00

Note. The principal stressors are rank-ordered as a function of the mean Item Index scores for the total judicial community. The ranks of the top three stressors are noted for each professional group following the mean item score. Inadequate salary ranked third for the administrative staff, but was not ranked among the top 10 stressors for the total judicial community.

RT44367_C017.indd 327

7/13/05 9:32:33 AM

328

A.M. O’ROARK

on devising ways to reduce the number of their interruptions, an intervention could be suggested that included all groups except the judges. Dealing with crisis situations was also ranked as one of the top two stressors by four of the five groups. Although this item was ranked fourth by the lawyers, their Stress Index score of 7.36 was much higher than the scores of three of the other four groups. These findings suggest that it would be advisable to include all five groups in sorting out a stress management action plan for this item. Critical incident reports associated with Dealing with crises and Frequent interruptions, which were reported in follow-up discussions, reflected an organizationwide need to give attention to employees’ work schedules and to revising management practices that were out of date, uncoordinated, or inefficient. Taking critical incident reports into account in the context of the results of the analyses of the JSS scales and subscales facilitated CC recommendations of interventions and action plans for the five work groups. Judges and judicial assistants. Making critical on-the-spot decisions was the most problematic stressor for the judges and was also ranked relatively high by their judicial assistants. The judicial assistants ranked Frequent interruptions as their number two stressor, whereas their bosses, the judges, who are protected from interruptions by their assistants, are relatively free from this stressor. However, both judges and their judicial assistants might be included in an intervention that focused on making on-the-spot decisions that could open up the judges’ awareness to how much the assistants were protecting them from interruptions. The judicial assistants enjoyed the lowest overall level of job stress, as indicated in Table 17.2 by their mean Stress Index score of 3.77 and very low Frequency score of only 3.16. One-on-one discussions with judicial assistants indicated that they felt protected by the judges on matters relating to adequate salaries and job security, which likely contributed to their lower stressor ratings. Judges and lawyers. For the judges, the stressor mix of their top three stressors, Making critical on-the-spot decisions, Dealing with crisis situations, and Fellow workers not doing their job, suggested that they would benefit from stress management recommendations directed toward helping them find effective ways for coping with these stressors. In follow-up discussions in which they reported critical incident examples, the judges attributed Fellow workers not doing their job to the lack of case preparation by lawyers or their failure to adequately explain matters to their clients. On the other side of the bench, the lawyers, who rated Meeting deadlines, Frequent interruptions, and Dealing with crisis situations as highly stressful, complained that judges brought stress into the court situation by their demands for perfection, by not maintaining time schedules, and by too little communication. Changes in the scheduling of trials increased the frequency that lawyers experienced interruptions and crisis situations. A conflict resolution recommendation for the judges and lawyers to address stress associated with Fellow workers not doing their job and Meeting deadlines could lead to an improved collaborative climate for the entire judicial community.

RT44367_C017.indd 328

7/13/05 9:32:33 AM

CALIBRATING PRODUCTIVITY IN THE WORKPLACE

329

Court administrators. The court administrators work for the state as the chief executive officers of a judicial community. Looking at their top four stressors, which are listed in Table 17.3 (Deadlines, Interruptions, Crisis situations, and Excessive paperwork), the mean scores for these stressors of 7.5 or more were much higher than those of the other professional groups, except for the lawyers, who also had very high scores for these stressors. For the court administrators, the CC consultant would likely advise that a major stress management intervention be built into follow-up activities that did not include other groups, at least in the initial sessions. Administrative staff. For the administrative staff, Frequent interruptions and Dealing with crises were their top-ranked stressors. Inadequate salary, their third strongest stressor with a mean Stress Index of 5.20, was not highly rated by any of the other groups. Lack of opportunity for advancement, which was the fourth strongest stressor for the administrative staff, was also among the lowest rated stressors for the other four groups. Although the benefits of good working conditions kept the overall Stress Index scores below average for the total judicial community, Inadequate salary and Lack of opportunity for advancement formed a combination of stressors associated with the lower pay status of the administrative staff and judicial assistants. Gender Differences in Occupational Stress in the Judicial Communities Gender provides an important perspective for interpreting job stress in the workplace. The mean JSS Stress Index, Severity, and Frequency scores for the men and women in the judicial community are reported in Table 17.4, in which it can be noted that both genders had below-average JSS Index and Severity scores, and the women had very low Frequency scores. The men’s slightly higher Stress Index scores resulted from their higher Frequency scores, indicating that they experienced the 30 stressor events considerably more often than women. Although the women had somewhat higher perceived Severity scores than the men, they reported experiencing fewer encounters with most of the 30 JSS stressor events, which might be attributable to the Table 17.4 Mean JSS Stress Index, Severity, and Frequency Scores for Women and Men in the Family Courts Judicial Community Gender

Index

Severity

Frequency

Women (25) Men (18)

4.10 4.41

4.55 4.33

3.64 4.48

Note: In the Job Stress Survey: Professional manual (Spielberger & Vagg, 1999), mean Stress Index scores are more precisely computed by adding Item Index scores and dividing by 30, the number of stressor items. The method for determining JSS scores in this and in previously reported consultations (O’Roark, 1995) is a time-saving variation that facilitates the consultant’s work with client groups in scoring their own results in a workshop or in feedback calibration.

RT44367_C017.indd 329

7/13/05 9:32:34 AM

330

A.M. O’ROARK

smaller proportion of women working at higher-level positions with more arduous duties. In previous studies with the JSS (Spielberger & Reheiser, 1994a, 1994b), gender differences observed at the item level were lost in averaged Stress Index scores for which no significant differences were found. For the judicial community, significant gender differences were found for almost half of the Severity items and 60% of the Frequency items. Some stressors were rated stronger by women, while others were rated as more stressful by men, enough to level out gender differences in the JSS scale and subscale scores. However, both women and men tended to perceive the same stressors as highest and lowest in severity. Lack of Support stressor items relating to supervisory and interpersonal relationships and organizational policy were rated highest in perceived Severity for both genders. Women’s stressors. The perceived Severity scores of women in the judicial community were significantly higher than those of the men for Inadequate salary, Frequent interruptions, Competition for advancement, Making on-thespot decisions, Covering work for another employee, Performing tasks not in the job description, and Insufficient personal time. The Frequency ratings of women were also greater for Inadequate salary, Personal insult from customer/ colleague, Noisy work area, Covering work for another employee, Assignment of increased responsibility, and Insufficient personal time. Most of these stressor events were Job pressures associated with workload and job structure, rather than relationships with supervisors and organizational policies. Men’s stressors. Men had higher perceived Severity scores for Lack of participation in policy decisions, Experiencing negative attitudes toward the organization, Poor or inadequate supervision, and Conflict with other departments, and higher Frequency ratings for Insufficient personnel, Lack of recognition for good work, Lack of participation in policy decisions, Experiencing negative attitudes toward the organization, Making critical on-the-spot decisions, Conflict with other departments, and Working overtime. Five of these seven stressors are Lack of support items, suggesting that the strongest sources of occupational stress for the men were relations with supervisors and organizational support, rather than workload and job structure. These outcomes, which would not likely be intuitively deduced, appear to be strongly influenced by more men working at higher organizational levels. A confounding factor in interpreting gender differences in organizational cultures is the ratio of women and men working at each organizational level. In many organizations, there is a preponderance of men working at seniorlevel, higher-paying positions, and more women in lower-paying clerical or nonprofessional positions. A study comparing stressors in university, corporate, and military work settings found substantially higher scores for corporate personnel than for university or military groups (Spielberger & Reheiser, 1994a, 1994b). Nevertheless, the military women reported perceiving stressors of greater severity than any other subgroup, an interesting statistic in light of subsequent legal actions on behalf of these women. In investigations of differences in the sources of stress experienced by managerial, professional, and clerical personnel (Spielberger & Reheiser,

RT44367_C017.indd 330

7/13/05 9:32:34 AM

CALIBRATING PRODUCTIVITY IN THE WORKPLACE

331

1994a), professionals rated the JSS Job pressure subscale stressors as more severe than the other groups. Managers reported experiencing job pressures more frequently than professionals, who reported more frequent job pressures than clerical personnel. In the calibration consultation with the judicial community reported in this chapter, there was a prevalence of women in the lower organizational level groups of judicial assistants and administrative staff. Thus, the finding of Inadequate salary as a leading stressor for women was consistent with the results cited in the normative studies. Because more women were working at lower wage and salary levels in this and in most previous studies, further research will be required to tease apart the gender and hierarchical work factors that contribute to the magnitude of occupational stress. DISCUSSION AND CONCLUSIONS This chapter reports the results of using the Job Stress Survey in a collaboration consultation with clients interested in cultivating optimal performance conditions in a professional work setting. Assessing individuals and work groups in order to identify targets for possible stress management interventions provided a foundation for a consulting partnership with clients in the judicial community by actively engaging them in specifying the special circumstances and unique challenges that might contribute to successful outcomes of intervention efforts. It was also noted that assessing stressors judicially requires that attention be given to both overstimulation and understimulation of uniquely trained and talented workers. Three important features of the calibration consultation method were noted: (1) the advantage of providing a report of assessment results that can be given to the client for joint analysis, (2) examining and integrating a broad range of observations and assessment information (e.g., psychometric tests, current problems, imminent choices or decisions), and (3) not overgeneralizing. The mean JSS scale and subscale scores that were obtained for the total judicial community proved to be less pertinent for intervention planning than the partitioned ratings of the JSS stressor items for the five professional work groups. Evidence is mounting that global results may be misleading and that considering the effects of gender and cultural differences on occupational stress is also essential. The JSS profile of the prevalent stressors in the judicial community revealed why global prescriptions often lead only to partial or minimal success. Although multilevel interpretation of the JSS data for the total judicial community helped to identify strategies for addressing the principal stressors that were felt by most employees to inhibit productivity, the major stressors for a particular work group or task unit may not have a negative impact in other parts of the organization. Partitioned JSS data, taking gender differences and the organizational level of specific work groups into account, will help management to make informed decisions about where and how to invest time, effort, and resources for improving the quality of work life.

RT44367_C017.indd 331

7/13/05 9:32:34 AM

332

A.M. O’ROARK

REFERENCES Atkins, S. (1973/1999). LIFO training and OD analyst guide. Beverly Hills, CA: LIFO Associates. Bass, B. M. (1997). Does the transactional-transformational leadership paradigm transcend organizational and national boundaries? American Psychologist, 52, pp. 130–139. Combs, A. W., Miser, A. B., & Whitaker, K. S. (1999). On becoming a school leader. Alexandria, VA: Association for Supervision and Curriculum Development. Combs, A. W., & Snygg, D. (1959). Individual behavior. New York: Harper & Row. Goldman, D. (1995). Emotional intelligence. New York: Bantam Books. Harrison, A. F., Bramson, R. M., Bramson, S., & Parlette, N. (1992). Your thinking profile: A self-help guide to understanding the way you think. Berkley, CA: Holland-Parlette Associates. Myers, I. B., & McCaulley, M. H. (1989/1995). Manual: A guide to the development and use of the Myers-Briggs Type Indicator. Palo Alto, CA: Consulting Psychologists Press. O’Roark, A. M. (1995). Occupational stress and informed interventions. In C. D. Spielberger, I. G. Sarason, J. M. T. Brebner, E. Greenglass, P. Laungani, and A. M. O’Roark (Eds.), Stress and emotion (Vol. 15, pp. 121–136). Washington, DC: Taylor & Francis. Schein, E. H. (1999). Process consultation revisited: Building the helping relationship. Reading, MA: Addison-Wesley (also check Schein, 1965/1969/1985/1987). Schnell, E. R., & Hammer, A. (1993). Introduction to the FIRO-B in organizations. Palo Alto, CA: Consulting Psychologists Press. Selye, H. (1974). Stress without distress. Philadelphia: J. B. Lippincott. Spielberger, C. D. (1983). Preliminary manual for the State-Trait Personality Inventory. Tampa: University of South Florida, Psychology Department. Spielberger, C. D., & Reheiser, E. C. (1994a). Job stress in university, corporate, and military personnel. International Journal of Stress Management, 1, pp. 19–31. Spielberger, C. D., & Reheiser, E. C. (1994b). The Job Stress Survey: Measuring gender differences in occupational stress. Journal of Social Behavior and Personality, 9, pp. 199–218. Spielberger, C. D., Westberry, L. G., Grier, K. S., & Greenfield, G. (1981). The Police Stress Survey: Sources of stress in law enforcement. Human Resources Institute Monograph, Series 3, no. 6. Tampa: University of South Florida, Psychology Department. Spielberger, C. D., & Vagg, P. R. (1999). Job Stress Survey: Professional manual. Odessa, FL: Psychological Assessment Resources.

RT44367_C017.indd 332

7/13/05 9:32:35 AM

Author Index Bagés, Nurì, 253 Biaggio, Angela M.B., 31 Bushman, Brad J., 77 Chacón-Puignau, Grace, 253 Dreman, Solly, 179 Egloff, Boris, 97 Feldman, Lya, 253 Fiksenbaum, Lisa, 271 González, María Jesus, 233 Greenglass, Esther R., 123, 271 Grinder, Robert, 32 Håseth, Kjell, 293 Hiew, Chok C., 143 Hodapp, Volker, 209 Korunka, Christian, 209 Krishna, Archna, 197 Krohne, Heinz Walter, 97

Laungani, Pittu, 3 Maier, Ehrentraud R., 209 Man, Frontisek, 45 O’Roark, Ann M., 313 Peréz, María Gabriela, 253 Pestemer, Iris A., 209 Prieto, José M., 233 Rayworth, Glendon, 143 Sharma, Sagar, 197 Spielberger, Charles D., 45, 77, 155, 197 Stanley, Julian C., 31–32 Táborský, F., 45 Tanzer, Norbert K., 155, 209 Toubiana, Joseph, 61 Vagg, Peter R., 77

333

RT44367_C018.indd 333

8/15/05 7:54:02 AM

RT44367_C018.indd 334

8/15/05 7:54:03 AM

Subject Index A Abilities, mismatch with job demands, 294 Absenteeism, 210 Academic achievement negative correlation with worry, 80 and problem behaviors in adolescent males, 143–145, 146–147 and test anxiety, 78–80 and types of problem behaviors, 149 unrelated to emotionality, 80 Action, vs. state orientation, 54 Action plans, formulating from JSS results, 322–333 Action potential factor, 192 Active coping in anger/hostility management, 133 correlation with vigilance, 109 Adaptability levels and mothers’/children’s perception of behavior problems, 184 in studies of maternal anger and children’s adjustment, 182 Adjustment outcomes, 191 and marital discord, 193 sons vs. daughters of divorced mothers, 180 Advancement opportunity, 271, 273, 275 in Canadian government vs. American corporate workers, 281 gender differences in perception of, 287 Advances in Test Anxiety Research, 82 Aggression, overlap with anger and hostility in AHA! Syndrome, 145

Agreeableness, and avoidance coping in females, 113 AHA! syndrome, 145, 151 relevance to school-related behavior, 146 Alarm stage, 20 Alcohol-drug disengagement gender differences in coping styles, 112 and passive coping/avoidance, 133 Ambiguity anxiety triggered by, 99 and vigilance/avoidance coping behaviors, 99 Analyses of variance (ANOVAs) Family Adaptability and Cohesion Evaluation scales, 184 German adaptation of Job Stress Survey, 221 Mainz Coping Inventory (MCI), 105 TAI across culture and gender, 86 Venezuelan adaptation of Job Stress Survey, 257–260, 259 Anesthesia, doses and trait vigilance, 102 Anger among school dropouts, 146 and anxiety during cross-cultural examinations, 155–156, 161–165 at being ignored, 15 during classroom lessons, 173 communicative expression of, 133–134 and coping, 132–133 in disabled persons, 42 experience, expression, control of, 36–42, 134 335

RT44367_C018.indd 335

8/15/05 7:54:03 AM

336

expression and cholesterol/lipid levels, 124 gender differences in management of, 134–136 in hypertensive patients, 42 maternal, 179–180 mediating role in adolescent problem behaviors, 145 and moral judgment, 31–33 overlap with hostility and aggression in AHA! Syndrome, 145 as predictor of progression of carotid atherosclerosis (PCA), 125 psychological implications of expression and management, 123–124 and risk-taking behavior, 40–42 role of social support in buffering, 131–132 as signal for problem-solving need, 318 and social context in CHD, 128–129 suppression and CHD in women, 135 suppression vs. disclosure in management of, 132 temporal and situational specificity during examinations, 165–167 Anger-Control, 37, 145, 147, 182 cultural differences in, 202 Hindi version, 197, 198 Anger control/suppression, and risk for CHD, 126 Anger expression, 37 association with occupational stress and Type-A behavior, 203 and children’s behavior problems, 188 class differences in, 42 cultural history and norms of, 200–201 equivalence of Hindi and English scales, 199–201 Hindi version, 197–198 maladaptive nature of, 128

RT44367_C018.indd 336

SUBJECT INDEX

predictor and criterion measures, 147–148 and school stress/problem behaviors, 143–145, 145–146 study methods, 146–148 and systolic blood pressure, 126, 127 Anger-Expression, 147, 182 Anger-In, 37, 131, 145, 147, 182 and bronchial asthma, 203–204 and coronary heart disease (CHD), 123, 124 Hindi subscale, 197, 198 in India, 201 inhibiting effect on performance and competitiveness, 41 negative correlation with extrapunitiveness, 39 and peptic ulcers, 203–204 positive correlation with impunitive responses, 39 Anger-Out, 37, 145, 147, 182 and bronchial asthma/peptic ulcer patients, 204 in coronary heart disease (CHD), 123–124 Hindi version, 197, 198 and hypertension, 42 positive correlation with extrapunitiveness, 39 and progression of carotid atherosclerosis (PCA), 125 Anonymous self-reports, in studies of delinquent behavior and school stress, 144–145 Anxiety ambiguity as trigger for, 99 and anger during cross-cultural examinations, 155–156 during classroom lessons, 173 cognitive theories of, 47 and depression as predictors of CHD, 125 failure to distinguish between trait and state, 77–78 and fluctuating coping mode, 101 impact on athletic competition, 46

8/15/05 7:54:04 AM

337

SUBJECT INDEX

and irrelevant cognitions, 47 and moral development research, 33–336 and moral judgment, 31–33 negative correlation with game performance, 45, 47, 53 non-interchangeability with stress, 78 relationship to self-esteem and locus of control, 33 and risk-taking behavior, 40–42 as signal for problem-solving need, 318 STAI measures of, 32–33 temporal and situational specificity during examinations, 165–167 Anxiety-provoking situations, in Mainz Coping Inventory (MCI), 103 Anxiety scales, 33 Arousal-motivated coping, 100 At-Risk Students, 146–147, 148 Athletic competition correlations between anxiety levels and performance in, 54, 55 female handball players in, 47–51 flow-experience in, 53–55 Trait-State anxiety, worry, and emotionality in, 45–47 Attention-rejection construct, 98 Attentional theory, and test anxiety, 79 Austrian teachers, Version 1 Job Stress Survey, 212 Automatic Thinking (AT) questionnaire, 253, 255–256 descriptive statistics for Venezuelan workers, 260 Average students, 146–147, 148 Aversive situations, 113 and coping behaviors, 99 predictability and controllability appraisals of, 105–107 Avoidance, 97–98. See also Cognitive avoidance and acknowledgment of alcoholdrug use, 112

RT44367_C018.indd 337

as arousal-motivated coping strategy, 100 correlation with positive problem orientation, 112 correlation with rational problem solving, 112 defined, 98 dissociation of verbal-autonomic responses in, 102 and intolerance of arousal, 100 and somatic arousal, 100 theory and measurement, 97–98 B Bacon, Francis, 4 Bank employees comparison with customs officials, 228 correlations of Norwegian JSS with Cooper Stress Check, 304 Job Stress Survey and, 209, 210 JSS participants and procedure, 222–224 JSS study results, 224–226 and Norwegian adaptation of Job Stress Survey, 293, 298 Behavior problems. See also Problem behaviors mean and standard deviations, 185 and respondent source, 185 Benchmarking, through Job Stress Survey, 316 Bettelheim, Bruno, 7–8 Bias, myth of freedom from, 4 Bilingual subjects, Hindi Anger Expression scale, 200 Blame, external vs. internal attributions, 173 Blood pressure. See also Systolic blood pressure and anger expression, 126 and avoidant coping style, 102 Blue-collar workers gender differences in stress severity perceptions, 259, 263 perception of stressors by, 253, 259

8/15/05 7:54:05 AM

338

Boredom, correlation with low stress levels, 315, 318 Bronchial asthma, and Anger-In scores, 203–204 Buchenwald concentration camp, 8 Burnout, 313 reduction through calibration consultation, 314 Burnout scale, for Austrian secondaryschool teachers, 217 Bus drivers, and Norwegian adaptation of Job Stress Survey, 298 C Calibration consultation intervention planning, 313 group level, 323–331 individual level, 319–323 Calibration procedures, 314–319 for individual special-courts judge, 319–323 Canadian government employees, job stress and gender among, 271–276 CBCL Externalizing scale, 183. See also Externalizing behavior problem scores CBCL Internalizing scale, 183. See also Internalizing behavior problem scores CBCL Total Behavior Problems scale, 183. See also Total Behavior Problems scale Chaotic adaptability level, 182 Child Behavior Checklist (CBCL) Hebrew adaptation, 183 maternal State-Anger and children’s problem behaviors, 189 Children, self-report of behavior problems by, 180 Cholesterol and lipid levels, relationship to anger expression, 124 Chronic anxiety gender differences in, 41 and T-Anger, 41

RT44367_C018.indd 338

SUBJECT INDEX

Clerical workers Spanish adaptation of Job Stress Survey for, 233 in Venezuelan adaptation of Job Stress Survey, 255, 259 Clever Hans, 6 Cognitive Anxiety subscale, 47, 48 and player performance, 52, 53 Cognitive avoidance, 97–98 correlation with positive affectivity, 113 correlation with positive reinterpretation and growth, 111 defined, 98 and depression, 133 and extraversion in males, 113 measurement of, 102–113 principal-components analysis, 109 in response to ego-threat (CAV-E), 103 in response to physical threat (CAV-P), 103 Cognitive conflict and postconventional thinking, 35 relation to S-anxiety, 36 Cognitive Interference Questionnaire (CIQ), 45, 48, 54 Cognitive switch, as response to strangeness of hospital environment, 21–22 Communication issues, LIFO and Myers-Briggs Type Indicator for, 316 Communicative expression of anger and hostility, 133–134 gender differences in, 135 Competition for advancement severity as stressor among Venezuelan workers, 260 in Venezuelan adaptation of Job Stress Survey, 256 Competitive sports. See Athletic competition Competitive State Anxiety Inventory2 (CSAI-2), 47, 48 and player performance, 52

8/15/05 7:54:05 AM

SUBJECT INDEX

Concentration test, 160 Confirmatory factor analysis, limitations of, 81–82 Conflict resolution with other departments, 323, 330 State-Trait Personality Inventory and FIRO-B for, 316 Confrontation, maladaptive nature of, 128 Connected level of family cohesion, 182 lower level of child behavior problems in, 191 Conscience, development of, 34 Consistent cognitive avoidance, 100 Consistent vigilance, 100 Construct validity job stress frequency factors in Venezuelan workers, 266–267 job stress severity factors in Venezuelan workers, 268–269 Mainz Coping Inventory (MCI), 108–112 Spanish adaptation of Job Stress Survey, 239, 243 Venezuelan adaptation of Job Stress Survey, 253, 261 Consultant-client partnership approach, 313, 319 Consulting psychologists, 313 Controllability and coping behaviors, 105 gender differences in threat situation appraisals, 107 Cook-Medley Hostility scale, 125 Cooper Stress Check, 297 and concurrent validity of Norwegian adaptation of JSS, 308 correlations with Norwegian adaptation of JSS, 304 COPE Inventory, 108 Coping modes construct validity as job stress frequency factor, 267

RT44367_C018.indd 339

339

construct validity as job stress severity factor, 268 differences between American and Venezuelan JSS, 261 frequency scores in Venezuelan adaptation of Job Stress Survey, 258 gender differences in, 262 model of, 98–102 Coping techniques. See also Cognitive switch active vs. passive in anger management, 133 anger and gender differences, 135 and anger/hostility management, 132–133 cognitive avoidance, 97–98 in concentration camps, 8 deficits in psychopathic persons, 101 depersonalization, 129 emotion-focused, 111 in face of medical prognoses, 18–19 for fear of death, 25 habitual vs. actual, 98 and recovery rates, 26–27 in severe illness, 8, 15, 22, 24–25 in test situations, 66 trait vs. state, 98 vigilance, 97–98 Coping variables gender differences in, 110–111 influence on repressive emotional discreteness, 102 MCI relationships with other, 108–112 principal-components factor analyses (PCA), 108, 109 Coronary heart disease (CHD), 274 and anger control/suppression, 126, 204, 273 anxiety and depression as predictors of, 125 contribution of high job demands and lack of control to, 261 gender tendencies towards, 275 interactive approaches to study of, 125–128

8/15/05 7:54:06 AM

340

job stress as risk factor for, 272–273 multiple emotions and, 124–125 and reactivity, 126 risk factors for, 123 social context of anger/hostility and, 128–129 and Type-A Behavior Pattern, 123 Corporate employees, job stress comparison with government employees, 271–272 Court administrators, stressor ratings, 325, 329 Covariance matrices, in confirmatory factor analysis, 81–82 Crisis management, 322 construct validity as job stress frequency factor, 267 construct validity as job stress severity factor, 268 stressor rating in judicial community, 326, 328 as stressors for judges and lawyers, 328 in Venezuelan adaptation of Job Stress Survey, 256 Criterion variables, 116 gender differences in correlation of MCI with, 114–115 Critical incident technique, 55 Cross-Cultural Anxiety, 82 Cross-cultural equivalence, 198 of Hindi and English Anger Expression scales, 199–201 Cross-cultural research addition of items to Norwegian JSS, 298 anger and anxiety during examinations, 155–173 Anger Expression scale, 198 ANOVAs of TAI, 86 Examination Anger and Anxiety Scales (EAAS), 156–157 German adaptation of Job Stress Survey, 211–213 Hindi version of Anger Expression scale, 197–198 letter cancellation test, 160

RT44367_C018.indd 340

SUBJECT INDEX

in maternal anger and children’s behavior problems, 193 Norwegian adaptation of Job Stress Survey, 293 reasoning test, 160 Spanish adaptation of Job Stress Survey, 233–250 spatial ability test, 160 in test anxiety, 61–63, 81 Three-Dimensional Cube Comparison Test, 160 Venezuelan adaptation of Job Stress Survey, 253–254 Viennese Matrices Test, 160 Cultural differences, in test anxiety, 67–68, 69–70 Customer service reductions, 210 Customs officials, job stress survey ratings, 228 Cynical Distrust scale, 125, 128, 135 Czechoslovakian National Female Handball Team, 45, 48 D Dachau concentration camp, 8 Daily activities, lack of control over, 21 Danger, psychopathic hyporesponsivity to, 101 Danger control, vigilance and, 100 Death fear of, 16–17, 26 repression of fear of, 25 Death rates, and lack of social ties, 130 Decision making stressor level in judicial community, 326 as stressors for judges and lawyers, 328 as top stressor, 322 Delinquent Activities Questionnaire (DAC), 147–148 Delinquent behavior. See Problem behaviors Demands/Control model of job stress, 294

8/15/05 7:54:07 AM

341

SUBJECT INDEX

Denial among professionals and senior management, 324 in COPE subscale, 111–112 influence on JSS scores, 319 JSS assessment of, 318 Dentist scenario in Mainz Coping Inventory (MCI), 103 predictability and controllability appraisals of, 105 Dependency, and hostility in coronary patients, 125 Depersonalization buffering effects of social support on, 131 and job stress, 128–129, 130 Depression and anxiety as predictors of CHD, 125 and cognitive avoidance, 133 correlations with Norwegian JSS, 305 transformation from chronic anxiety, 41 and vigilance/avoidance coping styles, 113 Descartes, method of systematic doubt, 4–5 Disability anger reactions of patients with, 42 fear of permanent, 26 Disclosure, vs. suppression in anger/ hostility management, 132 Disengagement and behavior problems by respondent source, 185 as degree of family cohesion, 182 Dissertation Abstracts International, 82 Divorced mothers family cohesion relation to behavior problems, 180 family structure and children’s behavior problems, 179–180 Doctor-patient relationships, 3, 23–24

RT44367_C018.indd 341

Domain-specific trait instructions, 157, 167–168, 171 Downsizing, 271, 273 Drug use, 148, 149, 150. See also Alcohol-drug disengagement correlation with anger expression in adolescent males, 143 and proneness to anger, 146 E EASI-III Temperament Survey, 112 Educational Resources Information Center (ERIC), 82 Ego defense, 39 Ego-threatening situations, coping strategies for, 97–98 Emotional arousal, intolerance of, 100, 115 Emotional blending, in test situations, 157, 169–170 Emotional expression, vs. emotional experience, 123–124 Emotional support, seeking behaviors and vigilance, 109 Emotionality, 77 and academic achievement, 78–80 in athletic competition, 45–47 cross-cultural research in test anxiety, 81 factor structures for TAI, 83 increase with somatic arousal, 99 mean factor cosines as function of culture and gender, 86, 87 in test anxiety, 64, 65 Emotions and coronary heart disease (CHD), 124–125 gender differences in focus on and venting of, 112 universal recognition of, 62, 69 Enmeshed level of family cohesion, 182 and behavior problems by respondent source, 185 mothers’ perception of fewest behavior problems at, 184

8/15/05 7:54:07 AM

342

Environmental work factors for Canadian vs. American workers, 281 construct validity as job stress severity factor, 269 differences between American and Spanish JSS, 248, 261 gender differences for Canadian government workers, 286 gender differences in stress perceptions, Venezuelan workers, 259, 263 noisy work areas, 271 severity in Venezuelan adaptation of Job Stress Survey, 258, 260 as sources of occupational stress, 273 in Spanish adaptation of Job Stress survey, 241, 242, 249 in Venezuelan adaptation of Job Stress Survey, 256–257 European Foundation for the Improvement of Living and Working Conditions, 235 Exam scenario coping reactions to, 104 in Mainz Coping Inventory (MCI), 103 Examination Anger and Anxiety Scales (EAAS), 156–157 English state version, 167 English trait version, 159, 163 German state version, 160, 166 German trait version, 159, 162, 164 mean and standard deviations for English trait studies, 161, 162, 165 Examination stress anger and anxiety in cross-cultural situations, 155–173 blended positive-negative emotions during, 156, 157 domain-specific, 157 rating scale formats, 158 statistical analyses, 158–159 subjects and procedures, 159–161 trait-state distinctions, 157–158

RT44367_C018.indd 342

SUBJECT INDEX

vs. test anxiety, 78–79 Exhaustion stage, 20 Experience-Sampling Form, 54 Experimental subjects, preconceptions of, 6 Experimenters, separation from objects of study, 4 Exploratory factor analyses (EFAs), German adaptation of Job Stress Survey, 213 Externalizing behavior problem scores, 185, 186, 189 state anger by gender interaction, 188 Extrapersonal factors, in experience of stress, 20–22 Extrapunitiveness, 39 positive correlation with Anger-Out and T-Anger scales, 39 F Face validity, 238 Factor-analytic studies, on test anxiety, 83 Factor loadings, 92 in cross-cultural test anger/anxiety studies, 158–159 Examination Anger and Anxiety Scale (English), 162 Examination Anger and Anxiety Scale (German), 164, 167, 168, 172 Job Stress Survey for Austrian bank employees, 225, 227 Job Stress Survey for Austrian primary school teachers, 213–214 Job Stress Survey for Austrian secondary school teachers, 218 Job Stress Survey for Canadian government workers, 277–278 oblimin-rotated, 213 Spanish adaptation of Job Stress Survey, 233, 241

8/15/05 7:54:08 AM

343

SUBJECT INDEX

Venezuelan adaptation of Job Stress Survey, 256–257 Factor structures comparison of Spanish and American JSS, 246–248 factor cosines, 84, 87 gender comparisons for Norwegian JSS, 302 gender comparisons for TAI, 83 gender comparisons with more than two factors, 85 German adaptation of Job Stress Survey, 217, 220, 221 Kaiser’s method for comparing, 82, 91–95 Norwegian adaptation of JSS, 298–301 pure oblique vs. orthagonal, 95 Severity and Frequency items for Norwegian JSS, 300 Spanish adaptation of Job Stress Survey, 239–242, 243 Factorial validity, anger-anxiety in test situations, 170–171 Family adaptability and behavioral problems, 190 and children’s adjustment, 179–180 Family Adaptability and Cohesion Evaluation Scales (FACES III), 182–183 Family cohesion and children’s behavior problems, 184–186, 190 curvilinear relation to adaptability and adjustment, 179 four degrees of, 182 Family cohesiveness and flexibility, 179–180 and maternal anger, 179–180 Family courts agency mean JSS stress index, severity, and frequency scores, 329 stressor calibration for, 323–325 Family stability, 190 Family structure anger and gender differences, 192

RT44367_C018.indd 343

and children’s behavior problems, 179 Fear of danger, 99–100 Fear of fear, 99 Feeling states, in threat situations, 99 Fellow workers stressor level in judicial community, 327 as stressors for judges and lawyers, 328 Firefighters and Norwegian adaptation of Job Stress Survey, 293, 298 Severity scores in Norwegian JSS, 306 FIRO-B, as assessment instrument for conflict resolution, 316 Flexible adaptability level, 182 Flow-experience, and athletic performance, 53–55 Fluctuating coping, 100, 101 Framingham Heart Study, 134 Frustration coping techniques for, 24–25 at lack of progress, 18 G Gender differences in aggressive behaviors, 135 in agreeableness correlation with avoidance coping, 113 anger and family structure, 192 in anger expression, 39–40, 201 in anger/hostility management, 134–136 Anger-In scores in India, 202 behavior problems in children of divorced mothers, 180 in communicative expression, 135–136 in correlations between MCI and criterion variables, 114–115 in cross-cultural test anxiety scores, 61, 64 in divorced mothers’ children’s behavior problems, 191

8/15/05 7:54:09 AM

344

SUBJECT INDEX

in effects of anger suppression on CHD, 135 in extraversion correlation with cognitive avoidance, 113 factor cosines for TAI with two factors, 84 in focus on and venting of emotions, 112 German adaptation of Job Stress Survey, 220 and Hindi Anger Expression scale, 202–204 in job stress among government vs. corporate employees, 271–276 in job stress among military employees, 330 in job stress in judicial community, 329–331 JSS item severity and frequency in Canadian workers, 286–288 lack of support and environmental problems as stressors, 253 in MCI correlations with coping variables, 110–111 minimization with equal work and organizational responsibilities, 288 in negative affectivity, 113 in Norwegian adaptation of Job Stress Survey, 293 in occupational level and job stress in Venezuela, 253, 255, 262–263 in perceived Severity scores in judicial community, 329, 330 in perception of deadlines as stressor, 258 in predictability and controllability appraisals of threat situations, 106–107 in ratings of specific JSS stressor events, 280–281 in response to job demands, 262 in responses to job stress, 275

RT44367_C018.indd 344

restraint and control of emotions in India, 201 in risk-taking behavior and anger expression, 40–41 in severity and frequency scores for Norwegian JSS, 299 severity of occupational stressors in Venezuela, 253 in social support, 136 in specific JSS stressor events, 280–281 in stress severity perceptions, Venezuelan workers, 257, 259 in T-Anxiety and moral maturity, 36, 70 in test anxiety, 67–68, 70–71 in trait anger and Anger-In relationships, 41 in wage and salary levels, 331 Gender factors in English-language TAI, 83 invariance and weighted ANOVAs, 86 more than two factors, 85 two-factor studies, 84 General Adaptation Syndrome (GAS), 19–20, 272 General Health Questionnaire, correlations with Spanish adaptation of JSS, 242 General Reactions Survey, and Spanish version of JSS, 236 Government employees Canadian participants, 276–277 increasing demands on, 274 job stress comparison with corporate employees, 271–272 Group of people scenario gender differences in controllability appraisals, 107 in Mainz Coping Inventory (MCI), 103 predictability and controllability appraisals of, 105 Guilt, transformation from chronic anxiety, 41

8/15/05 7:54:09 AM

345

SUBJECT INDEX

H Habitual coping behaviors, 103 S-R inventory of, 105 vs. state coping behaviors, 98 Health, role of social support in, 129–132 Health care workers and Norwegian adaptation of Job Stress Survey, 298 Spanish adaptation of Job Stress Survey for, 233–234 Health-Intensity, correlations with Norwegian JSS, 305, 306 Heisenberg, Werner, 5 High achievers, 146–147, 148 denial of stress by, 324 need to be free of workplace hassles, 314 High Frequency/Low Severity items, negative consequences of, 321–322 Himachal Pradesh University, 199 Hindi Anger Expression scale, 197 construction and validation, 197–198 means, standard deviations, and alpha coefficients, 200 research studies in, 202–204 Homes-Rahe Social Readjustment Rating Scale, 296 Hospital staff Job Stress Survey and, 209, 210 JSS study results, 220–222 JSS subjects and procedure, 219–220 necessity of re-educating, 27 Hostility communicative expression of, 133–134 and coping, 132–133 and dependency in coronary patients, 125 gender differences in management of, 134–136 and interpersonal stressors, 127–128 and job stress, 128–129

RT44367_C018.indd 345

overlap with anger and aggression in AHA! Syndrome, 145 as predictor of progression of carotid atherosclerosis (PCA), 125 psychological implications of expression and management, 123–124 role of social support in buffering, 131–132 and self-involvement, 133 and social context in CHD, 128–129 suppression vs. disclosure in management of, 132 and systolic blood pressure, 127 in transactional model, 129 Hypertension and anger control/suppression, 126, 203, 204 anger responses and, 42 Hyporesponsivity, in psychopaths, 101 I Impunitiveness, 39 positive correlation with Anger-In, 40 Inadequate Personnel/Equipment, 271 construct validity as job stress frequency factor, 267 in Venezuelan adaptation of Job Stress Survey, 258 Inadequate salary, 281, 322 gender differences in perception of, 287 as stressors for family court administrative staff, 329 Independence, loss in severe illness, 9, 10 Indeterminacy principle, 5 Individuality, loss of, 9 Inductivist fallacy, 5 Inexperienced driver scenario, in Mainz Coping Inventory (MCI), 103 Infinite regress, danger of, 7 Informed Heart, The, 8

8/15/05 7:54:10 AM

346

Institutionalization, psychological responses to, 21 Instituto Nactional de Seguridad e Higiene en el Trabajo, 235 Insufficient personnel, 280, 286 gender differences in perception of, 287 Interactive processes, in stress, 20 Interamerican Society of Psychology, 31–32 Interdepartmental conflicts, 287 Internal consistency reliability German adaptation of Job Stress Survey, 220 Guttman split-half coefficients for Venezuelan JSS, 256–257 Hindi Anger Expression scale, 197, 201–202 Norwegian version of Job Stress Survey, 301 Severity and Frequency scales, Venezuelan JSS, 257 Venezuelan adaptation of Job Stress Survey, 253, 256–257, 261 Internalizing behavior problem scores, 185, 187, 189, 190 and adaptability, 186 Interpersonal factors in experience of stress, 20 with hospital staff, 22–25 Interpersonal stress, and hostility/ reactivity, 127–128 Interruptions as stressors for judges and lawyers, 328 as stressors for judicial community, 326 as top stressor, 322 Intrapersonal factors in experience of stress, 20 in severe illness, 25–26 Intrapunitiveness, 39 Intrinsic arousal levels, and risk-taking behaviors, 40–41 Introspection, role in psychological research, 7

RT44367_C018.indd 346

SUBJECT INDEX

Investigator as patient, 14 Isolation Perception among Venezuelan workers, 261 gender differences in Venezuelan workers, 263 Item-remainder (IR) correlations, 202 J Jewish mothers, 193 in studies of family cohesion and children’s adjustment, 181 Job Annoyance, in Spanish adaptation of JSS, 249 Job demands, mismatch with worker abilities, 294 Job dissatisfaction, 274 overload and, 262 in Venezuelan adaptation of Job Stress Survey, 256 Job interview scenario, in Mainz Coping Inventory (MCI), 103 Job latitude, importance in controlling stress, 295 Job Pressure, 209, 227, 296, 316 among Austrian bank employees, 228 among Austrian hospital staff, 220, 221, 223 among Austrian primary school teachers, 214, 218 among Austrian secondary school teachers, 217, 219 among Canadian government workers, 277 comparisons between Spanish and American JSS versions, 246, 247 correlations with Spanish JSS, 242 equivalence with work content, 308 factor structures by gender for Norwegian JSS, 302 gender differences in Norwegian Job Stress Survey, 301 intercorrelations with other job stress factors, 215

8/15/05 7:54:11 AM

SUBJECT INDEX

in Norwegian adaptation of Job Stress Survey, 293 Severity and Frequency in Norwegian JSS, 300, 301–302 in Spanish adaptation of Job Stress Survey, 233 for special courts judge, 317 and State-Anxiety in Norwegian workers, 306 Job Satisfaction scale Austrian bank employees, 224 Austrian secondary-school teachers, 217 Job stress assessment instrument selection, 314–319 calibration methodology, 316 calibration procedures, 314–319 comparative frequency ratings for Canadian vs. U.S. employees, 284–285 comparative severity ratings for Canadian vs. U.S. employees, 282–283 comparison between Canadian government workers and U.S. corporate employees, 278–281 consultant-client partnership approach to diagnosis and treatment of, 313 costs of, 274, 293 Demands/Control model of, 294–295 differences between American and Venezuelan workers, 261 frequency of occurrence among Canadian government workers, 277–278, 278–279 Frequency ratings, 211 gender differences in, 275–276 and gender in comparison of government and corporate employees, 271–276 and hostility, 128–129, 130

RT44367_C018.indd 347

347

in judicial community, 316 and occupational level/gender differences in Venezuela, 253–254 perceived severity and frequency by judicial community, 325 physical consequences for men, 275 and physical illness, 315 severity levels among Canadian government workers, 277, 278–279 severity levels as function of gender and organizational level, 259, 260 Severity ratings, 211 signs of denial, 319 sources and consequences, 273–275 time pattern for specific sources of, 250 unclear definitions of, 254 Job stress reduction, in judicial community, 327–329 Job Stress Survey (JSS), 128, 271, 274, 276, 277, 295–296 adapting stressor items to specific disciplines, 319 additional elements in Spanish version, 234–235 alpha coefficients, 222, 223, 233, 238 for Austrian primary-school teachers, 213–216 for Austrian secondary-school teachers, 216–219 as benchmarking tool for calibrating optimal productivity, 313, 316 calibrating severity for individual clients, 321 Canadian participants in, 276–277 choice of complementary tests in Spanish adaptation, 235–236 comparison of Norwegian and American responses, 301, 307

8/15/05 7:54:11 AM

348

SUBJECT INDEX

comparison of Spanish and American versions, 246–248 concurrent validity in Norwegian adaptation, 308 concurrent validity in Spanish adaptation, 235–236, 242–246, 248 construction and development of Spanish version, 234–237 content validity of Spanish version, 238–239, 243 convergent and discriminant validity, 215–216, 217, 219, 234 correlations with General Health Questionnaire in Spain, 242–246 descriptive statistics for Venezuelan adaptation of Job Stress Survey, 257, 258 development of German adaptation, 211–213 development of Norwegian adaptation, 297–298 differences between American and Spanish study results, 247–248 explaining individual profile and calculations, 320 factor structure, 217, 234, 235 factor structure of Norwegian adaptation, 298–301 Frequency and Severity correlations in Norwegian adaptation, 293 frequency response category, 250 Frequency scale in Spanish version, 240, 245, 246 gender differences in ratings of specific stressor events, 280–281 German adaptation of, 209–210 Guttman split-half internal consistency reliability coefficients, 256–257

RT44367_C018.indd 348

implications for future studies (Spanish version), 249–250 intercorrelations among subscales for Norwegian adaptation, 302–303, 304 intercorrelations with Organizational Climate scales, 226 intercorrelations with other measures of occupational stress for Norwegian adaptation, 304 internal consistency reliability, 220 interpreting high and low ratings, 318 multistudy validation, 210 need to include discipline-specific items, 319 negative consequences of high Frequency/low Severity items, 321–322 Norwegian adaptation, 293, 297 perceived severity and frequency of stressor occurrence, 296 psychometric properties of Norwegian version, 298–301 psychometric properties of Spanish version, 237–238 reliability and factor analyses of Venezuelan adaptation, 256–257 report for special courts judge, 317 sample description, 236–237 sample selection in Spanish study, 236 Severity and Frequency ratings in judicial community, 316–317 Severity and Frequency ratings in Norwegian adaptation, 299 Severity index in Spanish version, 238, 240, 245, 246 severity ratings of anchor item, 216, 219, 222, 224, 226, 227 Spanish adaptation of, 233–234 stress severity factors in Venezuelan adaptation, 258

8/15/05 7:54:12 AM

SUBJECT INDEX

test application in Spanish study, 236 validity of Norwegian adaptation, 303–306 Venezuelan adaptation, 255 Job turnover, 210 Judges, stressor ratings, 325–326, 328 Judicial community gender differences in job stress, 329–331 low Severity and Frequency of stressor ratings, 325 principal stressors for five professional work groups, 327 professional work groups within, 324, 325 recommendations for reducing job stress in, 327–329 stress assessment for, 323–324 Judicial support staff and assistants low stressor levels of, 326 stressor ratings, 328 K Kaiser’s method for comparing factor structures, 82 geometrical description of, 91–95 L Lack of Organizational Support, 209, 227, 274, 296, 316 Austrian bank employees, 225 Austrian hospital personnel, 220, 221, 223 Austrian primary school teachers, 213, 214 Canadian government workers, 277 comparisons with relational factors in Spanish JSS, 242, 246, 247 construct validity as job stress frequency factor, 266 construct validity as job stress severity factor, 268

RT44367_C018.indd 349

349

correlations with Cooper Stress Check among Norwegian workers, 304 correlations with OCQ scales, 217, 219 equivalence with work context, 309 factor structures by gender for Norwegian JSS, 203 gender differences in Norwegian JSS, 301 gender differences in Venezuelan employees, 258, 259 gender ratings in judicial community, 330 intercorrelations with other job stress factors, 215 in Norwegian adaptation of Job Stress Survey, 293 relative frequency among government vs. corporate workers, 280 Severity and Frequency in Norwegian JSS, 300, 301–302 severity in Venezuelan adaptation of Job Stress Survey, 258 in Spanish adaptation of Job Stress Survey, 233, 260 for special courts judge, 317 in Venezuelan adaptation of Job Stress Survey, 256, 258, 259 Lawyers, stressor levels, 325 Layoffs, 271 Learning, inhibited by lack of control, 295 Learning Anxiety Scale (LAS), 66 correlations with PETER-TQ, 65 Learning Skills Questionnaire, 66 Letter cancellation test, 160, 166 Librarians, and Norwegian adaptation of Job Stress Survey, 298 Life-threatening illness, stress in, 3–8 LIFO, as stress assessment instrument, 316 Literature search procedures, in TAI studies, 82–83

8/15/05 7:54:13 AM

350

M Macroanalytic analysis, 115 Maintenance workers, Spanish adaptation of Job Stress Survey for, 233 Mainz Coping Inventory (MCI), 98 appraisal of MCI situations, 104–107 construct validity of, 108–112, 116 construction and validation of, 97 Cronbach alpha coefficients, 107 design of, 102–104 eigenvalues, 108 eight scenarios of, 103 gender differences in correlations with coping variables, 110–111 internal structure, 107–108 macroanalytic analysis and, 115 participants in, 104–105 predictability and controllability appraisals, 105–106 relationships with criterion variables, 112–113, 114–115 statistical and psychometric properties, 107 subscale correlations, 108 subscale means, standard deviations, and reliabilities, 107 Managerial positions gender imbalances in, 276 and Norwegian adaptation of Job Stress Survey, 298 Marital discord, and poor adjustment in children, 193 Marital status, and perception of stress, 254 Marlowe-Crowne Scale, 112 Maslach Burnout Inventory, 131 Maternal anger and behavior problems, 180, 187–190 and children’s adjustment, 179–180 and daughters’ behavior problems, 188

RT44367_C018.indd 350

SUBJECT INDEX

and family cohesiveness/ adaptability, 179–180 long-term effects of, 181 Maternal assertiveness, 192, 193 and children’s behavior problems, 191 Medical problems, job-related, 210 Meeting deadlines, 322 gender perceptions of stress levels, 258 stressor level in judicial community, 327 as stressors for judges and lawyers, 328 as top stressor, 322 Mergers, 271 Method of Quantitative Evaluation (MQE), 52 Military employees, gender differences in job stress, 330 Mistake on the job scenario gender differences in predictability, 107 in Mainz Coping Inventory (MCI), 103 Monitor-Blunter Style Scale (MBSS), 103 Monitoring-blunting construct, 98, 99 Moral development, 32 Moral judgment and anger, 31–33 and anxiety, 31–33, 33–36 relation to state and trait anxiety, 35–36 Moral law, higher than legal law, 35 Moral maturity, correlation to Tanxiety in females, 36 Myers-Briggs Type Indicator as stress assessment instrument for communication issues, 316 use in judicial community, 316 Mykletun Stress Check, 304 N Need persistence, 39

8/15/05 7:54:13 AM

SUBJECT INDEX

Negative problem orientation correlation with vigilance, 112 gender differences in Canadian government workers, 287 Negative thinking gender differences in, 263 and perception of stress, 254 NEO Five Factors Inventory, 113 Neuroticism, and vigilance, 116 Neutrality, myth of, 5 Nonserious delinquency behaviors, 148, 150 Norms, internalization of, 21–22 Nurse-patient relationships, 3, 23 Nurses and Norwegian adaptation of Job Stress Survey, 293, 298 Severity scores in Norwegian JSS, 306 Type-A behavior and anger aggression in, 203 O Objectivity myth of scientific, 5 in severe illness, 15 as unattainable ideal, 6 Objectivity-subjectivity dichotomy, 6 Observations goal of testing hypotheses through, 5 myth of neutrality, 5 Obstacle dominance, 39 Occupational group differences in German adaptation of Job Stress Survey, 22 in Norwegian adaptation of Job Stress Survey, 293, 307, 308 Occupational Stress Indicator, 308 Occupational Stress Inventory (OSI), 297 Occupational stress theory, 294–295 and three-factor analysis, 244–245 One-size-fits-all stress assessment, 323 Optimal productivity, 314

RT44367_C018.indd 351

351

Job Stress Survey as benchmarking tool for calibrating, 313 Optimal-stressor conditions, importance of striving for, 313 Optimism, and disclosure behaviors, 132 Organizational Climate Questionnaire (OCQ) bank employee version, 224 intercorrelations with Job Stress Survey, 226 teacher version of, 217 Organizational level and age in Venezuelan employees, 255 gender differences for stressors among Venezuelan workers, 258–259, 262–263 and job stress in Venezuelan employees, 253–254 and perception of Overload in Venezuelan workers, 253 Overload, 313 construct validity as job stress frequency factor, 267 construct validity as job stress severity factor, 268 frequency scores in Venezuelan adaptation of Job Stress Survey, 258 gender differences in perception by Venezuelan workers, 263 organization level differences in perception of stress from, 259 severity in Venezuelan adaptation of Job Stress Survey, 258, 260 as stressor in Venezuelan adaptation of Job Stress Survey, 256, 261 Overtime factor loading in Spanish JSS, 241 in Spanish adaptation of Job Stress Survey, 234–235, 239

8/15/05 7:54:14 AM

352

SUBJECT INDEX

as top stressor for men in judicial community, 330 P Paperwork, stressor level in judicial community, 327 Parent-child role boundaries, 190 Parental anger, increased family conflict with, 180 Parental control, difficulties in divorced mothers establishing, 180 Partial-area theory, 115 Participant observation, 6–8 Passive coping, in anger/hostility management, 133 Patients, preoccupation with own fears and anxieties, 22–23 Peak performance zone, 316 Pearson product-moment correlations Anger variables and problem behaviors, 150 German adaptation of Job Stress Survey, 213 in studies of problem behaviors in adolescent males, 148 Trait-Anger and Anger Expression Scales, 149149 Pennebaker Inventory of Limbic Languidness, 113 Peptic ulcers, and Anger-In scores, 203–204 Performance, correlations with anxiety, self-confidence, and task-irrelevant cognitions, 51–55 Person-Environment Fit theory, 294, 295 Personality traits effects on recovery rates, 19 vs. psychological states, 78 Pictorial Evaluation of Test Reactions-State Questionnaire (PETER-SQ), 63

RT44367_C018.indd 352

previous research with, 65–66 Pictorial Evaluation of Test Reactions-Trait Questionnaire (PETER-TQ), 61, 63 alpha coefficients for eight countries, 67, 68 correlations with Learning Anxiety Scale, 65 cross-cultural validity of, 69 mean and standard deviations, 71 previous research with, 65–66 user-friendliness of, 67 Pictorial Experimental Paradigm, 62 Picture Frustration Test (PFT), 38 Planning and active coping, 111 link with vigilance coping strategy, 109 Police officers correlations of Cooper Stress Check with JSS, 304203 and Norwegian adaptation of Job Stress Survey, 293, 307 Police Stress Survey, 277 Poorly motivated coworkers, 280 Positive and Negative Affect Schedule, 112–113 Positive problem orientation, correlation with avoidance, 112 Positive thinking gender differences in, 263 and perception of stress, 254 Postconventional thinking, 34, 35 Postdivorce criteria, 181 Practice game, 55 performance and anxiety measures in, 53 Predictability and coping behaviors, 105 gender differences in threat situation appraisals, 107 Primary-school teachers Job Stress Survey and, 209, 210 JSS participants and procedures, 213

8/15/05 7:54:15 AM

353

SUBJECT INDEX

JSS survey results, 213–216 Problem Behavior Theory, 144, 151 Problem behaviors adaptability and, 186–187 and anger expression/school stress, 143–145, 145–146 children’s perceptions of own, 183 Delinquent Activities Questionnaire (DAC) for, 147–148 and family cohesiveness/flexiibility, 179–180 as maladaptive coping responses, 145 and maternal anger, 179–180, 187–190 means and standard deviations, 187 mothers’ vs. children’s perceptions of, 184 predictor and criterion measures, 147–148 study methods, 146–148 study participants, 146 Problem-solving instruments, 316 Procrastination of Learning Assignments, 66 Productivity calibrating in workplace, 313–314 and rational problem solving, 315 Productivity improvements, 314 Productivity reductions, 210 Progression of carotid atherosclerosis (PCA), anger and hostility as predictors of, 125 Property damage behaviors, 148, 150 Psychiatric hospital job stress studies, 219–222 Psychological Abstracts, 82 Psychological assessment, 331 role in consultant-client relationship, 316 Psychological factors, influence on recovery rates, 19 Psychological states, vs. personality traits, 78 Psychopaths, deficits in coping resources, 101

RT44367_C018.indd 353

Q Quantum mechanics, 5 R Rational problem solving, 315 correlations with vigilance and avoidance, 112 Reaction to Tests (RTT) inventory, 61 Reactivity and coronary heart disease (CHD), 126 and interpersonal stressors, 127–128 Reactivity Hypothesis, 126 Reading tests, trait anger and anxiety during, 167–168 Reciprocal determinism, 34 Recognition and Advancement construct validity as job stress frequency factor, 267 frequency scores in Venezuelan adaptation of Job Stress Survey, 258 in Venezuelan adaptation of Job Stress Survey, 256 Recollected states, 163–164 in exam anger/anxiety studies, 158, 160 Recovery rates and demands on coping resources, 26–27 effects of psychological factors on, 19 Religion, correlation with MCI scales, 111 Repression, 25 guilt as dominant emotion after failure in, 102 Repression-sensitization construct, 98, 99 Repressive emotional discreteness, 102 Researchers, as own subjects, 7, 14 Resistance stage, 20 Resistance to extinction, 34

8/15/05 7:54:15 AM

354

Respondent bias, and ratings of children’s behavior problems, 194 Response factors, in stress, 20 Restraint coping, link with vigilance, 109 Rigid adaptability level, 182 Risk-taking behavior and intrinsic arousal levels, 40 relation to anxiety and anger, 40–42 Role ambiguity addition to Norwegian JSS, 298, 307 in Norwegian adaptation of Job Stress Survey, 294 in Spanish version of JSS, 235 Role conflict addition to Norwegian JSS, 298, 307 and Norwegian adaptation of Job Stress Survey, 293 in Spanish version of JSS, 235 S S-Anger, as emotional state, 36–37 S-Anxiety scores after dissonance induction, 33 correlations with T-Anger, 41 relation of moral judgment to, 35–36 Salary freezes, 275 Sample size of 1, 4 Sample sizes, in factor structure studies of TAI, 84 Sampling validity, 238–239 School failure, 144 School stress and adolescent problem behaviors, 143–145 and anger expression, 143–145, 147 predictor and criterion measures, 147–148 study methods, 146–148 study participants, 146 teacher ratings, 147 and Trait Anger scale, 147 Secondary-school teachers Job Stress Survey and, 209, 210

RT44367_C018.indd 354

SUBJECT INDEX

JSS participants and procedure, 216–217 JSS study results, 217–219 Self-confidence among Spanish workers, 244 and player performance, 51–55 precompetition measurement of, 48–49 and use of cognitive avoidance, 113 and vigilance/avoidance coping styles, 113 vs. anger suppression, 132 vs. anxiety, 47 Self-Confidence subscale, 47 Self-Control Schedule, 66 Self-fulfilling prophecy, 19 Self-preoccupation, effect on performance decrements, 79 Self-report questionnaires behavior problems in children of divorced mothers, 180 German adaptation of Job Stress Survey (JSS), 210–211 limitations in cross-cultural research, 62 in studies of delinquent behavior and school stress, 144–145 Youth Self-Report (YSR), 183 Selye, Hans, 19–20, 272, 313 Separated, as degree of family cohesion, 182 Separation Anxiety Questionnaire, 62 Severe illness fear of death in, 25–26 loss of independence in, 9–10 stress related to diagnosis, 12–13 terror in, 11, 14 Single-participant research methods, 4 Situation-related coping, 100, 101 Situational ambiguity, and vigilance/ avoidance coping behaviors, 99 Skinner, B.F., 4 Social Desirability Scale, 108 Social Problem-Solving Inventory (SPSI), 108 Social Science Citation Index, 82 Social stress model, 145

8/15/05 7:54:16 AM

SUBJECT INDEX

Social support and anger/hostility, 129, 130, 131–132 gender differences in, 136 and health, 129–132 positive effect on job stress problems, 130 as stress buffer by organizational level in Venezuelan workers, 260–261, 263 Social Support Survey (SSS), 253, 255 descriptive statistics for Venezuelan workers, 260 and organizational level in Venezuelan workers, 261 Sociedad Interamericana de Psicologia (SIP), 31–32 Societal change, fostered by postconventional thinking, 35 Societal values, internalization and development of conscience, 34 Somatic Anxiety subscale, 47 and player performance, 52 precompetition, 48 Somatic arousal and fear of fear, 99 and increase in emotionality, 99 perception of, 99 Spatial ability test, 160, 166 Special courts judge individual stressor calibration, 319–323 JSS report for, 317 Speech scenario, in Mainz Coping Inventory (MCI), 103 Spice of life, 313 Sport Competition Anxiety Test (SCAT), 47 Staff Cohesion scale Austrian bank employees, 224 Austrian secondary-school teachers, 217, 219 State-Anger, 37, 161–165, 182, 191 among Austrian students, 170, 171 and children’s behavior problems, 188 correlation with assertiveness, 193

RT44367_C018.indd 355

355

correlations with Norwegian JSS, 305 and daughters’ behavior problems, 179, 188, 189 and daughters’ empathy with mothers’ internal distress, 191 during examinations, 165–167, 169 and sons’ behavior problems, 188, 189–190, 191 State-Anxiety, 161–165 among Austrian students, 170, 171 in athletic competition, 45–47 correlations with Norwegian JSS, 305, 306 effects of stress correlated with T-anxiety, 49 during examinations, 165, 169 failure to distinguish from Trait-Anxiety, 77–78 influence of vigilance and cognitive avoidance on, 102 in letter cancellation and spatial ability tests, 166 precompetition assessment of, 48–49 State-Curiosity, correlations with Norwegian JSS, 305 State-Depression, correlations with Norwegian JSS, 305 State orientation, vs. action, 54 State-Trait Anger Expression Inventory (STAXI), 123, 127, 157, 191 construction and validation of Portuguese, 37–40 Hebrew adaptation of, 182 internal consistency reliability of adaptation, 38 in Israeli studies of children of divorced mothers, 182 Portuguese adaptation of, 33 translation adequacy, 38 validity of Portuguese version, 38 State-Trait Anger Inventory (STAI), 156, 296 State-Trait Anger Scale (STAS), 146

8/15/05 7:54:17 AM

356

State-Trait Anxiety Inventory for Children (STAIC), 62 State-Trait Anxiety Inventory (STAI), 80, 112, 125 correlation with worry cognitions in athletic competition, 47 for Czech female handball players, 48 Hindi version, 198–199 Portuguese adaptation of, 32–33 State-Trait Personality Inventory (STPI), 316 Statistical significance, in factor structure comparison, 82 STAXI. See State-Trait Anger Expression Inventory (STAXI) Stimulus factors, in stress, 20 Strangeness experience of, 22, 26 of the familiar, 21 Stress. See also Job stress acute or chronic nature of exposure, 245 away from home, 26 conceptualizing, 19–20 due to strange environment, 26 duration of exposure to, 245 effects on players with high/low T-anxiety, 49 emotional reactions in competitive sports, 45 in fear of death and pain, 16–17, 26 frequency of exposure, 245 health-related consequences of, 272–273, 294 impact on athletic competition, 46 from lack of communication, 13–14, 24 from lack of control, 21, 23 in life-threatening illness, 3–8, 14, 19–26, 26 from linguistic and cultural differences, 24 from loss of independence, 9 medical consequences of, 254

RT44367_C018.indd 356

SUBJECT INDEX

non-interchangeability with anxiety, 78 from pain and suffering, 26 from uncertainty of recovery/ prognosis, 26 in undetermined diagnosis, 12–13 from unfamiliar rules and norms, 26 Stress assessment instruments in consultant-client relationships, 316 for family courts agency, 323–325 selecting and modifying, 323–324 Stress Index Impact scores, Austrian primary school teachers, 213–214 Stress levels consequences of excessive, 314–315 consequences of low, 315 Stress-management planning, understimulation or denial and, 318 Stressor ratings, low with understimulation or denial, 318 Stressors, 78 Structured adaptability level, 182 Substance abuse. See also Alcohol-drug disengagement; Drug use relationship of school stress to, 143–144 Supervisory Support scale Austrian bank employees, 224 Austrian secondary-school teachers, 217, 219 Suppression maladaptive nature of, 128 vs. disclosure in anger/hostility management, 132 Surgery, effects on vigilance and avoidance on hemodynamic responses during, 102 Symptom reporting, correlation with vigilance, 113 Systolic blood pressure

8/15/05 7:54:17 AM

SUBJECT INDEX

and Anger-In scores, 203 and avoidant coping style, 102 T Task-irrelevant cognitions and anxiety levels, 47 correlations with player performance, 51–55 effects of stress correlated with high/low T-anxiety, 50–51 Teacher Stress Survey, 277 Teachers and Norwegian adaptation of Job Stress Survey, 293, 298 Severity scores in Norwegian JSS, 306 Teaching Stress scores Austrian primary school teachers, 214 Austrian secondary school teachers, 218 intercorrelations with other job stress measures, 215 Technological changes, as job stressor, 273 Temperature problems in workplace, in Spanish version of JSS, 235, 239 Test anger, 155 Test anxiety and academic achievement, 78–80 and anger in cross-cultural situations, 155–173 attentional theory of, 79 cross-cultural prevalence of, 79 cross-cultural research in, 61–63 cultural and sex differences in, 67–68 gender differences in, 61, 70–71 pictorial evaluation with PETER-TQ, 61–63 situation-specific nature of, 46, 79 vs. examination stress, 78–79 Test Anxiety Inventory (TAI), 61, 63, 66, 80–81

RT44367_C018.indd 357

357

cross-cultural factor structure comparisons, 81–83 Hindi version, 199 internal consistency reliability of, 65 meta-analysis of, 77–81 Test Anxiety Scale (TAS), 80 Threat situations, and coping behaviors, 99 Threats, 78 Three-Dimensional Cube Comparison Test, 160 Total Behavior Problems scale, 185 and adaptability, 186 Trait affectivity, MCI measures of, 112–113 Trait-Anger, 37, 151, 161–165, 182 and anger expression in adolescent males, 143 and chronic anxiety, 41 during classroom lessons, 168–169 correlation with S-Anger, 37, 41 correlations with adolescent problem behaviors, 149 correlations with Norwegian JSS, 305 gender correlations with extrapunitiveness and intrapunitiveness, 40 as individual differences in disposition, 37 during reading and math tests, 167–168 Trait-Anxiety, 33, 161–165 in athletic competition, 45–47 during classroom lessons, 168–169 correlation with vigilance, 113 correlations with Norwegian JSS, 305 and effects of stress on taskirrelevant cognitions in athletes, 50–51 failure to distinguish from StateAnxiety, 77–78 gender differences in, 36 influence on frequency and intensity of State-Anxiety, 78

8/15/05 7:54:18 AM

358

and irrelevant cognitions, 48 negative correlation with maturity of moral judgment, 35 and player performance, 51–55 during reading and math tests, 167–168 relation to moral judgment, 35–36 and risk-taking behavior, 40 threat perceptions and levels of, 55 Trait coping behaviors, 105 vs. state coping behaviors, 98 Trait-Curiosity, correlations with Norwegian JSS, 305, 306 Trait-Depression, correlations with Norwegian JSS, 305, 306 Trait-State anxiety theory, 55 in athletic competition, 45–47 fundamental tenets of, 45 Transactional Process model, 272, 295 Turbulent flight scenario in Mainz Coping Inventory (MCI), 103 predictability and controllability appraisals of, 105 Type-A Behavior Pattern (TABP), 123, 272–273 among high-level managers, 254 Indian relationships of anger expression to, 203 Type-B personality, and low stressor levels, 326 U Uncertainty experience of, 99 intolerance of, 100 Uncertainty-motivated coping, 100, 113 Understimulation, JSS assessment signs of, 318 Unit vectors, 92 Ursin Health Inventory (UHI), 306 V Variable vectors, transformation to unit vectors, 92

RT44367_C018.indd 358

SUBJECT INDEX

Varimax rotation, 299 Viennese Matrices Test, 160 Vigilance, 97–98 and advice-seeking, 109 correlation with active coping, 109 correlation with negative problem orientation, 112 correlation with rational problem solving, 112 and danger control, 100 defined, 98 and intolerance of uncertainty, 100 link with planning, 109 link with restraint coping, 109 measurement of, 102–113 and neuroticism, 116 principal-components analysis, 109 in response to ego-threat (VIG-E), 103 in response to physical threat (VIG-P), 103 theory and measurement, 97–98 as uncertainty-motivated coping strategy, 100 Violence, 150 correlation with anger expression in adolescent males, 143 as problem behavior, 148 W White-collar workers gender differences in stress perception, 259, 263 perception of stressors by, 253, 259 Work content, equivalence with Job Pressure, 308 Work context, equivalence with Lack of Organizational Support, 309 Work Stress Inventory (WSI), 308 Workplace stress, 210. See also Job Stress Survey (JSS) Worry, 77 and academic achievement, 78–80 in athletic competition, 45–47 cross-cultural research in test anxiety, 81

8/15/05 7:54:19 AM

359

SUBJECT INDEX

factor axes for males and females, 83 mean factor cosines as function of culture and gender, 86, 87 in test anxiety, 64 Worry-Emotionality Scale, 157

RT44367_C018.indd 359

Y Youth Self-Report (YSR) scale, 183 maternal State-Anger and children’s behavior problems, 189

8/15/05 7:54:19 AM

RT44367_C018.indd 360

8/15/05 7:54:20 AM

RT44367_C018.indd 361

8/15/05 7:54:20 AM

RT44367_C018.indd 362

8/15/05 7:54:21 AM