1 A Rationale for Studying Persuasion John S. Seiter and Robert H. Gass
Some lime ago. one of (he author ... \\ i.I\ imitcd 10 hi" "on· ... ,",ccond grade c1a ...... rool1l 10 tal"-
about
'·pcr~uasion.'·
the :-.ubjeci the author
tC~lchc ...
and rC\l!i.Jn.:hel" at a univcr..,ity. The
t;CC-
and graders. the: author wa ... told. were hu\ illg ··care!.!f year:- and parent' with all 'orts of professions were making appearallce:... The author said he would he happy to do it. though privately he had his douhts. Pcrhap" hc'u \cell Ino man) mo\'ic\ \\ ilh "career day" scenes-featuring o\,crzc<.Ilous parent ... speaking to cla ... sroorns of \quirmy. fidgety children. Wh;Hcvcr the ca..,c. he wa ... ,ure about nne thing: He was going to hu\ 'c to adapt to an audience prone to "ants in the pant .... " On presentation day. aftc;" a quid. definition of pep'Iuasioll (involving a chocolate bribe for all those who li . . tcned do\cly), the author and <;;;econd grader, tried a few simple experiments. In one. they watched eommerciab that made certain products (c.g., dolls, car ..... cerea l) appear much more dalliing and irre,i . . tib lc than they were when ...een or ta . . ted in the classroom. Onc ad featurl!u a child \I.' ho. IIllIllCdi .. lIcly after eating a particular brand of cereal. became popular. hmfy. and ... k.il1ed at ... katehoarding. But when \"cvcnll of the sl!cond graders "'
1
2
Chapter 1 • A R01;OIwh' for Studying Per.w(lsioll
villain did hi!o. !-Ion think he was? Of course. if hi5, son or classmates had been thinking such thoughts, they wouldn't be the fI,,1. Indeed, from time to time, those who study and teach social inlluencc have been criticized harshly. In the fifth century B.C., for example. Plato derided the first teachers of persuasion for "making the worse appear the better reason" (Corbell, 1971, p. 598). Later. in the mid-1970s, Simons ( 1976) noted: From a number of quarters Ihc~c day .... persuasion is under attack for being a manipulative acti\ity. Its highest critic ... equate notjusl some persuasion. but all persuasion with decep~ (ion and rote-playing, domination and exploitation. (p. 35)
More recently still. femini:-.{ scholars have characterized traditional rhetoric. that is. persuasion. as a type of communication that devalues the lives and perspectives of others. and a means by which persuaders attempt to feci good about themselves by denigrating others. By way of example, Foss and Griflin (1995) wrote: The value of the ~elr for rhelor~ in this rherorica l system comes from the rhelor's ability to dClllon.'ttrate .'tuperior knowledge. ski lL.... and qualifications- in other words. authority-in order to dominate the perspectives and knowledge of those in their audiences .... The act of changing others not only establishes the power of the rhelor over other!:. but also devaluc" the lives and pep"pcctivcs of other". (p. 3)
With such criticisms in mind (and reputations at stake!). we thought it essential that the iIHroduclory chapter 10 Ihis text set forth a rationale for studyi ng persuasion as well as provide a brief ethical backdrop for exami nin g such a study. Before we do so, however, we think it important to address two related issues. First. considering that critics of persuasion seem to emcrge and recmerge with some regularity. you might be wondering whether the study of persuasion has suffered as a result. At first glance. one might be tempted to conclude thai this is the case. When examining this issue. for example, Miller and Burgoon (1978) initially noted: While il would be hyperbolic to state Ihal Ihe gum. arc silent on the persuasive battleground. their roar ha~ grown ... poradic and muted. No longer are the pages of journals glutAs a result of these disciplinary trends. led with the results or per~lIasion studies. bedrock pessimists proclaim thaI per~lIasion research is a dying enterprise. while skeptics content themselves with the observation that it has become an area of limited, secondary import. (p. 29)
Though observations such as thes.e paint a grim picture. such skeptics and pessimists are mistaken. To be certain. upon funher inspection. Miller and Burgoon (1978) concluded that the study of persuasion has not wavered; it has simply changed focus and direction. Thus. while traditional studies examining linear. "one-to-many " persuasive attempts are less in evidence now, you'lI find a host of srudies on new persuasion topics. including those on compliance gaining and deception. In other words, skeptics concluding that persuasion re~cnrch has diminished were looking in the wrong places. Persuasion is a dynamic area of srudy. On the one hand. scholar ly interest in topics may change as a result of socia l. per\onal. or political climates- hence the large amount of research on mass per-
Chapter I • A Rationale for Studying Persuasion
3
suasion and propaganda during the I940s, especially in World War II, followed by examinations of social protest and resistance to persuasion during the 196Os. when there was a
greal deal of political unreSI and distrusl of governmenl On the other hand. scholarly interest in specific topics may not fade forever. Instead, interest in various aspects of persua-
sion may be cyclical. This ebb and flow of interesl is evidenced by research on cognitive dissonance, which flourished in Ihe 1950s and I 960s, entered a period of Ihe doldrums in Ihe I 970s and 1980s.' and then reemerged as a vilaltheory with importanl implicalions for persuasion in the 1990s. At one point. Leon Festinger, who created Cognitive Dissonance
Theory, Slopped conducling research on the Iheory altogether. In a Iranscriplion of his remarks al a 1987 symposium (cited in Harman-Jones & Mills, 1999), he said, "I ended up leaving social psychology .... I left and Slopped doing research on the theory of cognitive dissonance because I was in a total rut. The only thing I could think about was how correct
Ihe original slalement had been" (p. 383). In a recent volume by Harmon-Jones and Mills (Eds.) devoled to cognitive dissonance, however, Aronson (1999) underscored the imporlance of Ihe theory and "ils reemergence in the 1990s as a powerful means of predicting and changing human behavior in a variety of areas, including those thaI have abiding social importance" (p. 103). In short, then, despite what skeplics and pessimisls might say. persuasion research has never gone out of fashion; it has just changed iLs style and has on
occasion broughl back inlo fashion styles of old. In addition to noting the unwavering nature of interest in persuasion research. a second point we would like to make before offering our rationale for studying persuasion is that we do not disagree with all that critics of persuasion have to say. That is. we agree that plenty of people have used persuasion for the wrong reasons, sometimes with tragic consequences. Focusing on such instances alone. however, strikes us as seeing the glass half
emply. Think, for example, as the author and his son did for the rest of Iheir fishing Irip, of all Ihe good Ihings that persuasion mighl accomplish. Withoul persuasion, how does a physician urge a diabetic patient to layoff sweets or get more exercise? How does a friend
gel her drunken buddy 10 accept a ride rather than drive himself home? How does a mother warn her five-year-old child never 10 take rides from strangers? How do civil righls aClivislS speak OUI against racism or world leaders lobby for peace agreements? We hope you see our pain!. The list of good things that can be accomplished through persuasion is endless. The arguments of some critics, however, focus less on the ends of persuasion and instead point accusing fingers at the means by which persuasion is accomplished. For example. the feminist scholars we mentioned earlier take issue with traditional persuasion because it embodies an adversarial view of communication encounters in which one person is trying to do something to another. In contrast, their approach, an "invitational approach" to rhetoric, emphasizes cooperation and dialogue. One person is trying to communicate with another.
We respecI this poinl of view. Indeed, we would be among the first to acknowledge that in our patriarchal society, people often fail to recognize incentives for cooperative communication. They presume that communication encounters are competitive or adversarial in nature. They overlook their interdependence and view communication as a
win-lose process. They neglecI shared or communal approaches 10 problem solving and decision making.
4
Chapter I • A Rationale for STlldyillg PnslI(lsioll While cooperative. dialogic encounters may be the ideal to which we should all aspire. we believe that there are many situations in which people have to roll up their sleeves and persuade. 2 Imagine. for example. that you observed an injustice being committed by one person against another. You could begin by inviting the aggressor to engage in a dia-
logue in the hope of arriving at a mutually satisfactory outcome. But what if the aggressor spurned your invitation'? Would you simply say, "Oh well, I tried" and resign you rself to the fact that the world is full of injustices? We think it would be better to resort to persua:-.ion. to engage in an active effort 10 change the mind of the aggressor. We see dialogue and persua~ion as complementary, not alllilhclical, forms of communication . There are lime~. we submit. when one has {J moral obligation to try 10 change others' minds ami behm'ior. And as wc·ve argued elsewhere (Gass & Seiter, 2(03), we think the motives of persuader!-. have as much to do with the ethical defensibility of their innuence attempts as the means of persuasion used. The women's suffrage movement was a persuasive campaign. So was the women's rights movement that began in the 1970s. So are the efforts to ...ecure basic human rights for women throughout the world. We don't think that women who lOil in sweatshops. or who live in poverty. or who are denied the most basic education ... or medical care. or who are forced to marry against their will, mind too much if others lise persuasion to try to improve their lot in life. Persuasion is essential precisely
becau,. dialogue sometimes fails. From an ethical standpoint. then. we ... ide with Plato's student. Aristotle. who had this to say about persuasion: If it j!,! urged thai an abuse of the rhetorical faculty can work great mi.!.chief. the same charge can be brought against all gOQ{lthings (save virtue itself). and especially <'lgainstthe most useful things such as strength. health. wealth. and military skil l. Rightly employed. they work the greatest blessings. and wrongly employed. they work th e greatest harm. ( t355b)
In other words. we take the position that persuasion is amoral. or as McCroskey (1972. p. 269) called it. "ethically neutraL" Persuasion. much lik e any tool. can be used for good or bad. It all depend!<" on the motives of the tool u~er- Hitl er versus Gandhi. Osama bin Laden \ersus Martin Luther King. Jr.. Jim Jone~ versus Mother Teresa. In short. the gla~s
i, neither half empty nor half full-it's both. Although we believe that this point addresses the previously mentioned critici ... ms of persua ... ion. there are additiona l and perhaps even more compelling reasons for reading a book such as this. First. critics who argue against the ~lUdy of social influe nce by pointing their finger~ at unethical pen.uaders provide what we consider one of the best reasons for studying per~uasion: self-defense. The fact that unethical persuaders are lurking around corners. plying their trade on unsuspecting wrgets. is no reason to stop studyi ng persllu... ion. On the contrary. we believe that ~lIch a response on ly makes them more s uccessful. Instead. by learning the tricks of their trade. we are in a better position to defend ourselves against unethical persuaders. This, of course. was what the author was lrying to accom plish with his second-grade audience. By expos in g some of the bells and whistles advertisers use to appeal to children, he hoped to make the chi ldren more crit ical consumers of the persuas.ive messages they receive. What better way to di~arm those who might try 10 ma-
Chapter I • A Ratiollall' for Studyill/-: Persllwioll
5
nipulate or coe rce us into do in g ~o m c thin g thal '~ not in our best inte rest? Learn ing about pe rsuasion is an exccllent dc fense ! We reali ze. of course. th at by stud yin g persuas ion people may beco me more e ffec · ti ve per:-, uaders the mselves. Th is. we be lieve. is anothe r va lua ble reaso n for studyin g so· cial intlucnce. Whil e so me critics mi ght shudder at th is prospcc t, we view the abilit y to influence oth ers as a fund ame nt al ingredi ent of communi cati on competence. And we are not a lone. For example. in an ex te nsive rev iew o f communi cati on competence researc h fro lllilluitiple fi e ld,. Spitzbe rg and C upach (1 984) reported th at an indi vidual' s abilit y to adapt e ffec ti ve ly in order to ac hi eve goa ls is perh aps the most universall y accepted aspect of communi catio n competence. Th is abil ity, th ey noted . j, a c ruc ial part o f be ing we ll adj usted that , when ab ~e nt , " is oft en associated with abn orm al or eve n patho logical ori enta· ti ons" (Spitzbe rg & Cupach. 1984. p. 36). Co mpete nt com mu nicators are persuas ive. They know how to adapt successfull y in orde r to ac hi eve thei r goa ls. With the above in mind . it makes no se nse to us to suggest th at the stud y of such an impo rtant co mmuni ca ti on abi lity be avo ided . Imag inc. for exa mpl e, a co mmun ica ti on t e~lc h e r te lling a stude nt. "I wan t you to be ab le 10 organ ize your ideas: usc solid reaso ning and ev ide nce: have good cOlllmand of lan guage : and lise move ment, ge~ tures, and eye co ntact efTec ti ve ly- hw 11 01 if you are going 10 persuade anyolle of onylhillg!" Imagi ne a teac her te lling a c lass, " Yo u need to be able to e ngage in ac ti ve lis tening, to paraphrase othe rs' ideas. to e ngage in pe rce ption·chec king. to self-di sc lose in the here and now- bUl 11 01 if yO/l •re /I~rill g 10 c(}IH'iIlCe someoue of .wmelhillg.'·· Bei ng an e ffec ti ve persuader is pari a nd parce l of be ing a co mpete nt communi cator. Ha vin g made Ollr bias c lear, we feel it 's important to pa u ~e for a mome nt 10 point o ut. as others have (see Kellerman. 1992: Spitzberg & Cupac h. 1984). thaI cO mpelel1l co mmuni cators are not onl y e ffect ive in ac hi ev in g the ir goa ls but a lso do so in appropriate ways. Here. we emphasiL.e the word approp riate. Whil e we have no qu a lms abo ut teac hin g stud ent s to become more effec tive persuaders. we be li eve th at th e power 10 persuade ca rri es with it an o bli gati o n 10 persuade ethi call y. Happil y. the re see ms to be in c rea~ in g and co ntinucd interes t in ethi cal issues surroundin g persuas ion. By way of exa mple. Robert C ialdini ( 1999) has co mpared ethi cal pe rsuade r"> 10 "s leuths:' Unlike ·'bu ng lers." who use ineffecti ve strategies. or "smu gg le rs:' wh o re ly o n un scrupulou s tac ti cs. "s leuths" stud y a persuas ive silUati on in order to de· ve lo p the most e ffec ti ve tactics. ye t they also make sure th at the tacti cs they use are elhi · ca l. S imilarl y, Broc kri ede (1974) co mpared e thical pe rsuaders (or arg uers) to " lovers." Unlike ·;scducers." who use tri cke ry and deceit to achi eve the ir goals, or " ra pi ~ l s, " who u ~e th rea t ~ or force, " Iove rs" res pect others' d ig nit y, are ope n to oth ers' arg ume nts. and stri ve fo r eq ualit y. Finall y. as we've already noted. pers uas ion itself is amora l. Thi s sugges ts to U ~ thal decisions regardin g how ri ght or wrong any give n influe nce aHe mpt is depe nd largely on the natu re o f the situati on fi nd on th e motivc:-> of Ihe pe rsuader.' Eve n so. we believe thai a few general guide lines appl y to a lmost eve ry influence att empt. We can· not overemphasizc, for exampl e. the importan ce of res pecting others a nd looking out for the ir welfare. especiall y when th ey are morc vulne rable th an ot hers (e.g .. children). We believe th at persuasion is ge nerall y more e thi cal whe n people are made aware that they are be in g influe nced and whe n th ey ha ve unconditiona l freedo m to say " no" to influence att empts.
6
Chaple r I • A Rmiollalefor STudying Persllasion
In addition to its defen~ivc and instrumental benefits, there is another reason for studyi ng persuasion. Before offe ring our final reason, thou gh. we wish to point out one mo rc thin g about persuasion's detractors. Specifically, those who argue against the stud y of persuasion are themselves com mittin g a Til quoque fallacy, that is. accusing another of a
simi lar wrong. In the process of criticizing persuasion . such critics are relying on persuasion themselves. They would like to persuade you 1/01 to study pe rsuasion, not to use persuasion. or to use persuasion less. This approac h raises an interesting dilemma: How can one commun icate one's beliefs. op ini o ns. va lues. views. positions, preferences or 'dru ther<.; withoU/ employing persua!\ion?4 We have nOled elsew here (Gass & Seiter, 2003) th at influence attempts, whether implicit or explic it. are ubiquitous. Look around yo u. Eac h day you are bombarded w ith messages, adverti sements. logos, and countless ot her influe nce attempts. Consider, for example. ~ome of the following figures: • More Ihan $200 billion per year is spenl o n advert ising in Ihe Uniled Siaies (Berger, 2000), • If all the money spent o n advertising were divided up. it woul d work oul to about $800 per person inlhe Uniled Siaies per year (Berge r, 2000). • The average person in the United States is ex po!\ed to more than 3.000 advertising message, per day (Bo rchers, 2002: Dupo n!. 1999: S im o ns, 2001: Woodward & Denlon. 1999). Besides th al. think of the ot her co nt exts in whi ch you are exposed to inilue nce attempts. Per!\uasion is an obv iolls and indispensable co mpone nt in a number of professions, including cou nseling. law, management. po litics. sales. social work . teaching. and the ministry. It can be found ab undantl y in the sc ie nces. th e arts, interpe rsonal inte ractio ns. and secondgrade classrooms, just to name a few. Moreover. depending o n the way one defines persuasion (see chapter 2), it ITiay be virtua ll y im poss ible flOf to influence others. Appeara nces, for example . eve n if unintentionally communicated. ca n be influential. Thus. babyfaced people lend 10 be perceived as honesl (Brownlow, 1992), and bald candidales are less likely 10 gel elecled Ih an Ihose wi lh a full head of hai r (S ige lman, Dawson, NiIZ, & Whicker, 1990). Peop le who use a 101 of eye conlaCI lend 10 be more persuasive (Segri n, 1993), and attract ive people are mo rc believable than unattractive ones (Seiter & Dunn , 2000). Heav ier people are less likely 10 earn hi gh salaries Ihan slim people (Argy le, 1988), and tall people are more likely Ihan sho rt peop le 10 gel jobs (A rgyle, 1988), In short, avoiding persuasion wou ld be diffi cult. if not impossible. Even critics of persuasion cannot avoid it. Clearly. e ngagi ng in persuasion is an inex tricable part of being human . This statement. to us. provides the most co mpe llin g reason for studying pers uasion . Specifically. if we humans ever hope 10 understa nd ourselves, how can we ig nore one of the major underlying impulses for human com muni cation ? Studying persuasion can and doe~ te ll us a great deal about how hum ans produce. shape. perceive, interpret, and respond to l11e~sages. It provide!\ insigh ts into the ~ocja l and c ultural forces that give rise to influence att em pts. among them pres ide ntial debates, social protests, religiou s cults, and health campaigns. It di spe l ~ vari o us "common sense" assumptions about th e ways III wh ich persuasion " really" works, In s hort, as Miller and Burgoon ( 1978) argued:
--,
Chapter I • A Rationale for Studying PerSftasiOIl
7
No matt er hO\\ ferve ntl y !',ollle well-meaning indi vidual s try to deny or wish the fact away, concerning contro l a nd influe nce will remain an integ ral aspect of humanity 's daily cOlllmunicati ve ac ti vities. Gi ven the centrality of these questi o ns. it seems both foolhardy and frui tl c!-.s to a~~UIllC that Ollr unde rstandin g of human com munication can go on advancing wi thou t conti nu ed research attention to the persuasion process. (p. 45)
que!-. ti o n ~
It is in this spi rit th at we present and hope you will approach this collection of readIngs. It contu ins what we bel ieve are so me of the finest ideas from so me of the most important pe rsuasion sc holars of our time. The collection is organized into four parts.
I. Part I lays the grou nd work for study by examining conceptuali zations of persuasion, the hi story and nature of the tield. and the ways in which persuasion functions. 2. Part II exa mines important variabl es in the process of persuasion. Specifically, it discusses co mmuni ca tor characteristics-source credibility, argumentativeness, verba l agg ressive ness. and ge nder-and how they affect the creation, se nding, and receiving of persuasive messages. This section also examines elements of persuasive messages the mselves. both verbal and non verbal. 3. Part III explores the strat eg ies and processes by which people seek and resist compliance. It includes a theoretical disc ussion of how compliance-gaining messages are produced. as well as an explorati on of sequ enti al persuasion tactics, fear appeals. decepti on. and inoculation. 4. Part IV exa mines soc ial intluence in seve ral contexts. Specifically. it analyzes persuasio n
Noles ________________________________________________________ I. In 1978. Milkr and Burgoon argued that attribution and equity theories had ended the reign of cognili ve dissonance and OIhcr cogni ti ve cons i ~tency theories. 2. In fairnt" ....... \\c ... llOuld note that some femini~1 scholars admit that pe rsuasion is sometimes neces· .. ary (Fo .... & Griflin. 1995). J. Clearly. the number of per.-.uasive si tuations that cou ld be debated is endless and could fill vol~ uJ1le~. Though we do not havc rool11 to consider suc h "i tuati ons here. for a more detailed di sc ussion. see Ga .. ~ & Se iter (200J). chapter 16. 4. Again. to be fair. we should note th aU~ome feminist critics (see Foss & Griffin. 1995) say that their appro;lch. knov.n a ... ill\i tational rhetoric. iii not designed 10 "change" anything (indeed. that would be per· <"ua<"ion) . In<.,lead. thcy ~ay: Although invil;lIional rhetoric i~ nol designed to create a specific change. such as the transformation of ..ystem .. of oprre~~ion into ()ne~ that va lue and nurture individuals. it may produce such an outcome. (p. 16)
Moreo\er. \\ hile ,>uch cri ti cs cla im they are not attempting to characteri ze trad itional approac hes to rhetoric (i .e .. PCNI
8
Chapter 1 • A Rarimw/t' for Smt/\'illg Per,HW.\';OIl characterize pcr'>uasion. for example. thai 11 "devalue'> live ... and per... peclive..... (p. J I. "infringc~ nn OIhcr... · rights 10 choose" (p. 3). comititute\ "a ~lI1d of lre<.,pa\sing on the per-iollal integrity of OIhcr..," (p. 3), and Icad ... to the "denigralion" of other,' per..pecti ... C5 Cp. 6~is persua"i ... e. \\ hcthcr it \\;1\ intt!ndcd to be or nol.
Referellces_____________________________ Arg) Ie , M. (1988). Bodily cOIII/mmicmion (2nd cd.). Madi,on. CT: International University Pre' .... Aristotle. (1932). The rltetoric (L. Cooper. tran .... ). Englewood Cliff..,. NJ: Prcnlicc-Hall. Aronson. E. (1999). Dis!'>onance. hypoc ri,). and Iht: ..elf-concept. In E. lIarmon-Jonc!-. & J. MilJ<.. (Ed ... ), Cognitive di.\"SOIlWICl': Proj!.l"f'!>s Oil a pi I'O/(l{ theon' ill .wcia/ psydwlogy (pp. 103- 126). W a~hi ng ~ ton, DC: American P,ycholog} . Berger, A. A. (2000). Ads.fads, (/1/(1 COIISIII1U'r CI/It/lre. Lanham, MD: Rowman and Littlelield . Borchers, T. A. (2002). Per.wa ..itm 111 th(, media (lgl'. Bm,ton: McGraw-HilI. Brockriede. W. (1974). Arguer-. a... lo . . er'. Philosophr lind Rhetoric. 5. I - II. Brownlow, S. ( 1992). Seeing i~ belie\ mg: Facial appearance. credibilit) and mtitude change. Jmmllli of Nma'ubill Bellm·ior. 6. 253-259 Cialdim, R. B. (1999). Of trick, and tumor, : Some littlc-recogni/cd co'as of dishonest uo;e of cffcctl\c "ocial innuencc. PsycholoKY & Markelillg, 16(2),91-98. Corbell, E. P. J. (1971 J. Cla.fJi{'{/1 rhewncfortlle modem Sl/Idf'lIt (2nd ed.), New York : O"ford Univl!r:-.ity
Press. Dupont. L. (1999). Imll[.:l's lhal wl/: 500 way.\· to crt'ate gre(1/ ads. Ste~Foy. Quebec, Canada: White Rock Publi"hing . Fc\tinger, L. (1999). Reneclion\ on cognitlvc di,,"onance: 30 year ... later. In E. Harmon-Jones & J. Mill . . (Edf\.), Cngnitil't' di.w}IJ(mce: Progre ,\'s Oil {/ pil'OllIl theon' in social psychology (pp. 381-385). Wao;hington, DC: American P,ychology . Foss, S. K.. & Griffin, C. L. (1995). Bcyond per!>u;l'.. ion: A proposal for an invitational rhetoric, Commllni· {'olioll M01/0Krap},,\', 62.2-1 R. Gas~,
R, H.. & Seller. J. S. (2003). PerwasI()// ..wl'ial,1!I7I1ellce, alit/ compliance Kaining (2nd ed.). Boo;ton: Allyn & Bacon. Hannon·Jones, E.. & Millo;. J . (1999). COf{lIiln'e cli\'wJf/(lI/a: Prowess on (I pi\'owl theory ill .wcilll psy· chology. Washington. DC : Amcrican P"ychology. Kellerman, K. (1992). COlllmunlcation: Inhercmly ... trat~gic and primarily aUlOmatic. ComllllmicOIio1/ MOl/ograph:;. 59. 288-300. McCro ... key. J. C. (1972). AI/ intmdllC'fioll to rhetoriCllI cOImmmicCllio1l (2nd ed.). Englewood Cliffs, NJ: Prentice-Ilall. Miller. G. R .. & Burgoon. M. (1978). Per... ua\iOll rel'>earch: Re\iew and commentary. In B, R. Ruben. (Ed.). Commlmicari(m yellrl)(J(I~"2 (pr 29-47). New Brunswick. NJ: International Communication A"sociation. Scgrin, C. (1993). The effect!'.. of non\ocrbal bcha\ ior on outcomes of compliance gaining attempt ... , COIIImunication StIIt/ies, -14, 169- 187. Seiter. J. S .. & Dunn. D. (2000). Beaut)' and bclic\'ablhty in f\exual harassment cases: Doe~ phy!'..ical at~ tractiven!.!ss affect perception" of veracity and the likel ihood of being harassed? Cmm/ll/lliwtiol1 Research Reports. 17 (2). 203-209. Sigc1man, L., Dawson, E., Nitz, M .. & Whicker, M. L. ( 1990). Hair loss and electability: The bald truth . JOlln/lil o/Nofll'erbaIBehm'ior. 14.269-452. Simons. 1-1 . W. (1976). Persua.\·ion: Um/el"sullldillf{. practice. lind lInlilysis. Reading. MA: Addison Wesley. Simons, H. W. (200 I ). Persulision inmciet.\'. Thousand Oak.... CA: Sage. Spitlbcrg. B. H., & Cupach. W . R. (1984). Illterpersollal commullication compete"Cl', Beverly Hil l\, CA: Sage. Woodward. G. & Denton. R. E. (1999). Per.w(lsiml lIIId influence ill American life (3rd ed.). Prospec t Heights. IL: Wadsworth .
c..
Part
I Preliminaries Definitions, Trends, and Theoretical Underpinnings in the Field of Persuasion
Like most "part I s:' this first ~ect ioll sets the stage for all that follows. It does so by adsome preliminary questions involving the s tudy of persuasion----queslions of the "what?," "when?," "where ?," "why." and "how?" variety. We hesitate. however, when ca ll ing these queMions "preliminary." Indeed. to us the term preliminary implies the techni cal and detailed material that one mu st labor through before getting to the "good stuff:' in mu ch the sa me way that one needs to finish one's spinach or broccoli before e njoying dessert. Contrary to s uch connotations, we find the material covered in this sec tion appealing and fascinating . Thus (at the risk of pushing metaphors too far), rather than think of this sec tion' ~ chapters as "warm-up aCls" or "appetizers" we consider them "main events" or "main cour~c~" in and of themselves. And (a t the risk of mixing metaphors), we view thi~ section as the first reel of u movie that grabs one's attention and lays Ollt the basic characters and foundations for toe plot that follows, Although on the one hand the chapters in Ihi~ sec tion provide a background for understanding mnterial later in the book. they address issues that not only cOl1linue to intrigue (and vex) per~uasion scholars but in many ways. dictate the ways in which rc~earch and theori zing about persuasion are done. As we've already not cd, eac h chapter in this section addresses basic questions about persuasion, social int1uence, and compliance gain ing. Chapter 2 examines the "what" of persuasion . As the saying goes, "Before beginning a hunt, it is wise to ask someo ne what you are looking for before you begin looking for it" (Milne, 1995, p, 55), This chapter, then. addresses questions about " the nature of the beast." In short, it offers a definition of (h'es~illg
9
10
Pan I • Preliminaries: Definitions, TremJ.~, and Theoretical Ufl(lerpi""ill~r\
persuasion and related terms such as influence and compliance gai ning. As you wi ll see. the study of persuasion has undergone a dramatic transformation in the last few decades, This chapter argues that such changes invite a reexamination of how pcrsuasion is conceptualized, It focuses specifically on two criteria- whether persuasion i~ intentional, and whether it is successful- that scholars have used to define pef'ua,ion. It also discu,ses the implications of using or not using such criteria to limit what is studied in the field. After reading the chapter. we think you will see that understanding "what" pcr'\uasion is involves a lot more than consulting a dictionary. We also hope the chapter will help you better understand the terms "persuasion." "social influence." and "compliance gaining" when you encounter them later in the text. If chapter 2 is the "what" chapter, chapter 3 is the "when and wherc" chapler. In it. Daniel O'Keefe provides a road map for research and theory in persuasion. explaining when and where the field got its start, where it has been, and where it may be heading. Besides showing us "the lie of the land," we think this chapter is important because it invites us to explore the ways in which the field has expanded and how we have as a result developed new understandings about persuasion. At the same time, this chapter place~ the complexity of variables related to persuasion into perspective and underlines the importance of understanding how specific context~ influence the proces~ of persuasionconcepts that, as you will see, become important later in this text. Although the "trends" and "prospects" laid out in thi s chapter represent just one author's perspective 011 the field. we like this portrayal and find it a compelling and inforJ11alivejourney. Finally. chapters 4 and 5 focus on theories of persua~ion or. more specillcally. the "how" and "why" questions of social influence. What is a theory? According to Littlejohn (1996), in its broadest sense, a theory is simply an explanation of a phenomenon. Thl!ories are comprised of a set of related concepts or propositions that help us underMand how something functions or why it works the way it does (Infante. RanceI'. & Womack. 1997). Theories not only explain how or why persuasion works. they al~o offer the prospect of prediction and control. By way of illustration. Kim Witte's theory and model of fear appeals, described in chapter 13, not only explains how fear appeals function but a lso predicts the specific c ircumstances in which they will or will not be effective and offers insights into controlling their use for maximum persuasive effect. In addition, good theories are heuristic-they generate research, aid in discovery, and are "vital to the growth of knowledge" (Littlejohn. 1996. p. 32). The theories presented in chapters 3 and 4 meet all of these criteria. Chapter 3 provides a brief overview of some of the moM influential theories in the discipline. wherea~ Richard E. Petty, Derek Rucker, George Bizer, and John T. Cacioppo. devote chapter -l entirely to one theory, the Elaboration Likelihood Model (ELM). which i, arguably one of the most important contemporary theories of persuasion. The ~copc of its influence i\ reflected in the number of places where it is cited in the other chapters of thi, book. The ELM helps us understand the ways in which multiple variables (such as communicator characteristics, credibility, message factors, (he nalUre of contexi. and so forth) come into play in the process of persuasion. As you'll see, persuasion is rarely a simple. linear process, like a cue ball knocking an eight ball directly into a corner pocket. Instead, persuasion involves interactions among a number of variables. like pool balls banking off the cushions of a pool table or ricocheting off one another before finally dropping into a
Part I • Preliminoril'.v: Definitions. Trnlll\. lIlId Theon'tical UlldellJillllillg.\
11
poc ket. This chapter not only demonstrates such co mpl exi ty. it lays out the key proposition s of rhe ELM. and also addresses recelll criticis ms of the theory. We hope you find the chapters in this part informative and enlighten in g. After reading them. you should have a much better understanding of what pe rsuasion is. where it slands. and how it functions. Moreover. although we said carlier that we would like you to co nsider thi s section more of a main course than an appetizer, we hope that this first batch of chapters will whet yo ur appetite for all th ose that follow.
Referellces_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __ Infan te. D. A .. Rancer. A. S .. & Wom;'lck. D. F. (1997). BlIildillg ("{)/J/II//mic(//ioJ/ theory Ord cd.). Prospect Height s. IL: Waveland Press. Inc. littcjohn. S. W. (1996). Theorie.\ of human communication (5th cd.). BOl,ton: WnJs-.worth Publishing Compnny.
Milne. A. A .. ( 1995). Pooh's lillIe illstrtlclitm book. New York: DUHon Book<..
2 Embracing Divergence A Definitional Analysis of Pure and Borderline Cases of Persuasion Robert H. Gass and John S. Seiter
Scholars in various disciplines devote considerable attention to defining their terms. Consider the field of communication, for example. Dance and Larson ( 1976) reported well over 100 different definitions of the term "communication" in the literature. Much of the wrangling over what constitutes communication has centered on issues such as intentionality (Is communication necessarily intentional?), symbolicity (Is communication limited to symbo li c action?), and the number of participants required (Does communication require two or more persons?). Similar concerns have surrounded definitions and conceptualizations of the term "persuasion," though admittedly on a lesser scale. More than a dozen and a half definitions of persuasion have appeared in writing over the last two decades (see table 2.1). Yet ahhough articles addressing the merits of competing definitions of communication have appeared with regularity in communication journals (e.g., Andersen 1991; Bavelas, 1990:
Beach. 1990: C levenger. 1991: Cronkhite. 1986: Dance, 1970: Gerbner. 1966; Milier. 1980; Motley. 1990a. 1990b. 1991: Shepherd. 1992). rew or no recent articles have addressed issues related to definin£ and conceptualizing persuasion or its closely related terms, influence and compliance gaining (Di ll ard, 1988). This is hardly because the issue has been settled. To the contrary. many current definitions of persuasion are incompatible. It appears that as persuasion research has evolved over the last two decades. notions about w hat constitutes the study or persuasion have become less and less distinct. Still in evidence are "trad itional" studies or persuasion, typified by public or one-tomany investigations a imed at belief. attitude, or behavior change. Such traditional studies are apparent in heahh awareness campaigns designed to reduce smoking, increase seat belt usage, promote safe sex, or prevent drug use (see for example, prau, Kenski , Nitz. & Sorenson. 1990: Pfau, Van Bockem, & Kang. 1992: prau & Van Bockem. 1994; Witte.
13
14
Part I • Preliminaries: Definitions, Trends, and Theoretical TABLE 2.t
Underpiwljng!1'
Assorted Definitions of Persuasion over the Past Three Decades
"A conscious uucmpt by one individual or group to change the attitudes. beliefs, or behavior of another individual or group of individuals through the transmission of some message."
Bettinghaus, E. P., & Cody, M. J. (1994). Persuasive communication (6th ed.). Fort Worth, TX: Harcourt Brace, p. 6. "Persuasion is the coproduction of meaning thaI results when an individual or group of individuals uses language strategies andlor visual images to make audiences identify with that individual or group."
Borchers, T. A. (2002). Persuasion in the media age. Boston: McGraw-HilI. p. 15. "Persuasion is the name we give to the type of communication that brings about change in people."
BOSlrum, R. N. (1983). Persuasion. Englewood Cliffs, NJ: Prentice-Hall, p. 8. "We define persuasion as a conscious symbolic act intended to form. modify. or strengthen the beliefs. opinions, values, aUitudes, and/or behaviors of another or ourselves."
Burgoon. M., Hunsaker, F. G., & Dawson, E. 1. (I 994). Human Communication (3rd ed.). Thousand Oaks, CA: Sage, p. 177. "Persuasion is intended communication that affects how others think. feel. and/or act toward some object. person. group or idea." Cegala, D. 1. (1987). Persuasive commLmication: Theory and practice (3rd cd.). Edina. MN: Burgess International, p. 13. "Persuasion is the process by which language and symbolic actions infiuence Choice-making by others." Cooper. M., and Nothsline, W. L. (1992). Power persuasion: Movillg an allciellf art illto the media age. Greenwood, IN: Educational Video Group, p. 2. "Persuasion involves one or more persons who are engaged in the activity of creating, reinforcing. modifying. or extinguishing beliefs, aUitudes, intentions, motivations, and/or behaviors within the constraints of a given communicat.ion context."
Gass. R. H., & Seiter 1. S. (2003). Persuasion, social influence, and compliance gahling (2nd ed.). BaSIon: Allyn & Bacon/Longman, p. 34. "Persuasion takes place when a motivator is able to either change or con finn an existi ng attitude in the minds of listeners."
Hazel, H. (1998). The power of persuasion (2nd ed.). Dubuque, IA: Kendall-Hulll, p. 2. "Persuasion is a transactional process among two or more persons whereby the management of symbolic meaning reconstructs reality, resulting in a voluntary change in beliefs. attitudes. and/or behaviors."
Johnston, D. D. (1994). The art and science a/persuasion. Madison, WI: William C. Brown. p. 7. "Persuasion is the co-creation of a state of identification or alignment between a source and a receiver that results from the use of symbols."
Larson, C. U. (2001). Persuasion: Reception and responsibility (9th ed.). Belmont, CA: Wadswonh, p. 9.
15
Chapter 2 • £IIIhracillg Dil'ergellce
TABLE 2.1
COlllillued
"Persuasion is a complex. continuing. interactivc process in which It ... ender and receiver arc linked by symbols, verbal and nonverbal. th rough which the persuader attempt" to innuencc the pcrsuadee to adopt a change in a given all itude or behavior becaU'>e the pcr\uadce ha~ had hi, perceptions enlarged or changed."
O'Donnell. Y., & Kable . J. (1982). Per.\'[/{/siOf1: All il11eractil·e·depelldellcy approach. New York: Random Hou se. p. 9. "A successful intentional effort :u innuencing ,mother'" mental ,tate through communication in a circumstance in 'Which the pcr,uadee ha.'. some mC:I'iure of freedom ."
O·Keefe. D. J. (2002). Persuasion: Theory and re.\earcl! (2nd cd.). Newbury Park, NJ: Sage. p. 17. "Persuasion is an activity or procc:-'i in which a communicator ancmph to induce a change in the belief. alti tude. or beha\ ior of another pe r~on or group of per,on ... through the tran,mi"ion of a message in'l context in which thc pcr..uadee h;]" ,>ollle degree of free choice:'
Pe rlofr. R. M. ( 1993). The dynamics of persuasioll. Hillsdale. NJ: Erlballm. p. 15. " We use the teml persuasion to refer to any in ... tance in which an active attempt is made to change a person'~ mind." Pe n y. R. E .. & c'lcioppo. J. T. ( 198 I ). Allitudefl lIlIll persuasioll: C/assic alld ('olliell/pora,.y per· spec/ilres. Dubuqu e. IA : William C. Brown, p.--t. "We define [persuu"ion[
c"" including atlitude~. emotions. intention .... and behavior ... :' Prau. M .. and Perot. R. ( 1993). p(JI'.\lIasil'e cOlJlflllllliclf!;oll campaigns. BaSIOn: Al lyn & Bacon.
p.6. " Persua sio n is, in all case:., the activity of demon<.;trating and allempting to change the behavior of at l ea~t onc person through symholic interaction. It i... consciou ... ami occurs (a) when a threat to at least one person'" goals is ob... erved and (b) when the "ouree and degree of thi .. threat are suffi· cientl y important to warran t the expenditure of effort involved in IlCr... ua ... ion."
Reardon, K. K. (1981). Persllasion: Theory and cOlllext. Be\erly Hills. CA: Sage. p. 25. "The phrase 'being persuaded' applies to silllation~ where behavior ha ... been modi lied by symooli c transactioll'i (messages) which arc sometime:.. but not always. linked with coerc ive force (indi· reetly eoe rci\e) and which appeal to the rea ...on and emotion:- of the pcr'>on(s) being per'iuaded."
Roloff. M. E., and Miller, G. R. (1980). Persl/wioll: Nell' direcrimu ill theory lIfid research. Beve rl y Hills. CA: Sage. p. 15. " Persua~ i on
is human communication designed to innuence the autonomous judgments and action ...
of Olherl!. ...
Simons. H. W. (200 1). Persuasio/l i'l society. Tholl!'!and Oak!'!. CA: Sage. p. 7. [Persuasion isl "any message that is intended to ... hapc. reinforce. or change the responses of an· olher, or other...:· (based on Mill er's 1980 definition)
Stiff. J. B .. & Mongeau. P. A. (2003). Persurlsil'(' communication. New York: Guilford Press. p.4. (con/iI/lied)
16
Pari I • Prl'limillurie\ LJej;lIilioll\, r'Hld"
L -\ III .E 2.1
tllltl
11Il'(!Idfntl Clldl'lpillf1 illg.\
Continued
Without ulli:ring a ... peedl\.: ddulIUUIl. Irl'nholm ,I.lle ... Ilhll Pl'r'>Ua~i{)n cmtxldic\ the 1011u"lI1g char'H;tt.:ri<;tlc,,= Pcr<,ua ... iol1 i ... ",;,-ml"loll(' .lml nOllt'ocn.:i\c:' It · ...:reatc<;. reinforce .. , or dl:lIlge, re"pon\c .. :' it i .. "tran,aCIHlnal," anLl it i .. ··uhiquitou,."
Trenholm. S. (1I.)X9) Pl'nllll\iollll//(/ weial i,lllul'IIn'. Englewood Cliff... , NJ: Prentice-Hall. ·Pcr ... ua ... 101l cncompa ...... c ... the proct.: ...... c ... h) whidl lan guage and aclion'" IIltluC/H.:c the choice-
makmg of \lup,d'>c'" ,JIld nlhc ......
Williams, M. R.o & Coopcr. M.
I),
(2002). Power pt'rwlIsion: Mm'ing (III (If/ci£'111 or! illfo lite
IIIt,dia Ogl' (JnJ cU,), Greenwood. IN' Educational Video Group. p. 4. ··Pcr .. ua,>ulIl I' thc prot:c,>,> of prcpann!! Jmltkli\cring \-crbal and /loll\'erbalme",,,age<., to aUlOnomOll' Indl\-idual..; in onkr to ,lltcr or ,>1!"cngtht'll their attitude..;, helief", and behavior..;."
Woodward. G C. & Denton. R I (1992).1'('1"\//(/\/011 Will mj1ut!flce in American life (2nd ed.). Pro ... pect I kighl'>. IL: Waveland, p, 21
199-'- 1995). O\'CriaYlng ... uch traditional invc ... llgmionl" art: more recent. "nontraditional" ~tlHJII!!o, of per... ua~inn. empha"'l/tng IIH~rpt.'r ... onal or face-tn-face intluence. The wealth of ... lUdic ... on compliance gaming. cOlllpliance re ... l ... tlng. deception. and deception detection rcnect thi ... recent trend. \\hich can he charactcri/cd ~h a \\'::uep,hcd era in persua ... ion research. A ... Bo~tcr (1995) cOllllllented . "arguahly. in the 1a ... 1 15 year... the study of compliance-gaining mev·~age beha\ ior ha" heiLi the alLcnlioll of cOllllllunication scholars a~ much a .... if not more than. any other ... in glt.! topic III the di~l:ipli[lc" (p. 91). More recently, Wil ... on (199X) noted that ... cholarly inten.: ... t ill t'omplianl:e gai ning "is H!ry much alive" (p. 273). Thi ... evoluti on. or di\ergencc, in rc ... carch interc~h ha~ ...eryed only to muddy the definitional water.... Thu~. it remain ... unclear whether ... tudies of compliance gaining and deception repre ... cnt an extcn ... lon of per ... ua ... ion rc ... ertn:h or ... eparate lines of inquiry altogether (Burgoon & Dillard. 19(5). lhi ... lack of l:larit) i ... unfortunate when one considcr~ the lI11portl.lnce of definlllOlh 10 theOJ) huildlllg ~tnd "'l:hoJarly inquiry. In this regard, Di.IIKe ( 1970) noted that in the pmce ...... of l'on ... trm.:ting theories. a definition determines the behaVIOral field ob~c['\ed. which In turn affect" th e pnnciplc\ deriycd, the hypothe~es generated. and the system of law ...... tilted. Gl!orge Gerbner ( 1966) !o,imilarly argued that "thc choice of' a problem for ... lUdy and re ...earl:h, the allocation of resources. and the assessment of the rc\eyance of contrihutlon .... depend upon delinition!':o." (p. 99). Tht! fal:t that per ... ua ... ion re ... carch ha ... undergone a major transformation, and that a dh ergence in research method ... and foci has developed. invites a reexamination of how per ... uasion ... hould be defined and concep tuali/ed. Because definitions limit what is studied in a field. they may also limit \ariahlcs that arc given attention and in turn the ways in which we think and built! theoric ... of l:olllmunication. As Burke (1966) noted, onc's choice of terminology not only rcncch attention, it ~elects and deflects attention as well (1'.45). To this end. \\ie provide an anaJy . . i~ of t\NO fundamental criteria upon which prevail ing definition ... of pcrsua ... ion are based. illustrated by an accompanying djagrammatic rep-
Ch;.tptcr 2 • Embracing DiI'ergeflce
17
resentation. and suggest useful ways in which persuasion may be distinguished from closely related concepts such as influence. or social influence as it is often called, and compliance gaining.
Boundaries and Limitillg Criteria We wish La acknowledge from the outset that we maintain no illusions about there being a "corrcct" definition of persuasion. Various scholars and researchers conceptualilc persuasion differently and therefore subscribe to varying definitions of the tenn, And although there are some commonalities among some definitions. there are as many differences as there are similarities. The current transition from "public" persuasion to "interpersonal" persuasion has done little to clarify boundaries. Rather, as Boster (1995) noted. the emphasis on compliance-gaining research has had the effect of broadening the focus of study. Studies on the use of touch as a compliance-gaining strategy, for instance. have broadened the persuasion construct to include nonverbal behavior, as opposed to more traditional studies focusing on language and discursive symbols. Studies on deception have similarly examined the behavioral and physiological correlates of deception, rather than symbolic forms of influence. The emergence of such nontraditional research interests has had the cffect of expanding the swath of human activities that potentially may be regarded as "persua-.ion."
Pure versus Borderline Cases of Persuasioll Our position is that many of the definitional vagaries can be clarified, if not resolved. by focusing on two considerations. The first is whcther a given scholar or researcher is attempting LO define "pure" persuasion- what Simons (1986) and O'Keefe (1990) have labeled "paradigm" cases of persuasion-versus all of persuasion, including its periphery. which we term "borderline" cases of persuasion, By pure persuasion. we refer to clear-CUI cases on which almost all scholars in communication and related disciplines would agree. As examples. nearly everyone would include a presidential debate, a television commercial. or an allorney's closing remarks to ajury as instances of persuasion. Other instances. though, lie closer to the boundary of what we normally think of as persuasion. Not everyone would agree that a derelict's mere appearance "persuades" passersby to keep their distance. Nor would everyone agree that when city planners install speed bumps 011 a street where speeding is common. they are "persuading" motorists to slow down. Such cases are less clear-cut. Much of the disparity in definitions. then. is rooted in the fact that some scholars and researchers are concerned with "pure" persuasion. whereas others are concerned with borderline cases as well. The perspective, shown in figure 2.1, illustrates this distinction. ' As the gradation or shading in the figure suggests, the threshold between pure and borderline persuasion is fuzzy rather than distinct. What implications does the shift in emphasis from traditional to nontraditional research interests have on "pure" versus "borderline" conceptualizations of persuasion? Among other things. there appears to be a need to expand the scope of persuasion to encompass nonverbal behavior and implicit social cues that accompany face-to-face encoun-
18
Part I • Preliminarie.f: Dljinition.f, T"('II(I.f, lind Theoretical Underpinllings
Borderline Persuasion
Intrapersonal
Unintentional
Interpersonal
Intentional
Symbolic Nonsymbolic
No efIects
Free choice/Awareness
Coercive/Unaware
FIGURE 2.1
A Model oJ Pllre l'erSIlS Borderline Cases of Persuasioll
ters. Becau'e a good deal of what happen~ in interpersonal encounters occurs at a low level of awareness (Langer, 1989a, 19 ~9b), we suggest, as has Roloff (1980), that much of the influence that takes place operates at a similarly implicit level. As an illustration. cultural factors may influence an individual's choice of compliance-gaining strategies without the individual's consciou,,> awarene~s (Wiseman, Sanders, Congaiton, Gass, Sueda. & Ruiqing. 1995). As deception detection research has shown, nonverbal cues, such as blinking or smiling, may al;o affect perceptions of veracity (Feeley & Young, 1998; Hale & Stiff. 1990; Seiter. 1997; Vrij, 2(00), even though a good deal of nonverbal behavior is nonmindful. If these kinds of studies arc to be included within the purview of persuasion. then the delinition must be broadened accordingly.
Choice of Limiting Criteria A second consideration in defining and conceptualizing persuasion involves the limiting criteria that form the basi, ror a given definition. EI ... ewhere we have identified five key
Chapler 2 • Embracing Dil'ergence
19
criteria that constitute the basis for nearly all definitions of persuasion (Gass & Seiter. 1997; 2003). These are (I) whether persuasion is intentional or unintentional, (2) whether persuasion must be effective or successful. (3) whether free will or conscious awareness must be involved, (4) whether persuasion necessarily occurs via language or symbolic action, and (5) whether persuasion can be intrapersonal as well as interpersonal. In this chapter. we focus on two of these five criteria in an effort to differentiate the types of influence studied in our field. Although each limiting criterion is discussed separately for convenience's sake. many definitions incorporate both criteria simultaneously.
Intentionality and the Issue of "Accidental" Persuasion.
Many definitions adopt a "source-centered" view by focusing on the sender's intent as a defining feature of persuasion. Bettinghaus and Cody (1994), Burgoon, Hunsacker, and Dawson (1994). Cegala (1987),0' Keefe (1990), Pcrloff (1993), Petty and Cacioppo (1981). Reardon (1981), Stiff (1994), and Woodward and Denton (1992) all adhere to this requirement. making it the most common characteristic of standard textbook definitions (see table 2.1). Certainly, "pure" persuasion would seem to fall into this cmegory. When one considers traditional cases of persuasion. one tends to think of conscious. intentional efforts. along the lines of the classic fear arousal studies. Compliance gaining, too. would appear to satisfy this requirement. to the extent that compliance gaining is conceived of as planned. goal-directed communication (see Dillard. 1989. 1990; Dillard. Segrin, & Harden. 1989). But whm about other borderline cases of persuasion? We suggest that "accidental"' influence often takes place with little or no conscious awareness on the part of the influential individual. Appearance cues, for instance, may affect credibility judgments, without the source's awareness that he or she is conveying such impressions to others. Research on social modeling (Bandura, 1977) illustrates this phenomenon as well. As just one example, parents commonly instill beliefs. impart values. and model behavior for their children. Yet as any parent can allest, many of the lessons parents "teach" their children are completely unintended. Even when persuasion is intentional. many of the cues conveyed by a source may be unintentional. such as the appearance of nervousness or the lise of a powerless language style. Such unintended cues nevertheless carry considerable persuasive weight. In this regard, Cooper and Nothstine (1992) have commented that "those we try to influence through our persuasion may well respond to aspects of our persuasive actions that we are not aware of" (p. 3). This certainly holds true for research on nonverbal and physiological correlates of deception. h is precisely those cues over which deceivers have little conscious control that may serve as the most reliable indicators of deception . Another way in which an "intent" criterion is troublesome involves the prospect of unintended receivers. Two studies (Greenberg & Pyszczynski, 1985: Kirkland, Greenberg, & Pyszczynski, 1987) clearly demonstrate the operation of the "unintended receiver effecL" In these studies the researchers created a situation in which third parties overheard ethnic slurs directed against African Americans. The results of both studies revealed that the slurs they overheard led the third paJties to evaluate the individuals at whom they were directed less favorably. Of course, even traditional. public persuasion can involve unintended audiences. An advertising campaign might target one group of consumers but provoke a backlash among nontargeted consumers. Traditional laboratory studies, though, have typically relied on clearly defined target audiences and carefully constructed mes-
20
Pan I • Preliminaries: DefilliTions. Trefld~·. and Theof'('rical Underpinnings
sages in the form of st imulus videos or booklets. th ereby affording few or no opportunities for unintended receivers to be affected. Thi s lack of emphasis on unintended receivers. however, can be seen more a~ a limitation of traditional research paradig ms than a rationale for restricting the scope of persuasion. Basing a definition of persuasion o n intentionality makes good se nse if one w ishes to focus primarily on "publi c" persuasion. From an " inte rperso nal" perspect ive. howe ver, such a limiting crit eri on implies a rather linear view of communi cation. An intent requirement is problematic insofar as co mpliance gai ning is concerned. becausc compliancc gaining in the interpersonal arena is mutual or bilatcral. Berge r and Burgoon (c ited in Burgoon & Dillard, 1995) have und erscored this point in criticizing so me researchers fo r "their failure to acknow ledge the reciprocal nature of social influence processes in interpe rso nal commun ica tion" (p. 398). Interpersonal Deception Theory (Buller & Burgoon. 1996; a lso see chapter 14 of this vo lume) also stresses the importance of acknowledging the transaction a l nature of deceptive co mmuni ca tion . Where influence is reciprocal , whose intent cou nt s? Do both parties' influence attempts ha ve to be inte ntional ? Does compliance re."isling constitute a persuasive inte nt ? Standard textbook definitions that are based on an intent criterion do not provide clear-cut answers to th ese questions. In fact less than a third o f th e definitions listed in table 2. 1 ex plici tl y acknowledge the reciprocal nature of pers uasion. In o rd er to accommodate nontraditional persuasion studie s. some clarification or modification o f the intent require me nt would see m to be in order. At a minimum. adherellis to an intent requirement shou ld explain how the requirement is to be applied in twoway. transactional setti ngs. We suggest th at one pragmatic approac h toward clarifying the situation and establi ~hing some modest boundaries is to use the term "per~uasion" to refer sole ly to inte ntional effo rt s and the term "intlue nce" to refe r to e ither intentional or unintentional outco mes. Thus. conscio us modeling of behavior co nstitutes persuasion, while uncon sc i ou~ modeling sho uld be con~ id e rcd a form of influence. According to this scheme, " influence" can be understood as an umbrella term that e nco mpasses any and all fonn s of persuasion, wheth er pure or borderline. ~ " Persuasion" shou ld thu s be considered a special ca~e of inilucllcc that sat isfies an inte nt requ ire me nt. Thi"i conceptu ali zation is useful because it prese rve s the intent requirement embodied in man y definitions of persuasion. while simultaneously acknowledging that at times beliefs, attitudes. and behaviors can be modified accidentally. Although this distinction may ~eem obviolls, it is wOl1h noting lhat most sc holars and researchers use persuasion and influence interchangeably. Few of the stand ard tex tbook definitions make any di stinctions whatsoever between these terms. This di ~tin c ti oll still leaves o pen the qucstion of whether compliance gaining is best thou gh t of as a part o f persuasion or soc ial intluence. We tend to favor Dillard 's conccptualiLatioll of co mplian ce gaining as a type of planned, goal-directed communicati on (Di llard. 1990; Dillard. Scgrin. & Harde n. 1989). At the same time, however, we can envision situations in wh ich one person gains another's compliance without intending to do so. For example. a shopper might dec ide to buy a specific brand because she/he obse rved another attractive shopper selec tin g that brand. A pedestrian who crosses the slreet when th e li g ht is red may e nco urage other pedestrian s to follow, based on status cues associated with the first indi vidual 's dress or appearance. The most practical way to resolve thi s ambiguity, we s u gge~ 1, is to return to the distinction between pure and borderline
Chapter 2 • Embracing Dil'ergence
21
cases of persuasion made earlier: In its purest form, compliance gaining is an intentional, effortful activity, ''Typical'' compliance gaining can thus be conceptualized as a subset of persuasion. a subset that takes place in face-to-face settings. In some borderline instances, however, compliance may be secured unintentionally or accidentally. In those "nontypical" instances. compliance gaining can be conceptualized as a subset of influence that takes place in face-to-face settings. Intentional compliancc gaining should therefore be considered part of persuasion, whereas unintentional compliance gaining should be regarded as an aspect of influence.
Effects and the Issue of Unsuccessful Persuasion.
In addilion 10. or instead of. adopting an intent requiremcnt. some scholars and researchers have opted for a "receiveroriented" definition by restricting persuasion to situations in which receivers are somehow changed. altered. or affeeled. The definilions by Bostrum (1983). Cegala (1987). Cooper and NOlhstine (1992). Johnston (1994). Larson (1995). O'Keefe (1990). and Pfau and Perol (1993) embody Ihis requiremenl in varying degrees (see lable 2.1). An effects requirement would seem part and parcel of compliance gaining as well. The very term compliance gaining. suggests that a persuasive outcome is being sought-although the form the compliance must take is less readily apparent. In fact. if one wishes to focus on pure cases of compliance gaining. it seems sensible to combine both an intent and an effects criterion: The prototypical case of compliance gaining is a planned. purposeful effort to secure compliance. usually in the form of behavioral conformity. in response to a request or other message recommendation. Despite the intuitive appeal of the above. we cannot help mentioning one or two reservations about restricting all compliance gaining. both pure and borderline cases. to an effects criterion. We submit that a person can be engaged in an activity. whether or not he or she is performing the activity well. A salesperson. for example. could be engaged in selling without necessarily closing a deal. A parent might try to persuade a child to say no to drugs but might fail in the effort. Similarly, we suggest that a person can be engaged in the activity of persuasion even if it is ineffective persuasion. The difference in approaches hinges on whether one's focus is on persuasion and/or compliance gaining as an olllcome. or as a process. This distinction mirrors the discussion some years ago about the conceptualization of argument as a product (argument)) versus argument as a process (argllment~) (0' Keefe. 1976).-1 If persuasion or compliance gaining is viewed as an outcome. then limiting the use of either term to successful intluence attempts makes perfect sense. If. however, one is interested in studying the process or activity of persuasion. such a limitation is questionable. Consider the wealth of studies on compliance-gaining strategy selection. which we believe tell us a great deal about persuasion. Granted. the seeming preoccupation with strategy selection studies. at the expense of studies focllsing on actual compliance, has been rightly criticized (Dillard. 1988). Nevertheless. investigations into the dynamics of the persuasion process. such as failed or foiled persuasion. serve a legitimate research purpose. For inslance. Ifert and others (Iferl & Bearden. 1998; Ircn & Roloff. 1996; NeeI', 1994). examined how sources respond when targets resist initial influence attempts. And any number of studies on compliance resisting have focused specitically on how targets avoid complying altogether (see for example. Kazeoloas. 1993: Kearney. Plax. & BUIToughs. 1991; O' Hair. Cody. & O·Hair. 1991). It would be an odd slale of affairs,
I 22
P;lrt I • Prelimin(lrie.\': Dejin;fiol/.\, Trel/ds. and Thcurcrical UlldelpillllillKS
indeed. if sli ccessful compliance gaining were considered a bona fide area for persuasion research but successful compliance resisti ng were not. Nor does it help to argue that successful resistance is in itself a type of effect. Because successful compliance resisting implies unsuccessful comp liance gaining, v irtually every compliance-gaining e ncounter cou ld be construed as a , ucce" for one side or another. Much th e sa me may be said about research on deception and deception detection. Vrij (2000). for insrance. has c riti cized others' definitions of deception as incomplete for failing to include unsuccessful a:-. we ll as successfu l altcl11pts at deception. Based on an effects crite ri on, success full y duping someo ne would clearly seem lO constitute not only deception but persuasion as well: it is puq)Oseful. and it achieves ils intended effect. But what about success ful deception detection? Success in detecting deception spells failure for the deceiver. He nce. re liance on an effects criterion would exclude studies on deception detection from the scope of persuasion, eve n thou gh such s tudie ~ make up a significan t portion of the literature (M ill er & Stiff. 1993). Such a limitation appears to "cut the baby in half:' in Solomon ic fa,hion, by focusing on on ly half of the deception-- influence peddlin g is cOlllmonl y a two-way street. Do effects have to be observed in only one or in both interactants? And who dec ides if the effort was a !oI u cces~? Moreover. com pliance i~ rarely an all-or-nothing affair. The participants may succeed in ~ome rc~pec ts but fail in ot hers. or succeed only partially. The difficulties are co mpou nded when one i ~ faced with participants whose goab may be multiple, sketchy . and changi ng. and whose w illin gnes~ to acce pt various outcomes or compromi ses may fluctuate during the com muni cation enco unt er. In the case of research on ueception or deception detecti on, it see ms much Illore practical to inc lude all sti ch investigations under thc rubri c of persua sion. In facl. we believe the case has already been made. rath er (;ollvi nci ng ly. that dece ption is a form of persuasive activity (S tiff, 1995). Delibe rate falsifications. omissionloo. or di stortion s all sati sfy an intent requ ireme nt. and the goab of deceiver~ co rre~po nd with those traditionally associated with pe r<.,uas ion, for exam ple. affec tin g beliefs. attitudes. and behaviors. Although hypothetically ~ome forms of deception, suc h as unco nscio us omissions, may be unintentional. virtuall y a ll of the deception Iiteratu re to date has foc used on dece ption as conscio us. effortful act ivit y.
A Graphic Represelltatioll of Persuasion The preceding discussion leads us to the diagram of persua sio n shown in figure 2.2. whi ch e ncompasses research on compliance ga ining. co mpli ance resisting. deception. and dece p-
23
Chapter 2 • Embracillg Dil'ergellce Effective Pure Persuasion
Borderline Persuasion
• Successful, purposeful persuasion • Successful, purposeful compliance gaining • Successful deception
• Social influence • Accidental persuasion
Intentional
Unintentional • Failed or foiled persuasion • Compliance resisting • Deception detection
• Other communication
Borderline Persuasion
Nonpersuasion Ineffective
fiGURE 2.2
Four Quadrants Model of Persuasion, Based on Intentionality and Effectiveness
tion detection. The model consists of four quadrants, segmcl1led according to the two limiting criteria we have presented: intentionality and effects. The upper left quadrant represents pure persuasion, or persuasion in its most prototypical form. Pure persuasion. as we noted earlier. satisfies both an intent and an effects requirement. The upper right and lower left quadrants represent borderline persuasion, or more "iffy" instances of persuasion. The upper right quadrant. labeled "influence," reflects those situations in which persuasion '"succeeds without trying," that is. where there are unintended effects. Influence here includes unintentional social modeling. nonverbal, or appearance cues that are conveyed without conscious awareness, as well as other forms of "accidental" innuence. As per our earlier discussion, we've classified all such unintentional alterations of beliefs. attitudes. and behavior under the umbrella term "influence." The lower left quadrant represents unsuccessful persuasion or, conversely, successful attempts at foiling persuasion. Thus, successful compliance resisting and successful deception detection are included here. Although one could argue that slIccess in resisting compliance or detecting deception qualifies as pure persuasion, we believe there is merit in distinguishing between proactive and reactive influence attempts, with compliance resisting and deception detection representing the latter. An admitted weakness of our fig· ure, however, is that the physical separation of the quadrants implies that activities taking place in one quadrant are somehow separate and distinct from those taking place in another quadrant. Concepwally, and in practice, interpersonal persuasion is transactional; interactants are operating in several quadrants simultaneously. Thus. interactant A may be actively engaged in attempting to persuade interactallt B. while at the same time trying to resist being persuaded by B. The lower right quadrant represents "nonpersuasion." that is, words or deeds that neither seek to persuade nor manage to influence. Since every definition of persuasion included in table 2.1 includes either intentionality or effect::; as a requirement for persuasion, it ::;cems safe to say that communication that fails to satisfy either of these tests
24
Part I • Preliminaries: Defllliriol/s. Trends. alld Thcorerical Ullde,pillnillgs
should not be considered part of persuasion, or influence. o r compl iance gaining. Communication that was neither planned nor goal·directed, for instance, and had no effect on another's beliefs. attitudes. or behavior would fall into this category. A good deal of noncommmunicative behavior would fall into thi s category as well.
Discussion and Implications This approach, we believe, offers a practical. workable scheme for conceptualizing persuasion and ih related terll1 ~. We acknowledge that it represent s all approach toward conceptualizing persuasion. not the ollly approach. NeveI1heless. we suggest that conceivin g of persuasion in this manner offers a number of benefits. FirM. thi s approach makes meaningful distinctions between the terms persuasion, inf7uence. and compliance gaining. Bused on criteria cen tral to most definitions o f these term s, our conceplUaliz.alion clarifies where one concept lea ves off and anmher begins. Hitherto, these term~ ha ve been used with considerable ambiguity, often interc hangeably. and ofte n with different or even contradictory meanings even within the sa me article or text. This approach makes it relatively easy to classify most investigations as falling primarily into one of the four quadrants we've identified. Second, this approach accounts for and integrates both traditional and nontraditional approaches to ~tudying persuasion. The role and relationi)hip of more traditional "public persuasion" studies to more nontraditional "face-Io-face" research on compliance gaining is clarified: If both are intentional and successful. they are one and the same thing accord· ing to our conceptualization (the upper left quadrant). However. if compliance is secu red unintentionally rather than as a form of planned, goal·directed activity. it is better c1assi· fled as a form of influence rather than persuasion (the upper ri g ht quadrant ). If compliance gaining is intentional but unsuccessful. the event may be better thought of as successful compliance resisting (the lower left quadrant). Finally, this approach clearly identifies what isn 'f persuasion: communication or behavior that ne ither seeks to, nor succeeds in, moving another person to think or do somethin g. The graphic representation we've offered is subject to some limitations. One ~ u c h limitation. mentioned previously, is that persuasion, especially face-to· face pers uasion. often operates in several quadrants simultaneously. A persuasive interaction also might begin in one quadrant and end in another. Because persuasion is a dynamic process, thi s is to be expected. Our conceptualization incorporates this overlap. but the figure cannot. Another limitation with our figure is that the boundaries between the quadrants don't represent distinct categories so much as characteristics that are matters of degree. It isn't really the case that a given interaction is or isn't persuasion. Rather, a given interaction possesses more or fewer of the characteristics that persuasion comprises. This is especially true insofa r as an effects criterion is concerned. Persuasion is rarely wholly successful or unsuccessful. Usually. there are degrees of success or failure. A persuader might succeed in changing another's attitudes, but not as much as shefhe intended. Or a pe rsuader might change another's altitudes, but not the other's behavior. And whenever an effec ts criterion is in use. it rai ses the question of when. or at what point, the effects should be measured. Some effects may be sho rt ·term, so me long-term. and so me hoped -for
Chapter 2 • Embracing DiI'ergellce
2S
effects may never materialize. We would thus suggest that the two criteria that make up the quadrants, intentionality and effects, be viewed as dimensions along a continuum rather than as discrete characteristics. Some persuasive intentions may be more clear-cut or obvious. while o lh ers may be less carefu ll y formulated or planned. Some persuasive effects may be easily measured or quantified. whereas others may be more subt le or difficult to detect. A third limitation with the figure is that compliance-gaining studies focusing 011 "strategy se lection" or "strategy preferences," and using hypothetical scenarios, are sti ll difficult to classify according 10 this scheme. While such studies involve intentional efforts to per~uade, inasmuch as strategy selection entails conscious planning, they include no measurable, dis.cernible outcomes. The goal of strategy selection is comp li ance, but in many studies succes~ in achieving the goal is never measured. Should slich investigations be classified as part of pure persuasion. given that the objective is to secure compliance. even if success in achieving this objective is never considered? We subm it that the problem in classifying such swdies resides not in our approach to conceptualizing persuasion but rather in the inherent ambiguities of these investigations themselves. The reliance on hypothetical scenarios and the absence of even hypothetical measures of compliance, we suggest. is the source of the difliculty. We suggest that hypothetical strategy selection studies are parl of borderline persuasion rather than pure persuasion, since they involve intenrionai, but not necessarily successful , efforts to persuade. We would locate such invc~ligations midway between the upper left and lower left quadrants (see figure 2.3). Such a classification acknowledges that in these inveMigations compliance is an open question: there is neither success nor failure. because the issue of success is never raised. Of course. compliance-gaining studies
Effective
Intentional
Pure Persuasion
Borderline Persuasion
• Successful, purposeful persuasion • Successful, purposeful compliance gaining • Successful deception
• Social influence • Accidental persuasion
hypothetical, "strategy - - - selection" studies on compliance gaining
---1--------------
• Failed or foiled persuasion • Compliance resisting • Deception detection
• Other communication
Borderline Persuasion
Nonpersuasion
Unintentional
Ineffective
FIGURE 2.3 Four Quadra"ts Model Differentiating Persuasioll, Influence, Compliance Gaining, alld Complia"ce Resisting
26
Part I • Preliminaries: Definitions, Trends, and Theoretical Urulerpirmings
that do measure tangible effects in the form of raffle tickets sold, tips left by restaurant patrons, willingness to relinquish a photocopy machine, compliance with physician recommendations, and so on would qualify as pure persuasion. A final consideration, though not necessarily a limitation, is that our approach seems to suggest that any intentional communication that produces an effect must be regarded as persuasion. To the contrary, however. we suggest that communication can be both intentional and effective without necessarily being classified as persuasion. One could, for example. desire to have an enjoyable conversation with a friend and succeed in doing so. If the goal were simply to relay information, swap stories, or share experiences, and not to change the other's mind or behavior. one would be hard pressed to classify such an interaction as pure persuasion. Perhaps, then, it would be useful to distinguish between a communicative intent and a persuasive intent, the latter seeking the traditional goals of attitude, belief. or behavior change. We would suggest that when a communicator's primary goals entail attitude, belief, or behavior change. the interaction would best be characterized as pure persuasion, but when these constitute a communicator's secondary or tertiary goals. then the interaction would best be characterized as borderline persuasion, or nonpersuasion. In any case, we're not terribly bothered by this concern, because it seems to us that the ingredients for persuasion are present in most forms of human inter-
action. Most communication is, al least in part, instrumental in nature. Most language usage is, as Richard Weaver (1970) has surmised, sermonic, that is, inherently normative or evaluative. We believe that the approach offered here serves to clarify uncertainties and ambiguities involved in definitions of persuasion, intluence, and compliance gaining. The two limiting criteria we have identified can assist in framing a discussion of how each of these terms should be conceptualized. Regardless of whether one agrees with our conceptualization. the discussion provided should serve to clarify some of the central issues involved in any attempt to define persuasion and closely related terms. Definitions shape fields of inquiry. They direct attention toward some research foci and delleet attention away from others. As Shepherd (1992) observes, "definitions are consequential" (p. 203). They affirm or deny. encourage or discourage. permit or omit. Definitions contain assumptions, they embody values. and they espouse points of view. "These perspectives." writes Andersen (1990), "launch scholars down different theoretical trajectories, predispose them to ask distinct questions, and set them up to conduct different kinds of communication slUdies" (p. 309). We believe that our approach to conceptualizing persuasion is both meaningful and practical because it is inclusive. It encompasses and affirms both traditional and nontraditionaJ approaches to understanding persuasion. Such inclusiveness is desirable to ensure that nontraditional studies, such as those focusing on compliance gaining and deception detection, are recognized as instances of persuasion research, and to ensure that the term "persuasion" is not equated solely with traditional topics and variables. such as fear appeals. source credibility, and the like. We hope the viewpoint we have offered prompts greater interest in conceptualizing about persuasion. Conceptual discussions perform a heuristic function insofar as theory building is concerned. Conceptualizations not only innuence what theories are formed, but how the process of theorizing takes place. Conceptualizations also influence the kinds of research questions that are asked. as well as the methods used in trying to answer them.
---------------------------------------------------------------------------------------------,
Chapler 2 • Em/}racillM DiI"ergel/C('
27
Notes ___________________________________________________________ 1. The degree of fUllincss or di"'lim.:livcncss in the lhrc"hold between pure persu~sion and borderline depends largely on which limiting criterioll one is employing. Intentionality offer.. a cle'lf demarcation between pure and borderline persuasion. but only if there i ... agreement on how persmlsivc intent i ... c:-.tablished. (J<., it a persuader .... tated inll!l1t thaI count .... or a rcecher" ... perception of an intent to per.. uade. or a third pany's inference of un intent \0 pCf'.uadc?) An cffeel ... criterion is more continuous: .,w.;cc ... sful persuasion is Iypically a mailer of degree. 2. We sec lillie difference between the term" ·'inllucm.:c" and "social inllucnce:' other than the ooviou ... implication that the latter entaiJ.; ~ocii.ll factor" of \Ollle kind. Onl! cuuld ea<.;ily argue that all innuence entail .. social factor" in one form or another. Even "dr-per"uasion \\(Julll email ccnain 'iocialization procc"'''e<;;. J. Ironically. 0' Kcefe!. who highlighted the di"tinction between argument as a product and argument a" a process, hi rmel !' \lIb!o.cribe~ to an dTect<; criterion for defining pcr\u:.\ion (O'Keefe, 1990). 'The notion of .w('cess, " he wrote. ·'i .. embedded in the concept of perslIay, 'I pcrwaded him, hut failed.' One can 'lay. 'l/ried to per~u;Jde him. but failed,' but to ,ay 'limply, '1 per).uaded him' i, to Imply a \ucce"''Iful ;lttempt (0 innucrll.:e" (1990. p. 15). In our view. 0' Keefe begged the quc<.;tion in the abme exampk, by U'IlIlg the term "llCNHlded:' Ir one u'les the term "pcr<;;u
Referellces_____________________________________________________ Andersen. P. A. ( 1991). When one c:lI1not not cOllllllunic;lIc: A cilal1cngr.: to Motley'" traditional communication po,tulates. CO/ll/l1II11il'lIt;OIl SlIuli('.I, .J2(-t), 309~325. Bandura. A. (1977). Sociallearll/Jlg tlll'Orl'. Engle\~ood Cliffo.;. NJ: Prentice Hall. Bavela~. J.B . (1990). Bchaving and communicating: A reply 10 fl.lotlcy. Wellem )ol/mai oj S'peech Cammimica/itlll. 5.J(-t). 59J~602. Beach, W.A. ( 1990). On (nou ob~cr\ ing behavior intr.:ractionally. Wi'Hem )olll'llal o/Speedl COfllmllllica· lio/l, 5.J( -t). 60.1-612. Berger. C. R .. & Burgoun. M. (1995) . Pn:face. In C. R. Berger & M. Burgoon (Eds.). COflllllllfl;nuioll and weial ill/l,li'I/{'e prna.I.\(', (pp. ix-xi). Ea't Lano.;ing, MI: Michigan State Unr\er"rty Pres..,. BCliinghaus. E. P .. ~lI1d Cody, M, J (]99-t). Pn.\llll\il'e ('OI/IIIIIIII;nllioll (6th cd.). Fort Worth. TX : Harcoun Bracc. Bo"ter. F. (1995). A commentary on compliance-gaining ll1e, ...agc beha\lOr rt:'icarch , In C. R. Berger & M. Burgoon (Ed .... ), emill/lllllic(ltitlll lind wwial il/jlllellcl' Jlroce,He,l' (pp. 91-11.1). East Lansing. M[: Michigan Stale Univer:o.ity Pre..,..,. Buller. D. B.. & Burgoon. J. K. (1996). lnlerpcNlIlal deception lileor). COlI/mlmii'atiOlt TlleOl:\,. 6(3).
203-242. Burgoon, M .. & Dillard. J. P. (1995). Communication and 'Iocial inlluence: A prolegomenon. Communi· catioll Rel'eard,. 22(-t). 397--401. Burgoon. M .. lIun\acker. F. G .. & Daw!o.on. E. J. (19(4). !-I,IIIIW, CfJllllllllllicatio/l (3rd ed.). Thousand Oaks. CA: Sage. Burke. K. (1966). Lllllglluge u.,nmbo/ic (lc/;oI1. 1.0... Angcic-:. CA: Univcr'iity ofCaJifornia. Berkeley Preso.;. Ccgala. D. J, (1987). PerlIlCHi\'(' ('Ol/lIIlIlIIicatioll: Tlll'OII' alld prac//n' Urd cd.). Edina, MN : Burgess intcrnational . Clevcnger. T. (1991 ). One cannot nOI communicate? A connicl of mo(kl". COlllllllmicllliml Studies . .J1(-t),
340-353, Cooper. M .. & Noth ... tinc, W, L. (1992). Power 1'(,/,,1/1(/1/(1//: Mm'il//.: all allcielll or/ il1/o lire medi(l age. Greenwood. IN : Educational Video Group. Cron~hite. G. (19~61. On the foeu,. \t'()PC, and coherence of the ,IUd) of human >i),llloolic acti, ilY. QlIarterly )01/1'110/ (~r Speech. 72. 23 1-2-t8.
28
Part I • Preliminaries: Definitio//I, Tn'mh. amI '/I,('ol"l,tical Undl'l"pillllillgs Danr.:c. F. E. X ( 1970), The concept of cllIllInu nic.ltlon . )m/rlUi/ oj Cmllll/llflicmiofl, 20. 201-210. Dance. F. E. X. & Lar... on. C. E. (1976). TIl{' }/IIu(iol/\ oj IWlIlllll nmlllllll/;Cal;O", Nc'W York: Holt.
RlIlchart. and Win\;ton. DiltanJ. J. P. ( 1988). Compliuncc-g.aining mCI;sagc "election: What is ollr dependent variable? Commul/i("(IfiOIl MOllflKraphf.
55. 162- lln.
Dillard. J. P. (Il)Xl). Types of inllucncc goal" in pcNlIlal rclalion,hlp", lUI/mal of Social mill Persollul Rl'/ariOlHhipI, 6,293-308. DIlIanJ, J . P. (1990). A goal-dri\c'~11 modd of inlerpcNmal inllucncc. In J. P. Dillard (Ed.). See!.:./"R complilIllCl''- Ihe product of ill/erpu.\OIw! illfit/elln' /lwHflgei} (pp. 41 -56), Scothdalc. AZ: Gorsuch Scari .. bricJ.. . Dillard. J. P .. Segrin. C. & Harden. J . (ItJ89). Primary and ..econdar) goab in the intcrpcr. . onal influcncc proccv... CommulI;c(l/ioll Mmw).:rtlp/I.\. 5ri. 19 .19. Feeley. T. H .• & Young. M. J. ( 199~) . Iluman:.- it, lie detcctor.. : Some more .. ccond thought ... Commllll;ca'irm Qu(/nt'rh. -16. 109-116. 0:1'''' lUI.. & Scitcr. J . S. (2003). /'t'I".IIWI·;(II/. \OC;(I/ ill/TIIl'lIce, (II/(/ ("OII/pli(lllce gaillillg (2nd cd.). Boston: A llyn & Bacon/Longman. Cia' ... R H .. & Scitcr. J. S. (1997. Novcmber). On defining per-ua .. ion : Tov,,'ard a contcmporary perspective_ "lOp thrcc" paper pre ..cntcd atlhe annual cOll\cntion of the Wc.,tern Communication Association. ~1onterc). CA. Gcrbncr. G. ( IlJ(6). On defining. ("olllmunicatlon: Still another vicw. Jourl/al of CommJlni('(llif)1l, 16. 99-
IOJ. Greenberg, 1.. & p) vC/yn"ki. T. f 1(85). The effect of an overheard ethnic "Iur on e\ ahmtions of the targel' 110"' to .. pread a .. ocial di ..ca .. c. 10111"1101 of Etpenmellwl SocioJ P~yc"{Jlogy. 21. 61-72. lIak . J S .. & Still. J. B. (I9<Xl). NOllvcrhal primal:Y III \eracit) judgmellh. Communicalion Reports, J, 75-8.1. Ifer! DE .. & Bearc.kn. L. (1998). The Influence of argumentativcne .... and verbal aggressi ...·eness on re, .. pon .. c.. to fc..'fu ..ed reque .. ",. COllllllllllicaliml Rt'lmrl.\". 11(2). 145· 154. Ifen. D. E., & Roloff. M. E. (1996). Responding to rejected rcque~h : Pcr"istence and response type as function of ob .. tacles to compliance. Journal of l..l./Ilgulige ami Social P~ych{)l()g\". 15. 40--58. John .. ton. D. D. (1994). The (lrl ami I("iellu' ofperwll.Iioll. Madison. W I: William C. Brown . Ka/cola .... D. (1991). The imp::u.:1 of argumcntati\·cnc .... on resistancc 10 persuasion. Htlf1um Comnllmicalioll Rt'H'(/nh. ]0. liS-Ln. Kearney. P .. Plax. T. G .. & Burroug.h .. , N. F. (1991). An attributional analysis of college students' resistance deci .. ion ... Cml1ll11f11;("(lIiO/l Edlfctuio/l. 40. .115--341. KirJ..land. S. L.. Greenberg. J .. & PY ...7czynsi..i. T. (1987). Further e\ idellcc of the deleterious effects of o\erheanJ dCnl£ator) ethnic labek Derogation hcyond the target. PenOlIllJily alld Social Psychology Bullt,tlll. /3(1).216--127. Langer. E. J. (19g9a). IHl1Idfulneu . Reading. MA : Adthsun- We,Ic) . Langer. E. J. (llJ8lJh). X11nding nHlIter... In L. Bcrkowlil (Ed.). A"mlll ·t'.1 ill t'xperimenroi social psycho/, OK" (\01. 21. pp. 137-173). New YorJ..: Addi.,on-Wc ... lcy. LINin. C. U. ( 19lJ5). Per~u(lsi(Jl/: Rl'("('l'lioll (/Ild I"t'''l'0/llibilin' (7th ed.). Belmont. CA: Wadswon.h. Llln (1990). The mfluence of recclver,' re .. istancc on pcr.. uader.. · \erbal aggressiveness. Communicatioll Quanah. 3,.;. 170--188. t-.lillcr. G. R. (19~O). On being pcr.. uaded : Some ba .. ic di,tlllction:-.. In M. E. Roloff & G. R. Miller (Eds.). Perll/(/Iioll: Nt'II'direcl;O//\ ill IlIm,-y (/Ild I"('ware" (pp. 11 -28). Beverly Hill..;. CA: Sage. Miller. G. R.. & Stilf. J. B. (1993). Def"l'fuil't' col/lllllmiclItion. Ncwbury Park. NJ : Sage. Motley. M. T (1990a). Communication .. interac.:tion: A rcply to Beach and Bavelas. Western JOllmaJ of Spt'l'Ch Camllll/ninllioll. 5-1H). 61 3~62J. Motlc). M. T. (1990b). On whether on\! can(not) nol conlillunicate: An examination via traditional communicat ion po .. llIlate ... Wl'Ml'I"lI JOllmal of Spl't'ch COIIIIIIUllinlli(J/I, 5-1( I). 1-20. Motley. M. T . (1991) Ilow one Illay nol communil:ate: A reply to Andcr.;en. Commu1Iication Srudies. -12(4). ~26 -3.19. Neer. M. R. (1994). Argumentative nexihility tl'" a facl(lr influencing mcs.,age response style to argument;tli\c and aggrc .... i\ c argucr... A1"1-:1/111['111£11;01/ lIml A(/\'{)("(l("\', 3/. 17-33.
Chapter 2. • Embracing lJin! rgellce
29
O·Donne[l. V .. & Kable. J. ([982). Per'llU/sioll: All illtL'r{/ctil'e-dl'pelll/t'IIC\ approach. N!.!\\ York: Ran dom House. O·Hair. M. J .. Cody, M. L & O'Hair. D. ( [991). The impact of situational dimen~ion ... 011 cornp[ianceresisting !ootfatcgics: A compari"on of methods. C()IlIIl1IlIIiclIflOII Qllant'ri.\'. 39(3). 226-240. O' Keefe. D. J. (1976). Two eoncept~ of argument. JOlin/til of IIII' Ameril"llll Fore/Hie A \.\ocillfiol/ In:named Ar/.:lIl11t'II/(ltiolltlllt/ Adl"Occl£:rl. 13. 121-128. O'Keefe. D. J. (1990). Per,w asioll: Theory alld re.H'(lrcil. Newbury Par~ . CA: Sage. Perloff. R. M. (1993). The dYliall/ics ofper.\·II{/.~io". Hillsdale. NJ: Erlbaull1. Pelly. R. E.. & Cacioppo. J. (1981). Altill/de~' and per.\"IU/sirm: Owsic (llId C{)lIfem{Jomry (/pprollch('.I. Dubuque. lA: Will. C Brown. Pfau. M .. Kenski. H . C. Nit7. M., & Soren~on, J. (1990). Efficac) of inoculation !oolratcgie ... III promoting re"i~(ancc to political allad Illc"'''ages: Application to direct mail. COlI/lIllllli,tllim/ Mmwgraphl, 57.25--13. Pfau. M .. & Perot. R. (1993). PerJ/UI.I·;l'e mmmllll;m/ioll (·oll1paigm. Bo~\()n : Allyn & Bal:on. Pfall, M .. & Van Bockern. S. (199-1.). The persistence of inoculation in confirming resistance to smoking initiation among adole~cents: The "econd year. 1-111//1(/1/ COIlIIlIlIlIinllioll Research, 20. 413--BO. Pfau, 1\1 .. Van Bockern. S .. & Kang. G. J. (1992). U'e of inoculation to promote re"'l ... tancc to smoking initiation among adolescent .... Comllllll/iClllirm MOllogmplr'i. 59, 213-:!30. Reardon. K. K. (1981). Per.wmioll: Theory lIlId COIlIt'xl. Be\crly Hilb" CA: Sage. Roloff. M. E. (19~O) . Self-awarenc<, ... and the pcr~lIasion pnKcss: Do we rc .. t1ly know what we're doing? In 1\1. E. Roloff & G. R. Mi lia (Ed~.). Per,I//(/l"ioll : New direuiol/I ill them:,' (lnd n.Il'lIrch (PI'. 2966). Beverly Hilt.., CA: Sage. Roloff. M. E.. and Miller. G. R. (Ed .... ) (1980). PnH/(I.\·irm: Nell" tlin'('l;mH III theo/'\" (llltl rnearch. Bc\erly Hi lh, CA: Sage. Seiter. J. S. (1997). Honest or deceitful? A study of peNJIl ... · mental mod!.!h for judging. \!.!racity. HIIIII(lI/ COfllmllllil"(lIioll Re.'ieal'c/J, 24(2). 216-259. Shepherd, G. J. (1992). Communication a. . influencc: Dclinitional exclu ... ion. COIIIII/lllli("(ltioll SllIdies. 43. 203-219.
Simons. H. W. (1986). Per~lI(Jlioll : UIIl/er.\llllldillg. practice. lind (IIw/ni'i (2 nd cd.). Ne\\ Y()r~ : McGraw-HilI. Stiff. J. B. (1995). Conceptuali/ing deception as a pcrMIa~ive acti,lly. In C. R. Berger & M. Burgoon (Eds.), COll/l1lll11icarioll lind ,weial illj1l1ell("(.' pmct'Sses (pp. 73-90). East Lan ... ing, M1: Michigan State University Press. Stiff. J. B. (1994). Per.\l/{uil'e COllllllllllimtioll. New Yor~: Guilford Pre ...... . Trenholm. S. (1989). Per.\It{lsion {lnd .weilll illflJl('I/(·e. Englewood Clift... , NJ: Prentil:c-lIall. Vrij, A. (2000). Dl'Il'c ting lies and deC"t-iI: The ps\'chologr of lying (I/J{I Ihe implicilI;ml.1 for profe.nimllli I'l"(Icrice. Chithc,ter. England; John Wiley & Son .... Weaver, R. M. (1970) Language i~ <'crlllonic. In R. L. Johannesen. R. Stri{;~land. & R. Euhank<, (Ee"'.), Limgll{lge il' \'etll/ollie: Richard M. Welll'er 01/ the I/(ltllre (If rhe/()ric (pp. 179-2 11). Baton Rouge. LA: Loui~iana State Univer"ity Press. Wibon. S. R. (1998). Introduction to the ~pccial i<.,.,ue nn \ecking and rc~i~ting compliance: The \
Witte, K. ( 1995).
U~ing
\Care tactics to promote "afc -.ex among juvenile detention and hl£h ...chool youth.
jOIll7lal of Applied Comlllllnicafion Rese{ll'f"h, 23( 2). 128-1 42 . Woodward. G. C. and Denton. R. E. (1992). Per.wasioll (llId illfllU'I/CI' il/ Americal//if(' (2nd cd.). Pro<,-
peet Heights. It: Waveland .
3 Trends and Prospects in Persuasion Theory and Research Daniel J. O'Keefe
Systematic th ink ing about persuasion dates at lea~L as far back as the fifth century B.C.E., (Q Greek scholars such as Protagoras. Gorgins, Plato. isocratcs. and (especially) Aristotle. And in subsequent years persuasion received continuing altcntion from a variety of scholars within a broadly humanistic tradition (for a review. see Conley. 1990). But in the first part of th e twentieth century. the development of social scientific methods provided new avenues to the illumination of persuasion. The social scientific study of persuasion is often traced to Carl Hovland, a Yale University psychologist who. following World War 11 , initi ated a systematic program of persuasion research (see. e.g .. Hovland. Janis. & Kelley. 1953). On a great many topics in pcr!-.ua!-.ion research, the first work was done by Hovland or his associates (though historical accuracy compels some notice of the persuasion research that considerably predated Hovland·s. e.g .. Knower. 1936: Lund. 1925: Schanck & Goodman. 1939), Over the last 50 year.... social scientific persuasion research has flowered. Indeed, relevant research has been conducted in a great many academic fie lds. Nearly all the social sciences (i ncluding communication. psychology. sociology. political science. and anthropology) and related applied endeavors in which social scientific questions and methods appear (such as advertising. marketing. public health. medicine. law, bus iness, educati o n, env ironme ntal studies. and so on) contain research focused on persuasion. This surely refl ects th e pervasiveness of persuasion in human affairs. The marketplace. the courtroom. the campaign trail. the boardroom. the family- in a ll these selLings (and more) human dec isio n making is shaped by persuasive communica ti on. This chapter discusses three broad recent developmellls in the social sc ientifi c study of persuasion and social influence. Some aspects of these developments are al ready rather well along and have begun to bear significant fruit: others are on the horizon and offer s ubstantial promise. But a ll testify to the continuing vi ta lit y of persuasion theory and research.
31
32
Part I • Preliminaries: Definitions. Trends, ami Theoretical UI/derpinl/il/g.\
Beyond Attitude Change In its most basic form, persuasion involves changing persons' mental stales, usually as precursors to behavioral change, Of the various mental states that migfll be implicated in persuasion, attitude (understood as a person' s general evaluation of an object) has been the center of research anemion. Correspondingly. per~uasion has often been conceived of as fundamentally involving auilude change. This might involve a change in the valence (positive or negative) of the evaluation or a change in the extremity of the evaluation (as when an attitude changes from extremely negative to only slightly negative), Thus understood, auitude change is obviously an important aspect of persuasion. All sorts of decisions, from what products to buy to which candidate 10 support. arc plainly subject to changes in attitudes, But in various ways persuasion research ha . . seen a broadening of interest beyond this usual foclls. as reflected specitically in interests in three other sorts of outcome variables.
Other Attitude Properties First. properties of attitude other than valence and extremity have come to be recognized as potentially important foci for persuasive efforts. That is. rath~r than inlluencir'lg the direction or extremity of an altitude. a perf..uader might want to inlluence ~ome other attribute of the altitude, such as its salience (prominence. acce~sibility), the confidence with wh ich it is held. the degree to which it is linked to other attitudes. and so forth (for discus sions of some such attitudinal properties, see Bromer. 1998: Eagly & Chaiken. 1998: Petty & Krosnick. 1995; Roskos-Ewoldsen, 1997). For in!'.tance. where receivers already have positive attitudes toward one's product, the persuasive task may be to ensure that those attitudes are salient (activated) at the right time, perhaps by somehow reminding receiver:.. of their attitudes. So, for example. a manufacturer of food products may not particularly care whether people are thinking of its products while driving. but it does care what attitudes are activated when people are shopping for groceries-so it places advertising displays in sllpermarkets precisely in order to engage existing positive attitudes at the point of purchase. A number of such attitudinal properties have been grouped together under the general heading of "attitude strength" (for some discussions. see Bassili. 1996: Eagly & Chaiken, 1998. pp. 290-292: Petty & Krosnick. 1995: Raden. 1985). Conceptualizations of attitude strength vary, but a useful illustration is provided by Krosnick and PeLty's (1995) proposal that attitude strength is best ulld~rstood as an amalgam of persiqence (stronger attitudes are more persistent than are weaker ones). resistance (stronger attitudes are more resistant to change than are weaker ones). impact on information processing and judgments (stronger attitudes are marc likely to affect slIch processes than are weaker attitudes), and impact on behavior (stronger attitudes will have more effect on behavior than will weaker ones). Consider. for instance, (hat (wo persons might have attitudes toward a particular political candidate that we re equally positive (say. with a rating of 6 on a scale of I to 7) but differed in strength: Pat's positive attitude is weakly held, liable to nuctuate from momenlto moment. not very resistant to persuasion. and not very strongly connected to behavior. whereas Chri s's (equally positive) attitude is more strongly held, more stable
Chapter 3 • Treml., alit/ Prmpeu\ ;11 Penumiml n'e(lrY (lml R{,.,eaniI
33
over time. less likely to be altered by counterpcr ... ua ... lon. Jnd more likely to be cxpre ...... ed in corre ... ponding beha\ ior ( ... uch a ... votlllg for the candidate. working in the candidate· ... campaign. and ... o on). Therefore. t.!\Cn though Pat and Chri<., have identical attitude ... in one sem.e (they ha\e the same o\'erall cyalllatlOn). their attitudes arc rather different in other way ... (Chris's is ... tronger than Pat·\). In \uch a <.:in.:um\tance. obviou ... ly. the candidate would like to ... trengthen Par's altitude not nccc\\arily to make the evaluation more extreme. blll to make the attitude better anchored. l11or~ stahle. more connected 10 behavior. more resi\lUnt 10 counterpersua ... ion (in ... hon. to makc it more like Chri\'s attitude). To put the matter mon; generally. pcr\uadcr\ \ol11clil11c'" will have an interc\t in innuencing not merely thc valence and extrcmity of an attitude hut abo it\ ... trcnglh.
Other Mel/tal States Second. mental ... lale\ other than nttitude have hcen rccogni/cd a ... potential persua ... ion larget .... Two examples of such state\ arc IHmmlllVC cnn ... ideration ... and \elf-efficacy.
Normative Consideratiolls,
Variou\ kinds of belief.... anout norm!'. can be relevant target ... for p~r\uaders. For in .... ance. peoplc'\ hellef.... ahout "t.le ... cripti\·e norm ... ··-perception ... of what most people do--may innuence actHlI1\ and thu ... he a focus for per... u<1\ive effort ... (Cialdini. Kallgren. & Reno. 1991). For ilbti.lJ1l:e. l'ollege ... tudents appear commonly to overestimate the frequency of drug ami ah.:oh()1 u\c on Iheir ci.\mpu\t!\ (Perkin .... Meillllan. Leichliter. Ca ... hin, & Pre ... h.!),. 1(99). Slll:h o\'crl!"'llInation can in lurn lead ... Iudenh Ihcmsclve!'. to engage in exccs ... i\,c drug and alcohol u\e (hecau'ic of a belief that "everybody is doing it. "'0 it Illu ... t bc okay"). OhviOl"ly. Ihen. pt.!r'illa!'.I\'C intcn'c ntion ... aimed at correcting ... uch misperceptions of de ... criptivl' norlll'i might he helpful in reducing drug and ah.:ohoi abu,e (Haines & Spear. 1996: Miller. Monin. & Prent ice. 2000: Stefl!an. 1999). Similnrly, what the Theory of R~asoned Action (Fi..,hbcil1 & Aj/,en, 1975) term\ Ihe "\ubjective norm"- the pcr. . on·\ perception that ,ignific:lI1t other... de ... ire the performance (or nonpcrforJ1l~lI1ce) of the beha\ ior may hc a per . . uf.lsioll target. For in"'lance. one way of pep,lIading a 'InlOker to quit 11'<.1) ne to (on, inl'l! him thai other\ who are important to him (hi ... \pou\e. hi ... children. hi<., be ... t friend) think that he \hould quit. That i... , by altering Ihc receiver'\ conception of ,>,:hal ,ignifi(
Self-Efficacy.
Self-efficacy (or pen.:eived neha\'ioral control). the per\on's perceptIOn of hi ... or her ability to perform the b(:ha, lOr, i ... another mental \tate that hns come 10 be "cen a~ an important potential focu ... for per\ua ... i\'e etrorh (see Ajzen. 1991: Bandum. 19R6). Somelimes the barrier In n receiver· ... (olllplmnce \eel11\ not to be a negative attitude or negative norms. but rath~r a pl!rc~ivl!d inahility 10 perforl11lhc aClion succe\sfully. For in"'lance. a per\on might have a po\ithc altitude toward engaging in regular excrci\c and have posilive nonnative beliefs ahout thm Hctivity. hUlmight nevertheless not even try to exerci ... e regularly necau . . e ~he helic\ C\ Ihat 'ihe i ... inciJpabk of exerci ... ing regularly (because exercise is too timc-conMlming. doe ... n't fit her "'l'heduh::, require ... too much specialized equipment. and "'0 forth) . It i\ ('a ... y to imagine how a perceived inability to perform the behavior might underlie failure . . to C\('I'l·i,e. tI\~ condnlll\. quit ~ll1okiJ1g. and so forth.
34
Part I • Prelimillaries: Dejillitions, Trends, arid Theorerical Underpinnings
Research is on ly begi nning to acc umu late co ncern ing how persuaders mi ght address such self-efficacy conce rn s: there is so me indi ca tion. for exa mpl e. that mode ling (showi ng someone s ll cces~full y pe rforming the behavior) and rehearsal (g ivi ng persons an opportunity to practice the behavior) can be useful avenues to innuencing self-efficacy (A nderson. 1995.2000: Hagen. Gutkin. Wil so n, & Oats. 1998: Maibach & Flora, 1993: Weisse. Turbiasz. & Whitney. 1995).
Behavioral Outcomes Third, some linc!>. o f research ha ve focu5ed direc tl y on behavioral outcomes. as in studi es of the foot-in-the-door and door-in-the-face strateg ies. (For a ge neral di sc ussio n of slich strategies. see c hap ter 12.) Research that is focused on behavioral outcomes serves as a reminder that even when persuaders seek to change me ntal states (s uch as attitudes, normative beliet\. or self-effi cacy perceptions). influencing mental states is only a mean s to an e nd. Behavioral change is common ly the ultimate goa l. In a se nse. studies emphasizing behavioral out come!-. approach per~mbion effec ts from a direction exactl y opposite to that taken by research emphasizing mental !-.talcs: Instead of first centering on attitude change and subsequelllly taking lip the que~tion of how altitudes are related to actions. these lines of research cen ter on be havioral effects and the n take up th e question of what me ntal-s tate mechanisms might accou nt for the ob!-.c rved behavioral effec ts.
SUlllmary In sum. per!-.uas ion re~earc h ha:-. co me to recognize thar attilUde change is not the only ou tcome variable of intere st. Although persuasion researc h has typicall y focused on atti tude cha nge, increa~ing atte nti on is being give n to other outcomes-and, correspondingly, to new mechanisms of persuasion. A fter a ll. the means by which one might influence attitudes are not necessarily the same as th e mea ns by which one might influence other outco mes. One may hope that the cOlllinuing attention of resea rchers to the se additional outcome va ri ab les will lead to new understandings about means of social influence.
Context-Specific Research Persuasion research has ge nerall y been aimed at de ve loping co ncepts, finding s, principles, and theories that are useful across a wide ran ge of persuasio n settings. Of course, any parti cul ar persuasion study commonly involves so me parti cular context of persuas ion : The slUdy exa mines co nsu mer advertisements. or appeals on so me public policy question, or arguments abou t a legal case. and !ooo on. Still. the research aim has typi cally been the developme nt of find ings not bound to a ny part ic ular persuasive c ircumstance. However, an increasi ng amoun t of persuasion researc h has been appearing in studi es addressed at specific contexts of applica ti on. with co rres ponding development of context-specific co ncepts and modeI!-.. (For so me di sc uss ions of such contexts. see chapter 16 by KJingle. chapter 17 by Baxter and Bylund. c hapter 18 by Hirokawa and Wagner, and chapter 19 by Seiter and Cody.)
Chapler 3 • Trends and Prospects in Persuasion Theory alld Research
35
A useful example is provided by the articulation of various "stage" models of health
related behavior, exemplified by the transtheoretical model of health behavior (so named because putatively it integrates a number of different theoretical perspectives). The tran,theoretical model (sometimes called the "stages of change" model) identifies a number of distinct stages in a person's adoption of a given health-related behavior such as
engaging in an exercise program (see Prochaska & DiClemente, 1984; Weinstein, Rothman. & Sutton, 1998). In the precontemplation stage, a person is not even thinking about undertaking an exercise program anytime soon: in the contemplation stage, she is at least seriously thinking about doing so; a person in the preparation stage is ready to change and
may have undertaken planning or other preparatory action (such as signing up for a health club); in the action stage, she is undertaking the exercise program; finally, a person who continues to engage in exercise for some time is said to be in the maintenance stage.
Stage models offer the prospect of shedding light on persuasion, because of their potential usefulness in suggesting how best to tailor persuasive efforts to a particular audience. For example, for persons in the precontemplation stage. the persuader's challenge
will presumably be to get receivers thinking about the target behavior (i.e .. moving persons from precontel1lplation to contemplation). By contrast, for people in the preparation stage, the persuader will want to help people translate their plans and intentions into actions. (For some examples of investigations of the effectiveness of stage-matched health interventions, see Jamner, Wolitski, & Corby, 1997; Naylor, Simmonds, Riddoch, Velleman, & Turton, 1999; Quinlan & McCaul, 2000.) As another example of context-specific research, consider investigations of the persuasive effects of negative political campaign advertisements that attack a political candidate without necessarily even mentioning the preferred candidate. Studies of the effects of negative political advertising are commonly not especially concerned with contributing to general cross-context understandings of persuasion processes, but rather reflect a specific interest in illuminating this one facet of political campaigns (see, e.g., Basil. Schooler, & Reeves, 1991; Garramone, 1985; Haddock & Zanna, 1997; for a review, see Lau, Sigelman, Heldman, & Babbitt, 1999). Indeed, quite independent of whatever light such studies might shed on persuasion processes generally, they are valuable contributions to an understanding of how persuasion works in this particular setting. One may detect in these developments an implicit recognition of the potelllial limits of general 1110dels of persuasion. No single theoretical view of persuasion is likely to
provide a complete, wholly detailed account of every single possible persuasion circumstance-and such should not be asked of a persuasion theory. ){"s enough that a
general theory of persuasion offers broadly useful concepts and principles that are helpful in a variety of circumstances, even if insufficient to answer every possible question about any given persuasion setting. But this in turn suggests that particular persuasion contexts may demand correspondingly particular treatment----{;ontext-specific concepts, contextspecific principles.
Of course, the study of a specific persuasion context may both reed and be fed by general theorizing about persuasion. A nice example is provided by research on inoculation mechanisms, that is, mechanisms for making receivers resistant to counterpersuasion. Several studies have examined how general understandings of inoculation might be applied to the specific problem of creating resistance to negative political advertising; the
36
Part I • Preliminaries: Dejiniliolls. Trends. and Theoretical UI/{/erpillllillgs re~ea rc h ~ u gges t s that the c ffcct~ of such ads ca n be blunted if. before they appear, the candidate "inocul ates" voters by engaging in appropriate rebuttal of the attack s (sec. c .g .. prau & Burgoon. 1988; prau , Ke nsk i. Nitz. & Sorenson. 1990). Prev ious general researc h on inoculation gu ided the contex t-s pec ific work concerning inoculation against negative campaign ads~a nd the con tex t-specific work in turn has provided addi ti onal general information abo ut inoculati on processes. (For more on inoculation. see chapter 15 .)
The Complexity of Persuasion Effects Per!o.tl<.lsion phenomena are complicated. making the development o f dependable generalizations rather difficult. For exa mple. it is difficult to identify any particular persuasion tactic th at i!o. effec ti ve in all situations. Indeed. the researc h literat ure on persuasive effect!o. co nt ai ns many exa mpl es of appare ntl y inco nsiste nt findin gs. One researcher's study finds that beller liked co mmuni calOr~ are s ignifi cantly more persuasive than less well liked ones, whereas another st udy finds no such effect; one investi gation reports that statin g the message's conc lusions explicitly sig nifi ca ntl y e nhances persua sion. whereas a subsequent study fail ~ to obtain a significa nt effec t, and so on. But several rece nt developments in the study of persuasion have he lped to identify so me of the so urces of suc h complexities, thu s prov id ing a basis for better und erstanding how and why suc h diverse e ffects mi g ht arise. These develop ments are expressed briefly in the following three subsection s.
Moderatillg Factors First. a given persuasion variabl e ca n produce different effec ts under diffe rent conditions: a variable might sig nifi can tl y influence persuasive outcomes in o ne circum stan ce. but ha ve relatively little effect in another. For instance, acknowl edg in g potential counterargument . . (argument s aga inst the advocated view) has different e ffects depending on th e message's topic: It redu ces the persuasiveness of messages co ncernin g public policy questions. bu t not the persuasiveness of co nsu mer product advertisements (O' Keefe, 1999a). Many stud ies of persuasive e ffec ts can be desc ribed as a sea rch for possible moderating facto". that is. factors th at alter the impactlhat one variable has o n another. This ge neral idea is particularly prominent in dual-process mode ls of persuasion such as the Elaboration Likelihood Model (E LM) (Peny & Cac ioppo. 1986; see also chapter 5). The ELM sketches two broad avenues to persuasion: ( I) a "cenlral route" in which receiver~ ca re full y process message arguments and (2) a "periphcral route" in which rece ivers rely on me ntal short c ut s (" heuri stics") as a means of reaching a conclusion. One import an t moderating variabl e that affects which route is activated is th e receiver's degree of in volvcmen t with the topi c. As involvement increases. reliance on heuristi cs decreases and c lose message processing increases. From th e point of view o f th e ELM, it makes perfect ly good se nse that. for exa mpl e. the communi cator's likabilit y will so metimes ha ve a subs tan ti al influence on persuasiveness but on other occasions will playa very small ro le. When invo lve me nt is relatively low. communicator likability may have conside rable impact (because receivers decide whe th er to agree with lhe message by usi ng the short cut
Chapter 3 • Trends and Prospects ill PersIJasion Theory lind Research
37
of whether they like the communicator). When involvement is hi gh. however. the persua~ sive effect of likability will presu mabl y be muted (because receivers will be paying more attention to th e detail s of the message's arguments). One of the signal co ntributions of the ELM is to have systematized a number of apparentl y inconsistent findings by di stinguishing central-route and peripheral-route persuasion processes.
Multiple Roles for Variables Second-and also closely connec ted with the ELM-it has become clear that a given variable might play different ro les in persuasion in different ci rcumstan ces (for ge neral di s~ c ussion of thi s point. see Pett y. 1997; Petty & Wegener. 1998). As a si mpl e example. consider the impact of variations on the sheer le ngt h of a written message. Message length might playa role in persuasion by servi ng as a c ue for a heuri stic such as " longer messages probably have more good arguments." That is, receivers mi ght rely on message length as a shortcut for deciding whether the advocated view has merit (see, e.g .. Wood. Kallgren. & Preisler. 1985 ). Alternatively. message length mi ght innuence how much attention the receiver pays to the message ( more specificall y. mi ght influence the audience's motivation to process the message' s arg uments closely). For example. on a co mpl ex technical subj ect. recei ve rs might decide not to pay much attention to a short message (reasoning that it wouldn't be likel y to pro vide the necessary detail). whe reas a longer message would engage their attention (see. e.g .. Soley. 1986). Again, notice that suc h variati ons in the role played by a give n variabl e can lead to apparentl y inconsistent results across studies. Longer messages mi ght produce e nhanced persuasion when receivers re ly on message length as a shortcut to reaching a conclusion about the advocated position. but length mi ght have little sys tematic effect on persuasive outcomes when it influences the audience's moti vation to process the message (because close r processing of the message does not necessarily gU£lrantee greater persuasivenes~ of the message, and th e audience's close scrutin y mi ght uncover weak nesses in the advocate's argumentation).
Message-lo-Message Variability Third. above and beyond the first two complexities, accumulating empirical evidence suggests that there is message- to-message variation in the persuasive effects of message variables. That is, even taking into account kn own moderating factors. a given message variable (for exa mpl e. high versus low fear appeals. as discussed in c hapter 13 by Cho and Witte) will not necessarily have identical effects in every message; rather, the effect is likely to vary from message to message (see Jackso n & Jacobs. 1983: O' Keefe. I 999b). The existence of such variability po ints to a potential weakness in the kind of resea rch design that has co mm only been used in ex perimental per!}uasion st udies. In the 1110st co mmon sort of design, an abstract message category is represented by only one COIlcrete message. So, for instance. in studi es of the re lati ve persuasivenes~ of hi gh and low fear-appeal messages. researchers have usually co mpared one particular low fear-appeal message against one particular high fear-appeal message (the experimental co unterpart of th e low fear-a ppeal message, identical in every way except for the fear-appeal variatio n).
38
Part I • Prelilllil/{/rie.\': Defilliriolls. Trends. lind Thl'oretic(I/ Underpinnings In this "single message" re~earch design. each message category (such as "high fear ap-
peal") is repre~ented by only one sample message. But the effect of variations in fearappea l level is likely to be different from one case to another and from one message to another. To put the point more generally. the effect of a given experimental message manipulation (such as fear-appeal level) in the one particular message being studied is likely to be different from the effect of that same variation in other messages. Thus a si ng lemessage research de~ign leaves someth in g to be desired insofar a~ ge neralization is concerned: gauging the overa ll effect of a message variat ion requires exam ining its effect in multiple messages. not just a single one. In other words. dependable ge neralization across messages requires multiple instances. One way to obtain multiple-message research evidence is through th e inclu!o.ion of multiple Ille~sages in a sing le study. For examp le. suppose a researcher wanted to investigate the relative effectiven!.!ss of two different kinds of political att ack ads: those focused on the opponent's issue positions and those focu~ed on the opponent's image or character. Rather than comparing just one particular issue-attack ad with just onc particular imageattack ad (where the results might renect pecl1liaritic~ of the ads in question). the researcher could gather a large number of examp les of each kind of ad and compare the average per~uasiveness across the two sets of ad~. Such multiple-message ev id ence wou ld obviou~ly provide a better basis for ge ne ralization than would a si ng le-message design. A second way to obtain mliltiple-l1le~sage research ev idence is through colla tin g results acro~s a large number of existing ~ingle~message ~tudie~. The most sys temat ic proccdure.s for ~lIch collation are to be found in meta-analytic sta ti s ti cal procedures. Meta-analysis i.s a family of procedures for producing a quantitative summary of a ~e t of existing rcsearch .studies (for a genera l introduction. sec Rosenthal. 1991 ). In a sense. a meta-analysis is a ".supcrstudy" that combi ncs the results of ear li er separate investigations. A meta-analysis can provide information not only about the overa ll average result (acros~ all the o;,lUdies) but a lso abollt the results within subse ts of studi es. In particular. the exist in g studies can be subdivided based on levels of a ~uspec t ed moderator va riable. ancl the results within these ~ubgroups can then be compared. For exam ple. cons ider the question of whether the effectiveness of the door-in-the-face (D ITF) stra tegy depends on whether the two requests were made by the same person or by two different people. In some DITF studies. the same person made the two requests. whereas in o th er DITF stud ies different people made these requests. Thus. some indication of the role of this variation as a potential moderator of the effect of the DITF st rategy can be obtained by subd iv idin g the studies. assessing the results within each subset. and then comparing the resu lts. (As it happens. this moderator variable does make a difference: The DITF strategy is more successful when the same person makes the two requests than when different persons make them: O'Keefe & Hale. 1998.) Meta-analyses are not ea~y to do. and ana lyzing multiple-message ev idencewhether obtained across studies. that is. meta-analytically. or within a s ingle study-raises some complicated i~~lIes concerning the approp riate stat istical analysis to be employed (for some discussion. see Brashers & Jackson. 1999: Hedges & Vevea. 1998; Jackson, 1992: Jackson & Brashers. 1994: Jackson. Brashers. & Massey. 1992). Obviously. however. meta-analysis otTers an appealing way of synt hesizi ng the results from many indi~ vidual persuasion studies (for <,;ollle examp les. see A llen & Preiss. 1998).
---------------------------------~
Chapter 3 • Trel/ds lind Prospects i" Persuasion Theory and Research
39
Summary Taken together. the three complexi ties mentioned here--the importance of moderator factors in persuasio n, the multiple roles that a pe rsuasion variable can play, and the existence or message-la-message variability in persuasive effects--quite naturally underscore the problems of ge neralizing abo ut persuasion processes and effects. Each of these complexities sugges ts Lhal th e results of any sing le persuasion study may need to be held rather ten tati ve ly, while the deve lopme nt of in creasi ngly deep understandings of persuasion phenom ena will require continuing systemati c research attention.
The Future To some extent. th e near- term future in persuasion research will like ly involve further articulation of so me o f the developments discussed here: increasing attention to outcomes o ther than attitude chan ge, e nco uraging more context-specific s tudies, and developing great er se nsiti vity to matters of generali zation (multiple-message studies, meta-analyses. and <"0 fo rth ). A~ in any researc h endeavor. there will undoubtedly be developments that cannot be fore~ecn. but two !>.pecifi c researc h s ubjects bear watching: ( I) visual aspects of persuasion and (2) compute r-med iated pe rsuasion. Persuasion resea rch has typically foc used on lingui stic aspects of messages, such as whether the message di sc lisscs co unterarg uments or cxplicitly states the advocate's conc lusio n. By co mpari son. rela ti vely little attention has been given to nonlinguistic features ~ u c h a\ visual images: ye t vis ual message e le ments might substantially influence pers uasive effec ts (for some general treatments. see Messari s. 1997: Scott. 1994). This is a particularly co mpl ex subject, especially as printed linguistic messages are also v i ~ ual image!>.. That is. printed text is itself a visual object (even if there are no accompanying pic tures): there appe~lrs to ha ve been little sys te matic pers uasion-related research addressing the!>.c visual aspects of text. (The idea that printed tex t is a visual object is certainl y familiar to an y student who has fiddl ed with a term paper' s margins so as to affect th e ap pare nt le ng th of th e pape r. or to any job applicant who has chosen a particular font so as to make a rcsu me look more professional.) Additionally. of cou rse. a printed message might contain non textual (that is, nonlingui~tic) visual mat erial , Sti ch as pictures or drawing s. Analyzing such images for argumelllutive content is notoriou sly difficult (for some discuss ion and examples, see Birdsell & Groarke. 1996: Blair. 1996: Fleming. 1996 ; Lake & Pickering . 1998; Nelso n & Boynt on. 1997: Oestermeier & Hesse. 2(00). Moreover, when a message contains both ling ui sti c and (non tex tual ) visual material. the relation ship between the two may be important. Within a print advertisement. for example, the relationship between the linguistic and no nling ui st ic vi!>. ual aspects of the ad may play an important role in influencing persua~ i ve effec ts. Within a television commercial, the rel ation ship between visual images, vo ice-over lin g ui sti c content , and printed linguistic content may playa similar role. Th e persua sive contributions of visual message elements (or of different relation!>. hip~ of visual and verbal c lc ment s in a message) are only beginning to be explored, and co nfid en t conclusions are some way off. For some cxamples of re levant studies, see Areni
40
Part I • Preliminaries: DejinitiollS, Trends. (llld Theoretic(li Ufu/erpinl1if/!<s & Cox. 1994: Figueiras. Price. & Marteau. 1999: Miniard. Bhatia. Lord. Dickson. & Unnava. 1991: Morrison & Vogel. 1998. Plainly. though. the study of visual aspects of per:-.uasion will be an important focus for future researc h. A second. and not unrelated. potential focus for fUiure research is persuasion and computer-mediated communication. Widespread access 10 computing is a relatively recent phenomenon. The personal computer was introduced in the 19805, and the first Web browser appeared in 1993. Correspondingly, there is as yet relatively little empiric;]1 evidence concerning aspects of computer-mediated persuasion. But obviously, a variety of relevant questions can arise. For example. what makes expert systems (computer-based reasoning systems that model human expert problem -solving) persuasive to users? (See Dijkstra, Liebrand, & Timminga. 1998: Jiang. Klein. & Vedder. 2000.) What elements make interactive or Web-based advertisements effective? (See Bezjian-Avery. Calder, & Iacobucci. 1998; Li & Bukovac. 1999.) How might the physical properties of computermediated communication systems innuence pers uasion processes? (See Moon, 1999.) Though relatively little can yet be said with much certainty about such matters. computermediated persuasion is likely to receive increasing research attention in the future.
Concillsion The developments surveyed here suggef..t that persuasion research is at once becoming broader (in expanding beyond attitude change as an outcome of interest). deeper (by developing context-specific concepts and principles), and more complex (i n recognizing the complexities of persuasion processes and the attendant challenges to genera li zation). Systematic thought about processes of persuasion can be traced back to the ancient Greeks, but as these developments attest. the study of persuasion co ntinues to be a locus of exciting theoretical. empirical. and methodological developments.
References_______________________________________________________ Ajlcn. I. (1991). The theory of piannl!d bdlavior. Or!:(U1i::'lItiulIlIl Bellm'lor lI"d Human Decisio" Pmcesses, 50. 179-211 . Allen. M .. & Prei s~. R. W. (Ed~.). (1998). PerSflasiOIl: Adl'll1lces through mela~Q1wl)'sis. Cresskill. NJ: 1·lampton Press. Anderson. R. B. ( 1995). Cognitive appraisa l of performancc capability in the prevention of drunken driv ing: A tcst of se lf-efficacy theory . JOl/mal of Public Re/atiolls ReJearch, 7,205-229. Ander~on. R. B. (2000). Vicarious and persua~ivc innucnce .. on efficacy expectations and intentions to perform breast self-examination. Puhlil' Rehlliom' Re,·;ell. 26,97-1 14. Areni, C. S .. & Cox, K. C. ( 199-1-). The pero;;uasivc effects of evaluation, expectancy and relevancy dimen· sions of incongruent visual and vcrbal information . Adl'(lIIces in Consumer Research, 21,337-342. Bandura. A. (1986). Social fOllndations (If /I/oUNlu (/Ild ac/ion: A social cognitive theory. Englewood Clifl":-.. NJ: Prentice-Hall. Basil, M., Schooler. C .. & Reeves. B. (1991). Po:-.itivc and ncg.lIive political advertising: Effectiveness of ads and perceptions of candidates. In F. Biocea (Ed.), Televisio" alld political advertisillg, 1'0/. J: Psychological pmn.'Hes (pp. 245-262). Il ilbdalc. NJ: Erlbaum. Bass ili. J. N. (1996). Meta-judgmental versus operativc indexes of psychological attributes: The case of measures of attitude strength. Journal of Perwl/olity alld Social Psychology, 7 J, 637-653 .
Chapter 3 • Trends (llId Prospects in Pas/unio/l Theory lind Rese(/rel!
41
Bt!l.jian-Ave ry. A .. C;.ilder. B.. & Iacobucci. D. (\998). New media interactl"e adverti .. ing versus traditional advertising. JourI/al ofA(iI'errisillg Research. 38(4). 23-32. Bird<.;ell. D. S.. & Groar~e. L. (1996). Toward a theory of vi:-ual argument. ArXllmenf(ltiofl (llId Adl'()C(lCY, 3.1. 1-10. Blair. 1. A. (1996). The possibility and actual it) of visual arguments. Argllwelllafiul1l1l1d Admc(Il·Y. 33.
23-39. Bmsher<.;, D. E .. & Jackson. S. (1999). Changing conception .. of "message effects": A 24-year overview. Hllmall Communicalion Rl'_\i'arclt, 25.457-477. Bromer. P. (1998). Ambivalent altitudes and information proce .... ing. SlIiH Journal of Psychology, 57. 125-~34.
Cialdini, R. B., Kallgren. C. A., & Renu. R. R. ( 1991 ). A focus theory of norlllalive conduct: A theoretical rclinemerll and reevaluation of the role of norm .. in human behavior. In M. P. Zanna (Ed.). Ad1'(111('('.\' in experimenral social ".\yc!wlogv (vol. 24. pp. 201-234). New York : Academic Pre .. !'.. Conley. T . M. (1990). Rlle/orit' ill lhi' European IradilhJII. New York: Longman. Dijkstra. J. J .. Liebrand, W. B. G .. & Timminga. E. (1998). Pt:rsuasivene<;, of expert sy!'.tem,. Bel/m'iollr alld Ilifomlt/lion Tt.'cirllology, J 7. 155- J 63. Eagly. A H .. & Chaiken. S. (1998). Auitude .. tructure and function. In D. T. Gilbert, S. T. Fiske, & G . Lindley (Ed<;.). f1wlllb(J(J~ of_weilll jJryclwlogy (-Ith cd .. vol. I. pp. 269-322). Bo~ton: McGraw-
II i II. Figueira ... M .. Price, II .. & Marleau. T. M. (J 999). Effects of textual and pictorial information uJ~)n perccption:-. of Down 'yndromc: An analoguc .. tudy. PsydwlogYlllld flealflr, 14.761 -77 1. Fi~hbein, M .. & Aj/en. I. (1975). Belief. lilli/lUll,. illlelilioll, lIml bt'hal'ior. Reading, MA: AddisonWc<;ky. Fleming. D. ( 1996). C.Ln picture .. be arguments? Arglwll'lIullioll and At/I'oclin', ]], 11 -~2, Garramone. G . M . (19K5). Effects of negative political advt.!rtising: Tilt.! rolc~ of sponsor and rebuttal. Jmmral of Br()(U/c(/Slilr};: al/{I EIt'ctrmric Media. 29. 1-17- 159. Haddock G .. & ZilUla. M. P. (1997 ). Impact of negativc advertising on t:valuations of political candidates: The 1993 Canadian federal election. Basic lIlId Applied Social Psw·lIology. 19.205-223. Hagen. K. M .. Gutkin. T. B .• Wibon, C. P .• & Oats, R. G . (1998). Using vicariou~ expt.!rience and verbal per"u:l<;ion to enhance ..elf-efficacy in pre-service tcacher.. : "Priming the pump" for consultation. Schoof Psycfwlog" QUllrter/\'. 13. 169- 178. lIaine ... M .. & Spem. S. F. (1996). Changing thc perception of the norm: A strategy to decrease binge drinking among college .. tudent!'. . JOllrnal of Ameri{'{/n Col/ege fleall". 45. 134-140. Hedge". L. V .. & Vevca. J. L. (1998). Fixcd- and random -e ffect" model, in mCla-analy:-is. Psy('holo~iCltl Methods. 3. -186-504. HO\ land. C. I.. Janis. l. L.. & Kelley. H. H. ( 1953), Commlllrimtioll lind penllasioll: PI'w:lw/ogictll III/d· ie,1 of opinion dUlIIge, New Havcn. CT: Yale Univer.. ilY Prcss. Jackson, S. (1992). Mt'Hage ej)t'cI.\ rl'searc/J: Principles of dt',~iRIl (11/(/ mralYJiJ. New York: Guilford Pres<;. Jad.!'oon. 5 .. & Brashers. D. E. (199-'). M > I: Analy<.;i .. or treatment x replication designs . Hl/lI/tlT! COII/mlllIi('(lrion Re,I't'(lrdl. 20. 356-3K9. Jachon. S .. Bra~hers, D. E., & Massey. J. E. (1992). St:.ltistkalte .. ting in treatlllenl by replication deo;igns: Three option .. reconsidered. COlll/lllmi('(llioll Quarterl\'. -10. 21 1- 227. lachon, S .. & Jacob.. , S. (1983). Grnerali/ing about mc<;!'.agcs: Suggc ..tion .. ror de .. ign and analy .. is or experiments. 111111/(111 Commllll/carioll ReWlIrch. 9. 169-18 1. Jamner. M. S .. \volit .. ~i. R. J .. & Corby, N. H. (1997). Impact of a longitudinal community HIV intervention targeting injecting drug users' ~tage of change for condom and bleach usc. AIIIl'rictln Journal ofllealth Prolliotioll. 12, 15--24. Jiang, J. 1.. Klein, G .. & Vedder. R. G. (2000). Pcrsua:-.ive expert system .. : The influence of confidence ill1d di ..crepuncy. Computers illllllman 8ell(ll·;or. 16.99- 109. Knower, F. H. (1936). Experimelllal studies of changes in atllllides: II. A study or the effect of prinled argument on changes in altitude. Jourl/ol of Abllormal(llld Social P.lyclwiogy, 30.522-532. Krosnick . J. A .. & PellY. R. E. (1995). Attitude .. trength: An over.'ie\\. In R. E. Pelly & J. A. Kror.nick (Eds.). Alfilllt/l'H rengrh : Allfet'l,deIlH {/lui col/sequence!)' (pp. 1-24). Mahwah. NJ: Erlbaulll.
42
Part I • Preliminaries: Definitioll5, Trend!i. alld Theoretical Underpinnings Lake, R. A .. & Pickering. B. A. (1998). Argumentation. the vi .. ual. and the po'i!>ibi l ity of refutation : An exploration. Argumentatio", 12.79- 93. Lau, R. R .. Sigelman. L.. Heldman . c.. & Babbitt. P. ( 1999). The e ffec\<; of negative poli ti cal ad\'cni~e menls: A meta-analytic asses~ment. American Poli/ieal Science Rel'iell', 93.85\-875. Li. H" & Bukovac. J. L. (1999). Cognitive impact of banner ad characteristics: An experimen tal study.
Journalism alld Mass Communication Quarter/\', 76. 34 1-353. Lund. F. H. ( 1925). The psychology of belief: A study of its emotional ;md voli tional determinants. journal of Abnormal and Social Psychology. 20. 174-196. Maibach, E.. & Flora , J. A. ( 1993). Symbolic modeling and cognit ive rc:hear<;al: Using video to promote AIDS prevention se lf-efficacy. Comllllm;CW;01l Research. 20, 5 17-545. Messaris. p, (1997). Visual persuasion: The role of images in advertisillg. Thou>;and Oaks. CA: Sagc. Miller, D. T., Moni n, B .• & Prentice. D. A. (2000). Plurali stic ignorancc and incon"i"lcncy between private altitudes and public behaviors. In D. J. Terry & M. A. Hogg (Ed".). Al1itllde.\·. behm,ior, and social context: Th e role of lIorms al/d group membership (pp. 95- 11 3). Mahwah. NJ: Erlbaum . Miniard. P. W .• Bhatia, S., Lord, K. R .. Dickson, P. R., & Unnava, 1-1. R. (1991). Pi ctme-ba ..ed per"u;.I .. ion processes and the moderating role of involvement. JOllma/ of CO/lSlllller Research. IR. 92- 107. Moon , Y. ( 1999). The effects of physical distance and respoll',e latency on pcr~ua!"ion in compute r-med iated communication and human-computer cOlllmunication. jO/lma/ of Experillll'/llal P.\·yc!w/ogy: Applied. 5. 379-392. Morrison. J .• & Vogel. D. (199R). The impact!- of presentation visual~ on persuasion. "lj"rmmlliofl arid Mallagemem, 33.125-135. Naylor. P. J., Simmonds. G.. Riddoch. C. Velleman. G., & Turton. P. (1999). Comparison of <,;{agcmatched and unmatched intervenlion ~ to promote exen.:ise behav iour in the primary care .. etling. Health Education Research. 14.653-666. Nelson. J. S., & Boynton. G. R. (1997). Video rhetori('.~: Teled.\ ed ad\·ert;.\;I1~;n AI1If'rjc(1I/ politic.l. Urbana. IL: University of Illinois Press. Oestenneier. U .. & Hesse. F. W. (2000). Verbal and vi .. ual causal argu ments. CO!:lIition. 75.65- 104. O·Keefe. D. J. (1999a). How to handle opposing arguments in pen.uasive mcv"ages: A mCl:l-ana lyti c review of the effects of one-sided and two-sided messages. In M. E. Rol off (Ed.). Communication yearbook. 22, 209- 249. Thousand Oaks. CA: Sage. O'Keefe, D. J. (1999b). Variability of persuasive message e ffects: Meta-anal ytic evide nce and implications. Document Design, J. 87-97. O'Keefe, D. J. , & Hale. S. L. ( 1998). The door-in-thc-face inOuence strategy : A random-cffech l1letaanalytic review. In M. E. Roloff (Ed_). Communication year/wok. 2/. 1-33. Thou sand Oak ... CA: Sage. Perkins. H. W .. Meilman. P. W .. Leichliter. J. S., Cashin. J. R.. & Pn:f. ley. C. A. (1999). Mi spcrcept ions of the nonns for Ihe frequency of alcohol and olher drug u ...e on college cam pu ..es. JOllmaJ of Amedcarl College Health, 47. 253-258. Pelty, R. E. (1997). The evolution of theory and research in ...oc ial psychology: From single 10 mUltiple effect and process model s of persuasion. In C. McGarty & S. A. Ha .. lam (l!ds.). The me.Hage of social psychology: Perspectiw':, on mind in society (pp. 268- 290). Oxford. England: Blackwe ll. Petty, R. E.. & Cacioppo, 1. T . (1986). Commullication and penlUl\ion: Central lind peripheral rollle:, to altitude change. New York : Springer-Verlag. Petty. R. E., & Krosnick . J, A. (Eds.). ( 1995). Altitud(' streng,I!: Antec('dellls {lilt! ('OIlsef{'Il'/ICl'S. Mahwah . NJ: Erlbaum. Petty. R. E .• & Wegener. D. T. (1998). Attitude change: Multiple roles for pcrsua~ion variables. In D. T. Gilbert. S. T. Fiske. & G. Lindzey (Eds,), Handbook ojs()cial psychology (4t h ed .. \01. I. pp. 323390). Boslon: McGraw-HilI. Pfau , M . & Burgoon. M. (1988). Inoculation in polilicul campaign communication . NUlIIlIlI Cml1ll/lll/ica (ion Rese{lrch, 15.91- 111. Pfau. M., Kenski. H. C. Nitz. M .. & Sorenson. J. (1990). Efficacy of inoculation Mralegics in promoting resistance to political attack messages: Application 10 direci mail. Commullication MOI/o,~f(/ph\',
57.25-43.
Chapter 3 • Trnul\ lIml Prmpeu.f in Pn\/w\;oll TItt.'Of'\' (Ifill Rt'.\e(/rch
43
Procha!'.\..a, J. 0., & DiClemente, C C. (1984). The fr(/I/I/hl'ol'l'rjm/ approach: CI'(III/II,1: fhl' IHidifirmal b()llIul(lril',~ of fherap.\'. Homcwood. I L: Dow Jonc .. In in. Quin lan. K. B.. & McCaul. K. D. (1000). Matchcd and mi .. matchcd intervention .. \1 Ilh young adult .. mo\..o. er:-.: Te .. ting a :-.tagc theory. Nt'(/I,h P,rl'ltology. II), 165 ~17 1 . Raden. D. (1985). Strength-relatcd altitude dimcn .. ion ..... SoC'/al PHc/lli1ogr Qtlll/"lt't'h. -IX. J 12-'\10. Ro<;cllIhal. R. (1991 J. MeW-llllllh/it· procedlll"e.I'.for If}cial I"e.leal"ell (re'" ed.). :-.Jell nury Par\.. . ("A: Sage. Rmko~-E ....,old ..en. D. R. ( 1997). AuiUldc accc..,.,ibilIlY and pcr,ua"ion: RC\ iell and it twn,aclll e model. In B. R. Burle ... on (Ed.). Comm/illicalioll .\'l'lIrhoo( 20 ( 185 ~225,. Thu u'and Oab. CA: Sage. Schand. R. (" .. & Goodman. C. (19.W). Rcaction ... to propaganda on txnh ... ide'i or a conlrU\cr'iwl i..... uc.:. P/lh/il ' Opill/Oll Quarter/I', 3. 107-111. Scott. L. M. (199..0 . Images in advcni .. ing : The need ror a theory or \ I,unl rhetoric. jOlfmal (II ("01llWIU'I" He.Il'an·". 2 I, 252-273. So ley. L. C. (1986). Copy length ami indtl,trial ad\eni<;ing rcadcr,hip. /I/chwrilll M(/r/..e'rin.s.: M(/l/lI~('l1Iel1f,
15. 14.')--151. Steffian. G. ( 1999). Correction of normati .. c nmpcrceptilln .. : An alcohol anu<;c pn.!lcmion rmgrillll, jouI"' I/(il (~I f)rug Edunllioll. 29. 115 138. Wl!inslcin. N. 0 .. Rothman. A. J .. & Sulton, S. R. (1998). Stage thenrlt.: .. of heallh bcha\iur' Conccptual and methodological is .. tlc, . IIl'allh Pf.\'(·/wlog.". 17. ;:!t)O ·199. "ei<;sc. C'. S .. Turhia"7. A. A .. & Whitney. D. J. ( 1995). Behav ior,ll traimng and A IDS fI\~ reduction' Olcrcol11ing barrier .. to condom tI,e. AIf)S Edllct/lio" lIml PI"{'I'f'llIirlll. 7.50 ·59. Wood. W .. Kallgren. C. A .. & l)rci~lcr. R. M. ( 1(85). Ar.:ce .... to :IUilude·rclc\ ant information in memory a .. a dc.:terminant of per,ua .. iun: The ruh: or me .... agc altrihutc, jOllmal (~f E\l'l'I"imt'l//lIl Sm'lal Pn cfwlog\', '!t. 73·85.
4 Theorizing about Persuasion Cornerstones of Persuasion Research Robert H. Gass and John S. Seiter
Persuasion theories arc a lot like n o~es; eve rybody has one. some big. some !-'l11all. a nd no two exactly alik e. This c hapl er rev iews and evaluates some of th e be tt er-kn ow n theori es and models o f persuasio n, soc ia l inlluc nce. and co mp liance ga in ing. The goa l j.., to prm ide yo u with a ge neral framework for unde rstand ing these theories and models as they re late 10 the chapters that fo llow. That said. not a ll theori es or mode ls are re prese nted here. This rev iew is. of necessity. selec tive rath er than ex ha ustive. There are simpl y too ma ny th eories. espec iall y smalle r, spec ia lized lheori c!o. rela ted to specific COIHcx l s or strategies 10 address the m a ll. This does no t mean we rega rd such theories as unimport ant. We co nsider th e th cOI' i e~ covered here to be Ma ples of pe r~ u asi o n resea rch. They have bee n ~tudied ex ten..,ively and are wide ly c ited in th e sc holarly literature. They ha ve heuri ... ti c va lue in th at they've sparked further th eori zing and resea rch on pe r s u a~io n . We wo uld the refore incl ud e th em in any " must know" list for stude nts of pe rsuasion. We should also note th at th e e mph as is here is on "social sc ie ntific" th eo ri e~ and m od e l ~ o f persuasion, rath er th an "rhetori cal" th eorie~ of pc r\ u a~ i on or ot her ~y m bo li c perspec ti ves o f how people influe nce each oth er. Our aim is not to devalue such th eori es by e xc ludin g the m. but simpl y to ack nowledge thm prac tical constrai nts must gove rn any e ffort of thi s sort. W ith the foregoing ca veats in mind, perh aps the best place to beg in i ... by stat ing that th ere is not so much (l theory of persuasion. as there are theories of persuasio n. Pe rsuasion research is based upon a patchwork quilt of perspecti ves, some that can be c harac terized as " umbre lla" theories accountin g for a wide varie ty o f persuasive phe nome na and ..,ome re prese nt ing "micro" theories that ex plain spec ific fOfm s of influence in ve ry limited ci rcumstances. Be rger and Burgoo n's observation that " By th e late 1960s O, tI"UI11 (1968) counled some 34 differe nt pers ua ~ i o n theories, none of whi ch co uld provide a plausible ex planation for the broad ra nge of ob ... erved co mmuni cati on-pe rsua..,ion re l a t io n ~ hip "'''
4S
46
Pan I • Preliminaries: Definitions, Trefld~;, and Theoretical Underpinnings
(1995. p. x) is telling in thi s regard. This situation has changed little. If anything. the expansion of traditional, one-lo-many persuasion studies 10 encompass interpersonal or one-on-one influence attempts has served to increase the number of theories und er conside rati on. While some theories have generated more research interest or found more favor within certain camps, at present no si ngle, unifying perspective can be identified in the literature, at least not one upon which all scholars and researchers agree. What's more, at least so me persuasion research, including so me studies in the area of nonverbal influence and some in the area of compliance gaining, appears to be largely atheoretical in nature. That is, either there is no clearly defined theoretical framework that guides the research, or a guiding theory, such as it may be, is implicit at best. One should not be overly bothered by this state of affairs, however. A cert ain amount of overlap among theories is to be expected, along with gaps and occasional contradictions. Human communication is complex. There are a lot of variables to consider in the persuasion equation. It should come as no surpri se. then, that researchers ha ve found it difficult to explain how people come to be persuaded all in one go. Theories are always provisional. They are subject to revision, modifi cation , re finement , and extension. Empirical findings may increase social scientists' confidence in a given theory, but it remains just that, a theory . And even if one could embrace a si ngle theory of persuasion today. it wouldn't be the same theory 10 years from now. In short, we are a long way from proclaiming anything like a "unified" theory of persuasion. All told, eight theories andlor models are examined in thi s chapter. They include attitude change processes, Mere Exposure Theory (ME), and a variety of related theori es grouped under the heading "Consistency Theories," such as Cognitive Dissona nce Theory (COT). We also examine the Theory of Reasoned Action (TRA), Social Judgment Theory, Expectancy Violations Theory , the Elaboration Likelihood Model (ELM). and th e Heuri stic-Systematic Model (HSM). We begin by examining what is not so much a specific theory as it is a mosaic of theories and research on attitudes and attitude change.
Attitudes and Attitude Change Processes Attitudes are to persuasion research what Elvis Presley is to rock-and-roll. Modern persuasion research is rooted in the study of altitudes, work that originated in the I940s under the stewardship of Carl Hovland, founder of the Yale Attitude Research Program . Attitude research continues to flourish today. A more thorough discussion of recent literature on attitude change processes can be found in se veral excellent reviews (Eagly & Chaiken. 1998; Petty, Wegener, & Fabrigar. 1997; Wood. 2(00). There is now general agreement that an attitude is "a learned predisposition to respond in a consistently favorable or unfavorable manner with respect to a given object" (Fishbein & Ajzen, 1975, p. 6). That is, attitudes are learned rather than innate, they reflect tendencies to respond or react in predictable ways, and they represent favorable or unfavorable evaluations of things. It is this last feature of attitudes, their evaluative nature (that is, appraisals of things as good or bad, right or wrong). that represents the hallmark of attitude research (see Dillard, 1993).
Chapter 4 • From the standpoint of per~ua~ion,
allitude~
Thenrbllg aiJolII Perll/lisioll
47
are important because they are thought
to corre late with and predict behavior. Just how well attitudes correlate with behavior.
known as the '"A-B relationship," has been the subject of intense inveMigation, Earlier stud ies were pessimistic about the extent to which attitudes predicted behavior (Dillehay. 1973; Wicker. 1969), At least some of the incon~istencies. however. were due 10 methodological shortcomings in researchers' measurement~ of attitude~ and behaviors (Ajzen & Fishbein. 1977; Kelman, 1974), Recent meta-ana lyses suggest that the correlati on between attitudes and behavior ranges from moderate (r ;;;; ,30) (Kralls~. 1995) to strong (r;;;; .79), if methodological artifacts are exclu ded (Kim & Hunter. 1993a). to stronger still (r; ,87). once the moderating role of receiver involvement i . . taken into account (Kim & Hunter. 1993b). Regardle~s of the exact correlation involved. the importam point is that altitudes do seem (0 predict behavior reasonably well. Another important poinl is that because attitudes and behavior are related, changing;:111 individual's allitude(s) should lead to changes in her/his behavior(s), Thus. attitude change research. in a nutshell. attempts to identify ways of modifying receivers' altitudes in order to bring about corresponding changes in their behavior, A good deal of research on attitude change has focused on the ro le of moderating variables that mediate the strength of the A-B relationship. A number of ~uch moderating variables have been identified to date. While we do not have sufficient space to devote to a ll of them here. we highlight some of the key moderators in table 4.1. Attitudes remain central to the study of persuasion. although, as Daniel 0' Keefe noted in chapter 3. researchers have moved beyond the attitude con~tnlct to explain a variety of othe r persuasive phenomena, SlUdie ... focusing on compliance gaining. for example. are primarily concerned with behavioral conformity, with or wilhout any accompany ing change in attitudes. Because attitude change research has occupied researchers' attention for more tha n 50 years. th ere isn't sufficient space here to examine all the ways in which atti tudes can be modified. However. one well-known way is through what is called Mere Exposure Theory, which we discuss next.
Mere Exposure Theory Mere Exposu re Theory. also known as the Mere Exposure Effect. ... tales thai repeated exposure 10 an unfamiliar stimulus can in and of itself increase po~itive affect toward the stimulus (Zajonc, 1968), Stated simply, some messages "grow on liS," Thus. a consumer who e ncountered a product logo on several different occasions would tend to evaluate that logo more favorably than he or she wou ld other. unfamiliar logo~, It wouldn't matter whether the consumer knew what Ihe familiar logo represented. The reason the theory is called "mere" exposure is that. unlike Clas\ical Conditioning. repeated exposure to a stimu lus produces increased liking in the absence of any reinforcement. You may recall that. according to Classical Conditioning (think Pavlov's dog here), a n unconditioned stimu lus (food) will produce an unconditioned response (salivation), ThaI is. Pavlov's dog naturally drools at the sight of food. When an unconditioned
48
Part I • Prdimillwin. Ih1/lIiIlO1/\, 71"£'/1(/.\, lIlItl /Jlt'orl'lint! l 'lU/erpillllillg.\
Moderating
TABLE 4.1
rariable,~
Affecti"g the A-/J Relati(Jllship
Atlilluit-HlIi('//('(' Of c£'IIIFo/in", Attitude" that are ~enlral 10 onc· . . core belier", and values arc
Inore Iikel)
s4uare \\ 11h behavior than <.lltilUdc ... that arc morc marginal Of tangential. llfliwdc(,\J ullel ht'llllrior( .\ ): Attitude ... arc more likely to corre ... pond with behm ior when 'pccific altitude ... and ",pcl..'ific behavior.. arc involved.
SpnUkitr
10
(~rllle
\ui/uti!',\ hawd 011 din'u ni1t'1"if'IH'l': AttitUlk ... formt!d \-ia per...onal experience; corrc!o.pond Inore do,cly to hcha\inr than altitude'! that afe formed ~ccondhand (Fa7io. 1986: Fa/io & Zanna. 19H1) So('i,,/ dl',;rohifiIY him. Pcople lend 10 hehave in ways Ihey consider to be socially polite or
corn:ct. especially in public . . clIing:-- (Furnham. 1(86). For this reason. ~omc rC~l!an;her!-. have 'ld\'OG.ilcd the u,c of' unohtru!-.Ive mea ... ure ... (Andcr!-.cll. 1(89) and indireci que ... tiolllllg (Fi~hcr. 19(1) a ... a mcam. of reducing . . ocial dc!-.irabilit} bia .... SeU~"/(milorilig. The A B relatiom-hip i!-. !-.tronger for lo'A self·monitor~ than high ,elfmonitor:--. hCl'au,t: the laller an: more ··ch'lllldconlikc" and more inclined 10 tailor their beha\,· 101' to spel'liic Slluatlon . . (Snyder. 1974. 1(79). ,\Ui\'(lfiOI/ of rl'i('l'lJlIllIllilluln , Attitude ... tend to prcdkt behavior more accufiltcly when they 'Ire ;.H,.'li\<.lled. thai i.... hroughl 10 the forefront oj an 1n<.Iividual·s con'Sciou ... :)warene!-is. Some· tllm!s peopk' need to he reminded ",hat thcir altitude ... are in order for them to adju ... , their be· ha\ior ilccordingl~ Lastly. the A- B reiation ... llIp "" likely to he strengthened when "multiple act criteria" are em· ploycd. Oi\ ing people more than one 0PP0l1unily. 10 11li.lIlil'e . . 1 their atlilude . . through their beha\ior imprmes the fit hc!\\een thl' 1\\0.
stimulu ... (fond) i... rcpc:ltcdly paired with a conditio ned stimu lu s (be ll). the conditioned qimulu ... (hell) \\-'ill c\'cntually clicit the conditioned response (salivation) all by itself. Thu .... hy ringing a bcll e\er)" timc the dog ... cc:o. food. Pavlov eventuall y gets hi s dog to drool at the ... ound of the hell alnne. The difference het\\-:een Mere Expo~ure Theory and Cla ...... lcal Conditioning. thcn. I'" thai mcre cxpo:o.urc doe ... n·t requirc the presence of any additional rClnforccmclll 10 \\ork lis magic. !\1ere r::\posurc Theor} (ME) po~tu l ates that repeated expo ... ure to a ... tilllulu ... (hell) \\ilJ n.~"'lllt in more favorable elvalu<.llion ... of that stimulu ... (han of other unfamilHlr stimuli (wtlJ"'tlc. bUller. !.:hime, etc.). As an illustration of ME in a labof
Chapler 4 • Theorizing ahoUl Persuasion
49
Moreland & Zajonc, 1977; Zajonc, 200 I: Zajonc & Rajecki , 1969). Moreover, the phenomenon has been demonstrated using a wide variety of stimuli, including general knowledge statements, works of art, yearbook photos. musical compositions, product name~ , nonsense words, and geometric figures (see Bornstein, 1989: Harrison, 1977), ME has also been demonstrated to work across different cultures. There are differing explanations as to why or how ME works. Some scholars believe that cognitive processing. or mediation, is involved. Others assume that ME is a more automatic, unconscious process. One of the most common cognitive explanations involves learning. With each additional exposure, a person acquires additional information about a stimulus. The additional information enhances the person's appraisal of the stimulus. However, the assumption that the more you learn about something. the more you'll like it is not without limitations. Some studies have shown that ME decreases liking if the initial >timulus is evaluated negatively (Amir, 1969: Brickman, Redfield, Harrison, & Crandall, 1972: Perlman & Oskamp, 1971), Another common explanation is based on a misattribution involving "fluency," According to this view. individuals mistake "fluency:' or the proficiency with which they process a ;timulus, with positive affect or liking (Jacoby & Kelley, 1990). The nuency is a result of their previous exposure to the stimulus. but observers don't realize this. This explanation accounts for the fact that in slUdies in which participants are aware of their previous exposure, reduced ME effects are observed (Bornstein, 1989). There is also some evidence that fluency has a gremer effect on cognitive judgments compared with affective judgments (Lee, 200 I), Other explanations sllggest that ME takes place at a low level of awareneS!l or even unconsciously (Moreland & Zajonc. 1977). One sllch explanation involves familiarity, According to this view, familiarity tends to enhance liking. A familiar stimulus is perceived as more inviting or attractive than a novel or unfamiliar stimulus. Based on this explanation. if a voter saw a candidate's name repeatedly on bumper stickers and lawn placards. that candidate would enjoy an advantage over unfamiliar opponents in that voter's mind on election day. In support of the unconscious-processing explanation, Bornstein's (1989) meta-analysis revealed that when ME takes place without conscious awareness, it tends to be more effective than when it takes place with conscious awareness. A recent study offers some intriguing evidence that different types of ME can take place through different brain hemispheres (Compton, Williamson, Murphy, & Heller. 2002). Words or text may be processed in one hemisphere. images in the other. This finding may explain some of the inconsistencies in previous studies. In addition. complex stimuli appear to be evaluated more favorably with increasing exposures than are simple >timuli (Bornstein, Kale, & Cornell, 1990; Heyduk. 1975; Zajonc et aI., 1972). Research also indicates that ME may facilitate a preference for familiar over unfamiliar brands, but not over equally familiar, or more familiar, brands (Baker, 1999), Other studies suggest that ME works best if the exposures are brief in duration (Hamid, 1973). Finally, some research indicates that there are diminishing returns to increasing exposure, with a leveling off or drop-off in effectiveness after 10 to 20 exposure< (Stang & O'Connell. 1974: Zajonc et aI., 1972), In the real world, it is difficult for persuaders, such as advertisers, to control the number oftimes consumers are exposed to a message. A commercial might air
SO
Part I • Preliminaries: Definitions. Trellds. all(/ Theoretical UndeI1)illllillg.\
100 times. but so me viewe rs may see the co mmercia l only a few lim es or no t at all. whereas other viewe rs may see it dozens of times. At prese nt. it is unclea r whether ME operates v ia co nscio us or uncon sc ious processing, or both. The litcrature seem s to indi cate that ME is more effective when it take s place at a low le ve l of awareness or un consc iou sly. Whatever the underlying mechani s m, however, the lit erature sugges ts that M E works and works we ll. Although questions remain as to exactly how and why M E works. it is a relatively s impl e theory of pe rsuas ion: Re peated, unreinforced exposu re to a stimulus facilitates likin g for th e stimulu s. even in the absence of awareness.
Psychological Consistency Theories A variety of theories fall under the rubric of what have come to be known as "consistency theories:' These include Heider s Balance Theory ( 1958). Newcomb's Symmetry Theory (195 3), and Osgood and Tannenbaum's Co ng ruit y Theory ( 1955 ; Osgood, Tannenbaum, & Suci, 1957). among others. These theories share th e co mmo n assumption that individuals have an innate desire to hold consiste nt beliefs, attitudes. and behaviors. Holding di sparate belief;, atLitudes, or behaviors is thou ght to be psyc holog ically uncomfortabl e. A person who had trouble "looking him se lf in the mirro r" wou ld be experi encin g psychologi ca l inco nsis tency. as would a perso n who fc lt !\he W3!\ "between a rock and a hard place" insofar as a n importa nt dec is io n was co ncerned. Thi!\ psychological tensio n mo ti vates indi viduals to adjust their thou ght s, fee lin gs. or actions accordingly. As an example, a hea lth-orie nted person might be bothered by the fa ct that he or she eats red mea l because red meat is associated w ith increased cho les terol levels. Per~uasivc messages can be desig ned e ither to create or res tore consistency by bringing such incons istent cog nition s into line. Using the sa me example. a pe rsuas ive message cou ld differentiate between eating red meat occasionall y, as opposed to frequently, and eating lea n red meat rather th an meat that is high in fat. Such dif/erellliarioll is one of the p~ychological mechan is ms for bringing about co ns istency . Other co mmon me chani s ms for preserving or restoring consiste ncy in clude denial (" I don't believe th ere is a we ll -esta blished link between red meat and cho lesterol o r cholesterol and heart disease"). bolSTering (", I don ' t eat that much meat anyway: red meat is a good source of protein"). altitude or helief modificatioll ("I think eating red meat is okay as lon g as I ha ve small portion!\"). and transcendence (Hey, we've a ll got to die of so mething"). Consistency was originally viewed as a dri ve·reduct ion theory. That is, inconsistency creat es a psyc hologica l dri ve to maintain or re store consiste ncy. More recent thinking. however. suggests that consiste ncy is as much an e ffort to manage one's se lf image or maintain face in th e eyes o f others as it i~ an internal drive state (Aronson, Cohen, & Nail. 1999; Greenwald & Roni s, 1978; Scher & Cooper. 1989). We may know we are not bein g consistent , but we wan t others to think we are. How much inconsis tency an indi v idual ctl n tolerate is related in large part to the cefltralily of the beliefs, attitude s, or behaviors in question . Incons istenc ies involvi ng core beliefs are more tro ubling than those in vo lvin g tangential beliefs. Thus, the notion of cheatin g on a diet wou ld both er most peopl e far less than the nOli on of cheat ing o n a spouse. C urre nt rese<.l rch also suggests that there are indi-
I Chapter 4 • 71leori:illg a/una Persullsiol/
51
vidual differences in people'\ tolerance for inconsistency. An inconsislency Ihal "bugs" one persall might nol faze another. Research furlhcr suggests Ihal there is a cultural COI11poncnl involved in the degree 10 which people "rive for con~i~lency (Cialdini. Wosinska. Barret. Butner. & Gornik-Durose. 19991. All of Ihe consislency theories sutTer from Ihe drawback of being unable to accommodate more than three cognitive relations at one time. If we know, for example. that Timmy like~ Popcye. the cartoon character (favorable altitude), and Popeye is positively as\ociated with eating spinach (favorable attitude). then Timmy's attitude, will be psychologically cOI1,i\tent if he too like\ "ipinaeh. Yet such a model Lends to oversimplify the complex associative networks in which beliefs. altitude\. and other cognitions exist. Timmy might like Popeye and. therefore. spinach. but his best friend, Emile. might hale spinach: or Timmy might nol like the way his mother serves spinach (Popeyc. after all, gulps it straight from the can): or he might like other vegetables more than spinach: or he mighL identify with Wimpy. the character who loves hamburgers. more than he identifies with Popeye. Attitudes don't exist in isolation. but rather in clusters of beliefs. altitudes. and values. The.",c clustcrs arc in turn interrelalcd with other clusters. Even consistency theories that lake into account degrees of attitude and not just their valences, such as Congruity Theory. suiTer from thi~ limiwtion. COIl~i"itency theories are useful as far as they go, but (heir explanatory and predictive power i" Iypically confined 10 one Iriad al a time.
Cogl/itive Dissol/al/ce Theory A specialized ver:-.ion of consi"tency. known as Cognitive Dissonance Theory (COT). was developed by Leon Fe\linger (1957. 1964: Fe\linger & Carlsmith, 1959). The theory. which fell out of fa\or in the 1970s and 1980s. has come roaring back since the 1990s. Nearly 100 artiele"i have been published on cognitivc dissonuncc in the last decade. While FC'ilinger's original theory has been modified and extended by others (see Cooper & Fa!.io. 1984: Scher & Cooper. 1989). many scholars argue that the original theory remains viable today (Beauvois & Joule. 1999: Harmon-Jones. 1999: Mills. 1999). Similar to other consistency 'hcoric\. CDT postulates that holding di~sonan( cognitions (beliefs. attitudes. perceptions. eLc.) is an aver~ive p"ychological .state. Cognitive di!o.sonance isn't an all-ornothing phenomenon- it occurs in varying degrees. An important decision evokes more di"sonance than an unimportant one. Persuasive messages can be aimcd at either increasing or decreasing dissonance. On the one hand. a persuader might wanllO increase dissonance in order to get another person to rethink his or her position on an i!o.slle. A parent might tell a college-age son, for example. "Are you ",ure you want to get a new car. rather than a used one? Your car payments will be highcr and you'll have to pay more for insurance too." On the other hand. a per ... uader might want 10 minimi7e dis.sonance by reassuring another person that the decision ... he or he made was (he right one. A parent might tell a college-age daughter, for inslallce. "You made the right choice in buying a car. rather than leasing one. If you take care of it. you'll havc reliable transporlation long after it is paid off." Four common research paradigms have been llsed to study cognitive dissonance: the free-choice paradigm. the bclief-di"iconfirmation paradigm. the effort-justification para· digm. and the induced-compliance paradigm (Harmon-Jone, & Mills. 1999). The free-
52
Part I • Prelimillaries: D({illiTiom, Trends, and Theoretical Underpinlling.1
choice paradi{f11l focuses on the psychological angst a person experiences following a freely made decisioll. For this reason. COT is often referred to as a "post-decision theory." Once a person makes a decision, the person worries about whether she or he made the right choice. This phenomenon is commonly referred to in sales as "buyer's remorse:' The person seeks to reduce her or his dissonance by justifying the decision that was made. Allcmpts to justify or reinforce the dccision can take plaec through the individual's thought processes. words, or actions. For a more detailed discussion of some of the modes of dissonance reduction. see Steele (1988). Stone. Wygand. Cooper. and Aronson (1997). and Burris. Harmon-Jones. and Tarpley (1997). A second paradigm for research on COT involves belief disconjirmation. Dissonance is arou~ed when an individual is exposed to information that is inconsistent with her or his belicfs. The theory predicts that a person will reject. distort, or avoid information that arouses dissonance. When there is a bear market. for instance. some investors can't stand opening their quarterly statements, much less reading them, because they don't want to know how much money they've lost. When there is a bull market. the same investors can't wait to tear open the maillO see how much money they've made. CDT also predicts that. having made a dccision. individuals will engage in selectil'e exposure by seeking alit information that is consonant with their choice and avoiding information that is dissonant with their choicc. Before buying a car. for example, a consumer might look at ads for a variety of makes and models. After purchasing a car. however. the same consumer will tend to look for ads or favorable reviews of the specific car purchased. Nevertheless. people don't always stick their heads in the sand when confronted with dis.'.onanl informatioll. Empirical studie!-. suggest that people can and do tolerate a certain amount of disso· nant information (see COllon. 1985: COllon & Heiser. 1980). A third avenue of research for studying CDT is known as the induced compliallce paradiXI1l (originally called the forced-choice paradigm). This is perhaps the most widely studied of all the dissonance paradigms (Devine, Tauer. Barron, Elliot. & Vance. 1999). When a person is induced to engage in behavior that is contrary to his or her attitudes or self-image. the magnitude of dissonance is less. When a person performs a counterattitudinal action of his or her own volition. however, the magnitude of dissonance is greater. By way of illustration, an attorney who was paid handsomely for defending an unsavory client would find his or her actions easier to justify than if he or she took on the case pro bono. The greater the external incentives. whether positive (promises of reward) or negative (threats of punishment). the easier it is for an individual to ascribe his or her behavior to the external inducement. Research based on this paradigm has also revealed a negative illcefltil'e effect (Fe!-.tinger & Carlsmith, 1959: Harmon-Jones & Mills, 1999). The larger the external incentive for engaging in counterattitudinal behavior, the less attitude change there will be in the direction of the behavior in question. The smaller the external incentive, the greater the change in attitude, Thus. people who volunteered to go door [Q door to raise money for a charity would tend to have more favorable attitudes toward the charity than people who were paid to do the same type of fund-raising, The fourth paradigm that has prompted COT research is the effort justification para· digm (see Beauvais & Joule, 1999). The basic notion here is that when a per!-.on has to earn ~oJllelhing, he or she appreciates it all the more. The greater the sacrifice thal i~ required 10 achieve an outcome, the more an individual will value the outcome. Conver~cly,
Chaptl!r 4 • Thl'ori::illg ahollf Permas;OI/
53
the less sacrilice involved. the le% value the individual will attach (Q the outcome. By way of illustration, a college student who pledged" fraternity or sorority and wa~ required to perform embarrassing or humiliating acts during "hell week" would tend to rationali/e the behavior by valuing the outcome of membership in the fraternity or sorority even morc. Why? Because performing unpleasant acts would arouse dissonance. but the dissonance could be reduced by enhancing the perceived value of the outcome. A college student who did not have to undergo embarrassing or humiliming initiation rituals. however. would not experience as much dissonance
Theory of Reasolled Actioll The Theory of Re",oned Action (TRA) was developed by Martin Fishbein and "aac Ajlen in the late 1970s (AjLen & Fishbein. 1977. 19RO: Fishbein & Aj7en. 1975) and was followed by AjLen's Theory of Planned Behavior (1991). The TRA is often referred 10 as a "rational" thcory of persuasion. bl!cau~c it fm.: use!-! 011 the deliberative process all individual engages in when shc or he is prescntcd with a persuasive message. The opcrating assumption is that individuals systematically analY7e messages. evaluate all availablc information. and actively weigh the benefits and rbks associated with compliance beforc making a decision. The linchpin of the TRA is an individual's /Je/lt/viortll illlelIlioll. That is. the most reliable indicator of what a person will do is what he or shc intends to do. Although behavioral intcillions don't always correspond (0 behavior- for example. the time a pcrson in tends to wake lip when she or he sets an alarm clock at night is not necessarily (he actual time the person gets up in the morning- a number of studies have confirmed that behavioral imcl1tions are, by and large. predictive of beha\ ior (Ajzen & Fishbein. 1973: Kim & Hunter. 1993a. 1993b: Sheppard. Hartwick. & Wa"haw. 1988). Behavioral intentions arc in turn guided by Iwo major f"ctors: a per...on' ... attitude toward the behavior and \ubjcclivc flonns. The first major factor. afliwde IowaI'd fhe be/wl'ior, refers to a person's evaluation of the bencfits and ri . . ks associated with performing the action requested in a persua . . ive message. Favorable attitudes lead to approach behavior. and unfavorable attitudes Icad to avoidance behavior. For exam pic. the Jikclihood that Naomi would join a carpool to commute to work would be based on her intention to join a carpool. Her intention would in turn be based on her attitudes toward carpooling, If Naomi held favorable attitudes toward carpooling ("'I get to work fa"tcr." '"I'm helping the environment'") then her intention to join a carpool would be stronger. If ... he held unfavorable attitudcs toward carpooling ("I can't comc and go as I please:' "I don't enjoy talking to people on the way to and rrom work"). her intention to join a carpool would he weah.er. The combination of po,itivc and
54
Pan I • Prt'limirwries: Dejillitio!l\', Treml\, (J1Il1 Tltl'oretU'(J/ U",lerp;l/l/i"KS
negative attitudel., toward the behavior i:-. one factor that determine ... a person's behavioral intent. The TRA also sl<.Ites that an individual's altitude toward the behavior. in this case carpooling. will be based on his or her beliefs ahow the outcome of performing the behavior and his or her el'olu{If;OIl of the olllcome. The more a person believes thaI performing an action will produce a favorable outcome. and the more favorably the person evaluates that outcomt!. the stronger the person's allitudcl., toward the behavior will be. Continuing with the same example. Naomi might believe that one outcome of carpooling is that she would save money on gasoline. Her eva lu ation of that ou tcome wou ld be positive. But she might also believe thut another ou tcome of carpooli ng i:-, that :-.he would be unable to leave work on lunch brl!ah to run errands. Her eva lu ation of that ou tcome would be nega tive. According (0 the TRA. the tota lity of Naomi·1., bclicl\ about these ou tcomes and her evaluation of them would guide her attitude toward joining a carpool. A second major factor that determines a person's behavioral intent is subjecfil'e norm.\. SUhjecti\'e norms are made up of a person· ... 1l0rll/{lfi\'e beliefs about what significant other... think. along with the pcrl.,on· ... mmil'Cllioll to co",,,ly with significant o th ers' opinion .... Thul.,. If Naomi·!o. friends were cll\ironmclllali\h. I.,he would tend to have favorable beJiet\ toward carpooling. If Naomi al ... o had a strong desire to live up to her friends' normative expectations, she would be even more inclincd to join a carpool. The various components of the TRA- beliefs about the outcome. cvaluation of the outcome. normative beliefs. and motivation to comply- are Iypically measured usi ng self-report rating ... ca les. The TRA has been tested on a wide variety of topics and issues. including AIDS risk reduction (Cochran. Mays. Ciaretla. Caruso. & Mallon. 1992). belief in extraterrest ri a ls (Patry & Pellctier. 2(01). condom u'e (Albarracin. Jo hn '''n. Fi'hbein. & Muellerieile. 200 I: Greene. Hale. & Rubin. 1997). dental hygiene (Tonea llo & Binik. 1987), drinking and driving (Gaslil. 2(00). expo\urc to the sun (Steen. Peay. & Owen, 2<X>O). mental practice (Trafimow & Miller. 1996). moral behavior in sports (VaJlera nd. Deshaies. Currier. Pelletier. & Mongeau. 1992). recycling (Park, Levine. & Sharkey, 1998), and voti ng beha\ior (Granberg & Holmberg. 1990). The TRA ha, been e'peciall} usefu l in predicting the role of intention!o. on health-related behavior. In addition .... tudies have show n that there me individual differences. gender differences. and cultural differences in the weight or importancc people attach to the \-ariou ... components of thc TRA (Greene. Ha le. & Rubin. 1997: Godin et al.. 1996: Lee & Green. 1991).
Social Judgment Theory Social Judgment Theory. pre,ented by MUlafer Sherif. Ca ro lyn Sherif, and Robert Nebergall (Sherif & Sherif. 1967: Sherif. Sherif. & Neberga ll. 1965), focuses on how people evaluate per~uasive me!o.sagc ... and how ~uch evalua ti ons affect whether or not persuasion occur'i. According to the theory. on any given topic a person might hold a range of possible positions. By way of cxample. con~ider the debate among vegetarians. meat eat-
Chapter 4 • Theori::Jng abollt Persllasion
SS
ers, and some of those who are "in between." Here are several positions. so me extreme, some moderate, that a person might embrace on this issue (see Corliss, 2002): J. Sproutarianism- You should build your diet around beans, wheal, and other sprouts.
2. Fruitarianism- You should eat plant parts that the plant can easily replace (e.g .. berries, apples, tomatoes. grains. seeds). 3. Veganisl1l- You should eat plants and avoid meat, dairy, eggs. honey, or any other animal product. 4. Ovo-vegetariani sm- You may eat plants and eggs, since hens would lay the eggs even if we didn't eat them, S. Ovo-lacto-vegetarianism- YOli may eat vegetables. eggs. and dairy products. since doing so kill s no animals. 6. Pesco-vegetarianism- Yoll may cat fish because fi sh don ' t have sophisticated nervous systems. 7. Pollo-vegetarianism- You may eat chi cken. but not red meat. 8. Meat caters- You may pm just about anything in your mouth . Social Judgment Theory argues that on this, or any continuum. a perso n has a most preferred position, called an anchor point. This anchor point functions as a reference point or "psychological benchmark" against which other posi tions and viewpoints are evaluated. If, for example. Dean agrees most with position 2- that you should eat replaceable plant parts-position 2 is Dean' s anchor point, but that doesn't mean that it is the only position he may find acceptable. For instance. though preferring to stick to position 2, he may think it is all right to eat honey, eggs, and ice cream (positions 3. 4. and 5), Such positions. together with his anchor point, represe nt Dean' s latirude of acceptance. that is. the range of positions he finds acceptable, But what about the remaining positions on the continuum. which fall olltside a perso n' s latitude of acceptance? According to Social Judgment Theory, a person may fee l ambivalent about so me of these positions and strongly opposed to others. Those positions in the first category (a mbivalent ) are said to fall within a person's latitude of I/ol/commitment. Those in the second category (strongly opposed) are said to fall within a person' s latitude of rejection. Thus. if Dean feels neutral about eating fish, position 6 fall s within his latitude of noncommitment. If he is adamantly against eating chicken and beef, positions 7 and 8 fall within hi s latitude oIrejecrion, How is this important lO persuasion ? First, Social Judgment Theory suggests that a person's anchor position is used as a standard to evaluate all other positions, As such, it may be difficult, if not impossible. to persuade a person to accept a position that is too disparate from his or her anchor point. In fact. when a persuasive message advocates a position that is highly di sc repant from a perso n's anchor posi tion. Social Judgment Theory predicts that the persuadee will perceive the position advocated in the persuasive message to be farther away from the anchor (han it really is. This outcome, known as the contrast effect, makes rejection of a persuasive message more likel y, In contrast, a persuasive message advocating a position that is not too far away from a person's anchor position , that is, one that falls within the perso n's latitude of noncommitment , may be deemed tolerable. In this case. Social Judgment Theory sugges ts that the persuad ee may end up perceiving the
56
Part I • PreJimil/orie.\: Definitiolls. Trellds. and Theoretical
U1/{lerpil/llif/~s
advocated position to be closer to the anchor than it really is. This phenomenon, known as the assimilation effect. makes acceptance of a persuasive message more likely. Not surprisingly. the breadth of any particular person's latitudes of acceptance. noncommitment, and rejection influences how difficult it may be to persuade that person. For instance, Social Judgment Theory conceptualizes an ego-involved person as someone with a narrow latitude of acceptance and a wide latitude of rejection. In our example. a person who makes a living selling cattle and who loves to eat beef with every meal would be such an ego-involved person. Social Judgment Theory suggests that trying to persuade such a person to become a !otproutarian would probably be a waste of time. On the other hand, just because a person is ego-involved in one issue does not mean he or she will be that way on all issues. The cattle rancher. for example. may have a narrow latitude of rejection on issues such as gun control or capital punishment.
Expectancy Violatiolls Theory A number of persuasion theorie!ot focus on expectations, or, more precisely, what occurs when people's expectations are violated. Three such theories-Language Expectancy Theory (M. Burgoon. 1995). the Nonverbal Expectancy Violations Model (1. Burgoon, 1994), and Reinforcement Expectancy Theory (Klingle, I996)-are discussed later in this volume (see chapters 9. 10. and 16) and will therefore be highlighted only briefly here. Although there are SOme unique differences among these theories, aJi share the COIllmon assumption that people have expectations about what constitutes normal behavior. When slich expectations are violated, it catches receivers off guard. Someone standing too close or using extremely intense language. for example. might violate a person's expectations for normal behavior. According to these theories. such violations cause receivers to shift their attention from the message to the source of the message. Whether such violations hinder or facilitate persuasion, then. depends on the receivers' perceptions or the person violating the expectations. For example. when perceptions are positive, that is, when sources are perceived as credible. attractive. or likable. sources may be more persuasive when violating expectations than when behaving in accordance with expectations. On the other hand. when perceptions are less positive, sources who violate expectations for normal behavior are probably less persuasive than they would be without such violations. These related theories also suggest that some sources have a relatively wide bandwidth of acceptable behaviors, whereas others are permitted a relatively narrow bandwidth. By way of example. Reinforccmcl1l Expectancy Theory argues that because of social norms, female physicians can't get away with using aversive strategies in the same way that male physicians can. When female doctors use such strategies, they violate patients' expectations about what is appropriate. and as a result, patients are less likely to comply. According to the theory, then, male physicians can increase compliance by using either positive (e.g., "Regular eating will make you feel so much better") or negative ("You have two choices: change your diet or spend the rest of your life wishing you had") influence strategies, but female physicians can increase compliance only by using positive ~trategies.
1 Chapter 4 • Theori:illg about Per.was;ol1
57
Dual-Process Theories of Persuasioll Two of the m05t recent theories of persuasion are known as dlla/~proces.'i theories (Chaiken & Trope. 1999) because they postulate thai persuasion operates via two basic paths. Both Peny and Cacioppo's Elaboration Likelihood Model and Chaiken and Eagly's Heuristic-Systematic Model maintain that people employ two qualitatively different modes of information processing when they are exposed to a persuasive message. Though the two theories differ in their particulars. both assume that one mode or route is more cognitive. deliberate, reflective. effonfu!' and generally slower than the other, which is more automatic. reflexive. habitual. affective. and generally faster. People rely on one mode when they need to think through a decision and rely on the other when they need to expedite their decision making. Other dual-process theories of persuasion also exist. For example. Kim Wine's Extended Parallel Process ing Model. which is the subject of chapter 13, presumes that fear appeal., are processed along one of two basic routes.
The Elaboratioll Likelihood Model Richard Peny and John Cacioppo's (1986a. 198Gb) Elaboration Likelihood Model of Persuasion (ELM) is one of the 1110st widely cited models in the persuasion literature. 8ecause the ELM is the focu s of chapter 5. our di sc ussion of it is truncated here. The ELM postulates that there are two basic routes to persuasion that operate in tandem. The first of these. known as the central roWe. involves cog nitive elaboration. that is. effortrul thought and deliberation about the con tent of a message. The second route. known as the peripheral rO/lIe. emphasizes mental shoncuts. ~ uch as a reliance on source attractiveness or argument quantity in evaluating a persuas ive message . According to the ELM. the predominam form of processing on which an individual reli es depends on a number of factors, including the individual's involvement with the topic or issue. ability to process the message. motivation to process the message. and need for cog nition . As an illustration of the ELM in action. suppose that Alex is interested in joining a health club. If hi s determination to gel in shape is high. meanin g that he has high personal involvement in the issue. he will tend to rely on central processing. Thm is. he'll tend to read different ads for health club..,. compare their features and prices. ask question s to c lariry points of information. and l11ull over the advantages and di sadvantages of different plans in hi s mind. If. however, hi!-t motivation to get fit is low, thaI is. his doctor encouraged him to exercise more regularly but Alex isn't panicularly eager to do so, he will tend to rely on peripheral processing. He might join a health club on the recommendation of a friend. or simply join the gym with the catchiest ad. He would tend to use central processing if he wanted to make a thou ghtful decision and peripheral process ing if he wanted to make an expedient deci sion. A so urce of controversy surrounding the ELM has been the iss ue of parallel or simultalleO/u processing. Although Petty and Cacioppo acknowledged the possibility of parallel process ing. that is. usi ng both routes at once (Pett y. Ka smer. Haugtvedl. & Cacioppo. 1987), the ELM has been criticized for not making thi s assumption more explicit (see Hamilton. Hunter. & Boste r. 1993: Mongeau & Stiff. 1993: Stiff. 1986; Stiff & Boster. 1987). While Peny and Cacioppo argued that the ELM docs not preclude the
58
Part I • Preliminaries: Definitions, Trend.\·. and Theoretical Ullderpillflifl!:.I'
pos~ibility
of parallel proces~ing. Ihey suggested thal there is usually a trade-off, in that a person tend~ to favor one route over the other. The controversy over parallel versus exclusive proces~ing may involve a "difference that makes no difference," however, because to date there is limited empirical evidence from laboratory studies that ~imu ltan eous processing ever occur> (Booth-Butterfie ld ct al.. 1994: Chaiken. Liberman. & Eagly. 1989 ). Petty & Cacioppo address this and other criticisms of the ELM in chapter 5.
The Heuristic-Systematic Model of Persuasion Another model of persuasion that bear~ many similarit ies to the ELM is Shelley Chaiken and Alice Eagly's Heuristic-Systematic Model (HSM) (Chaiken. 1980. 1987: Chaiken. Lieberman. & Eagly. 1989: Chen & Chaiken. 1999; Eagly & Cha iken. 1993). As with the ELM. the HSM operates on the a:-.:-.umption that individuals rely on two different modes of information processing. One mode. called .'1ystemotic processing, i:-. more thoughtful. deliberate, und analytical. This mode focuses on the content of the message and is roughly analogous to central processing in the ELM. The other mode. called heuristic processing. is more renexive or automat ic and is anal ogous to peripheral processing in the ELM. Heuristic processing is based on the application of what Chaiken and Eagly call decision rules or heuristic cues. such as mental shortcuts, which simplify information processing and decision making. A person who always tips 15 percent when dining ou t. regardless of the quality of the food service. would be employing a decision rule. Other examples of decision rules would be "size matters" when buying an SUV. "experts ca n be trusted" when evaluating a scientific study. or "never pay retai'" when shopping for jewelry. According to the HSM. decision rules are Mored in memory and activated under th e appropriate circumstances. Heuristic cues operatc ~imila rl y but tend to be based on appearance cues and subjective preferences. A consumer who selected one product or service over another because a celebrity endorser was more attractive. an advertising jingle was more catchy. or a commercial was more vivid than its competitors would be relying on heuristic cues. The activation of systematic proces~ing via decision rules and heuristic cues places fewer cognitive demands on the individual. Consistent with the ELM. the HSM !'.tates that a person's moth'arion and ability to process a message are both key detcrminants of whether a person wi ll rely on systematic or heuristic processing. or both. Chaiken and Eagly also maintained that individuals are '·economy-minded·' when deciding which mode of processing to use (Chaiken. 1980). Thus. the theory postulates the operation of a sufficiency principle. which states that people ~eek to strike a balance between not thinking enough aboll t a decision and thinking too much about it. According to the HSM . people attempt to expend as much menIal ene rgy as they need to, but no more (Chen & Chaiken, 1999). A difference between the ELM and HSM is that Chaiken and Eag ly' s HSM explicitly acknow ledges th e prospect of simulraneolls processilll: of messages. That is. messages travel the heuristic and systematic route~ concurrent ly. While the ELM doesn't rule thi s out. neither does it incorporate simultaneous processing as an explicit as~umption of th e theory. The HSM assumes that the two route~ can operate separately. in combination, or in opposition to one another. The additi\'it.\' hypothesis (Maheswaran. Mackie. & Chaiken. (992) states that systematic and heuristic processing can complement one another. For ex-
c------------------------------------------------------------------------------------------------ ,
Chapler -l. • TlU'ori:ill).: Clbolll Penumioll
59
ample. if a co n~ul1ler preferred (0 buy name brand product~ (heu ri sti c cue) and th en read an arti c le in
person who read a novel and admired the author's idea . . and U"ie of la nguage (syste mat ic proces!'>ing) might not care what the author looked like (hl.'!uristic cue) if she or he later saw the author on a television talk ... how. The ELM a nd HSM ha ve been shown to have both practical and heuristic vul ue insofar as th e ir ability to explain and predict people's reactions to persuasive messages is concerned. D07cns upon dozen.., of studics based on the!-c dual-process theories have been ca rri ed o ut on a vari ety of topics. receivers. and seltings. The resu ll ~ to date have ge nerally uphe ld both theori es' utility as comprehensive. integrative explanat i on~ of how persuasion functions. While the theories arc not withollt their critics (sec Kruglan ... ki. Thompson, & Spiegel. 1999: Mongeau & Stiff. 1993: Stiff & Boster. 1987 for criticisms). it is safe to say that they enjoy considerable support in the li terature.
Co netu sio II In this chapte r we have examined eight different theori\!s of per~uasion. social influence. and compliance gaini ng. Although ab"orbing eighllhcories in one fell swoop is enough to make anyone "theory weary," we want to st re~s the importance of acquiring a solid g rasp of basic theoretical frameworks when qudying persuasion. Good research, we believe. is o r should be theory driven. TheOl'ic!- inform research, and the results of empirical research in turn aid in ex tcndi ng. modifying, relining. and in somc cascs refuting theories. It is not enough 10 know th at a panicular ... tudy found a particular result. Theories and models help us to understand not o nl y H'}Ill! the results of a study were but al ... o why those re\ults were obtained. Thu\. {O fully under... tand persuasion, soc ial influence. and compli ance ga in ing. we be lieve it is importa nt that you learn not on ly about research finding~ but abo about their theoretical underpinnings. Even at e igh t theories, we have o nl y sc ratched the surface when it comes to the mul titude of theoretical explanations of persuasive phenomena. Anyonc's li ... 1 of "greates t baseball players" or "most imrortant lilms" will likely vary from another person's list. so we apologize in advance if we ...... tiffed .. your favoritc theory in our list. Some theories that are \:ipec ific to particular per\uasivc strategies (such as a "guilt-ba\:ied" explanation for the door-in-the-face ... trategy. o r Interpersonal Deception Theory as an explanation of deception detection) are covered in later chapte rs in this volume.
60
P
Hopefully. lhi\
iew of pcr,ua,ionlheorie\ ha:-. provided you with a ba..,ic fOllnda· lion from which (0 underl"llmd varioU\-, terms. conceph. principlc!oo. proCCI"..,cs. :-,trHtcgic .... and phenomena you will read aboul a~ YOLI study per~uasion. social innuencc. and compliance gaining. There arc those \vho would argue that "all a theory and a dollar will gel you j, a cup of coffee," meaning that theoric.., aren't worth much. We con..,ider theoric ... 10 be valuable. however. becau ... e it is not the finding'> of studies that lell U:o. how persua,iol1 worh. but rather what the finding ... mean or how they are interpreted in ligh t of prevailing (heoTie'" of per<.,ua'tion. fe\
Ref erellces _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __ AJI~J1. I. ( 19(1). The thcory of planned bell:1\ lOr. O/"glllll::lItiol/al lJel/(/l'ior (IIld Ihol/(/II Dl'('i,\io ll Proce.lwl.50.17lJ-:! II . Aj/en. I.. & Fi"hbein. :vi . (19HO). Unde'HulI/tIillg (llIilllt/e'I' (llId /Jred;CI;If.S~ .l"Ociallu-lllIl·ior: Allitlldi'l. 111ft"I/f/r"'.1 alld p('I"("('in'd lu'h((l'loral ("(ml/"(ll. Englc\Vood Cliff!">. NJ: Prentice-Hall. "J/('n, L. & Fi"hhein. M. (1977). Anitw.lc-bch'l\inr rdation,,: A thcorl!tical analy ... i" and rc\-ic\V oj emplril:ai n:"earch. Psychological/JIII/i'lill. 8.J. HH 91 X. A.l/en. L & Fi .. hbeLn. \ '1. ! 197J). Allituthnal .md nOrLllatilc ,ariahlc ... a ... predictor" ()f "f'CCLriC beh .. , I(1r". )ollmal (lr Pt'no/lalil,· ami Pwd/(l!o~, '. 27. -ll :;7. Alharracin. D.. John ...on. R. T.. Fj"hhcin. M.. & Mucllcrlclle. P. A. UOOI). Theoric .. of rt.\l ..om.'d actIOn amI cnmlom u ..agc: A met..l-analy .. i .... PndwlogiwIIJllllt'till. 127! I). 14:!-161. AIlIir. Y (1%9). Conlact hypothc .. l.. 111 cthnic relation". Pnl'h%giml Blilleril/, 71. J 19 ·J42. Andc .....en. J W. ( 1989). Unontru .. i\c me ..... urc ... In P. Emmert & L. L. Barker (E(h.). Mi'lI.\ //rl'lIIt'lII 0/("(/111IIIl1l/iwrioll bellm·jor (pp. 2-l9-2ti6). White Plain ... NY; Longman . I\.rnrNlIl. J .. Cuhen, G .. & Nail. P. R. (1999). Sdf-allirllliition theory: An update and an apprai .. al.ln E. ilarmon -Jonc .. & J. Mill .. (Ed!->.). Cogllllil·t, (/;1\011111/('(': Progrnl 011 (/ (J/\'olailltl'on- ill .\(Idal/Hrl'holog\' (pp. I :!7- 1-l7). Wa .. hingtlln. DC : I\meric"lll P.. ydlOlogical A""ociation . Ba~cr. W. E. (1999). When can affectr\ e t;oT1(litrorung and Illere cxp{}~urc dircctly inllucnce h rand chorl.:c? )oumal o{Adl'Nli lillg. 2X! -l J. J 1-1-7. ReaU\oi~. J. L.. & Joule. R. V. (1999). I\. radil.:(ll]x)IIlI of ,ic\.\. on di .. :-.onancc theory. In E. I-Iarmon-J()[lc .. & J !\1 i II:-. (Ed .... J. Coglliril·t, (/i\WlltllIl'(': Pmgrt'5J 011 a (Ii \"Owl ,lit'ory ill I(W;al,Hrcholo.lp I pp , -U 70). Wa .. hington. DC: Amcrican P..,ychologll:a l A .... ocialion Bcrgcr, C. R .. & Burgoon. M (1995). COII/II/IiI/;nlfioll tllIlI v}('illl ill(1'I('IIC(, /'ton'He,l. Ea ... t Lan .. ing. ~11. Michigan State L'ni,cr.. it) Pre .... BUrLhtcin. R. r~. (1989). E\po .. urc and •• ffcct: O\cnlc\\, andmeta-ilnaly .. i,> of re"cilrch. 196X-19S7 PHdwlo,t.:iwl Bul/elm. /I)(j. 165· 1X9. Born .. tein. R. ,... .. Kale. A. R .. & Cornell. K R. (199()). Boredom a... a limiting wmlition on the mere cxpu.. urc d Icc!. )ournal 01 PenOllllliry alltl Sodal P,Hdw!ogr. 58. 791-8IXl. Booth -Buttt:rtkid. S .. Cooke. P. Andrighelli. A.. Ca .. tccl. B.. Lang. T .. Pear..on. D .. & Rodriguc/, B (199.t). Sillwltallcou .. \cr.. u.. c,xdu .. i\c procc .... lI1g of IlCr .. ua .. ive argumcnh (Illd cue ... ell/IIIII//Ilinllioll Qllurluly. n. 21 ·35. Brickm~ln . P .. Rcdtidd. J. Harrr"on . A A. &.: Crandall. R. (1972). Dri'e and prcdi .. pt).. itiun ... a ... factor .. in Iht.: attitudinal effcch olmcre ex pn:-.un: . )(1/1/"1/(/1 0/ L.ll'l'rim(,lIwl Social P.\ \'chologr. X. 31-1-4. Burgoon. J K (199-l). NoO\erbal .. i.gnaJ... In M L. Knapp & G. R. Miller (ELI,>.). HwulhooJ\ o{ill/t"1)('r· \//1/(/1 ('OlllllllllliC(uioll (2nd ed .. pp. :!:!9- 285). Tbou .. and 0[1""'. C A: Sage Publica tion ... Burgoon. M. (1995). Language l:.xpccwncy Theory : Elaoor;r tion. exp lication, and Cxh.! lhion. In C R. Bcrger & M. Burgoon (Ed ... ). COIIIIII/minl/iOlI al/(I weial illfluellce I'mCl'.IJ('\ (pp. :!l) SI). Ea .. , L:IIl~ing. Ml: Michigan State Uni\'cr .. rty P rc:-.~. Bum ... C T .. Harmon-June ... , E.. & Tarpil.~) . W R. (1997). "By faith alone" : Rcligiou .. agltatiun and r.:o!,!niti \'e di ... ,onancc. Bmit" ApplicaliO//\ or Social P,n-dwlogy. J9. 17-31 .
61
Chapter 4 • Theori:ill{? ahollf Persllasio/l
Chaike n. S. (1980). Heuristic vcrsus ,\ystcmatic information procc ..sing and the u~e of ~ourcc \'crw'" mc .. ..age cues in persuasio n. JOllrnal of Per.WJII(I/iIY (///(1 Social Plych%gy. 39. 752-766. Chaiken. S. ( 1987). The heuriMic model of per<;ua~ion. In M. P. Zanna. J. M. Obon. & P. Hl!rman (Eds.). Social influence: The Ontario Symposium (vol. 5. pp. 3- 39). lIillsdale. NJ: Erlbaulll . Chaiken. S.. Libenllan. A.. & Eag ly. A. H . (1989). llt:u ri .. tic and .. ) ... tematic information procc ...... ing within and beyond the persuasion context. In J. S. Ulcman & J. A. Bargh (Eds.). Unilliellded thol/ghl (pr. 212- 252). New York; Guilford Pre ...... Chaiken. S .. & Maheswaran. D" ( 1994). Heuristic proce .. ..,ing can hi as systematic proce .... ing: Effect... of source credibility. argument ambigu ity. and ta,,1.. importance on altitude judgment. JOIII'I/(li of PersOIullit)' and Social Psychology. 66. -\.60~73. Chaikcn. S" & Trope. Y. (Eds.). ( 1999). 01/(/1 process Iheorie.\ ill social psycllO/ogv. Ncw York: Guilford Pre .. s. Chen. 5 .. & Chaiken. S. (1999). The heuri ... t i c-sy~tcmat i c model in it s broadercomext.ln S. Chaiken & Y. Trope (Ed".). Dlla/ pmce.u thl'Orie.\ in .w cial psych%Xy (pp. 73-96). New York: Guilford Pres ... Chen. S .. Shechter. D .. & Chaiken. S. (1996). Gelling at the truth or gelling along; Accuracy VCNt .. impression motivated heuristic and "y"tematic information proces~ing. Journal of Penonality (llId So-
c.
cial Psychology. 71.262-275.
Cia ld ini. R. B .. Wosinska. W .. Barrelt. D. W .. Butner. 1.. & Gornik-Durose. M. (1999). Compliance \\ilh a request in two cultures; The differential inllucncc of ,>urial proof. commitmentfcon .. i .. tem;) on col lectivists and indi\ iduuli"I\. PenOl1l1lit\ (fllll Social p,l.w·/wlogy /Jullelill. 25( 10). 12-\.2- 1253. Cochran. S. D .. Mays. V. M .. Ciaretta. 1.. Caru,;o. C .. & Mallon. D. (1992). Efficacy of the theo!') of reasoned action in predicting AIDS rc lated ri,,1.. reduction among g.ay men . jOlmlll1 oJApplied Sodal Psychology. 22. 1481 - 1501. Compton. R. L Williamson. S .. Murphy. S. G .. & Jieller. W. (2(X12). Hemispheric difference,; in aflccti\t~ response: Effects of mere expo .. urc. Social Cogllilioll. 20{, I I. 1-16. Cooper. J.. & Fa/io. R. H. (198-\.). A new look at di ......onancc theory. Adl'{lI/cCJ ill /:.\pl'rimellllli Social Psychology. /7. 229-266. Corli,;". R. (2002. July 15). Should we all be vcgctarian,\'? Tillie, /60 (3). 48- 56. Cotton. J. L. (1985). Cognitive dis .. o nancc in selective exposure. In D. Zillman & J . Bry:mt (Ed .... ). Se/ecliI'e exposure 10 ('omlllll1liclltioll (pp. 11 - 33). Hill ... dalc. NJ : Erlbaum. COIIOI1. J. L.. & He i...er. R. A. ( 1980). Selective expo . . ure to information and cognitive dissonance. JOllrl/a/ oj Research ill Persolllllily. 14. 518- 527. Devine. P. G .. Tauer. J. M .. Barron. K. E.. Elliot. A. J.. & Vance. K. M. (1999). Moving beyond altitude change in the study ofdis .. onancc-rclatcd prOCl!~'\.t: ... In E. Harmon-Jone~ & J. Mil).., (Ed ... . ). Cogllilin' di ....HJl1(lI1ce: Progre.\'S 011 II pi1'()wl rht;'or\' ill social p,lycllOlogy (pp. 297- 323). W:l'"hington. DC: American P\ychological As . . ociation. Dillard. P. (1993). Per.. uasion past and pre ..ent: Altitude .. aren't what they u..ed to he. Cmlll/llllli(,{lIiOiI MOl/ogmpll.l'. 60( 1).90--97. Dillehay. R. C. (1973). On the irrele\ancc of thc ela" ... icalnegativc e\idence concerning the cffech of attitudes on behavior. Americal/ P~.\'dlologi.\{. 28. 887- 891 . E.'lgly. A. H .. & Chaiken. S. (1993). The 1'.\\"('/10/08.\' OIl/lfilutie,';. New York: Harcourt. Bmct:. Jovanovich . E
62
Part I •
Prelilllill(Jrie.~·:
Definitiuns, Trends. and l1u'oretical Underpinnings
Festinger, L. (Ed.). (1964), COl/flicl. deci~iun, (/nd di"sOI/(U/ce. Stanford. CA: Stanford UnivcP'!ity Pres .. , Fishbein, M., & Ajz;cn. I. (1975). Belief, (ltrifll(/e. illtem/oll. (llId hc/w\';or: All illirodllctioll/o theory (llId research. Reading. MA: Addison-Wesley. Fisher, R. J. (1993). Social desirability bia~ and the validity of indirect quc<,tioning. jOIlrJ/a/ of CO/lSlllller Research. 20. 1993. Furnham. A. (1986). Response bias, socia l dc..,irability. and as"imilation. Persullality al/d Illdidr/i/(/l DU~
Jerence, 7.385-400.
or
or
Gasti!. J. (2000). Thinking. drinking, and driving: An application the theory reasoned action to DWI prevention. jOllmal of Applied Social P.lyc/wlogy. 30( 1 1). 2217-2232. Godin. G .. Maticka-Tynadale. E .. Adrien. A .. Man..;on-Singcr. S .. William..;. 0 .. & Cappon. P. (1996). Cros<"-cultural testing of three social cognitive theorie ... : An application to condom use. JOIlrlwl of Applied Social Psyclwlogr. 26. 1556--1586. Granberg. D .• & Holmberg. S. (1990). Intention-hehavior relatiorl'.;hip among U.S. and Swedi<.,h \Olers. Sucial Psyclwlog\' Quarlerly. 53. 44-54. Greene, K .. Hale. J. L.. & Rubin. D. L. (1997). A test of the Theory of Reasoned Action in the context of condom use and AIDS. CommlllricaliOlr Repo(I.\". JO( 1).21-33. Greenwald. A. G .. & Ronis. D. L. (1978). Twenty years of cognitive dissonance: Case study of the evolution of a theory. Psychological Review. X5( I ). 53-57. Hamid. P. N. (1973). Exposure frequency and stimulus prcference. British Journal of Psychology. 6.J.
569-577. Hamilton. M. A .. Hunter. J. E .. & Bo"ter. F. J . (1993). The elaboration likelihood model a..; a theory of attitude formation: A matllematical analysis. Comm/tnication Tlll'or\'. 3. 50-65. Harmon-Jones, E. (1999). Toward an under"itanding of the motivation underlying dissonance el"fect,.,: Is the pnxluclion of aversive consequences necessary? In E. Harmon-Jones & J. Mills (Ed,>.). CoWritil·e di.Honwlce: Progre.u· on a pil"Of{/1 theory in social psychology (pp. 71-99). Washington. DC : American Psychological Association. Harrison, A. A. (1977). Mere exp0'iure. In L. Berko\\ilz (Ed.). Ad)"{lIIces ill experimell/(ll social psycholug\, (vol. 10. pp. 39-83). New York: Academic Pres'i. Heider. F. (1958). The psychology (~f il1lerperJOIral re/tlliOl/s. New York: John Wiley . IIcyduk, R. G. (1975). Rated preference for musical compositions a,., it relates to complexity and expo<;LIt"e frequency . Perceplioll and P.lyc!wphy.1 ie,. 17. 84-9 [ . Jacoby, L. L.. & Kellcy. C. M. (1990). An episodic vic\\ of motivation: Unconsciou'i influence" on memory. In E. T. Higgins & R. M. Sorrentino (Ed".). The hallliboo/.: of lI/o/i)"{llio/l alld cognition (vol. 2. pp. 451 -58 1). New York: Guilford Pre"s. Kelman. H. C. (1974). Allitudes are alive and well and gainfully employed in the sphere of action. Ameri('{J1r P.~yclJ()l()gi.\·f. 29. 3 I0-324. Kim. M-S .. & Hunter. J . E. (1993a). Allitude-behavinr relations: A meta-ana[y"i . . of alli tudinal relevance and topic. JOIIl"f/al ofComlllllllimlioll. ~3, 101-[42. Kim. M-S .. & Hunter. J. E. ( [993b). Relationships among attitudes. behavioral intentions. and behavior: A meta-analysis of past research. pari 2. CO/ltfflrlllim/ioll Research. 20. 331-364. Klingle. R. S. (1996). Physician cOIlllTlunication as a motivational tool for long-term patient compliance: Reinforcement expectancy theory. COlllm/il/j('alion Studies. 47. 106--217. Krauss. S. J. (1995). Attitudes and the prediction of behavior: A meta-analysis orthe empirical literature. Personality {lnd Social Psychology BIII/elin. 21. 5R-75. Knrglanski. A. W. o Thompson. E. P .. & Spiegel. S. ( 1999). Separatc or equal?: Bimodal notions of persuasion and a single process "ltnimodcl." In S. Chaiken, & Y. Trope (Eds.). Dllal pfOCI'.I'S theories ill social psrC//Ology (293-313). New York: Guilford Pres'i. Lee. A. Y. (2001). The mere e,xposure effect: An uncertainty reduction explanation revisited. P(>I'wlllaliry alld Social P.lycllOlogy Bllllerill. 27( 10). 1255- 1266. Lee. c.. & Greene. R. T. ( 199 [). Cro"s-cultural examination of the Fishbein behavioral intention model. JOllmal of Jlllemario//al Busilless S/udic.I. secolld q/tal"/a. 289-305. Maheswaran. D .. & Chaiken. S. (1991). Promoting systemalic proces~ing in low- motivation setting,,:
Ch,lptcr 4 • Theori:ing about Persuasion
63
Erfect of incongruent information on proccs ... ing and judgmenl. jOllrnal of Personality lind Social Psw-/wlogy. 6/. \3-25. Mahc<.,waran. D.. MacKie. D. M.. & ChaiKen. S. (1992). Brand name as a heuristic c ue : The effects of ta<;k importance and expectancy confonnation on con"ulller judgments. journal of Consumer Psychol· ogy. 1.317-336. Mill .... J . (1999). Improving the 1957 ven.ion of cogniti ve di ....>onancc theory . In E. Hamlon ·Jones & J. Mill s (Ed<;.). Cogllitil'e di.\:1OIIallc(': Progl't'xs 0/1 (l pil'oral rl,eory ill .wcial psychology (pp. 25-42). Washington. DC: American r ... ychological A" . . ociation. Mongeau. P. A .. & Stiff. J. B. (1993). Specifying ca u . . al relationship ... in the Elaboration Likelihood Model. CO/I/mllllic(l/ioll Theory. 3. 65-72. More land. R. L. & Zajonc. R. B. (1977). Is stimulu<., recognition u necessary condition for the occurrence of ex pasure effects? JOIl 1'1/(/1 oI Pl'I'sO/wlilY (I"d Social Psychology . 35. 191 - 199. Newcomb. T. M. (1953). An approach 10 the <"Iudy of communicalive acts . Psyclwlogieal Review. 60.
393-4()4. O<.,good. C. E.. & Tannenbaum. P. H. ( 1955). The principle of co ngruity in the prediclion of attitude change. PsyclwloRicaf Rel·iew. 62. 42-55. O ...good. C. E.. Tannenbaum. P. II.. & Sud. G. J . (1957). The meawreme1l1 of meaning. Urbana. IL: Uni· versifY of Illinois Press. Park. H. S .. Levine. T. R .. & Sharkey. W. F. (1998). The theory of reasoned aClion and self·construals: Under-.. tam.l ing recycling in Hawaii . CO/lllllllni('{llioll SI/Idies. 49(3). 196-208. Patr) . A. L .. & Pelletier. L. G. C:~OOI). E;xtraterrc"trial belief... and experiences: An application of the theory of re
R('.\etm'll. 15.325-343. Sherif. M .. & Sherif. C. W. ( 1967). Altitude" a... the imli .. idua l' ... own categoric . . : the socia].judgment approach to alii tude and attitude c ha nge. In C. W. Sherif & M . Sherif (Eds.). Attiflltle. ego· ill\'()II'emelll. alltl challge (pp. 105- 139). Nc\\ York : Wiley. Sherif. C. W .. Sherif. M .. & Nebergall. R. E. (1965). Allill/de and (lrtitlltle change: The weilll j udgmem· ;1I \,oll'{'mnll al'pr{wch. Philadelphia: W. B. Saunders. Smith. E. R .. & Zarate. M. A. ( 1992 ). E:.:cmplar·ba'ied model of ~ocial j udgment. PsyclJ()lo~ical Rel'iew. YY, 3-2l. Snyder. M . (1974). Self·monitoring of exprc<,si . . c bchuvior. jOllmal of Personlliity alld Social P~)'cI/Ol· ag.\'.
30. 526-537.
Snyder. M. (1979). Self· monitoring proce ......e ..... In L. Bcrkowitl (Ed.). At/wl1Ices ill experimental alld so· cial psychology (vol. 12. pp. 85-12R). New York : Academic Press. Stang. D. 1.. & O'Connell. E. J. (1974). The co mputer a . . ex periment in social psyc hology research. Be· 1I1l1'ior ReulIr('h Methods and IIi Sll'lllllellwt;oll. 6. 223-23 I.
64
Part I • Preliminaries: Definitions. Trends, and Theoretical Underpinnings
Steele, C. M. (1988). The psychology of self-affirmation: Su~taining the Integrity of the Self. In L. Berkowitz (Ed,), Adl'QIICeS ill Experiment(iI Social Psychology (pp. 261 - 302). San Diego, CA: Academic Press. Sleen, D. M .. Peay. M. Y .. & Owen. N. (2000). Predicting Australian adolescents' intentions to minimize sun exposure. Psychology alld Healrlt, 13( 1), 111 - 119. Stifr. J. B. (1986). Cognitive processing of persuasion message cues: A mew-analytic review of supporting infonllation on attitudes. Communication MOllographs. 53, 75-89. Stiff,1. 8. ,& Bosler, F. J. ( 1987). Cognitive processing: Additional thoughts and a reply to Petty. Kasmer. Haugtveldt. and Cacioppo. ComllmlliC(I{iOIl Monographs. 54, 250-256. Slone. J .. Wygand. A. W., Cooper. L & Aronson. E. (1997). When exemp li fication fails: Hypocrisy and the mOlive for self-integrity. Journal oj PersollaliTY alld Social Psychology, 72. 54- 65. Tonealto, T.. & Binik. Y. (1987). The role of intentions, social norms. and alii tudes in the performance of dental nossing: A test of the theory of reasoned action. Joumal ojApplil'd Social Psycholo,;,'·. 17.
593-603. Trafimow, D .. & Miller. A. (1996). Predicting and understanding mental practice. Journal of Social Ps.vellology. 136. 173- 180. Vallerand. R. 1.. Deshaies, P.. Cuerrier. J. P., Pelletier, L. G .. & Mongeau. C. (1992). Ajzen and Fishbein's theory of reasoned action as applied to moral behavior: A confirmatory analysis. JOIII"I/oi of PerSOflaliTY and Social Psychology. 62, 98- 109. Wicker, A. W. (1969). Attitudes versus actions: The relationship of verbal and overt behavioral re:-.ponses to attitude objects. Journal of Social Issues. 25, 41 - 78. Wood. w. (2000). Attitude change: Persuasion and social inOuence. Annual Rel'iell' of Psycho/OR.\'. 51.
539- 570. Zajonc, R. B. (2001). Mere exposure: A gateway to the subliminal. CurrenT DirecTions in Psyrhological Science. JO(6). 224-228. Zajonc, R. B. (1968). Attitude effects of mere exposure. Journa/ of Personality alld Social Psychology Monograph Supplemelli. 9. (2, Pt. 2), 1- 28. Zajonc, R. B., & Rajecki. D. W. (1969). Exposure and affect: A field experiment. Psychollomic Science, 17,216-217. Zajonc. R. B .. Shaver. P.. Tavris. C, & Van Kraveld, D. (1972). Exposure. satiation, and stimulus discriminability. Journal of Personality and Social Psychology, 21. 270-280.
5 The Elaboration Likelihood Model of Persuasion Richard E. Petty, Derek D. Rucker, George Y. Bizer, and John T. Cacioppo
The Art of Persuasio1l: The Early Years The study of pep.)uasion anti rhetoric date ... hack to the time of the ancient Grech. In those times. persuasion wa~ seen a... the in\trulllcnt by whkh debate ... could be rc~olved. indiv idual s cou ld be educated. and idea~ could be comlllunicated to an audience. Given thal persuasion WaS such a vital a ... pect of Greek society. understanding the factors re ... ponsible for social inOuence wa ... crucial. Recogni/ing thIS. Ari ...totle, one of the great thinker... of the time. provided a theory that specified what a speaka needed to know in order to un derstand how to persuade others. Ari ... totlc reasoned that to be successful at persua,ion, one had to under... tand charm.:teri ... tic ... of the ...ource (erhos). the message (l0Ros). and the emotions of the audience (pathos: Ari ... totle. 1954). For exampk. Ari ... totle remarked that if a ..,ource werc well respeCltx.I. It would be t:asier to persuade others of his \'ie\\s than if he was not well respected. In the 2.400 years that have passed since the time of the ancient Greeks. the art of per ... uasion has become an even more i11lcgral part of society. Per... ua ... ion ha ... become the chief tool by which important legislation get<., pas ...ed. products get sold. and parenh influence th eir chi ldren. Furthermore. it i... Ari"!lOtlc· ... ideas thai provided the foundation for much of the ea rl y work on per"!uasion in the twentieth century. Nowhere is thi"! morc ev ident tha n in the work and thcori/ing of Carl HO\];.lI1d and his colleagues who began asse"'''!ing the effech of variablt:s related to the source. thc message. and the audience on the impac t of per<.,uasive attcmpts (Hovland. Jani .... & Kelley. 1953; Hovland. LUl1lsdaine. & Sheffield. 1949: Hovland & Wei". 1951). Early research in persuasion was guided by th e belief that any given variable. for example, ... ouree credibility. had a sing le and unitary effect on persuasion: A \ariablc \\a:-. thought either lU enhance the ~uccess of a persuasive attempt or reduce it. Furthermore. there was an av",Ul1lptioIl that there was one mechani"'im
65
66
Part I • Prelimillaril'.l: Ol'jillilioll.l. Trellds. and 71leorefiml Underpinllillgs
by which the effec t wa~ produced. for example. source credibility enhanced persuasion by increasing learning of th e message. In essence, this re~ earc h followed a "single effect" and "single process" approm:h to understanding the impact of variables on persuasion (see Petty. 1997). Thus. the goal of this research was to determine what the single effect of a variable was and what the process was by which this variable worked. Initial endeavors following thi s approach appeared promising. For example. follow· ing Aristotle· ... noti on of ethos. researchers found that credi ble so urces increased persua· sion ( Hovland & Weiss. 1951). Following Ari~totle's concept of logos. research ers found that increasing the number of arguments in favor of a position increased the overall amount of persuasion (Calder. Insko. Yandell. 1974). Finally. researchers following Ari~totle· ... concept of p~lIhos found that placing the audience in a negative emotional state red uced per'iuasion (Zanna. Kiesler. & Pilkoni s. 1970). Furthermore. (he researc hers tied th e effect~ of these va riables to single processes. For example. negative emotion was said to reduce pcr~lIasion because of classical co nditioni ng (Staats & Staats, 1958). Although some carly research wa~ co n!'!istent wi th the idea that a variable had a single effect on persuasion via one mechanism. the single·effect and single· process up· proach soon became untenable. Resea rch on persuasion began to experience a period of chaos and turmoil because subseque nt research findin g~ con tradicting early results began to appear in the literature. For example. subseque nt researc h on increasing th e number of argument!'! in a mcssagt! found thai more arguments did nO( always lead to gremer altitude change (Norman. 1976). Subsequent researc h 011 source credibility and negative emotions found that some times highly credible sources cou ld be associated with reduced persuasion (e.g .. Sternthal. Dholakia. & Leavitt. 1978) and that negative emotions could be used to increase per'iuHsion (Roge rs. 1983). Uncovering different findings led researchers to pos· tulate dirferent procc~se~ by which the variables worked. Even when resea rchers could agree on th e !'!i ngl e effec t that wa!'! to be observed. they often di sagreed on the process by wh ich the effec t came about (c.g .. was it dissonance or self·perception?: Greenwald & Ranis, 1(78). This slate of affairs c rippled the approach of searching for the ~ ingle effect of a g iven variable and its sing le process. However. co nflicting findings did more than simpl y destroy thi s approach: They placed the en tire field of attitude change in a state of confusio n (e.g .. Himmc lfarb & Eagly. 1974). This left the state of attitude researc h in need of a resolution of these apparent contradictions . The Elaboration Likelihood Model (ELM) (Petty & Cacioppo. 1981. 1986b) was developed to explain and orga ni/c past con ni cts in the persuasion literature as well as to guide new research. The goal of this c hapt er is to provide an underslanding of the basic tenets of the ELM as a framework ror understanding and investigating the effects of per· suasive communicat ions. To thi 'i end. th e key postulates of the ELM are reviewed , and the utility of (he model for rc!'!olving con Oicting findings in the literature and guiding research is highli gh ted. Furthermore, misconceptions, misinterpretations, and challenges to the mode l are co nsidered and addressed.
Overview o/the Elaboration Likelihood Modef.1 As articulated in more detail short ly. the ELM outli nes a finite number of ways in which variables can impact judgments. and it spec ifics when variables take on these roles, as
Chapter 5 • 1'11(' t:lulmralirJlj Likelihood Model oIPer.\I/(l\ioll
67
well a~ the con ...equences resuhing from these different role .... That i.... the ELM is a theory about the proces..,cs underlying changes in uuitudcs. the variables that induce these pro~ cesses. and the strength of the judgments resulting from these processes. Unlike the single-process and single-effect approache ... described earlier. the ELM docs not hold that a given variable has only a single eficl:1 on persuasion or inlluences persua ... ion by only one proces .... Instead, the ELM posits that anyone variable can inllucncc attitudes in a number of different way..,. The same variable. depending on the role it plays. can act either to increa ... e or decrease per. . ua..,iol1. Furthermore. whether the variable serve.., to increase or decrea ... e per... uasion. it can do so through ..,e\,cral diffcrent mechanisms. At the core of the ELM is the elaboration continuum. The elaboration continuum is ba..,ed on a person· ... JII01il'll{iol1 and ahility to think about and as..,!!s.., the qualities of the i. . sue-rele\'unt information a\'ailable in the persuasion context. When both motivation and ability to think are high. individuals an: indined to scrutini/e carefully all issue-relevant information stemming from the ... ource. me\ . . age. context. and them\elve\ (e.g .. their emotions) in an attempt 10 make an accurate judgmcnt about the merits of the is\ue (called the cellfral rOllle to persuasion). However. when either motivation to process is low (e.g .. if personal rckv~lnce is low) or ability to proces\ is hindered (e.g .. if a per\on is distracted) attitudes can be changed by onc or more of it family of relatin!ly low ~e ffort processes (called the p£'riplleral rowe 10 persuasion). Thu\. the ELM posits that for the \akc of \implicit}. pcrs1l3\inn can be thought of as following one of two route!>, to persuasion: cenlral and peripheral. More specifically. in their pure form the two route ... to attitude change correspond to anchoring points on an t!laboration continuum. The central route entails attitude change that requires much effort and thought to reach a deci\ioll. For example. carefull) ...crutini/ing the merits of the \ub... tantive information in a message and integrating one's thoughts into a coherent position are pratotypi<..:al actions based on the central route to pcr..,lIa..,ioll. The second route. the peripheral route. entails attitude change that occurs primarily when elaboration is low, and it (3n involve thought processes that arc qU<.llltitativcly or qualitatively different from the high-elaboration central route. For example. a low-elaboration processor might carefully scrutini/c only the first argument or two rather than all of them (quantitative difference in proces\ing) or might prace\s all of the arguments by cQullting them rather than scrutiniling them for merit (qualitati\,c uiffcrence: see Petty. Wheeler. & Bi/cr. 19(9). What these t\\'O proccs..,or, ha\c in common i, thc relatively 10\\ amount of thought il1\"o hed in attitude change. The ELM ..,pecities that \\ hether attitude change occurs by the central or the peripheral route has impol1ant Implication~ for the ~trength of the rc'-.ulting attitude. That is, attitudc changes brought about through high-elaboration processe.., tend to be more pcrsi..,tent, resistant. and predictive of behavior than changes brought abuut because of low elaboration proCl:..,ses (Petty. Haugtvedt. & Smith . 1995). Thi\ issue j.., discussed further later in thi.., chapter. Of coursc. since elaboration is a continuum. atlitude change is SOl1le~ times brought about by a medium amount of thought (rather than \'cry high or low amounts) and can be detennin~d b) som~ combination of cCJ1lral and peripheral route proces ...e..,. A key idea of the ELM is that multiple persuasion proce\\e\ operate along the elaboration continuum. and different pcr ... uasion proces..,c" require different amounts of thought. That is. the ELM recognilC\ that attitude change i\ influenced by a variety of ~pecific processes such as (ognitive rcsponse.., (Greenwald. 1968; Petty. Ostrom. & Brock.
r
68
Part I • Prelilllill(lries: De.f/lliliol/.\. Trends. (lnd Theoretical Ul/de1l);III/;l/gs
1981). integration of beliefs (Fi,hhcin & Ajlen. 1981 J. self-perception (e.g .. 8cm. 1972). classical conditioning (e.g .. Staah & Staat~. 1958). reliance on heuristics (c.g., Chaiken. 1987). and cognitive dissonance (e.g .. Festinger & Carlsmith. 1959). Some of these processes <:Ire more likely to influence altitude~ at low leveb of elaboration (e.g .. clilssical conditioning), othcr~ require ~oll1e minimal amount of thinking (c.g., self-perception). and still others an:: morc likely to influcnce attitudes at high levels of elaboration (e.g .. cognitive responses). In short. the ELM is a multiproccs~ theory of persuasion that views persuasion processes a~ falling along an elaboration continuum. When attitudes change as a re~ult of relatively high amounts of i~sue-releval1l elaboration. people are said to follow the central route. but when attitudes change as a re ... uh of relatively low amounts of is~ue-relevant elaboration. they arc said to follow the peripheral route. Whether persuasion occurs through the central or peripheral route is determined by a person's motivation and ability to think aboLlt the is~uc-relevant information available. A schcmatic representation of the ELM is depictcd in figure 5.1. Having provided the basic outline of the model. we now discuss its specific postulates.
Postulates of the Elaboration Likelihood Model Petty and Cacioppo (1986b) pre ... cnted the ELM in seven postulates (sec also Petty & Wegener. 1999). We do not present the full formal po~tlliates here. only the gist. That is. we explain the essence of the postulates. along with a :-.ampling of research relevant to each.
Postulate J: The Correctness Postulate The first pO~llIlatc of the ELM statc~ that people are motivated to hold what they believe to be "correct" attitudes. Correct attitudes need to be correct. not necessarily logically but in the sen.,c of an individual's o.;ubjeclivc appraisal. Correct attitudes are helpful because they often allow people 10 gain rcward~ and avoid puni.,hmcnts by approachi n£ helpful objects and avoiding dangerous one .... Holding correct attitudes i~ imponant if people want to act on their altitudcs. People can determine which attitude is most correct in a number of way.,. When motivation and ability to think are high (such as when the issue is an important one), perhaps the most obviolls way for a person to gain confidence in the correctness of one's view is to con~ider carefully all of the issue-relevant information available. However, if either motivation or ability to think is low, one might attain sufficient confidence. for example. by ... imply relying on an expert source. If the issue is important but there is insufficient time for proce~sing right now. the per...oll might tag the iso.;ue for later scrutiny (~ee Petty. Jarvi;. & Evans. 1996). The assumption that people want to be correct doe~ not imply that people cannot be biased in their assessment of evidence, however. In fact. being certain that olle is correct and wanting to maintain one's correct attitude can lead to defensive processing of contrary information (Petty & Wegener. 1998). The first postulate of Ihe ELM merely assumes that people are rarely explicitly motivated to be biased. Rather than explicitly being motivated
Chapter 5 • The Elaboration Likelihood Model of Persllw,ioll
c§"'rsuaSive Communication
Motivated to Process? (personal relevance , need for cognition, etc.)
YES Ability to Process? (distraction, repetition, knowledge, elc.)
~ NO
~
YES
Peripheral Attitude Shift Changed attitude is less enduring, less resistant to counterpersuasion , and less predictive of behavior
I'
YES Is a Peripheral Process Operating? (identification with source, use of heuristics, balance theory, etc.)
What Is the Nature of the Processing? (argumenl quality, initial attitude, etc.)
NO
More More Unfavorable NO Favorable Thoughts Than Thoughts Than Before? Before?
Is There a Change in Cognitive Structure? (thought rehearsal , ease of generation, etc.)
Attitude does not change from previous position
NO
YES
YES
(favorable)
(unfavorable)
Change
Retain Initial Attitude
1 YES
YES
Central Positive Attitude
69
Central Negative
Attitude Change
Changed attitude is relatively enduring, resistant to counterpersuasion, and predictive of behavior FIGURE 5.1
The Elaboratioll Likelihood Model of Attitude Change
Adaptcd from Pctt y & Cacioppo, 19S6h.
to be biased. indi viduals may possess various goals or motivations that will promote bias. For example. people are so metimes motivated to be co nsistent over time (which can cause th em to de fend their attitudes). or th ey can be motivated to impress others. which might lead them Lo try to sec the merits in whatever posit ion a liked individual has (Kelman.
196 1).
70
Part I • Prelillli//arin:
D(~fil/itiom,
Trel/ds, alld Theorerh'(ll Ullderpillllillg.'i
Postll/ate 2: The Elaboratioll COlltinllll1ll Postll/ate Thc second postul<Jte ~tate~ that the amount of mental processing (i.e., elaboration) in which a per~on engage~ regarding a message varies along a continuum. A~ stated earlier. at one end of the continuum, the per~on engages in no thought whatsoever about the i~sue relevant information available in the pcn.uasion context. Conversely. at the other end of the continuulll. the per,oll engages in extensive elaboration of all information available. Because a continuum exists between extrcmc and zero claboration. people can engage in any middling degree of thinking about a persuasive communication. The elaboration continuum i~ a qlllllllifclfive dill1en~ion. For cxample. two different people may be elaboraling on an advertisemcnt for a mutual fund. One person may be evaluating the l1les~age because he is con,idering where to invest his entire life's savings; the other may be cvaluating the message because she is considering where to invest a $50 check. The two people may both bc using the arguments in the message to determine if the fund is a wi"ic inve~tlTIcnL but one may be evaluating the message more carefully or evaluming more information than the other. In such a case. although both people are engaging in the.! :-.ame qllll/if(llil'£' proces,", (i.C!., thinking about i~~ue-relevant information). they may be exhibiting qU({fllitafil'e difference,", in the extent of processing (cf. Friedrich. Fethcr!--lOnIHlUgh. Casey. & Galli.lghcr, 1996), The:>te quantitative differences in the amount of thought can lead to difference~ in the strength of the attitudes formed. It is also possible that people may engage in different qllalillltil"e processes along the elaboration continuum. Returning to the above example. the individual who is carefully evaluating the l1les~age may u:-.e the quality of the arguments as the primary determinant of ""hether 10 invest in a particular fund. However. the individual who i~ not carefully evaluating the mes,",agc may lise the mere number of arguments rather than their quality to decide whether or not to invest her money in the fund. For example. if there arc nine arguments in favor of the fund. the individual may conclude that since there are so many arguments. the fund must be good. Here. the two potential investors arc processing the information in a different way both qualitatively and quantitatively. One person is effortfull) asses!'>ing the merit of the arguments. whereas the other is using the less cognitivc1y demanding strategy of counting the arguments. This might lead the potential invcstors to reach very different conclusion.s. For instance. if the nine arguments are all specious. the person processing the information by the counting mechanism should be more persuaded by nine we.lk Jrgllments than by three strong argumenh for the fund. Conversely. the per!-.on '" 110 is evaluating the merits of the arguments ~hould be less per~uaded by nine weak rather than by three strong arguments (see Petty & Cacioppo. 198-lb). Thus. at different level~ of elaboration. the samc processes may be operating in varying degrees (a quantitative difference). or entirely different processes may be operating (qualitative difference). Rcturning to the elaboration continuulll. what detcrmines where on the continuum the l11e~sage recipicnt will fall? A.., noted earlier. when a person has a high degree of l1loti\'atioll and abi/ily to elaborate carefully on the message arguments, processing will be further IOwaI'd the central end of the continuum. When the person lacks either motivation or ability. the processing will be clo~er to the peripheral cnd. Consider an advertisement with a long list of arguments detailing the merits of a particular type of car. Whether people pay
Chapler 5 • The Elaboration Likelihood Model of Persuasion
71
close attention to and elaborate on those arguments has much to do with the degree of 11l0/iHllioll they have to do so. For example. if a person is in the market for a new car, he or she may be highly motivated to scrutinize carefully and think about the relevant information pre~ented. If, however, a person has no interest in purchasing a car in the near future, he or she will lack the motivation to engage in effortful processing. This person is more likely to follow the peripheral route to persuasion (Petty, Cacioppo, & Schumann, 1983). Next. consider an article about a new antihistamine in a medical journal. If a doctor has spent years in medical school and understands medical jargon, he or she has the ability to process the arguments in the article carefully. However, regardless of motivation, a first-year undergraduate student likely lacks the ability to understand and process the medical jargon. This undergraduate will therefore be more likely to resort to the peripheral route {Q persuasion. Whereas personal relevance serves as a motivational variable in the example involving the new car, knowledge of medical jargon serves as an ability variable in the example involving the antihistamine, A lack of either motivation or ability will move people toward the low end of the elaboration continuum. It i~ important to point out that the distinction between high and low elaboration :-.hould not be viewed as a distinction between "good" versus "bad" persuasion. For example. the usc of the peripheral route can be an adaptive, necessary tool in people's everyday lives. When motivation or capacity is low, one might forgo decision making-which is not always possible-or postpone it until conditions foster it (Petty et aI., 1996). It is also imponallllo note that thinking does not ensure an optimal outcome, as one's thoughts can be bia~ed by various contextual factors. For example, when people are spending a lot of time on active thought, their assessment of arguments is biased by their mood states (e.g .. DcSteno. Petty. Wegener. & Rucker, 2000; Wegener, Petty, & Klein, 1994).
Postulate 3: The Multiple-Roles Postulate The third po:-.tulate of the ELM state~ that variables can play multiple roles in persuasion contexts. The role in which a variable is most likely to serve depends on the situation. First. when elaboration is not constrained by other variables to be high or low. variables tend to serve as determinants of the amount of thinking that takes place. Second, when elaboration i:-. low variable:-. tend to serve as cues or input to low-elaboration processes such as clas:-.ical conditioning and use of decision heuristics (Chaiken, 1980). Finally. if the elaboration likelihood is set at a high level by other variables in the persuasion context (e.g., high personal relevance, high knowledge, few distractions). variables tend to serve in yet additional roles and lead to persuasion through high-elaboration processe~. For example. when people arc actively thinking, the variable can be processed as an argument or can bias the ongoing information-processing activity. Thus. depending on the likelihood of elaboration in any context, variables can influence attitudes in a multitude of w<.lys. As an example of a variable serving in multiple roles depending on context, consider a person's mood state. Depending on the situation, mood can serve in a variety of roles specilied by the ELM. First. a person's mood can serve as a determinant of the extent
72
ParI I • Preliminaries: Definitions. Trends. and Tll('orelicul Underpinnings of elaboration when thinking is not already constrained to be high or low by other variables. Based on the Hedonic Contingency Model (Wegener & Petty, 1994) which holds that people in a positive mood are especially motivated to maintain this state, Wegener. Petty. and Smith (1995) hypothesized that being in a positive mood should enhance message elaboration relative to a sad mood if the message recipient believed that process ing the message was likely to make people feel happy. Conversely. being in a pos itive mood should lead to less elaboration than a sad mood if an individual believed that processing the message would be likely to make people feel negative. To test this idea. Wegener and colleagues (1995) told so me individuals that processing an upcoming message would be a generally positive or a negative experience. In actuality, everyone received the sa me message. When happy people expected processing the message to be uplifting, they processed the message more carefully than did people in a sad mood. However, when people expected processing the message to be unpleasant, happy individuals did not process the message as carefully as did individuals in a sad mood (see also Schwarz, Bless, & Bohner. 1991). According to the ELM , when the likelihood of elaboration is low, mood can serve as a simple cue to decide whether or not to accept a message. This could be the result of a number of processes such as classical conditioning (Razran. 1940) or mood misattribution (Schwarz & Clore. 1983). In the case of mood misanribution , people mistakenly infer their attitude from their mood (e.g., "If I feel good, I must like it"). When the likelihood of elaboration is high, however, mood can serve as an argument (Martin. Abend, Sedikides, & Green, 1997) or bias the ongoing thoughts (Petty et aI., 1993). Research by Petty. Schumann, Richman, and Strathman (1993) provided an illustration of the multiple roles for mood under high- and low-thought conditions. In one study. Petty and colleagues had participants view a series of commercials, one of which contained an advertisement for a pen. Some participants were led to believe that they would get to select a pen as a gift at the end of the st udy (high-elaboration likelihood). whereas others were led to expect they would select an alternative gift (low-elaboration likelihood). The critical ad for the pen as well as other commercials was placed within a televi sion program that invoked either a positive mood in the participants or invoked no mood. In both high- and low-e laboration conditions. participants rated the adverti sed pen more favorably when placed in the context of the television program that had invoked a positive mood. Although the attitudinal effects of mood were the same in the high- and lowelaboration condition, the underlying processes were quite different. Using path analyses. Petty and his colleagues (1993) showed that, whereas mood had a direct effect on attitudes in the low-elaboration condition, the effect of mood on attitudes in the high -elaboration condition was mediated by the valence of thought s generated. That is. being in a positive mood biased the type of thoughts people generated under high elaboration. Here, mood was not used as a simple cue: instead , mood influenced the valence of thoughts that were generated, and these thoughts in turn influenced attitudes (see Petty, Dcstcno, & Rucker, 200 I, and Pelly. Fabrigar, & Wegener. in press. for further discussion of the role of mood at different levels of elaboration). In addition to the work on emotions noted above, a variety of source. message, and recipient factors have also been shown to work in multiple
Ch.tptcr 5 • The Elaboration UkelillO(I(/ Model of Penllasioll ways in different situations for rcview~).
(~ee
73
Petty, Priester. & Brino!. 2002: Petty & Wcgcner, 1998,
Postulate 4: The Objective-Processillg Postlliate The fourth po~tulate of the ELM addresses situation~ in which people are engaged in objective procc~sing, that is, they are interested in (lchie\'ing the "truth" from a message rather than achieving a particular attitude toward a target. In such situations, the fourth postulate states that variablc:.- impact a person's motivation and/or ability to process a message by either enhancing or reducing the scrutiny of message arguments. Some variables affcct a person'.Io. overall motivation to think about the message, whereas others affect his or her overall ability (0 think about the me.lo.sage. The processing is considered to be "objective" if people follow the evidence wherever It leads. That is, the information proces\ing does not favor one particular outcomc over another. To demonstrate this notion. consider a person who wants to know whether an insurance policy is good or bad. Prior to processing. the person has no stake in the outcome. She may simply hope to learn whether the policy is a scnsible one. If motivation and ability in this scenario are high, attitudes will be impactcd by how compclling the issuerelel'{Jllt arguments within the mel)sage arc. Thus, if the arguments within the message are compelling. the recipient will generate favorable thought~ and develop a positive attitude toward the policy. If the arguments within the message arc wcak. however, .Io.he will generate unfavorable thoughts, leading 10 a relatively unfavorable attitude toward the policy. If. however. motivation or ability i~ low, her attitude likely will not be affected by scrutiny of the arguments. Rather, aUitudes may change because of a peripheral process. Research on the ELM has identified a large number of variablcs that influence the amount of thinking people do when confronted with a persuasive message. For example, Petty. Wells. & Brock (1976) demonstrated that di.lo.tfUction call either enhance or diminish attitude change depending on what kinds of thought~ the uistraction disrupl~. When a message contained compclling arglllnents. di~lraction disrupted the favorable thoughts that normally would have been elicited. thereby decreasing persuasion. However. when a message contained specious argumcnts. distraction dbruptcd the unfavorable thoughts that normally would have been elicited and thereby increased persuasion. Thus. distraction itself did not impact persuasion directly. Rather, it impacted the extent to which the arguments within the message were proccsst!d and thereby influenced the extent of attitude change. In addition to distraction. other variables that have been ~hown 10 influence a person's ability to proce~s a Il1c~. sage include the message's complexity (Hafer, Reynolds, & Obertyn,ki. 1996), the time a person has to procc" the message (Kruglanski & Freund. 1983), the number of opportunities a person ha~ to ,)crutinize the arguments (e,g., Cacioppo & Petty. 1979), and a person's knowledge of the message topic (e.g., Wood, Kallgren. & Preisler. 1985). Variables that have been shown to affect a person's overa ll motivation to think about a message include the personal relevance of the communication (e.g., Petty & Cacioppo, 1979. 1990). an individual's need for cognition (Cacioppo. Petty, & Morri.lo., 1983). one's per\onal rcspon.lo.ibility for evaluating the message (Petty. Harkins.
74
Part I • Preliminaries: Definitions, Trends, and Theoretical Uf1derpil/l1ill!r~
& Williams, 1980), the expectation of having to discuss the message with someone else (Chaiken, 1980), presentation of the message in an unexpecled formal (Smilh & Pelly, 1996), and presentation of a message on a topic about which people feel ambivalent (Maio, Bell, & Esses, 1996),
Postulate 5: The Biased-Processing Postulate Variables not only affect the amount of thinking that takes place btu can also inlluence the nature of the lhoughl process. Thus, lhe fifth postulate of the ELM deals wilh biased processing. Some variables affect a person's motivation to generate certain kinds or thought s, whereas other variables affect a person's ability to generate certain kinds of thoughts. Consider, for example, a situation in which a person has just purchased a new computer. She likely holds a positive attitude toward lhat com puler and probably wants 10 maintain that positive attitude because it would be dissonance-arousing to believe that one's choice was incorrect (see Harmon -Jones & Mills. 1999. for a recent review of dissonance work). If she reads a Consumer Reports article shortly after purchasing her new computer, it is not likely that she will process the article in an objective manner. Because the person wants to hold a positive attitude, she will be motivated to think positive thoughts about the message. She will try to see any arguments presented in regard to the computer she purchased in the most favorable light possible. Other variables can induce a desire to reject the message. For example, forewarning people of a speaker's persuasive intent can motivate counterarguing and resistance to the message (Pelly & Cacioppo, 1979). Ability factors can also be important in producing resistance. For example, negative emotional states might make negative thoughts and ideas more readily accessible (Bower, 1981; Forgas, 1995). On the other hand, having a great deal of knowledge in support of one's attitude might make it easier to counterargue messages against one's viewpoint (Wood et aI., 1985). Often, people are not aware of the biases that influence their information processi ng. However, in some cases people may become aware of a bias that they consider inappropri ate and attempt to correct for it (see Pelly & Wegener, 1993; Wilson & Brekke, 1994). For example, in one study Petty, Wegener. and White (1998) gave students a persuasive message in favor of a policy requiring senior comprehensive exams that came from a source that either praised the students' school (likable source) or disparaged the students' sc hool (unlikable source). In addition, half of the participants were told lhatthe exam policy was for their own university (high relevance) and half were told the exam was for another university (low relevance). Finally, half of the participants were told not to let their personal opinion of the speaker influence their evaluation of the message. Petty el al. (1998) found that when participants were not cautioned about using their persona l opinion of the speaker to evaluate the message, and the issue was low in personal relevance, they were significantly more persuaded by the likable source than by the unlikable source. However. low-involvement participants who were cautioned about using their opinion of the source to form their evaluation were equally persuaded by both the likable and unlikable source. That is, they corrected for the source bias. When the issue was high in personal relevance and participants were not cautioned about the possible source bias. attitudes were not influenced by the source (since, as expected by the ELM, under high-relevance conditions
Chapter 5 • The Elabororioll Likelihood Model (~rPerHllIsirJ/l
7S
people focused on evaluating the sub~tantive b"ue-relevant arguments). However. when issue relevance was hi gh and people were forewarned of a possible source bias. people st ill corrected for a pre~lIl11ed him•. leading them to be morc persuaded by the unlikable than by the likable source. This and other rese'lfeh (e.g .. Wegener & Petty. 1995: Schwafl & Clore. 1983) has demonstrated that in ~OI1lC circum ... tance~. people will attempt to bias their judgments. If a potential bim. is made salient. people can and do correct their attitudes. This can lead them to remove the bias. though if overcorrection occurs. a reversc bias can become apparcnt.
Postulate 6: The Trade-off Postulate The sixth postulate predicts a trade-off bctwcen the impact of argument elaboration and peripheral route processes on attitudes. That is. as the likelihood of is!\lIc-relevant thinking is increased, the impact of cen tral route processes (e.g .. examining information for merit) on attitudes incrcases. and the impact of peripheral routc processes (e.g .. counting arguments) on attitudes decreases. Conversely. as the likelihood of issue-relevant thinking decreases. the impact of peripheral route processes on attitudes increa~es. and the impact of central ro ute processes decreases. It is important not to interprct the trade-off postlliate as suggeMing that certain variables (e.g .. sources) are proce~'ied only when e laboration i ... low and others (e.g., message factors) only when e luboration is high. Rather. this postulate holds that variables are more likely to have their impact as a re ... ult of a low-effort prece.ss when the elaboration likelihood is low but by a higher effort process when the likelihood of elaboration is high. For example. a !-.ource variable can be proce . . scd under highc l;.Iboralion condit ions. but it is evaluated for its evidentiary value rather than working by invoking a si mpl e dccision heuristic or other mean ..... Likewi . . e. message argument ... can be processed under low-e laboration conditions. but the processing is either nol as thorough as it is under high elaboration or reprcsents a qualitatively different low-effort mechanism (e.g., counting the arguments rather than evaluating them for merit). It is also important to note that at most points along the elaboration continuum. both central and peripheral processes influence attitudes.
Postulate 7: The Attitude Strellgth PostuLate The tinal postulale of the Elaboration Likelihood Model deals ,I, ith the outcome of mc ... sage processing. Specifically. this po . . tulatc states that attitudes created or changed by the central route will be more persi . . tent over time. will remain more resistant to persuu\ion. and will exert a greater impact on cognition and beha,·ior than will attitudes changed or created through the peripheral route. That b. although attitudes can be changed to the samc degree under the central and peripheral route ..... the central route produce . . "stronger" attitudes. When attitudes are based on high levels of e laboration, people have the necessary "backi ng" to defend their attitudes against later counterattitudinal persuasion attempts and to maintain the attitude over timc. Thcse attitudes will also tend to be more accessible and held with greater confidence. Because of this higher acces . . ibility and CO/1fidence. people will bc more like ly to act on central route attitudes. Attitudes based on peripheral processes and simple clles. however. arc less likely to demonstrate the ...e char-
76
Pan I • Preliminaries: D(finiTions. Tr{'ful.\', mul Theoreriw/ Underpinning.,
acteristics. Evidence that altitudes formed under high elaboration are stronger than tho:-.e formed under low elaboration has been found in several studies (e.g .. Cacioppo. Petty. Kao. & Rodriguez. 1986: Chaiken. 1980: Haugtvcdt & Petty. 1992: see Petty. Haugtvcdt. & Smith. 1995. for review and analysis).
Puttillg It All Together: Resolvillg COllflictillg Filldillgs with the ELM As ~tated at the out~et of this chapter, the ELM was developed in part to orgalliLc and explain apparent contradictions in the persuasion literature. Having reviewed the elaboration continuum and the multiple roles postulates. readers may already have a good grasp of how this i:-. accomplished. Still. an illustration is worthwhile. Con~ider the effects of sourcc credibility. Recall that early re~earch found that credible sources typically increased persuasion (Hovland & Weiss. 1951). whereas later research found that this wa:-. not always the ca:-.e (Stemthal et al.. 1978). Using the ELM as a framework, one can derive specific predictions regarding when a credible source is likely to lead to more. less. or equal persuasion relative to a source of questionable credibility. Consider a situation in which a person is given a message containing either weak or strong arguments and presented either by a source with high credibility or by one with low credibility. How might the credibility of the source impact persuasion? This depends on the amount of elaboration involved. First. consider the situation in which elaboration is low due to lack of effort or ability. In this case. the individual will not devote much effort to processing is:-'lIc-rclev<1nt information and will instead rely on simple cues to decide whether to accept the message or not. In particular, a high-credibility \ource may be used as a cue to trust and accept the message, whereas a low-credibility source may be used as a cue to mistrust the message and reject it. Thus. when elaboration is low. the credibility of the source may serve as a peripheral cue invoking a persuasion hellri~tic (i.e., "experts can be trusted"). leading to more or less persuasion in the absence of much issue-relevant thinking. As a result. regardless of argument quality. people may have less favorable attitudes when the message is presented by a low-credibility source than by a high-credibility source (e.g .. Petty. Cacioppo. & Goldman. 1981). Now consider a situation in which elaboration is high and. as a result. people arc motivated to proce:-.s the arguments of a message. In this example. the credibility of the source may be relatively unimportant as a cue for deciding whether to accept or reject the message. Instead. if the quality of the arguments is unambiguow,. only the substance should matter and source expertise is likely to have little impact (Petty. Cacioppo. & Goldman. 1981). However. if the arguments arc ambiguous and open to multiple interpretations. expertise might bias the interpretalion of the arguments. leading to more favorable interpretations of the arguments when expertise is high (e.g .. he must have meant this) ratherthan low (Chaiken & Mahcswaran. 1994). Finally. consider a situation in which elaboration is moderate. In this situation, learning that the source is credible may cause people to decide the message is worth pay-
Chapter 5 • Tlu' Elabora/ioll LiJ..elihood Model (~r Penlwsiofl
77
ing close attention ro. Icading to an increase in the amount of elaboration given to the me
COilfusions and Misillterpretations of the ELM Although the ELM has proved useful in resolving contradictory findings in the per~uasion literature and continues to serve as a useful framework for guiding re search. it has not escaped some critici . . m. On the one hand. criticism that point . . to logical flaws in a theory or a mismatch between theory and daw can be useful in fixing or advancing a theory or. in some cases. for putting a theory to rest. On the other hand. criticism that arise . . from misunder~tandings can lead researchers to reject or modify
Sillgle- versus Muitichallllelllljormatioll Processillg In one of the earliest questionings of the theory. Stiff (Stiff. 1986: Stiff. 1994: Stiff & Boster. 1987) suggested thai the ELM docs flot accurately rellect the way in which people process information. Stiff argued that the ELM depicts humans as :,ingle-channcl information proce . . sors. capable of proce%ing only peripheral clles or message arguments. even
78
Part I • Prelimi/laries: f)(~/i/l itiollS. Trel1ds . and TheoreliClJI Ul1derpinnings
though prior research seemed to indicate thut humans are capable of parallel information proce"ing (e.g .. Kahneman. 1973). The assumption that the ELM does not allow for dual-channel (or parallel) information processing i~ simply wrong. Although early presentations of the ELM (Petty & Cacioppo. 1981. 1986a. 1986b) did not comment explicitly 00 the distioction between single versus parallel processing. the ELM never portrayed information processing as prohibiting parallel proce~sing. This misunderstanding arose from Stiff's (1986) apparent view that because some ELM research has ~hown that argument quality had an impact on attitudes under high-processing conditions whereas source allractivene":.s did not (e.g .. Petty. Cacioppo. & Schumann. 1983). people could process only arguments. but not ~ources, under high-elaboration conditions. In stark contrast to this assumption. the ELM holds that people process as much information as possible (including source and message factors) under high-elaboration conditions. This information can be proces~ed either serially or in parallel. Just because information is processed, however, does not mean that it will affect attitudes. Thus. people might be cognizant of the mere number of arguments or the attractiveness of the message source under high-processing conditions but still might not view this information as a valid basi~ for attitude inference (Petty. Kasmer. Haugtvedt. & Cacippo. 1987: Petty & Wegener. 1999). Thus. as aniculated io our discllssion of the trade-off postulate of the ELM. it is flot the case that people process only peripheral cues when elaboration likelihood is low and only central arguments when elaboration likelihood is high. Rather. both types of information may be processed. The trade-ofT postulate addresses the impact of central and peripheral processes on attitudes (see Pelty ct al.. 1987. for further commentary on this critici~m).
COllfusioll over Source VerSlls Message Factors Perhaps the most COlll11l0n misllndcr~tanding of the ELM can be traced to the multiplepostulate- the idea that anyone variable is capable of influencing attitudes by different means in different situations. Several researchers (e.g .. Stiff. 1986: Kruglanski & Thompson. 1999) have mistakenly viewed the ELM as c1as~ifying all message variables (e.g .. number of arguments. argument quality) as lIr~WllellfJ influencing attitudes under the central route and all non message variables (e.g .. source credibility. a person's mood) as peripheral cue.\· influencing attitude~ only under the peripheral roule. This has led researchers to claim that the theory cannot explain results of studies in which non message factors (such a~ source credibility) influenced attitudes under high-elaboration conditions. or where message factors influence 311irudes under low elaboration conditions (KrugJanski & Thompson. 1999). However. as explained in the multiple-roles postulate, the ELM holds that the same variable can serve in different roles. depending on the extent of thinking. For example. early ELM research showed that the attractiveness of the message source could serve as a simple cue and influence attitudes by a heuristic process when thinking is low. but Lhe same manipulation could influence attitudes under high-claboralion conditions if analysis of the variable as an argument provided cogent evidence for the merits of the ;'luitude obrolc~
Chapter 5 • The Elahom(ioll Likelihood Model of Penum;oll
79
ject (e.g .. an attractive spokesperson for a ~hampoo might provide cogent vi~ual tCMimony for th e effecti veness of the product; Pelly & Cacioppo. 1984a). Furthermore. ju>! as early resea rch !o.howed that source variab lc!o. could !o.crve in multiple roles. so too did early researc h show that message variables cou ld be proccs,ed in a heuri~tic manner (cou nting) or a more cen tral manner (eva luating quality: Petty & CCicioppo. I984b). Thu .... source. message. reci pie nt. and contextual vClriab les can innuence attitudes under high, low. and moderate levels of elaboration. btltthe underlying IllcchClnislll will vary (see Petty & Wegcncr. 1999: Pelly et a l.. 1999. for further d iscussion).
Misullderstalldillgs of the Use of Argumellt Quality as a Methodological Tool There are several confusions regarding the use of argument quality (i.e., strong versus weak messages) in research-testi ng predictions made by the ELM. Some re,earcher~ (Mongeau & Stiff. 1993; O' Kecfe. 1990) have critici/cd the ELM for manipulating socalled strong versus weak argument!) withou t !o.pecify ing the underlying factors that make an arg ument stron g or weak. These crit icisn1!) fa il to recognile that ELM studies use argument quality primarily as a methodological tool to help differentiate the differen t ro les for vari ables. For example, if a variab le (e.g .• source expertise) produce~ the pattern in the top panel of fi gure 5.2. it suggest!o. thai the va ri ab le is serving as a simple cue as the variable increCises persuasion regard l e~s of argument qua lity. On the other hand. if a variable produce!) the pattern in the bottom panel of figure 5.2. it ~uggests that the variablc is se rving to inOu ence th e extent of information proces~ing activity (sec Petty. Wegener. Fabrigar, Priester. & Cacioppo, 1993). Other researchers have assumed that argument quality is defined strict ly in tcrms of logical quality or in terms of how likely an att itude object is: to possess some attribute (A reni & Lut z. 1988). However. as a methodological tool. manipulations of arg um e nt quality refer 10 any feat ures of the argumcn ts th;:It get people to think favorable th oughts (strong arguments) or unfavorable thoughts (weak argume nts) to the advocacy. Petty and Wegener (1991) suggested that stro ng argument!) were those that pointcd 10 highly desirable conseq ue nces that would most certainly occur if some advocacy was acceptcd. These arguments could be made weaker either by pointing to less de~irable consequences that wo uld occ ur if th e advocacy wa~ adopted or to de~irable conseque nce!>' that were less like ly to occur. That is. arguments cou ld be weClkencd by reducing either the de!o.irability or the likelihood of the co nsequences propo,cd in the argument (see also Fi!o. hbcin & Ajzen, 198 I).
Assertions That the ELM Is Not Falsifiable Some researchers have argued that the ELM's Illultiple-ro l e~ hypothesis "all ows th e ELM to ex plain all poss ible outcomes of an experimental study" (Stiff & Boster. 1987. p. 25 I). More recently, Stiff ( 1994) re mark ed: "Unti l the ELM specifics a priori the conditions under which important st imulus va ri ab les reneet central processing, a peripheral cue. or both. it wi ll remai n impos~jb l e to falsify" (p. 188). As discussed earlier. the multiple-roles
80
Part I • Preliminaries: Defillitions, Trends, (lI1d Theol'elical UIlt/l'11Iillllillg\ 5
Strong Arguments Weak Arguments
"
4
'0
::0
:; ct
3
2 L-______~----------~~---------Low
High Variable Level Expected effects when a variable serves as a positive peripheral cue
5
~::0 :;
Strong Arguments
4
-
ct 3 Weak Arguments
2 L-______~----------~~---------Low
High Variable Level Expected effects when a variable serves to enhance information FIGURE 5.2
Argumelll Quality Vasil.\' Qualllity with High or Low ["miJ'emellt
hypothe!o> is is necessary 10 unclcr..,land fully the dynamics of persuasion. However. (his postulate does not make thl: ELM devoid of a priori predictions. In fact. the ELM clearly specifies when variables take on the different roles. For example. the predictions for a person's mood and source credibility at different levels of elaboration were discussed earlier. and available research supports these predictions (see also Petty et aI., 2002). Although the ELM postulates multiple roles for variables, it does not say that any role can be assumed at any time. Thus. contrary to Stiff ' s c la im, the ELM does make II priori predictions regarding when a given variable (e.g., source atlractiveness, mood) has an impact on attitudes by different processes. For example. the ELM ho lds that variables are more likely to influence attitudes by biasing processi ng when the elaboration is high and invoking a heuristic when elaboration is low, and Lhat they are more likely to affect (he extent of thinking when it is not already constrained to be high or low (moderate elaboration). Therefore. th e ELM could be falsified, for example, if a variable produced a
Chapler 5 • The Elahorllrioll Likelihood Model of Persuasion
81
grearer bias to the ongoing information processing under low- than under high-elaboration conditions. or if simple heuristics (e.g .. "more is better") had a larger impact under high than low elaboration conditions.
Restricted Rallge of Topics Some researchers (O'Keefe, 1990; Stiff, 1994) have criticized the ELM for relying on a limited number of message topics and message arguments. Specifically, they assert that research on the ELM is confined to message topics involving comprehensive exams and wition increases. They further contend that a reliance on such a small number of topics challenges the generalizability of the ELM. We agree that confining message topics to comprehensive exams and tuition increases, while not darnaging the validity of the theory, could limit its gelleralizability. However. this criticism fenecrs a lack of appreciation of the broader literature on the ELM rather than the topics lIsed in some of the mosl widely cited studies. While it is true that initial research on the ELM involved studies using the topics of
comprehensive exams (e.g .. Petty & Cacioppo. I 984a. 1984b; Petty. Cacioppo. & Goldman. 1981; Puckett. Petty. Cacioppo. & Fisher. 1983) and tuition increases (e.g .. Cacioppo. Petty, & Morris. 1983). subsequent research has used a variety of diverse topics. A cursory review of the literature reveals experiments testing and confinning hypoth-
eses of the ELM using topics such as condom use (Helweg-Larsen & Howell. 2000). the city sales tax (Desteno et al.. 200 I). the foster care system (Petty et aI., 1993. ex peri ment 2: Wegener. Petty. & Smith. experiment I). nuclear power (Fabrigar. Priester. Petty. & Wegener. 1998. expo I; Haugtvedt & Wegener, 1994), vegetarianism (Fabrigar et al.. 1998. expo 2), and environmental conservation (Wood, Kallgren, & Priesler, 1985). In addition to these social issues. research using the ELM framework has also been conducted with a variety of advertising messages for goods and services such as answer-
ing machines (Haugtvedt & Petty, 1992). bicycles (Haugtvedt & Strathman, 1990), cameras (Laczniak & Carlson. 1989). detergent (Shavitt & Brock, 1986). low-alcohol beer (Andrews & Shimp. 1990). food additives (Haugtvedt & Petty, 1992). pens (Petty et aI., 1993. experiment I), restaurants (Shavitt, Swan, Lowery, & Wanke, 1994). shampoo (Petty & Wegener. 1998). and vitamins (Smith & Petty. 1996). The ~lbove research represents only a scant number of the diverse topics that have been studied under the framework of the ELM. Consequently. upon examination of the
breadth of the literature on the ELM, it is clear that the ELM generalizes to multiple topics beyond senior comprehensive exams and tuition increases.
Replacillg the ELM with a Sillgle-Process Model of Persuasioll Perhaps the most ambitious critique of the ELM-and other multiprocess models of social
judgment-comes from Kruglanski and Thompson (1999), who argued that a single route to persuasion was a more parsimoniolls way to account for the various persuasion findings generaled by the ELM (und the related Heuristic·Systematic Model: Chaiken. Liberman.
& Eagly. 1989). In proposing their unimodel. Kruglanski and Thompson (1999) argued that there are no qualitath>e differences between the two routes to persuasion. Instead. all
82
Pan I • Prelimillaries: Definitions, Trellt!\, allli Theoretical Ullderpinnillgs that is necessary to account for persuasion is the elaboration continuum that ranges from minimal processing on one end to maximal processing on the other. The central and peripheral routes to persuasion are then solcly a function of depth or extent of proccssing, and no qualitatively different processes operate along this continuum as specified by the ELM. To make this point. Kruglanski and Thomp.')on (1999) proposed that both cues and arguments could be conceplUaliL.ed as types of "evidcnce," In this conceptualization, all forms of evidence can fit into Kruglanski·s Lay Epistemic Theory (LET; Kruglanksi. 1989) in which evidence. when considered with its paired relevance. leads to attitude change. According to Kruglanski and Thompson. because both arguments and cues can be considered as "cvidence," there is no need or theoretical rationale to differentiate them. At first glance. the unimodel may seem attrilctive due to its parsimony. On the other hand. if it is useful to distinguish two or more qualitatively different processes of persuasion, the unimodel, though parsimonious. would not be accurate. As should be clear from our presentation earlier. the ELM highlights a continuum based on the extent of careful examination of the relevant evidence (the elaboration continuum). In facL a con~iderable amount of persuasion results can be accounted for with just this continuum. However, in contrast to the unimodel, the ELM holds that different persuasion processes operate along this continuum and that some or the!-.e processes arc qualitatively different from each other. Many variables can be viewed as "evidence," but how a pcrson processes this evidence is what determine!-. the effect that the variable has on per!-.uasion. For example. an attractive source can be input to the heuristic ;'if she likes it. so do L" in which case anything that the attractive source endorses will be more persuasive (peripheral roUle). Alternatively. the attractive source can be evaluated as an argument so that the attractive source is persuasive when attractiveness is relevant and cogent (e.g .. "if his hair looks great from using that shampoo. it must be a good product, so I'I! buy it too") but is llllpersuasive when irrelevant (e.g .. "her hair is nice but what does that have to do with this refrigerator?"). Or to return to an example we used earlier. evaluating the arguments in a messi.lge by simply counting them (i.e., lIsing the arguments as a numerical heuristic) versus carefully scrutinizing those argulllents for mcrit can lead to different persuasion outcomes (Petty & Cacioppo. I98.Jb). Though Kruglanski and Thomp,on (1999) may contend that there is no interesting qualitative difference between counting and scrutiniling messages. proponents of the ELM argue that counting and elaborating are more than simple opposite ends of a quantitative continuum: they are two distinct, qualitatively different processes that can produce fundamentally different outcomes when applied to the same evidence. Furthermore. the ELM holds that there are other qualitatively different psychological processes (c.g .. dissonance versus self-perception) that are of interest in understanding persuasion (Petty & Cacioppo. 1986a). The ELM 'pecifically postulate, that many of these processes operate at different points along the elaboration continuum and require different amounts of cognitive effort (i.e .. a quantitative difference). But the difference in cognitive effort is not the only difference among these processes. Because of their qualitative difference~. these processes often specify different mediators and moderators of persuasion effects (e.g., dissonance focuses on the presence of aversive arousal, whereas selfperception does not). This richness and predictive power is lost by focusing :-.olely on quantitative differences in amount of thought. In sum. because of its accommodation of
Chapter:=; • The Elahoration Likelihond Model oj Penlltl:;iotl
83
both qualitative and quantitative differences in persuasion processes. we believe that the ELM remain ... a more satisfactory model for accounting for persuasion effects than the unimodcl.
Directiol/s for Future Research So far we have explained the ELM and addrcs~ed various controversies that have arisen regarding the model. Before concluding thi ... chapter. we turn to a discus~ion of some curren! direction ... in persua ... ion re ... earch that have stemmed from an appreciation of the ELM.
Self- Validatioll Processes The postulate of the ELM that has received the least research atten!ion to date is the firM one. that people seek correct attitudes. This postulate has been used in two ways. First, it helps to explain why people engage in greater information processing in certain circum..,Iances than they do in other",. That i~. becau ...e careful I.,crutiny is oftcn a good way to determine correctness. people will engage in more effortful scrutiny when it is important to be correct. ~uch as when a I11c~sage has high consequcncc~ for the self (PclIY. Cacioppo. & Haugtvedt. 1992; Petty. Wheeler. & Bi7er. 2000). Second. this postulate implies that in the absence of competing motives (e.g .. consi\tency, reactance. impression management). the defauh goal is to be corrcct and to auemplto process messages in a relatively objective way. Recent rc\earch has sugge\ted another consequence of the correctness nlorive. Specifically, in "'OIl1C circum~tance~ people will reflect on the validity of their thoughts and attitudes (referred to as se/F1'(t/idarioll processes: Petty. Brinol, & Tonllala. 2(02). For c:\umplc. with n!"'pccl 10 aflirude \'a/idalioll, consider a person who has just processed a message and rejected it. Because the attitude survived an attack. the person might gain confidence in this attitude. In facl. in a seriel., of studic .... Torl11ala and Petty (in prcss) . . howed that thc ... tronger the attack people believe their attitude resisted. the more confidence Ihey gain in it. Furthermore, this enhanced confidence led people to be more willing to act on their altitudes. Thus, attitude validation processes can increase attitude strength. People may also sometimes reflect on the validity of the indh idual thoughts that they have in re~pon\e to a perl.,uasive message. For example. in a series of studies on I/u)//glir l'Cllidatiol1. Tormala. Petty. and Briiiol (in press). showed that the easicr it was for people to generate thoughts on a message. the more confidence they had in them, and the more they relied on them in forming their attitudes. In particular. Ihese inve"tigators made "ome people feel that generating fa\'orable thoughts to a me..,sage was easy. because their lask was to gcnerme only two favorable thought~ to the argumcills. Other, wcre made 10 fcel that gcnerCJting favorable thoughts was difficult becau"e eight were requested (~ee abo Schwart et al .. 1991). When it was easy to generate favorable thoughts. people had more confidence in thes..e thoughts and were more per~uaded by the message than when it felt more difficult to generate the thoughts. Conversely. when people felt it was easy to generate counterarguments to the message because only two were requcsted. Ihey were
84
Pan J • PI"l'limiliaries: J){~finitiO/I.\ . 7i"ellll.\, alld Theorelim/ Ullderpinnings
le:-. ... pcr ... uaded than when they fell it wa~ hard to generate counterargument .... Thu:-,. rc...earch on :-,elf-validation proce:-, ... e:-. ~ugge _... ts that another role for variables in per~ua5,ion ... cttillg~ i~ that thcy can help people asscs~ the validity of their thoughb and attitudes. Numerou\ variables may influence attitude ... and attitude Mrength in this way (e.g .. people
lJlay be more confident in Iheir favorable Ihoughls if Ihey were generaled in response 10 a ... ource of high rather than low credibility).
Consequences of Objective Versus Biased Thinking The ELM makes an important dhtinction between processing that is relatively objective ver ... us that which is biased. Understanding whether a variable (e.g .. source credibility. mood) i:-. enhancing objective proce~sing or imparting a bias (0 the processing is important for under ... tanding both the valence of thoughts produced and the ultimate persuasion outcomc. Recently. Rucker and Petty (2002) have discu~sed :-,ome implications of biased \er... u\ objective processing for the ... trength of the altitudes resulting from persuasion . In an initial . . eries of ... tudie~ on this i....... ue. Rucker and Petty (2002) compared . . ituation ... in which indi\iduah were in..,tructcd to process a message in a relatively objective manner (try to generate thought..,) or in a bia~cd manner (try to genefilte ncgative
Ihough,,). These participanls were exposed 10 a message favoring a brand of aspirin Ihal t'ontained very strong argument ..... Although both the objective and biased group~ ~howed equivalent attitudc change to these ... tTOng argumen ..... and equivalent amounts of cogniti\'e effort in processing the message. individuab who had focused on linding fault with the Illc~~age reported more certainty in their changed attitudes. Furthermore. the attitude\ of individuab who had tried but failed to find fault were more predictive of subsequent behavioral iJltelltion~ toward the aspirin. Rucker and Petty (2002) argued that people who tried to find fault but failed were more coglliL.ant of the fact that the aspirin had no falllt~ than wcre people \\ho simply proces\ed the message objectively. The lauer group wa~ t'ogni/ant mostly of the favontble aspects of the a~pirin. Subsequent analyse~ confirmed that perceptions about the lack of negative features of the aspirin mediated the increa:-,ed t'crtainty in the changed auiwdes. Thi~ research :-.hows that qualitatively different approache . . to proce:-,sing the ll1e~sagc can lead to differences in attitude !o.trength even when the cxtCI1t of mes\age elaboration appears to be conMant.
Conclllsioll Thi~
chapter began by noting the chaos in altitude and persua~ion research reflected by numerou ... conflicting findings in the litentture and then focused on explaining how the ELM can organi/c past inconsistent finding . . in a manner that allow~ researchers to predict when a variable might have a given effect. anti when a particular process respon ... ible for that effect might occur. Furthermore. the chapter has provided representative examples of research ba\cd on the ELM framework throughout. Finally. in addition to stressing the utility of the model in explaining pa . . t conundrums in the literature. thi\ chapter has focused on clarifying misconception~ about the ELM and pointing to direction!) for future rc ... earch. In particuliJr. recent work on people· . . as ... essl1lent of the vulidity of their thoughts
85
Chapter 5 • JIll' 1,!a!"mIlUJI/ UkdiJ/Ood Modl'l (~I Pl'nt/wioll
and altitude ... ha ... pro\ided ne\\ IllCi.lIl ... by \\ hlch \ariabk: ... can affect pcr ... utl . . ion (by affecting thought confidence) and altituue ... trcngth (hy affecting altitude confide ned.
Notes ________________________________________________________ I, Although not di . . cu ......ed in Ihl' eIl.lpter, olher IIltxich uf pcr~ua<.,illn and ,ocial Judgment havc hccn dc\clopetl that :-.hare a number of the fcature, 01 thc ELM, A di"-Cu ... ,iol1 01 the'c moJcl ... i. . oc)'om.lthe <.,copc of thl' Ch:Lptcr. but the intcrc'tcd n.:adcr i... rc:h,,:rrcd to :L compendium of dual-procc" model<.. 01 <"(lcLal Judgment LIl Chailo.en and Trope (1999) ami il compari<"lln of the ELi\\ with ,omc of thc,e modch hy Pell) and Wegcner (1998. 1999) and PClly, Fahrigilr. and Wcgener I in pre,,).
Rej erell ces'_______________________________ Andrc .... '. J C .. &. Shl mp. T. A. ( Il)l)()), I Jh..'l·t ... of in \u\\ emcnt. argument ... trcngth .•1111.1 'llurl'C charactcri,· tlc<., on centml ami peripheral p1'lK'C'\llIg . I'lnlio!ol.p am! Mllrkl'li"g. 7, 195 21-l Areni. C. S.. & LUll. R. J. (19RR), The rols.: 01 argument qualit) In the ElahoratH)f} Lil..clihnod \!lodd Adl'ill/( 'I i" Om,llI/lIl'/' Ni'H'anll. 15. 197 · 20.'. An'totlc, I (9)-l). RlleluriC', In W Rohen'ltran"', ), \n.IW1W, rlU'/oril" (//U/ I'(ll'/il".l. I\c ..... Yorl.. : Modern Lihrary BC1Tl. I), J (197:!1. Sdf-pen:cplioll theory In I BerI..U\\lt/ (1.<.1),\(/\'(1//('('" illl',\/Il'rIll/l'//It1/ \lIcit/IIJI\ dl%.l:\' (\01. 6. pp. 1-62). \lc\\. Yorl..: Acadcnm: Prc ..... 130\\ cr. G. I 19X I ). Mood and memory \lIIl'1"in/1/ " ., ,('/I(I/o.l,:i,'/, 36. 129-1-l8, Car.:lOpJ)(l. J T .. & PeIlY. R. E. (1979). File!:" ulllle .... age rcpctllJOll and J)(Nuon un wgnltl\c rc'pon,c<". rCl·;tll. and ~r.ua,i()n. JOIIl'llol (lj Pl'rHll/tIlil,1' (/1/(1 Sona/ P,n'/llIlog\, 37. 97- \09 CIl'I(lPpO. J T. PCII). R. E.. Kao. C.. & Rudngue/. R (19X6). Ccntral .lIld pcnpherJ.l route ... tn pcr... u'l,ion An l11di\ldual dil'ferencc pcr'pe!:II\I.". j01l/"l/(/1 (~/ Palll/Wli/r tllIll So('/al PnlltoIOl{'. 51. I(H2
10-1.1. Cal:ioppo. J, T .. PeIlY. R. E .. & Morri,. K. J, (19S3). ErIe!:t ... 01 neeu far cognition nn me ..... agc I.!\aluatmll. recall. ,tllli per'lIa .. ion. JO/lI'II(/f (~r 11n,I(///(/lilr tllld Sol'ialll,H'c/lOlogr, 45, X05--81 X, Caldcr, B, J.. In ... l..o. C. A .. & Y'l1ldl.!ll, B. (1974). The rel:ltilln 01 cognition ;.lIld ml."morial proec ......e' to pcNI:I,ioll 111 a <.,imulatcdjllry trial jot/mal o/,\I'I,/il'd ,')ocia/ PHdwlo,I,:\', 4, 62 ·93. Ch;ul..ell. S, (1981)). HClIri,ric \er,u, 'y'tematK inlurmatlon pro();:"lllg III the lI'C 01 ...ourl'e \cr,u, me ... • 'CNLa,ion, In
or
86
ParI I •
Prelilllillarie,~:
D('/illitiom', Trel/ds, alld Theoretical UII(/e'1Jillllillg,\
R. E. PCtl)'. T. M. O .. trom. &. T. C I3md, lEd .... ). Cogllliin' respollSes '" penlllaio" (pp . :n9- .'\59). 1-J1l1 ...dalc. 1'\J: Erlbnum . Forga ... J. P. (1995). Mood ~md Judgment : The aflect infU",lun model (AIM). Pn'{"/lO/oKim/ Blil/Nill, 117. 39-- 66.
Friedridl. L Fether"lonhaugh. D., C3"C),. S .. & Gallaher. D (1996). Argument integration and atlllude chang!!: Supprc~ .. ion t.!ffcCI\ in lhe 11lI1!~rali()n 01 o!lc · .. idcd argument ... Ihal vary in pcr... ua .. l\cnc ..... P\ych%gical Blil/l'lllI. 22.179- 191 Greenwald, A. G. (1968). Cognitive tearnlng. cognili\c rc"pon ...c to per'..ua"'lon. and atlitu(.lc change. In A. G. Greenwald. T C. Brock. & T. M O .. trom (Ed ... ). Pn-cilOlogil'{lljolilu/ari(ms o/atllllltit· , (pp. 147- 170). New Yor!..:: Academic Prev,. Greenwald. A. G .. & Roni .... D. L. (1979). T~cnty year,> of cognili\e dissonance: ('a,e ..,lUdy of Ihc c\oluliun of a theory. Psvc/IO/ogica/ Rel·ie\\". 85. 53- 57. lIafer. C. L. Reynold,. K.. & Ohcrtyn..,1..1. M A. ( 19961. Mc"age comprehelhibility and per~ua ... on : Efft.!Ch of complex languagt.! in countcrallilUdinal appeal" to laypeoplc. Soda/ Cognition. /-I, 317337. Harmon-Jones. E.. & Mill,. 1. (Et!...). (19lJ9). Cognitive di" .. omUlce: Progres, on a pivotal theory in ,()Cial p"'ychology. Wa .. hinglon. DC: American P'ychological A,>,.;ocimion. Haugn.cdt, C. 1' .• & Pelt). R. E. (1992) Pcr,omllit) and per... ua,ion: Need for cogmtion mooeratc, the pcr... l... tence and n!,i'tancr: of atllludc changc ... jOllmll/ of PersOIwlin alld Socia/ Pn-dlOlogr. 63. 10K- 319. Hec,ackcr. M. H .. Petty, R. E .. & Cacioppo. J. T. (19H3). Field dependence and attitude change : Source cn.:dihility can alter pcr,ua,ion h) allc:cling mc\\agl>rcJc\ant thinking. journal of Per.\olllllif.\·, 5/.
653-666. Helwcg-I...ar-cn. t\L & Howell. C. f ~()()O). Ellcch of crotophobla on the pef',ua,j\ ene ..... of condom ad"crtl ...ement., conttlining .. trong or \\c.lk argulllcnt .... B(lIi(" amI Applied Social Pn·c/lOlo.l.p , ]]. 1 I 1117. Himmdfarb. S .. & Eagly. A. II. (Eds.). (1974). R('(ulillgl ill (It/itl/de challge. New York: Wiley. Hovland. C. I.. Janis. I. L.. & Kelley. II. II. (1953). Co/ll11/lmitatioll wul pi' o//(n;,,,,: P\\'d/OI()~i("(lI\lI/(I il'.1 of0l'illiofl £"IllIl/gl'. New Haven. CT: Yale Univcr .. ity Pre"s. Hovland, C. I.. LlIImdainc. A. A .. & Shclflcld. F. D. (1949). E.\perimelll.1 OIl man (·Olllllllllliwtiol/. Pnnceton. NJ: Princeton Univcr:o.ity Pre ... , . Il0vlantl, C. I.. & Wei ... ,. W . ( 195 I ). The Intlucm:c of .,oun:e cn.::dibility on communication effectl"enc ... '. Pl/hUI ' Opinioll QII(lnl'l"I\'. 15, 6.15 650. Kahnernan. D. (1973). Aflellliolllllld eflort. Englewood Cliff.... NJ: Prentice-I·lall. Kelman ( 1961 ). Proce .. "c, of opinion change . Puhlic Opilliol/ Quarterly. 25. 57· 78. Kruglan .. ki. A. W .. & Freund, T (1981). The Ireeling. and unfreezing of lay-inference .. : Elfech of Impre ..... ional primacy . cthnlc ,tcreotyplng. and numerical anchonng. journal of £\"pt' flml'lIIal So· cial PH'clloloRY, /9, +H~--t68. Kruglml'~ i. A, w. ( 1989). L(lr ephtell/in o//(/ 1111/11(/11 klloll/('(lgt': Cogllitil'e (III(/ motiwliolla/ haw' f . New York : Plcnum Pre ..... Krug.Jan,ki. A. W .. & ThomlNm. E. P. (](1)t)). Pcr... ua ... iun h) a single roule: A \iew from thl! unil1lodcl. Pn'c/w/ogiuli IlIl/l/in, /0.83- 109. Lacznial... R. N .. & Carl ..on. L. (1989). El.::lmlning the intluenc.:e of attitude·tow.ard-lhe-ad on orand atli· tude .... journal oj !J/I\;III'.H Re.H'lIrch. 19, 303- 311. Maio. G. R .. Bell, D. W .. & E... "e', V. M. (191.)6). Ambivalencc and per-Ui.\sinn: The proce,,,ing of mc~ "age, about immigrant group... jOl/l'1/a/ (If L".werill/(,'ltlll S(ldal Psychology. 32. 5 13- 536. Martin. L. L.. Abend. T. A .. Sedi!..:idc,. C. & Green. J . (1997). How would I feel if . . . ? Mood a" WpUI to .1 role fulfillment e,"aluation procl!"". jmmwl of Pennl/alif.\" (llId SOl"lal PWc-/lOlogy. ~42253. Mongl!ilU. P. A .. & Sliff. J. B. f 1993). Spcr.:d·Ylng r.::lu,>tll rcluliUlI,>hip" in Ihe cl .. boralion likcliho<xt model. COII/mullicatioll Theon'. J, 65- 72. Moore. D. L.. Hauskncr.:ht. D .. & Thamodaran, K. (1986). Timl! compres ... ion. rc'pon .. c opportunity. and pcr"uasion. JOI/I'1/(// (!{COIISl/IIWI" R(' I'I'al"ch, /3,85- 99. Norman. R. ( 1976). When whal i.... aid is imporlant. A compari"on of expcn and allraclive ...ourcc!.. jour""I of £rperime",a/ Social Pnc/wlogr. 12, 295- 300.
n,
Chapter 5 • The Elaborarion Likelihood Model of Persuasion
87
O·Keefe. D. J. (1990). Persuasion: Theory and research. Newbury Park, CA: Sage. Petty, R. E. (1997). The evolution of theory and research in social psychology. From single 10 multiple effect and process models of persuasion. In C. McGany & S. A. Haslam (Eds.), The message of social psychology: Perspectives on mind in society (pp. 268-290). Oxford. England: Basil Blackwell. Petty, R. E., Brinol, P .. & Tormala. Z. L. (2002). Thought confidence as a determinant of persuasion: The self-validation hypothesis. Journal of Personality and Social Psychology, 82, 722- 741 . Petty. R. E.. & Cacioppo, J. T. (1979). Issue-involvemem can increase or decrease persuasion by enhancing message-relevant cognitive responses. Journal of Personality and Social Psychology, 37. 1915- 1926. Petty, R. E" & Cacioppo, J. T. (1984a). Source factors and the elaboration likelihood model of persuasion. Advances in Consumer Research, I I. 668-672. Petty. R. E.. & Cacioppo. J. T. (1984b). Thc cffects of involvement on response to argument quantity and quality: Central and peripheral routes to persuasion. Journal of Personality and Social Psychology, 46,69-81. Petty, R. E .. & Cacioppo, J. T. (1986a). Communication and persuasion: Central and peripheral rOUTes 10 attitude challge. New York: Springer-Verlag. Petty. R. E .. & Cacioppo. J. T. (1986b). The elaboration likelihood model of persuasion. In L. Berkowitz (Ed.), Advances ill experimental social psychology (vol. 19, pp. 123- 205). New York: Academic Press. Petty, R. E. , & Cacioppo. J. T. (1990). Involvement and persuasion: Tradition versus integration. Psychological Bulletin. 107,367- 374. Petty. R. E., Cacioppo, J. T., & Goldman. R. (1981). Personal involvement as a determinant of argumentbased persuasion. Journal of Personality and Social Psychology. 41. 847- 855. Petty, R. E., Cacioppo, J, T., & Haugtvedt. C. (1992). Involvement and persuasion: An appreciative look at the Sherifs' contribution to the study of self-relevance and attitude change. In D. Granberg & G. Sarup (Eds.), Social judgmem and i1lfergroup relatiol1s: Essays in honor of Mu::.afer Sherif (pp. 147- 174). New York: Springcr- Verlag. Petty, R. E .. Cacioppo, J. T.. & Schumann, D. W. (1983). Central and peripheral routes to advertising effectiveness: The moderating role of involvement. Journal of Consumer Research, 10, 135- 146. Petty, R. E., DeSteno, D., & Rucker, D. (2001). The role of affect in persuasion and attitude change. In J. Forgas (Ed.). Handbook of affect and social rognitiofl (pp. 212- 233). Mahwah. NJ: Erlbaum. Petty. R. E .. Fabrigar, L. R.o & Wegener. D. T. (in press). Emotional factors in attitudes and persuasion. In R. J. Davidson, H. H, Goldsmith, & K, R. Scherer (Eds,), Handbook of affective sciences. Cambridge: Cambridge University Press. Petty, R. E .. Harkins. S. G .. & Williams, K. D. (1980). The effects of group diffusion or cognitive effon on attitudes. An information processing view. Journal of Perso1lality and Social P.~ychology, 38. 81-92. Petty, R. E .. Haugtvedt, c.. & Smith, S. M. (1995) . Elaboration as a determinant of attitude strength: Creating attitudes that are persistent, resistant. and predictive or behavior. In R. E. Petty & J. A. Krosnick (Eds.), Attitude strength: Anfecedents {If1d comeqlleflce.\· (pp. 93-130). Mahwah, NJ: Erlbaum. Petty, R. E., Jarvis, W. B. G., & Evans, L M. (1996). Recurrent thought: Implications for attitudes and persuasion. In R. S. Wyer (Ed.), Advances ill social cognition (vol. 9, pp. 145- 164). Mahwah, NJ: Erlbaum. Petty, R. E .. Kasmer, J. A .. Haugtvedl. C. P., & Cacioppo. J. T. (1987). Source and message factors in persuasion: A reply to Stiffs critique of the elaboration likelihood model. Communication Monographs. 54, 232- 249. Petty, R. E., Ostrom. T. M .. & Brock, T. C. (Eds.). (1981). Cognilil'e responses in persuasion. Hillst.lale. NJ: Erlbaum. Petty. R. E.. Priester. 1. R.. & Brino!. P. (2002). Mass media and attitude change: Advances in the ELM. In J. Bryant & D. Zillmann (Eds.). Media effecrs: Advance.\· in theory and re.5earch (2nd ed .. pp. 155198). Hillsdale. NJ: Erlbaum. Petty, R. E., Schumann. D. W" Richman. S. A .. & Strathman, A. J. (1993). Positive mood and persuasion:
88
Part I • Preliminarie!i: Definitions. Trends. and Theoretical Uf/derpil/llill!(s Different roles for affect under high- and low-elaboration conditions, jOllrnal of Personality allli Social Psychology, rH, 5-20. Pelty. R. E.. Tomala. Z .. & Rucker (in press). Resistance to per,,>ualiion: An attitude ~Irength per<"pcclivc. In M. R. 8anaji. J. T. Jost. & D. Prentice (Eds.). The yill and yang of :wcial cognition: Fesu'hriftfor William J. McGuire. Wa'lhinglon , DC: American P~ychological Association. Petty, R. E., Wells. G. L. , & Brock, T. C. (1976). Distraction can enhance or reduce yielding to propaganda: Thought di~ruption versus effort justification. jOlll"l/a/ of Personality and Social Psvclw/0RY. 34.874-884. Petty, R. E., & Wegener, D. T. ( 1991 ), Thought systems, argument quality. and persuasion. In R. S. Wyer & 1'. K. Srull (Eds.), Adl'(l/lce.\ ill social cONl/irion (vol. 4. pp. 147- 161). Hillsdale. NJ: Erlbuum. Petty. R. E .. & Wegener. D. T. (1993). Flexible correction processes in social judgment: Corrccting for context-induced contrast. JOIi/"flal of Experimental Social f>~ycl/Ology. 29. 137- 165. PClty, R. E.. & Wegener. D. T. (1998). Auitude change: Multiple roles for pcr~uasion variable .... In D. 1'. Gilbert, S. T. Fiske. & G. LindLey (Eds.). The handbook of social pHdlOlogy (·trh ed .. \01. I. pp. 323-390). New York: McGraw-Hill. Petty. R. E .. & Wegener. D. T. (1999). The elaboration likelihood model; current status and controver... it.:s. in S. Chaiken. & Y. Trope (Eds.). Dual-process lhear;t'!!. i" Jo("icli p.\ycllOloXY (pp. 41 -72). New York: Guilford Press. Petty. R. E., Wegener. D. T.. Fabrigar. L. R.. Priester. J. R.. & Cacioppo, J. T. (1993). Conceptual and methodological issues in th!.! Elaboration Likelihood Model of persuasion: A reply to the Michigan State critics. Communicatio/l Theory. 3. 336-362. Pelty, R. E., Wegener. O. T.. & White. P. ( 1998). Flexible correction processes in social judgment: lrnpli cations for persuasion. Social COXl1ilioll. /6.93-113. Petty. R. E.. Wheeler. S. C. & Biler. G. Y. (1999). Is there one pcrsua\;ion process or marc? Lumping versus splitting in attitude chi.lnge theories. PsycllOlogical bUlll/n·, 10(2). 156-163. Pelly. R. E.. Wheeler, S. C, & Bizer. G. (2000). Matching effects in pcrsU
Chapter 5 • The Elaboration Likelihood Model of Persuasion
89
Stifr. J. B. (1986). Cognitive proces<,ing of persUilsivc message cues: A meta-analytic rc\iew of thc effects of supporting information on attilUdes. Commullicatioll Monographs, 53, 75-89. StilT. J. B. (1994). Persuasil'e Communicatioll. New York: Guilford Press. Stiff. J. B .. & Boster. F. J. (1987). Cognitive processing: Additional thoughts. and a reply to rellY. Kasmer. Haugl\edl. an(] Cacioppo. Communication MO/lOgraphf. 54, 250-256. Tormala. Z. L. & Petty. R. E. (in press). What doesn't J...ill me maJ...e<, me stronger: The effecb of resisting persu:l,ion 011 attitude cenainty. journal of Per.l(Jfwliry alld Social Psychology. Tormala. Z. L .. Petty. R. E.. & Brino!. P. (in press). Ease of retrieval effects in persuasion: A selfvalidation analysis. Penol/ality mId Social Psychology Bul/et;I/. Wegener. D. T .. & Pcuy. R. E. (1994). Mood-management across affective states: The hedonic contingency hypothesis. lOl/mal of Personality lind Social Psychology, 66. 1034-1048. Wegener. D. T., & Petty. R. E. (1995). Flexible correction processe... in social judgment: The role of naNe theories in corrections for perccived bias. jUllmalof Per.wl/ulit.\' & Social PlrcllOlogy, 68, 36--51. Wegener. D. T .. Petty. R. E.. & Klein. D. J. (199..1-). Effects of mood on high elaboration altitude change: The mediating role of likelihood judgments. Europe"" jOllrllal of Social Psycholo!?y. 24. 25-43. Wegener. D. T .. Pctty. R. E.. & Smith. S. M. (1995). Pm.itive mood can increasc or decrea~e message scrutiny: The he(]onic contingency view of mood and message processing. 101lma/ of PersOIwfjry wul Social P,\~\"cI/()logv. 69. 5-15. Wilo;on. T. D .. & Brekke. N. (1994). Mcntal contamination and mental correction: Unwanted influences onjudgmcnts and evaluations. Plyclwlogical Blllletin. 116, 117-142. Wood. W .. Kallgren. C. A .. & Preisler. R. M. (1985). Access to attitude-rclcvant information in memory as a determinant of per'luasion: The role of message attributes. Journal of Experimel/tal Social P.'1Ycll%gy. 2 I, 73-85. lanna. M. P.. Kil!sler. C. A .. & Pilkonis. P. A. (1970). Po,>ilive and negative attitudinal affect establi~hed by clas ... ical conditioning. lOIln/al of Per.w"alir.\' and Social P.\·ycJlOlogr. /4, 321-328.
j
Part
II Persuasion Variables Perspectives on Sources, Receivers, Channels, and Messages
Chapter!o. 3 and 5 of this volume mentioned Carl Hov land. a
p~ychologi\1
from Yale Uni-
ve rsity who is often credited wi th initiating the r.,ystcmatic .... oeial scientific study of persuasion during World War II. After the war. Hovland and his colleagues (Hovland. Janis, & Ke ll ey. 1953) continued to develop and investigate their "mc!"sage learning approach." which among other thing\ made it clear lhal per<.,uasion i"i no simple process. Their approach suggested that in order to be effective. a persuasive message had to capture the audience's attention. be comprehended. be yie lded to. and be remembered. h also suggested that this process could be facilitated with incentives for agreeing with the persuasive message. More pertinent to this pan or the book. however. are the moderating variab les cXamined by Hov land and his group. Moderating variables can be thought or as the "it depends" of persuasion. In other word,. whether persuasion b errective or not often depends on several ractors. As just one example. whether the use or touch tcnds to facilitate or inhibit persuas ion depcnds, in part. on whether the touch is perceived as a positive or negat ive vio lation of the recipient's perception..,. Hovland and his colleagues focused their attention primarily on four variab les that might moderate the procc!>.s or persua1-.ion: the characteristics or the person sending the persuasivc message (the source). the nature of the mes!>.age itse lf'. the c hannel or medium by which the message is sent. and the characte ristics of the person receiving the persuasive message. Likewise. the chaptcr!o> in this scction d iscuss thc rolc or communicator characteristic!>.. I1lc!-.sages. and channe l!'. in the proce!'.s of persuas ion. T he chap l er~ not only examine the role these variables play in effective pcrsuasion. as Hov land and his colleagues Liid. they abo discuss the ways in which such variab les inlluence the prod uctio n of per,u
91
92
Part " • Pl'rslw.\iol/ Variab/n: Per,\ lN'Clil'cs OIl SOl/rCl'S, Rcccil 'ers. C"alllll'ls, alld Messa!!,!'s
There are, of course, several source charaCI(Tistics that might affect persuasion. such as how likable, attractive. or similar 10 the audience a source is perceived 10 be. As impor ~ tant as such characteristic~ are, however. they have not received nearly as much research attention as the topic of credibility. which is without doubt one of the most thoroughly studied topics in the field of persuasion. Some of this research has attempted to discover the underlying dimensions that make up credibility (e.g .. trustworthiness, competence. charisma. etc.), The first chapter in this section. chapter 6. hy William Benoit and Alan Strathman. examines such research and expands it to broaden our understanding of how and when credibility works persuas ively . The chapter does an excellent job of showing how theory (e.g .. the Elaboration Likelihood Model) helps us predict the conditions under which credibility mediates persuasion. It is also important to note that this chapter dis+ cusses the ways in which a tarnished image may be repaired. In addition to source credibility, other cOllllllunicator characteristics mediate the process of persuasion. An enormous body of research, for example. has examined the ways in which demographic variables (e.g .. age, etilnicity, intelligence) and personality traits (e.g .. self-monitoring, ego-involvement, dogmatism) influence both the sending and receiving of messages. Given space considerations, however. this part of the book focuses on just three such characteristics. First, chapter 7, by Andrew Rancer. discusses two communication traits. argumentativeness and verbal aggrcssivcness. that affect people's orientations toward argumentative encounters. As you will see. individuals who possess one trait or the other have vastly different approaches when trying 10 persuade others or when being confronted with the inlluence attempts of others (e.g .. others' attempts to resist persuasion). Similarly. chapler 8. by Linda Carli, examines the ways in which gender affects persuasion. While the lion's .;;hare of previous research on this topic has concentrated on gender and persuasibility, chapter X asks whether onc gender is more influentiallhan the other. and if so, why. The last two chapters in this ~cction examine whether messages and challileis affect persuasion. A channel. of course. is a medium for communication. Two channels for persuasive mcssages are language and nonverbal communication. The persuasive potential of language is well known. Most of us, for example, are probably aware of situations in which euphel11i ... m~ and politically correct language arc necessary. Moreover, research tells us that using profanity or "powerless language" (e.g .. "uh," "you know") Can have damaging effects on credibility (Bostrom. Baseheart. & Ros!-.iter. 1973: Haleta. 1996). While such issues arc important, we believe that when trying to understand the role of language in persuasion. perhaps the most significant factor to keep in mind is the old maxim, "Meanings are in people, not in words." This principle is one of the key tenets of Language Expectancy Theory, which is the subject of chapter 9, by Michael Burgoon and Jason Siege\. Indeed, the specific words a speaker uses when attempting to pers uade an audience may be importal11. but not nearly so much as the way in which the !-.peaker's audience perceil 'es those words. This chapter. as you will see, suggests that understanding audiences' expectations about language usage is crucial 10 predicting how persuasive language will be. Finally. chapter 10. by Peter Andersen. examines nonverbal influence. Interestingly. some definition:.. of persuasion suggest that a chapter on this topic might not belong in a book such as this. A number of authors (Larson, 2001: Miller. 1980; Reardon, 1981:
Part " • Persuasio" Voriables: Penpe('lil'e.'i 011 Sources, Receh'en,
ClUlllllds,
alld Me ....\age.'
93
Roloff & Miller. 1980) have argued that per~uasion inv o lvc~ only ~ymholic expression, including language and other meaning-laden act~ ~uch at, civil disobedience and prolest marches. Gerald Miller (1980). for example. asserted that "persuasion rel ies upon sy mbolic transactions ... the sc holarly endeavors of persuasion researchers-and for that mat ter, the ordi nary language usages of the term 'pcrsuasion' - have consistently centered on the manipulation of symbols" (pp. 14-15). Similary. Cooper and Nothstinc (1992) argued. "Persuasion is the process by which language and symbolic actions innuence choiccmaking by others" (p. 2). Finally. according to Larson (2001). "Persua",ion is the cocreation of a state of identification or alignment between a source and a receiver that resuhs from the use of symbols" (p. 10). Though we understand the point or such definitions. we do not agree with them. In fact. we have argued elsewhere that limiting the study of persuasion to words or symbols leaves out too much (see Gass & Seitcr. 2003). We hope that aftcr reading chapter 9, you will agree. In our opinion this chapter demon strates that so me of the 1110st intriguing aspects of persuasion can be found in nonverbal behavior, which lie~ on the periphery or symbo li c action. While some lexls' di sc llssions of nonverbal persuasion are largely athcoretical in nature. wc believe thai this chapter does an excellent job of eX<1mining models and theories that explain why non\erbal behavior can be so intlut.!nlial.
Referellces_____________________________ Bostrom. R. N .. Basehearl, J. R .. & Ro . . sitcr. C. M. (1973). The c ffer.:t ... of three type ... of profane language in pcr<.,uasivt: messages. JOII,."al of C0I1I1mmicalioll. 50.415-420. Cooper. M .. and Noth<;tine. W. L. (1992). Po,,'t'r penulHiulI: MOIing till (Illdelll an illlo filt' media (lge. Greenwood. IN: Educational Video Group. Gass, R. 1-1 .• & Seiler, J. S, (2003). Persl/asio/l, social illpllence. lind colllpliallce gainillg (2nd cd.). Bo ... · ton: Allyn & Bar.:on. Haleta. L. L. (1996). Student perception-. of teacher. . · u ...e of language: The effect ... of powerful and powerless langunge on impress ion fomlation and uncertainty. Comlllllllicaliull EdliclIfirm, .J5( I ). 16 -28. Hovland. C. I.. Jani<." I. L.. & Ke ll ey, J. J. (1953). COlllmlllliclIfioll ./Ild pa_HUI'ii(m. t\C\\ ]Iaven. CT: Yale University Press. Larson. C. U. (200 I). Persuasion : Receplioll and respolIsibilifY (9th cd.). Relmont. C A: Wad<;worth. Millcr, G. R. (1980). On being pcr"uaded: Somc ba~ic di ... tinclion ~. ln M. E. Roloff & G . R. Millcr (E("'.). Persuw,'io//: Nell' direClio/1!j ill theory wltl re"earch (pp. 11 -28). Bevcrly Hil].., . C A: Sage. Reardon. K. K. (1981 ). Per.mmio,,; Theory a/ld CO/lleXI. Bc\'crly tillb. CA: Sage. Rolo fr. M. E.. & Miller. G. R. (1980). Per~/uHi(}//; Nt,\\, directions illlll('(lr.\' alltl rew'oreh, Be,crly Hil]..,. CA: Sage.
6 Source Credibility and the Elaboration Likelihood Model William L. Benoit and Alan Strathman
For thou~ands of years scholars who !'Itudy persua!->ion have recognized thai some message sources are more pcro;;uasive thi.J1l others. The ancient Greek philosopher Aristotle. for ex-
ample, explained in the fourth century B.C.: "We believe good men more fully and more readily than others: this is true generally whatever the question is. and absolutely where exact certainty is impossible and opinions divided" (1954. 1356a6-8). Isocrates, a conlemporary of ArislOtie. was known morc as an educator than a philosopher. and Cicero declared that from Isocrates' ··school. as from the Horse of Troy. none but leaders emerged" (1942. 11.9-1). Isocrates indicated his belief in the importance of credible sources when he asked. "Who does not know that word~ carry a greatcr conviction when spoken by men of good repute than when spoken by men who live under a cloud. and that the argument which is made by a man's lifc is of more weight than that which is furnished by wordsT ( 1976. p. 278). In the twentieth century several literature reviews have conc luded that so urce credibility is an important element in persuasion (Anderson & Clevenger. 1963~ Benoit. 1998: Hass. 1991: Littlejohn. 1971). Petty and Cacioppo ( 1981 a) wrote. "The expertise of the source of a message is one of the most important features of the persuasion situation and one of the earliest variable!-. to be investigated. It remains, however. one of the least understood manipulations" (p. 235). This chapter is devoted to explaining how and when ~ourcc credibility influences persuasion. Me!-.Mlge sources have multiple dimensions. including the source's physical attractiveness (see. e.g .. Bersheid & Walstcr. 1974: Chaiken. 1979: DeBono & Harnish. 1988 ; Kahle & Homer. 19~5). similarity to the audience (see, e.g .. Berscheid, 1985; Simons. Berkowitz, & Moyer. 1970). and other demographic factors. The two principal elements of .wurce credibility are traditionally considered to be expertise (the level of the source's knowledge or the topic of the mes;age. typically established by education . training. or experience in the field) and trustworthiness (whether the source can be expected to provide
95
96
ParI II • Persuasion Variables: Penpectil'e\
Oil
Sources, Nl'ceil'er.\', Challne/s, (///(/ MesJrlgl>,\'
an objective or unbiased perspective on the topic). Wilson and Sherrell's (1993) Illetaanalysis found that the effect of expertise on persuasion is greater than the effect of trustworthiness, attractiveness. or . . illlilarity. Fewer of the srudies they reviewed manipulated trustworthiness. but the effect of trustworthine . . s aha appeared to be stronger than that of attractiveness or simi larit y. McCroskey and Teven (1999) argued that thcre are three dimensions of credibi lity: expertise. trm,tworthiness. and goodwill. However. most research has focused on the first two dimensions. ACl:ordingly. this chapter will focus on expertise and trustworthiness. The Elaboration Likelihood Model (ELM) will be employed as a theoretical framework for understanding the nature of source credibility effects in persuasion. After elucidating the ELM. we lise it to explain how the nature of a messagc's source intluences the process of persuasion. We then take up several topics explaining how (and when) ... ource credibility affects attitude change. Next. we evaluate the rl!search on ...ource credibility. and finally. we discuss options for repairing damaged credibility.
The Nature of the Elaboration Likelihood Model The cognitive response Illodel (Greenwald, 1968: Pcrloff & Brock. 1980) and the ELM specifically (Petty & Cacioppo, 1981a, I986a. 1986b) portray auditors as active participants in persuasion (see also Chaiken, 1980). Receivers produce cognitions (thought~. elaborations) in response to the stimulus of persuasive discourse; attitude change does not result direc(ly from persuasive messages bur arises from (he thoughts of audience members when they are exposed to persuasive messages. Petty and Cacioppo (f986a. 1986b) posit two "routes" to persuasion: central and peripheral. The eelllral route consists of thoughtful consideration of the arguments (ideas. content) in the message and is adopted when a receiver has both motil,lItiOIl and ability to think about the message and its topic. The peripheral rollle is llsed when the listener bases reaction 10 the message 011 cue:-- other than message content. such as attractiveness of the source. the number of arguments in the message. or the length of the message. This route is adapted when (he auditor is unable and/or unwilling to engage in much thought on the message. The ELM is important in part because it enjoys considerable empirical support: The a . . sumption that systematic or central route pn.1Ccs ... ing require . . motivation and ability has been documcnlcd in many stlldics. u... ing a variety of motivational and ability variables: Persuasive argumentation is a more importnnt dcterminnnt of persuasion when recipient!. are motivated and able to procev~ attitude-relevant information than whcn they arc not. There is also ~ubstantial empirical support for the hypothe ... is of Ihe~e mode!... that heuristic or peripheral cues exert a siza ble pcrsua~ivc impact when motivation or ability for argument proce ... l,ing is low. bUllittle impact when motivation and ability are high. (Eagly & Chaiken, 1993. p. 333) This distinction between central and peripheral procl:ssing is important because these two routes are asserted to have different effect"i on receivers: "Attitude changes that resuil mostly from processing issue-relevant arguments (central route) will show greater temporal persistence. greater prediction of behavior. and greater resistance to coulltcrpcrsu
Ch::lpter 6 • Source Credibility
(lI/(J
the Elaboratiol/ Likelihood Model
97
than attitude changes that result mostly from peripheral cues" (Petty & Cacioppo. 1986a. p. 2 1). Attitude change can occur via either process. but the route (central or peripheral) makes a difference. Petty and Cacioppo selected a netlle~ome metaphor in describing these as "route!<," to pcrsua<;ion. Ordinarily. when there are two routes to a destination. only one is taken. A person traveling from St. Louis to Detroit could take two routes (via Chicago or Indianapolis) but would not take both routes on the same trip. However, "central" and "periphera l" processing are not two dichotomous choices but two endpoints of a continuum of message processing. Different people listening 10 the same persuasi\e message might engage in var) ing amounts of central and peripheral proce . . sing (depending upon each receiver's motivation and ability), So the metaphor of two "routes" is in some respects unfortunate, because it implies that auditors will take one route or the other. whereas cognitive proces~ing can actually occur at any point on the elaboration continuum. Petty and Cacioppo (1986a) do acknowledge that elaboration is a continuum: "We view the extent of elaboration received by a message as a continuum going from no thought about the issue-relevant information presented. to complete elaboration of every argument" (p. 8). According to the ELM, the key to persuasion is understanding the thoughts about. responses to. or elaboration, of a message. Cognitions vary on many dimensions. but two arc particularly important: valence and amount. Fir ... t, thoughts can agree or disagree with the message (or be irrelevant to the message). Agreeing. po..,itive, or favorable thoughto;, facilitate per..,uasion: disagreeing. negative. or unfavorable thoughts inhibit pero;,uao;,ion and ca n ca use a backla:,h. Persuaders who elicit favorable thought~ are therefore morc likely to create the desired attitude change. Receivers can experience a mix of favorable and unfavorable cognitive response.., to a message. Second. receivers cun generme muny or few thoughh 111 respolhe to a pcro;,ua..,ive message. The number of thoughts also shapes the kind of inlluence that is likely to occur. Attitude change i.., more likely to ensue from many than from a few favorable thoughts. On the other hand. large numbers of unfavorable thoughts reduce the likelihood that ames· ~age will be per..,uasive. As noteo above. a receiver', cognitive rc\ponses can be a mixture of favorable and unfavorable thoughts. All thingo;, being equal. when the "net" favorable cognitive response is larger (more favorable than unfavorable thought..,), persuusion is more likely to occur in rc~ponse to a mc..,sage. A~ the margin of favorable 10 unfavorable thoughh increases. the likelihood (and amount) of pcrsua..,ion should increase. When the "net" favorable cognitive respon..,c i...... maller (or a negative number. with unfmorable thought<., outnumbering favorable one ... ). persuasion is les ... likely. If the unfavorable thoughts outnumber favorable ones. a "boomerang" effect of altitude change in the opposite direction of that advocated by the message becomes more likely. Thi ... explains why both the IIlIlIlber and l'alel1('e of cognitive respon<.,es are so important to persuiJsion.
Number oj Cogllitive Respollses The number of cognitive rc~ponses is determined by two factors: the receiver's motil'llliOI1 and ability (0 process a message cognitively. Motivation i... directly related to involvement in the message: The morc important a me ... sagc topic is to a receiver (the greater the
98
Part II • Persuasiotl Variables: Persp(!('(il'{'.\" all Sown's. Receivers, Chal/l/()/s. and Messages involvement), the more motivated that person is to think about the message (and produce thoughts). The less a topic matters to a listcner. the less motivated that person is to think aboulthe message (and fewer thoughts should occur). Research confirms that auditors are more likely to scrutinize a I11cssage-or use the central route to persuasion, or produce more cognitive rcsponses-on an involving (salient, relevant. important) than a non involving topic (see. e.g .. 10hnson & Eagly. 1989: Peuy & Cacioppo. 1979a. 1979b. 1981b. 1984; Petty. Cacioppo, & Goldman, 1981: Peuy. Cacioppo. & Schumann, 1983: Roser & Thompson. 1995). Thus. topic involvement increases receivers' motivation 10 process a message and increases the number of cognitive responses they are likely to produce in response to that message. However, the audience mllst also have the ability to process centrally. One variable that has frequently been employed to study processing ability is distraction: Receivers who are distracted produce fewer thoughts in response to a persuasive message than those who are not (O,terhouse & Brock. 1970: Peuy. Wells. & Brock. 1976). Other factors also influence the receiver's ability to process a message. Messages that are more difficult to understand should produce fewer thoughts. Message repetition can increase the audience's ability to process a message. although too much repetition could create boredom or tc(liulll. reducing that message'!o. effecliveness (Cacioppo & Petty. 1985: Petty & Cacioppo, 1979b). If the audience know, lillie about a topic. it should be difficult to produce many thoughts about that message. Thus. many variables can influence the extent of people's ability to process a message and. therefore. the number of cognitive responses produced when processing a mcssage. Distraction, for example. can either hinder or help persuasion depending upon the valence of the thoughts it suppresses. If a receiver is likely to disagree with a persuasive message. distraction means that this auditor would probably havefewer unfavorable cognitive responses (compared with undistracted listening). In this case. the auditor who is diMracted is likely to experience more persuasion with distraction than without it. On the other hand, if a receiver is likely to agree with a message, distraction would mean fewer fal'()rable thoughts and thu!o. less persuasion (compared. of course. with undistracted reception). Distraction thus interferes with the listener's ability to process a message and reduces the number of cognitive responses. When fewer unfavorable thoughts are "suppressed" by distraction, the listener should be more persuaded: when fewer favorable thoughts occur, the listener should be less persuaded.
Valellce of Cogllitive Responses Several factors influence the valence of cognitive responses. First, the valence of thoughts is influenced by the position advocated by the persuader. Listeners are likely to produce favorable thoughts when they agree with the messages they receive. Conversely. receivers are likely to produce unfavorable thoughts in response to messages they do not accept. The position of the message relative to the audience's attitudes therefore influences the valence of thoughts. Second, forewarning auditors about a persuasive message can intluence the thoughts they are likely to produce in response to that message. When people are warned that they
Chapter 6 • SUllrn' O w lihililY amI rill' tJtI/Jorar;{J/1 LIJ..l'Iilwod MOt/l' l
99
will disagree with a message (counterauitudinal mc:-. ... age). th~y are likely to produce negative thought ... in anlicipation orth:'11 message (Brock. 1967: Pctty & Cacioppo. 1977). Third. argument quality inlluenccs the valencc of cognitive responses. Stronger argument ... are 1110re likely to elicit favorable thought ... [him weaker messages. wherea ... weak arguments are more likely to e\(lkc ncgalive thoughts than stronger or high-quality message .... Petty and Cacioppo (19K..1) found that auditors produce more favorable cognitive responses to Ille ...... age ... \\ ith strong argulllenh than tho ... e with weak arguments and generate more unfavorable thought:.. in regard 10 message ... compri ... ing weak rather than strong argumcnts. Benoit (1987) found that mc ... sagcs with strong arguments produced more ravorable thoughts, fewer unfavorable thoughts. and more attitude change than messages with weak ;Jrgul11enh. These effcct... arc more pronounced on involving than uninvolving
topic, (Andrew, & Shimp. 1990: Petty & Cacioppo.
198~).
Thu:-. ....eventl faclOr~ can inlluc.::ncc the \'i.llencc of cognitive re~pon:-.es. Messages that disagree \\ith the audience are more likel) to produce unf,.I\ arab le thought.... whereas agreeing me ... :-.;Jges arc more likely to encourage favorable thoughts. Strong me:-.sages arc morc likely to produce favorable thoughts. whereas \\·cak messages arc prone to elicit unfavorabh! thoughts.
Peripheral Processillg Mo ... t of the discu\,ion thus far ha ... fm:u ... eu on central pn.lCc . . sing. although the amoullI of central proccs ... ing employed (h.;pentis on the motivation and ability of the receiver. Peripheral processing is associated with fewer thought .... Howc\'cr. peripheral processing is thought 10 be qualitatively different from central procc:-.sing. Thc receiver who engages in peripheral processing lIse.s a Clh.! or a uccision rule to decide whether to agree with the message (instead of thinking about the arguments in the message), Peripheral processing occurs when receivers lack the ability and/or motivation to think about the arguments in the message: however. they may rind other ba ... c ... for deciding \\ hether to accep t the mes~age.
Severnl peripheral cue ... (\\hich may be thought of a~ mental shortcuts) have been identified. For example. whcn receiver ... pcrcei\e the source a~ physic;Jlly attractivc. they may usc attractiveness as a pcripheral cue (Petty & Cacioppo. 198Ia). An audience mcmber mightthillk. "This is a very attractive source. I think I should agree wi th him (or her)." Remember that periphcral procl!s ... ing i. . morc likely to occur when the topic is less imporlant or Ie ...... invoh ing. ~o deferring to an attracti,e ... ource (in ... tead of expending the effort to think about the ideas and argument ... in the mc ...... age ) rnay not be problematic for the listener. Sel:ond. if a message contuin . . a large number of argument .... a receiver Illay decide lO accept the rnes ... age 011 the ba ... is that any message with so many arguments is probably correct (Petty & Cacioppo. 1984a). We may have it decision rule that. all things being equal. a I11c:-...,age with many argulllcnt ... is Illore likely to be true than one with few arguments. Third. when a listener belic\c~ that several ... ourcc.:s col\ecti\'el) endor..,e a message position. the liqener ma) be more likel) to accept that me ...... age. Harkin ... and Petty (1981)
100
ParI II • Persll(l.lio/l Variables: Perspeclh'e.l·
(lfl
Sources. Receil'(' f"s. Challl/l'I.\".
{/I/(/
Me.\'slI!!,e.'I
found that more arguments and more !-oources eac h generate more favorable cogni ti ve re sponses and more attitude change than ll1e~sages wi th fewer argument..; and sou rces. All things being equal. an idea that many people accept is more like ly to be true than one that few people believe.
Source Credibility With this understanding of how persuasive messages arc processed, we ca n turn to an analysis of source credibilit y. which can influence persua~ion by influe nci ng hoI\' receil'ers process messages- the number of cogn iti ve responses. the va le nce of cog niti ve respon ... es_ and the object of their thoughts: message ve rsus source. The claim that source credibility influences persuasion by affec ting message processing is suppon cd by research "uggesting that source credibil ity affects persuasion only if the source is identifi ed hefore the Illes sage has been processed ( Hu sek. 1965 : Mills & Harve y. 19 72). Ward and McGinnies ( 1974: see also Greenberg & Tanne nbaulll. 1961 ; Sternthal. Dholakia. & Leavitt, 1978) found that there was no difference in attitude change between high- and low-credibility source\" when those sources we re ident itied ajfer the message (see also O' Keefe's meta-analy,is. 1987). In addit ion. Greenberg and Miller ( 1966) reported that low-credibi lity source\" hind ered persuasion only when they were identified as disreputable before the message. These findings all suggest that source c redibilit y innuences persuas ion by altering how people process (or elaborate) messages. Wh e n th e source is identified ajier the message. that mcssage has already bcen processed. Any thought s the audience may have abou t the message h ~we already occurred. so it is too late for identifi cation of the source to have any e ffect. Si milarl y. Rhine and Kaplan ( 1972) found that when there is no ll1 es~age to process (mere ly an assertion or c laim ), th ere i\" no persuasive effect from va riati ons in source c redibilit y. Thus. research indicates that c red ibility cues mediate persuasion by influencing holl' messages (/lfrihllfed 10 that SO/free (Ire processed. Second. the persuasive effec t s of source credibility are more like ly to manifest themselves on less involving topics. suggesting that credibil it y can also serve as a peripheral cue. Petty. Cacioppo. and Goldman (1981) found th at on highly involving topics. message argume nt s produce attitude c ha nge but so urce c red ibilit y does not. On unin volvi ng topics, both arguments and credibi lit y mediate persuasion, but credibility has more influence (see also Benoit. 1987: Chaik en. 1980: John son & Scileppi. 1969; Petty. Cac ioppo. & Goldman. 198 1: Petty. Cacioppo. & Schumann. 1983). The authors of the ELM have specificall y addressed the relationship between sou rce f<.ictors and persu<.l!-oion. Pe tt y and Cacioppo explain how charac te ri sti cs of message <.,OllfCCS can inlluence attitude change: " In th e ELM. source factors can influence auitude change in three ways: They can serve as arg ume nt s [an attractive mode l is ev idence for a beauty product I, th ey ca n se rve as [peripheral [ cues, and they can affect argumellt processin g" (1986a, p. 205). It is possible that th ese factors co uld come into play in situation s in volving expert ise as well as attractiveness. For exa mpl e. the cigare tt e smoke r who is dying from cancer co uld be seen a"i the embodime nt o f the argument to quit (or never to start) ... moking. However. we were unable 10 locate any research on the first method of inlluence, in which the nature of the source actually functions as an argument {Q support
Chapler 6 • SOl/rce Credibility alld the Elaboratioll Likelihood Model
101
the claims advanced in the message. Our rcview will discuss how expertise and trustworthiness can inllucnce argumcnt processing and serve as peripheral cucs.
Number of Cogllitive Respollses Source credibility can influence persuasion by inlluencing both important components or cognitive responses. First. source credibility may inlluence the nll/llber of tho/lghts produced in response to a message. Here. belief that the source is an expert on the topic of the message could encourage receivers to "relax their guards:' or feel le">s motivation to scrutinize the message (produce fewer cognitive responses). In contrast. if the source is thought to be disreputable. that belief may lead receivers to be more wary. subjecting the message to greater scrutiny (produce more cognitive responses). Gillig and Greenwald (1974) found that fewer unfavorable thoughts are produced in response to a message attributed to a high-credibility source, whereas more unfavorable thoughts are elicited by a low-credibility source. Conversely. auditors are more motivated to think critically. and tend 10 produce more counterarguments. in regard 10 mes,agcs from apparently nonexpert sources. reducing persuasion from such sources (see Benoit, 1991; Cook. 1969: Gillig & Greenwald. 197-1: Has;. 1981: Perlorr & Brock. 1980).
Valellce of Cogllitive Respollses Source credibility also can innuencc the I'{llenee of cognitil'e responses. Benoit and Kennedy (1999) found that trusted sources produced more favorable and fewer unfavorable thollght~ (and more attitude change) than untruslcd sources. However. there was no difference in the total number of thoughts produced frol11 these sources. That is. source trustworthiness influenced the distribution (valence) of thoughts in this study, but not the number of thoughts. The morc favorable thoughts. and the fewer unfavorable thoughts, the more persuasion should occur. Conversely. cognitive responses that arc more unfavorable than favorable should inhibit attitude change.
Source Credibility Effects In this section we will describe and explain three specific topics: involvement. moderately credible sources. and di..,crepancy. These are important areas in which the ELM helps to explain source credibility effects.
Involvemellt.
Several studies <.,uggest that credibility is most important on uninvolving lOpics. which suggests again that it often functions as a peripheral cue. Johnson and Scileppi (1969) report that high-credibility sources were more persuasive on uninvolving than involving IOpics. Pelly. Cacioppo. and Goldman (1981) found that argument quality but not credibility inlluences persuasion 011 involving topics. while on uninvolving topics credibility was more inlluential than argument quality (see also Chaiken. 1980). Benoit (1987) found that experts were no more persuasive than nonexperts on an involving topic (of. Stiff. 1986). Neimeyer. Guy. and Metzler (1989) found that a credible source elicited more favorable and fewer unfavorable thoughts. but only in a low-involvement condition.
I 02
Purt II • Pn,\If(H iOI/ Vo riahlt's: Per,\!)(!{'{i!'{'S ()/I SOli rce,\", Rea; I'crs, Chunnels, and Mes,wges
The mcta-analy,is by Wilson and Sherrell (1993) found that in two-thirds of the studies examined. source factors had a ~ignificallt effect on attitude change only in a lowin\'olvemcnt condition. On involving topics. audience members are motivated to .s crutinize the message. engaging in central proces:..ing of the idea:.. and arguments in the message, Thus. source credibility ..,hould have a minimal impact on sllch topics. As noted above, Wilson and Sherrell"" (1993) study identified a number of studies in which source credibility had effects in high-invol\'ement conditions. but these effects were in the minority.
Moderately Credible Sources.
Ordinarily. highly credible sources are more persuasive than sources who appear le~s credible, However. therc are circumstances in which a moderately credible source is more persuil'live than a highly credible source. Several studies have found that when a message is proaffilUdiJla/ (attempting to reinforce or strengthen rather than change attitudes), moderately credible sources are more persuasive than highly credible one, (Bochner & In,ko. 1966: Bock & Saine. 1975: Dean. Austin. & Watts. 1971). Sternthal. Dholakia. and Leavitt (1978) reported that in such situation,. the moderately credible "ource elicits 1110re fa\'orable cognitive responses than the highly credible ...ource. Here. the high-credibility source reduces motivation to scrutinize the message (to engage in central processing), However. hecause a proatlilUdinaimessage agrees with the audience. the kind of cognitive responses that are ~uppresscd are favorable. resulting in less attitude change. This is. of course. because valence of cognitive responses is so important Thi.., effect of source credibi lity is analogous 10 the effect of distraction di"clIssed earlier. Receivcr"i who are distracted produce fewer thoughts. If undistracted they would have produced primarily fill 'orable thoughts. when they are distracted they produce fewer favorahle thoughts and are less persuaded (than if they were not distracted). If they would have produceJ primarily UI~f{lI'of{/h1e thoughts when undistractcd. distraction would reduce the numher of unfavorable thoughts and they will be more persuaded (than those who arc undi . . tracted). Similarly. high-credibility sources decrease the audiencc's motivation to think about a message. reducing the number of thoughts on proattitudinal topics and producing Ie ... .., persuasion than a moderately credible source (a disreputable source is likcl y to pro\'okc unfavorable thought .... by the way).
Discrepancy and Attitude Change.
If persuaders wish to change an audience's auitude, they I11U ... t disagree with the audience. If the persuader parrots their own attitudes back to thcm. thcre will be 110 reason for the audience to change those attitudes. Of course. if persuadcr" di..,agree too much. the audience may consider their message to be unreasonable. and it may produce unfavorable cognitive responses and no attitude change. Thus. the degree of discrepancy between the message position and the audience's attitude is an important variable in persua ... ion. Considerable research has documented a curvilinear relation..,hip between discrepancy and per",u3sion. Disagreeing with the audience a little produces little attitude change. disagreeing moderately produces morc persuasion. and disagreeing a great deal produces little or no persuasion (Bochner & Insko, 1966: Insko. Murashima. & Saiyadain. 1966: Johnson. 1966: Osgood & Tannenbaum. 1955: Peterson & Koulack. 1969: Tannenbaum. 19(7). Note that Stiff (1994) claimed SUpP011 for a linear
Chapler 6 • Source Credibility alld the Elaboration Likelihood Model
103
relationship between discrepancy and attitude change, but he then provided three reasons why high levels of discrepancy would be unpersuasive, undercutting his own argument. However, some research has found a different pattern of results for highly and moderatcly crediblc sources. Aronson. Turner, and Carlsrnith (1963; see also Bochner & Insko, 1966) reported that for moderately credible sources, the relationship between discrepancy and atlitude change was indeed curvilinear (little attitude change for low and high discrepancy: more attitude change for moderate discrepancy). However, for highly credible sources there was a direct relationship between highly credible sources and discrepancy: the greater the discrepancy, the more attitude change. Although these researchers did not measure cognitive responses. the likely explanation is that highly credible sources reduced the receivers' motivation, so the audience produced fewer unfavorable thoughts to high levels of discrepancy. Thus, without the inhibition of counterarguments. the more discrepancy advocated by highly credible sources, the more persuasion results. Moderately credible sources, on the other hand, do not reduce motivation to think about messages. and highly discrepant messages are thus likely to evoke more unfavorable thoughts than less discrepant messages, resulting in less persuasion. Rhine and Severance (1970) presented data suggesting that these results are likely to occur for uninvolving. but not for involving topics. This finding is consistent with (he ELM. which would suggest that receivers who are highly involved in the topic are likely to engage in central processing regardless of the credibility of the message source. Furthermore, Choo (1964) failed to replicate the finding that for moderately credible sources there is a curvilinear relationship between discrepancy and persuasion but found that highly credible sources have a direct relationship. However, this study operationalized credibility as trustworthiness rather than expertise, and trustworthiness may not influence the number of thoughts. It is clear that we would benefit from studies that replicated this research, clearly distinguishing between expertise and trustworthiness, and measuring cognitive responses.
Evaluation of Work on Credibility There can be no doubt that source characteristics are capable of influencing persuasion. Probably the most basic question here is, what is source credibility? Research manipUlates and measures credibility in troubling ways. Research does not always distinguish between different source manipulations. Some research employs celebrity sources (Petty. Cacioppo. & Schumann. 1983), but it is not clear in all cases whether such source manipulations concern expertise or trustworthiness (or attractiveness). Furthermore, expertise. trustworthiness, liking, and attractiveness are often confounded. For example. Johnson & Scileppi (1969) manipulated credibility by telling message recipients that the source of (he communication was either a medical authority described as an expert on the topic or a "medical quack" who had served a prison term and who had written the article for a sensationalist publication. Tn contrast, Ward and McGinnies (1974) described their high credibility source as an expert on the issue and someone who was considered trustworthy. Their low-credibility source was presented as having no expertise and as being devious and calculating. In both studies the high-
104
Part II • PerslI(uioll Variahfes: Per.\pl'clit'es
OIl
Sources. Receit'ers, Chal/llels. and Mes.Hlges
credibility source was associated with greater altitude change. However. given that expertise and trustworthiness were confounded in these manipulations. it i~ impossible to ascertain whether the effects are due mainly to expertise or to trustworthiness. Similarl y. Chaiken and Maheswaran (1994) attributed a message either to COIl.\'Umer Reports or to a K-Mart circular. Subjects perceived Consumer Reports to be the more credible source. but it isn't clear whether this was due to greater perceived expertise or greater perceived trustworthi ness. Some research contains other potential confounds. For example, Chebat, Filiatrault, Larouche, and Watson (1988) employed a fairly common method of manipulating expertise: The expert was described as a professor (in a discipline relevant to the topic of the message), the nonexpert as a studenl. However. this could confound expertise with simi larity, because the low-expert source, a student. is probably seen as more similar to the subjects than the high-expert professor. Future research ll1u .... be clear about which \ource characteristics are being studied and how they are manipulated. Concerns can also be raised about how credibility is mea~ured. Some studies do not report the items used in manipulation checks (Chebat. Filiatrault. Laroche. & Watson. 1988). Others combine arguably di"imilar traits. Chebat, Filiatraull. and Pen'ien (1990), for example, operationalized credibility as four questions (expertise, trustworthiness. attractiveness, and prestige) thHt were averaged into a single credibility score. It is nol clear exactly what construct is being meusured here. The role of cognitive respon..,es in attitude change is not always investigated (or reported). For example. PeIlY. Cacioppo, and Goldman (1981) did not reporl measures of cognitive responses. Other studies failed to report effects of manipulations of the independent variable on cognitive responses. Peuy, Cacioppo. and Schumann ( 1983) found no effects of the manipulations on cognitive responses, possibly because the thought-listing mea~ure was administered after a series of other messages. Given the importance of cognitive responses in the ELM. and the fact that the ELM is clearly a process model of persuasion. this is not a trivial concern. We speculate that the common method of measuring cognitive responses~thought listing~ may not be a reliable indicator. Other approaches. like concurrent verbalizution. might yield better results.
Repairillg Damaged Credibility The discussion so far has not addressed the question of whcther a persuader's credibility is static. Politicians like President Bill C linton and Representative Gary Condit. public figures like Tonya Harding or Martha Stcwart, and corporations like Firestone and Enron have learned the hard way that public perceptions of a person or organization can change-and change precipitously~ for the worse. Research reviewed earlier (e.g., Benoit & Kennedy. 1999) makes plain that influence attempts from persuaders with low credibility are likely La produce more unfavorable thoughts. fewer favorable thoughts. and les~ attitude change than attempts from persuaders with positive reputations. This rai\es the question of whether persuaders who ha\'e suffered damage to their reputations (and who can Lhu~ expect to be less effective perwaders) can do anything to remedy this situation.
Chapler 6 • Source Credibility lind the Elaboratioll Likelihood Model
lOS
Typology of Image Repair Optiol/s Benoit (1995a. 1997c, 2000a) has developed a theory of image repair discourse that discusses the options available to people or organizations who need to recover from a damaged reputi.ltion. He begins with the assumption that a threat to an image has two components: oJlellsi!'eness and responsibility. First. there must be a problem or breach of expected conduct. If nothing bad happened, there is no blame to apportion. Second. a person (or organization) accused or suspected of wrongdoing must have caused (or encouraged, permitted. or failed to prevent) the problem. If a bad thing has occurred but I had nothing to do with it. I cannot reasonably be blamed for that problem. Benoit and his associates (Benoit & Domes, 1996: Benoit & Hanhcock. 1999: Benoit & Wells. 1996) have argued thaI the damage from an attack or criticism can be exacerbated by increasing the apparcl1l offensiveness of the act or the accused's apparclll responsibility for that act. Furthermore, Benoit (1995. 1997) has explained that image repair effons ean proceed through three general approache~: Reduce or eliminate the apparent offensiveness of the act in question, reduce or eliminate the accused's responsibility for the act, or concede both offensiveness and responsibility with an apology and request for forgiveness (which. following Kenneth Burke. he calls "mortification"). Thu~, the potential strategies or options for restoring a tarnished image grow out of the two components of an attack, criticism, or accusation. Table 6.1 lists five general strategie~ and 14 specific tactics for image repair. The first two general categorie~, denial and evasion of re~pol1sibility, attempt to reduce (or eliminate) the accused's apparent re~ponsibility for the offensive act. The next two general strategies. reducing effectiveness and corrective action. attempt 10 reduce the perceived offensiveness of the act in question. Finally. as indicated earlier. mortification admits performing an offensive act but apologiLes and asks for forgiveness. All of these strategies have the potential to rehabilitate, at least in part, a damaged image or reputation.
Effectiveness of Image Repair Strategies Some quantitative research has investigated the effectiveness of attempts to restore a dam· aged reputation. Benoit and Drew (1997) reported that offering corrective action and apologi.dng (mortification) are perceived to be the most effective and appropriate strategies. However. a good deal of this work has used rhetorical criticism or a casc-study approach. Benoit (1997) summarized this work and offered several suggestions for repairing a damaged reputation. First, it is important to identify all of the key accusations; a message that ignores key allegation.., will not repair one's credibility. Second, it is vital to identify the key audience, The persuader need nol necessarily restore lost credibility for everyone: the ones who matter the most are those he or she will try to per~uade (the target audience for a persua~ive mes~agc should be the target audience for repairing a tarnished image). Benoit also offered suggestions for constructing a message to repair a damaged image. He begins by cautioning that the image repair message itself must be persuasive, using strong arguments. providing adequate support for claims. and so forth. He recommends that people who have commilled wrongdoing confess and apologize: for example,
106
Part II • Persuasion Variables: Perspectives 011 Sources, Receivers, Channels. and Message.\'
TABLE 6.1
Image Restoration Strategies
Strategy
Key Characteristic
Example
Simple denial
did not perform act
Tylenol: did not poison capsule
Shift (he blame
another performed act
Tylenol: a "madman"' poisoned capsules
Denial
Evasion of Responsibility Provocation
responded to act of another
finn moved because of new
taxes Defeasibility
lack of infonnation or ability
executive nOllOld meeting changcd
Accident
mishap
tree fell wreck
Good intentions
meant well
Sears eager to provide good auto repair service
011
tracks causing train
Reducing Offensiveness of Event Bolstering
stress good traits
Exxon's "swift and competent" cleanup of oil spill
Minimization
act not scrious
Exxon: few animals killed in oil spi ll
Differentiation
act less offensive than simi lar acts
Sears: unnecessary repairs preventive maintenance, not fraud
Transcendence
more important values
helping humans justifies testing animals
Attack accuser
reduce credibility of accuser
Coke: Pepsi owns restaurants, competes direcLly wi th you for customcrs
Compensation
reimburse victim
disabled moviegoers given free passes after being denied admission to movie
plan to so lve/prevent recurrence of problem
AT&T long-distance upgrades; will spend billions more to Improve service
apologize
AT&T apo logized for service interruption
Corrective Action
Mortification
Derived from Benoil. 1995, 1997.
Chapter 6 • SOllrce Credihility (/Ild tlil' £I(1/){)ralioll Likelihood Model
107
President Clinton probably got into more trouble from deceiving people about his relationship with Monica Lewinsky than from the relationship itself. On the other hund, those who are truly innocent (e.g" Tylenol apparently was not responsible for tainted medication) should proclaim innocence and if possible identify the "true" culpril. However, blame should be shifted away from the persuader. Nixon blamed his closest, hand-picked subordinates for Watergate, Clearly, this did Ilot clear him rrom blame. At Limes factors beyond one's comrol are to blame. Exxon could have blamed the slow oil spill cleanup on poor weather. However, they chose to blame the U.S. Coast Guard and the state of Alaska (implausible targets of blame). It is important to implement corrective action, cither by fixing the problem or preventing its recurrence (or both). Even though Tylenol established its innocence. it introduced tamper-resistant packaging and then gradually replaced capsules with caplets to prevent future episodes of poisoning. Some strategies (c.g .. minimiLation. provocation) have not been found to be effective. It is also possible to combine some of these strategies, but not all combinations are equally effective. For example. mortification and corrective action seem to complement one another: "1 am SOITY for the damage I caused and I will fix the problem" is a plausible response. On the other hand. "1 did nothing wrong and I apologize" seem:.. to be an awkward combination: Why would you apologize if you (truly) did nothing wrong? It is therefore important to develop strategies that work in harmony. Thus, while some reputations may be beyond repair (e.g., Jeffrey Dahmer). it is possible for persuaders to improve their credibility. Research has identified a number or options available to those who need to repair a tarnished reputation. We are beginning to develop an understanding of how and when to lISC these options. IL is clear. however. that more empirical work would be helpful in this area.
Conclusion In this chapter we took the ELM as a theoretical standpoint for understanding source ef-
fects in persuasion. Unlike aLtractive sources. expert and trustworthy source:.. influence persuasion only when they are identified as slIch before a mes~age is processed. We hold that attitude change is a function of the number and valence of thoughts produced in response to persuasive messages. Because it often functions as a peripheral cue. credibility is more likely to intluence persuasioll Oil less involving topics. Credible sources may reduce motivation to process messages. which means that they arc likely to facilitate persuasion when thoughts are likely to be unfavorable and may actually impede persuasion when thoughts are likely to be favorable. Disreputable sources are likely to encourage unfavorable cognitive responses and result in less persuasion. Highly credible :..ources may increase a message's persuasiveness by permitting higher levels of discrepancy (without an increase in unfavorable thoughts). We also expressed several reservations about current research on source effects in persuasion. Some studies do not make it clear how credibility is manipulated or measured. Other studies clearly confound multiple constructs (expertise, trustworthiness. attractiveness, and prestige). Some research does not measure (or report) the effects of independent variables Oil cognitive responses, which arc conceptualized as a mediating variable in lhe
108
Part II • Persuasion Variables: PerspeClil ,cs on Sources, Receivers. Channels, and Messagl' s process of persuasion. It may be useful to employ alternative methods of measuring cognitive responses (concurrent verbalization). Despite literally centuries of inquiry into the effects of source variables on persuasion, this construct would benefit from morc focused research. Finally, we addressed the topic of changes in perceived credibility. Persuaders who suffer damage to credibility do have a chance of repairing their tarnished reputations, and thereby enhancing future attempts at persuasion. However, it is clear that more empirical invesligation into this topic would greatly enhance our underslanding.
References__________________________________________________________ Andersen. K.. & Clevenger. T. ( 1963). A \ummary of experimental research in ethos. Spf'lyh /110110 ~raph.~. 30, 59-78. Andrew ... J. C. & Shimp. T. A. (1990). Effect .. of involvement. argument strength. and source charat..:teristic!-. on central and peripheral processing in advertising. Psyclwlog\' & Marketing. 7. 195- 214. Aristotle. ( 1954). The rhetoric (W . R. Roberts. trans.). New York: Random House. Modern Library. Aronson. E .. Turner, J. A .. & Carlsmith. J. M. (1963). Communicator credibility and communication di~ crepaney as determinants of opinion change. jOllrl/ol ofAlmor/1w/ alld Social P.\)'choloKY. 67. 3136. Benoit. W. L (19R7). Argument and credibility appeab in persuasion. SOllthern SpC('ch Comlllllllicatioll jOllrnal, 52.181-197. Benoit, W. L (1991). A cognitive respon!-.c analysis of ~ource credihility. In B. Dervin & M. J. Voigt (Eds.). Pro~re!Js ill commullicatioll sciences (vol. X. pp. 1- 19). Norwood, NJ: Ablex. Benoit. W. L. (1995). Arco/fl//s. excuses. apologie.\·: A theory of image /'('.\10((111011 sll'{llegies . Albany. NY: State University or New York Press. Benoit. W. L. (1997). Image re~toration discourse and cri..,is communication. PlIblic Re/ariolls Rel'iell', 23. 177- 186. Benoit. W. L. & Domes, B, (1996). Darcline NBC's pcr:-.uasive altack of WalMar!. COIlIlIIllI1i('(llio/l Quarfer/\', -t4, 463----477. Benoit W. L, & Drew, S. (1997) . Appropriateness and effectivencs:-. of image repair strategies. COIl/IIIUnicalioll Reports, 10. 153- 163. Benoit. W. L. & Harthcock. A. (1999). Funetion~ of the Great Debates: Accluim:-.. a!tacks. and defenses in the 1960 presidential debates. COII/mlll/icalioll MOl/ogmphs. 66. 341 - 357. Benoit. W. L. & Kennedy, K. A. (1999). On reluctanl1estimony. Commllllicatioll Quarterly. -t7. 367387. Benoit, \Y. L.. & Well~. W. T. (1996). Calldid(lleS ill conj7ict: Pl'r.\IIW'il'e mUle/.. alld defcllJe ill Ihe /1)92 presidenlial debmcs. Tuscaloosa, AL: University of Alabama Pres ... Berscheid. E. (1985). InterperMlnal allraction. In G. Lindsey & E. Aronson (Eds.), Halldbook 0.( social psychology Ord cd .. vot. 2. pp. 413--484). New York: Random House. Bcrscheid. E .. & Walster. E. (1974). Physical attractivene;.,s: In L. Berkowitz (Ed.). Admllce.\ ille.r;perilIIenllll social psychology, 7. I 57 ~215. New York: Academic Press. Bochner. S .. & Insko. C. (1966). Communication disnepancy. source credibility. and opinion change. journal of Pcrsollality alld Social Psychology, -I. 614-621. Bock. D .. & Saine. T. (1975). The impact of source credibility, attitude valence, and task :-.en:-.itivity on trait errors in speech evaluation. Speech MOl1oKmplis. 37, 148- 155. Brock, T. C (1967). Communication di~nepancy and intent to persuade as determinant.. of counterargument production. jOl/l'I/a! of £x{JerimclIla/ Sorial Psychology. J. 296-309. Cacioppo, J. T., & PellY. R. E. (1985). Central and peripheral rOlltes to per~lIasion: The role of message repetition. [n L. F. Alwin & A. A. Mitchell (Eds.). PS'ycholoKical processes and adl'('ni.\·illg effects: Theory. research. lIlId applicatioll (pp. 91-1(1). Hillsdale. NJ: Erlbaum. Cacioppo. J. T.. Petty. R. E .. & Kao. C F. (1984). The efficient assessment of need for cognition . .Iolll'lllll of Personality AssessfIlcl/I, 48. 306-307.
Chapter 6 • Source Credibility (llId fhe Elaboration Likelihood Model
109
Cacioppo. J. T.. Petty. R. E.. & Morri .... K. (1983). Erfcct:-. of need for cognit ion o n me~ . . a£~ eva lu ation. recall. and per"ua:.ion. )olll"llal of Pawmlliin (/lui Social P\.H·/wloK\,. -15. 805-818. Calder. B. J .. 111 . . "0. C. A .. & Yandell. B. (1974). The rdatio n of cogniti\e and memorial proce:-.<,c . . to r)er . . ua~ion in a . . imulated jury trial. }mm",1 oJ Appli('ti Social P.I".\dwlogy. -I. 62-93. C hai"cn. S. ( 1979). Communicator phy . . lcai attmcti\'cne . . s and peNla . . ion . )011/7/(/1 oJ PerSOIulfi1.\' lIlIlI So· cial Plyclw/ogy. 37. 1387-1397. Chaiken. S. (1980). I-I clIriqic ven.u . . sy . . tcmalic in formation proce~~ing and the u.se of :-.ource \cr . . u~ me.s:-.age cue~ in per!o.ua~ion. )Ollrl/(I/ of PawmalilY alld Social P'~y(""(JI()gy, 39. 752~756. C haiken, S .. & Maheswaran. D. (1994). Hcuri:..tic procc.;,sing can bias sy.stemat ic proce" . . ing: Effects of .;,ou rce credibility. argument ambiguity. and Ia.sk importance on attitude judgment. )ourl/ol of Per· .\olUdity and Social Psychology. 66.460-473. C hebar. J-e., Filiatraull. P.. Larouche. M.. & Wat ...on. C. ( IY88). Compen ... atory effeus of cognitivt.! char· acteristics of the \oun:e, the rnC'i~agc. and the reccivcr upon
/22.609-62 1.
Cheba!. J-C .. Filiatrault, P .. & Pcrrit.!n. J. (1990). Limit'> of c redibility : The ca~e of polllical per~ua:-.ion. )011/"1/01 of Socia/ P.rychology. 130. 157~ 167. Chao, T. (1964). Communicator credibility and cornmunication discrepanq a~ determinant ... of atlltudt.! change. )olfmal (~rSocial Psychology. 6.J. 65 ~76. Cicero. ( 19-1-2). De oratore (E. W. Sutton & 1-1 . Rackham. tran" .). Cambridge. t\tA: Harvard Uni\cr . . ity Pre"s. Coo\"' , T. D. ( 1969). Competence. countaargui ng. and anitmle change. )olll"l/af oj Perlollulil,l'. 37. 3-1-2-
358. Dean, R .. Aust in . J.. & Wans. W. (1971). Forewarning clrec t ~ In pen,ua..,ion: Field anti das<.,room e\pcri· menl ... )ol/mlll of Persollality lIlI(l Socio/ P.H·c!lOlog\,. IN. 210- 221. Debono. K. G .. & I-I arni ... h, R. (1988). Source experti ..t.!. ~nurce allracti\ ene ... ~. and the praee<;sing of pcr... ua .. ive information: A functional approach. PerllJ//ollty lIIlll Social P~ychofoK.\· Bllllelill. 17.245251. Eagly. A. 1-1 .. & Chai ken. S. ( 1975). An attribution analy\is of the effect of communicator characteristic ... on opi ni on change: The ca ...e of communicator Htlractivenc ...... )olll"llal oJ Pa,\"OlIlIlily (lml Social PsycllOlog\'. 32, 136-4..L Eug ly. A. H .. & Chaiken, S. ( 1993). The IHYC/JO/ogy of auitlllle.l. Fort Worth. TX : lI arcourt Brace Jovanovich. Gi Ilig. P. M .• & G reenwald. A. G. (1974 J. ]..., i[ time [0 lay the !-;\eeper effect [0 re.,r.' )olfrl/(/I of Penollalir,\' alld Social Psychology. 29. 132~ 139 . Greenbe rg. B. S.. & Miller. G. R. ( IY66). The cfk<.:[" of low ·crooib le source:-. on mc ..... age acceptance. Speech MOllogl"(/{Jhs. 33. 127~ 136.
Greenberg. B. S .. & Tannenbaum. P. 1-1 . ( 1961). The cffr.!ct .. of byline ... on altitude change. )olll"lluli.HII Quarterly. JR. 535-537. Grr.!e nwa ld . A. G. ( 1968). Cognitive learning. cognitive rc:-.ponse to persuasion. and attitude change. In A. G . G recm';ald. T. C. Brae". & T. M. O ... trolll (Edq. P~ydw/og;cal fO/llulll1;ol/s of (lfllllld{'~ (pp. 1.t7- 170).Kew York: Academ ic Pre,,\ . Har"in.s. S. G .. & Pelt). R. E. (198 1). Effect.;, of .;,oun:e magnificallon of cognithc effort on attitude.s: An information process in g \ iew . .Iolll"llal oJ Perwmafity lind Social P,lyclwlogy. -10. 40 l-t 13. Has .... R. G. (1981). Effect<., of source eharactcri<.,tic:-. on cognitive rC'\pon ..es and pcrsua:-.ion. In R. E. PI!\[Y. T . M. Ostrom. & T . C. Brae" (Ed~.). Cogllitil'{' I"l'JpOIIH'\ ill pt:r<:ilasiol/ (pp. 44-72). Hilhda1c. NJ : Eribaulll. Ihl...cJ.... T. R. ( 1965). Per.. ua~ive irnpal.:t:-. of early. latt.!. or no mention of the negative .source. )olll"llal (!f Persol/ality (lnd S(wial P.lydwlogy. 2. 125 ~ 128. In ~ko. C. A .. Mura ... hima. F .. & Saiyadain. M . (1966) . Communicator di ..crcpancy . .stilllu lu ~ ambiguity. and intluence. )0111"11(/1 oj P('I"w!I/(/lily. 34. 262~274. l ... ocrate:-.. ( 1976). Antidos is. 1.lOcnlles (G. Norli n. tran~.). Cambridge. MA: Harvard Univer ... ity Prt.!s". Loeb C lass ical Library. vol. I . John ...on. B. T .. & Eagly. A. 1-1 . (19~9)' The dlt.!Ch of in\olvement on pcrsua" ion: A l1leta-ani\ ly\i~. P.\y· ('/JnloRical Bulletill. /06.290- 314.
110
Part II • Persuasioll Variables: Perspectil'es Oil Sources, Reail'ers, Challllels, all(/ Messages
Johnson, H. H. (1966). Some cffc(;ts of discrcpan(;y le\cl on responses to negative information about one's self. Sociolllelry. 29,52-66. Johnson. 1-1. I-I. , & Scileppi. J. A. (1969). Effect" of ego-involvcment conditions on attitudc change to high and low communicators. Journal of Persona/ily and Socia/ Psychology. /3, 31-36. Kahlc. L. R .. & Horner. P. M. (1985). Physical attractiveness of the celebrit), endor~er: A social adaptation perspective. Journal of ConSlllller Rese(lrch, II, 954-961. Littlejohn. S. (1971). A bibliography of qudies relatcd to variables of SQur(;e credibility. In N. A. Shcarer (Ed.), Bihliogmphic Allllilal ill Speedl COllllllllllimlioll. 2. 1-40. M(;Croskey, J. C .. & Teven. J. J. (1999). Goodwill: A reexamination of the con'otruct and its mcasurement. Communicmioll MOl1o[?raph .I-. 00. 90- I 03. Mills, 1.. & Harvey, J. (1972). Opinion change as a function of \\ hen information about the communicator is received and whcther hc is attractivc or exper\. JOllrnal (l Pa.wllaliry (lild Social Psyc/wlogr,
21.52-55. Neimeyer. G. J.. Guy. J.. & Metzler. A. (1989). Changing attitudes regarding the treatment of disordered eating: An application of thc Elaboration Likelihood Model. jOllmlll of Sucial alld Clillical Psrclwlogy, 8. 70- R6. O'Keefc, D. J. (1987). The perslla~ive effects of delaying identification of high- and low-credibility communicators: A meta-analytic rcvic\\. Celllml States Speech .Ioumal. 38. 63-72. Osgood, C. E.. & Tannenbaum, P. II. (1955). Thc principle of congruity in thc prediction of attilllde change. Psych%gical Rt'l'ic\\'. 62. 42-55. O~terhouse. R. A .. & Brock, T. C. (1970). Di:-.tractioll in("fcase:-. yielding to propaganda by inhibiting countcrargui ng. JOIln/al (1 PenOllllliry alld So('/al Psychology. 15. 344- 358. Pcrloff. R. M .. & Brock, T. C. (19RO) .... 'A nd thinking makes it ... o': Cognitive responses to persuasion. In M. E. Roloff & G. R. Miller (Eds.). Per.I'lIasioll: Nell" direclioll.l' ill ,heory alld res('(lrch (pp. 6799). Bevcrly Hill<;. CA : Sage. Peterson, P. D .. & Koulad,. D. (1969). Attitude change a~ a function of latitudc ... of acceptance and rejection. Jourl/al of Personality alld Socia/ Psychology. 11.309-311. Petty, R. E.. & Cacioppo. J. T . (1977). Forewarning. cognitive rC'oponding. and resistancc to persuasion. jOllrnal (!(P('/".\ollalily and Social Psychology. 35, 645- 655. Petty , R. E.. & Cacioppo. J. T. (1979a). Effect:-. of forcviarning of pc.NJa~ive intent and involvement on cognitivc responses and persuasion. Per.IO/wliry alld Social Psych%gy /1l1l/elin. 5. 173- 176. Petty, R. E., & Cacioppo, J. T. (1979b). Is:-oue involvement can increasc or decreasc pcrsuasion byenhancing me~'oage-relcvant cognitive proce'osc~. Journa/ Pl'I"wllafitv afld Socia/ PSTC!IO/0f.:Y, 37. 1915- 1926. Petty, R. E., & Cacioppo, J. T. (19R la). AtIilHde,I' lIlId persuasioll: C/a.n/r (llId ('OJllempomry approaches. Dubuque, IA: William C. Brown. Pctty. R. E .. & Cacioppo, J. T. (198Ib). IS'oue involvcment as a modcrator of the efli:cts on attitude of advcrtising content and context. In K. B. Monroe (Ed.). Adl'w/ces ill ConslllI/er Rl'.~(>arch (vol. 8, pp. 2(}-24). Ann Arbur. MJ: Association for CO[ll,umer Research. Petty. R. E .. & Cacioppo. J . T. (1984). The effects of involvcment on responses to argument quantity and quality : Central and peripheral routes to persuasion. jO/lI"l/(I/ (~f Persol/ality and Sodal Psy{"/w/O[?Y, ·M. 69-R I. Petty, R. E.. & Cacioppo, J. T. (19R6a). Communication (llId per.I·lIl1siol1: Cl'nlral and penj)heral /"()lIte.~ /() altitllde change. New York: Springer- Verlag. Pett),. R. E.. & Cacioppo. J. T . (I 986b). The claboration likelihood modcl of persuasion . In L. Berkowitz (Ed.). At!I'(//lces ill experiml'll/a/ s(lcial psychology (vol. 19. pp. 123-205). San Diego, CA: Academic Press . Petty. R. E.. C\('."ioppo. J. T .. & Goldman. R. (1981), Personal involvcment a~ a dctcrminant of argumentbased persuasion. journa/ of PersoJ/ality {llId Social P,I·yc/wlogy. 41, R4 7- R55. Petty, R. E.. Cacioppo. J. T.. & Schumann. D. ( 1983). Central and pc.ripheral routcs to advertising effectiveness: The moderating role of in\'olvcment. journal ofC(m.I/lmer Research. 10, 135-146. Petty. R. E.. Haugtvedt. C. P. , & Smith, S. M. (1995), Elaboration a~ a determinant of attitude strength: Creating attitudes that are per~i~tent. resistant. and predictive of behavior. In R. E. Pctty & J. A.
or
Chapter 6 • Sour£'(' Crt'tlibi/iry (lml file EI(I/uJrtllicm Likt' /ihood Model
III
Kro ... nid (E(j.,.). Arrillld(' \lrt'lIt:tli : AlIIl'cetlt'I1t.\ lIl1ti {·(IIIWqlll'l1ct'.\ (pp. 9]· lJO). Mahwah. NJ : Erlbaum . Petty. R. E.. O'Mom. T. M .. & Brock. T. C. (Ed ... . ) (191'1 I). Coglli/ile re.~p(mle.( ill pen/IOrio". Hilhdale. NJ: Erlhaulll . Petty. R. E.. OMrolli. T. M .. & Brock. T. C. ( 198 I). lI i ... torical foundation ... of the cognitive respon ...c approach to aHitllde~ and pcNw . . ion. In R. E. PeIlY. T. M. O~trol1l. & T. C. Brock (Eds.), Cv!tni/il't' respo//Il'l ill pl'rIUa.'iioll (pp. 5- 29). Ilill ...dale. NJ : Erlbaum . Petty. R. E .. Schumann. D. W .. Richman. S. A .. & Strathman. A. J. (1993) . Po ... iti,,·\! mood and pe~ua ... ion: Different role ... for affect under high- and 1()~-clahor.lIl()n condLtion .... Journal oj Perm1l11lin' alld Social Pryt"iw/ol-:Y, 6-1. 5 -- 20. Petty. R. E.. Well .... G. L.. & Brock. T. C. (1976) . Di ... traction can enhance or reduce yielding to propa· ga nda: Thought disruption ver ... u ... effort ju ... tilicalion. JOIII'II(l1 of Pt'r.IOl1l1lirl ami Social P.l yc!wl08.1'. 14. g74-884. Rhine. R. J .• & Kaplan. R. M. (1972). The dTcct of incn.:dulity upon evaluation uf the ... ourcc of a COllununication . Jmlnwl oJSocia/ PnC"lwlogy, 88. 255- 266. Rhine. R. J.. & Sc\crancc. L J. (1970) . Ego.in\"ohement. di ..erep..1nC) ....ource credihllity. and altitude change. JOfln1ll1 oj Persollali,,· amI Social Pn'c/w/og\", 16. 175- 190. Ro ..cr. c.. & Tholllp ...on. M. (1995). Fear appeah and Ihe formation (If acti\"e puhlic ... Journal oj Commun;cathm, 45. 103- 121. Shav itl. S .. SW(ln. 5 .. Lowrey. T. M.. & Wan~c. M. (1994). The intcraCli()Il ofcmlnr.. er attractivene ... s and invo l\ement in pcr ... uasion depend ... on thc goal that guide .. I1lc ...... agc proce ...... mg . JOllrnal of COl/ slImer Psrdwlogy, 3.137- 162. Simons. H. W .. Berkowitz. N. N .. & Moyer. R. J. (197()) . SlIniiarity. credibiht} .•md attitude change: A review and H theory. PsycllOlogiml8ullerlll, 42. 21'15- ] I'" Sivacek. J., & Crano. W. D. (1982) . Vc~ted intere .. t a .. a Illotlt.:ratur of altitude-behavior con ... islcncy. Jourl1al of Pl'r.H)//lllity Will Social P\H'lwloKY. 43. 210- 221. $tern thal. B .. Dholakia. R .. & Leavitt C (1978). Thc pc.:r... ua:.ivc cfft.!cI of ..ource credibility: A situational analy<;i<;. Pflbiic Opillioll QIIMIl'rh. 41. 285- ] I... . StilT. J. B. (1986), Cognitive pnlCc ...... lng nf pcr... ua ... i\c me ......agc r.:ue ... : A met .. -analytir.: re\ icw of Ihe effect ... of \Upportlllg infonnation on allllude .... Commlllliwtio" MOfwgrapil.\, 53. 7.5 -·89. Stiff. J. B. ( 199... ). Per.\/ws;l·e ("(JllllflWlimliol/. Nc\\. YorJ... : Gui Iford Prcs,",. Tannenbaum. P. II (1967). Th\! congruity theory rc\ i... lled : StudlC'" in the reduction. Induction. and geller· aliza t ion of per~ua~ion. In L. Ber~owit7 (Ed . I. Adnlll("('1 11/ t' lf1erimt'lfw/ we/al psychology (\ 01. 3. pp. 271 - 320). New YorJ...: Academic Pre ....... W'lrd. C. 0 .. & Mt:Ginnie~. E. (1974). Per...lw .. ive dTc(:t of early and lale mention of r.:redible and noncredible !-()urce~. Journal oj P\\,cllOl0!ty. 86. 17- 23. Wi l<;on. E. J .. & Sherrell. D. L (19931. Source effel'''' in l'onlillullIcation and peNIa .. ion re~earch: A meta· anal) ... i<.; of effect size. Jo"mal of Ihe Act/tie",,· of MlIrkl' linl: S("it'nct'. 21. 101 I 12.
7 Argumentativeness, Verbal Aggressiveness, and Persuasion Andrew S. Rancer
Illtroductioll Over the la~t several decades. scholars in the communication discipline have advanced many models of persuasion. For several years, the ;'onc to many" model prevailed. In that model. scholars were generally concerned with how source and message variables influence large groups of people. Since then, a new emphasis has emerged. The publication of a seminal study on compliance gaining by Miller. Boster. Roloff. & Seibold (1977) foclised renewed interest and attention on illferpersollal persllllsion. This model concerned itself with issues such as how we persuade people to do something we want them to, how we resist attempts by others to influence us, and the role of aggressive communication in interpersonal relations (Infante, Rancer, & Womack, 1997). Very often, such compliance-guining artempts involve the use of argumentative beIUII'io,.. That i~, when attempting to persuade other~, sometimes people present arguments r.,upporting the position(s) they arc advocating. while attempting to refute the position(s) of others. These compliance-gaining attempts can vary considerably. For example, they might involve arguing with a variety of people. including strangers or intimates. They can also involve arguing over both "major" issues. such as where you and your partner ~hou ld take your annual vacation, and more "minor" ones. such as what flavor toothpaste you should buy. It should be noted, however. that a minor issue to one person may be seen as a major issue to another. The manner in which partners communica te during these compliance-gaining efforts can help determine whether their relationship will be seen as sat isfying or unsatisfying. or whether the compliance-gaining attempt will be successful or unsuccessful. For example, a person who communicates aggressively, attacking the self-esteem of his or her partner, might damage the relation~hip and be less persuasive as a result. Accordingly. an
113
l14
Pan II • Per.Hw.lioll Varia hIes: Per,I/H' criw>.1 Oil Sou fee ,I-, Recein>r.I. Chwllleis. and Me.lsage.1
understanding of argumentativeness and verbal aggressiveness can facilitate a better understanding of persuasion in one-on-one settin gs. With thi~ co ntext in mind, the focus of this chapter is a discussion of interpersonal (or informal) persuasion as it is ex hibited via aggressive commun ica tion.
A Conceptualization of Aggressive Communication A commun ica tive behavior is "aggressive if it applies force ... symbolicall y in order. minimally. to dominate and perhaps damage, or maximally, 10 defeat and perhaps destroy the locus of the attack'" (In fante. 1987a, p. 158). To pres~urc someo ne inlO behaving a particular way, for example, a persuader mi glll attack that person's se lf-estee m by using profanity or by calling the other person mimes. Though Illany behaviors are aggressive in nature, not all arc as "bad" as this example suggests. To be sure, some aggressive behaviors are "bad" while ot hers are "good." Let us say. for exa mpl e, that you hire an attorney to defend you. More than likely you would not want hin) or her 10 sit passively in the courtroom while th e opposing attorney "trounces" you. Rather. you would want your own attorney to respond aggressively, offering a barrage of arguments and convinci ng legal points. In thi s example, aggre!
Assertivel/ess alld Argumel/tativelless: The COllstructive Traits As~ertivelless
and argumentativeness arc the constructive traits. AsserliI'ellf:!.B is the more global of the two. If you are asseflive, you tend to be interpersonall y asce ndant. dominant. and fo rceful, using this behavior 10 ach ieve personal goa ls wh ile creat in g positive affect in people. If you are highly ar~wllefl1l11il 'e, you tend to advocate and defend positions. 011 controversial issues while attempting to refute other people's positions o n thos.e issue~ (Infante & Rancer. 1982. 1996), Argumentativeness is a subset of assertiveness. because all argument is assertive, though not all assertiveness involves argument (c.g., a request). Depending on their predisposition to argue, people ca n be considered either high 01 low argumentativcs. In addition. th ere are at least two types of moderate argumentatives: those with con llicting fee lin gs and those who are apat heti c. in each case. the sit uati on influences whether slich people c ng
Ch.lplt:r 7 • ArgUlI/l'flflllil'ell(' \', Ferhal Aggn' nil'etleH. (llld PerHUHio/l
liS
"apathetic" generally argue \.\ hell the inccnti\'c of :-.ucce!>.\ b high. They ncither like nor di ... li"- e arguing and l!ngage in it mainly for utilitarian rea~on:-..
Hostility amI Verbal Aggressivelless: The Destructive Traits When peop le "argue:' thcy can move from constructh c. argumentative behavior to another form of aggrl! ... ~i\c communication bcha\ior. That 1~. with consciou\ intent. interactanl ... can get "mean" or "hurtful." Thi ... switch in interaction \tyles can occur when people believe that they an: losing the argument and/or when they have trouble genera ting additional argument ... to cmplo) in \UPport of thcir po ... ition on a controversial is ... ue. The t,..'o dc ... tructi\'c trait ... in the model of aggrcS\ivc communica tion arc hostility and verba l aggres ... h-enes .... /-Imliliry is more g lobal. People with this trait lise messages to exprl!:-. ... irritability. ncgativi ... m. rescntmcnl. and ",u"'picion. People high in I'erhal (lIU.:re\si!·el1e.\,\ tend to attac~ the ... elf-concept\ of othe r people in order to inflict psychological pain such as humiliation. el1lharra~sl11cnt. dcpres~ion. and other negative fl!cling~ about ... e lf ( infaille & Wigley. 19X6). The model of verbal , were thought to be equally liJ..cly 10 be high verbal aggressive". Although thiS a~~umption wa~ supported by several earlier "'lL1dil!s (c.g .. Infante & Rancer. 1982: Infante & Wigley. 1986). more recent empirical re\earch has ... ugge\teu that at lea ... t for somc populations (e.g .. adolesccnts). argul11ent:.1livcnes ... and \erbal aggressivenc\S arc moderately correlated (Rancer. Whitccap. Kosbl!rg. & A vtgis. 1997: Rancer. A vtgi .... Kosbcrg, & Whitecap. 2000: Roberto & Finucane. 19(7).
116
Part II • Per,HUlS;OI/ Var;ahles: Perspectil'es 011 Sources, Receivers, Channels, alld Me.\'Sage.\·
Ullderstandillg Aggressive Commullicatioll ill Persuasioll: Situational Factors Also Matter So far. we have seen that peoplc's traih influence th e deg rec to whi ch their be hav ior is aggres~ive. In addit io n to the tra it component. s itu ation a l factors are ass um ed to influe nce aggressive commu ni cation behavior. In ot her words, aggressive com muni cati on ca n be understood as a joint product of si tu atio na l factors and th e characteristics of th e person (Ande rscn. 1987; Atkin son, 1957; Eps te in , 1979). This constitute s an inrel"{lclioliisl ap~ proach to persOIwlitv (e.g .. Epstein, 1979; Mag nu sson & Endler, 1977). Co nsiste nt w ith this approach, Infante's ( 1987 b) model of argum cmati ve ness in cluded both trait and situatio nal factors to predict more accurately how moti va ted a person will be to argue in any givcn situation. This model suggests that a perso n's moti vat ion to argue in a give n situation is determined by hi s or her trait argumentativeness (A RG g l). as we ll as percept ions o r how likely he or s he is to succeed or fail and how important it is to ~ucceed in the g iven si tu at ion. A n example of this interact ioni st approach wou ld be the following: If Brandon thinks his a rgument will probably fail. and he doesn' t mind failing. his motivation 10 arg ue may be low. O n the o ther hand, jf he think s he will succeed. and success is important in the parricular s ituation. hi s mo ti vation to argue may be hi gh. A high motivation to argue. coup led w ith hi g h trait argumentativeness, s ho uld increase argume ntative ness cve n funhcr. Several stud ies have slipporrcd th e cent ral co nten tion of thi s interactionist model. that predictio ns based o n traits or situ atio ns alone are not as accurate as predictions ba~ed on both traits and situati o ns ( Infante, 1987b; Infante & Rancer, 1982, 1993; Stewart & Roach. 1993). For example. one s ituatio na l faCIal' Ihat might influcnce argumcntative behavior is the nature of th e top ic itself. To be ce rtain , Onyekwere. Rubin , and Infant e ( 199 I ) found that ego-invo lvcmcll t in th e top ic of an argument innllenced the bchavior of high and low arg um en tati ves. Specificall y. when people were highly in vo lved in an iss ue (e.g .. a gu n store ow ner arguing about g un con trol). they were not only more mo tivated to argue. the ir argumentat ive bc havio r was improved as well. Similarly. Infant e and Rancer (1993) fOllnd th at high argumenlatives argued more about certain topics than did moderate and low argumcntatives (c.g .. soc ial. political. pe rsonal behavior. others' behavior, and moral-cthical issues) but behaved no differently th an low and moderate arglll11entatives o n olher topics (e.g .. sport s. emcl1uinm cn l. rel ig iolls issues). Finally. based upon thi s interactioni,( perspec ti ve. Stewart and Roac h ( 1993) in vest igated how a situational faclor (e.g .. the topic of an argumen t) and a characterist ic of the communicator (e.g., rel ig io lls o ri e nt ation) influenced arg um entat ive behavior. They found that in ge neral. trait argu lll enta~ tivcnes'l was the most impol1ant fac tor associated with a person's willingness to argue. However. ext rin sicall y rel ig ious persons (th ose who view religion mainl y as an instrullle ntal soc ia l conventio n and are more utilit arian and irregular in th eir churchgoing behavior) were morc argumentative than were e ither intrinsically religious perso ns (thosc who view religion as the source of mean in g in life and arc regular in their churchgoing bchavior) o r pro re li gio u ~ persons (tho . . c who have hi g h leve ls of intrinsicness and hi g h leve ls of extri nsicness) (S tewaJ1 & Roach. 1993. p. 28). Nonreligious persons were also mo re will ing to argue than were proreligiou~ pe r~o n s.
Chapter 7 • A rglll1lt!lI/mil'cllcss. Verha! AKK ressil 'elle,~.,-. {/nd Persl/asio/l
117
In addition to the topic of argumelll. characteristics of the communicators can act as determ inants of argumcnlat ive behavior (Wigley. 1998). One such factor is a person's gender. For example. Infante (1989) examined what types of messages males and females used when trying to persuade pcoplc. Male and female participaJ1l~ were asked LO imagine themselves trying to persuade a friend. After each attempt. the friend responded with either argumentative or verbally aggressive mcs~ages. Moreover, after each of the friend's rejections, participanh were asked to choose one of four follow-up strategies: rewarding, punishing. argumentative. or verbally aggressive messages. Results indicated that when the target was verbally aggrcs~ivc. males werc more likely than remales to respond with verbal aggression . whereas females were more likely that males to respond argumentatively. Another situational factor Ihat may influence one'~ argumentativeness includes the nature or behavior of onc's partner. For example. Rancer and Infante (1985) found that the argumentativeness of an individual and that of his or her adversary interact to determine motivation for arguing in a specific situation. Waggenspack and Hensley (1989) examined how the interpersonal situation (socially oriented or conflict oriented) and gender influcnce preference for having a partner who is argumentative. Lim (1990) found that situational factors. such as the friendliness and intensity of resistance by a target of persuasion. are important in identifying whether a persuader will exhibit verbal aggressiveness. Specifically, persuaders were more verbally aggressive when targets of the persuasive attempt were unfriendly than when they were friendly. In addition. persuaders resorted to higher levels of verbal aggressiveness more quickly when targets exhibited more intcnsc resistance to the persuasive attempt. That is, when individuab fclt that their efforts at persuasion were going to fail. they were quicker to resort to higher levels of vcrbally aggressive comlllunication (Lim. 1990). Along these samc lincs, Infante, Trebing, Shcpherd. and Sceds (1984) found that verbal aggressiveness is dependent upon the individual's argumentativeness and the obstinacy (i.e .. stubbornness) of his or her opponent. A study conducted from the perspective of the receiver of the message (Infante. Wall. Leap. & Danielsoll. 1984) found that the gender of the message sOllrce determined whether more argumentative people received verbal aggression. In summary. these studies indicate that situational factor" do indeed intluence aggressive communication and as slich should be considered along with trait predispositions when attempting to explain a person·s behavior during social influence.
Argumentativeness, the Interactiollist Perspective, and the Theory of Reasoned Action Stewart and Roach (1998) tested two competing theoretical framework" for determining a pcrson·s ··intentions to argue'·: the interactionist perspective and the Theory of Reasoned Action (TRA) (Aj7en & Fishhein. 1980: Fishbein & Aj7en. 1975). Recall that the ·' intcractionist perspective·· maintain . . that people's mOliv<.uion to argue is dctermincd by the interaction of situational factors and traits (i.e .. argulllentativene"is) as well as the
118
Part II • Pa\'/ws;()// Var;ab/e,\ : Per,\ f}l'cti!'e,\
01/
Source.\ . Rece;I'l'r\'. ChallllelS. alld MeH'age,~
perceived probability and importance of success and failure in a given situi1 ti oll. The TRA maintains that a person's decision to engagc in a purposeful activity depends on several factors. some of which are situational and somc of which arc mediated by personal dispositions or tmits (for a morc detailed description ofTRA. see chapter 4), Stewart and Roach (1998) argued that the TRA might have "greater explanatory power" th:m the interactionist framework for understanding argumcntat ivc intention~. because the TRA takes into account a greater number of situational factors (Stewart & Roach. 1998. p. I 82)-for example. people's belief... about arguing (e.g., Rancer. Baukus. & Infante. 1985; Rancer. Kosberg. & Baukus. 1992). the importance (ego-involvement) placed on the issue of the argument (e.g .. Infante & Rancer. 1993: Onyekwere. Rubin. & Infante, 1991 J. the effects of other e1i'positional factors (e.g .. Stewart & Roach. 1993). and the influence of other people. It WHS reasoned that all of these factors might combine to influence a person's intentions to argue in a given situation. Since arguing is an intentional behavior under the control of the arguer, it can legitimately be called a form of "reasoned action" (Stewart & Roach. 1998). The re,ults of Stewart and Roach', ( 1998) study showed that the TRA was no better than the intcractiollist model at predicting argumentative intentions. Indeed, the TRA was less powerful than the interactionist model in prcdicting motivation to argue. This prompted the researchers to state. "As such. the interactioni't model warrants further use as a framework for the study of trait vcrsus "iiluational dcterminants of argumentative intentions" (Stewart & Roach 1998. p. 191). However. Stcwart and Roach found that peoplc's attitudes toward arguing in a particular situation. coupled with their beliefs about what people who arc important to them think about arguing. are the primary determinants of intentions to argue. Their findings abo revealed that whereas high argumenlatives generally had a more positive attitude about arguing than did low argumentatives, the direct effect of those attitudes on intentions to argue came from sources other than trait argumentativeness (i.e .. the normative component and the perceived behavioral cOlltrol portions of the TRA-see chapter 4), This finding sugge ... ts that beliefs and motivations to argue may be more socially driven than individually deterillined. Stewart and Roach (1998) speculated that because high argumentativcs are more competent communicators. they may be more open 10 nonnative pressures in deciding whether to engage in an argument. as we ll as experienc ing strong social pressure to perform well in
Identifying Beliefs About Arguing One way tn understand aggressive communication is to study the beliefs people havc about arguing and employing verbal aggression. For example, Rancer, Baukus, and Infante (19X5) found that people maintain several be lief Mructures about arguing. These include the following:
• Hostility- the b~lief that argu ing is a combative encounter. • Actil'ilylprocess- the belief that arguing i~ a mode of intcraction. like having a conversation.
Chapter 7 • Arglllllelltati\'eness, Verbal Aggressil'efJe~'s, lind PerSllasion
119
• Control/dominance-the belief that arguing is a way of having onc's opinIOns prevail. • Conflict/dissollallce- the belief that arguing is a source of conflict or dissonance between antagonists. • Self-image- the belief that arguing is an encounter that impacts on a person's sense of self. • LeaminR- the belief that arguing is a way to learn about self. others, or the environment. • Skill-the belief that arguing is an indicator of one's verbal skills. These belief structures were found to distinguish individuals who vary in argumentativeness. A greater proportion of high argumentatives had positive beliefs about activity! process, control/dominance, conflict/dissonance, self-image, learning, and skill. More low argumentatives held negative beliefs of hostility, control/dominance, and conflict! dissonance. In addition, high, moderate. and low argumentatives were found to have different perceptions of two functions of arguing: cilitiwaioll and {l/Hagolli.<;/ll. High argumentatives view arguing as a source for cultivating information. whereas low argumentatives see it as a behavior that reveals their lack of argumentative and rhetorical competence. Moreover, while high argumentatives view arguments as a means of reducing conflict. low argumentatives see them as unfavorable and hostile acts to be avoided at all costs (Rancer, et aI., 1985). Understanding such beliefs about arguing is significant because it may help individuals function more effectively in compliance-gaining. persuasive, and conflict management situations. For example, some spouses have difficulty managing their communication when they are trying to influence each other. Il may be that spouses' differing beliefs about arguing contribute to the tension they experience. The husband. for example, may possess a positive belief about arguing as learning (a way to gather information), and manifest this belief whenever he tries to understand his wife's position on an issue on which they disagree. His wife, however. may possess a negative learning belief that causes her to become agitated and uncomfortable when her husband begins "arguing" with her (i.e., when he asks her probing questions concerning the position she holds, or continues to probe her position beyond what she feels is a "reasonable" amount of time). Following this same vein, Rancer, Kosberg, and Ballkus (1992) examined which beliefs best predict argumentativeness and which best discriminate between high and low argumentatives. Results of their study indicated that high argulllentatives believe that arguing has enjoyable, functional. and pragmatic outcomes, as well as having a positive impact on their self-concept. Low argumentatives' beliefs about arguing lay in the opposite direction. Using this "beliefs about argui ng" framework. RanceI' and Baukus (1987) concluded that males and females differ in belief structures about arguing, with females holdi ng more negati ve bel iet's than males. A related line of research has investigated beliefs about engaging in verbal aggression. For instance, Infante. Riddle, Horvath. and Tumlin (1992) round dirrerences in beliefs about verbal aggression between those who vary in the trait. Specifically, high verbal aggressives believed that competence attacks, physical appearance attacks. and threats were less hurtful to others than did individuals low in verbal aggressiveness.
120
Part II • PerSI{(u;un Variables: Perspect;\'es Oil Sources. Receil'ers, Chal/lle/s. anrl Messages
A Transactional Approach to Argumentativeness and Persuasion A "transactional approach" to personality sugge~ts that an individual':-. personality may inlluence olhers' behavior as well as his or her own. This approach a"ume, Ihal one person's me!-lsage behavior affects the other's message behavior and vice versa. Levine and Boster (1996) expanded our understanding of argumentati vencss and persuasion by incorporating a transactional approach to examining argumentative bchavior. Their study involved observing per'\uasive interactions (i.e .. five-minute arguments over controversial issues) between individuals who were matched and mi!'!matched on argumcntativenes!'! (i.c., high argumentatives talking with high argulllcntatives. low argumentatives talking with low argumentatives, and high argumentatives talking with low argulllentatives). The researcher!'! studied the impact of each actOJ"s argumentativeness 011 the number of arguments and the type of resolution generated in these arguments. The results revealed that both convcr!-lational partners' levels of argumentativene!oo!-l interacted to influence both the number of argument!-l generated and the extent to which one person yielded to the po~ition advocated by the other. In general, the "high argumentativeness- low-argumentative other" condition emerged as the one in which persons generated the greatest number of arguments and showed the most resistance to yielding to the other. More specifically . highly argumentative individuals were more argumentative when paired with a low-, rather than high-argumentative partner. The result~ also revealed that the argumentativeness of the adversary did not substantially affect the argumentative behavior of low argulllentativcs. In order to explain these findings. Levine and Boster (1996) speculated that high argumenlatives like winning and. when paired with a low argumentative partner. seile thi s opportunity to demonstrate their superior argumentative skill. They also sllggc~ted that when paired with an equally argumentative adversary, the high argumentative might experience some frustration in his or her inability to dominate his or her adversary. which would then be reflected in his or her less assertive behavior. These finding!-l. however. are conlrary 10 earlier research by Rancer and Infanle (1985). who discovered Ihal highly argumentative individuals reported more motivation to argue when paired with a !-limilar adver!'!ary. While the Levine and Boster (1996) findings may have more limited general izability (only males were included in the study). they do underscore the value of lIsing a transactional approach to studying aggressive communication and per!-luasion.
Processing Persuasive Messages: The Influence of Aggressive Communication Predispositions In addition to investigating how argumentativeness and verbal aggressiveness affect the manner in which people seek compli ance, research has explored how these traits affect people's processing of persuasive messages. FirM, Hample and Dallinger (]987) examined how people "edit" their own argument!, before they actually deliver them_ According
Chapler 7 • Argllmelltatil'e"ess, Verbal Aggressil'elless. a"d Persllasio"
121
to tradition, there are two pans lO the process of argumentation: people create or invent arguments, and then they select which ones to present during a compliance-gaining attempt. Hample and Dallinger's study sought to investigate the second part. that is. "why is one argument offered, and another suppressed?" (p. 124). In previous research, Hample and Dallinger (1985a. 1985b) identified four general categories of cognitive editing standards that most people use:
1. Ejjectil'elless-people reject certain argument!'! because they feel the argument won't work or might backfire. 2. Principled objections-people reject arguments because they have disdain for the type of argument strategy (e.g., threats. bribes). 3. Persoll-ceflfered issues- an argument may be rejected because it violates the arguer's self-image. might injure the adver~ary, or might do irreparable harm to the relationship. 4. Discourse competence- people might reject an argument because it is judged to be false. too easily refuted. or irrelevant to the conflict. Hample & Dallinger's (1987) study examined how argumentativene". verbal aggressiveness, interpersonal orientation, and gender innuence cognitive editing. Their results demonstrated that people who vary in argumentativeness and verbal aggressiveness use different cognitive editing standards in suppressing arguments. More specifically. females. people high in interpersonal orientation (those who are 1l1Ore sensitive and attuned to the personal characteristics of a relationship and more responsive to a partner's behavior. Swap & Rubin. 1983). and people low in verbal aggressiveness are less likcly to use the effectiveness catcgory or to endorse arguments in general. Instead, they are likcly to use principled objections and person-centeredness (e.g .. harm to others) to suppress the use of persuasive appeals. Results of the study also demonstrated a tendency for people high in argumentativeness to cndorse more compliance-gaining messages. In short, it seems clear from this research that the argumentativeness and verbal aggressiveness traits do influence the cognitive editing of arguments. Second. Mongeau (1989) studied how argumentativeness and need for cognition impact persuasive message proce:-osing. Need jor cognition has been defined as "enjoyment individuals derive from engaging in effortful information processing" (Cacioppo, Petty, Kao, & Rodriguez. 1986. p. 1(33). In other words. people high in the need for cognition like to scrutinize messages more than those who are low in the need for cognition. An experiment was conducted to explore whether argument quality (i.e .. high quality versus low quality). need for cognition (high, low). and argumentativeness (high. low) influenced participants' attitudes toward a proposal (e.g., comprehensive exams for undergraduate students) and their behavioral intentions (to work for or against the proposal) (Mongeau, 1989). Results of the study suggested that while low and high argumentatives did not differ in their perceptions of the higher-quality arguments, high argumentatives did perceive the lower-quality messages to be weaker than the low argumentatives did. In addition, high argumentatives exhibited significantly greater attitude- behavioral intention consistency than did low argumentatives. Finally. trait argumentativeness influenced persuasive message processing and the relationship between altitudes and behavior in a very similar
122
ParI II • Penmasioll Variables: PerspeClil'e.\ 011 Sources, Receit'ers, Channels, and Mes.m ges
fashion to the need for cognition. This finding further ~upports the notion that argumentativeness has a cognitivc as well as behavioral componcl1l (Mongeau, 1989). Third. Kinney and Segrin (1998) discovered that people's ability to process information, their sensitivity lO feedback, and their beliefs about themselves can make them susceptible (or impervious) La the negalive effects of verbally aggressive messages. Specifically. people who are less certain about themselves are more likely to experience negative emotions when friends who are usually supportive behave in verbally aggressive ways (Kinney & Segrin, 1998). This finding was explained by Expectancy Violations Theory (Burgoon & Hale. 1988) as follows: When supportive friends engage in verbal aggression. the emotional effects of this behavior are significant. because these friends have violated expectancies, Sensitivity to feedback may be a characteristic that can "protect" individuals from verbal aggression and as such may be another cognitive moderator of the effects of verbal aggression (Kinney & Segrin. 1998. p. 66).
Argumelltativelless, Verbal Aggressiveness, alld Compliallce-Gaillillg Behavior Research examining the types of compliance-gaining messages that pcoplc tend to :-.elcci has been based on the assumption that per~uaders are generally aware of the choice ... they make. Such choices might be affected by situational factors. For example, HUllter and Boster (1987) argued that a per:-.uader's choice of compliance-gaining ll1essage~ depends in part on the cxpected emotional impact on the target. They presented a model arguing that "the more negative the emotional impact of a compliance-gaining message on the li~ tener. the tess any given persuader will want Lo use the message" (p. 65). Allhough this model might describe most people's selection of compliance-gaining strategies. what if the person selecting a strategy is argumentative or verbally aggressive? The amOLint and type of strategy preferences a person uses may depend on whether he or she i~ high or low in these traits (Hunter & Boster. p. 82). Indeed, Hunter and Boster (1987) advanced several hypotheses examining the relationship between aggressive communication traits and compliance-gaining behavior. These suggested the following conclusions: I. Individuals high in argumentativeness and verbal aggressiveness would be likely to transmit numerous compliance-gaining messages that vary widely in emotional impact. 2. Individuals high in argumel1lativeness but low in verbal aggressiveness would be likely to send numerous messages of relatively homogeneous emotional i mpaci. that is, the messages would be predominantly positive. 3. Individuals low in argumcntativeness and high in verbal aggression probably would send few messages, most of which would be relatively negative in emotional impact. 4. Individuals low in both traits would be likely to send few messages: they might simply request compliance and. if it were not forthcoming. cease the effort (Hunter & Boster. p. 82).
Chapter 7 •
Aq~lIl11efl/{llil'efle.H, Verbal Ag~rl>Hi,'efles .... alld PenutHiol/
123
While these speculations await empirical examination, they did lead the way for an examination of aggressive coml11unication traits and their impact on compliance-gaining behavior. One of the tirst efforts was conducted by Reynolds (1987). who studied the effects of argumentativeness. assertiveness. and neetl for cognition on the selection of compliance-gaining strategies. Among the major findings was that proaclil'e assertil.'elless (being forceful and ascendant in support of self. c.g .. "When .,tanding in line and a person pushes in front of me. I tell them to go to the back of the linc") and argumem (l\'oidance (the general tendency to avoid arguments) are associated with the use offeu'er compliancegaining strategies. This suggests that proactive assertiveness and the tendency to avoid arguing may be associated less with overt compliance-gaining stralegies than with "subtle or nonverbal suasory message strategies" (Reynolds. 1987. p. 15). Boster and Levine (1988) and Boster. Levine. and KalOleas (1993) replicated as well as extended this line of research by examining how argumentativeness and verbal aggressiveness correlate with compliance-gaining message choices. In general. both studies found that. compared with low argumentative~. high argumcntatives u~ed a grealer variety of strategies and were generally more persistent (8m.ter & Levine. 1988. p. 117). In addition, verbally aggressive individuals used more negalhcly oriented compliallce-gaining messages (Boster. Levine. & Ka7oleas. 1993). perhaps due to their lack of skill in arguing, which impedes their ability 10 create and use compliance-gaining strategies thal are more "po~ilive" in nature. In an interesting twist on this line of research . Ifen und Bcarden (1998) explored whether argumentativeness and verbal aggre~sivenc!'>s inlluenced the types of appeals individual~ say they would usc when responding lO refusals of interpersonal requests. The researchers argued thai in persuasive situations. individuals often rcspond to refu ... als with two types of mcssages: evidentiary and nonevidentiary . E"idell1iary (/ppel/ls are often referred to as rational appeals beciluse they are arguments (claims) that contain information 10 support a claim (Cody. Canary. & Smith. 1994: Ifert & Bearden. 1998: Reardon. 1991). As such. evidentiary appeals are seen positively and judged more favorably and effectively. NOl1el'idel1lim:r appeals are argument~ (claims) that contain liule or no \upponing material but in~tead rely on simple assertions. These types of appeals tend to be more emotional in nature and are seen less favorably than evidentiary appeals (lfert & Bearden. 1998).
In the study by Ifert and Bearden (199R). participants were asked to imagine themselves trying either to persuade a professor to change a low grade or to innucnce a meter enforcement officer to refrain from ticketing their car. Participants were then given a hypothetical statement that the professor or officer might give to reject their persuasive request. They responded to various refusal statements (e.g .. "You didn't fulfill the assignment guidelines" or "The law ...ays I have to ticket you") by writing out what they would say in response. Participants al ... o completed both the argumentati\enes\ and verbal aggressiveness scalcs. The results of the stutly showed that both types of aggressive cOllllllunication traits influenced participants' choices of appeals 10 the refused requests. People higher in trait argumentativeness reponed conMructing more evidentiary appeal~ in re~ponse to these refusals. In addition, people high in verbal aggrcssi\ene~s reported cOI1~tructing a greater number of nonevidentiary appeals than did those lower in verbal aggressiveness. Ifert and
124
Pan II • Per,\"lu/.\;Oll V(/riable.,: Pa.'pl'ctil·{J,,· 011 SOUfce.\. Re('C'il·l'rs. Challnels.
(/I/d
Mes.w ges
Beardcn (1998) ... uggc~tetl that the ... e rc:--ull~ ~upport the notion that ve rbal :.tggre~si\'enc ... :-Illay be the re ... ult of an argumentativc skill deficiency (scc Infante & Rancer. 1996).
Aggressive Communication and Resistance to Persuasion Somc research in per<.;ua ... ioll ... tlggesl:-- that high argulllcntatives may cnjoy all advantage ovcr low argumentatives in persuasion. For example, whcn compared with low arguillenla· tives. highly argumentative individual<.; arc more cumpetent communicator ... in a variety of contexh (Rancer. Kosberg. & Silvestri. 1992: Richmond. McCroskey. & McCroskey. 1989). have internal loci of control (A\tgi ... & Rancer. 1997: Canary. Cunningham. & Couy. 1988). anu have marc soilltion·oriented contlict styles (Nicotcra. SmilowitL. & Pcar",oll. 1990). Bc<.;ide ... that. argulllcntative people may h<'l\e charactcriqic ... that make thcm more resistant to persuasion. With that in mind. a serie ... of studies has invcstigated the.:: intlue.::nce of argull1entativenes ... (and verbal aggressivencs..,) on resistancc to per..,ua . . ion. One factor that may make an indi\ idual more re"i"tant to per .. uasion is the ability to generate or construct counterargument.... A \tud) by Ka/o!ca.., (1993) ... uggc ... tet.l that high and low argumentative ... Inight differ in thi" ability. Spccificall). K'ILolea ... found high argull1cntati\e ... more resi",tant to pcr",ua<.;ion bccau ... e they generate more countcrargumcnts when they think about a cOllnteraltitudinal Il1c~sage (i.e .. 11 message incon ... i..,tent with one· ... existing attitude). This cognitive activity can be thought of as "another nk" ... agc" in the pcrsuasion context. Thu .... the re ... ults of Ka7olea,,'~ (199~) research may mean that high argumentative.., tentl to "pe.::rsuade thcm ...clvc<.;·· even if they ... hmlld disagree with the position taken b) an "dme"te (ll1f0l1lC. Step. & Horvath. 1997. p. RO). This "cognitive re.::sponse per ... pectivc·· (Ka7olcas. 1993) repre ... e.::nted a ~hift toward ill(m,,£1r.\o11ol argul1Il'lIfa(;OI/ (Infante ct a!.. 1997) and formed the basis or another study. In it. Infante. Step, and Horvath (1997) suggested that arguing with onc ...clr might be mort:! enjoyahle for high argumentative ... than it is for low argumentative<.;. In the ... tudy. participant ... were asked to create eithe.::r a proattitlldinal or counterattitudinal me ... sage on a propo ..... 11 (regi ... tration of all firearm ... in the United State ... ). After encoding these message". attitudes toward the proposal v.'ere me.::asured. Infante and colleagues (1997) found that argumentatives who encoded a cotJlltcrattitudinal mes ... agc (a mes..,age illcon~istent ""ith their existing attitudes toward the proposal) were no les" favorable toward the task than people who engaged in proattitudinal advocacy. Thu .... for high argumcntative .... enjoying an argument i... possihlc e\"en when one·<.; opponent i... oneself. Thi ... wi l lingne~<.; to comtruct argumcnts that oppose one\. currentl) held po ... itions may help explain why argumentativenes ... has been a ... snciate.::d with numerou ... pos itive outcomc ... ( Infante et a!.. 1997; Johnson & Johnson. 1979). Moreover. the.:: ... c findings support the notion that high argumentativc:-. differ from olhe.::r people both cognitively and affcL·tively during arguments. These findings. howc\"er. were not replicated b) Levine and Badger (1993). Using the Cognitive Response Model. the.::y predictcd that highly argumentative individual ... \l,Iould be more re ... istant to per ... ua ... ion than low argumentatives. To test this as~umpt i on. Levine and Badger (1993) had high and low argumcntatives lislen to several persuasive
Chapter 7 •
ArgIl11lel11ar;I'eness, Verbal Aggressil'ellen. and Pt'I"HI(lS;O/l
125
prese ntations on a vari ety of t o pi c~ ( fore~ t co nse rvati on, orga n donat io n. preve nti on of heart att ac ks, aborti on. votin g, better eati ng habits. e tc.). The re~u lt s of th e study ~ ur p ri s ingl y contradi cted what was expected: Hi gh argumc ntat ives de mOn\ lrated ~ i g nifi ca n l l y greater attitud e change in th e directi on orthe message than low a rg u mcn t a ti ve~. T his fi nd in g was al so true across the different message to pi cs. Several ex planations fo r these surpri \ ing a nd co ntradictory fi ndings were offe red. First, since sources c hose th eir own to pi cs. they may have selec ted to pics or posi ti ons th at they alread y favored. Thus. th ey may ha ve heard onl y proattitudinal pers uasive m c~sages. Second. hi gh argullle ntati ves may also have ge ne rated more "pro" me\sages when faced with an acce ptable message. Thus. it was sugges ted that highl y argume nt ative indiv id uals may indeed be more o pen-minded (Levine & Badge r. 1993. p. 76). Finall y, Lim ( 1990 ) in vest iga ted whe th e r rece ive rs' res i!o-tance to compl ia ncegainin g e ffort s led pe rsuade rs to be more ve rball y agg ressive in subseq uent pers uasive effort s. More specifica ll y. th e stud y exa min ed whe th er frie ndlin ess (the ex tent of liki ng toward the receiver) and in te nsity o f resista nce a ffected pe r~ u aders' verba l aggres\ive n e~s, The result s demonstrated th at persuade rs who encou ntered ~tro n g res istance to a pe rsuasive atte mpt ex hibit ed ve rball y aggress ive be hav ior mo re q uick ly th an th ose fac in g weake r res istance. In add it ion. persuade r, engaged in more ve rbal aggressio n when encounte rin g un fri endl y targets. An outri ght rejecti o n of a pe rsuas ive appea l by the targe t made pe rsuade rs the most ve rba ll y agg ress ive of a ll. Thi s study furth er underscores the tran sacti onal nature o f interpe rsonal pers u a~ i o n by ~ h ow i ng tha t pe rsuader.., dec ide on a persuas ive co urse o f acti o n not onl y based on si tu ati ona l and pe r,ona l pre fe re nces but also as a res ult of the res ponses that they receive from the targets of th e ir pe rsuasive atte mpts (Lev ine & 8 m tcr. 1996: Lim . 1990).
The Display of Aggressive Communication During Persuasion The ways in whi ch a perso n perceives con, tructi ve or destructive att ack.., can inll uc nce th e outco mes of a co mpli ance-gaining c lTort. For exa mple. based on No [ton'~ ( 1978. 1983) resea rch. Infante and Gorde n (1 989) ide ntifi ed two cOl1lmuni cato r styles. The first. an aJfi rming cOIlIl11f11riCalUr style. is highl y re laxed . fri endl y. and atte nti ve. and is accompa ni ed by re lati ve ly low levels of ve rba l agg ressive ness. Th i.., style see ms to mediate pc rce pt io ns o f agg ress ive co mmuni cati on so as 10 y ie ld more pos iti ve than negative o ut co mes (Infant e. Anderson. Martin. Herin gton. & Kim . 1993: Infante & Gorden. 1987. 1989). For instance. indi vidual s w ho e ngage in argume ntati ve be hav ior a nd do so with an affi rmi ng communicator style sec m to make the argume ntati ve behavior appear more "pa latabl e," Con versely. if ..,omeone e ngages in arg ume ntativeness accord in g to the seco nd. J/OJ/affirming communicator slyle, whic h is hig hl y agi tated. un fri end ly. a nd inatte ntive. he or she Illay be mistak enl y rega rded as engaging in ve rhal aggress ion. Empi rica l suppo rt for th ese ass umpti o ns was obtained in a stu dy co ndu c ted by Infante. Ranee r. and Jordan ( 1996). They suspected th at o bse rve rs wou ld be Ie" like ly to overes tim ate ve rball y agg ressive messages and more likely to pe rceive argume nt ati ve behavior when they read tra n sc ript ~ of a co mpliance-gai ni ng effo rt marked by an affi rming
126
Part II •
Per.HllI.\iOIl Vllriahk\: Penpee(/I'es 01/ SO/lrcC'.\. Rl'('('il'en. Clulllllel.\,
(Jill/
M('\\tlgn
vers us a nonaffinnin g cOlllmunicator sty le. In the ... rud y. participants read a transcript of a conversation depicting a co mpliance-gaining effort between two roomlllate .... The cO llye rsation desc ribed a conflict that arose. c lima xed. and then en ded in a re"olution or so lu tion. All Matemcnts in the tran scripts perlained to th e i\,>ue und er co nt rove r"y. exce pt for several state ment s made by both parties lh at were in fact ve rball y agg ressive. Result" of th e stud y found more arg umentati ve behavior. alo ng with le'>s ve rbal aggression. when th e messages were presented with an affirming cO llllllunica tor .\ lyle ( Inbnt e. Rancer. & Jordan. 1996). In addition. fewer mistakes were made in identifying \crbaJ agg rc ... s ion in that text. A similar study (Rancer. Jo rdan, & Infante. 2000) supported the earli er finding ..... Parti c ipants \ ign ifi can tl y overestimated th e amount of ve rbal aggre'>sion whe n individ u,l l, communicated with a nonaffirming co mmunicator style. rega rdl ess of whether conve rsations we re prese nted on videotapc~ o r o n written tra ll ~cripl '>. These findin gs haH: c lea r implications. underlinin g the important role of nonverbal behavior in mediating pen.:eption~ of construct ive and destructive behavior during compl iance-gaining atlempt .... That is. when involved in a pcr~uas i ve crfort. individu als shou ld engage in argumentative bchavio r but makc sure to do so in an affi rmin g (i.e .. hi ghly relaxed. friendl). and attenti\'e) way.
Aggressive Communication and Persuasion ;11 the Organizatiollal COlltext: What Makes for a Persuasive Supervisor? Th e quc!'.tiol1 of what makes a pcrsua ... ive \ upeni ... or ha\ bc.!cn inve'>tigated in a number of studies (Go rde n & Infante. 1987. 199 1: Gorden. Infante. & luo. 1988: Infante & Gorden. 1987. 1989. 1991). Much of thi s research has been ba"d on the Indepe nden t- Mindedne" Theory of organizational co mmuni cat ion . Th e basic tenet or thi s th eory s u gge~ ts that the values held by th e ge neral society s hou ld b(! affirmed in the work pl ace, A\ such. "A meri ca n" corporations should e ncourage free speec h and promote individualism and independent-mindedness. This body of research hLl ... re vea led th at superiors who afe high in argul11cntativenc ... s and low in ve rbal aggressiveness (and who comm un icate w ith an afiirming style) h,ne a numbef o f positive qualities. They arc perceived a ... encourag in g subordinates to cxpres ... th ei r vicws on co ntro ve rsial iss ues (Gorden et a l .. 1987); th ey are more e ffecti ve in upward influence si tuation s (Infante & Gorden. 1985a. 1985b. 1987): they are judged by th e ir s ubordinate~ as having mo re co n ~trll c li ve persuasion sty les (Gorden. Infant e. & 1110. 1988): and th ey enjoy hi gher credibility perception, (Infante & Gorden. 1987. 1989).
COIlc/usioll This chapter ha ~ sugge:-. t\!d that argument i ... inhc.!rcnt in the process of pcrslILlsion. A':> such. an individual difference perspl.!ctivc regard ing aggre~si\'l.~ co mmuni cat ion beha\ior and pcr:-.uasion is illuminating. Research on argumentat ive ness and verbal aggrcssivcne~:-. has examined how the persuasion process i ~ inO uenccd by th c:-.c trait .... Thi\ re ... earch has stud-
Chapter 7 • Ar~fI",ellt{{tirelle,\'Ij. Verbal AgKres.\'h'(Jl/e, ~.\',
(lnd Persll{Isiotl
127
ied both the source and receiver. As dyadic persuasion is a more common form of persuasion and one we need to understand beller. incorporating •.111 understanding of aggressive communication traits helps us under:-.t3nd the process more fully, The research reviewed in this chapter suggests that argumentativeness and verbal aggressiveness are predispositions thai warrant further research. since it has been demonstrated that these two traits have considerable bearing on interpersonal social influence. Whm should the future of this line of research be? What are some types of research that should be conducted? First. this line of research should be extended frol11the laboratory into the field. That is. while a few studies cited in this review asked people to engage in "interactive" persuasive efforts. the majority did not. Much of the research also involved persuasive and compliance-gaining efforts conducted under traditional "Iaboratory" conditions and in nonspecified or "Mranger only" contexts. This approach i!o. beneficial to understanding how these aggressive communication trailS function in per"uasive contexts and helpful in building theoretical frameworks that attempt to explain the interaction of these constructs. However. the generalizability of these findings is limited. Thus. exploring the effect of argumentativeness and verbal aggressiveness ill field settings and in specified persuasive contexts seems appropriate. For example. Infante et al. (1997) found that high argumentative!> indicated the greatest self-persuasion under conditions involving counterattitudinal advocacy. This willingness to argue with one's own position might be observed in the dating/court~hip contcxt!o. and in the organi/ational context (e.g., in employee interviews and job-related meetings). This speculation should be subject to empirical scrutiny. A better understanding of the low-argumentative individual, especially during different advocacy situations. is also worthy of empirical examination (Infante et al.. 1997). In one study. Infante and hi~ colleagues found that a proattitudinal task was liked equally by slightly. moderately. and highly argumentative individuals. It was speculated that proattitudin:.tl advocacy. especially if not engaged in via a face-lo-face encounter with an adversary. may not be . . eell as dista:-.teful by low-argumentative individuals (p. 85). Thus. alternative channels of cOllll11unication, especially e-mail. might engender more favorable proattitudinal (and po~sibly even counterattitudinal) advocacy feelings for lowargumentative individuals. As sllch, research that varies the channel through which persuasive efforts are attempted (by those who vary in argumentativeness and verbal aggressiveness) i... also worthy of study. Along these lines. Levine and Badger (1993) observed an interesting yet counterintuitive finding: Low argul11cntatives were more resistant to persuasion than highly argumentative individuals. The researchers suggested that argumentativeness leads to less resistance to proatlitudinal mes!o.ages (p. 76) and proposed that the effect or argumentativenes . . . in instilling resistance to persuasion is valid primarily for "positions that individuals would not readily endorse" (p. 76). Again. this speculation awaits empirical scrutiny. Further. the relationship between argumentativeness and dogmatism, as well as the relation . . . hip between the trait and initial attitudes. deserves increased attention. In summary. this corpus of resean.:h underscores the importance of aggressive COI11munication predi . . . positions in understanding the persuasive and compliance-gaining processes. Additional research conducted in varied communication settings and contexts should be helpful in developing guidelines to enhance persua,ive and compliance-gaining outcomes.
128
Pan II • Persuasion Variables: PC'r.~fJecli1'es 0/1 Sources. Receil'ers, Challlle/.\, (/Ild Mes.wge,\
Referellces_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __ Ajzen, L & Fishbein, M. (1980). UlldentmuliliR altitudes alld predictiliR .weial b('//(Ivior. Englewood Cl iffs, NJ: Prenlice-llali. Andersen, P. A. ( 1987). The trait debate: A critica l examin ation of the individual differences paradigm in interpersonal comm uni cation. In 8 . Dervin & M. J . Voigt (Ed:--.), Progres\' ill C(IImlllllliCllli(1II sciellces (vo l. 7, pp. 47-52). Norwood, NJ: Ablcx. Atki n ...on. J. W. (1957). Motivational dctenninant .. of risJ...-I:lI\ing behavior. P.\ydw/ogica/ Rel'iell'. 6.J. 359-372. Avtgis, T. A .. & Rancer. A. S. ( 1997). Argumentativeness and verbal aggressiveness as a function of locus of con trol. Communication Research Report.l. 14. 441-450. Bcauy. M. 1.. & McCroskey. J. C. (1997). It' s in our nature: Verbal aggre ..si\e n cs~ a .. tcmpcramcllIa l expression. Cummul/icatiun Quarterly, 45,446-460. Beatty. M. J .• & McCroskey. J. C ( 1998). Inte q)C r ~ona l communica tion as tcmpcr:lIllcntal ex pression: A communibi ologica l paradigm. In J. C. McCroskey. 1. A. Daly, M. M. Martin. & M. J. Bcatty (Eds.). Communicatiull and penol/alit.\": Trait perspecfil'es (pp. 41-67). Cre ..... kill. NJ: !'!amplOll Press. Boster. F. 1.. & Levi ne. T. ( 1988). Individual differences and i.:Olll pJiance-gai nin g message selecti on: The effects of verbal aggressivcness. a rgumentativeness. dogmati!>lll. and negativi .. m. Communicatio/l Research Reporu", 5. 11 4--119. Boster, F. J.. Lev ine, T .. & Kazoleas. D. (1993). The impact of argumentative ness and verbal aggrcs~ive ness on strategic diversity and persistence in compliance-ga ining hchuv ior. C(III/I1l1l11iclllio/l Quarter/y.4/.405-414. Burgoon, J. K .. & Hale. J. L. (1988). Nonverbal expectancy violation1oo: Model elaboration and application to immediacy behaviors. Communication MO/lograph,\, 55. 58-79. Cac ioppo. J. T .. Petty. R. E .. Kao. F. C, & Rodri gue/. R. (1986). Centrul and peripheral routes to pcr .. uasian: An individual difference perJo.pective. )ollmal of Persollality lIrld Social Psychology. 51. 1032- 1043. Canary. D .. Cunni ngham , E. M .. & Cody, M. J . ( 1988). Goal types. gender. and locus of control In managing interperso nal con ni e!. Communication ReJearc/i. 15.426-446. Cody. M. J.. Canary. D. 1.. & Smith. S. W. (1994). Compliance-gaining goals: An ind ucti\c analysis of actor.. ' goal type,,>, strategic<." and wcces<.,e ... In 1. A. Daly & J . M. Wiemann (Eds.). Sfrate.s.:ic illferpenO/wl commlllli("alioll (pp. :B-90). Hill sdale. NJ: Erlballm. Costa. P. T .. & McCrae. R. R. (1980). Still stabl e after all these years : Pep,unalit y as a key to somc i!>sues in adulthood and old age. In P. B. Baltes & O. G. Brim (Eds.). U(e·.lpall del'l'lopmew (llId bellm'ior (vol. 3, pp. 65-102). New York: Academic Pre~:-.. EpMcin. S. (1979). The stabi lity of behavior I. On predicting most of the people mudl of the time. jml/"1w! of Pel".\()/Illlity alld Social PsydwloRY. 37. 1097- 11 26. Fi<;hbein. M .. & Ajzen. I. ( 1975). Belief, (l(lillld('. ;/IIellf;OI/, tlml hehm'ior: All illlroriucrio/l to tlieor'l" alld research. Readi ng, MA: Addison-Wesley . Gorden. \V.!.. & Infante, D. A. ( 1987). E mpl oyee rights: Context. argumentati\enl!s .... \erbal aggressiveness. and career sati sfuction . In C. A. B. Osigweh (Ed.). COlllllllllliullillg ell/plon'e re.\pOII,\·ihililie.\ (llId rights (pp. 149- 163). We<.,tport, CT: Quorum . Gorden. W . I. . Infante. D. A .. & lLl.o. J. (1988). Variutloll'> in \oicc p~rtainll1g 10 di<; ... atisfac tionl satisfaction with '>ubordinate .... Mal/ agement Communicatioll Qllarterly. 2.6-22. Gorden. W. I.. & Infantc. D. A. ( 1991 ). Test of a commun ication mode l of organitational cO lll mitment. Communicatiol/ Quarterly. 39.144- 155. Hample, 0 .. & Dallinger. J. M. (1985a). Unused complia nce gaining ... trategie .... In J. R. Cox. \1. O. Sitlar.... & G. B. Walker (Eds.). Argument alld sodal praerin': Proceedillgs a/the fOllrfh SC;VAFA ("(mfer· ellee 01/ arglllllel/f(/liol/ (pp. 675- 691 ). Annandale. V A: Speech Com muni cat ion As .. oc iiltion. Hample. D .. & Datlinger. J. M. ( 1985b, November). Cogllitil'(' editing ojargumellt\tnttegies. Pal)Cr presented at the annual meeting of the Speech Communication As~ociation, Denver. CO. Hampl e. D .. & Dallinger, J . M. (1987). Indi vidual differences in cog niti ve ed iting <., Iandards. Hll lllall COII/municlilio/l Research. 14, 12 3-144.
Chapler 7 • Argllmell(a/il'f'lIeSS, Verl/al ARf!,rf'Ssil'elle.H, lind Penuasiotl
129
Ilunier. J. E.. & Bo~tl!r. F. J . ( 1987). A model of compl iance· gaming me .... age selection. Cmmmmiclflioll MOl/og1'llphl. 54.63-84. Ifert. D. E.. & Bearden. L. (1998). The IIll1ucnce of argumcntativene~'i and verbal aggression on respon'-C~ 10 refu ~ed rcquc\l!>. Commllllil'llliOfl Rep0rl\, II. 145- 154 . Infantl!. D. A. ( 1987a). Aggre~sivene~ ... In J. C. McC'ro.. l-.ey & J. A. Dal) (Ed .... ). Persoflalul" lilld mlerperwlllal com11l1Illiwt;(J/l (pp. 157 · 192). Nc",hury ParI-.. CA: Sage. Infante. O. A. (1987b). Enhancing the prediction 01 rc\ponsc to a commun ication situation from COIlUllUnication traib. C011l1l1tlllicutio/l Quarterly, 35. 308-316. Infante. O. A. (1989). Rcspon\c to high argulllenlalive~: Me!o.\age and "cx differcnces. Soutlu'rII Commll"ieario/l joumal. 5oJ. 159-170. Infantc. D. A .. Andcr ...on. C. M .. Manin. M. M ., !lerington. A. D .. & Kim. J. K. (1993). Subordinatc ... · ...ati ... faction and perceptIOn .. of .. upcrior\· compliance-gaining I<:IcIIC\. argumenlatlvenes\. verbal aggre ... :-.ivcne ....... and style. M(II111gl'lIIent C0I1II1IIWil"tlfion QUl/rierly. 6, 307-326. Infantc, D. A .. Chandler Sabourin, T .. Rudd. J . E.. & Shan nun. E. A. (1990). Vcrbal aggrc'i ... iun in violent and nonviolent marital disputes. COlllmll"j('£lfioll Quarter/v. 3M, 361-371. Infante. D. A .. & Gorden. W. I. (1985a). Bcncfil'i vcr ... u'" hia\: An lIlve"ugation of argumcntativcnc<.; .... gender. and organizational cOlllmunication Ouh:nmc .... COl1l11l1l11icllti(m Research He/lOrts. 2. 196201. Infant!!. D. A .. & Gorden. W. I. (1985h). Superior.. ' iirgumcnlalivene ... \ and \crbal aggre'i\ivcne ... \ a:. pn:· diclor~ of ... ubordinates· ...ali ... facllon. HIIIII{III Olllllllllll;('lli/Oll R(!wureh. 12. 117-125. Infante . D. A .. & Gorden. W. I. <191-17). Superior and ... uhordinalc communication profile ... : Implication ... for independenl-mi ndedne<; ... and upward cffccti\"l.:nc\ .... Cel/lral SllIte.\ Speech }otmwl, 38. 73-80. Infante. O. A .. & Gorden. W. I. ( 1989). Ar£umcntau\cnc ..... and aflirnllng communicator ... tyle a~ prediCtor ... of \ati<.;factionldissati<;factlon with ... ubordtn
130
Part II • Per.\/w.I'im/ VarialJ/t>s: Perspecfil'e,\' Oil Sources. Receivers, Chal/I/efs. and Messages Kinney. T. A. (1994). An indUl.:ti\ ely derived typology of \crbal aggression and it~ relationship to distrcs~. HIlII/{/1I Commullication /(e\l!arclr. 21. 183-222. Kinney. T .. & Segrin. C. ( 1998). Cognil ive moderators of ncgativc reaction~ to verbal aggrc~sion. Communication SflIdies. 49. 49Levine. T. R.. & Badger. E. E. ( 1993). Argulllcnt.lti\'ene .. ~ ,Ind resistance 10 persuasion. CommulliCC/tirm Report.\'. 6. 7 1-78. Levine. T. R .. & Bo:o.ler. F. J. (1996). The impact of ~clf and other~' argumentativencss on talk about controvcr.. ial i",sue~. Communication Quarterly. 44. 345-358. Lim. T. S. (1990). Thc innuencc of receiver.. ' rc",i~tance on per~uaders' verbal aggrcssivene .. s. COl1lmllllic(Jtio/l Qllarterly. 38. 170- 188. Magnu~!>on. D .. & Endler. N. S. (1977). Interactional psychology: Presen t .. tatu .. and future pro .. pect~. In D. Magnu ..~o n & N. Endler (Ed~.). Pas(lI/ality at till' l'rmsroad.l: Cllrrl'll1 isslle.1 ill illferactiollal psych()I(I~y (3- 35). Hilhdal c. NJ: Erlbaum. Miller, G. R .. Bo .. ter. F. J., Roloff. M. E.. & Seibold, D. R. (1977). Compliance-gaining message strategies: A typology and sing and attilUde-bchavior rclation,. COllll11l1l1ictlliOl/ Research R('{wrts. 6. 1- 6. Nicotera. A. M., Smilowitz. M .. & Pearson. J. C. (1990). Ambiguity tolerance. conflict management style and argumentativeness ib predictor~ of innovativenes:o.. CommulIication Research Reports, 7. 125131. Norton. R. W. (1978). Foundation of a communicator style con<;lrUCI. HUll/an COl1ll11lllliratirJ/1 Resellrch.
n.
-1,99- 112. Norton. R. \V. ( 1983). Comll/lll1iwlOr style: Theory. applicatioll. and mea.wres. Beverly Hill..;, CA: Sage. Onyekwere. E. 0., Rubin. R. B.. & Infante. D. A. (1991). Intcrper<;onal pen:cption and commun ication sali .. r,u;tion a ... a runction of argumcnt •.Iti\cncss and ego- in VOlvement. Conll1/lmiC(l(;OI/ Qllarterly.
39,35-47. Rancer. A. S .. AVlgis. T. A .. Ko,hcrg. R. l.. & Whitecap. V . G. (2(x)(). A longitudinal assessment of trait argumentativeness and verba l aggre .. sivcncss between ",cventh and eighth grades. Communicatioll Educatioll. -19. 11.+-119. Rancer. A. S .. & Bau~u~. R. A. ( 1987). Discriminating malc .. and female~ on belief struct ures about arguing. in l. B. Nadler. M. K. Nadler, & W. R. Todd-Manc illas (Eds.). Adl'{lI/ces ill gender alld COlli' II11111icatioll reward, (pp. 155- I 73). Lanham. M 0: University Pres .. of America. RanceI'. A. S .. l3aukll~. R. A .. & Infante. D. A. (1985). Relations between argumentativenes", and belief ~tructure .. about arguing. tomml/Ilicarioll Educalioll. 3-1, 37--47. Rancer. A. S .. & Infante. D. A. (1985). Relation .. between motivation to argue and the argulllentativeness of advef\arie,. Commullicarirm Q/lClrtl't1y. JJ. 109- 218. Ranccr. A. S .. Jordan. F. F.. & Infante. D. A. (2000. November). Obsen'f'rS' perceptiolls (~r(lll ill1erperso//(II di,lplIle liS a filllctioll oj lIffi rlllin~ ,I·tyle and mode of pre.\·el/((Ilioll. Paper presel1lcd al the annual meeting of the Natiomd Commu nicatioll Association. Seatlle. WA . Rancer. A. S .. Ko .. berg. R. L.. & Bauku~. R. A. (1992). Belicfs about arguing as predictor~ of trait <.Irgument3ti\ eness; implicat ion s for training in argumcnt and conflict management. COII/III/micatioll Etillcati()//, 4/. 375-3R7. Rancer. A. S .. Kmtx:rg. R. L, & Silvc<;tri. V. N. (1992). The relationship between self-estecm ~lIld aggres:o.ive communication predi .. p()~itions. CO/llmunication Reuarch Report.l. 9. 23-32. Ranccr. A. S .. Whitccap. V. G .. Kosbcrg. R. L. & Avtgis. T. A. ( 1997). Te .. ting the cfficacy o f a communication training program to incrci.l ..c argumentativene~s and argumentative beha\ ior in adolc~ cenls. CmlllllllflicalirJ/l Edt/Clition, 46.273-286. Reardoll. K. K. (1991). Per.II/{/,\·iO/l ill pmctin'. Newbury Park . CA; Sage. Reynolds. R. A. (1987. May). A '"MwI/ellfatil'('IIt'.I·S. /leed for (,o~1/itjoll. alld wisertil'elle.I·S (/.1 predictors of compliance xaillillg 11I{'Hagl' .Hrategy .1·electioll.S. Papcr presented at the annual mecting of the International Communication Association. Montreal. Canadi.l. & McCro,key. L. L. (1989). An invc ~t igalion ofsclf-pcrceived comRichmond. V. P .. McCroskey, J. munication compete nce and personality oricntations. ComlllUllicatioll Re,warch Rep()rt.~, 6. 28-36.
c..
Chapter 7 • Argllmellrllril'Cfless, Verl)(ll Aggressil'clless, and Persllasion
13 ]
Roberto, A, J .. & Finucane. M. (1997). The a<;<;c<;<.;ment of argumentativenes," and verbal aggre<;r;ivcne .....; in adolescent populations. Commullication Quarterly. 45. 21 - 36. Stewart. R. A .. & Roach. K. D. (\993). Argumentativeness. religiour; orientation. and reactions to argument '>ituatiom. involving religious vcr<,u ... nonreligious i~<;uer;. Communication Quarterly. 4/. 26--
39.
Stewart. R. A .. Roach. K. D. (1998). ArgumentativcnesJ.. and the theory of rea~oned action . Commllllicatirm Quarterly. 46.177- 193. Swap. W. C .. & Rubin. J. Z. ( 1983). Mea<;urClllcnt of interpersonal orientation. lOl/mal uJ Penol/alit\' and Soda/ P.\Yl·hology, 44. 208- 219. Valcncic. K, M .. Bcany. M. 1.. Rudd. J. E.. Dobos. J. A .. & Heisel. A. D. (1998). An empiricallCst of a cOllll11l1nibiological model of trait verbal aggrc<;sivenc<;s. Commllllicarioll Qll{lrferly. 46. 327- 341. Waggcnspack. B. M .. & Hensley. W. E. (1989). Perception of the argumentativeness trait in interpcrsonal rciation"hip situations. Social Beh(ll'iol' alld P(!I',\'()1/ality, 17. 11 1-120. Wigley. C. J. (1998). Verbal aggres~ivcncss. In J. C. McCro"kcy. J. A. Daly. M. M. Martin, & M. 1. Beatty (Eds.). COll1l1l11l1icalioll {lml per.wfwlity: Trait penpecti,'es (pp. 191-214). Cres~kilJ. NJ: Ilamrlon Press.
8 Gender Effects on Social Influence Linda L. Carli
The
'lalU!o.
of
wOlllen
hm. improved in recent years. The presence of women \uch
<.1\
National Security Advj,er CondoleC7/i.l Rice. former atlorncy general Janet Reno. and
producer and media mogul Oprah Winfrc) in highl) visible po~ition, of power i\ emblemalit: of thi~ improved statu .... in fact. currently about 47 percenl of workers in the United State ... arc women. up from 40 percent in 1976 (U.S. Bureau of Labor Statistic!-.. 200 1). Where:.!:.. 25 year!o. ugo 25 percclll of managers v..'ere women. now women fill nearly half of all managerial and admini ... trative po..,ition:-. (U.S . Bureau of Labor Slatistic!o.. 2001). The salary differential between men and women ha ... also shrunk. Today. on average. women carn about 74 percent of what men earn. whcrca . . in 1976 they earned only 58 percent of men· . . income (U.S. Bureau of the Cem.u ..... 20(0). evert he less. in ~pite of the advances that women have made and the presence of a ,mall but highly visible number of women in pm.ition . . of authority. wom~n continue to be underreprc . . ented in the upper echelons of power. In Fortune 500 companic ..... fc\\'er than I percent of CEOs are women and women hold only 5 percent of the top executive po,ition, (Catalyst. 2000). In the United States governmcnt. only 13 percent of senator.... 14 percent of congressional repre!o,entative~. and 10 percent of . . tate governors are women (Center for the American Woman and Politic ..... 200 1). '-"'omen are likewi . . e ab . . enl from the highest po"ition~ of power in the legal profession (Rhode. 2001). higher education (Chronicle of Higher Education. 1998). medicine (Reed & Buddcbcrg-Fi . . cher. 200 I). broadcasting and tclccoml1lunication~ (Jamieson & Slass. 2(X) I). and the U.S. military (U.S. Department of Defense. 1998). The persistence of women's exclusion from the 1110 .... t powerful positions underscores the continued re . . istance 10 women· . . illiluence and authority. Indeed. the literature on gender and social influence ha ... typically reported that women are les ... influential than men. This chapter will review research showing gender differences in socia l influence and will argue that these difTerence~ occur a ... the resuit of gender ... tereotypcs. In particu-
133
134
Pan II • Per,ll/usiO/l V(lriahln: Pt'r,lpl'Cfil'(:,.I' OIl SOl/rees. Receil'f(I·.
CI/(IIIIIt'l.I.
(lnd Mess(lgl'.I'
lar. it \\ ill ... Il(l\\ that \~Ol1lcn and ,girb exert le:-.s influence than men and boys. because female ... more than male ... muM c~tablish themselves as competent and likable sources in order to be influential. Likable ... ource ... appeal to their audience because they are similar to thcrn, are physicall) allnlctivc. or possess other socially de:-.irable characteristics. Compelent ..,ource.., appear knowledgeable. intelligent. and articu late. convey ing competence and cxpcrti,e, Influ ence agents who establish themselves as competent (B radl ey. 1980: Dri,'eli. Olmstead. & Sal",. 1993; Erickson. Lind. Johnson. & 0·8arr. 1978; Holtgraves & La,'y. 1999; Son & Schm itt. 1983) and likable (Carli. 1989: C haiken. 1980: Chaiken & Eagly. 1983: Wood & Kallgrcn. 1988) exert greater influence than those who do not. Pcople trust competent. likable influence agents and y ield to their influence. The present analysi ... suggests that mcn excrt greater inOuencc than women because. according to gender ,tereOlype..,. male ... are more competent than female!o.. Moreover. based on s[ereotype~. people expect fell1ale~ to be warmer. nil:er. and more likable than males and consequently arc mOl'\! likcl) to re..,i ... t the intluence of females than [hat of males for not being likabl e enough.
Gender Stereotypes and Socia/lnfluellce The Stereotype of the Competent Male Re~c:lrch examining people· ... gender stereotypes about the Iype~ of traits th aI men and women po~sc~s reveab that men arc con:-.idered to posse~s more agenlic qualities, which n::tlect greater competclll:Y and instrumentality. than women , who in turn are thought to po..,..,cs~ more communal qualitic.., than men. Spccifically. men are cons ide red more IC~ldcrlike. illlcllectuai. analytical, capable of abstract thinking. and able to so lve problems. whereas women are considered kinder. warmer, morc expressive. more s upporti vc, and gentle r (Broverman. Vogel. Broverman, Clarkson. & Rosenkrantz, 1972; Dc1.lux & Kite. 1993; Eagly & Mladinic. 19B9; Fiske & Ru scher. 1993: Ruble. 1983: Williams & Best. 1990). Similar :::.tereotypes have been reportcd in work ~c tting s: execllt ives consider malc managers 10 be morc competent than female manager~ (He ilman. Block. & Martell. 1995). and management ability and compelcncc is considered morc c haracteri sti c of men than of women (Schein. in press). Other research examining cvaluation of men's and women's performances has likewi\e rc\'ealcd lhis same stereotype. For example, a sma ll bias favoring male expertise wa!o. reponed in a mcta-analytic review of studies using the "Goldberg paradigm," in which participant<., eva luated identical behaviors or products but were so metimes told that those behaviors or products were produced by men and some times that they were produced by women (Swim, Borgida. Maruyama, & Myers, 1989). The review revealed that male performance is rated more favorably when the stimulus matcrials are either gender neutral or in a stereotypically ma~culine domain. but that men and women receive equal evaluation!o. when the domain i~ stcreotypically feminine. These effec ts. although small. indicate that men arc pre<;,ull1ed to be more competent than women unless the task favors female experti..,e, but even then women arc cono;;idered equal to. but not more compe tent than men. Studies examining stereotypes usually provide ~ubjecls with lillie information about the target indhiduals whol11 they are evaluating and often simply ask participants to de-
Chapter 8 • Gender Effecis on Socia/ Influence
135
scribe a typi ca l man or a typical woman. Studies employing the Goldberg paradigm generally prese nt brief desc ripti ons of the target, such as a resu me, or something the target has writt en or created before being evaluated. Perhaps parti cipants rely on stereotypes under th ese condi ti ons because th ey have little objective information upon which to base their ev aluation s. If tru e, then the bias in evaluating male and female competence should disappear when participants are exposed to actual behaviors by men and women and can base thei r eva luati ons on th ei r firsthand observations of male and female performances. Unfortunatel y, even in face-to-face interactions with men and women with no objective differenc es in performance. undergraduate subjec ts rated men as having performed more
co mpetent ly than wo men (Ca rli. 1991: Carli, 1997: Wood & Karten, 1986). Indeed, re~earc h
has revealed a double standard in the evaluation of men and women. Women must di spl ay greater ev idence of ski ll than men to be consi dered equally competent (Biernat &
Kobrynowicz, 1997: Foschi. 1996). Moreover, in order for people to perceive a woman as more competent than a man. they must be given ve ry clear and explicit evidence of the woman 's substantial superiority relative to the man 's (S hackelford, Wood , & Worchel,
1996: Wagner, Ford. & Ford. 1986: Wood & Karten, 1986). The double standard for competence has also been revealed in research with children . A study of fourth and fifth graders playing a cooperati ve game revealed that girl s were perceived by other children to be less co mpeten t at it th an boys were, eve n th ough objec ti ve analysis of the children's actual performance did not reveal a gender difference (Lockheed, Harris, & Nemceff, 1983).
Status, Social Roles, and Gender Wh y do people perceive men to be generall y more co mpetent and agenlic? According to
Alice Eagly's ( 1987) Social Role Theory. men and women are distributed differently into soc ial roles. First. based on the traditional division of labor in the family , men have more often had the role of financial provider and women the rol e of homemaker. Second, paid occupa tion s are highl y ge nder scgregated, with men's posi ti ons conferring higher level s of status and power than womcn's. Typically. (he higher-status occupational roles to which men have been as:-.igned require agentic behaviors, such as task co mpetence, leadership, and dominance. Conversely. women's domesti c roles and lower- status occupational roles more often requ ire comm unal behaviors. such as nurturance, kindness, and selflessness.
Eagly argues that people have deduced the ge nder stereotypes through observation of men and women in these highl y segregated roles and as a re..~ult have come to expect men to behave in a more agentic manner than women. The association of men with powerful. high-status roles has resulted in their generall y gaini ng higher status than wo men . According to Expectati on States Theory (Berger,
Fisek. Norma n, & Zelditch. 1977), ge nd er acts as a diffuse status characteristic, a general attribut e that is associated w ith an indi vidual 's relati ve statu s in society. Diffuse status characterbtics incl ude gender, race. degree of physical attractiveness, and education. C haracteristics that are valued or considered desirable, such as being male, white, physica ll y attractive. and well educated, confer high statu s. Research indicates that a high- statu s indi vidual is assumed to be more competent than someone of low status, and as a result, people seck the opi nions of high-statu s people and yield to their influence more than to
people of low status (Berger et aI., 1977). This tendency to encourage high-status people
136
Part II • Persuasion Variables: Perspeclives on Sources. Receil'ers, Channels. and Messages
to contribute their ideas and act as task leaders creates a self-fulfilling prophecy: the more individuals make task contributions, the more they enhance their status, increase their influence, and emerge as leaders (Hawkins, 1995: Ridgeway, 1978: Stein & Heller, 1979: Wood & Karten, 1986). Therefore, high-status individuals are not only expected to exhibit higher levels of competence and performance. but these expectations lead them actually to be more successful in influencing others, In the same way that high-status individuals are given opportunities to exert influence. low-status individuals are denied these opportunities. Individuals' diffuse status affects not only their perceived competence and expectations about their future performance but also their expectations about what constitutes appropriate behavior in the group. People perceive low-status individuals. because of their presumed lower competence. as lacking in legitimacy as authorities; as a result, they are more likely to resist the inOuence of low-status than of high-status individuals (Meeker & Weitzel-O' Neill, 1985: Ridgeway & Berger, 1986). Instead. when low-status individuals behave in a status-asserting man~ ner. oveltly attempting to influence others or taking on leadership roles, they are ignored. penalized, or rejected, which drops their status further (Meeker & Weitzel-O'Neill. 1985). This analysis suggests that because men generally possess higher status than women. more men than women would be given opportunities to make task contributions in groups and would be more likely than women to enhance their status, influence others. and emerge as leaders when doing so. In fact. a recent meta-analytic review of gender differences in group interactions revealed that men do make a higher proportion of task contributions than women do (Carli & Olm-Shipman, 2000). In addition, although men's task contributions in mixed-sex groups predict their ability to influence other group members, women's task contributions are unrelated to influence (Walker, Ilardi. McMahon, & Fennell, 1996). Further. women's task-related behavior is more likely than that of men to evoke negative reactions from others (Butler & Geis, 1990). Further support for this analysis can be found in Eagly and Karau's (1991) meta-analytic review of gender differences in leader emergence. which revealed that in initially leaderless groups, men emerge more often than women as leaders. In summary. the presumption of greater male competence is based on the different distribution of men and women into social roles and the relatively high status of men's roles compared with women's roles. Because competent individuals exert greater influence than less competent individuals, women and girls would be expected to exert less influence than men and boys. The male advantage would be expected to occur except in contexts that favor female expertise and competence, either because a particular female has demonstrated clear superiority over her male counterpart or because the domain of the interaction is stereotypically female, such as child care.
The Stereotype of the Communal Female Just as men are perceived to be more competent than women, women are perceived to be nicer and more communal than men (Broverman et aI., 1972; Deaux & Kite, 1993: Eagly & Mladinic, 1989: Fiske & Ruscher, 1993; Ruble, 1983: Williams & Best, 1990). Research indicates that people value communal traits highly enough so that attitudes toward
Ch~lp{cr 8 • Gender t.1/eflJ
011
Socia/Influence
137
women tend to be morc favorable overall Ihan uttitude . . IOward men, a finding Ihat has been labeled the "women are wonderful" effect (Eagly & Mladinie, 1989, 1994; Eagly, Mladinie, & Ouo, 1991). Yet even though women are held in esteem for po"cssing desirable communal traits. thi~ esteem docs not confer them with increased innuence in task-oriented grouP!\. On the contrary, people view cOl11mul1ul trail!\ as important in stereotypical femininc contexts. so that communal individuals are seen as best suited for domestic roles and traditionally female-dominated profession, (Cejka & Eagly. 1999; Eagly & Steffen. 1984). Consequently. womcn's greater communality ... hould enhance their innuencc in stereotypical female context . . but should provide no particular advantage to them in gender-ncutral or masculine domains. Although the "women are wondcrful" effcct may be viewed as generally beneficial. this Mereotypc is not merely descriptive but also highly prescriptive. Descripfil'e gellder stereotype.\ renect beliefs about the way men and women are perceived to be. whereas prescr;pli\'(' gender steremype.\ delineate how men and women ought to be. that is. the behaviors considered appropriate for each gender. In the case of the stereotype of female communality. people not on ly believe that women are niccr than men. they reqllire womcn to be so (Burge" & Borgida, 1999; Eagly. 1987). This pre,cription demands that women be warm. nurturing. and selfless or be perceived as violating gender-role norms. This finding suggests that observers may dislikc and penalize a woman whom they consider lacking in communality and resiM her innuence as a rcsult. NOlably. the prescription for women includes avoiding behavior that is too Matusassening. thrcatening. or directive. In essence, people do not consider it appropriate for women to seeJ... leadership or ,latus overtly or to attempt to influence others too directly or forcefully (Carli. 1999). This attitude i, based on the lower diffuse status of women relative to men and on the dome . . tic and lower-status occupational roles thar women more often hold. which involve a greater amount of seillessllcss and other-directed ness than men's roles 00. Certainly .... talll' theorists have argued that low-status individuals must ... how warmth and communality more than high-status individuals in order to be innuential. This is because tho~e of low ... tatus lack legitimacy and do nO[ have the right to take charge, direct other~. or act as Icaders. In ... tcad, low-stutm individuals must communicate that they have little desire to take charge or lead others and are motivated merely by a desire 10 help other member; of their group (Meeker & WeitLCI-O'Neill, 1985). As a rc... lIlt. people arc generally morc open to a man's than a woman's influence. regardless of the man's inllucncc style. but would gi\'c greater scrutiny to the style of innuence of a woman and penalize her for beha\ ior that is too status-as\cning or insufficiently communal. Indeed. research with children (Connor. Serbin. & Ender, 1978) and adults (Carli, LaFleur. & Loeber. 1995) has confirmed that people prefer females who are indirect, agreeable. and cOlllmunal to those who are direct. threatening , and status-asserting, whereas they likc males equally well regardless of communality or status assertion. Clearly. in most !\iruations. women's ability to influence others. compared with that of men. would be more dependent on the use of an influence style that corresponds prescriptively to thc stereotypical female role. Displays of warmth and communality should therefore be expected to enhance the innuence of women and girb. wherea~ dominant or a~senivc beh.n;ior ... should be expected to reduce their innuence.
138
Part II • PerJlws;on Var;ables: Perspeclh'eJ
011
Source,\. /(ecein'n. ClIlIIIIICIs. lIlId Me,\ \{/Kl',\
One of the unfortunate effect:;, of gender stereotypes is that highly competent bdli.1\ior in women may be viewed as too status-asserting and incompatible \.\ ith the traditional female gender role. Therefore. unlike men, women experience a double bind. On the one hand their competence is more likely to be questioned than the competence of men. and on the other hand, behavior that clearly conveys competence may be considered inappropriate in women. Competent women are often not liked as much a~ competent 1l1t.!11 or less competent women (Carli. 1991: Falbo. Hazen. Linimon. 1982). Status theorish have argued that the problem of low-statu:;. individuals appearing too status-asserting i~ 1110st pronounced in interactions with high-!ky. 1992~ Schein. in press), and men are more likely than womcn to endorse tradition~1 gender roles (Twenge. 1997). In summary. the presumption of greater female communality i. . based on the different distribution of men and women into social roles, with women more often in Jome~tic roles and lower-status occupational roles. In addition. women. like othc" of low ,tatu,. have less legitimacy as leaders and consequcntly are pcnJli/cd for status-asserting behavior. Because female communalit y has become prescriptive. women who do not behave in a warm communal manner are likewise penalized. As a re!o.ult. women's inlllll:ncc j:-. perceived to be more conditional than men·s. with women excrting greater influcnce while displaying communal behavior and les~ influence in re!o.pon ...c to !->tatu ... -a"'!o.erting behavior. Funhermore. it is men more than women who should re~i ... t the influence of compctent females.
Gellder Differellces ill Illfluellce As one would expect, given the greater perceived competence and legitimac) of' male influence agents. research confirms that men exert greater influence than women do. Lockheed (1985) conducted a meta-analytic review of 29 'tudie, examining gender differences in task-oriented mixed-gender groups. She rep0l1cd thai men exert greater inlluence and exhibit more leadership behaviors than do women. Moreover. re . . earch reveal ... that the gender difference in social inlluence is not merely due to higher-quality performances by men. For example. Propp ( 1995) reported that in group interactions member. . were more likely to attend to ideas contributed by men and to u\e them in ...olving group problems than to the identical ideas contributed by women. Similarl). rc ...carch has ... howl1 that men remain more influential than women, even when the pcp.,un . . ive messages of the male and female agents are manipulated to be identical (Altemeyer & Jones. 1974: DiBerardinis, Ramage, & Levitt, 1984) or when the performances the male and femalc ugellb are manipulated to be equally good (Schneider & Cook. 1995: Wagner. Ford. & Ford. 1986). The same pattern of results has been reported in research on children. In interactions with peers, boys are more influential than girls (Jacklin & Maccoby. 1978).
or
Chapter 8 • Gl'II
139
Competellce, Gellder, alld Sociallllfluellce Further research ~pecifically links gender differencc~ in perceived competence to gender differences in social influence. In one ">lUdy. women and men attempted 10 influence others by speaking either in a competent manner (by supporting their argumenh with evidence) or in u le% competent manner (with no ~uppofling argumenl"» (Bradley. 1981). Consistent with the double ~Iandard for competence. re~ult.'. rl!vealed that men were perceived to be equally competent and were equally influential regardless of their communication sty le. whereas women were perceived to be more competent anu exerted greater influcnce when using a competent ~tylc than when u~ing the less competent style. Other researc h revea ls that women exert less influence than men in ~tcreotypically masculinc and gender-neutral domains. in which men are expected to show higher competence than women. but are more influcntial than men in ~tereotypically feminine domains and contexts in which women are expected to be more competent. For example. men cxert greater influence over the opinions of others for masculinc topics sm:h as sports, gun control, and military affairs. whereas women exert greater influence for feminine topics 'illch a~ women's fear of crime and child care (Falbo. Hazen, & Linimon, 1982: FeldmunSummers, Montano. Kasprlyk. & Wagner. 1980; Javornisky. 1979: Gerrard. Breda, & Gibbons. 1990). Moreover. evidence of clear femalc superiority at a task increa'ies women's influencc and decreases men'!>. (Pugh & Wahrman. 1983: Shackelford. Wood. & Worche l, 1996). As already noted, although competence generally facilitates influence. thi~ is not always the case for women. whose competent behavior may sometimes be perceived as too statu!>.-asserting, A number of ~tudies have revealed that ",omen can be di~advantaged by competence displays. For example, in one study, male and fcmale influence agent'i attempted to persuade others u~ing either a direct. competent ... Iyle of communication or a more indirect style. Results showed that men were equally per~lIa~ive regardle~s of their communication style. wherea~ women exerted greater influence when communicating in a more indirect manner (Burgoon. Jonc~. & Stewart, 1975). In another stlldy, corporarc exeCLItives were asked to evaluate the competence of male and female job applicants and indicate whether they would hire a candidate after reading the applicant's resume and a transcripl of Ihe job inlerview (Bullner & MeEnally. 1996). Results revealed Ihallhe executives were J110~t persuaded by and preferred to hire men who comll1unicated in a highly competent manner, showing directness and initiative. rather than men lIsing a less competen t style. The reverse was found for women applicants: the executi\cs reported bcing leaM persuaded by and lea')t likely to hirc a woman using the highly competent Mylc compared with women using other, less competent styles. Research shows that men are much more inclincd to resi~t women's inlluencc than men's (Ridgeway, 1981). Moreover. men also particularly resiq the influence of competent women. A Mudy examining the efTeclivene"ts of asserti\'c ver~lI"t tentative speech revealed that women who used tentative ~pecch. which involved verbal qualifiers such as disclaimers (e.g., " 1 may be wrong" or 'Till no expert") and hedges (e.g .• ''sort of," "kind of"). were perceived to be less competent than those using a~~ertive ~pcech that did 110t contain such verbal qua lifiers (Carli. 1991). In that study, males were perceived to be equally competent regardless of their speech style and, not surprisingly. equally influential
140
Part 11 • Penua ,'1ioll Variable.\': Pt!r\fJc{'/i\ 'cs Oil SOllrce~. Rl'ceiL 'en. Cltalllleh. ami Ml'.Hagl'.\
using either type of speech. Of particular interest was the effect of perceived competence on women's influence. When speaking competently rather than tentatively. women exerted greater influence over 11 female audience but less influence over a male audience. In essence, men were more influenced by a woman they perceived to be lacking in competence than one who appeared highly competent. rating the competent woman as le!o.s trustworthy and less likable than her less competent counterparts. Other research confirm s the finding that men resist the intluence of competent women. In one study, women were equally influenced by competent male and female intluence agents and liked them equally well, whereas men were Illore influenced by a competent man than by a competent woman (Carli. LaFleur. & Loeber. 1995). In that study. Illen reported that they felt more threatened by a competent woman and liked her less than they liked a competent mall. Similarly. a recent study revealed that a woman who presented herself as a feminine woman who preferred traditional gender roles wa\ perceived to be less competent than a woman who presented herself as less traditionally feminine (Matschiner & Murnen. 1999). As expected, the traditional wOlllan exerted more intlucnce over men and less influence over women than the less traditional woman did. Again. men, but not women. judged the more competent. nontraditional woman to be less likable and were more resistant to her influence than a woman with lesser competence. In a similar sllldy. participants listened to an audiotape of a male or female expert who presented a speech advocating nontraditional gender roles; results revealed that womcn were equally persuaded by male and female speakers. but men were less persuaded by women than by men (Rhoades. 1979). Male resistance to female competence has also been found crossculturally. In this study. the re\earcher examined the responses of male and female officials working in Israeli bureaucrmic organizations to the requests of male and female confcderates (Weimann. 1985). In general. confederates were not particularly effective when their requests conveyed helplessness and dependence on the official, with one exception. Female confederates exerted greater influence over male officials when using this relatively incompetent \lyle of cOlllmunication than when using other, less helples ... and more competent appeals. The tendency of males to resist female inlluence has been found not only in re\carch on adults but also in research 011 children. including toddlers and preschoolers. Jacklin and Maccoby (1978) examined the intluence patterns among mixed- and same-sex pairs of 33month-old toddlers. They found that boys and girls were equally likely to issue verbal prohibitions (e.g .. "no" or "don't") when another child attempted to take their toys. but girls issuing prohibitions exerted less influence over their mule playmates than over female playmates and less influence than boys exerted over either mules or female~. Indeed. the boys' behavior was completely unaffected by girl s' prohibitions. which the boys ... imply ignored. Similar findings have been reported with a slightly older sarllple of childrcn. In that study. researchers studying the innuence of preschoolers when issuing direct requests reported that girls exerted les~ influence over boys than over girls. but that boys were equally influential with bOlh male and female classmates (Serbin, Sprafkin. Elman. & Doyle. 1982). In a study of middle school children. boys and girls attempted to persuade their peers to eat bitter-tasting crackers (Dian & Stein. 1978). Although the authors reported that atlractive children were generally more influential with the opposite sex than unattractive children. overall. boys were more inclined to eat the cracker\ after being per-
Chapler 8 • Gender E!fect.\ on Socia/lnflllellce
141
~uaded
by a male than female peer. wherea ... girl~ wert! equally influl.!nced by bmh gender.... Finally, research reveals thal boys resiM the influence of adult females, as wcll. A study assessing the effectiveness of parent~' imperatives and request~ 10 their two- to sixyear-o ld children revealed that girls were equally likely to comply with their J11other~ and father~. but boys complied le ... s with their mothers' influence attempt ... than with tho ... e of their father (Power. McGrath. Hughes. & Maniro, 1994).
COlIIlIIl/nality, Gender, and Social Influence Women's influence depends not only on their apparent competence but al\o on thl! extent to which they display comlllunal behavior. conveying a concern for others and a lack of interest in asserting their status. Men's influence doc~ not. Instead, re ... earch indic:.HC~ that men are often cqually influcntial, regardless of how communally they behave. In one study, male and female confederates communicated either in a communal style. by agreeing with others, or in a dOlninant. status-as ...erting ... tyle. by overtly di~agrecing "ith others. Results revealed that women exerted greater influence when communal than when dominant. wh ile mcn were equally influential in both cases (Carli, 1998). Moreover. in this study people dis liked the dominant woman and responded to her dominance with anger. irritation, and hostility, whereas they did not express ho~tility toward men who were equally dominant. Other re~earch confirml., that women ul.,ing a self-a" ... crting. dominanl. or threatening ~tyle exert les~ influence than men using the same style (Burgoon. Dillard, & Doran. 1983; Perse. Nathanson. & McLeod. 1996) and less innuencc than women using a group-oriented. communal style (Burgoon. Birk. & Hall. 1991: Shackelford. Wood. & Worchel. 1996). Likewise. research reveab that a ... serting one's status through ... elfpromotion is perceived more favorably in men than in women. For example, women who describe thcir achicvcmcnts in a self-promoting manner arc perceived a . . less descrving of recognition or ,upport than less self-promoting women. whereas men are not penali/ed for self-promotion (Giacalone & Riordan. 1990; Wosins.a. Dabu!. Whetstone-Dion. & Cialdini. 1996). Research abo reveals that women who self-promote generally exen less inlluence than more modest women and are less wcll liked. even though self-promoting wOlllen are considered more competent than their marc modest counterparts (Rudman. 1998). In effect. women who appear to be too ... liuus-asserting. directive. or aggres ... ive in their communications are penalized for their gender-role violations. People di ... li~1.! 'llch women and resist their influence. Even nonverbal self-assertion has costs for women. For example. visual dominance. which involves maintaining a relatively higher level of eye gale while speaking than while listening and which is associated with possessing status and authority, i" more acceptable in men than in women. Women who show high amounts of vi'.;ual dominance are le ... s well liked and less inlluentialthan less vi~lIally dominant women (Copeland. Dri ... kell. & Salas. 1995; Mehta. Dovidio. Gibbs. Miller, Huray. Ellyson. & Brown. 1989. cited in Elly~on. Dovidio. & Brown. 1992). although high amounts of visual dominance are accepwble in men and do not reduce men·s influence (Mehta et al.. 1989, cited in Ellyson. Dovidio. & Brown. 1992). Similar findings have been reported in research on children. Killen and Naigles (1995) examined the effectiveness of dominant and comlllunal influence attempt, by boys
142
ParI II • Persuasion Variables: Perspel'lil'eJ ol/Sources, Receil'eI"J, CI/(ml/l'I.~, and M{''\.'WEW~
and girls who were interacting with peers, They found that girls exerted greater influence when using communal behaviors~agreeing. collaborating. and compromi~ing-than when using dominant behaviors---commanding others, issuing orders. or disagreeing, Two very recent studies examined preschoolers' reactions to female and male puppets exhibiting competent and communal behaviors (Carli. Olm-Shipman. & Kishore. 200 I). The first study revealed that boys disliked g irl puppets that displayed leadcrlike and competent behavior more than boy puppets displaying the same behaviors. but girls liked competent boy and girl puppets equally; both boys and g irl s had eq uall y favorable reactions to communal boy and girl puppets. The second study revealed that boys, but not girls. considered direct influence anempts by girl puppets to be less influential th an indirect attempts when the girl puppet was attempting to influence a male puppet. whereas both boys and girls considered boy puppets to be equall y influential in either direct or indirect mode. regardless of whom the puppet was influencing. In general. then. the research on child ren reveals that. just as with aduhs. males' influence is unaffected by whe th er they use communal or dominant behaviors. Moreover. these findings, along with those discussed earlier comparing the effectiveness of communal versus dominant commun ications by females. suggest that boys in particular re~ist the influence of dominant or competent females. Finally. a study examin ing adult reactions to the commu ni cations of one-year-old infants revealed that adulls were three to four times more likely to respond to girls who talked. babbled. or gestured than to girls who demanded attention. cried, or screamed (Fagol. Hagan, Leinbach. & Kronsberg. 1985 ). This same study revealed that adulls responded to boys about the same amount. regardless of their behavior. Clearly. even in childhood. girls' abi lit y 10 influence depends on their use of a communal style of interaction and avoidance of a dominant or self-asserting SlY Ie. whereas boys' ability to influence i~ relatively unaffected by their style of communication. The research reviewed so far indicates that the prescriptive gender stereotype requiring communal behavior in women and girls is endorsed by both males and females. Because being warm and likable is prescript ive for women but not for men. likability is associated wi th social influence for women more than it is for men (Carli. 1989). That is. people are more influential when they are likable. but the link between being likable and influence is stro nger for women than for men. However. there is evidence that men. in particular. prescribe communality for women. Men respond unfavorably to women who communicate self-in terest rather than friendliness. wa rmth. and other communal characteristics (Ridgeway, 1982) and like communal women more than women who are not communal (Carl i. LaFleur. & Loeber. 1995). Similarly. a meta-analysis of research on evaluation of leaders indicates that womcn leaders are denigrated for using an autocratic rather than democratic leadership styl e, especially by men. whereas male leaders are perceived to be eq ua ll y effective regardless of how they lead (Eag ly. Makhijani. & Klonsky, 1992). As this research on women leaders suggests, men's resista nce (0 the influence of competent women can be overcome when the women display communality as well as competence. One study spec ifically (esting this revealed (hal men were less influenced by women who spoke in a highly competent manner. using rap id clear speech. (han by men who spoke in the same manner (Carli. LaFleur, & Loeber. 1995 ). In this same study. with a male audience. women exerted as much influence as men when they combined competent speech with warmth. by smi ling and nodding. and more influence than women who
Chapter 8 • Gentler E{ft.'rf\
011
Sm:ill/lnfluence
143
were merely competent. Warm and competent women wen! perceived as more likable and less threatening to men than women who were competent but not warm. These results clearly demol1\trate that women who adhere to the prescription for female communality and combine competence with warmth reduce male rl!~i ... tance to their innuence. Essentially. communal behavior reduce\ the threat of female competence.
COllclusioll The different distribution of men and women into social role~. according to which women arc more oftcn found in dome~tic and lower-\tatus occupational roles and generally have lower overall statuo;; than men. ha\ rc\uJted in descriptive gendcr stereotype~ that women are less competent and le\\ legitimate than men as authorities and leader.... In addition, prescriptive ~tcreotypes require females to exhibit greater communality than males. These descriptive and prescriptive stcreotypes create an unfortunate double bind for women, who must both demonstrate exceptional competence to be seen as equal in ability to mcn and simultaneous ly avoid threatening others with their competent behavior. As this review has shown. although people who are perceived as competent und likable excrt greater innuence than tho~e who are not. achieving this balance of competence and likability is more or a challenge for women. Behavior that increa\es a man's perceived competence may enhance. or at least not reduce. his likability. competence being con!)istent with stereotypes about men. In contr;l\1. competent behavior can enhance a woman'\ innuence by increasing her perceived compctence. which Illay be in doubt as the re~ult of gender stereot ypes. while at the ~amc time reducing her influence by lowering her likability. This twin phenomenon occurs becau~e behavior that appear .. competent often also appears status-asserting and lacking in the coml11unal qualitie\ prescribed by stereotypes about women. Certainly. women \\ho appear to be direct. competent. and as~ertive may also be penalized for being ,een Z\\ illegitimately ... eeking \tatus. Icader,hip. or innuence. As a result . in order to exert innuence. women Il1U~t ~omehow combine competcnce with behavior that conveys a lack of desire for self-gain. Communal behavior serves this purpose. Women who combine competence with communa lit y can overcome resistance to their innuence while still adhering to traditional gender-role expectations. For women. innuence depends more on being likable than it does for men (Carli, LaFleur, & Loeber. 1995). The greater importance of li~ability for women', than men's innuence i~ under...cored by research on gender differencc!o. in powcr. Compared with men, women typically possess lower levels of expert power. which is ba~ed all perceived competence. and legitin1ate power. which i~ ba\ed on status and legitimate authority: women do, however. possc!o.s relatively high levels of referent power, which is based on women's perceived warmth and communality (Carli. 1999). Clearly. women have relatively less access to sourcc!o. of power that arc more available to men. A ... a result. women l11u~1 rely on their referent power. or likability. more than men do in order to be innuential. Indeed, thi s may account in part for the greater coml11unal behavior shown by women than men. This behavior include, higher levels of po,iti,e social behavior (Carli & Olm-Shipman. 2(00), nonverbal warmth. (Hall. 1984) and democratic leader\hip (Eagly & Johnson, 1990).
144
Part II • Persuasion Variable!)': PenpecliI'eJ Oil SourceJ, Receh·ers. C"a"nel.~, and Mes,mgl'\
Being likable is especially imponam whe n women interact with men. Resistance to fe male influence is panicularly pronounced in men and boys. who are more like ly to dislike and negatively sanction females who are see n as too competent and direct. It is primaril y in interaction s with men that women lack authority and legitimacy. and it is therefore not su rpri sing that men respond less favorably than women do to starus-asse rrin g be havior in women. Given the resistance to worne n's influe nce. particularly by men. how should women behave in order to be influen ti a l? First. women can increase their influe nce by co mmunicatin g in a warm and 01 her-directed manner and avoiding di splays of highly dominant or self-asserting behavior. In addition, in order to overcome the double standard in evaluation . women can enhance th e ir inllucncc by combining a warm communication styl e with outstanding le ve ls of competence. Clearly. the need to exhibit competence combin ed with warmth places an additional burden on women that is not shared by men. In co ntrast to women, the manner in which men and boys communicate has little apparent effect on their likabilit y or influence. Studies show that men are given the be nefit of the doubt and are presumed to be co mpete nt, eve n when their behavior might be see n as incompetent if ex hibited by women. Sim ilarl y. men who lack communality. self-pro mote. or behave in an overtl y directive or dominant manner are perceived more favorably tha n their female counterparts. Indeed. much of the research in thi s revi ew reveals that male influence is re lati ve ly unaffected by how much com petence or communality they display. As influence agents. m ale~ seem to have greare r behavioral latitude than fe ma les. Becau!'lc stereotypes dictate that female~ lack com petence and should be warm and comm unal , th e behavior of fem ale innuence age nts receives greate r scrutiny than that of mal es, and their influence depends much more on their di!'lplaying a careful balance of competence and warmth . Unfortunately, the path to influence is less easily navigated for women than it is for men, as a result.
Rejerellces,____________________________ Ahemeyer. R. A .. & Jone .... K. ( 1974). Sexual identity. ph), ... ic:.!1 attractiveness and ~e:.! t ing p
Broverman.1. K.. Vogel. S. R .• Broverman, D. M .. C larkson, F. E.. & Rmenkrantz. P. S. ( 1972). Sex role stereotypes: A current appraisal. journal oj Sodal Issues, 28(2).59-78. Burgess, D .. & Borgida, E. ( 1999). Who women are, who women should be: Descriptive
Chapter 8 • Gender Effects
(HI
Socia/Influence
145
Burgoon. M .. Dillard. J. P.. Doran. N. E. (1983). Friendly or unfriendly persuasion: The effect!. of violation ... b)' m'lle ... and female .... Humllll Commllllicmioll Reulm-h. 10. 283-294 Burgoon. M .. Jone ... S. B.. & Stew .. n. D. ( 1975). Toward a message-centered theory of perma.,ion: Three I!mpiric... 1 invc.,tigmions of language intcn\lty. Hllman Communication Research. 1.240-256. Butler. D.. & Gei .... F. L. (1990). Nonverbal affect responses to male and female IC<.lders: Implications for Icadcr... hip evaluation~. Journal of PerSOlwlitv alld Social Psye/wlog)·. 58.48-59. Bunner. E. H.. & McEnally, M. (1996). The interactive effect of inlluence taclic. applicant gender. and type of job on hiring recommendations. Se.\ Role.f. 34. 581-591. Carli. L. L. ( 1989). Gender differences in interaction style and inlluence. JOllmal of Perso/laiity ali(I Sodell Psycholog\'. 56, 565-576. Carli. L. L. ( 1991). Gender. ~tatus. and innucnce. In E. J. Lawler. B. Markovsky. C. Ridgeway. & H. A. Walker (Ed .... ). A{h'allCt'J ;1/ ~I"OIIP prOCeSH!s: rheory and resellrcll (\"01. 8. pp. 89-113). Greenwich. Cf: JAI Press. Carli. L. L. (1997. October). t.Yfeu of gender compo,l·irion all selj·I.'I'(lllIarioll. Paper presented at the meeting of the New England Social Psychological A~sociation. Williams College. William'itown. MA. C ... rli. L. L. (1998. June). Gender efferrs ill social illflllem·e. Paper pre ...emed at meeting o r the Society for the I>"'yc hological Stud) of Socialls~ues. Ann Arbor, MI Carli. L. L. (1999). Gender. interpcr-.onal po'Wcr. and I;ocial inllucncc. Joumal of Social Issues. 55. 81-99. Carli. L. L.. LaFleur. S. J., & Loeber. C. C. (1995). Nonverbal behavior. gender. and influcnce. Jourt/(/I of PerHJIIllli(,' (lnd Social P.n'd/OI()~y. 68. 1030-1()41. Carli. L. L.. & Dim-ShIpman. C (2(XX»). Gender differences in ta ~ k and .,ocial behavior: A meta-analytic rC\'lcw. Manu ...cript in preparation. WelJe'iley College. WelJc,ley. MA . Carll. L. L.. Olm-Slupman. C .. & KI ... horc. S. c:~OO I). Gender. interaction. and innuence among preschool chi ldrcn . Manu ...cript In pn:pnratlon. Wellesley College. Wclle ... ley. MA. Catalyst. (2(}{)(»). ("('//.1/1.1 afwomell mrporare officer.1 alld top eaml."·s. New York: Catalyst. Ccmer for the Amcric:m Woman anll Politics. (2001). FCict sheet. New Brun'iwid... NJ: Eagleton Inl;titute of Politic .... Rutger ... Umvcr-.it). Available online: hllp :llw'Ww.rct.rUlger~.edu/-eawp/pdf/e lecti\e . pdf Cej~a. M. A .. & Eagl)'. A. H. (1999). Gender-.. tereotypic image!' of occupation .. correspond to the sex ~egregation oJ employment. IJer.\OlIlllit)' llml Social PSJchol()~y Bllfleri". 25. 4 13-423. Chaiken. S. (1980). Ilcuristic vcr ... u.. sy ... tcmatic information proccssing and the use or source ver.. us Illes.,age cue ... III pcr'iua<;ion. Jmlrlwl vf PerlOIl(lhtl' lllUl Social PJ.\'{·JlOlogy. 39. 752-766. ChalJ..en. S .. & Eagly. A. H. ( 1983). Communication modality as a detemlinant of pcr\ua ... ion: The role of communicator "'allenee. irmma/ oj Personalit\' alld Social Pwchalogy, 45. 241-256. Chronicle of Higher Education. ( 1998). Almal/lic (vol. 45. no. I). Wa~hington. DC: Chronicle or Higher Educ:uion. C()nnor. J. M .. SerbIn. L. A.. & Ender. R. A. ( 1978). Responses orboy'i and girl ... to aggre~~ivc. ,,,"'cnive. and pas\l\e nchavior... of male and female Ch.lfilclers. JOIln/al of Gelleti!' Pn-clwlog\'. 133.59-69. Copeland. C. L.. DmJ..dl. J. E.. & S;lIa .... E. ( 1995). Gender and reat:tiom. to dominance. JOllnwl of Social Bt·h(/\'Ior allli P('rmllalitr. /0, 53-68. Deaux. K.. & Kite. M (1993). Gender ... tereotype'i. In F. L. DenmarJ.. & M. A. Paludi (Ed.,.), Psychology of womell: A lumdlwok ol;nlli'\, allll Theories (pp. 107- 139). We.,tport. Cf: Greenwood Pre~s . D1Bcrardini\.1.. Ramage. K.. & L.C\ltt. S. (1984). Risky shift and gendcr of the advocate: Information theory \CN1 ... normative theury Croup & OrXllI11:cu;Oll StudU'I, 9. 189-200. Dion & K. K.. & Stelll. S. (1978). Physical allractivenC\1; and inlerpcr...onal innuencc. Joumal oJlixperifI/('IIwl Social P,n·c!wlogr. N.97-108. Dri<;kell. J. E.. Ohmtead. B.. & Sal:I.,. E. (1993). Ta\k cues. dominance cues. and innuence in ta~k groups. Journal olApl,lied P.wc/wlogr. 78.51--60. Eagly. A. II . (19R7). Se.t dWen'IIC'(,f ifl .wcial "ellm'ior: A social-role ;ruerprewt;orl. Hill side. NJ: Erlbaum. Eagly. A. II.. & John",ol1. B. T (1990). Gender and leadership style: A meta-analy .. i:-.. Psychological 811/let;fI, lOS. 233 ,256. Eagl). A II .. & Kar:tu. S. J. (1991). Gender and the emergence or leaders: A mew-analy .. i, JOIln/al of PUWmali,\· (Ifill Social P,H'IIolog\,. 60.685- 710.
146
Part II • Per.l'IlaSiml Variables: Pnspectil'es all Sources. ReCl'iI'ers. Challlleis. Clnd Messages
Eagly. A. H .. Makhijani. M. G .. & Klon ... k)'. B. G. (1992). Gcnder and the evaluat ion of leaders: A me taanal)'~i~. P.\·ychological Bll l/elll/. II J. 3-22. Eagly. A. H .. & Mladinic, A. (1989). Gender ~tereo t ypc!o. and allitudc~ toward women nnd men . PersollaliTY and So(:jal Psychology 811l1etill. 15.543-558. Eagly. A. H .. Mladini<.:. A .. & Otto. S. ( 1991). Are women evaluated more favorably than men? Psychology of Women Quarter/v. 15. 203-216. EagJy. A. H .. & Steffen. V. J. (1984). Gender qereolYpc!o. ... tcm from the di ... tribution of women and men into social roles. Journal of PerSOI/(l!il\' lind Social PJycholagy. 46, 735-754. Elly~on. S. L.. Dovidio. J. F .. & Bro\\n. C. E. (1992). The look of power: Gcndcr differenccs in visual dominance bch~tvior. In C. L. Ridgeway (Ed.). Gel/fit'r. iflleracTion. and ineqJ/ality (pp. 50-80). New York: Springer~Verlag. Erickson. B.. Lind. E. A .. Johnson. B. C .. & O·Barr. W . M. (11)7H). Speech style and impression formation in a court ...clting: The effects of "powerful" and "powerIes!'>" speech. Journal oj Experimelltal So· cial Pjyc!wlo8Y. 14.266-279. Fagot, B. I.. Hagan. R .. Leinbac h. M. I) .. & Kron~berg. S. (1985). Differe ntial reaction~ to assenive and communicative acts of toddler boy .. and girls. Child Dr?I·(l/opmellt. 56. 1499- 1505. Falbo. T .. Halen. M. D .. & Linimon. D. (1982). The co<; t ~ o f 1'>cJecting power ba ...e\ o r messages associated with thc oppo ... ite sex. Sex Roln N. 147-157. Fcldman~Summer~. S .. Montano. D. E .. Kasprzyk. 0 .. & Wagner. B. (1980). Intluence attempts wher competing views are gender~rclatcd: Sex as credihility. Psychology oj Womell Quarterly. 5, 3 11 -
320. Fi~ke.
S. T .. & Ru <"cher. J. B. (1993). Negat ive interdepende nce and prejudice: Whence the affect? In D M. Mackie & D. L Hamilton (Ed .... ). Affect. co!?,nitirm, lIIul sTereoryping: /1f1t'rl1ctivc processes ii, grollp perceplioll (pp. 239-268). New York: Academic Press. Foschi. M. (1996). Double standard:;. in the evaluation of men and womcn. Social P.fyc!wlogy Quarterly,
59.237- 254. Gerrard. M .. Brccht. C. & Gibbons. F. X. (1990). Gendcr crfech in couples' deci:;.ion making and contra· ceptive u"'c. Journal oj Applied So('ial PsydroloK\', 20. "'49--464. Giacalone. R. A .. & Riordan. C. A. (191)0). Effect of ... elf~ prc ..enl:.Jtion on perceptions and recognition ir an organiz.uion. Journal oj P.wclw/og\·, 12-1, 25-38. Hall. J. A. ( 1984). N01/l'abal .~ex tiijJ('I't'lI('es: COIlIl1IlIlIi('llrion accuracy and e.\lu·('.\·j; ,'e sryle. Balti more MD: Johns Hopkin, Univcrsity. Hawkin s. K. W. ( 1995). Effects of gender and communication co ntent on leadc r ... hip t:mergencc in slllal task -orient ed groups . Small Croup Research. -16. 234--249. Heilman. M. E .. Block. C 1.. & Martell. R. (1995). Sex ... tcrcoty pes: Do they influence perceptions 01 managers: In N.J . Struthen, (Ed.). Gender ill II 'O/'~phl('e (special i\s.uc). Journal oJSocial Be· JrQl'ior wul Pen(mality. 10(6). 137-252. I-Ioltgraves. T .. & La,ky. B. (1999). Lingui ... tic po\\er and pcr.ua!'>ion. Journal oj LlI/lguage alld Socia, Psyc!Jnl0K,\', JN. 196-205. Jacklin. C. N .. & Maccoby. E. E. (1978). Social behavior at 33 months in same·!'>cx and mixed ~sex dyads
'"l'
Child Del'l!/opmellt. 49. 557- 569. Jamie son. K. 1-1 .• & Sla ... !'>. L. (2001). Pmgre.H or I/O mom at lire lOp? Thl' role OJII'O/II£'II ill telecommllfli· caliOTlS. medill. lI11d e~compallit'.\. Philadelphia: Anllcnberg Publi c Policy Cl!ntcr of the Universit) of Penn"'y lvania. Javorni sky. G. (1979). Task contclH and ... ex difference,," In COnfOnllity. Jourl/aI11PlydlOlogy, 108.213-
220. Killen. M .. & Naigle .... L. R. (1995) Pre ...c hool children pay attention to their addressee~: Effects of gendel composition on peer disputes. Di.H·OIll'w' Pl'{lI'e.\se.\·. 19.329- 346. Lockheed, M. E. (1985). Sex and 1'>ocial inlluence: A meta~analy!'>is guided by theofY. In J. Berger & M Zelditch. Jr. (Ed s.). Statu!), rell'(jrd.~, Will illjluellce: Noll' expectalio/l\' orxani:.e behllvior (pp. 406429). San Francis.eo: Jossey·Ba ... ,. Lockheed. M. E .. l-l arri .... A.M .. & Nemccfr. W. P. (1983). Scx and socia l influcnce: Does sex function a,r statu<" c haracteristic in mixcd~~cx grou ps of c hildren? JOllrtllll oj EdllClIIiofllll Psychology, 75. 877888.
Chapter to! • Gnuler Lf(l'ct' OIl Socillllfl{1l1etlC('
Mal,chincr.
~1..
147
& Mumen. S. K (1999)_ lIypcrfeminlllilY ami influcm:e. P.fvcholog\ of Woml'fI Qllor-
lerk 23, 6:\ I (H2. ~1cc"l!r.
B. F., & \Veil/d-O'Neil. P. A (198)). Sex role ... and intr.:rpcr'onal beha\-ior in I:I'''-oricnlcd group'. In J_ Br.:rger & \ I Zcldlh:h (Ed,_). Stallll. l"('lwrl/\. will il/fluence (pp. 379--40), Wa ... hlngtun. DC: Jo, . .e)-Ba<;s. Pcr"'L'. E. ~1 .. r\ath;m,on. A. I.. & "kLc\ld. D. ~1 (1996). Ellech 01' "'po"e"'pcr.on <;ex. public announce!ment apretll. and involvement on C\aIUallon ... 01 ,afe- ...c\ PSA,. Hi'alll! Commlmicallofl, 8, 171 189.
Power. T Go, ~kGralb, M . P.. Hughe,. S, 0" & ~ 1 ;iI1ire. S II. (1l)9~). Compliance and .,df-a ....';enion: YOUllg children', re . . ponl;\!s w mothcr, vcr ... u, father ..... /)('I '{'I0l'mellwl P_~ycl/OloXY. 6.980---.989. Propp. K. M. (1995). An experimental eX:llnination or bio logical ... ex :11; a ... talus cue in deci ... ioll-Illaki ng group ... :wd it ... influcncc on inlormalion u...e. Smal/ Grou{J Re.l l'lI !'ch, 26, .tSI--474. Pugh. M . D .. & Wahrman. R. (19S3). NClItr:lli/ing ,exi ... m in mixed-,cx groups: Do wmncn have to be reller than men '! Americall jOllm{l/ oj So('/olo,lp-. 88,746··762. Reed. V .. & Bllddeocrg-Fi<.,cher. B. (2(X) I). C.reer ob'laclc'" for womcn In medicine: An O\-Cf\ ic\\. . Ml'dimll::tlllctllirlll, 35, 139-1.t7. Rhuaue ... , (\1 J R. (19g I). A <.,ocial p'yl:hologlL'al in\ e ... tig:ltlon nllhe differential intlucnce or m•• le and fcmak .llh(K:atc, of nomradllional ,cx rule ... (dOI'IUral dl"Crltltioll, Ball Stale UIll ... er"lIy, 1979). Dint'l"laljol/ Ah,llracl\ 1/lIi'mlll/(l/UlI,
.J I,
~7.t7
Rhode, D_ L. (!(XII). The 11I1(iIl/lhl't! OI,:t'l/(lll. \\'Oll/t'li lllld lilt' It'gal pm{eHio/l. Chicago: American Bar A"ociatulIl, C{)mmi"ion on Women in the Prufc",·.ion Ridgewa) , C (1978). Conformit). gruup-urientcd motivation, and ... talu, •• lIainmenl in ... mall group.... So(·ial P"dlolog l- QlUmerly . .J I, 175- 188 Ridgc\\a) . C L (19SI). !\um:unformily, cOIll(X'tenL'c and intlucnl·C In gmup ... : A le"'l of t\\.o Ihcofle .... 'Illainm Son·ologica/ Rl'I'it'lI' -If>, Ttl.- ",47 Rldge\\.-.IY. C L. (19X2), Statu ... III gfllUP': Thc llllponancc olmOli' IIlCOn'l,tenC) ;lI1d gcndcr, SII/al/ Group Resf'lIrch. 26, J72-399. Serbin. L, Spraf"in, C. Elman. r-.-L &. Doyle. A (1982). The carly dr:\elupmt:nt or ...cx-diffcrcmiatcd pal tern, 01 ,ociallnllucnce. Cc/I/(/(lIl1l1 jrl/ll'lllli of' /Jehtll'io/'{/I S<"II'IIC·(', 1./, JS0--36J. Shnc"elfnrd, S., Wood. W .. & Worchcl. S. (1996). Bcha\lural ",ylc ... and thc intluencc of "omen in mixed-,c\ group" Social PH-dlology (JulIrlal,·, 59. 2X+.-29J Son. L.. &. Schmitt. N (1983). The inilue nce nl ,cx bw, upon cnrnpHance \~ilh cxpcn power. Sex Roll',\, 9, 2.tl l.t6 StelTl. R T, & Ikllcr. T. (197!}). \n cmpiril'al anal),i, of Ihc correlation, bct\~ccn Icadcr-hip ,talll'" and PJrtlL'lp~lII()n rate ... reponed III Ihe I1ter.llurc. JOllrnal (~r Pl'nOfw/iI\' alld Social Vncholot:" , J7. 1991-lIX)2. Swim, J , Borglda. E., Maruyama, G .. & ~1yer .. , D G. (1989). J()~1ll McKay ver"'u'> John McKay : Do gender ... tercotype, bia<, c\aluallOn ... ? Pwd"l/o.t:;('lI/ /JIII/t'l ill , 105, .t09--429. Twengc, J. M. ( 19(7). Altillldc ... toward women, 11)70-1995. P.H'(""ol()~y of Womell Quarter/.\', 2 I, JS-51. U.S. Bureau of Labor Slatbtic<,. (20(H). / ,1IIwrjim 't'1I111;'\ lin./I"OI1I lite ("111"1"('111 popu/olirm \'UrI'e.\": Amlll(l/ m·l'm,!:t',' //(1//,eliold daw. Table 2: Emphl) ment 'latll" 01 thl! civil ian nonin<;litutionallXlpulation 16 year, and mer b) 'iCX, 1969 to dale_ A\ailnbk onlinc: http ://ww\\..ce nsu<,.gov/hhc ...lincomc/ hi,ti nc/p.'6.ht Ill!. C.S Bureau nl the Cen ... u~. (lOOO). f!/\/or;ml iI/Will!' w/J/('\. Peoplt·. Table P36: Full tlllle. yeilr-round \\.-or"cr.. (all racc') by mcdlan incmm! .lIld ... r.:x: 197010 :!()(X), A\ailable online: Interm:1 http:// \\ \\ \\ _bl"',gmkp"homc.hlm#charcmp
148
Part II • Persllasion Variables: Perspeoil'es on Sources. Receivers, Channels. and Me ssages
U.S. Depanment of Oefen!".c (1998). i\Cli\'e dUlY miliulI)' personnel by sen'ice by rank/grade (for September 30.1997). Ava ilable online: http://webl.whr.;.osd. mi l/mmidlm ilitary/miltop. hlm . Wagner. D. G .. Ford. R. S .. & Ford. T. W. (1986). Can gender inequalities be reduced? American Sociolo;:ical Rel'iel\'. 51. 47-61. Walker, H. A .. Ilardi. B. McMahon. A. M" & Fennell. M. L. ( 1996). Gender. interaction. and leader~hip. Social Psychology QIUlrlerly, 59, 255-272. Weimann, G. (1985). Sex difference!". in dealing with bureaucracy. Sex Roles, 12,777-790. Will iam,. J. E.. & Be~t. O. L. (1990). Mea.w";"1: sex slereotypes: A mll/lillmiollal sWdy. Newbury Park. CA: Sage. Wood, W .. & Kallgren, C. A. ( 1988). Communi cator attributes and persuasion: Rec ipient s' access to attitude relevant information in memory, Personality and Social Psychology Bullerin, N, 172- 182. Wood. W.. & Kanen. S. J. ( 1986). Sex differences in interaction style as a product of perceived sex differences in competence, JOllrnal of Personality and Social Psychology. 50. 34 1-347. Wosinska. W.. O;'lbu l. A. J .• Whetstone-Dian. R.o & Cia ldin i. R. B. (1996). Self-presentational responses 10 ,uccess in the organizali on: The costs and benefits of modesty. Basic and Applied Social P.~y dm/ogy, IR.229-242.
c..
9 Language Expectancy Theory Insight to Application Michael Burgoon and Jason T. Siegel
Lallguage Expectancy Theory: Explication The book you presently have open is replete with a variety of theoretical framcworh designed to increase your understanding of persuasion and social intluencc. Unfortunate ly.
frequently students of socia l intlucnce arc presented with theories of the literature in a f'a"hion thai falsely suggests a ralher simp li stic journey from thought to insight \0 theorie s
100
of human behavior. However. of obvious import to those studying persuasion. and of vita l imporl to those who wish one day to be on the appa-,ile side of the metaphoric textbook . is how theories of persuasion are spawned from what sometimes begins as liLtle more than a \ccd lin g of thought. As Michael Burgoon staled in creating Language Expectancy Theory (LET): " ... thi~ theoretical formation ha ... I4Ikcn
149
~O l~
Part II • Penu(I.\ioll Variahle.\: Perspeclil'es OIl Suurces, Receh'ers. Challne/s. and Messages
paradigm. which centers on how the language and expectations can work in tandem to
decrease or inhibit the persuasive strength of a future persuasive attempt, a message yet to come. However. rather than addressing all three paradigms. this chapter will just present a detailed explanation of the passi ve paradigm. the most studied and practical use of persuasion principles (for a detailed review of all three paradigms of LET sec Burgoon. Denning. Roberts. in press).
Language Expectancy Theory: The Core Propositions Two studies have been credited by M. Burgoon and Miller (1985) with being catalysts for the core propositions of LET: Brooks (1970) and McPeek and Edwards (1975). While the story of how Brooks's study came to the attention of Burgoon has been told elsewhere (Burgoon. 1995). due to its ~cll1inal importance to the development of this perspective. this story will be repeated one more time. Burgoon sent off one of hi~ al1icle~ inve~tigating what would become part of LET'~ ae/h'e paradigm (Burgoon 1970) to be considered for publication in a journal. In response. the editor of the journal. RobcI1 Brooh. sent him a copy of another article. which had been accepted for publication but not yet published. While the study itself was of interest 10 Burgoon. it was the concluding paragraph of this piece that he credits with the insightful discussion that eventually led 10 the development of LET. This conc lud ing paragraph reads as follow\: , , . the possibility of contrast effeels should be considered. This principle assumes that we carry stereotypes into sllch social situations as the public speech. There. the speaker\ behavior may be discrepant with stereotyped expectations. If the discrepant expec tati ons still cannot be assimilated or ignored, they are likely to b~ exaggerated in a li Mener's perception. So viewed. mere civil behavior on the parI of Malcolm X may be perceived as extraordinarily genteel by an auditor who expecb barbaric actions from a black (sicl nationalist. One explanation, .. is this: unfavorable (or favorable) speakers may be per~ ceived more (or less) favorably not becau..,c their behavior is intrinsically persuasive (or dis..,uasi\,c) but becau\c it contrasts with \tcreolyped expectations which audiences hold for notoriou\ (or popular) public figure.." (p. 155) This article by Brooks (1970) led Burgoon to begin thinking abollt the nature of stereolype~ and also the ingredients that comprised and determined expectations. Burgoon began to ponder whether there arc expectations that are specific not only lO a person but also to a culture. Also. if expectations about behavior exist. are there also enduring expectations about the pattern':. of ordinary language? La.'.tly, do expectations of language use, if they do indeed exist. differ based on a speaker's gender. socioeconomic c lass, and/ or credibility? Burgoon and Chase (1973) took thc.'.c skeletal notion.'. of expectations violation.'. and investigated what occurs when receivers' expectations of language intensity are intentionally violated. Specifically. Burgoon and Chase (1973) hypothesiLed that if participants in
Chapter 9 •
LnllgfW!W Expectcmc\, Theon'
151
thc ~lUdy were pre~ented with a pretreatment message (message I) of high intensity. they would expec l a follow-up message 10 be of al lea" equal inlensily (see appendix A following thi\ chapter for an example of a low- and high-intensity me\\age). However. if the participants' expec tations were violated with a message of moderate intcnsity. they wou ld see the ~ pcak er as more ·;reasonab le." A':t a resu lt , the speaker wou ld be more successfu l wi lh a follow-up appeal (message 2). Mo reover. Burgoo n and e ha" predicled Ih al people who were initiall y ex posed to a message of low intens it y would s imilarl y expect th e follow -up me s!'!<.Ige to be o f equal intens ity. It wa!'! rcasoned. therefore. thai if these people initiall y received a message thai wa\ ei ther of mode rate or hi g h intensity, they would not ex pect a follow-up message of even g reate r inte m,il y. Thm, th ey wo uld be most persuaded by the mod e rate or hi gh-inten si ty :-.econd message. Burgoo n and C hase further predicted that if people expected a foll ow-up persuasive appea l of moderate intensity and rece ived \ uch a me\sage. their expectati ons would not bc violated. Thu\. they wou ld perceive the mes:-.age a\ more reasonable. Howe\'er. because they wou ld expect this levc l of intensi ty, the per\uasive strength of the message would bc minimized. To test th ei r hypo thes is. li ve wcek':. prio r to an actual experi me nt , stude nt s were prese nted with a pretest questionnai re asking them the ir op in ions on a varie ty of campus and nati o nal i\\ues. Considering that th e hypothesi:-. was predicting the studen ts' respo nses when prc~ented wi th a (second ) refutational message (a message they wou ld unquesti onably dio;;ugrce with). it was cruc ia l for the experimen ters to find an is\ue to which students were ununimous ly opposed. The i\\uc that wa\ most offens ive to the \tudents wa'\ a propo\cd plan by th e uni \'e rsity to admi t o nl y \cniors and juniors. Specifically. o n the day of the aCllla l experime nt. a ll subject\ who tool-.. pa l1 were unanimously opposed to such a plan. In the ex perimental sessio ns, 11 4 partic ipants were ra nd o ml y as!>.igned to seve n diffcren t cond itions: one con tro l and s ix experi me ntal. All su bj ect!>. received a hi gh-, moderatc-, or low- int e ns ity prc treat mc nt message arguing for the policy disallowing fres hman or sopho mores from being admittcd. A ... discu ... scd. this pretreatment message was used to create expec tat ions on the part of the receiver. For examp le. if a panicipant received a s upportive mes':tage of low ime n... ity. it wa ... a ...... umcd that thi s would crea te an expectation on the pan of the participant that the next me':tsage would be of similar intensity. Following the pretreatment message. a ll \tudents recei\-ed a message of moderate intcnsity. A\ predic ted, those student\ who had heard a mcs\age of high intensity and then heard one of moderate intensity \. .'ere "iignilicantly more persuaded than those student s who had heard a message of moderate illl cn",ity followed by another message of moderate intcnsity (posi ti ve viola ti on ). Thc reason the vio lati o n is cons idered positive is that the me ssage was Icss intense than expectcd. S in ce the first ll1essHge was hi g hl y int ense, the audience prepared for a message o f simi lar intensity. Howeve r, when the message was of on ly modcrate int ensity. the audience's expectatio ns we re posit ive ly vio lated. The c han ge is cOIl\ide red a positive vio lati o n becausc the enacted behavior was better or mo re prefe rred thall that ex pected in th e silliati on. A positive violation can also occur whcn ncgative ly eva luatcd sou rces confo rm mo re c lo.!o.e ly than expected to cu ltural va lu es. ~ocietal norm ... , or \ ituation a l exigencies. Returning to the expe rimen t. th ose stude nt s who heard a messagc of low inte nsi ty followed by a mess~ge of moderate intensity we re also s.ignifi ca ntl y more pcrsuaded (a ltho ug h the vio lati on wa"i nega ti ve) than those (unprepa red)
152
Part II • Persuasion Variables: Perspecli\'es Ofl SOllrce.\. Receil'ers, Channels. and Me.{,(wgl" ~
students who did not have their expectations violated (i.e .. the students who heard a me~ sage of moderate intensity followed by another message of moderate intensity), The reason this re~ult occurred is that the student~ who had originally heard a message of low intensity dropped their counterarguing defenses. expecting another message of low intensity. However. when the second message used language of greater intensity than they expected, the students were unprepared to counterargue it. The violation is considered negative. because the language being used fell outside the bandwidth of socially acceptable or expected behavior in a negative direction. Two years later. McPeek and Edwards (1975) published a second major precursor to LET. While Burgoon and Chase (1973) investigated the effects of violating expectations that were created by the investigator. McPeek and Edwards (1975) investigated the outcome of speakers who argue for a position that is opposite of what would be expected. based not on expectations set by the investigator. but on expectations set by societal and cultural norm~ with the following claims and questions : Interpersonal perceptions and beh
McPeek and Edwards (1975) made predictions concerning the effectiveness of antirnarijuana messages based upon how "expected" the message source would be to make such an argument. It wa~ posited that a hippie arguing against marijuana. the obviously unexpected position. would be seen as more credible than a seminarian making the same. albeit expected. anti-marijuana argument. Conversely. it was hypothe~ized that when a hippie took the expected pro-marijuana position. he would be less persua ... ive than a seminarian taking an unexpected. pro-marijuana. position. When a study was conducted. the hypothese, gained only partial support. The hippie pUlling forth an anti-marijuana argument was more persuasive than the seminarian arguing the same position: however. the .scminarian arguing for marijuana was no more persuasive than the hippie putting forth Ihe same argument. However. it is important to note that the communicators taking unexpectcd positions. whether the hippie arguing against marijuana or the seminarian arguing for it. were perceived as being more honest and more sincere then the commun icators who took the expected position. At about the same time that the McPeck and Edwards' (1975) piece was published, a piece by Burgoon. Jones. and Stewart (1975) introduced what would become the initial roots of LET. This seminal article provided evidence that persuasive success can be mod-
Chapter 9 • Language Expectancy Theory
153
erated by the linguistic choices made by the communicator. Specifically. Burgoon. Jones. and Stewart put forth the following propositions in relarion to LET's passive paradigm: Proposition A: Attitude change is a function of the level of language intensity in a persuasive message. type of persuasive paradigm employed, and the receiver's expectatIOns of the source's communication behavior. Proposition B: Given the passive message reception condition. when a source uses a level of language intensity that violates the receiver's expectations in a positive manner. significant altitude change will occur in the direction advocated by the source. Proposition C: Given the passive message reception condition. when a source takes an unexpectedly intense position. it will result in minimal or even negative attitude change. At the time this piece was published, the investigation of language intensity was by no means novel, nor was the investigation into expectations, but this was the first piece to combine the role of societal expectations and language intensity in persuasion research on an a priori basis. As put forth by Burgoon and colleagues: '" ... all present evidence suggests that expectations develop in receivers about syntactic. linguistic. and pragmatic variables in persuasive messages. Most of the studies explained unexpected findings on a post hoc basis" (Burgoon, Jones, & Stewart, 1975, p. 243). In short, the realization that violating receivers' expectations could be a moderating variable in the equation of persuasion was stepping into the spotlight. The experiment to test these propositions was similar to prior experimental designs. Two messages were created. this time concerning a required GPA of 3.25 to enter the university where the experiment took place. These messages were identical except for key verb phrases that were either highly intense or low in intensity. One of the key tests of this experiment focused on societal expectations of the communicators of persuasive messages
based on gender. The hypotheses were as follows: Hypothesis I: Male receivers will demonstrate less altitude change than will female receivers. Hypothesis 2: There will be an interaction between language intensity and sex of the source such that a female source will be most effective with low-intcn!<.c language and a male will be least effective with low-intense language. This was the first lest in this line of research where expectations were not created by the experimenter but were expected due to societal norms. Explaining the logic behind their prediction. Burgoon. Jones. and Stewart stated: Bem and Bern (1970) <;uggest that the socialilation process. whether it be intentional or not, has programmed females to be complementary rather than independent of the male, submissive rather than dominant, domestic rather than business- or scientific-minded. and 10 generally consider themselves Jess knowledgeable than men. If the suggested stereotypes of submissiveness are correct. then certain communication behaviors would be expected to differ for men and women. Females :.Ire probably expected 10 be less intense than
154
Pan II • PerM/wioll Variables: Perspeclil'es Oil Sources, Receil'en, Chanllels. {Illd Me.'i.'wges males. lr highly inten~c encoding by a remale occur~. a "boomerang errect" should occur because she will be taking an "unconventionally and unexpectedly strong position." A male who is expected to be strong will likely be less erfective using low-intense language. (p.
245)
This hypothesis was supported. The female speaker was more slIccessful when using a low-intensity message than when using a highly intense message: and the male speaker was more persuasive when using a highly intense message than when usi ng a lowintensity message. This finding was crucial for two reasons. First it gave support to the notion that there are societal expectations concerning hlllguage and language intensity; second. it showed that violating these expectations could impair the persuasive strength of a message. Another area of intereM was the arena of fear appeals. Fear appeals are any messages that use fear or anxiety arousal in the hope of ')caring the audience into complying with the message. Obviously, knowing societal expectations in regard to the quantity of fear used in such messages is key to any speaker who wishes to scare his or her audience into compliance. Violating the expectations of the audience can either greatly increase or decrease the persuasive strength of the message. Unfortunately, carly studies of fear appeals presented often conflicting. seemingly confounded results. For example, one of the first studies on fear appeals found that mild fear appeals were more successful than strong fear appeals (Janis & Feshbach. 1953); however. other studies (e.g .. Leventhal & Niles, 1965) found the opposite resuh. while yet other investigations revealed no significant difference between level of fear and attitude change (Beach. 1966; Powell 1965). Fortunately, Hcwgill and Miller (1965) allempted to clarify the influence of fear appeals by combining the influence of fear appeals with the influence of source credibility. The study hypothesized the following; I. If a source has high credibility with a listener. appeals that elicit strong fear for persons highly valued by the listener will affect greater attitude change than appeals that elicit mild fear. 2. If a source has low credibility with a listener. appeals that elicit mild fear for persons highly valued by the listener will affect greater attitude change than appeals that elicit strong fear. Taken together. these two hypotheses predicted an interaction between level of fear and source credibility. To test these hypotheses. the investigators split up 90 subjects into four experimental groups and one control group: (I) high-fear message, high-credibility source; (2) high-fear message. low-credibility source; (3) low-fear message. highcredibility source; and (4) low-fear message. low-credibility source. Each group heard a message emphasizing the advantages of community shelters while stressing the disadvantages of family shelters. It's important to note that this study took place at the height of the cold war when people feared a nuclear attack. The four messages contained the same content, but the high-fear messages contained 13 statements concerning physical injury or death to "pouse or children. Specifically. the high-fear message stated the following:
Chapter 9 • Lallgll(l~e £~pectallcy Theon
J55
Unle:\" proper shelter precautions are taken against fallollt. the children of thousands of familie~ would be killed in nuclear war. The mo~[ dangcrou:\ fallout i~ the early fallout that ralls within 24 to 48 hour... after the thermonuclear explo\ion. and thi:\ is the primary hazard frol11 which we ~hould be prepared to protect our hu:\bands or wive .... and our children. since many of them will perish if ... lIch protection is not av'lilable. Even minimal community shelter precautions would spare the lives of thousand ... of adults and children. Chi ldren would probably suffer I.;evere radiation hurn ... while going frolll school to home and then have to endure radiation I.;ickne:\.., while confined in a ..,111 ..111 family shelter. (Hewgill &
Miller. 1965. p. 96) Credibility was manipulatcd by either attributing the message to a professor of nuclear research or lO a high school sophomore. After each group heard their message. they immediately underwent a series of teMs to assess their attitudes toward community and family shelters. As prcdicted. participants who heard a message containing a high-fear appeal attributed to a highly credible source werc more pcr..,uadcd than those participants who heard a message employing a mild-fear appeal attributed to a low-credibility source. Additionally. the participants who heard the message using high-fear appeals attributed to a highly credible source were significantly more per..,uaded than any of the three other groups: however. there were no differences in attitudes between the groups who heard messages attributed to a low-credibility source. regardless of the fear appeal used. Of import is that in a follow-up study. Miller and Hewgill (1966) hypothesiLed that a lowcredibility source using Mrong language would reduce hi" or her credibility further. whereas a highly credible source would increa~e his or her credibility by using a strong fear-arousing message. Their hypotheses were generally 'Iupported. Thc reason lowcredibility sources using highly intense language are problematic is that cultural expectations grant high-credibility sources a greater range of options in their language use. For example. a full professor giving a passionate. high-intensity ..,peech about the need for stricter college standards would likely be accepted: ho'Wcver. imagine a young teacher's assistant giving the same pa~sionate. high-intensity speech. The teaching aS50istant would be seen as unprofessional and possibly overemotional. Simply put. if someone of high credibility uses extremely inlen..,c language, his or her language use is perceived to be a necessary act. However. a low-credibility source using the same language may be pcrceived as "out of control" or immature. By incorporating re!o.carch in fear appeals. compliance gaining. language intensity. opinionated language. and <.l host of other message variables. amongst others. the following propositions of LET's passive paradigm were put forth (Burgoon. 1989: Burgoon. 1995: Burgoon. Denning. & Roberts. 2001: Burgoon & Miller, 1985): Proposition I: People develop cultural and sociological expectations about language behaviors that subsequently affect their acceptance or rejection of persua,ive mcs~ages.
Proposition 2: Usc of language that negatively violates societal expectations about appropriate persuasive communication behavior inhibits persuasive behavior and either results in no attitude change or changes in position opposite to that advocated by the communicator.
156
Part II • Persllasioll Varillbles: Perspl'clil'l'.,
Oil SOlfr('(',\',
Nt'ceil'as. CltalllU'/s.
(/1/(/
Message"
Proposition J: U~e of language that po. . itivcly violate . . ..,ocietal expectations about appropriate per..,uasive communication behavior facilitate~ per~uasive effectivcncs.." Proposition -+: Highly credible cOllllllunicators have the freedom (wide bandwidth) to select varied language strategies and compliance-gaining techniques in developing per. . ua . . ive messages, while low-credibility cOllllllunicators nUl..,! conform to 1110re limitcd language option, if they wish to be effective, Proposition 5: Because of the nonnative impact of ... ouree credibilit), highl) credibh: ..,ource . . can be more succes..,ful w.,ing low-intclbity appeals and more aggres\ive compliam:c-gaining Illc\sages than low-credibility communicators lIsing either strong or mild language or morc prosocial compliance-gaining strategies, Propo. . ition 6: Communicators percei\'ed a ... low in credibility or those unsure of their perccivc.::d credibility \\ ill usually be more persugaining Jlle~sages .
Application of LET: The Family SUI! Safety Campaigll Although the ability to explain pa~t experiments shil1e~ positive light on a theory, the spot light ,hincs much brighter when the theory can be applied to a real-world context. Such i.., the case with the LET pa\si\'c paradigm, In 1999 alone. -+-+,200 people were diagno<.,ed with life-threatening malignant mcb noma: 7,300 of them died. Thi" figure i<., e\-en more staggering in light of the fact that 90 percent of skin canccrs are preventable (America Cancer Society. 1999), \Vhilc there i<., no shortage of way~ to decrc<1sc these figures- sun<.,crcen" usc of shade. etc,-the difticulty lies in persuading individuals to comply with the afol'l!ll1cntioned preventive measures. Thc goal of thi~ Family Sun Safet y Project. funded by the National Cancer lil \titute. wa~ rather straightforward: create sun ,afety l1le~sage~ differing in language intcn\ity and logi cal :-.tyle while using LET'::, explanatol') power to hypothe . . ize v.hich me..,..,ages will be 1110..,t effective. Based on LET. it wa.., predicled that a sun safety campaig n containing more intense language would be more successflll than a campaign using low-intensity language. since pa~t empirical re<.,earch suggested that Sd100ls and pediatricians, the source.., to which sun
,,,rety message, \\ere attributed. arc highly credible (Builer. Calli"er. & Reichert. 1995). As discu~sed carlier. communicators of high credibilit) are granted a "wider bandwidth" of acceptable communication, wherea~ "'peakers of low credibility are relegated to a much narrower bandwidth. Practically speaking. a ..,peaker of low credibility is excluded. under
Chapter 9 • Language £rpecuUlcy Theon'
157
penalty of a potential boomerang e ffect. from lIsing aggressive strategies such as threats. hig hly intense language, and fear appeals. A low-credibility speaker who chooses such aggressive strategies over nonaggressive strateg ies such as promises. less intense language, and reassu ran ces will negatively vio late the receivers' expectations. On the contrary, speakers of high credibility who decide to use the more aggressive language st rat egies j ust mentioned will be treated to an increase in the persuasive strength of their message (e.g., Burgoon, Dillard, & Doran, 1983). Based on th is theoretical rationale and empirical findings, it was predicted that a sun safety campaign containing more intense language wou ld be more successful than a campaign usi ng less intense lan guage. The logi c was rather straightforward: Since highly inten se lan guage is not the norm for Illost campaigns focusing on health promotion. such intense language would be a violation of expecta ti o ns. Furthermore, since past research has put forth data suggesting that schools and pediatricians. the sources of the slin safety messages, are highly c redible (Burgoon, Birk. & Hall. 1991). there exists a greater bandwidth of acceptable language for these sources. It was predicted that since there was a large bandwidth of acceptance, the inte nse language would be perceived as a po<.,itive violation and would therefore produce greater compliance with the sun safety advice than the message lIsi ng low-int ens it y language. The low- intensity language messages do IlOt negatively violate expectatio ns, but rather conform. Therefore, while (here will not be a boomerang effec t, the messages using highly intense language were predicted to be more effective than the former. Statements with high-intensity language included the fo ll owing: "Skin cancer is a grotesque growth of skin cells." "Treatment of skin cancer involves removing tumors from th e skin," and "Tragicall y. about 7.200 Americans will die from melanoma, a very se ri o us type of sk in cancer, this year a lone." Statements of low intensity in c luded the following: "Skin cancer is an unusual growth of skin cells" and "Sad ly about 7.200 Americans will die from melanoma, a ve ry seriolls type of skin cancer, this year alone" (Bu ll er. Borland. & Burgoon. 1998, p. 450). Parents were recruited for the Family Sun Safety Project from elemenliJry schools and a pediatric clinic. Participants received sun safety prevention messages in the form of newsletters, brochures, and tip cards. The project accrued over 800 familie~ and lasted over a year. In short. as predicted by LET, messages llsing intense language produced more compliance than messages of low intensity. Specifica ll y. as reported by Buller, Burgoon. Hall , and colleagues (2000): Compa red to parents receiving low-intensi ty Illc!')!)age.''I. parcnts in the high illlcll,;ily message (a) dec reased the frequency of their own midday SUIl exposure during the current summe r. (b) were more likely to say that Ihey in general planned to protec( (heir children more next summer and planned to protec( thcmselves and their children more in the upcoming winter, and (c) reported a larger increase in thcir planned frequency or applying sunscreen, using protective clothing. and limited midday sun exposure with their children in the upcoming winler. (p. 108) In summary, by using the LET proposition framework. the experimenters were able predict successfull y whic h messages wou ld be more effective purely on the basis of the intens ity of the language used. Most important. multiple behaviors were changed and those changes persisted over a long period of time.
10
158
Part II • PerllU/sioll V(lriabh'J: Pn.\ pl'Cfh'(',\ 01/ SOIIIH'\. Rcceil'('/'\, CI/{lIIl1d'l. (/lId Me.\.\(/ge.\
LET's Passive Paradigm: New Directions To Mate that LET'~ passive paradigm is Just beginning 10 hit its !o.Lride might be an understatement. In the past decade the theory ha", been succes!o.fully used to (I) e.xplain the ... uccess of negative altack campaigns in politics (Pf3u. Parrott, & Lindqui",t. 1992). (2) explain compliance and satisf3ction with physician!o.' instructions (M. Burgoon. Bir"-. & Hall. 1991). (3) assess the etTcctivencs", of communication strategies de ... igned In improve both initial and long-term medical adherence (Klingle. 1993: Klingle & M. Burgoon. 1995). and (4) improve adherence to \un ",arety recolllmendation", (Buller. Burgoon. Hall. Levine. Taylor. Beach, Melcher. Buller. Bowen. Hun"iaker. & Bergen. 2000). H(J\\:ever. the theory continues to be ripe with opportunitic!o. for refinement and expansion. To COI1 clude. we shall present at least one venue where the theory'", framework can be expanded and one realm where the theory could hc fruitfully applicd.
Etllllicity and Culture JU\t as male\ have greater bandwidth", of acceptable ordinary language use than femalc.!"'. it "icems plau"ible that ethnicity might al",o be a ,ariablc that moderates expectation ..... Certain ethnic groups may be pcnalilcd. much a:-t female" arc. for using language of high intensity. This scenario could depend on the dynamics of the cuhure or cthnicity of the presenter and the culture or cthnicity of the audicncc. For example. a high-credihility source who i. . part of one . . ocial minority pre . . cnting to mcmber . . of a different !
Public Service Al/l/oul/cemel/ts Anyone who views a modicllm of tclt:vision ha"i undoubtedly . . een advertisements urging adolescents to avoid drug lise. While some of these advertisemcllts are unquc . . tiolll.lbly effective. LET cun be used to explain why ",ome ads (maybe 1110:-.1) may not be a", slHxc:-.sful as they could be (see Siegel & Burgoon. 2001 for a more detailed explanation). For example. one of the more popular ad"i highlight . . a teenager armed with a frying pan alerting the audience member... 10 \\ hat their friend"i und family '\-'ill go through if they u\e drugs. As the female teenager states in the commercial: 'This i\ your brain. This i", heroin. Thi . . i~ \\hat happen . . to your brain after "illorting heroin." The actre ....... then ~mashe . . the egg with the pan and exp0"ie"i the cracked egg ",Iiding off the back of the pan to the audience. The rest of the actress's dialogue is spoken 3", she U!o.CS the pan as a weapon to brea"- !
Chapter 9 • Language Expectancy Theory
159
highly credible by their target audience. If the audience doe, not perceive the creators of the commercial to be credible, the high-intensity fear appeal will at best be ineffective and at worst will cause a boomerang effect, persuading the audience in the direction opposite of tlmt advocated by the commercial. It is also possible that the actress in the commercial will lose some of her credibility.
LET: The Most Recellt Findillg Recent research supported by the National Institute of Drug Abuse through a multimillion dollar experiment using LET to explain adolescent reaction to antidrug inhalant and marijuana messages (Burgoon 2001, Grant #DAI2578) obtained interesting results about cxpectations regarding media commercials and interpersonal communication. This experiment involved over 1.300 adolescent students exposed to antidrug/inhalant messages. While the participants all saw the same message!-.. students were exposed to footage of one of four different scripted focus groups discussing the advertisements. Specifically, after viewing each of three advertisements. the students either saw a focus group consisting of peers or adults using either implicit or explicit antidrug language. This resulted in four different experimental conditions (adult explicit. adult implicit. peer implicit, peer explicit) discussing both the advertisements and the topic of adolescents and drug/inhalant use. In the implicit condition, the individuals made statements such as, "When it comes right down to it ... it's your own decision" or "Why would anyone want to do that stuff?" In the explicit condition the comments were more to the point and outspoken: "Don't do drugs ... it's as simple as that" or "Only complete idiots do drugs" (see Burgoon , Alvaro, Broneck. Miller. Grandpre. Hall. & Frank, 2001). The research literature, without much sound scientific support. has continually suggested that peers are more effective than authority figures in persuading adolescents about lobacco prevention/ce~~ation, alcohol use. and drug uptake. Language expectancy theorists wou ld question sllch an intuitively appealing, but simpli~tic asseSsment. Specifically, LET would posit that only those conditions Lhat positively violalcd expectations would produce desired changes. and those would probably not be 1110st likely in peer-la-peer situations. Ralher. it was predicted thal students would have their expectations positively violated by the adults who used implicit language. since the notion of adults offering ado lescents a choice when it comes to drug use is certainly not the social norm. Furthermore, it was predicted that adults using explicit language whcn discus:-.ing adolescent drug use wou ld conform to social norms. thereby confirming expectations. In ~hon. it was hypothesized under the LET framework that the adolescents would react more favorably toward the adults who used imp li cit language than to the adults who used explicit language. Preliminary results suggest that not only were students significantly more favorable toward the adults who used implicit language. but the adults' lise of implicit language was such a positive violation of expectations that the students preferred the scenes of adults using implicit language even more than seeing their own peers discussing drugs or inhalants, regardless of the type of language u,ed. Moreover. the students who viewed the adults using explicit language were less favorable toward the adults than were students who viewed any of the three other conditions (adult implicit. peer implicit. peer explicit).
160
Part II • Persuasion Variables: Perspeczil'es on Sources, Receil'ers, Channels, and Mes,'wges
The set of experimental questions focused on how participants hearing the focus groups discuss the advertisement would influence the students' perceptions of the ads themselves. Specifically, after viewing all the ads that were all followed by footage of one of the four aforementioned focus groups, the students were asked to evaluate the advertisements without taking the focus groups' comments into consideration. The results indicate, as LET would suggest, that the students who had their expectations positively violated by witnessing the focus group of adults or peers using implicit language to discuss drug use also reconstructed their evaluation of the advertisements. Even though all four conditions viewed the same advertisements (remember only the video of the focus groups differed, not the ads themselves), the students who viewed the implicit focus groups were significantly more likely to evaluate the advertisements positively than were any of the three other groups, Additionally, the students who viewed the ads supplemented with the adult-explicit focus group found the ads themselves significantly more controlling than the students who viewed the ads in any of the three other conditions, These findings suggest that not only can the violation of language expectations alter how we perceive the speaker, but the violation can also change our evaluation of what the speaker is speaking about. Once again, all four groups viewed the same advertisements; only the conversations heard after the ads were viewed differed. These results suggest that we construct or reconstruct our evaluation of a message based upon whether what is said about the messages confirms or violates our expectations. Additional implications for this finding can also lead one to muse about the political arena. After televised political speeches or debates. almost all news networks provide some form of commentary. These recent findings suggest that when a political speech is followed by a newscast of pundits, talking heads and analysts can positively or negatively violate expectations with their own conversation, thereby changing our perception or the political figure and his or her message content. That such commentary influences public opinion is not new information. However, this interpretation from LET puts in high relief how important such analyses can be not only in shaping perception of public figures but also in actually changing the evaluation of message content already processed. In sum, a theory must be judged not only by how well it fares under scientific scrutiny but also by the extent to which it informs us about how humans behave in everyday situations. It is certainly our contention that LET, in its main incarnations. has risen to these challenges in an admirable manner.
Referellces_____________________________ American Cancer Society (1999). Callcer facts and figures. Atlanta. GA: American Cancer Society. Aronson. E.. & Golden. B, (1962). The effect of relevant and irrelevant aspects of communicalor credibility on opinion change. Journal of Personality. 30. 135- 1--1.6. Beach. R. I. (1966). The effect of a "fear-arousing'" safety film on physiological. attitudinal and behavioural measures: A pilot study. Traffic Safery Research Ret'iell", 10,53- 57. Bern. S. L.. & Bern. D. J, (1970). Case study of a nonconscious ideology: Training the woman to know her place, In D. J. Bern, Beliefs, atritudes and hUlIlan affairs. (pp. 89- 99). Belmont. CA: Wadsworth. Brooks, R. D. (1970). The generalizability of early reversals of attitudes toward communication ~ources. Speech Monographs, 37. 152-155,
] 61
Chapter 9 • ulIIglIllge £l;pectallCY Theorv
Buller. D. B.. Borland. R.. & Burgoon. M. ( 199R). Impact of behavioral intention o n effectivenc<; ... of me<;<;age feature<;. EI
433-453. Buller. D. B.. Burgoon. M .. Hal1.1- R .. levine. N.. Taylor. A. M.. Beach. B.. Buller. M. K.. & Melcher. C. (2000). Long-term effects of language intcnsity In preven tlvc me""age .. on planned family !'olar protection. Hi'llflh comlmmimtion. 12(3).261 -275. Buller. D. B.. Burgoon. M .. Hall. J. R.. Le\lnc. N., Taylor. A. M .. Beach. B. H.. Melcher. Buller. M K .. Bowen. S. L.. Hunsaker. F. G .. & Bergen. A. (2000). U,ing language Intensity to Incre3!'e the ..,uccc.., ... of a family intervention to protect children from ultra\ iolet radiation: Prediction<; from language cxpectancy theory . Pl"el'ell/il'e Medic;'/(': AI/ Il/lemarimUlI Jounral Dl'l·oted 10 Pmctice &: Theory. 30(2). 100- II ..n. Buller. D. B .• CalliMcr. M . A .. & Reichert. T. (1995). Ski ll cancr.:r prcvention by parent<; of young chil .. dren: Health infonlmtion source ... ski n ca ncer knowledge. and ... un-protection practice .... Ollcolog\' Nllr\';lIg Forum, 22. 1559- 1566. Burgoon. M. (1970). The effect ... of re~pon<;e ..,ct and race Oil Ille<; ...agc Illlerprelaiion. Spe('ch Mot/o!:l"lIplu,
c..
37. 264-268. Burgoon. M. (1989). The cffect<; of me<;sagc variab les on opinion and :Hlit ude change. In J Bradac (Ed..,.). Mef'l(lges in ClJllUlllmi{,(lI;OIl fciellct': COII,elfll'Ol"llrYlipprotlChel 10 ,he\l/I(h of eITeC1\ Cpp. 1291(4). Newbury Park. CA: Sage Burgoon. M. ( 1995). Language expeclancy theory : Elabora ti on, explication. and exten"lon. In C R Bergcr & M. Burgoon (Ed .... ). COII/mlll/icarioll (IIlllweilll il/!1f1ellce I'ron'\'.\es (PP. :!9 -52). E
cial P'Iyclwlog\", 48. 79-9::!. Ke lly. G. A. ( 1955), 711e psychology of pen mill I COnHI"IICH. Nc~ York: Norton. Klingle. R. S. ( 1993). Bringing time into phy..,ician compliance gaining research: Toward :1 Reinforcement Expecta ncy Theory o f strategy effectivcnc..,." I-It,tll'h COIIIIIIJllliwfioli. 5. 2K3-308. Klingle, R. 5 .. & Burgoon. M (1995). Patlcnt compliance and .. ati<;faction \\lIh ph;..,ician influence
162
Part II • Pl'rwwioll Variables: Penpecrire.f 01/ Sources. Receil'eT.f. Channels, and Me.fSages attemph: A reinforcement expectancy approach to compliance·gaini ng over time. Commllnication Re,'ie(lrl'h. 22. I~8-187 . Leventhal. H .. & Nile<;. P. ( 1965). Per":oi":olcncc of inllucnce for varying durations of exposure [0 threat >.,timuh. PfH'hological Repon'i. 16, 223- 233 . McPee~. R. W .. & Edward .. , J. I). (1975). Expectancy disconfirmalion and altitude c hange. JOllrnal of Social /'Iyc!wlogy. 96, 193 -208.
Miller. G. R.. & Hewgill. M. (1966). Some retent rc,-earch on fear-arousing message appeals. Speech Monogrtll'hl , 33. 377- 391.
Pfau. M .. ParmI!. R.. & Lindqui>.,1. B. (1992t. An expectancy theory explanation of the effectiveness of polilica l allac~ Ielevi .. ion ,pOI<;: A C
235- 253. Powell. F. A . (1965). The effech of anxicly-arou,ing Ille~~age~ when related 10 personal. familial. and inlcrpcr\onal rt:rerenb. Sl'l,t'ch MOl/ograph~, 32. 102-106. Siegel. J. T., & Burgoon. J. K . (2(X)I). Expcctam:y theory approache~ to prevention: Violating adolescent expectations to incrca .. c the cffec li vencio.":o of public servicc announcemenls. In W. D. Crano & M . Burgoon (Ed ... ). Mm, Media mlfl Dru,~ Pn' I'emiolf: C/a.'·Jic {lml Cmllemportln Theories and Re{eardl (163- IH6). Mahwah. NJ: Erlbaum .
Appendix A
High-Intellsity Message (Hamiltoll, HUllter, & Burgooll, 1990, pp. 243-244) The la ws regulating the sale of heroin in this country have frequently done more harm than good. both to society and to the individual who mu!',t use heroin. The public is confronted with an astronomical number of crimc!oo committed each year in every major city
by addicts desperate for money to ,upport their habit. The addict suffer; not from heroin, but from painful I.,ccondary comp lication ... which are promoted by the drug's continued illegality. In England. where the government controls the legal sale to addicts, heroin-related crime!>. arc non-cxiMent. Crime is not actually caused by the drug itself. but by completely outdated laws which prohibit its use. In the United Stales. addicts are driven to commi t crime"> againM innocent citilcns to obtain money to pay exorbitant black market prices charged by their underworld ~lIppliers. As a resull of these hugely expensive (ransaction~. law enforcement agencies are co nstant ly tempted by graft. Many heroin addicts die needlessly from disease caused not by the drug, but from agonil.ing secondary complications. Medical au thoritie~ now Mrongly agree that heroin cause ... very little physical damage. Symptoms of heroin withdrawal are not nearly as dangerou~ as those associated with alcohol. Yet in New York City last year over 900 addicts died from tetanUI! and hepatitis caused by improper means of injection. Addicts almost alway ... re-use and share fillhy needles. or improvil!e with objects not designed for injecting <.Irug~ into (he bloodstream. becau\e hypodermic ~yringes are not legally available.
Low-Intensity Message (Hamilton, Hlillter, & Burgooll, 1990, p. 244) The law~ regulating the sale of heroin in this coulllry ha\'e sometimes done more harm than good. both (0 society and to the individual who mmt use heroin. The public is faced with a large number of crimes comm itted each year in most major cities by addicts searching for money to support their habit. The addict suffer~ not from heroin. but from unpleasan t secondary complications which are promoted by the drug's continued illegalit y. In England. where th e governme nt controls the legal sale to addicts, heroin-related crimes are almost non-existent. Crime i~ not caused by the drug itself. but by slightly outdated laws which prohibit its use. In the United States, addicts are forced to commit
163
164
Part II • Per..\" 011 Sources, Heceivers,
crime~
C/UlllfleiS,
and Messages
against innocent citiLcns to obtain money to pay high black market prices charged by thcir underworld suppliers. As a rcsult of thesc somewhat expensive transactions. all enforcement agencies are occasionally tempted by graft. Some heroin addicts die needles~ly from di!o,ease caused not by the drug, but from uncomfortable secondary complications. Medical authorities now tentatively agree that heroin causes very little physical damage. Symptoms of heroin withdrawal are not as dangerous as those associated with alcohol. Yet in New York City last year over 900 addicts died from tetanus and hepatitis caused by improper means of injection. Addicts from time to time re-use and share unsanitary needles. or improvise with objects not designcd for injecting drugs into the bloodstream, because hypodermic syringes are not legally available.
10 Influential Actions Nonverbal Communication and Persuasion Peter A. Andersen
... Ih~ effect ... of nOIl\ cfhal heha,iop" and \'ariou ... verbal me ....,agc variables on
pcr,ua'ioll and compliance \\cre compared. TIll'. comparison revealed that nonvcrb'll behavior... afe a~ powerful. in ~OI11C Ci.l'CS I11ml.! powerful. than some of the mes ... agc ... tratcgic, Ihal have bcc.:11 . . wdied ill proolH.: ing compliance frolll others. Scgrin. 1993. p. 1M3
. . . Ilollverbal cOlllllluniG.tlion i ... a... important " .... perhap~ more important than. verbal communication in pcr,u'luing uther... 10 ch'.II1gl! their :.mitude-. and bt:havior. Andcr~cn. 1999. p. 273
Pcrhap ... the mo ... ! common human enlerpri . . c is influencing mher people. People arc in\ohed in thousand ... of pero;,uu\ion attempt . . cach \\cd.. from the mundane acts of getting your roommate to turn down the ... terea or pcp.. uading your partner to arrive on time to Illore impol1ant is ... ucs ... uch as getting your partner to marry you or persuading a troubled friend 10 seek. counseling. When people think of per... ua ... ion they think. of talk: but there i~ more to persua!o.ion than word .... The un ... poken, unwrillen messagc!o. WI.! .... end and receive have as much to do with the ~ucce\s of our influence attempts as the words we uller. Like Illost people, scho lars have typically thought of persuasion as a verhal activity. From the rhetoric of Aristot le to <.:ontemporary social scienlitic studies. the va~t majority of research. inc luding most of the chapter ... in thi ... hook. ha ... focu . . ed on verbal communication. Verbal communication is what we "no\\ the most ~lbout. Thi ... chapter is an exception . for it will focus on the multiple ways in which nonverbal communication influences altitudes and changes behavior.
165
J66
Part II • P(' nIIlU;OIl Var;ahlcJ: Pl' r.' !wuil·(,\
Thi ... chapter will review the communication theories.
CffCCh
(II/
SOllrces. Receil'ers, C!u/fUlels. and Meuages
of nonverbal communication in the context of several
The Direct Effects Model of Immedincy Nmll 'l' r/Ja/ immediacy belul\'iors are nonverbal acts that simultaneously signal warmth.
dccrca ... c psychological or physical distance between communicators, arc interpersonally ... timulating, and signal availability for communication (Andersen, 1985). Behaviors like eye contact. touch, and close distances arc prototypical example ... of nonverbal immediacy behavior .... Overwhelmingly, per... U<.I... ion re ...earch ,\upports the Direct Effects Model of 1111mcdia<.:y. which suggests that warm, involving. immediate nonverbal behaviors signifi-
cantly enhance the persuasi\e effect; of a me "age (Andersen. 1985. 1999; Segrin. 1993). Ot1l.cn ... of ... tudies suppon the model. The Direct Effects Model of nonverhal pcr",ua ... ion ha", al<.,o been called the Social Me:1llJng Model (Burgoon. Coker. & Coker. 1986: Burgoon. Manu ... ov. Minco. & Hale. 19K)). BOIh l110dcb contcnd that warm. involving. immediate behavior results in incn.:a ...cll compliance on the part of the receiver. Some ~cholars sugge ... t that we may have
an inborn predisposition to comply with ,,"neone we like (Cialdini, 1984). Complying with tho",c whom we like may be the underlying basi~ of human cooperation that has cvoh-ed throughollt the millennia. However. as Cialdini (1984) notes: Fc\\ people would be ... urpri,ed to learn thaI. a ... a nile, we most prefer 10 ",a) yC!o. to re10 note, howcver. is that thi ... ,illlpic rule is used in humJred ... of way' by lotal ... trangers to get U!o. 10 comply with 1/I£' ir
4w:' .... of !-.omeone we !...now and Ii!.....:. Whal might be 'lartling
reque,lS. (p.163) The fundamental impulse 10 tl'u!>.t .,miling actor", on advcrtisemcm ... : friendly ",olicitors for charitiCS;
Eye Behavior Studies of eye behavior have provided sub~tantial support for the Direct Effects Model of Nonverbal Immediacy. particularly the per",uasive effects of gaze (looking at another per... on) and eye contact (mutual ga7e into one another's eyes). In field ~tudie~ research has ~hown that unacquainted per~uaders are more effective
if the) usc eye contact. Kleinke and Singer (1979) found that both male and female sub-
Chapter 10 • Influential Actions
167
jeets took significantly more leutlets from campaigners who held their gaze than from those who did not, and the effect was stronger in the absence of any accompanying verbal
communication. In a study by Bull and Gibson-Robinson (1981), poorly dressed solicitors for charity dramatically increased their effectiveness in obtaining contributions when
using direct gaze. Additional field research by Valentine (1980) reported that bystanders were more likely to assist a disabled victim with a broken arm if the victim looked at them directly. This finding held true when the victim was accompanied by a friend as well as when the victim was alone. One study found Lhat hitchhikers were more likely to obtain
rides when they used direct eye contact than when they failed to use eye contact (Snyder, Grether, & Keller, 1974). Laboratory studies likewise confirm the Direct Effects Model for gaze, Burgoon and
colleagues (1986) provided additional empirical support for the Direct Effects Model (or Social Meaning Model) when they found that a person was judged more likely to be hired for ajob when gazing than when not gazing. Specifically. they reported that gaze aversion carried very negative meanings and was very unpersuasive, whereas gaze was highly effective in interpersonal persuasion. G,:l7e l11i.ly be effective in gaining compliance because it is simultaneously powerful
and immediate (Andersen. 1985; Segrin. 1990). Linkey and Firestone (1990), for example, examined a group disclission task and found that influence was primarily a function of the visual dominance ratio (the degree to which one person looked at hislher partner divided
by the degree to which the partner looked at him/her). Research by Liss, Walker, Ha7elton. and Cupach (1993) showed that compliance correlated strongly with mutual gaze. sugge~ling that eye contact by a persuader is a potent predictor of compliance, In a study of televised debates in Denmark, all of the eleven speakers who gazed intensely at
their audience won their debates and lost none (Jorgensen, Kock, & Rorbeck, 1998) suggesting that the persuasive effects of eye gaze may be cross-cultural. In a statistical summary of research called a meta-analysis, Segrin (1993) combined the result:, of 49 nonverbal studier-.. including 12 on gaze behavior. Gaze produced greater campi iance than averted gaze in everyone of the 12 studies. Though the persuasive effects of gaze were not huge, they were quite consistent: Gaze and eye contact increase compliance. This entire body of research. in study after study, shows that gaze, dominant gaze, and mutual gaze all seem to have persuasive effects.
Touch alld Persuasioll Touch. like eye behavior, is generally perceived as a warm, friendly behavior except in situations where the touch is hostile or there is a preexisting negative relationship between the interactants. A large number of studies indicate that touch, even by a stranger, has positive effects on persuasion.
One study (Kleinke, 1977) tested whether sLrangers could be persuaded to return change left in an airport phone booth. When strangers who found the change were asked to return the change with a gentle tOllch. 96 percent of them complied. When no touch ac-
companied the request. only 63 percent complied. Similarly, in two field studies of compliance behavior. Willis and Hamm (1980) had experimenters ask strangers to comply by signing a petition or filling out a questionnaire. In both experiments half the strangers
168
Pan II • Per,HIlHioli Variahfe~: Penpec lil'('l 011 Source.", R('n'il'l'ro;. Chlllllle/\,
(III(/
Me.\WI/.:c.\
received touch. and half received no touch while other communication cue ... remained constant acros~ the two groups. In the first experiment ~trangcr~ signed the petition 81 percent of the time when touched but only 50 percent of the time when not touched. In the second experiment .... trangers completed the queMionnaire 70 percent of the time when touched but ani) 40 percent of the time when not touched. Re\uit\ of both experiment>, ,hawed significant. po ... itive effect ... of touch on compliance. Studie~ of service encounters have ~hown thill waitres ...es· touch increase~ compliance behavior. In two studies. when waitres~es touched Lhc hand or arm of customers. they received bigger tips (Cru,"" & Wetzel. 1984: Stephen & Zweigenhaft. 1985). A recent ~tudy found that when waitresses asked bar patrons if they wanted somcthing to drink. their touch resulted in significantly more alcohol consumption than when thcy did not touch the patron (Kaufman & Mahoney. 1999). One comlllon persutl ... i\,e ... trategy is called the fOOl -in-the-door technique. where a small request is followed by a larger request (see chapter 12 of this lext and Gass & Seiter. 2003 for a good summary). Several studies have showll that touch increases the fool-in~ the-door effect. An invcMigation into the effects of tOllch on volunteering for charity. employing 3 foot~in-the-door appeal, showed that touch increased compliance in most experilllcntal conditions (Goldman. Kiyohara. & Pfannen:-teil, 1985) and was effective even following an initially negative communication. Patterson. Powell. and Lenihan (1986) employed touch in an experiment in which students were asked to help ... core ex~ ams. After Ilcoring some exams, "'Iudents were asked to May and score additional lest .... Students who were touched !'.tf.lyed longer to score the exams than ~tudents who were not touched. The authors attributed the increased compliance to perceptions of greater liking and/or status of the experimenter who did the touching. In a meta-analysis of 13 studies examining the influence of touch on compliance, Segrin (1993) reported that touch showed a positive. consistent effect on compliance across the 13 studies. In virtually all these studies experimenters touched a !'.tranger on the arm while making a requc ... t und compared the response to that received when there was no touching. Linle i... known about the effects of per... uasive touch in close relation",hips. nor have we examined the effects of more intimate types of touch-such as ... trakes. caresses. or squeeles-or more intimate targets of touch such as the chest or face. Anecdotal evidence suggests that compliancc requests in intimate relationships are often accompanied by a touch.
Killesic Behavior alld Persuasioll Kinesics is the study of comlllunication via body movements. One kinesic behavior. the open body position. has been associated with greuter immediacy and approachability (Andersen. 1985). Mehrabian (1969) reported that open arm and leg po,ition, create positive attitudes in receivers. Morris (1977) has shown that kinesic '"barrier signals" communicate defensiveness and avoidance. the opposite of the attitudes indicated by open body postures. In a study of opinion change. McGinley. LeFevre. and McGinley (1975) reponed that open body positions result in more persuasion than when communicators keep
Chapter I () • I"fluef/tial Actio1lS
169
Ihcir knees and feel logelher. arms folded and held clo,e 10 Ihe body. The "udy of 37 Dani ... h televised political debates discw.. '1ed earlier (Jorgensen et al.. 1998) found that debaters with more open body pol.itures did ,ignilicantly better than those with closed body pm,tures. The researchers also examined what they called "a di~mil)sive attitude," conveyed by closed posture and an unfriendly facial expression. Five of the six debaters displaying this dismil.il.iive attitude lost. The smile is particularly persua'-,ivc. Receivers of communication message!.. are disarmed by a smiling person and more likely to comply wilh his or her request. Burgoon, Birk, and Pfall (1990) examined the impact of several kinesic behaviors on persuasion and found that facial pleasantness was mOSI predictive of persuasive success. Liss and colleagues (1993) examined the effect of o;,miling on compliance gaining and found that more smiling resulted in greater compliance. Part of persuasive immediacy is bodily animation. Burgoon and colleagues (1990) also report that more overall bodily mO\ement and animation correlated with persuasiveness. Similarly in the Danish debate study. winning debaters showed an animated and energetic kinesic style by a ratio of I I to I and gesturally animated debaters won by a ratio of 12 to 2 (Jorgensen et al.. 199R). Together these studies suggest that persuasion is facilitated by kinesic uClivity that includes open body position .... positive facial affect. and kinesic animation.
Vocal Cues alld Persuasioll SlUdic~ tend to ~how a positive link between vocal immediacy and persuasion. A series of sludies by Buller and his associate~ ~uggeSl that vocal immediacy cues. including a pleasant tone of voice and a fast delivery. are linked to greater compliance. Interestingly, these effects are particularly true for individuals who are skilled decoders of nonverbal communicalion (Buller & Aune. 1988. 1992: Buller & Burgoon. 1986: Buller. LePoire. Aune, & Eloy. 1992). Segrin', (1993) "ali"ical summary of Ihe research showed Ihal bOlh vocal rate and vocal pleasantne~s were associated with persua!-tion. A slightly different finding was reported by Burgoon and colleagues (1990). who found that persuasiveness increased with greater vocal nuency and pitch variety but not with greater vocal pleasantness. In Ihe Danish debale "udies Jorgensen and colleagues (1998) found Ihal speakers with modulated voices (e.g .. greater pitch variation) were more likely to win than speakers with monotonous voices. Likewise. vocally energetic !o.peakers were often debate winners and inarticulate debaters were often losers. Other slUdie~ generally support the ...erics of !\tudies by Buller and his associates indicating that faster rates of vOl:al delivery arc a ... sociatcd with greater persuasiveness (Apple. Slreeler. & Krauss. 1979: MacLachlan. 1979: Miller. Maruyuma. Beaber. & Malonc. 1976). These studies suggested that faster delivery is associated with competence and confidence and is unlikely to be used by an insecure or deceptive person. While morc research on vocalic inlluences on persuasion needs to be conducted. current research suggests that a fast. pleasant. vocally varied nonverbal communication style will make verbal communication more per!-tuasive. This may be particularly true on the telephone. where vocalic information is the only available nonverbal cue.
170
Pan II • Persuasion Variables: Perspectil'es on Source.!'., Rl'(,l'iI'l'r\,
C/UII/Ill'/,\,
al/{I Ml' \'\'aKes
Multidimensional Nonverbal Clles and PerSllasioll So far the discussion has focused on the per~uasivc impact of ~ingle nonverbal cues like touching or smiling. But research ha~ found that nonverbal immediacy is usually communicated though multidimensional di!>.plays that include ~cveral nonverbal immediacy cues simultaneously. Some nonverbal persuasion studie\ have examined the impact of \everal nonverbal cues in combination, The most cOl11mon experimental manipulation ha\ examined the simultaneous persuasive effects of touch and gale.
Touch·Gaze Combinatiolls alld Persuasion. A number of studies have examined the persuasive impact of both touch and gaze in combination, This approach permits an examination of whether one of these cues can sub!>.titute for the other. their relative persuasive impact. and whether they have additive or cUlllulative effect!>,. Kleink.c (1977) reported two such experiments that examined the effect of gaL.e and touch on compliance. In the first study, the experimenter left a dime in a phone booth. and when a !>!ubject found the dime the experimenter requested it back, The experimenter manipulated gale and touch to generate four experimental combinations: (I) louch and gaze. (2) louch alone. (3) gaze alone. and (4) neither touch nor gaze. Combined touch and gaze resulted in the greatest number of dimes being returned, indicating that lOut.:h and gaze have additive effect{" on compliance. In the second ~tudy. experimenters asked ... trangers in a !>'hopping mall if they would lend them a dime. As with the fin,t experimcnt. four combination ... of touch and gaze were employed and results again showed Ihat combined touch and gale produced increased compliance. The researcher's explanation b thm touch and gaze increase both attention and involvement. making noncompliance morc difficult. In another ~lnic1e. Kleinke (1980) reported two experiments, which replicated the 1977 "udy while also adding a legitimate request (dime for a phone call) and an illegitimate reque~1 (dime for a candy bar). Results replicated the 1977 study for the legitimate rcque ... t. blll for the illegitimate requests greater compliance was obtained through ahsence of tollch or eye contact. since the experimenter may have appeared more tactful or humble. A rcplil:ation of Kleinke's (1977) phone booth experiment by Brockner. Pre"man. Cabill. and Moran (1982) showed that both touch and eye contact independently increa ... ed compliance. The data from these five studies show cumulative effects of both touch and eye contact consistent with the Direct Effects, or Social Meaning. models. Tilese findings have substantial practical importance for persua~ioll, When trying to promote positive health hchaviors. soliciting for charity. or getting assistance from a stranger. the combination of tollch and gaze considerably increases the chances of complianl:c with one's request. Persuasioll and Other Muitidimensiollal l ml1lediacy Behaviors.
Studie ... of the persuasive impact of multidimensional nonverbal immediacy cues ~how a pattern !>!imilar to single cues and combinations of gale and touch: immediacy produces pcrsllilsion. One study of classroom compliance employing multiple indices of immediacy found that students were more likely to engage in communication practices ~uggestcd by immediale rather than nonimmediate teachers (J. Andersen. 1979). Burgoon. Birk. and Pl'au (1990) reported a similar finding in a study of persuasive spe.akcr~. Specifically. the) found that increased persuasive effects were associated with \ot.:alic beha\ior ... (\ouch a~ longer
Chapler J 0 •
'nj1I1l!lIfial ACliOllS
171
pauses, vocal pleasantness, and pitch variety). increased eye gaze. more smiling. greater facial expressiveness, and more overall movement. Recently. research has also ~ho\Vn that individuals use nonverbal communication to resist persuasion. In a study of rejection strategies for flirtatious advances. Trost and Engstrom ( 1994) reponed that rejecters avoid nonverbal contact. ignore the persuader, maintain larger personal space, act cold and unin terested, display alternative relational ties (e.g .. engagement rings). and act nervou s and uneasy.
Multichanlleled, Mediated, Persuasive Cues.
Many of the persuasive messages we receive come through mediated channels such as radio. television. and the Internet. Immediacy cues also seem to produce persua sive effects in mediated as well as face-lo-face communication. Studies of television newscasters, for example. have shown that nonverbal behaviors influence viewers in a variety of ways. We treat newsca~ters as virtual acquaintances who are nightly guests in our homes. For many years Walter Cronkite. the anchorman for CBS news, was considered the most credible man in America. Studies show that the nonverbal behavior of television newscasters affects voting preferences. Friedman. Mertz. and DiMatteo ( 1980) reponed that newscaste rs' facial expressions were consistently biased toward certain candidates at levels unlikely to occur by chance. In the 1976 election Walter Cronkite. David Brinkley, and Harry Reasoner showed more favorable facial expressions when reporting about Carter than about Ford. John Chancellor and Barbara Walters showed more facial positivity toward Ford than toward Carter. Verbal content showed no corresponding bias. Two studies replicated and extended these findings to the 1984 elections and showed that biased facial expression~ were associated with voting behavior of viewers (Mullen. 1986). In the first study, these researchers found that Dan Rather of CBS and Tom Brokaw of NBC showed no facial bias during stories about Reagan or Mondale. However. Peter Jennings of ABC exhibited strong facial bias toward Reagan over Mondale. The second ~tudy examined voters in four cities in Ohio. Misso uri. Massac hu sett s, and Penn sy lvania and showed that in eac h city viewers of ABC had significantly more favorable attitudes toward Reagan than viewers of NBC or CBS. While pro-Reagan viewers might have tuned in more to ABC. Mullen and colleagues believed that their study actually de monstrated a subtle. peripheral route to persuasion. occulTing without deliberate. consciolls consideration of arguments. In e ither case this study suggested a strong association among nonverbal behavior. televis.ion viewing. and electoral decisions. Recent studies of the Internet have suggested that perce ived proxemic cues affect compliance. Moon (1999) reports thai one proxemic variable, the perceived distance of one's partner during computer-mediated communication. correlated with the amount of persuasion. People who believed that they were communicating with someo ne several miles away complied more frequently than if they perceived that the person was. thousands of miles away.
Appearal/ce and Credibility: The Halo Effect People use simple. obvious, nonverbal cues as shorthand indicators of statu s and credibil ity. While evidence has shown that good looks or ni ce clothes are not an indication of
J 72
PUrl II • Per:ma.sioll Variahles: PerspeCli\'l!\
()/I
SO/lrces. Receil·en. CluJ/lI/eis. (llId Me .\,\(/gl'.\
greate r co mpeten ce or credi bili ty (Feingold. 1992). a large body of studies ha!o. "ugge~tcd that we belie\'e that well-d ressed, good looking people are s marter. warmer. more hone"t and the refore more deserving of compli ancc than less well dressed. unattractive individuals. This is called the "halo effec t," whereby one positive quality in a person call~e~ us to assume that the indi vidual has many positive qualities. Our clothing does more th an protect and conceal our bodies; it communicate" our status and credibil ity. Studies have show n that apparel is not an immediacy behavior like touch. gaze. and s milin g (Andersen. Andersen. & Jensen. 1979 ). Instead. rc~earch suggcsts that clothing is a statu s or cred ibilit y cue llmt has powerful persuasive effec t ~ (~ee Andersen. 1999). People are more likely to comp ly with rc~pcc tablc and conventional pers uader" than with those who appear to be weird and unreliable. Studies reveal that a conventional attire or appearance has more positive per~uasivc effects than does an uncon ven tional appearance. In general. "dressing up" is recommcnded for most persuasive situation s: people arc more likely to comply with hi gh-~ latu s people than with low-statu s ones. In one ~ tud y (Bic kman. 197 1). a stimulus person left change in a public phone booth and waited until the next caller found it. When the ~ timulu~ per~on wa~ well dressed. 77 percent o f the ~ ubj ec t s admitting finding the money and ret urned the change: when the stimulu s per~on was poorly dressed, only 38 percenl of caliers relUmed Ihe change. Similarly, Raymond and Unger ( 1972) found that pa~~ersby were mo re willing to make chan ge for cOl1\cntionally dressed indi vidual s than for unconventionally dressed ones. Likew ise. Kl einke (1977) found Ihal more people " Ienl" a dime 10 a neall y ralher Ihan a sloppil y drcs;ed expe rime nter. In a petition-signing study by MacNeill and Wil son (1972). experimemcrs obtained more signat ures when dressed conventionally (su it and tie with short hair) than unco nv entionally (faded army jacket with lon g hair) . Likewi se. Keasey and Tomlin son-Keascy (1973) found that conventionally dressed male and female petitioners were able to obtain more signatures on petition~ against the U.S. inva ~ ion of Cambodia than "hippic" pcti tioners. In a similar study , Darley and Cooper (1972) examined the impact of countc rcu ltural or "hippie" appearance on political campa ign e ffecti ve ness. They found that voters we re less likel y to take a leaflct from a " hippi e" than from a co nventio nall y drc~scd campaig ner-H nd were more likely to throw it away after having taken it. Moreover. voters allribuled more radical. less acccpwbic views 10 candidalcs supponed by hippi es Ihan to those supported by co nven tion all y drc~"ied campaigners. Similarly. people were found to be more likel y to donate money to charit y to a per~on w ith short hair dre~~cd in a jacket and tie than to a perso n with long hair dressed in jean~ and sandal s (Chaikin. Oerlega. Yoder. & Phillips. 1974). One sludy found Ihal passersby were more likel y 10 an"ver a survey administered by a co nven tionall y dressed interviewer than by an unconventionally dressed one (Walker. Harriman. & COSielio. 1980). However. allnlher "udy Ihmlooked "I the e ffect s of five types of womcn's clothing on willingness to complete a qu es tionnaire revealed no significant differen ces between women dressed in a formal ~kirt. formal pants. casual skin , casual pal1ls, and jeans (Harri s el al.. 1983). Clothing also signifies authority. Ho ~pi tal doctors. law enforcement officer~. and so ldie rs wear uniforms to instantaneously comm uni cate their authority: and people are exceedingly likel y to comply with suc h authority figures. For example. medical profes<.;ion-
Chapler 10 • IIlj1U('Iltill/ Actions
173
als, especially in hospitals, typically wear a white coat or dress to communicate authority. This enables them to make us wait, go to designated rooms, take off our clothes, and to submit to embarrassing or painful medical procedures. In Milgram's (1974) classic experiments, in which he successfully induced participants to provide shocks (that they believed were real) to students in laboratory learning studies, compliance was gained by an experimenter wearing a gray technician's coat over a white shirt and tie. Bushman (1988) conducted a study in which a female randomly stopped pedeMrians and said. "This fellow is overparked at the meter and doesn't have any change. Give him a nickel." This female wore different clothing in three experimental conditions. When dressed in 11 uniform. she gained more compliance than when in business auire or casual dress. The nondescript blue uniform produced compliance 72 percent of the time, whereas only 48 percent compliance was ga ined in the business auire condit ion and 52 percent in the casual attire condition. Similarly, well-dressed men receive better and quicker service in department stores than less well-dressed men (Stead & Zinkhan. 1986). One study found that salespeople were more likely to comply with complai nt -based requests in department stores from a welldressed person th an from a less well dressed one (Krapfe!. 1988). In his statistical summary of 19 studies, Segrin (1993) concluded that the more formal or high status the clothing, the greater the compliance rate obtained. Evidcntly. we are more like ly to permit appropriately and well-dressed individuals to approach us and to gain our tru ~ t than we are to allow inappropriately or less conventionally dre%ed individuals to do so. Moreover. Segrin found that the greater the status manipUlation in physical appearance studi es, the greater the magnitude of the effect. Clearly, stalU~ pcr'-,uades; and clothing acts as a highly salient !'.urrogate for a person's Matus.
Expectancy Violations Theory Expectancy Violations Theory emerged as an alternative explanation for thc persuasive impact of immed iacy behavior. particularly proxemic behavior (Burgoon & Jones, 1976). This theory claims that each pers-on has cultural and personal expectations about the normal distances people maintain during interaction. Attractive or rewarding individuals are more persuasive if they stand c loser or farther than th e nOl'm. whereas unattractive or unrewardin g individuals will be more persuasive if they maintain normal distances. Highly rewardin g individuals who vio late norms draw attention to their positive characteristics and enhance their personal persuasiveness. perhaps through the halo effcct'-, discussed previously. Several stud ies of proxemic behavior have provided support for the theory. Stacks and Burgoon (1979) found that rewarding communicators were more persuasive at either closer or farther distances than at nonnative distances. As predicted by the theory, unrewarding individuals produced no difference in persuasiveness at close, normal. or far distan ces. Two studies by Burgoon and Aho (1982) provided additional support for this theory. In both studie s. distance had no significant cffcct on compliance for either lowreward or high-reward communicators. though reward itself produced significant effects. Alben and Dabbs ( 1970) reported that actual persuasion was greater m far distances (14 to 15 feet) than at either close or lIIoderate distances (I to 5 feet). Buller (1987) conducted a
174
Part II • Persllasiml Variables: Perspectil'eson Sources, Receil'er,\. Challnel\, lIml Messages
study in which cxperimcntcrs as~umcd closc. moderatc. or far distanccs while attempting to get citizens to sign petitions. Findings showed that close distances resulled in the greatest compliance. while moderate and far distances resulted in lower levels of compliance. Neither Albert and Dabbs nor Buller manipulated the reward value. but their results suggest that non-nomlative distance can increase compliance. In an extension of Expectancy Violations Theory to voca lic communication. Buller and Burgoon (1986) showed that pleasant voices (a positive norm violation) produced more compliance. but only for good decoders. Poor decoders complied more with hostile voices (a negative norm violation) than with neutral or pleasant ones. While reward value was not successfully manipulated, the study provided some support for the expectancy violation model. Expectancy Violations Theory was expanded to eye behavior in two studie\ (Burgoon el al.. 1985. 1986). BOlh sludie, employed eilher rewarding (well-qualified job interviewees) confederates or nonrewarding (unqualified job interviewees) confedenHc ... and differing levels of eye gaze. Resuhs of Ihese studies provided lillie support for Ihe Expectancy Violations Theory, since the primary results showed direct persuasive effects for increased levels of gaze regardless of reward value. These results supported the Social Meaning or DirecI Effecls Model described previously. Expeclancy Violalions Theory has received some genem l confirmalion, bUllhe support for its persuasive effects is mixed. Moreover. since few of the studies that have tested Expectancy Violations Theory have actually co nfirmed whether expectations were violaled or nOI. a cenlrallenel of Ihe Iheory remains unlesled (Segrin. 1990). Proxemic behavior is best explained by Expectancy Violations Theory. which suggests that a rewarding individual .,hould probably stand closer or fanher than the normal interaction distance for maximal persua"ive effects. Standing at a "normal" distance may be t1l0~t persuasive for unrewarding communicators.
Distraction Models Distraction models share some similarity with Expec tancy Violations Theory. These models suggest that if a source's nonverbal behavior distracts. the receiver is more susceptible 10 persuasion. Slacks and Burgoon (1979) predicled Ihal dislance violalions would produce more persuasion because of their distracting propertic .... Unfonunately. they found that distance violations had no effect on a self-reported distraction mea:\ure. However. close and far distance violations were more persuasive than a normal distance. Why did distance violations produce more persuasion? Perhaps distraction worked even though subjects were unaware of the distraction and were therefore unable to self-report it. Distance violations may also produce arousal effects. Another ... tudy by Stacks and Burgoon (1981) reported a significanl effecI of bOlh dislance violalions and exlremes of physical attraction on distraction. Small persuasive effects were found for physical attraction and none for distance violations. thus providing better support for distraction models. In a test of Ihe dislracling effecls of rapidly spoken messages on per;uasion. Woodall and Burgoon (1983) showed that fast messages were more distracting than mes~age" delivered at a nomal pace. but no persuasive effects of faster messages were found.
Ch;}ptcr 10 • Inp/lemilll ACliom
175
Bulle r ( 1986) conduc ted a meta-analysis of 38 'itudies that examined the diMractionpersuasion re lationship. His \tudy indi cated that commu ni cation-irrelevant factors ~ u c h as noise or visual di stractions ge nerall y reduced persuasion and attitude change. because they impeded comprehension of the mes\ages. However. communicati on-relevant distractions. such as highl y attractive or credible sources. c3 u\ecl receivers to focu\ on these positi ve qualities and prod uced a positive effec t on persuasion. If the source had low attrac ti veness or credibility, foc using on the ~peaker red uced attitude change because these negative di stracti ons impeded believability and persuasion. In genera l. positively regarded source!o. may benefit by focusing the recei\er'~ ~Hlention on that source's po~itive charac teristi cs. a finding consistent with Expectancy Violations Theory , POllY and Cacioppo's (1986) work on persuasion and cog niti on may also help expla in the process of di\traction. Basically they have shown that distraction aids the persuasion process if the me\sage is of low quality. In thi s ci rc um stance a distracting appearancc or proxemic be havior. fo r example. would di stnlc t a per:;on so that the receiver might not be ab le 10 conccn trate on the many naws in the message. Hi gh-qualit y persuasive arguments. on the other hand. wou ld be less effective when a receiver was distmcted by con..,picuous nonve rbal cue~. for much the sa me reaso n. The receiver ca nnot concen trate on the hi gh-q ua lit y mes~age. so its effecti ve ness is reduced.
Communication Accommodatioll Theory Communication Accommodation Theory (CAT) deals primarily with the voca li c or paralinguistic effec ts o f nonverbal co mmunicat ion. It posits thut li..,tener~ perceive ~pcech similar to their own as more attract ive. pleasant. inte ll igible. and per,;"uasive than unfamiliar speech (S treet. 1982: Strect & Brady, 1982: Street & Giles, 1982), Furthermore. speakers typically adjust or accommodate their ~pecch (Q the ~ t yle or rate of th e other interactant. even th ough m o~ t ~peaker~ are unaware of thi~ accommodation (Street, 1982). Based on thi s theory. speake rs who adjust to the comm uni cat ion of their liste ners should be morc persuasive. Several recen t studi es pro vide .., upport for the pe rsuas ive effects of comm unication accommodation. Buller and Aune (1988) reported that good decoders were most likely to comply with faster message~ . Thi~ finding did not hold true fo r poor decoders, Why did good decoders prefer fast mes>"ge,'J The authors "bo found that good decoders spoke fa~ter. so faster ~pca k er~ see med to prefer to listen 10 fas ter speech. a position consistent with CAT, A seco nd slUdy by Buller and Aune ( 1989) partially repli cated the first study. Again. good decoders spoke faster th an poor decoders, though no effect was found for ac tual or perceived similarity of speech rate on compliance. However. poor decoders. who generally speak slow ly. complied most with the moderately slow voice. whereas good decoders (who gene rally speak fast) complied most with moderately fast and ve ry fa>t voices, These findings suggest that people are inl1uenceo by those who speak at the same rate as th ey do. Another study support ive of CAT, by Burgoo n and colleagues ( 1987), exa mined the effec ts of a number of comlllunication variables on patients' compliance with their physician's advice. Among the many cOl1llllun icmion \ariable~ tested. o nl y perceived similarit y corrclutcd signifi can tl y with co mpli ance. a fi nd ing co nsistcn t wi th CAT. A final
176
Pari II • Per.\'IIClS;01I Variables: Per.<.pecti\'es on Sources. Receil'(' rs. Challnels. and Message:, study by Buller and Burgoon (1986) showed that good nonverbal decoders complied more with pleasant voices than with neutral or hostile voices, whereas poor decoders complied most in the hostile condition, moderately in the neutral condition, and least in the pleasant condition. Since good decoders are more affiliative. composed, and social than poor decoders. Buller and Burgoon suggeMed CAT as the explanation. People are more likely to comply with tones of voice that are similar to their own. Several studies not based on CAT show that rapid speech enhances persuasion (Apple et al.. 1979: Buller & Aune, 1988; MacLachlan. 1979; Miller el aI., 1976). Rapid speech may enhance persuasion by increasing the effort it takes to perceive the message or by improving the perceived competence of the source. One study, however (Woodall & Burgoon, 1983). found that fast rates had the same persuasive effect as slow rates. One explanation for these results is the region in which the studies were conducted. As noted previously, both Ihe Apple el al. study and Ihe MacLachlan studies were conducted in New York, whereas the studies by both Buller and Aune and Miller and colleagues employed subjects from urban Southern California and Arizona, where fast speech is the norm. The Woodall and Burgoon study. in contrast, was conducted in the Southeast. where speech is considerably slower. Thus. it may be that slow speakers are more persuasive in the South whereas fast speakers are more persuasive in the North and Pacific West. CAT. as applied 10 vocalic behaviors and persuasion, has received suppon. Communicators would probably be well advised to use vocalic cues similar to their persuasive targets to m3ximiL.:e compliance. Speech accommodation is a complex dyadic process in which both interactants adapt to the other's speaking style. Since no study of this dyadic type of social interaction has yet been conducted, the theory remains untested in its most complete form (Segrin. 1990). CAT could be applied to other areas of nonverbal communication. However, substantial support for the Direct Effects Model, according to which increased gaze. faster speech, or more touch is persuasive. presents a problem for CAT. Unless most receivers of persuasive messages were high gazers, fast speakers, and frequent lOuchers- a condition untrue by definition-the Direct Effects Model would be more predictive than CAT. The likelihood that more immediacy is more important than morc accommodation also supports the Direct Effects Model rather than CAT. However, a combination of immediacy and accommodation may be highly effective. given that both forms of communication seek to enhance persuasion in prior research.
Elaboration Likelihood Model In an attempt to create a comprehensive model of persuasion that subsumed all prior findings on persuasion, Pelly and Cacioppo (1986) crealed the Elaboration Likelihood Model (ELM). As noted in previous chaplers, the cenlral concepl of the ELM is that two types of persuasion exist. The first type of persuasion results from the careful and thoughtful consideration of the merits of the information presented in a message and is called the central rowe. The second type of persuasion results from cues in a persuasive context. such as an attractive source, without necessitating scrutiny of the actual merits of the information; Ihis is called the peripheral rOLlte. Petty and Cacioppo (1986) assened thai persuasion pro-
l
Chapter 10 • Influential Actions
177
cessed through the central route appears to be more enduring than that which is processed "peripherally. " The ELM has becn widely criticized in communication for weak and circular definitions of key components of the model and insufficient a priori specification of whether cues will be processed centrally or peripherally (see Stiff & Boster, 1987). Perhaps even more important. the ELM privileges the central route of verbal and logical persuasion and, by relegating most nonverbal communication to the peripheral route. fails to recognize bOLh the power of nonverbal communication and the true intent of the persuasive message. First. the very terms cemral and peripheral suggest that one type of communication (i.e .. logical. thoughtful. verbal) has more va lidity than the other (intuitive. instinctual. nonverbal). Evidence suggests. however, that intuition abollt the character. expenise, or competence of a source is an equally valid type of persuasive assessmenl. Since Aristotle first introduced ethos as a central concept in rhclOric and pcrsuasion. peripheral me~sages such as source characterislics have bcen considered a valid and rapid means of assessing the merits of an argumenl. Detection of deception and subtle incongrucnt cues may be more important than discerning the syllogistic and cvidentiary structure of the arguments in human interaction. Indeed. the abundant literature on deception detection suggests that deception is often detected from subtle, incongruent cues. Second. the available summaries of literature (see Andersen. 1999; Segrin. 1993) suggest that nonverbal communication is at least as powerful as the traditional central route message variables such as evidence that has relatively wcak effects. Third. central processing, the preferred or favored route according to the ELM. may still entail numerous errors of logic and reasoning (Kahneman, Slavic. & Tverksy. 1982). Indeed, the primary purpose of so-called logical communication in human beings may be more to rationalize and harmonize competing cognitions (Cialdini, 1984; Stacks & Andersen. 1989) than to search for truth with a capital T. Persuasion is as like ly to be abollt harmonization of the relations among people and between disparate connections in the human brain as it is to be about the cold logic of'·truth." Finally, in their initial formulation Petty and Cacioppo insisted thm the central and peripheral routes were separate paths and that receivers of persuasive messages chose between them. Communication researchers have criticized the ELM for this dichotomization of the central and peripheral route (Mongeau & Stiff. 1993). Though the founders of the ELM have recanted on this position (Petty et aI., 1993) the best advice to receivers of a persuasive mc:-,sage is 10 u~c both verbal, logical processes and nonverbal, intuitive processes in making judgments about per~uasive messages. Indeed, research on aspects of the ELM has suggested that so-called central and peripheral cues actually interact in good decision making (see Puckett. Petty. Cacioppo. & Fisher. 1983).
Conclusion Research suggests that nonverbal communication has as much or more persuasive impact than verbal communication and overwhelmingly suppons the Direct Effects Model of Immediacy. More immediate. involving communication produces more persuasive impact.
178
Part II • PenlillSioll Variab/e,\: Penpeclil'(!.\ 011 50/lrc('\. Receil·l'Ys. CiuUllleiJ. (lml Messages
Whether nonverbal immediacy i~ increased in a \ingle channel or in multiple channels, touch. gaze, ~miling. and other nonverbal cues have a po!\itive impact on persuasion.
Rejerellces,_____________________________ Albert. S .. & Dabb .... J. M. (1970). Phy ... ical di ... l a predictor of tc'lI.:hing effcctivene~ .... In D. Nimmo (Ed.), Cm1//lIImicalioll yearbook 3 (pp. 543-559). New Brun ... \\'id.. NJ: Tran ... action Book<.,. Ander~en. J . F .. Andersen. P. A .. & Jcn ... en. A. D. (1979). The mcasurcment of nonverbal immediacy. JOl/mal of Applied COfllfllwlim/iofi Re.H'(Irch. 7. 153- 180, Andersen. P. A. (1985). Nonverbal immediacy in interpcr:.onal communication. In A. W. Seiglllan & S. Feld ... tein (Ed .... ). MJllIidullllleJ /lIlegraliul/s of NOIll'er/)(l1 Belun'io r (pp. 1-29). Hillsdale. NJ: Erlbaulll. AnderloCn. P A (1999). NO/II'abal COllllmmicllI;ol/: FonlH lllld flll/Cliom.. Mountain View. CA: Mayfield. Apple. W .. Streeter. L. A .. & Krau .. ". R M. (1979). Effecls of pitch and ... peech rate on pe ....onal attribution .... Journal of Per.wl/aliry (/11(/ Social Psychologl , 37.715-727. Bickman. L. ( 1971). The effect of social ... tatu ... on the hone ... ty of mhen.. Journal of Social P.~ycllOlogy, 85, 87-92 Brockner. J .. Pressman, B.. Cabill. 1.. & Mor~lIl. P. (1982). Nonverbal intimacy . ..,cx and compl iance: A field "tudy. JOllrnal (~rN(}/lIw/)//1 iJellm'iot". 6.253-258. Bull. R.. & Gihson-Robin ...on. E. (1981). The inlluencc of eye-gaLe .... Iyle of drc ...... and locality on the amount:. of money donated to charity. Humall Rel(l/;Of/J. 34. 895-905. Buller. D. B. (1986). Di ... lraction during pe~u[hive communication: A meta-analytic review. Commullicalioll M(Jllograph~. 53,91-114. Buller. D. B. (1987). CommunicatIOn apprehen ... ion and reaction .. tn proxemic \"iolations. Joumal of NOIII'erlm! Bellm'ior, J I. 13-25. Buller. D. B.. & Aune. R. K. ( 1988). The effect .. of \()Calics and nonverbal sen~itivity on compliance: A ~pcech accommodation theory explanation. Human COllll1lllnicmirm Research. 14.301 -332. Buller. D. B.. & Aune. R. K. (1989. May). Tlte ejfecl.\ of I'om/in and /IOIH'erbul ~nuil;l'ily 011 complialia: Fllrllta leSIS of II/e~pt'ed/ accommodatiol/ t'XpIWWlioll. P
I
Chapter 10 • Influential Actions
J79
Chaikin. A. L.. Dcrlega. V. 1.. Yoder. J .. & Phillip .... D. (1974). The effects of appearance on compliance. jO/lrtllll of Social P!>ycllOlo,'{\. 92. 199-200. Cialdini. R. B. (1984). /1Ij1lfellce.~ci('J/(,(, aud !,metice. Ne .... York: William Morrow. Cru!.co. A. 1-1 .• & Wetzel. C. G. ([984). Thl! Mid,h touch : The effect of inlerpcrsonaltouch on restaurant tipping. Permnllli0' and Social P.\·ycholog\· Bul/etill, 10. 512-517. Dnrley. J. M .. & Cooper. J. (1972). Thc "clcan for genc" phenomenon: The effect of <:.tudent)oo· appearance on political campaigning. jot/rtlal of Applin/ Social Psvc!lOlof:Y. 2. 24-33. Fcingold. A. (1992). Good looking pcoplc ar\! nOl what we think . P.I~\·chologicaI Bulletil!, 111.304-341. Fricdman.lI. S .. Mertz. T. 1.. & DiMaltl!{), M. R. ( 19kO). Pcrceivcd hias in the facial cxprcssion~ oftelc\i ... ion news broadl·a~ten., . jOllmal of COlllllllllliC£ltioll. 3~. 103- 11 I. Ga .. s. R. 1-1 .. & Seiter. J. S. (2003). Per\'lfl/sion ..wcial influence lllld COm(llilll!ce-gaillillg (2nd ed.). Boston: Allyn & Bacon. Goldman. M .. Kiyohara. 0 .. & Pfannen ... tcil. D. A. (19H5). Interpersonal touch, :-ocial labeling. and thc foot -in-Ihc-door effect. The jOllrnal of Social Pn'c!wlogy, 125. 143- 147. lI'lrri .... M. B .. Jame .... 1.. Chave/. 1.. Fuller. M. L .. Kent. S .. Ma:-sanari, Moore. C .. & Walsh. F. (1983). Clothing: Communication. compliance. :md choice. journal of Applied Social P!>ydwlogr. I J. 8897. Jorgensen. C., Kock. C .. & Rorbcck. L. ( 1998). Rhetoric that ... hift ... \otes: An exploratory !.tudy of per~u:l ... jon in i..... uc-orienlcd debate .... Political Commllllicatiofl. J5. 283-299. Kahncman. 0 .. Slo\'ic. P .. & T\'cr'ik y. A. (1982). JudRmellt //IIdn cerrailI0': Heurirtics and Bim·e,,·. Ncw Yorl.: Cambridge Uni\'cr"ity Pre ...s. Kaufman. D .. & Mahoney. J . ~f. (1999). The cffcct of \\(litre ..... touch on alcohol consumption in dyad .... JOIll'llal afSodal P'Iycholo.r.p', /J9.261-267. Ke;N~). C. B.. & Tomlin ... on-Kl:!a ... ey. C. ( 1973). Petition signing in a naturalistic .. elling. jOltl'/lal of Social
c..
P\\,(,hology. 89. .113-31-L
Klcinkc. C. L. (1977). Effect~ of dre ...... on cumpliance to reque .. t .. in a ficld setting. journtll of Social P.IYclwlogy. /01.123-22-1.. Klcinkc. C. L. (1980). Interaction between gale and legitimacy of requeM on compliance in a field sctting. jOllrl/al of NOIII'l'rlJllI Behal'ior. 5. 3-11. Kkinkc. C. L .. & Singer. D. A. (1979). intlut.!nce ofgalc on compliance with demanding and conciliatory fl!quC:-t ... in a field setting. Paw)//(/lity alld Social P~')tcJ/Ol()g\ Bllllefill, 5. 386-390. Krapfcl. R. E. (1988) CU'itomer complaint and ... alesper,>on re ... ponse: The effect of the communication ... ource. j01l1'l/(/1 oj'R()wilillg. 6-1. 181-19H. Linkey. H. E.. & Firestone. I. J . (1990). Dyad dominance cOO1po ... ilion effects. nonverbal bchavior~ .•lnd in nuence. joufl/al ~r Re,\carch i1/ Per.w"wUf\,. 24. 206--2 15. Li ...... B.. Walker. M" Ha;.-elton. V .. & Cupach, W, D. (1993. February). Mllfllal ga::.,e and smiling as correlates of ('omplillllC'e-glli1/ill~ \/In.-ell. Papcr prl! ... ented al the annual meeting of the Westem Slate ... Communication Association, Albuquerque. NM. Maclachlan. J. (1979). Wh<.l1 people really thInk 01 fast talkers . PJ.\'l'IJology Today. 13(6). 112-117. MacNcil1. L.. & Wilson. B. ( 1972). A .field ,\fud\' of lite eJIecH of cm/I'elllimUlllllld IlIIcmn'elltiOfUlI 011 petition ,'IiWlillg bl'lulI'ior. Unpubli ... hcd M:1l1u ...cripl. Illinoi s Slate University. Normal. IL. McGinlcy. 1-1 .. LeFevrc. R" & McGinley, A. (11)75). The inlluence of communicator's body position on opinion change in other.... jOlllal of Penol/alif.\" {/luI Social Psychology. 3 J. 686-690. Mchrahian. A. (1969). Significance of po~turc and po ... ition III the communication. altitude and ... tatu .. rclat ion ... hi p.... P~ycl/Ologic{jl BII/ll'till. 71 . .159-372. Milgram. S. (197-1.). Obedience to authority. New York Ilarper & Row. Miller. N .. Maruyama. G .. Beaber. R .. & Malone. K. (1976). Speed of spcech and persuasion. jOlll'l/ol of P('nonalit.\' and Social P.~yd/Ologl'. 34, 615-62-1.. Mongeau, P. A.. & SlitT. J. B. (199.1) Specifying the ELM: Specifying casual relation:-hips in the Elaboration Likelihood Model. COII/lI/Ulli('(I{iOIl Theon'. 3. 65-72. Moon. Y. (1999). Thc cffcct~ of phy<.;ical di",wlll:e amI re:-pOll',e lalency on persuasion in computcrmcdialed communication and in human -co mputer communication. journal of Experimental P.\ ydrology: Applied. 5, 379-392. Murri .... D. ( 1(77). Mallll'lltching: A field gllidc to III/II/wl IU"UII·ior. New York: Abram~.
d,.".\.,
180
Part II • Persuasion Variables: Penpecril'es
Oil
Source:), Recein!rs. Channels, and Messages
Mullen. B .. Futrell. D .. Stair..... D .. Tice. D .. Baumeister, R.. Dawson, K .. Riordan. C. Radioff. C. Goetha1 .... G .. Kennedy. 1.. & RO!:oCnfeld. P. (1986). New~casters' facial expressions and voting behavior of viewers: Can a smile elect a president? Journal of Perso"aliTY and Social Psychology. 5/. 291 - 295. 1"IUt.!f';on, M. L., Powell. J. L.. & Lenihan. M . G. (1986). Touch. compliance. and interpersonal affect. Journal of Nom'erbal Belu/I'ior. /0. 4 I- 50. Pelly. R. E., & Cacioppo. J. (1986). The elaboration likelihood model of persuasion. In L. Berkowitl (Ed.). Admllces ill Experimelltal Socia/ Psychology (vol. 19, pp. 123-205). New York: Academic Pres ... Pt.!lIy. R. E.. Wegener. D. T.. Fabrigar. L. R.. PreiMer. J. R.. & Cacioppo, J. T. (1993). Specifying the ELM: Conceptual and methodological issues in the Elaboration Likelihood Model of Per~uasion: A reply to the Michigan State Critics. Commullication Theor)', 3, 336-362. Puckett, 1.. Peuy, R. E., Cacioppo, J. T., & Fbher. D. (1983). The relative impacts of age and nttractiveness stereotypes on persuasion. Journal of GerofllOlop,y. 38, 340-343. Raymond. B. J .. & Unger. R. K. (1972). "The apparel oft proclaim~ the man": Cooperation wi th deviant and conventional youths. Joul'llal of Social Psychology, 87, 75-82. Scgrin, C. (1990. November). NOIII'erblll behavior (llid compliance: Affiliation, arousal or domina lice . Paper presented at the annual meeting of the Speech Communication Association. Chicago. Segrin. C. (1993). The effects of nonverbal behavior on outcomes of compliance-gaining attempts. COIIII/lImiClltiu" Swdies. 44,169- 187. Snyder. M .. Grether. 1.. & Keller, J. (1974). Staring and compliance: A field experiment on hitchhiking. Journal of Applied Social Psychology. 4. 165- 170. Stacks, D. W .. & Anderscn, P. A. (1989). The modular mind: Implications for intrapersonal communication. SOllthern Cum1l!wlicatiofl JOl/rna/. J. 273- 293. Stnd,,,. D. W., & Burgoon. J. K. (1979. April). The perJII£Isi!'e (1fects of l'iolatillg spacial disItIll('e expectations in small groups. Paper presented at the annual meeting of the Southern Speed Communi· cation As~ociation Convention, Biloxi. MS. Stacks. D. W .. & Burgoon, J. K. (1981). The role of nonverbal behav iors as distractors in resistance to persua~ion in interpersonal contexts. Celllral SllIte.'· Speech Journal. 32.61-73. Stead, B. A" & Zinkhan. G. M. (1986). Service priority in department stores: The effect.'. of customer gender ami dress. Sex Role.I·, 15, 601 - 611. Stcphen. R., & Zweigenhafl, R. L. (1985). The effect on tipping of a waitress touching male and femalc customers. JOllrlllll of Social P:.ychology. 126. 141 - 142. Stiff. 1. B.. & Boster. F. J. (1987). Cognitive processing: Additional thoughts and a reply to Petty, Kasner. Hauglvedl. and Cacioppo. CommunicaTion MOflographs. 54, 233-249. Street. R. L (1982). Evaluation of noncontcnt !o.peech accommodation. LaIlRlwge and COIl/Il1l11Jicalioll. 2, 13- 31. Strcet. R. L.. & Brady. R. M. (1982). Speech rate acceptance ranges as a function of evalw.ltive domain. Ibtener !)peech rate and communication context. Commullicatioll Monographs, 49, 290-308. Street. R. L .. & Gile~, H. (1982). Speech accommodation theory: A social cognitive approach to langll
c.
Part
III Seeking and Resisting Compliance Strategies and Tactics
The old phra . . t; 'There's more than olle way to "kin a em" may not actually be referring to the proce ...... of pcr<.,uasion. but it certainly applies. Indeed. there is flO "one way" to pcr. . unde another pcr ... on. The Ilumher of Jiffercl1I approaches i~ cndle ... ~. This fact. however. ha . . not prc\'cntcu rc ... earchcr... from tfying to docliment the nUl11eroll~ "tralegies and tactic ...
that people typi<.:ally u..-,e when If) ing 10 influence other.... An enorlllou", amount of research hi.! . . been devoted to identifying not only the different types of influence ..,trategies a\uil able to would-he persuaders btu al"lo to examining how likely pcr..,uudcrs are to select slich \Iralcgic\ and to dc ... cribing the ... illllJlional \ariahlcs that influence their choice~. For example. an earl) ...eminal effort hy Marwell and Sdllllin (1967) identified 16 \uch ... tratcgie .... including making promi",c\ and threatl-.. gClIlIlg other. . to like you. rewarding other. . heforc a . . l-.:ing a fa\or. and "'0 on. Latcr allcmpt ... \\ erc even more ambitiou"i. Kellerman and Cole (199-1). for in"'lance. integratell Ihe I) pologie ... of other rc<.;carcher", into 64 di:-.tinct comp liance -gaining ~trategie<.;. Although ... ueh cfrol1 ... ha\c ... hcd a good dcal of light on the topic~ of compliance gaining and compliance rc"i ... ting. they have al ... o generated con ... illcrablc heat " ... wcll. They ha\e little to <.;
or
181
182
Part III • Seeking {l/u/ Rl~"is lil/K Compliallce: SlrlIll!Xies alld Taclic",'
compliance-gaining resea rch has focu~ed on outcome effecl i venes~. for example, studies on food server behaviors and restaurant patrons' tipping. or o n sequen ti a l request strategies (e.g. foot-in- th e-door und door-in-the-face techniques) in re lation to charity contribu ti o n ~. Too many stud ies. however, have neglected to exa mine the actual effec tiveness of slratcgies on me"age recipients. re nd eri ng the app li cability of their results to th e real world ~omcwhat dubiou~. In contras t, the c hapt crs in thi s part arc designed to address these issues and others. Firs t. chapter II. by James Price Dillard. presents the Goals-Plans-Action Model of interpersonal influe nce. The model is va luable because it hel ps us understand how and why per. . ua . . ivc messages are created. It sugges ts that people ha ve different goa ls when engaging in persua . . ion and shows us how the con fi gu ration of such goals influences 110t on ly the per . . uasive plans and strateg ies that a person generates and selects but also how persuasion is enac ted and reacted to. In con trast, the following two c hapt ers focus on the effectiveness of two particular types of pcrsuusive strategies and tactics. Chapte r 12. by Robert Ciuldini and Rosa nna Guadagno, exa min es the topic uf seque ntial pe rsuasion . Here you will read about a num ber of tactics aimed at ge ttin g people to behave in a particular way. Such tacti cs are labeled "sequential." because all of th em involve in creasing onc\, persuasiveness by saying or doing something before actually making a request. Similarly. chapter 13. by Hyunyi Cho and Kim Witte. focuses on the effects of persuasion by examining the u ~c of fear appeals. This chapter presents the Extended Parallel Process Model. which integrates 40 years of researc h on fear appeals to explain the co nditions under which fear succeeds or fail s as an approach to persuasion. Both of these chapt ers do an outstanding j ob of examining the effectivene . . s of "'peci fi c approaches to persuasion. but more s ignifican tly. th ey help us und erstand the underlying proce~ses that explain why such tactics arc often effect ive. Chapter 14. by Judec Burgoon and David Buller. focuses 011 deception , a tac ti c that i~ frequently included in th e compliance-gaining typol og ies we mentioned earlier. Although some might view deception a . . part of the "dark side of communication." thi s chapter shows that deception serves beneficial social functions and is a part of communication competence. And although so me people might not co n ~ ider deception a form of persuasion. in our view it clearly is. As Mille r (cited in Miller & Stiff. 1993) argued: Decepti ve communication . . (rives for per~uasivc ends: or. ~ tated more preci . . ely. deceptive commun ication is a genera l pcr. . uilsive stralcgy that aims al inOuencing the be li efs. allitudes. and be haviors of other... by means of deliberate message distortion. (p. 28) For this reaso n. then , we ha ve included a chapter on this important form of influence in thi s book. As you will see, Burgoo n and Buller's c hapter unde rlin es the interactional and interpersonal nature of deceptive communication. Their chapter contributes not only to understanding the process by which deception i~ enac ted but also to th e manner in which it is perceived and detec ted . As such. this chapter is as muc h about resisting persuasion as it is about how people altcmpt to per. . uade. This last stat emen t leads us to the final chapter in thi s part of the book. Specifically. the st ud y of persuasion ha~ focu~ed not just on how persuasion occur~ but al so on how the
Part 111 • Seeking allli Resisring Compliance: Straregies and Tactics
183
pcr~uasive
attcmpts of others might be thwarted . Though thcre are many approaches 10 resisting persuasion. chapter 15. by Michael prau and Erin S7abo. is devoted to inoculation, which has received perhaps the most attention in thi s area. The chapte r explains how inoculation works. and more important. it demonstrates the broad and socially significant applications of thi s approach to resisting persuasion.
Referellces_____________________________ Kellerman. K.. & Colc. T. ( 1994). Classifying compliance gaining messages: Taxonomic d.<.,order and strategic confusion. CommllniCllfiotl Theory. 4. 3--60. Marwell. G .. & Schmitt. D. R. (1967). Dil11cn!.ions of compliance-gaining behavior: An empirical analysis. Sociometry. 30. 350--364. Miller. G. R.. & Stiff. J. B. (1993). Oeceplive cOIwlllmic(lIirm. Ncwbury Park. NJ: Sagc.
11 The Goals-Plans-Action Model of Interpersonal Influence James Price Dillard
The Goals-Plans-Actioll Model The Goals-Plans-Action (GPA) model is an attempt to shed light on the way in which messages are produced and on the effects that they have (Dillard, 1990a. I 990b). Because the structure of the model is quite general. it might be adapted to a variety of different communicative functions. For instance , it could be applied to self-disclosure. social suppan. or informarion seeking. De!\pite this versatility, it was developed specifically as part of an effort (0 better understand influence behavior. Accordingly. empirical research de-
signed to test and inform the GPA model has tried to ctnswer the question of how and why individuals influence one another. A lthough the foundations of influence behavior are complex. the GPA model begins with a simple and well-accepted idea: That message production can be modeled as a sequence involving Ihree components (Miller, Galanter. & Pribram, 1960). Goals are the rirst component. They are derined as future Slates of affairs that an individual is committed
to achieving or maintaining (Dillard, 1997: Hobbs & Evans, 1980: Klinger. 1985). Goals motivate plans. the second component in the model. Plalls are cognitive representations of
the behavior; that are intended to enable goal attainment (Berger, 1997). Whereas goals and plans are cognitive entities. actions exist "in the world." Acrions are the behaviors enacted in an efror! to realize the goaL The behavioral response of the message target COI1stitutes feedback to the message source that may produce changes in goals and plans.
These ideas provide the basis for the GPA model. Within this general framework, the model advances a number of specific claims regarding the nature of goals. plans, and actions as well as their relation ship to one another. The aim of this chapter is to explicate those claims and to present some empirical test~ of the tenets of the theory.
185
186
Pan III • Seeking arid Resisting Compliallce: Strategies (wd Tactics
Goals Types of Influence Goals It might seem that people try to persuade others for an unlimited variety of reasons. However, research on interpersonal influence goals reveals that perceptions of infinite diversity are illusory (Cody, Canary, & Smith, 1994; Dillard, 1989; Rule, Bisonz, & Kohn, 1985). The most frequently identified reasons for influencing others are listed in table 11.1, a long with a description and an example of each. Because this list is compiled from studies using a variety of methods (qualitative and quantitative) and data sources (e.g., recalled goa ls versus hypothetical goals), we can have some confidence that table 11.1 captures something of the psychological reality of natural socia l actors. In fact, the concept of psychological reality is important to this theory. Although there are certainly many mental and social processes that occur without awareness, the OPA model focuses on volitional behavior. This focus implies that individuals construct representations of situation s that are meaningful to them and that they are capable of exercising a significant degree of conscious control over their actions. Hence, information concerning what an individual is trying 10 achieve constitures a va lid and meaningful explanation for his or her behavior. Table 11.1 is not without certain limitntions. First, these findings are inherently bound by the characteristics of the research procedures and the samples employed. A ll of
TABLE 11.1
Influence Goals Description
Examples
Gain assistance
Obtain material or nonmaterial resources.
Can I borrow your car?
Give advice
Provide counsel (typically about health & re lat ionships).
I think that you should quit
Share activity
Promote joint endeavors between source and target.
Let's do something tonight. How about going to see that new band?
Change orientation
Alter target's stance toward a sociopolitical issue.
There is another, more realbtic. way to look at the abonion laws.
Change relationship
Alter the nature of the source-target relationship.
J think that we ought to have a monogamous relationship.
Obtain permission
Secure the endorsement of the (more powerful) target.
Would it be OK if I handed in the assignment one day late?
Enforce rights and Obligations
Compel target to fulfill commitment or role requirement.
You promised that you would keep the music down. So. how about it?
Type
using so much Prozac.
Chapter I I • The GOlll.,·PllIIl.\·AcTioll Model (~r IlllerpersOfwl ",{luella
187
the data are grounded in the subjective experience of young adults living in North America and attending public universities. Studies of influence in organi7ations, for examp le. indicate that individuals hold goals that do not appear in table ILl. such as initiating changes in work procedures and improving another's job performance (Kipnis, Wilkinson. & Schmidt. 1980). Thus. although the variety of influence goals is not infinite. il may be broader Ihan is reflecled in lable 11.1. Second. the goals described in table 11.1 exist at a particular level of abstraction thal is subject to debate. One can easily imagine more genenll groupings of goals involving a smaller number of categories. For instance. we might distinguish between goals that substantially advance the interests of the individual versus those that remedy a problem. a distinction that yields only two goal types. COllversely. anyone of the seven goab might be parsed more finely. The Change Relationship goal is a case in point. Even if olle assumes that the communicative processes involved in relational escalation and deescalation bear some similarity. the experience of the two event!'! is quite distinct and the utterances associated with each markedly different. Consequently. it may be desirable to consider several specific forms of the Change Relationship goal. Yet the Mudies that contribute to table I 1.1 suggest that these seven goals reflect how individuals generally think about their int1uence allempts. In olher words. whereas goals can be usefully conceptualized and researched at higher or lower levels of abstraction. the contents of tablc 11.1 are indicative of the level at which ordinary people typically conceive of their own and others' influence behavior (Cody el al.. 1994: Dillard. 1989: Rule el al.. 1985). In sum. ahhough individuab may have many different persuasive goals. research shows that this number is smaller than one might initially think. Although existing research is not without limitations. I believe that the !->even goals described in table I 1.1 represent cOlllmon and recurring influence aims. They also pos~ess particular properties that need to be explicated more fully. Those properties are di~cussed next.
Primary Goals In Ihe parlance of the GPA Illodel. Ihe goals li'led in lable 11.1 are primary goals. They are so named because the theory attributes several unique properties to them. For one. primary goals lie at the beginning of the GPA sequence. They are primary in the sense that they initiate the series of constructs that model message production. Hacker (1985. p. 278) made a similar point when she noted that goals "are reflections of a reality that does not yet exist. but has to be created. and they connect present with future" (I'. 278). From this perspective. primary goals are potential realities that individuals strive to construct. Because primary goals energize cognition and behavior. it can be said that they serve a mOlivat iOIl(l/ junction. A second feature of primary goals derives from the previous point. That is. primary goals allow one to bracket the interaction. to identify its beginning and ending point. Knowledge of what is being attempted permits "oeial actors to segment the stream of imeraction into meaningful units (von Cranaeh. Machler. & Steiner. 1985). Such segmentation is surely valuable for making sense of what might otherwise be viewed as an undifferentiated outpouring of behavior. Meaningful segl11cnt~ of social interaction have been labeled ",oeial episodes" by Newell and Siulman (1988). Bracketing i, possible because
I
188
ParI III • Set'king CIIItI Rt!\;Hillg C(JlIIl'lillll("(': Siraregil's (111(1 /{u-lin the primary goal imbue, the interaction with meaning (Dillard & Solomon. 2(XX)). KnO\\ ledge of the primary goal allo\\ s the interactant~ to ';;.. Finally. primary goals direct a number of mental opcratioll!>.. By providing all unders tanding of the intended purpose of an interaction. goal<., dc termine \'vhich a'pects of a situation are percei\'ed (Kanwishcr. Driver. & Machado. 1995: MandT. Dancken. Campi in. & Currie. 1999: Tipper. Wemcr. & lIoughton. 199-1.). The y also influcnce \\ hich pcn.:cption\ are encoded and retrieved (Cohen. 19R1). In fact. primary goals set into motion an ensemble of lower-level cognitive pro(';cs<.;cs that occlir in parallcl and align with the O\·crall aim represented by the primary goal. In this fa~hion. primary goal\ servc a guidul/ce.lilllclion that promotes tcmporary reorientation and unification of various mental \uhsystcm"i. 1
Secolldary Goals In the cour"iC of pursuing or planning tn pursue a primar) goal. other conct!fIlS may ari"ic. For example. one college ~lUdcllt who hopes to initiate
#1) rnighl recoglli/e Ihe ri'k of rcjeclioll alld wi,h 10 a'oid feeling hurl (goal #2). In a ~il1lilar
vcin. the parcnt who \Vant~ to prevent his young. child from inserting !>.ilvcnvare inlo electrical ...ocket.., (goal # I) might a1"io want to en"illJ'c that hi . . warning doc . . not crcatc a generali/cd fear that eXlcnd"i beyond lhi"i specific problem (goal #2). Such concerns arc called seemulan goals bec. for beha, ior. They arise from individuals' principles and values and, at the broadest level. their . . elf-concept.
or
Ahhough people generally desire 10 a"1 in accordance wilh Iheir principles_ il is probably not the case that individual ... <Jctivcly (';on"iider their identity goals in every interaction. Many social episodes arc routini/cu. and people have gcnerally already made far-reaching deci\ions about what docs and docs not con:-.titute ethical heh.lvior. 111 such cascs. il . . ccm ... unlikely that an identity goal will be
Chapter 11 • The CO(l/s·P/(lfls·Actioll Model oj illterperSOIl(l//lIjlflence
189
Hence, it is most often the case that people try to maintain or improve their relation~hips with others. or course, relational resource goals don' t really come into play unless one has a preexisting relationship with the hearer or hopes to establish one. Relational resource goals focus on the benefits that flow to the source because of the relationship itself. As a consequence. relational resource goals have a longer time frame than conversation management goals. Personal resource goals reflect the physical. temporal. and material concerns of the communicator. More specifically, they arise from the desire to maintain or en hance one's physical well-being, temporal resou rces. finance s. and material possessions. The desire to behave efficiently is viewed as a personal resource goal (cf. Berger. 1997; Kellermann. 1988). although the GPA model does not suppose that IIldividuals always prefer a high level of efficiency. Like so me of the other secondary goa ls. personal resource goals will not be relevant to every interaction. But when th ey are rel evan t. they ca n be important in determining how messages are created and uttered. By positing the existence of affect mallagemelll goals. the model assumes that individuals strive to maintain preferred affective sta tes. Significantly. affect management goals are not so simple as the wish to enjoy positive feeling and elude negati ve ones. For instance. individuals seek to increase their level of anxiety because it motivates vigilance or to enhance their level of anger so that they are emotionally aligned with a plan to take a hard interactional stance ..l The introduction of the concept of secondary goal~ has at least one broad implication for how we conceive of the task of interpersonal innuence. Namely. it suggests that most, and possibly all. interactions in volve multiple goals that individual[o, try to achieve more or less simultaneously. This premise is so broadly accepted among commun ication researchers as to be viewed as a trui sm (Berger, 1997. p. 23). Surprisingly. [o,ome writers in other fields claim that there is a paucity of data-based researc h underlying that trui sm: ;'Multipic goal striving appears to be the rule. yet lillIe empirical research addresse.) the fOpic" [emphasis added[ (Austin & Vancouver. 1996. p. 362). In reality. dOlen, of empirical studie s ha ve examined precisely that topic (e.g .. Dillard et aI., 1989; Meyer. 1997: O'Keefe & Shepherd. 1987: Saeki & O'Keefe. 1994: Tracy & Coupland. 1990: Waldron. 1990: Wil son. 1995). This solid and growing empirical base has helped to inform the GPA model and other theories of influence (e.g .. Meyer. 1997: Wilson. 1995). In sum. secondary goals arc wants that arise in response to th e co n ~iderat ion or adoption of a primary goal. Previous researc h support s the existence of fi ve conceptually distinct secondary goals. However. the exact number is not so important as appreciation of the fact that individuals are almost always attempting to sat isfy multiple goa l ~. The primary goal defines the si tuation. while secondary goals are the e ntailmen ts that follow in its wake. The GPA model holds that understanding the relations hip between primary and secondary goals is crucial to explaining planning and action. Consequently. the next sectio n begins an exploration of those iss ues.
The Relationships Between Primary Goals alld Secondary Goals Possibly the most fundamental co mmunication decision is whether to e ngage i:U1other perso n in interaction or not. The interplay of primary and ..,econdary goals can help to shed
190
Pari III • Seekillg and Rej.;j.tillg COlllplialll'e: Stm/e;.:ft'J ami rae/in light on this choice point in the me~~age production process. To simplify the illu~tration. it will be helpful to assumc a primary goal and just one ~econdary goal and then to evaluate the compatibility between the two (ef. Kellermann. 1988: Samp & Solomon. 1999). Logically,just three possibilities exist. In the first case, the two goals may be incompatible with one another. In this vein, Brown and Levinson (1987) a..,sert that influence aHcmpts are by their very nature intrusive (but ,ee Wilson. Kim. & Mcischke. 1991/1992). Irtruc, then any effort to produce behavioral change in another will necessarily run the risk of threatening that person's autonomy. A second logical possibility is that the .o;,econdary goals are irrele\'ant to the primary goal. For example. concern for a friend's physical well-being is not often an issue when asking the person if he or she would like to see a film with you. In the lhird case, the primary and secondary goals align or are cnmpafible with one another. Relational initiation offers one context in which this might occur. The norm of reciprocity demands that individuals repay favors provided to them by others. When one person asks another for help (e.g .. a ride to the grocery store) that he or ..,he cannOl immediately repay. the message source is signaling a willingne ... s to enter into a relationship in which reciprocity will occur over time. Such is a defining feature of friend!'o.hip" (Hatfield. Utne. & Traupmann. 1979). Thus. the speaker may obtain a ride and. in so doing. also solidify a nascent rclation~hip. Although the third case is clearly the most desirable of the three alternatives. I suspect lhal it is also the least common. Rather. most interactions can be characteril.ed as a blend of cases one and two. Because there are multiple secondary goa"'. it is likely that some of them creale opposition to the primary goal, while others will be irrelevant. Hence. in most instances the set oj relel'{lIIt sa·ol/dclIY goal.\· will constitute a counterdynamic to the primary goal. And. although I have drawn the possibilities in a categorical fashion. the degree to which primary and secondary goals are (in)compatible with one anoth er is more accurately viewed as a matter of degree. To the extent that concern for the seco ndary goals outweighs the desire to achieve the primary goal (and any compatible secondary goals). the individual may view engaging the other as unduly ri"ky and may therefore choose not to engage. Thus. knowledge of the relationship between primary and secondary goals can help explain why individuals make an inlluence attempt or not. However, to speak of the relationship between primary and secondary goal ... is "omething of an o\ersimplification. Actually. primary and secondary goals form a set oJ ... tructured relationships. Coal structure comp/exilv is the concept describing that set of relationships.
Goal Struetllre Complexity Primary goals lend Illation and meaning to social epbodt,;s. However. secondary goab also figure prominently in the message production praces ... in that they reflect other psychologically significant. but logically subsidiary, concerns of the individual. In combination, the two goal types constitute the intrapersonal goal structllre of the communication episode. Research reveals that influence episodes vary in goal stru cture comp lexity, in that various episodes comprise a greater or lesser Ilumber of active goa ls (Schrader & Dillard, 1998) Participants in the Schroder and Dillard (1998) 'lUdy were provided with one of 15 hypothetical scenarios, each of which represented a primary goal drawn from Cody and
Chapter II • The Goals-Plafls-Acrhm Model of IlIferpersoflalllljIuelice
191
colleagues (1994). They were asked to reciJlJ an interaction from their own experience that was similar to the example and then to re~pond to a series of questions intended to assess retrospectively the importance of the primary goal and five secondary goals. The resulting data were submitted to a statistical routine called cluster analysis. Ln essence. the routine looks for patterns of similarity across variables and then creates groupings (i.e., clusters) on the basis of those similarities. In this application. the cluster program tried to create groups of social episodes that were similar with regard to the perceived importance of the primary and secondary goals. Four such clusters emerged slIch that each possessed several notable features as a group. For one. the importance of the primary goal increased from cluster I to cluster 4. Generally. the secondary goals showed a pattern of increasing activation that corresponded to increases in the importance of the primary goal. In other words, as the importance of the primary goal increased. so did the importance of the secondary goals. The relational resource goal was the sale exception to this pattern. It declined in importance as the primary goal grew in importance. In addition to considering the goal clusters in these broad strokes. it is useful to examine each one individually. The first cluster contained primary goals representing innuence attempts that are cOl11mon but not particularly important. focusing on issues concerning the close and collaborative nature of the source-target relationship. Secondary goals were of relatively little concern in these episodes. with the exception of the relational resource goal (cL Wilson. Aleman, & Leatham. 1998). This grouping W3!o. called Il/ailllenallce episodes to emphasize the habitual and relational aspects of the goal"i in the cluster. Primary goals in the second cluster occurred mostly within close relationships. In this case the goals were not routine. but rather represented more important issues of a nonrecurring nature. Two primary goals. Change Orientation and Relational Escalation, were representative of the breadth of issues represented by this cluster. which was labeled .\'pecial issue episodes. The goal structure of this cluster was more complex than that of mail1lenance episodes, but less complex than the remaining two clusters. The third cluster. problem-soh'illg episodes. consisted of goals that represented either high need or high rights to persuade. Such episodes tend to occur within more distant social relationships (e.g .. strangers. neighbors. professors). a feature that was mirrored in the decreased importance of relational resource goals. However. overall goal structure complexity and the importance of the innuence attempt were higher than in either of the two previous clusters. The fourth cluster was called high-swkes episodes to reflect the fact that goals in this grouping showed a dramatic increase in the importance of personal resource and arousal management goals. Members of this cluster were flat only rated highest in importance but also showed the greatest flumber of active secondary goals, that is. the highest goal structure complexity. This cluster included two highly risky activities: initiating a relationship and dealing with a bureaucrat. It seems clear from the results of Schrader and Dillard's (1998) study that goal structure complexity is a useful concept for organizing our thinking about different types of primary goals. The findings also suggest that primary goals with complex goal structures are more difficult to achieve than those with simple structures. Furthermore, we might reasonably expect individuals to be marc reluctant to engage another person in a
192
Pari III • Seekillg alld Resisting Compliallce: Strategies and Tactin highly complex episode because of the many potential risks of failure. Goal structure com· plexity might also shape our next topic. the planning process.
Plans In this section, both the nature and content of plans are examined. Following that. atten· tion is given to the manner in which plans come into being and how choices are made among them.
Features of Plans Plans can be differentiated in terms of their hierarchy, complexity, and completeness (Dillard, I 990a). Hierarchy refers to the level of ab'traction at which the plan is cast. whereas complexity captures the number of steps and contingencies it contains (cr. Berger, 1997). Plan completeness is a measure of the extent to which the plan is fleshcd out. Because the behavior of others is sometimes difficult to predict. it is assumed that even when speakers engage in preconversational planning. the resulting plans are neces· sarily incomplete (Bratman, 1987). These three properties Illay be used to analyze plans of any sort, but it is the cOlltellt of influence plans that sets them apart from plans more generally.
The Content of Compliance-Seekillg alld Resisting Plalls Influence plans contain guidelines for the production of verbal and nonvcrbal behaviors. Whereas strategy level plans are concerned with lines of action and sequences of behav· iar, tactic plalls exist at a lower level of abstraction. 4 They are instructions for producing smaller units of behavior such as individual utterances. For example, though one might approach an influence attempt with the intention of implementing a liking ~trategy, there are many different ways to do this at the tactical level. The first move might consist of utterances such as, Hyou look great! Looks like you lost some weight" or, "That was a really smart thing that you said in our discussion group. I was impressed." Research on the perception of message tactics sugge't' that four dimensions are particularly important to understanding innuence plans (Dillard, Wilson, Tusing. & Kinney, 1997; Wiseman & Schenck-Hamlin. 1981). Any influence plan or behavior can be repre· sen ted as a point (tactic) or a vector (strategy) in this four-dimensional space. The first of these dimensions, explicitness. is the degree to which the message source makes her or his intentions transparent in the message itself. Whereas implicit messages require little or no guesswork regarding the speaker's wants, inexplicit messages necessitate more interpreta· tion (Blum-Kulka, 1987). Table I 1.2 presents examples of both types of action. Dominance references the relative power of the source vis-~I- vis the target as that power is expressed in the message. An expression of dominance in any sing le utterance need not accurately reflect formal differences in status nor a consensual definition of the source-target relationship. Rather, message dominance simply expresses the source's per· ception of, or desire for. a particular source·target power rclmionship.
Chapter 11 • The Gual.\-PlcmJ·-Acf;Oll Model oj IlIIerpersuflallllfillell<."e
TABLE 11.2
The Contelll of Compliance-Seeking Plaus and Actions
Dimension of Plan/Action Explicitlle!oos
193
Example of One Roommate Urging Another to Exercise High: " I would like you to come Low: "Hey, I'm going
to
10
the gym with me:'
the gym."
Dominance
High: "You ... aid Ihat you wanted to wor'" oul. Now, let· ... do it:' Low: " I would really, really appreciate it if you worked out with me."
Argument
High: " I !ooleep a lot hl!tll!f when I work oui. I'll het that you would too."
Low: "We ... hould go worl.: out."
Sou rce control
Iligh: "If you want to get ~oll1e exercise. )'11 go to the gym with you."
Low: ")f you don't get sume c){crcise. you arc probably going to die:'
Argllmellf is defined a~ the extent to which the message presents a ration:Jlc for the sought-afte r ac ti on and rerers to the degree to which the source prO\ides explicit rca~ons for why sthe is see king complia nce. rather than simply making an unelaboratcd request. Of course, messages may be structured argumentatively even though the evidence is less than compelling (e.g .. Langer. Blank. & Chano witz. 1978). Argument refer; to the pcrceived quantity rather than quality of rca"ion giving (d. Roloff. Janiszew!\ki. McGrath. Burns. & Manrai. 1988: Samp & Solol11on. 1999. on elllhel/i"/lIl1elll). COl1lro/ over outcomes is the fourth and final dimen"iion that charactcri/e"i influence plans. The property indexes the extent to which the "iource can exerci"c control over the reaso ns for compliance. Among other things, thi" di-.tinction makes clear the difference between a threat (e.g .. I wi ll hurt you. if ... ) and a wurning (e.g .. You could be harmed,
if ... ). A ... noted above. these four dimcn'iion:., are central to characterizing the content of influence plans. However, they are abo useful for undcp,tanding resistance. In fael. messages intended (0 refute compliance-seeking auemph can be analyzed in term.., of the same four concepts. Table 11.3 provide" an illustration of resistance messages Ihal might be made in response to one or more of the inilucnce mes"iage~ in table 11 .2. Before turnin g to the question of how plans are generated and selected, it i.., important to ask what mi ght be missing frum the charac terizat ion of plans along these four dimen sions. The answer is, "A grea t deal." There arc many elements to plans and co nve r· sal ion that are I/O! encompas!o.ed by exp li ci tness. dominance, argument. and source control. For instance. persons who see"- to innuence often anticipate various forms or re~islance and attempt to gain information about those obMac lc!o. prior to making a request (e.g .. "Are you busy right nowT "No'! Then you wouldn't mind helping me out with this, would
194
Pan III • Seeking and Resisting Complicma: Strategies and Tactics TA BLE 11.3
The Contelll of Compliance-Resisting Plans and Actions
Dimellsion of Plan/Action Explicitness
Responses to One RoomnuJle Urging Another to Exercise High: "1 don't want to." Low: "I'm pretty busy right now."
Dominance
Hi gh: '"I' ll decide when I exercise, not you." Low: " I really appreciate you helping me out in this way, but now is not a good time for me." High: "Can't do it now. I've got to study for an exam later 'oday."
Argument
Low: "Nope. I don't think so:' Source control
High: "I'm just going to take it easy right now, but 1 may go later." Low: "Can't do it. 1 have to wait for the telephone repair person to come."
youT) (lfert & Roloff. 1994; Paulson & Roloff, 1997). Even afler 'he 'arge' has complied, the ~ou rce may revisit the request later in the episode and attempt to sec ure further commitment, such as when a source says "50, you did agree to pick me up at 6 P.M .. right?" (Sa nders & Fitch. 2(01). Thus, the four dimensions should no' be viewed as offering an exhaustive account of the content of influence/resistance plans. Nevertheless, because there is such strong evide nce that social actors them selves view influence in these terms (Di llard. 1997). 'hey mu,. be considered essential aspects of any influence episode. Having now addressed the features and con tent of plans, we can tum our attention to where plans come from and how choices are made among them.
Gel/eratil/g al/d Selectillg Plal/s When 'he desire to influence another arises, individuals will initially search long-term memory for boilerplate plans tha, are likely 10 achieve the primary goal (Berger, 1997; Dillard. 1990.; Meyer. 1997; Waldron, 1997; Wilson, 1995). This search may yield plans ,hat vary in abstraction. complexity, and completeness. If the available plan(s) meet or exceed some individually de,ermined threshold of perceived plan adequacy, then the individual moves toward translating the cognitive representation of action into behavior. Of course, this involves a great many lower-level processes that must work in unison (and could fail do so) if 'he plan is to be successfully instantiated as behavior.' To the ex'en' that the preexisting plans are judged to be less than satisfactory and the primary goal is viewed as important, individuals will devote additional cognitive effort to ( I) making existing plans more complete or more complex and/or (2) creating new plans. This kind of top-down planning is constrained by the recognition that successful interac,ion partially depends on the behavior of the target. To the extent tha, the source views
'0
Chapter 11 • The Goals-Plam-Actioll Model oj lllU'rpt'r.H)Jwf IIIj7m'lIce
195
tho~c response~
a\ unpredictable. he or ..,he wi ll be less inclined to expend cognitive effort in the ~erv i ce of plan developme nt. Moreover, it should be the case that the number of plans viewed a~ adequate shou ld show a negative correspondence with goal ~tructurc comp lexity. Although there may be many different ways to achieve a primar) goal if none of the secondary goals is activated. it ,hould be more challenging to devi ... e CJ plan that will salisfy the man y competing de~ires that are present in an episode th at is high in goal structure comp lexity. When multiple plan~ or plan variations are available. the message ... ouree mu ... t select among them. The GPA model assumes that selection is made with regard to finding a ,atisfac tory confi guratio n of primary and seco ndary goals. This may be a drawn-out. con templative process but very often is not because conversation movef.. so quickly. Indeed. an opportunity for innucnce may open up during the course of interaction that demand ... plan deployment in the nex t co nver;at ional turn. Berger (1997). Meyer (1990. 1997). and Wilson (1990, 1995) all provide more detailed accounts of the cognitive operations involved in plan generation and se lection. Havi ng now described the fir)! two componcnt
Actioll alld Illteractioll GPA theory was de~igned to model the processes by which individuab produce action ... intended to alter or maintain the behavior of other~. The \'arious s ubcomponenh of the model and their relations to one another arc depicted in figure 11.1. Thi ... ,eglllent of the chapter addres~es theory and research regarding how individual ... produce influem:c behaviorf.. and how. together. two individuab create an interaction.
Message Productioll ill the GPA Model The model propOf..es two pmhways to the production of influence behavior. In the first. individuals af..se'''' their goals. decide to engage the target. and then mo\e to plan generati on and select ion. It is assumcd that this sequence is likely to obtain when the importance of th e primary goa l substantially ou tweighs the coullterdynarnic repre ...entcd by thl.! ... ct of relevant. incompatible secondary goals. 1I In the second sequence. the generation of one or more plans that arc viewed a~ likely to succeed encourages thc decision to engage. which is then followed by plan select ion. This path is more likely when Ihe approach and avoidance forces are fairly closely matched. It is worth emphasiLing that although thcse two path~ are logically distinct possibilities. they nced not be approached in a de libermive manner. Rather. because real-time conversation takef.. place very quickly. we might expect tra ve l time on these paths to be measured in milliseconds. Movement from plan ~e l ectio n to tactic implementalion is the tran ... lation of cognitive entilie~ into empirical action. Thi, process must necc, ... arily involve a host of very rapid. elemental processe~. many of which do n01 involve con,cious awareness. The link from tactic impl ementation to target response a~sume~ a target that proce ...
196
Purt III • Sl'l'J..illl-[ amI RI'\i\lil1g COlllplial/a: Strtl/{'Rie,\ al/(I Tactics
Decision to Engage
'"
Goal
Assessment
Plan
Selection
'" .
Tactic Implementation
Target
Response
,•
Plan
Generation
Buffer
F'IGLRE ILl
The Goals-Plans-Action Model of IlIIerpersonalluj1uellce
Thl! ua . . hed line .. ~hov. unt.! of Iht.! tWO p<)... ~ihk palhway~ to pl:m 'eicc tion, while the dolled line ... indie:llc the aJtematl\'c. Sulld line ...... ho" noncontingcnt . . . cquenced ... lage ... in the model
Nil/I':
the '1ourcc may return to the gou l av. arcnes~ '1wge and move throug h th e entire proc es~ again. In such a scq uc n<.:c. goa b are ree va luated in li g ht of the target' s behavior. Alte rnathely. the ... ource may ... to re a number of tact ic plans in a buffer and iterate only as far bad, a ... th e tacti c ~c l ec tion stage . Wh en the !-.ource encounters nonco mpliant behavior fro m th e target. th e first tendency will be to chan ge low level elenlents in the existing plan (Berger. 1997). But to the ex te nt th at the plan it self is ... ce n as having failed. Ihe so urce may di sca rd il wholesale and move to other available options. In the event that no plan is aV<Jilable that ca n sati sfy the primarY-'1ccondary goal array and thai the perceived value of de\bing o ne i'l low, individuals may attempt to exi t th e ep isode. The mos t ob\iou,,> means of accolllpli~hing an exit i ~ by c hangin g the topi c or physically leaving the interaction .
Research
011
Compliallce-Seekillg Message Productioll
Research o n the ll1e ... sage production proces'I permits severa l e mpirical generaliLations that are co n ~b t c nt with the logi c of the GPA model. For example. there is evidence Ihat indi vid ual s try harder to achieve inOuence goals that are important to them . As the importance of the primary goal in creases. so does the amount of planning and cog niti ve e ffo rt that indi vidual , expend in the ,crvi ce of that goal (Dillard et al.. 1989: Wilson & Zigu". 200 1).
Chapter II • 71", GOliIS-PI(/II\ -Anion M(J(JeI (d· JIII('I"f'erwlUiI Inpllt' m·e'
197
Primary goal imponanct! also has implication ... for mC"I ... age con"ltruction. For in... Lance. more important primary goab corre ... pond with l11e ... ~agcs that U!o,c higher level ... of argument (Dillard et 1.11.. 1989). In their ... tudy of computer-mediated communication. Wilson and Zigurs (100 I) showed that primary goal imjxlrtance is negatively a ...... ociatcd with the u ... e of images and emphatic text formatting (i.c .. the u ...c of bold. italics. underlines, font chilngcs. or type size ch~.IIlges). Evidently. important primary goals promote il focus on mc ...... age content and away from me"l"lagc Myle . Of course. a"l predicted by the model. secondary goals ... hape message production as well. For example. in face-to-face interaction .... incrca ... ed importance of the identity goal i ... a ... sociated with increases in the use of argumcnt and dccrea ... es in explicitnc:-,s (Dillard et 0.11.. 19H9). These findings sugge.." a desire to influence on principled ground:-i and to allow the target the option or refu ... al. In computer-mediated exchange .... heightened importance of the identity goal yields reductions in thl! usc or images but increased usc of emphatic tcxt formatting (Wilson & Zigur .... 200 I). The lowered lI'.,C of images might be "iccn as a move away from form in the direction of function. but the gro\"th in emphatic text i~ more difficult to under ... tand. Ccnainly. additional re~ean.:h in both face-to-face and computermediated contex .... i"l needed to paint a more complete picture of the operation of the idl.!n tit) goal. The data al ...o pre~ent a coherent pattern of rc ... ults for the affect management goal. Greater efforts to ll1ilnage one'!o, arousal arl! a~sociated with message~ thai are both more dominant (i.e .. lower in po~itivity) and les ... dcpl!ndeJ1l on argument (Dilbrd ct al.. 1989). In addition. as a!Teet management becomes more important. individuals become Ie ...... verbose and less concerned about conforming with rules of language u...e such a"l ... pelling. capitali/
Illteractioll Processes alld the Rebuff Phellomelloll The behavior of both the ... ourcc and thc target can be modl!lcd as GPA proces:-ic"I. Thu"l, whill! the GPA model is es ... entially individualistic. it allow ... for thc study of inter;'lction a~ a p<.Iir of collaborating GPA processes. Examination of a stream of behavior 1.11., the output of two cooperating psychological sy...,tcm ... rcvcal ... that LOp-dol).. n processc ... highlighted in the GPA :-.equcnce arc receptive to Jllany bottom-up influences. including the action of the other illtcractant. A brief an;.t!y ... i.., of one l!ll1pirical regularity hc::lps to make thi~ point. The rebuff phellolllnuJI/ can be ul!scribed a!o, follow ... : " ... when un initial persua..,ive effort i... rebuffed. follow-up persuasive mcs"Iage..., are ruder. more aggressive. and more forceful than the first one·· (Ilalllple & Dallinger. 1998. p. 3(5). Conceived in this way. the rebuff i"l clearly a pattern of interaction. Hample and Dallingcr (1998) ... ugge~ted that it may occllr for two reason ... ; Inui\idual ... may become more aggres ... ivc beci.lu ... e they
198
Part III • S(>ekillg ami Resisting Compliance: Strategies and Tactics
exhaust their ~upply of prosocial appeals. or message sources may adjust th e ir standards for behavior in such a way that more aggressive messages are seen as acceplable. Although their study did not address th e first explanation. it did offer support for the second. Peop le's concern for effect iveness increased as a positive function of numbe r of rebuffs, wh il e their concern fo r principles and desire to harm the hea rer decl ined. In the te rminol ogy of the GPA model. it Inight be ~a id that resistance (i.e .• rebuffs) in creased the importance of the primary goa l and decreased the importance of the identity and co nversati o nal management goab.
Illteraetioll Processes at Multiple Levels of Abstractioll Th e four variab les that describe tact ic and strategy pl ans and act ions can also be co nce ived of a ... features of th e imeract ion. In fac t. the va lu e of under:-,tanding the influence e pi sode at the level of hoth ullerance a nd episode can be seen in the following exa mple. Consider that a s ingle hint ... uch as. "Do you think that it 's a little breeLY in here?" is low in explicitness. However. a se ries of such messages- "Do you think that ir" s a lillie breezy in here'!"; 'Till kind of c hill y"; "Do you suppose th at someone meant to lea ve that window openT- has the cU lllulati ve effec t of co nvey in g the source's intent very clearly. Given that the ~alllc variab les are used to characteri ze both strategy and tactics. thi s exa mple implie ... a ... urpri\ing point: The impact of a series of utt erance . . that occupy o ne location in the four-dimen . . ional message . . pace may produce an o ut come opposite to that of a si ng le utterance in the sa me location. Or. in this partic ul ar ca . . e. a se ries of inexpli ci t messages is explicit. The imp li cat ion o f this illustra tion is c lea r: If o ur understanding of influe nce interaction ... ca n vary so dramatically a ... a function of level o f analysis, it may be important for future research to u . . e both perspectives whe never po . . sib le.
Message Effects Research 011 the CPA model has abo Mimulated research o n message e ffec ts. The first portion of th is segmen t co ns iders th e impact of influence messages 011 target e mo tions and relational judgments. Th e seco nd portion addresses how the perce ived co mpete nce of influence messages varies as ajoint fun ction of messagc form and goal structurc.
The Relational and EII/otionallmplications of Influence Attempts It seem ... sc lf-c\idcnt that what individuals say to one another has implication s for their feeling ... and the we ll -bei ng of their relationship. Surely, thi s is as true o f influence attempts as it is of com muni cati on mOTe generall y. Thus, it is not completel y surpri s in g that scvera l sludi e ... have shown that influ e nce atlemplS that arc high in dominance have nega tive rclational imp li cati ons. Source dominance corre lates negatively with perce ption ~ of liking for the targct (Dilla rd . Palmer, & Kinney.1995) and with perceived politeness (D ill ard & Kinney. 1994: Dillard et al.. 1997). Conve rse ly, hi g hl y dominant innu ence
Chapter II • The Gou/s-Phms-Acliofl Model oj 11lterpersonallnfluence
199
messages are viewed as illegitimate and as obstacles, two perceptions that typically result in anger (Dillard & Harkness, 1992; Dillard. Kinney, & Cruz, 1996). Explicitness shows quite a different pattern. Whereas some theories of message production contend that explicit requests result in unfavorable relational inferences (e.g., Brown & Levinson. 1987; Leech. 1983). it appears that the reverse is true at least among interactants who are in friendly relationships with one another. In this relational context, highly explicit request!o. seem to ~ignal solidarity between the interactants and correspondingly favorable emotions and interpretations of the influence attempt (Dillard et aI., 1996; 1997). Although explicitness and dominance themselves tend to occur simultaneously (Dillard. Henwood, Giles, Coupland, & Coupland, 1990; Schrader, 1999), dominance is responsible for the negative relational judgments and for feelings of anger. The relational meaning of explicitness !o.cems to be highly context-dependelll (Dillard et aI., 1996). Finally. there is the argument dimension. Individuals generally report a preference for messages that provide reasons for complying versus those that do not (Kipnis et aI., 1980). However. convincing evidcnce exists to show that reasons are not necessary for producing compliance in close relationships (Roloff et aI., 1988). Roloff and colleagues contended that the obligations inherent in close relationships substitute for persuasion. Nonctheless. messages high in argument seem to indicate positive regard for the target (Dillard et al.. 1997), and whether they are necessary or not, they may contribute to the long-term health of a close relationship. Research has yet to examine the relationship between argument and emotion, but on the basis of the research just reviewed, it seems quite likely that messages high in argument will engender favorable emotional responses.
Goal Structure Complexity alld Perceived ComlllUllicatioll Competellce The notion of communication competence hinges on the ability to comprehend the situation accurately and formulate messages appropriate to the circumstance. At the most general level. the Schrader and Di liard study (1998) on goal slructure complexity suggested that some primary goals arc much more difficult to achieve than others (because there are more secondary goals associated with them that speakers are trying to achieve simultaneously). Significantly. the study specified exactly which goals are more or less difficult and grouped them into four clusters. Knowledge of these clusters and their content provides individuals with advance knowledge about the difficulty of achieving various primary goals. In other words, a priori information concerning goal structure complexity can provide the basis for one aspect of communication competence. that is, accurate identification of the important aspect~ of the situation. This is the first step toward constructing effective and appropriate messages. Schrader (1999) studied the relationship between message behavior and perceived competence as a function of goal structure complexity. His work provides several useful picces of information concerning communication competence. First. higher levels of dominancc are associated with higher levels of perceived incompetence regardless of goal complexity. Second. whereas explicitness will nOl harm competence judgments in the low-complexity cluster!o.. explicitness correlates negatively with competence in the high-
200
ParI III •
Se('J.;.illg (lml Resi.will~ COII/piial/a: Strtlll' 1-!it' ,\ al/(I Tactin
complexity clusten•. 1 Here we see evidence suggesting that one's ability to formulate inexplicit me ... sages may substantially enhance effectiyenes"i in complex situalions (which, notably. contain highly important primary goals). Finally. the use of argument correlates positively with competence. except in the high ... takes cluster (i.e .. the most complex cluster). A ... Schrader noted, "In highly unpredictable circum ... tances wherein the target has the power to embarra ... s or humiliate the ;"ource. perceptions of argument use become decidedly negative" (pp. 196-197). As a group then. these conclusions present fairly specific guidelines for what qualifies as competent influence behavior across episodic Y<:lriarions in goal structure complexity. Knowledge of the way in which pal1icular message forms will be perceived encourages planning aimed at developing competent messages.
Frequently Asked Questions About the GPA Model Since the GPA model was first detailed in 1990. a number of questions ha\e arisen connection with it. Several of these are considered in the following section.
III
What Is the RoLe of Awarelless ill the CPA Model? Some of the language u~ed abm'c. such a... the "deci">ion to engage" and "plan ~election:' might be taken to imply a high degree of con ... cious choice in the GPA process. In large mca ... ure however. these lexical choices are only matters of exposi tory convenience. Many of the proce ......cs necessary to run the GPA model may take place with little or no con,,>cious awareness. In this vcin. people often find thcmsclycs embroiled in conversation without ever having made a reflective decision to cnter that interaction. They expel air from their lungs to power their vocal apparatu .... moving their tOngue and lips in (usually) well-coordinated way ... to produce sound ...-all without a\Varene:-.~. At the same lime. the Illodel docs assume that primary and secondary goals an.: ill principle accessible to consciou ... awareness. To make thi ... assertion is. in man) respect..,. equivalent to saying that people generally know what they are doing. When a college student ponder ... how to obtain permission 10 usc her roolllmate's car to gel to the grocery ... tore. she is aware of her goal and. depending on its importance. aspect ... of the planning procc ....... If the roommate in question unexpectedly mentions that ... he will be making a trip to the grocery store. the fir~t Mudent may quid.ly recall her own need for groceries and as"to accompany the car owner \\'ithout any appreciable forethought. In this in ... tance. it might appear that the first ~tlldent acted on a goal without a\\"areness. However. if she were queried about what she wa ... doing with that request. she wou ld very likely understand that portion of the interaction as attempting to gain compliance from her rooml11ate. Put differently. she would be able to recover her goal con ...cious ly despite the fact that she might flat have articulated it without being asked. This example hints at another interesting property of goals: Once an indi vidual begins to pursue a goal via il1lcraction. that goal can be monitored without much consciou~ effort. Consequent ly. when a child attell1pt~ to persuade a parent to quit sllluki ng (a Give Advice goal). he may not be acutely aware of th.H goal throughout the interaction. Rather.
Chapler II • Till' Goa!\ -Plmn -AuuU/ Mmll'l OJ'/IIIt'rI'l'rl"OlilIl h,{l/lell(,(,
201
he i... more likely to foctl . . on the ... econdary goab that are at play on a lurn -bY-lurn basi~. Once a per!-oon become ... beha\iorally commilled 10 a primary goal. awarenes ...... hift ... away from that ovcrarching concern and toward the meal1~ 01" accompli ... hing it.
Do Goals A rise from
Selj~/lIferesf?
It i... ~lIrely the ca ...e thai individuab often act out of sclf-inlerc:-tl. but it would be a mi!-otake to concludc from thai observation Ihat they always ... eck to ... atisfy only their own needs (cf. Shepherd. 1998). Goals CiJn be cgotistic or altrui ... tic. i'lclf-serving or philanthropic. As the existence of the Give Advice goal (table 11.1) implies. intlucnce goab may arise for the purpo ... c ofbcnefiting others (Dillard & Schrader. 1998). Self-interC\I i... not a defining fealure of goaJ... in the GPA modl:!.
Call Goals Challge Durillg COllversatioll? Much of the research conducled on the GPA model has relied on individuals' recollections of conversations in which they look part. Thus. they have been asked about their goals in the convcrsLition as a whole and nol a~kcd to report on their aim~ before Jnd after or on how their goab might havc changed during the conversation. However. a ... ~hould be clear from figurc I 1.1. interaclant ... may conceivably change their goal ... at variolls til11elo! within an intcraction. In principlc. thi, could occur at e\'cry wrn.M Waldron (1997) described the r"ults of a study that speaks to precisely this iSM,e. In hi!-o inve~tigalion. iJ1lcractant... took pari in an cighl-minule conversation in which they attempted to acquirc information about their punncr" ... religiou, or political belief,. Next. each intl:r:lctant reviewed a videotape of the conversation ,Ind rated the importance of their primary and secondary goab evcry 30 ~cconds. Waldron reported that approximately 30 percent of the comparison~ (from one time period to Ihc next) showed a significant shift in goal importance, and over half of thaI 30 percent involvcd multiple changes in importance . .\uch as an increa . . e in the instrumental goal and a decrea!'!c in one or more of the !-oecondary goals. Such result!-o offer ready evidcncc of the nuidily of the GPA proce.\.\. Thinking about influence epi,odes in terms of diMinct primary goal\ may offer a sati~faclOry account of an interaction after the fact. Moreover. characterizing interactions in terl11~ of individual goab is an effil:icnt means of explaining what lOok place and remembering the relevant detail~. Howcver. a~ Waldron's data indicate. when individuals arc engaged in interaction, goals can "hift quite rapidly, just as the GPA model !-.ugge.\ts.
Whaf Happens When a Secolldary Goal Becomes More Importallf Thall a Primary Goal? One an!-owcr 10 this question might be that when a secondary goal eclip~cs a primary goal. thc interaction is no longer about innuence and is consequently beyond the purvicw of the model. However. as noted at the oulSet of this chapter. the GPA model can be applied 10 many different types of intcraction. In fact. secondary goals can assume the status of primary goaI!-.. When they do. the intcraction may not be defined in terms of innuence. but it
202
Part III • Seeking and Resisting Compliance: Stmregies and Ta('(in
can be modeled as a GPA process nonetheless. Consider the following hypothetical ex change between a mother and her son: 01
Mother: "Did you remember to clean up your roomT
02
Son: ··Yeah."
03
Mother: "And did it get done ""
04
Son: ''I'll do it later. Mom:'
05
Mother: "You know that pan of your respon~ibilitie~ a~ a member of this household includes cleaning your room. We're having guests for dinner, and I would like you to have it done before they arrive"·
06
Son:"1 know, Mom. and I'll get it done. butl'lcjust got a lotto about right now."
07
Mother: "What? Is something bothering you?"
08
Son : "Yeah. I did terrible on m) geography exam yesterday. and I feel like dirt because of it. I gue~s I'mju~t dumb."
09
Mother: "Hmm. You know you said that you felt :-.ick yesterday at breakfast. I wonder if you weren't just having a bad day."
thin~
Although it i~ not certain what the aim~ of these particular individual~ might be, the interaction indicates that two primary goals might be in play. From cOIl\'er~ational turn~ 01 to 05, it looks as if the mother is trying to get the son to clean hir-. room. In fact. at turn 05, her explicit reference to his household respon~ibilities suggeMs that we might view the interaction up to this point as an instance of an Enforce Right ... IObligations primary goal. In turn 06, however. the son makes an explicit promise to comply but also ~hift:-. away from innuence LO somcthing else (in the la:-.t cIHu"ic). From that point on, it appem:-. that the mother's primary goal has changed. In turn 07 she seeks information that would allow the conversation to be defined along some other line. By turn 09 it appears that :-.he ha:-. adopted a primary goal of social support. This simple example i, intended to illu,trate two point s about the GPA model. One is that it can be applied to a variety of different Lypes of interaction. However. to move beyond influence episodes will require additional research aimed at uncovering the content of goals and plans in other communication domains. The second point bears on the utility of the notion of a primary goal. From the vantage point of the reader. one can ea ... ily view the sample interaction in terms of Ull influence goal and a :-.ocial support goal. But breaking the conversation into two neat piece~ may not be jll"~tified. Do we need to po ... it the existence of an information-acquisition goal to explain turn 07? Or. i~ turn 07 better understood as a point in the conversation that lacks a primary goal but representr-. a trnn"ii tion from one primary goal to another? In its current formulation, the GPA Illodel doer-. not provide a definitive answer to these questions. Indeed, rather than answer by theoretical decree, the best means of addressing this queM ion may be through the interplay of theory and data.
Chapter 11 .. The Coals- PI(/II\-Auiofl Mudd of IlIfe'1'('I"\OIlOI IIl/lUt'lll"l'
203
COllclusioll This c hapte r preselllcd a brief descri ptio n of the worJ..:i n g~ or the GPA model. ~l\ we ll a~ answers to so me question s that arc so met imes rai sct.l about it. LiJ..:e all current theori es. the GPA mod e l is an incomplc te framewo rk for unde rstanding the co mpl e'(i t ic~ of human comrn unic:ll ion behavior. No n ethe l c~..,. it achieves many of the goa ls that a theory shou ld accomplish. It provides trac ti o n on difficult conceptual i... sue'i. offers guit.lance for empirical resemch projects. and exp lain'i ho\.\ and \I"hy lI11li\ iduals attempt to influence o ne ano ther. By these standard s. th e GPA model has cOIl,idcrahle utilit y.
Notes ________________________________________________________ I. Whic h is nollo ,ay that primary goa l, arc il1\ ilriably 'lH.:l:e,-,flll oll. 1997. p. 11) 3. Thc 1990 vcr ... ioll o f the Iheory COllce:ptlwli/ed the:,e: more narrowl y ilo;; "aroll"'al managcmcnt goals." Affcctmanagement i3 no" preferab le IIhofar a ... it rccognll.e, the nlullidimen,innal ;'Ino multllunctional nature of fecling' . 4. Ce:rtainly o ne c:1I1 conceive: of pl .. n ... atlllany hlt:rarchlCallc\el, dllh:rcnt from tile two offered here. and it may pro\C u ...cfulto do ..,0 depending o n onc', re ... eiln::h quc,tion, Thu, far . t\\.o Ie:'vd, ha\e hce:n suffi cicnt for advanci ng re~carch on inte:rper... onal intluem:c. 5. Ahhough unde:r'tanding the:\e: pn>ce ...... c ... may be IInportanl. thl.!) arc not the: fOl:u, of the CPA model. Ad\andng o ur underqanding (If the role 01 these luwe:r·le:'vcl pnxe ......e' is left to tho..c ime,tlgator ... who find the m int erc,t ing (c.g" Greene. 1997). 6. or cuur ...c. any rde\ant 'iccondary goal'> thaI were cOlllp,lIib lc \\-ith the primary goa l would work again ... tthc ...c t o f relevant. incompinihJc ..,ccondilry goah. 7. I ha\'c wkcn ... ome ... mallllbcrty \\ith Schrader' .. (1999) Jala in thi , Inte:rprCI
Referellces______________________________________________________ Austin. J. T.. & VancoU\cr. J. B. (1996). Goall:On,trul:I" III p')L'holog) Structure. pn:)(.'c'". and l·ontcnt I'n'dwlog iclIl Blliletm. 12U. J38 ·375. Berger. C. R. ( 1997). Planning ,\·rmfexic m({'/"(I('r;oll: Arralllillg ~(Iltll r"mllgll C01l1l/1l11.il'aril·e (lerum. Mahwah . NJ: Erlbaum. Blum- K ulk .. . S. ( 1987). Indirectne .... and pol itene: .... in rcque,,,: Same or dillcrent'? jouf'lwl Pra,~lI/lIfi{'\. II. t3 t- t46. Bralman. M. E. ( 1987). IlIfellfimlJ, ,,(tI//I. alld "m('rical rem 011. CamhriJge. MA: lIarvurd Uni\t.!r'lt} Pre ...... Brown. P .. & Le vinson. S. ( 1987). /Joli re,,('I·,'(.· SOIl/(' rmil'{'r\(l(1 III /(lllgUCHW IImgt'. C:unhrid!:!e: Cunbridge L' ni\o'el"\lt y Pre',.
or
204
Part III • Sc'e/.;illg
{/I/(/
RC'.li ltillg COIIII'/i(IIICl': Slrcl/c'giC''\
(II/(/
Tanio
Cod)" t>.1. J., Canary, D. 1. &. Smith. S. W (11)94). Compliance+gaming goal .. : An inductl\c anal) ...... 01 'lctO('" goal t)pc ...... trateglt! .... and ... ul'CI.' ...... I.' ... _ln J. \\'Iemalln & J D.. I) (Ed ... ). C0I1II1Illmcclfing \lratt'gintllr Hilhdi.lle, :"JJ: Erlhaulll Cuhl.'ll. C. E. (19M I). Goal ... and ...chemala in pcNIIl perception: Ma]..lI1g ...cn ...e frolll the ... Iream 01 heha\ ior In N. Cantur & J F Kihl ... lrom (Ed-..). PuwlI(tfil\-. cognitioll. l/1ul II/cial illlereletioll (pp. 45- 6M). Itlll ...dale. NJ: Erihaulll. Dillard. J. P. (1989). T) pc ... of inlluence goal ... In clo...c n.:lation ... hip.... )olll"l/af (If PU,Wllalllllr! Socilll Rl'faliomlt;pl. 6, 293 ·]08. Dillan.l. J. P (llJ90a). A gual-dri\cn model 01 interpl'NlIlal influence. In J. P. Dillard (Ed.). SI'l'i..iIJg 1'11111· 1'//(11/("£': Thl' prodllClioll oj illft'rpl'rlmwl illjlllC'lIn' lIIi'HUIW' (pp. 41-56). ScolI<;dale. A/..: Gor ... uch S<,·ari ... hric]... Dillard. J. P. ( 1990b ). The nature and ... uh ... tam:e oj' goal ... in tactical communicati on. In M. Cod) & M . McLaughlin (Ed .... ). Till' p~yd/Olog\ (~II(f{'/ini/ ('(I1//11l11l1icllli(l1l (pp. 70-90). Clevedon. UK: Multi lingu,tl Mauer... Dillard. J. P. (1997). Explicating the goal (.;on ... truct Toul ... for thenri ... t .... In J. O. Greene (Ed.). Ml',\ \(/~(' prodUClioll: Adl'{l/It"e,\ ill cOmlllllllil"alioll tlw(//"I" (pp. 47-(9). Mahwah. NJ: Erlbaulll. DIllard. J. P.. & II"r]..nc ...... C. D. (1992). Exploring till' affcctl\e impact o f influence me ...... agc .... Jourl/ol of l..tlllglllllW (11/(1 Social P~n 'lwlo~,\', II. 179 191 DIllard. J P .. Heo\\-ood. K.. Gile .... 1-1 .• Coupland. N .. & Coupland. J (199t». C()mpliance galnlllg }oung and old: Belief.. ahout inlluencc in diffen.·nt age group ... , Comlllllllil"llfioll ReportJ. }. H4 9 I . DIllard. J_ P.. & Kinney . 1'. A (199.t). Expcncntml and ph} ... iological n: ... pon ...c ... to interper-onal influence H,lllum COI/II/IWlinlli(1/I ReH'anh, 20. S02·52X. Ddlard. J. P.. Kl1lne). T A .. & Cru/. M. G. (1996). Innuencc. "ppm, ... a].... and emotion... in do ...c rdatulIl .. hip .... COlllllllllliC"tlliol/ Monographs, 63, 105 1.10. Dillard. J. P .. Palrw.:r. M . T.. & Kinne). T. A, (199S). Relational inference in an influence COIllC\!. I/I/II/WI Commllllil"tltiol/ Rl'wan'll. 2 I. 331-35.1 DIllard. J. P.. & Schrddcr. D. C. (1998). On the utlllt) of the goai .. - plan"'-~Icti()n ...equem:c: Commentar) repJ)'. COIl/mUllin//irlll SI/ldie~. 4t;, .100--304 Dillard. J. P.. Segrin. C. & Harden. J. M ( llJXlJ). Primal') and ...econdtlr) goa].., in the interpeNlIlal inllll cnce prOl.:c ...... COIlllllllllicatioll MOllogrtll'!t.\. 56. 19-~8. Dillard. J. P .. & Solol11un. D. H. (2000). Conccplllallling context 111 1ll~"""lge " prodLiction rc ...earch. Com IIIlmiml;OIl The(//"\'. 10. 167~175. Dillard. J. P.. WiJ..,on. S. R .. Tusing. K. J .. & Kinney, T. A. (1997). Polltcncs ... judgll1ents in pCNmal 1'1.'1;1 tionl;hip .... )011,.,,01 Of/.,(III~I/{/g(' WId Social P\\'('/w/og\·. 16, 297~J25. Greene. J. O. ( 1997). A ...ccond genemti()n act inn a .. scmhly theory. In J. O. Greene (Ed.). Mel,\tlge I"·Or/IIC· lioll." AdmllU',\ /1/ COll1lJ1ll11icalioll theon' (pp, 15 I 17(}). i'>1ah\\ ah. NJ: Erlhaum. lIac]..er. W . (1985). Acthity: A fruitful concept in IIldu .. trial ps}chology . In M. Fre ...e & J. Sahllll (i: d .... ). Goal dirt'l"lf'd "elwdor: The ("(mapl of fiction ill p.ln·llOlo.~\ (pp. 261 -284). IIi1I ...dall.', ;"\IJ l::.rlbaulll. Hample. D .. & D""mgcr. J. i'>t. (199R). On the etlOlog) of the rl.'bllff phcnnmenon: Why :Ire pcr"'U;J ... iH~ Ille .....agc ... Ie ...... polite alter rehuff ... " Cm",lI/l1Ilmlioll Stlldie.\'. 4f}. 3()5-~21. I 1m field. E.. Utnc. M . K .. & Tmupmann. J. (1979). Equity theory and intimatc relationship .... In R. L. Burge ...... & T. L. lIu ... ton (Ed .... ). Sociaf ni"lulIlgl' ill dC'l·elopillg rt'ialioll.l/tip\ (pp. 99-133>. Nc .... Yor].. : Academic Prc ..... . Ilohh.... J. R .. & Evan .... D. A. (19~H»). COIl\·er"',lIion a .. planned bcha\'ior Cogllilil'l' Science. 4. 21J 1J:! Ii oneycutl. J. M ., Cantril!. J_ G .. Kcll~. P .. & L.;unh]..in. D. (1998). 110 .... tlo I low thee? lei me enn ... idcr Ill} option ... : Cognition. \erhal <;(r:lIegic .... and the cl;caiation of IOllInal:). 1-I,1IIUII/ CmlI11Wllil"(II;OIl
Rewardl. 25.
39-6~.
Ifcrt. D. E .. & Roloff. M. E. (199.t). Anticipated oh ... tacle ... to cnmpliancc: Prt.!dicling their prc ... ~nce amI cxprc ..... IOIl. COllllllllllim/;oll Swdit'l. 45, 110 I JO. Kanwi ... her. N .. Dnvcr. J .. & Machado. L. (19lJ5). Sp:ttial rcpctition hlindnc ... ~ i... modulated hy ... elccttlC attention to color antl . . hape. Cogni/i1't' Pl\'dw/ogr, 29. 30] :rn. Kellermann. K (19RK March). Ullt/{'/"\I(//u/illg {(1('/ic(l1 choire: MC'lago(ll\ in cO//I·{'/"wlioll Paper pre...ented at the Temple Uniler... it) Di ...c\lur\c Conh:rence. Philadclphia. PA .
Chapter II • Thl' Couh - PJwl.\-A('tioll Model oj illlt!rper.\ OIWIIIlj1//l'IICl'
205
Klmgcr. E. (19X5). Mi ... <;ing linJ... ... in act ion thr:ur> In ~ 1. Fre ...e & J . Suhmi (Ed .... ). Goal-dirt.'('tl'd hl'iwrior (pp. 311-3J.I). Hillsdale. J : Erlhaulll Kipm .... D .. Schmidt. S. M._ & Win,in ... nn. I (19HO). Intworgilnizalional mlluencr: tacUl· ... : E\ploratlon .. III gelling onc· ... way. Joul"llal ofApplil'd PI,,"wlog\ . 65. -+4Q--...lSJ. . Langer. E. 1.. Blank A .. & Chanowitl.. B. (1978). The mindle .... ne' .. 01 (l .. lcn .. ibl> thuughtful al·tlOn: The role or "p lacebic" infoml~lIion In tnh:rpcr.. onallntcraction. JOllmlll (~f PC'nOlwlirr (/11(/ Social l'\rcllOlag\'. 36. 635----642. Lec(; h. G. (1983). Principle:. o!prag/J/ufin. London: Lungman. MandT. P.. DancJ...ert. J .. Camplin. G .. & C UITie. J. (1999). Behavioral goal-. con .. train the ..election o j \ i .. u:tI information. Psych%giml S6l'II("('. 10. 522-525. Meyer. J. A. (1997). Cogn itive innucm:e .. on the :Ihtlity to addre .... intcrw.'tioll go.. I.... In J . O . Grec:ne (Ed .). M ('\,W/,IW prod/(ction: Ad\'{/nC"l'\ ill ('ol/lll/llI1i("lIlilln Illl'orr (pp. 7 1-(0). M al1\\ ah. NJ. Erlb:tulll. Miller. G. A .. Galanter. E .. & Pribram . K. II ( 1960). l'la//\ (//uJ lite .lfnlCf/(rl'o{ht'!wI·;or. Nc\\ YorJ... : Ii o lt Newell. S. E.. & Siulman. R. K. ! 19S8). The .. odal confrontatio n ~ r i~olk COII/II/lllllwtioll MOIwgmpll\.
55. 266-285. O·Keefe. B . J .. & Shcphcrd. G . J . ( 19R7). The PUNHI 01 mull iple objectIH'" In lace-\()-facc: pcr.. ua .. l\C: imCf:lctioll": Effect ... of con'.truc t diHercntiation un 1ll1!'''age organil
11.263 289. Sc hradt.: r. D. C. (1999). Goal com plc'\ ilY :tIllithe perceived compete nce o f interper .. nllal inllut.:I1I.:C: me .. ...agc .... Coml1lllllicarirm SllIdin, 50. I HH 202 Schrader. D. C. & Dillard. J. P. ( 1998). Go:d ... truclurc ... and interpcr ...onal innuc:nec . ('OIl/I/III11/{'(lIioll SllIdU',\,
.J9, 176-293.
Shepherd. G J . (1998). The trouble: \\ 1\h goal-.. COll1l11l11li("(lIioll Sllidil'l . .J9. 29-l 199. Tipper. S. P.. Wea\er. B., & Houghton. G. (llJ9-l). Ar:ha\ioural goal-. ddcrrninc inhihilur~ mcchani .. m.. in ...elecl i\c attention. Qllarrerl\ JOIIl"llal (If Etpt'riml'tlllll P~yd/(Jlogr . .J7A. gOl)·-R..Hl. TraC)-. K.. & Coupland. N. (1990). Mu ltI ple goal .. in di .. couI"'\c: An oven ie\\ of i ..... ue ... )0/11"//(// oj 1~1II .!:Wllfl' Cllld Social PsycllO/og\'. y, I 1-'. Tral.'Y. K.. & Tracy. S. J. (1991'0. Rudene ..... at911. ReconceplUali71ng face and l'i1ce :Hlad. /-1 //1//011 COlli/11ll11iwlioll Re.H'Clrch, 25, 225-15 1. "nil Cranal"il . M .. Madder. E.. & Steiner. V. (19H5). The organi \ati nn of goal-direc ted action: A re ...C::lrch rc:porl. In Gin .. burg. G. P., Brenner. M. & \'(1I1 Cranac h. M. (bh.). /)il('(I\ 'er\'\/r(llegie,\ ill Ihl' pHe/wlng\' (if {lct;OII (pp. 19~(1). London : Academic Pre ...... Waldron . V. R. ( 1990). Com,trained ralional it y: Sit uational inllllcnce .. on informatIOn :lcqll l.. itlOn p lan .. and tact ic ... COIl/l1Il11licarioll /I1ollogmplll, 57. 184 201. Wa ldron . V . R. ( 1997). Toward a theor> or interactl\c planning. In J . O . Greene (Ed .). Mt',I\Ug l' pror!m 'lion: Adwl/laf il/ COllllllllllinllioll II/l'on' (pp. 195-220). Mahwah. ~ J : Erihaul11. Wi l...on. E. V .. & Zigur\. I. (200 1). Intcrper..ona l inllucnce goal\ and com puter-med iated communicat ion. JOllmlll (If Organi:arjOlwl COl1lJllllillg lI/ul l:'/l'l'I/"fmic COif/man'. II. 59-76. WiI ..nn, S. R. (1990). Development and tc: .. t of a cog nit1\ C: rule . . model 0 1' i/llcracl1on goa!... COl1l",llfIimliol/ MOllographs. 57. 8 1- IOJ.
206
Part III • S(Jeking lind Resisting Compliance: StrCltegit's and Tactics Wihon. S. R. (1995). Elaboratin g the cogn iti ve rules model of imerac tion goa ls: The problem of account ing for individual differences in goal formation. In B. R. Burleson (Ed.). Communication yearbook IN (pp. 3- 26). Thousand Oaks. CA: Sage. WiJ..,on. S. R.. Aleman. C .. & Leatham. G. ( 1998). The identity implicat ions of influence goa ls: A revi~ed analy..,i~ of face-threatening acls a nd application to seeki ng compliance with same-sex friends. Hum(lll Commu1/icaTion Re'iearch, 25. 64-96. Wihotl, S. R .. Kim, M.-S .. & Meischke. H. ( 1991/1992). Evaluating Brown and Lev inson'., politeness theory : A revi<;ed an al ysis of directives and face. JOIll'f/a/ of ulll~J/(/ge alld Social/IIl('/'C/crion. 25. 2 15-252. Wiseman. R. L.. & Sc henck- Hamlin, W . J . ( 198 1). A multidimensiona l scaling validlltion of an inductively-deri ved set of compl iance-gai ning Mralcgies. Commullicatioll MO/lographs. 48. 25 1-270.
12 Sequential Request Compliance Tactics Robert B. Cialdini and Rosanna E. Guadagno
Have you ever gone to a store or auto sales lot with the intent of purchasing the "bargain" item that was advertised in the Sunday paper? If your experiences are anything like ours, when you arrived, the salesperson may have told you that the bargain item was sold out, but that a similar item at a higher price was available for you to purchase instead. Many people in such a situation find themselves buying the more expensive item and, once they walk out of the store. end up wondering why. The reason most people buy the more expensive replacement item is that they have already committed themselves to the purchase. As you will read later in this chapter, commitment can be a very powerful motivating force. The above scenario is an example of a compliance tactic called bait and switch, wherein an individual commits to purchasing one item, only to have it replaced by another, more expensive one. Often we find ourselves purchasing items or agreeing with requests made by friends. family, or salespersons when we may not have initially planned to do so. In these situations, the individuals around us may have used persuasive appeals to gain our compliance. Compliance occurs when an individual behaves or responds in a particular way because another individual is encouraging him or her 10 do so (Cialdini & Trost, 1998). For instance, individuals who end up purchasing the more expensive ilCm as a result of the baitand-switch procedure are complying with the salesperson's suggestion. At some point, they acquiesce. There are many types of compliance tactics that can be used to increase the likelihood that we will agree with another's request. The focus of this chapter will be on sequential compliance tactics, that is, tactics that require more than one step (usua ll y two) to be effective. These types of tactics can be successful in gaining compliance. However, before we move on to a discussion of why compliance tactics work, we need to introduce a few terms that we will use throughout the chapter. Because most sequential request compl iance tactics are two-step maneuvers, there are common terms for
207
208
PJn III • Seeking and Re,i,flinX Compliance: Srrmegie,\ lind TllUics
each request or stage in the process, The first \tagc of a sequential rcqueM tactic is usually rcfcrred to as eithcr the initial or first request, The next request or singe in thc process is u ... ually rcfclTeu to as the second or rarget request, becau:-,c it is lhe rcque:-, t on which the influence agent actually hopes to gain compliance, The first request is what helps to in-
crea;c the likelihood of the larget of innuence acquiescing to the larget request. This can occur for
Why Do Compliance Tactics Work? Cialdini CWO I) was interested in dctermining why influence practitioners sLich a.., salespersons are "'0 ..,ucces~ful in gaining compliance from their targets. To examine this question. he observed the mcthods used by ~ales practitioners in real s ituations such as advertising. fund -raising. recruiting. anu sales. Ba!o.ed on hi~ ob:-,ervatioll!o.. Cialdini determincd that ... ix key principle ... of influence underlie moM influence attempts: scarcity. reciprocity. con~isteIH..:y/commitmcnt, authori ty , social proof. and similaritylliking. When scarcity is u!\cd. an item or opportunity is presented as somet hin g that is not readily available. eithcr due to low quantity or because the offer is on ly good for a short perilxl of timc. For example. when the Mazda Miata was first introduced. it was released in slich low quantitie ... that the cars usually sold for several thousand dollars over the manufacturer" ... sugge ... ted reHul price. Because the Miata was both new and hard to get. its !\carcity increased it ... de ... irability. The second principle. reciprocity. de ... cribes intluence tactics that work because the inllut.!nce practitioner has done a favor for or made a concession to the ",rget of influence. Targets are more likely 10 agree with the request because they feel they "owe" the influence practitioner. Anyone who has received free address labeh with a request for a charitable donation shou ld be familiar with these types of tactics. Con:-,islcncy-and-commitment taclics work because the influence practitioner is able to get the target to commit to the transaction before the practitioner alters the deal. The example of the bait-and-switch tactic mentioned at the beginning of this c hapter illustrates this type of tactic. Next. experts CU ll influence us because they are authorities on a lopic. For instance. Senator Bob Dole is sliccessful at selling Viagra because he has experienced prostate cancer. Social proof is most successfu l in situations where we look LO ot hers to guide our actions. We choose to engage in a behavior because we believe th at others would do the same thing in that situation. Car salespeople take advantage of thi s when they emphasize how popular a particular car is. Internet service providers use this taclic when they empha ... ize the large number of subscribers to their service. Finally. simi larity and liking
Chapter 12 • Seqllenfilll Reqlltst Compliance TlICfiCJ
209
tactics emphasize that the intlucncc agent is likable or similar to us. For instance. a salesperson may statt! that he or she shares the ... ame hobbies or drives the same car as a potential cu ... tomer. The mention of this similarity is intended to make a customer more likely to purchase from the influence practitioner. The majority of scquential requcst compliance tactics fit into one of two categories from the list above: commitment and consistency or reciprocity. We will start off with a discussion of commitmcill-and-consistency tactics and then move on to reciprocal tactics. In each section we will cover the basic mechanism of each tactic. factors that affect the likelihood of its success in gaining compliance. and defenses against each type of tactic.
COl1ll1litl1lellt-alld-Collsistency Tactics Have you ever seen an adverti ...emcnt encouraging you to enter a contest concerning a certain household product? in such contests. individual ... arc asked to write a short testimonial on the product to explain why they usc a certain laundry detergent or toothpaste. The pril.e for the conte ... t i... frequently a one-year supply of the product. Well. that kind of conte ... t .,ound" like a good deal for the winner. but thc manufacturer of the product will have to give away a large quantity of its product. Why would the manufacturer want to givc away money that way? In actuality the manufacturer ha~ found a subtle way to increase sales: By a~king consumers to extol the virtues of the product. the manufacturer is ensuring that many people will make a public commitment to the product. This written and public commitment will increase most individuals' loyalLy to the brand. So while giving away a oneyear supply to one winner, the company is actually gaining thousands of customcrs who will become more loyal and more likely to purchase the product consistently than before the contest. And. commitment can be a powerful motivating force. A number of sequential request compliance tactics work by making the targets of inlluence feel committed to an action or a product. We will review three of these tactics in the sections below. Table 12.1 presents the bafo.ic mechanism of each tactic: the low-ball procedure. the foot-in-the-door technique. and the bait-and-switch tactic. Although each tactic is different in the v. ay it i... implemented. all rcly on the same underlying mechuni\lll ... to make them \ucces"ful: com",irmelll and consistency.
The Low-Ball Procedure The reason the ... econd author of this chapter studies social influence has to do with the low-ball procedure. During her senior year of college. she bought a car. At the time it seemed like a good deal. Two weeks later. however. the salesperson from the dealership called to say there was an error in the loan paperwork and that she had to return the car or pay an additional $2.000 for it. or course she chose lO pay the money. because she was already committcd to the I,;ar and had already shown it to her friends and fami ly. It was only after this stressful predicamcnt was over that she realized she had never received an updated loan statement and that the extra cash was never recorded on any of (he paperwork. It was then that she realized she had been duped.
210
Part III • Seekin!-: and Re.vi'iting Complicllll'l': Slmlt'xie\ and Tactics
Commitment-Based Sequential Request Compliance Tactics
TABLE 12.1
Tactic
Initial Request
How Initial CommiJment Is Created
Example
Second Request How the Initial Commitment Is Used to Trap t"e Target into a Less Desirable Outcome
Example Changing the deal due to an error in the financial paperwork !l0 that it co~t~ more .
Low-ball
Gelling the target to agree tn a <.,pecific dc .. irable arrangement.
Negotiating a deal to pun.:ha.,c a car.
The tcnll'~ of the agreeme nt change to be Ie!';" advan tageous for the larget.
Bait -a nd-Switch
Urging the target to cOll1mit to a certain bcha\·ior or action.
Gcttlllg the targct to commit to purcha .. ing a new 'Iereo ha ..ed on a 10\1( adverti ..cd price.
Informing the target that Making the behavior the desired .. tcreo is ~o ld or action unadvi .. ablc or una\ ail:lolc and out and offering an proposing an altemati\e. more alternative that i.. not expcn,ive one. a, de .. irable an outcome for the target.
Foot -i n-Ihc-Door
A .. king the target 10 agree with a small reque3t.
Ge(\lng the target to sign a petition ad\'oc::lIl1lg govcrnmt.!nt aid 10 a group In need .
Following up with a rcque<.,t for more a .... i3tance on a (u,ually) rdated reque ..!.
A<.,J..ing the targelto volunteer time or donate money to help the same group.
This experience is one example of what soc ial scien ti sts rerer to as the JOI\'-ba/l technique. This technique is used when an individual commi ts to one outcome. in thi s case purcha~ing the car for a specific price. Once the commitment has been made. the deal changes and becomes less desirable than the arrangement 10 which the target initia lly committed (i.e .. the price on the car is rai~ed). Howe\er. because of the existing commitment. many people will sti ll agree to it. Thl". in the low-ball technique, an individual agree\ to the first request because it i ... easy to agree to. or is advantageous to him or her. When the opportunity or deal changes to become less desirable. most individuals already feel commilted and follow throu gh on that commi tment. Cialdini. Cacioppo. Bassett. and Miller (1978) were first to demonstrate the low-ball technique. In their study. research participant ... were called and asked to participate in an experiment. Participants in the con trol gro up were told that the experi ment was sc hedul ed for a very undesirable time: 7:00 A.\I. In con tra ... !. participants in the low-ball group were first asked to participate in the experiment and th en, once they agreed, were informed that it would take place at 7:00 A.M. As can be see n in figure 12. 1. the initial commitment to the first reques t had the desired effect: Participalll~ in th e low-ball condition agreed to participate at a much higher rate than did control participants (56 percent to 24 percent respectively). In addition_ the commitment made by the low-balled participants had a
Chapter 12 • Sequential Request Compliance Tactics
211
100% 80% 60%
•
Controt (participants told the time of the experiment before being asked to commit)
•
Low-ball (participants told the time of the experiment after they agreed)
56%
40% 20% 0% Percentage of participants who agreed to participate in the experiment at
7:00 FIGURE 12.1
A.M.
Compliance with the Target Request ill the First Low-Ball Study
Source: Cialdini et al.. 1978.
powerful impact on actual behavior: 95 percent of participants in the low-ball condition who agreed to participate actually showed up for the experiment. Why is the low -ball technique so successful? It works because once someone agrees to a business deal or course of action. that individual feels committed and will stay committed to that course of action even when the details change. In addition, the initial request is usually perceived as a "good deal" by the target. and this perception helps enhance his or her commitment. As the initial demonstration of the procedure presented above clearly illustrates, the low-ball approach is a powerful compliance tactic. However. a number of factors affect the likelihood of its success. For instance. additional research has indicated that the degree of commitment between the target and the influence agent plays an important role in the success of this technique_ For example, Burger and Petty (1981) demonstrated that this tactic did not work when different people administered the first request and the target request. In addition. Burger and Cornelius (200 I ) demonstrated that both public and verbal commitment is necessary for this technique to work. In their study. participants who were interrupted before they had the opportunity to commit to the initial request were significantly less likely to agree to the less desirable target request than were participants in the low-ball condition. Additional research indicates that targets must also feel that they made the initial commitment freely in order to be sliccessfully low-balled (Cialdini et aI., 1978).
Bait alld Switch At the start of this chapter. we presented an example of the next consistency-based compliance tactic we will examine: the baif-and-switch technique. We are sure many people can remember instances when they have been drawn into a store after spotting the 40 percent off lag on a nice camera or pair of shoes. However. if your experiences have been anything like ours. you have probably been disappointed because the camera was out of
212
Par1 III • Seeking (llId Resisting Compliance: Stra tegies and Ta ctics
Mock or the bargai n shoes were available only in extremely large or small sizes. In most situations like these. the typical sales associate is only too happy to recommend an alternati ve s. hoe or camera that is very similar in appearance or features but is pricier. And if you are anything like us, you may have purchased the alternative item even though it was not what you wen! into the store for in the first place. If so, you have experienced the baitand-sw itc h procedure. The tactic works by ge tting a target to commit to an item or a course of action. Then, once the commitment has been made or the "bait" taken, the influence practitioner alters the silUation so that the item or action that the target has co mmitted to is no longer available. The influence practitioner the n offers an alternative option or course of action that is not nearly as good a deal for the target. Many target s in thi s situation will accept the alternative item or action eve n though they would not have done so if that had been the initial option. Although thi s. tacti c ha~ been around in the sales industry for quite a while. the initial publi shed cx pe rime ntal dcmonstration of the bait-and-switch tactic was authored by th e French social psyc hologists Joule. Gouiiloux. and Weber (1989). They referred to it as the lure procedure. In the ir study, they recruited partic ipants to take part in a study on film c lips-a very interesting experiment in the eyes of the typical re search participant. Thi s was the bait. Once willing participants arri ved to take part in the experiment, p:.uticipant <;, in the lure co ndition were told that the experiment had becn canceled. Then they were sw itched: They were offered an alternative experiment in which their task would be to me morize list~ of numbers-a ve ry boring task in comparison to the original experiment. The re~u lt ~ attested to the s u cce~s of thi s compliance tacti c. Among a control gro up of participants who were ju~t asked to do the number-matching task , only 15 percent agreed. whereas 47 perce nt in the bait-and-switch co ndition agreed to be in the memory study _ Wh y does it work? Much like the low-ball procedure, the bait-and-sw iICh techniqu e works by ge lling an initial co mmitme nt to comply, which makes people more like ly to accept a less attractive arrangement than they ordinarily would have accepted. It is diffe rent from the low-ball tactic in that the outcome the target commits to is not altcred to beco me something less des irable; rather, it is replaced by a different outcome that is less desirable than the one initially committed to. For example. a victim of th e low-ba ll procedure would agree to buy the sa me car for a hig her price, while a victim of the bait-andswitc h proced ure would agree to buy a different. more expensive. car.
Foot-ill-the-Door Technique A few years ago, while walking through a shopping mall , the second author was stopped by a clean-c ut young man who asked her a few short questions about her hou sehold demographics. Once she answered hi s questions, he thanked her for her time. Then, before she could walk away. he asked her to provide more information about her family demograph ics and consumption habits in a longer interview. An hour later. she walked out of thi s yo un g man's office carrying with her a box of ';fire-roasled grill snap s" that she had agreed to taste- test fo r a week and wondered what had hit her. Her family tried th e crackers. and a week later she participated in a 20-minute telephone interview to provide feed -
Chapter 12 • Sequel/tia! RequesT COlllpli(lI/C() Tactics
213
back on the taste test. She was amazed that agreeing to answer a few short questions had led her to commit to a course of action that she would have initially refused if she had been informcd of what the clean-cut young Illan really wanted from her in the first place. This is an example of the next consistency-based compliance tactic. the foot-in-the-door (FITD) technique. Essentially. the FITO technique works by asking for something small-usually a minor commitment-and building upon that commitment to gain compliance with a larger. usually related request. When two social psychologists. Jonathan Freedman and Scott Fraser (1966). initially investigated the FITO technique. they found that once an individual agrees to the small request. he or she is more likely to agree to a related. larger request. They labeled this effect the foot-in-the-door technique because the slllall request is like the proverbial foot in the doorway that makes it hard for a potential customer to close the door on a salesperson. In their initial demonstration of the FITO cffect. Freedman and Fraser (1966, study 2) asked participants either to ~ign a petition or to place a small card in a window in their home or car. The petition and the card advocated one of two prosocial messages: to keep California beautiful or to support safe driving. Both requests were easy to agree to. After all. most people living in California believe in safe driving. and 1110st Californians wan I to keep California beautiful. Approximately two weeks after complying with the initial request. participants were contacted by a second experimenter and asked to place a large sign advocating safe driving in their front yard. Although initial acceptance of the small card that advocated safe driving led to the greatest amount of compliance with the large request. all experimental conditions generated more compliance than the control group.
Why Is the FITD Techllique Effective ill Gaillillg Compliance? Freedman and Fraser explained their results in terms of self-perception. They concluded that compliance with a small initial request for a public service action causes a change in the individual's self-perception. This small act of compliance produces a change in self-concept in which the person "becomes in his own eyes, the kind of person who doe;., this sort of thing" (p. 201). Thus, the initial act of compliance with a small request. a request virtually no one would refuse. makes an individual more likely to agree to a later. larger rcquestparticularly if it is similar to the initial request. Othcr researchers have challenged this explanation of the FITD technique both because they suggest alternative explanations seem more likely and because the FITO technique has been notoriollsly difficult to replicate. In a literature review on thc FITD effect. Dejong (1979) concluded that support for the self-perception theory is weak. Similarly. in a meta-analysis of 120 FITD studies. Beaman. Cole. Pre>lon, Kientz. and Steblay (1983) concluded that the FITD is an effective compliance technique. but that the size of the effect is smaller than was suggested by the results of the Freedman and Fraser study. The results of the meta-analysis by Beaman and colleagues also suggest that support for the self-perception theory to explain the FITD effect i;., inconsistent. In addition, in another FITD study. Gorassini and Olson (1995) measured participants' self-perceived helpfulness between the first and second reque!->ts. Although participants in the FITD condition perceived themselves as more helpful than participant .. in the control condition, this greater perception of helpfulness did not predict compliance with the second request. In
2 J4
Pan III • Sl'ekinlt and Resi,Hillg Compliance: Srmfl'gie.\ allli Tacrin
sum. much research on the FlTD ha ... ind icated that o ther factors beyond ... elf-pe rcept ion may we ll inOuence susceptibil it y to it.
Factors That Affect the Likelihood of all FITD Effect.
More recc ntly. Burger (1999) conducted a meta-analysis on FITD lactics and found support for a number of factor ... that inOuencc whether an attempted FITD lUetic w ill be e ffecti ve in increasing compliance. A detailed listing of many of the facto" appears in table 12.2. The first. as we ha ve already me nti o ned. is self-pe rcept ion. Individuah. who see the act of com pl y i ng with the lirst req uest as indicative of the type of person they are will be 1110re like ly to comply with the target reque~t than will indi vid ual s who do no t experience thi s self-percepti o n. For instan ce. Burger and Caldwell (200 I) reported that a monetary reward for co mpliance with the first request reduces comp liance wi th the second request. They ex plained this findin g as indi ca tin g that a monetary reward leads people to believe that they agreed to the first request for the money. not because o f the kind of people they were. In addition. Burger and Guadagno (in press) reported that individuals who have a c lear or high sc lfconcept are more \usceptibl e to the FITD than are indi\iduals who have a less clear scn ... e TABLE 12.2
Factors Tho/Impact the Likelihood oJ a SuccessJlI1 FITD I\ttempt
Psychological Process
Potential Effect Oil
the FITD
Example
Self-percept ion
Enhances the effec t
H an individual sees him- or herse lf as the type of pcr<';ol1 who c n gagc~ in actiom ... lI ch a ... Ihe initial rcque ... t. he or ~he wi ll be more likel y 10 agree with the larget req uest because of thai se lf-percept ion.
Reciproc it y
Reduces the effec i
Indi vidua l... who comp ly with the initial request the requc ... lOr has done them .1 favor are unlilo..ely to co mply with the larget request because they perccive the favor a~ already havi ng becn returned. b CC.lU SC
May reduce or enhance the effect
Indi vidual ... for whom consistency i ... not a core need arc unlikely 10 be ... u~ceptib l e to the FITD. However. when cOlh i... tency is a core need. the oppo ... ite is true.
Anributions
May reduce or enhance the effect
If indiyidu;:lb allribute their comp li ance to an ex ternal factor (e.g .. payment for complying), 3n FITD effect i~ unlikely to occur. However. if indi vidual ... altrihute thei r co mpl ia nce 10 an interna l factor (they arc helpful by nature). they are more lik e ly to compl y.
In vo lve ment
Enhance~
The greater the involve ment requ ired to complete the initial rcque~t. the morc likely individual:. arc to co mpl y with the target request.
Cons iste ncy
SO/lrce:
need~
the effect
Adapted in part from Burger. t<)99.
Chapter 12 • Sequential Reqllest Compliallce Tactics
215
of self. because such individuals are more likely to aher their self-concepts when reacting to new information than are individuals with less clear self-concepts. Thus. after complying with the first request in the FITD manipulation, individuals with high self-concepts are morc likely than those with low self-concepts to experience the resulting change in selfperception that will incline them to agree with the second request Self-perception. however, is only one factor affecting the likelihood of a successful FITD. For instance. according to Burger's (1999) meta-analysis. if targets comply with the initial request because of reciprocity norms, that is. because they think they owe the influence agent something. they are less likely to agree with the target request. The perceived reciprocity produces a boomerang effect. because the targets come to believe that they agreed with the initial request only to return a favor. For instance. if an individual received a free gift for signing a petition to change the speed limit (first request). he or she would be less likely to agree to attend a demonstration on the topic (target request) than if he or she had not received the free gift. Individual differences in consistency needs also have an impact on the likelihood of success of the FlTD effect. Cialdini, Trost. and Newsom (1995) suggested that individuals might actually differ in the amount of consistency they prefer. In their study. Cialdini and colleagues introduced a personality scale that measured individuals' preference for COI/sistellcy (PFC). The PFC scale mcasures individual diffcrences in the desire to be both internally and externally consistent on three separate but highly interrelated subscales: the preference for consistency within oneself, the preference to appear consistent to others, and the preference for others to be consistent. Individuals who score low on this scale may actually prefer to behave in an inconsistent manner. Conversely. individuals who score high on this scale consider consistency very important. Cialdini and colleagues (1995. Study I) conducted an FITD study, examining whether PFC level would have an impact on the success of an FITD manipulation. Specifically, they predicted that high-PFC individuals would bc susceptible to the FITD and show the traditional increase in compliance after first agreeing to a small, related request. For low-PFC participants. they predicted no difference between the FITD and the control conditions because consistency was not important to them. To test this hypothesis, the researchers contacted experimental participants by telephone and asked them to answer three short questions about their television viewing habits (the first request). Next, all participants received the target request: to fill out a 50-item questionnaire on their television viewing habits and return it in two weeks. The results showed. as predicted. that highPFC individuals (people for whom consistency is important) were more susceptible than the 10w·PFC participants to the FITD effect. Figure 12.2 presents a graphic representation of the results. Thus, high-PFC participants who agreed to a slllall request were more likely to agree to a second. larger request. Conversely. low-PFC participants were just as likely to agree to the second request whether or not they had agreed to the small request first. These results were interpreted 1O support the hypothesis that individual differences in PFC are one reason the FITD effect is difficult to replicate reliably. Additional data analyses revealed that the low-PFC participants displayed a relatively strong tendency to say yes regardless of condition. This tinding suggests that low-PFC individuals tend to be interested in novel opportunities and experiences such as the chance to participate in an unknown survey about television viewing habits.
216
Part III • Seekil11: {lml Re:ii.Hillg COl1lpIUIIICe: Stratt'1:ie.1 a/Jd Tactics 100%
90% 80%
• Control 70%
•
60%
FITD
50% 40%
Percentage of participants who agreed to fill out the 50·item survey
30%
20% 10'%
0%
Low
High PFC
FIGURE 12.2 Individual Differences ill Compliallce with the FITD Technique. People with a low PFC are not ..,usceptibJe 10 the tactic. wherea.., people with a high PFC show the traditional increase in compliance after fir..,t agreeing to a ... mall initial reque .. t. Srmne. Claldini el al.. 1995.
In a replication and expansion of the ..,(udy by Cialdini and colleagues. Guadagno. A ... her. Demaine. and Cialdini (200 1. study 2) used the same method as in the previous study. with slightly modified FITD conditions. They found that for high-PFC individual>. reminding them of the concept of consistency between the first and second requests increased the extent of the FITD effect. For low-PFC individuals. however. reminding them of the concept of consistency had the opposite result. actually decreasing the FITD effect. In addition to co nsistency needs. several other processes influence the likelihood of the FITD effect (Burger. 1999). For instance. the greater the action or involvement required to comply with the initial reque .. t. the greater the effectiveness of the FITD. Han\en and Robinson (1980). for example. reported a stronger effect when participants elaborated on their answers to question!oJ than "hen they simply provided responses. Labeling the behavior a .. helpful also increases the likelihood of an FITD effect unless the target is low in PFe (Guadagno et
Defellse Agaillst Commitmellt and Consistency Tactics What can individuals do if they find them!oJclves trapped by their own consistency and about to hecome a victim of the low-ball procedure. the bait-and-switch tactic. or the
Chapter 12 • Sequential Request Compliance Tactics
217
FITD technique? Cialdini (200 I) recommended two options, both based on the premise that consistency generally is a good thing unless it is foolish and rigid. Prospective targets of influence (e.g., most people) should learn to recognize both when a commitment and consistency tactic is being used on them and when they are engaging in such foolishly and rigidly consistent responses. How can we tell the difference between healthy consistency and this perilous second variety? Sometimes when we are in a situation in which an influence agent is eliciting a foolishly consistent response from us, our instincts tell us that something is wrong and that we are being pressured to agree to a request with which we do not really want to comply. If this happens to you, we recommend thm you inform the influence agent that complying with his or her request would be a foolish type of consistency that represents behavior in which you choose not to engage. A good way to judge the situation is to ask yourself whether you would make the same commitment if you could go back in time and make the initial choice knowing what you now know. If the answer is no. we suggest you refuse to comply with the influence agent's request.
Reciprocity-Based Sequential Request Compliance Tactics While walking to class. have you ever been stopped by a friendly individual who offered you a free T-shirt or teddy bear if you would take the time to fill out a credit card application? Or, have you ever received free chocolate to entice you into a candy store? In either situation most of us have at some point in our lives found ourselves successfully influenced. We are influenced because we feel the need to "repay" the individual who has given us a "frec" gift. This feeling stems from the norm oj reciprocity, which states that it is appropriate for individuals to return favors (Gouldner. 1960). Although reciprocity is a prosocial behavior. influence practitioners can easily take adv31l1age of this norm to induce compliance. The next two sequential request compliance tactics, the door-in-the-face and the that's-not-all-technique, work by using the norm of reciprocity against us. Each of these tactics is reviewed in the next section, while table 12.3 presents the basic mechanism of each.
The Door-in-the-Face Technique Have you ever answered a knock on the door to find a salcsperson asking you to buy 10 two-year magazine subscriptions for a total of over $350? For most of us, that is far loa much money. Imagine you refused the request to buy the magazine subscriptions and the salesperson followed up by asking you to purchase a single two-year magazine subscription. This modified request seems so much more reasonable by comparison that you agree. As you shut your door, you immediately ask yourself why you just purchased a two-year subscription to Modem Fishing when you do not even fish. If you have ever had an encounter like this, you have experienced the door-in-the-faee (DITF) technique. In order for the DITF tactic to be successful, the influence practitioner has to come up with a request that is so large that most people would not even consider agreeing with
doing volunteer work.
218 TABLE 12.3
Part III • Seeking {lmi Resistin8 Compliance: Srrategies and TaClic\
Reciprocity-Based Sequelllial Request Compliallce Tactics Initial Request
Tactic
Second Request How Reciprocity Is
How Reciprocity Is Initially Created Door-in-the-facc
The target rejects an unrea ...onably large requc\t from the
innul!nce agent.
That's-not-all
The innuencc agcnt offer, to makc a deal with the target.
Used to Trap the Target into a uss Example
Desirable Outcome
Example
Asking the target to spend five hours a week for the next two years do ing volun tee r work.
The inOuence agent concede ... by making a . . econtl .... maller rcque . . t. and the target feels normative pressure 10 reciprocate and agree to the conce ...... ioll.
A,king the target to ... pend olle a ft ernoon doi ng vo luntee r work.
Offering to ...ell
Before the target has a c hance 10 re ... pond .
Adding ~l nc\\ CD changer to the offer without c hang ing the price.
...omeone a car for a certain price.
the innuence agent sweetens the dea l by offering more for the ... ame prict'!.
it. Once the targe t rejects {he inordinalely large reque~t, the influe nce practitioner concedes and asks the target 10 agree 10 a s maller reque st. In thi s case, more people w ill agree with the second req uest than if they had been presenled with that req uesl initially . Most researchers believe that the co ncession on Ihe part of Ihe innuence agent is essentiul for the DITF tacti c to work. That is becausc the target fee ls nonnative pressure to reciprocate the co ncess ion of Ihe target. Cialdini, Cacio ppo. Basscll. and Miller published the initial demonslration of the DITF tech nique in 1975. To exami ne the effeCliveness of the tec hnique. these ex perimenters approached student.s on a college ca mpus and asked Ihe m (Q vo lunteer to chaperone ju ve nil e delinquents on a day long trip to Ihe LOa on an upco ming Saturday. All participa nt s received this req ues t. For participanls in the DITF condilion. thi s see min g ly large request followed an even larger one. Pmtic ipanl s in the DITF condition were first asked to vo lunteer to cou nsel juvenile delinquents for two hours a week for two years. Once they refused to compl y with thi s initial reques l, targets were then asked to volunteer for the trip 10 Ihe zoo. As fi gure 12.3 shows. Ihe resulls , upp0rl ed Ihe elTeclivencss of Ihe DlTF: 50 percent of participants w ho received the requeM as pari of the DITF technique agreed to chaperone the trip to the LOa, co mpared 10 17 perce nt of participants who received only Ihe targe l request. In anoth er DITF study. C ialdini a nd Ascallli ( 1976) demonstrated that Ihi s technique wm. more successful at gelling indi vi dual s both to agree to donale and aClllally donate a pint of blood. Thai is. after refusi ng a request to donate a unit of blood every six wee ks for IwO years. participants who the n received a request to donate just one pint of blood were
Chapter 12 • Seqflemial Request Compliallce Tactics
219
1000/0 ,----------------, 80%
•
Control (participants asked only to chaperone juvenile delinquents at the zoo)
•
Door-in-the-Face (participants first asked to spend two hours a week volunteering with delinquents)
50%
60% 40% 20%
0% Percentage of participants who agreed to act as a chaperone on a trip to the zoo
FIGURE 12.3
Compliance with the Target Request in the First Door-in-the·Face Study
SOl/ree: Cialdini ct
ilL. 1975.
more likely to agree than were those who had received only the request for one pint of blood. In addition, of those participants who agreed to donate blood. those in the DITF condition were more likely to follow through on their agreement and actually donate the blood. Thus, the DlTF tactic ha!) been shown to produce both verbal and behavioral compliance. Another DITF study is of particular interest to students (Hami, Mohr. & Hosey. 1980). In it professors were the research participants. Initially, the professors were asked to spend two hours a week all term tutoring a student. The target request was to spend 15 to 20 minutes with the student. A~ preoicLeli, compliance was much greater after the professors heard and refused the initial request: 79 percent agreed to spend 15 to 20 minutes with the student. compared with 59 percent of the control participants. Thus. one tip for students is that if you want some extra help or attention from your professor, ask for more time or help than you think you need. If your professor refuses. retreat to requesting ajust adequate amount of help.
The That's-Not-All Technique While flipping through the channels watr.;hing late-night television, have you ever stumbled across a commercial for a knife that slices wood and soda cans and is still sharp enough to slice a tomato? After demonstrating the knife's prowess, the announcer mentions that it is available for an all-time low price of 519.99. Then. after a short pause, the announcer says: "But wait. that's not all. For the same low price you'll also get a paring knife, a set of steak knives. and a knife sharpener." Getting additional items for the same price makes this seem like quite a bargain. At the end of the commercial. you may feel compelled to pick up the phone and order now. This is a cla~~ic example of the Ihat's-lIolal/ technique. This technique works as follows: The inOuence practitioner offer., the target a product for a specific price and then sweetens the deal by throwing in a "free" gift or lowering
220
Part III • S('ekll/.~ alld R('.\;\lillg COIIII'/;(llIn':
SII"lIIt'V,II'.\ and I(/("Iin
the price. It "ork\ becam.e the anchor point i... adju<.,{ed. An anchor point i ... an initial \· ..lIuc that c ... tabli..,hcs the standard price of an iteill. tclling u\ what the item i ... worth. Once the initial price point \cl.., thc value of an item. the free extra items or the lowered price make the ..,ccond offer ..,CCIll like a bargain in contn.. ..,t to the first. Reciprocit) is al ... o a factor here. By Imvering the pricc or providing additional item ... for the :-.amc price. the influence practitioner appears to be doing the target a fa\Of. increasing the normative pres..,llre on the target to reciprocate by making the purcha ... e. In the initial demonstration of thi ... tac ti c. Burger (1986) gave re ... earch participant ... the opportunity to buy cupcake ... at a univer ... ity c<.lmpu:-. bake sale. They were offered a cupcakc for a 'pecifie pri ce. a:-.kcd to "wait a ... econd:· and then lold that the pricc abo inclmkd a ... mall bag of cookie .... The ...e participant... turned out to be significantly more lil..cl) to buy the cupcake and cookie:-. pacl..age than participant ... who \\'ere told at the out... et thaI the price \\a:-. for the cupcake and cookie:-. package. HO\l,.'cver. Burger. Reed. DeCe ... arc. Rauller. and Rozilil" (1999) dClllolhtrat~d that the that's-not-aI1tcchnique can boomcf
uropped from SIO to $3: othe" were told that it hau been reduced rrom S5 to S3. Compared 10 a condition in which participant' heard only about a 53 donation. the manipulation from S5 to $3 produccd a s ignificant increa ... e in donation .... Howe\er. thc :-.witch from $10 to S3 caul"ed participant ... to gin: ... ignilicanlly Ie ... :-. money than tho ... e in thc control condition. The re"earcher!-l rea ... oncd that becau ... e the initial reque . . t wa ...... 0 large. it led tn an imlllcdiate rejection of the requc"t. The experilllenter:-. were therefore unable to alter the participant ... anchor point. and the effccti\enc ...... of the tl1at' ... -not-o.1ll manipulation wa, lost.
De/elise Agaillst Reciprocity Tactics What can we do when a favor or conct.:\l"iol1 turn ... Ollt not to be the boon it was initially percei\'eu to be. but inqead merely the initial ... tep in the DITF or the that""'-l1ot-a ll tech nique ? The 1110 ... t extreme rc\pon"c \\'ould be to reject all favor, and conccs\ion ... that come our way to protect our:-.elve~ from thc potentialmi"u:-.e of the norm of reciprocity, While that approach may 'illcce!-l<..full) protect againl"t thi, t) pe of compliance tactic. it ma) abo end up hurting the fceling~ or people \\ho hone ... tly mcallllO do u... a favor or mal..e a conce~:-.ion. We therdore recommend a more moderate re"'poll\c: to rt.:frame the favor or COIlcc ... :-.iol1 a ... the trick that it actually i~ rather than ... ec it a~ a true act of gencro<.,ity. Once people recogni7c the trick for "hat it i". the normative pn:" ... ure to comply by reciprocating with a cOl1ce:-.:-.ion or favor J"rom their own cnd will di:-. ... ipate (Cialdini. 2(01). The author~ u:-.c thi" dcfen"c c\'ery time they get "frec" addn::-.:-. lubeb in the mail along with a reque ... t for a charitable donation. They rccogni/c Ihat the label:-. arc actually not a gift at all but merely a ~ale:-. lOol. While the) mayor may not choo"e 10 donale money to the charity. thb choice i... ba:-.ed on their opinion of the charitable cause. not the result of any pres:-.ure to comply and reciprocatc becaul"c they recei\cd label:-. at no co ... t. Although thi~ cxample i... not a ,equential request tactic. the same defen,c technique can be applied to situation ... in which the intluence practitioner i... u,ing the.! DITf- or the that'\-not-alltechnique.
Chapter 12 • Sequenfial Requesf Compliance Tacfics
221
Conclusion Sequential request compliance tactics require two steps to produce a successful influence attempt. The majority of influe nce tactics can be categorized into one or the six principles of influence: scarcity, reciprocity, co nsistency and comm itment, authority, social proof, and simi larity or liking. Most sequential request compliance procedures fall into one of two categori es of influe nce: commitment and consistency or reciprocity. There are three main com mitment-based procedures-the low-ball approach, the bait-and-switch technique, and the foot-in-the-door effect-and two reciprocity procedures-the door-in-theface tactic and the that's-not-all technique. Although more research in this area is needed. social scientists are slow ly learning more and more about the factors that compel us to say yes to a request when our initial inclination is to say no.
References___________________ _ _ _ _ _ _ __ Beaman. A. L.. Cole. C. M .. Preston. M .. Klentl, B.. & Steblay. N. M. (1983). Fifteen years of the foot-inthe-door research: A met'l-analysis. Per.wJl/ality al/d Socia! Psycholog\' BI/lletil/. 9.181 - 196. Burger. J. M. (1986). Increasing compliance by improving the deal: The that'~-not-all technique. JOl/mal of Persol/ality al/d Social P.I)'cho!ogy, 51. 277-283. Burger. J. M. (1999). The root-in-the-door comp liance procedure: A mu1tiple-proces~ analy~i~ and review. Persol/ality and SO£'ia! PsycllOlog\" Rel'iew. 3.303- 325. Burger. J. M .. & Caldwell. D. F. (20CH). The effects of l/uJl/etary illCf'lItil'es and labelillR all fill' .!oor-illfhe-door effect: EI'idellce for a self-perception proce.~.\·. Manuscript submitted for publication. Burger. J. M .. & Cornelius. T. (2001). Rai.~il1g fhe price of agreemel/t: ComparillR rhe illl('l"l"IIPlioll and low-hall compliance pmcedllres. Manuscript submitted for pllblication. l3urger. J. M .. & Guadagno. R. E. (in press). Self-concepi clarity and the foot-in-the-door procedure. Basic and Applied Socia! Psychology. Burger. J. M .. & Petty, R. E .• (1981). The low-ball compliance technique: Task or person commitment? Journal of Persollalily al/d Social Psychology, 40. 492-500. Burger. J. M., Reed. M., DeCesare. K.. Rauner. S .. & Rozilis. J. (1999). The effects of initial requcst size on compliancc: More about the that's not <111 technique. Basic lIl/d Applied Social Psychology. 21 (3),243- 249. Cialdini. R. B. (200 I). 11/j7/1ence: Science and practice (4th cd.). New York: IlarperCollins. Cialdini. R. B.. & Ascanti. K. (1976). Test of a concession procedufC for inducing verbal. behavioral. and further compliance with a request to give blood. Jourtlal oj Ar1plied Psychology. 61. 295- 300. Cialdini. R. B., Cacioppo. J. T .. Bassett. R .. & Miller. J. A. (1978). Low-ball procedure for proJucing compliance: Commillneni then cost. JOIIl"llal oj Persol/ality and Social P.IYc/W!OXY, 36. 463--476. Cialdini. R. B .. & Trost, M. R.. (1998) Social inOuence: Social norms. conformity. and compliance. In D. T. Gilbert and S. T. Fiske (Eds.). The handl}()ok of socia/ p.\·ychology. 1'0/. 2 (~th ed .. PI'. 15 I- I 92). Boston: McGraw-Hill. Ci
222
Pari III • Seeking and Res;'~til/s Compliance: Srraregies wul T(lcrin
Gouldncr. A W (!960). The norm 01 reciprodt)' : A prdimm.H) ,-talcml'nl. Americal/ SouoJog/('a/ Rel'ieu.25,161-178. Guadagno. R. E.• A<;;her. T., Dem~lInc. L. J .• & Cia ldin i. R. B. (l(X)!). When "aying yc.,lc:!t!<. to "aying no: Preference for consiste ncy and the rever.. e foot-in-Ihc-door c lTr.:cl. PerwJIllllity cllld Sm'illl P,\yclwl0KI' Blilletin. 27. 859-867. Han sen. R. A., & Robin'ion. L. M. (1980). Te'iting the effecli vcm: ...... uf ,t ltcrnati vc foot -in-lhc-do(Jf mampulatlon ... Joumal of MarJ..etIllR Rewarch. J7. 359-36-L Harai. H .. Mohr. D .. & Ho<.ey. K. (1980). Facult) hc1pfulne... ... to ... lUden",: A compari.,nn of compliance techniques. Peno"alin'lImJ Sonal Pn'dlOlog\, Blllln/n. 6. 373 )'77 Joule. R. V.. Gouilloux. F.. & Weber. F. (1989). The lure : A new compliance procedure . Journal o/"Social P.n'choJog\'.129.741 -749.
13 A Review of Fear-Appeal Effects Hyunyi Cho and Kim Witte
Since the beginning of recorded time and before. fear has been a powerful motivator. For
example. the Bible abounds with fear appeals: " ... but you must not cat from the tree of
the knowledge of good and evil. for when yo u eat of it you will surely die" (Genesis 2: 17), Death in thi s passage is a key constru ct of the fear appeal. represe nting what researchers have termed "unfavorable con,eque nces" (Hovland. Janis, & Kelley. 1953), "magni· tude of noxiousness" (Rogers. 1975). or "severity" (Witte. 1992b). By describing the terrible consequences that may happen unless people do what the message recommends. fear appeals attempt to per!\uade them to change their attitudes and behaviors (Witte,
I992b). The use of fear in persuasive appeals. however. is fraught with accounts of unsuccessfu l resu lt s and even adverse effects if lIsed incorrectly (see Hale & Dillard. 1995). For example. despite the evocation of death. Eve ate the fruit from the tree , marking perhaps the earliest incidence of an unsuccess ful fear appeal. Failed fear·appeal attempts can be unfortunate when you consider th e consequences. Fortunately. current theorie s show how
to develop fear appeals that work as we ll as how to avoid developing fear appeals that fail or even backfire (Rogers. 1983: Witte, 1992b. 1998: Witte. Meyer. & Martell, 2(01), For example, the Extended Parallel Process Mode l (EPPM). a fear appeal theory that integrates 40 years of research. clearly distinguishes the conditions under which fear appea ls arc like ly to succeed from those under which fear appeals are prone to fail (Witte.
1992b, 1998; Witte et al.. 200 I). This theory will serve as the guiding framework for this chapter. In addition. this chapter acknowledges that despite the best theoretical advice. no messages produce uniform effects across a spectrum of audiences (McLeod & Becker.
1974). Because communication is a social process by definition (Berlo, 1960; Shepherd, 1999). the effects of fear appeals may differ from one audience to another as each processes messages differently. Fear is a universal emotion. but different audiences may hold different perception s about what is scury and how scary a message is. The understanding of such audience differences is pi votal to the development of effective fear appeals. There·
223
224
Part III • St'f!~illg (lml Rl',<;i.\lillg COIIII,/iWIl"t': SlrtHt'gie,\ allll rac1in fore. the purpose of thi ... chaplcr diver<.;c audience<.;.
I~ 10 n:\'I~\\
the
r~~earch
on
th~
effects of fear appeah on
The Extellded Parallel Process Model Defillitiolls The IwO key conslrue" of Ihe EPPM (Wille. 1992h. 1998: Witte el al.. 2(01) are perceived Ihreat and perceived efficacy. Fir<';l. imagine Ihat you smoke two pach of cigarettes a day and ha\ejust heard an adverti<.;em~11l telling you that cigarettes cau<.;c cancer and thai che\\ ing nicotine gum can help you break your habit. Perceil'ed threat. the fin,t com,tnu.:1 of the EPPM. refer<.; 10 whether you perceive you afe in danger. It sugges .... that YOlI pcrc~iye danger If you belieye thai (I) you arc nllnerable 10 a threat anu (2) Ihal the threat i...... erious. In Ihi ... case. if you cOlhider that your chances of gelling cancer from "'Illo~ing arc high (i.c .. percein'd susceptibility) and that cancer i... a scyere disease (i.c .. pl'rceil'l't/ s{,I'frity). Ihe percei\-ed Ihreat is high. Pen'e;\'ed efficacy. the second con ... lruct in the EPPM. refers to a per... on· ... pen.:epliol1' aboul Ihe re'ponse Ihal i, recommended in Ihe fear appeal. In olher word,. "hal arc your perceptions regarding the ni<.:otin~ gum rec:olllmt.:ndation'? Perceived efficacy COI11pri ... e ... two elements: <.;elf-efficacy and re ... pnn ... e efficacy. Per('eil'ed se~f-eJJic(ln' refers to beliefs about whether or not you believe you can perform the recommended rc"'pol1 ...e. whereasperceil'ed respo"se effic(U'\' refers to your beliefs ubulit whether or not the recommended response work .... For example. are you capable of chewing nicotine gum? If so. self-efficacy is high. Do you believe that the gum will breal.. your habit'! If so. respom,e efficacy is high.
Appraisals lIlld Processes The EPPM \uggests that upon expo,U/'c 10 a fcar appeal. 'JUdience ... will appraise it in one of the IWO ways and then respond in one of the thn:e way ... a ... a re ... ult of Ihe apprai ... al (see figure 13.1). Fir ... 1. individual ... appraise the threat. Is it severe'! Are they ... usceptihlc to II'! The greater the perceived threat. the grl!ater the rnoli,
Message Processing Appra isals
External Stimul i
Protection Motivation
PERCE IVED EFFICACY MESSA GE COM PONEN TS Self-Efficacy Response Efficacy Susceplibllity Severity
f----'
(Self-E fficacy, Response Efficacy)
<-- ____
feedback loop
(Susceplibility, Seventy)
No Threat Perceived (No Response)
T T
N N
VI
The Ex/elided Parallel Process Model
Danger·
Control Process
FEAR
~ Defensive MotIvation
Individual DIfferences
FIGUR E 13.1
Message Acceptance
~ 1 -:1
PERCEIVED THREAT
4
Process
Outcomes
Message
Rejection
Fear· Control Process
226
Pan III • Seeking afld Resisting Compliance: Strategies and Tactics
If, on the other hand. individuals doubt their ability to carry out the recommended response effectively, or don't believe that the recommended response works, then they give up on trying to control the danger and instead attempt only to control their fear. Fear control is a nonconstructive or maladaptive response. In the fear control process, individuals engage in psychological defense mechanisms like defensive avoidance (avoiding thoughts about the threat), denial, or reactance. [n this condition, the audience's responses to fear appeals are focused on controlling the unpleasant arousal of fear (hence the name "fear control process"). It is important to distinguish between no effect, which occurs when no threat is perceived, and fear control effects, which occur when perceived threat is high and perceived efficacy is low. Both result in rejection of the fear appeals' recommended responses; but the former calls for stronger fear appeals whereas the latter calls for the avoidance of fear appeals, that is, efficacy messages only.
The Role of Individual Differellces As shown above, th e EPPM elucidates how audiences with different levels of perceived threat and efficacy may process a fear-appeal message differently and subsequently engage in either danger-control or fear-control processes. The EPPM delineates the role of individual differences in fear-appeal message processing, suggesting that individual differences affect one's perceptions of the message. For example, individual differences inOuence how a threat and a recommended response are perceived. A naturally anxious person may perceive a threat to be greater than it really is, whereas a naturally pessimistic person may perceive the recommended response to be less efficacious than it really is. Once threat and efficacy are appraised, however, the same theoretical tenets apply. Specifically, threat motivates action, any kind of action, and perceived efficacy determines whether or not individuals engage in danger control or fear control. In summary, individual differences can and do affect how a message is processed. Therefore, fear appeals must be designed with the nature and perceptions of the audience in mind, as the same fear appeal may have different levels of effectiveness and different kinds of effects among different audiences.
Fear-Appeal Effects: Message Topics The EPPM has been tested in a variety of persuasion contexts, including fear appeal messages advocating preventive behavior against AIDS (Witte, 1992a), genital warts (Witte, Berkowitz. Cameron, & McKeon, 1998), tooth decay (Berkowitz, 1998), bulimia (Smalec, 1996), skin cancer (Stephenson & Witte, 1998), electromagnetic fields (McMahan, Witte, & Meyer. 1998). tractor-related injuries (Witte, Peterson, Vallabhan, Stephenson, Plugge, Givens, Todd, Becktold, Hyde, & Jarrett, 1993), and others. Regardless of the study topic, messages that made individuals feel seriously at risk and made them believe that they could do something that would effectively avert the threat promoted the most attitude. intention , or behavior change (Witte & Allen. 2000). These results suggest that the most persuasive messages are those that promote perceptions of both high threat and high efficacy.
Chapter 13 • A Rel'ieu of Fe(lr~Appe(l1 £ffecfs
227
Fear-Appeal Effects: Audience Characteristics Age One demographic factor that has been shown to influence the effectiveness of fcar appeals is the age of the audience. A number of meta-analyse!! of fear appeals have been conducted (Bo,tcr & Mongeau. 1984; Mongeau. 1998: Sutton. 1982: Witte & Allen. 2000). These meta-analyses have consistently reported that such appeals are more effective for older audience!>. than for younger audiences. suggesting that the interaction of age with the perceived fear influences the attitude change. That younger audiences tend to believe they arc invulnerable to possibly seriou~ consequences of their behavior and environmental cvents is well documentcd (e.g., Weinstein. 1980. 1982). Researchers attribute this tendency to younger audiences' lack of experience and exposure to unfavorable life events relative to o lder audiences (e.g .. Mongeau. 1998). Witte and colleagues (2000) suggested that the key 10 effective fear appeals in young audiences is finding out what they truly find threatening or scary. Death or terminal illnesses may not be scary to a teenager. but losing one's friend s or facing changes in physical appearance could be. To create effective fear appeals for young audiences. message designers need to identify the kind of threats that the young can relate to. For example. adolescents may vie\\' the prospect of developing lung cancer as a result of smoking as quite remote. A more immediate or pertinent threat may be a better vehicle to persuade them. Hansen and Malotte (1986) noted that although adolescent smokers tended to deny the ultimate severity of the danger of smoking, they readily acknowleJged that as smokers they were at risk of being out of breath during exercise. presllmably becau~e of their own experience. Schoenbachler and Whinier (1996) found that for adolescents, the social threat of rejection by peers was more effective than physical threats in a fear appeal. persuading them to change their attitudes and intentions townrd drug u ... e. Therefore. it is critical to find out exactly what is threatening to a target audience about a given topic. The perceived harm may be physical, social. economic. spiritual. and so on.
Allxiety Some people are more chronically anxious than others. Trait anxiety refers to "one's charlevel of anxiousness in response to a threat thai leads one 10 react in either an avoidant or coping/sensitizing manner" (Wille & Morrison. 2000. p. 6). The very nature of the trait. characteristic anxiousness. has invited fear-appeal researchers' attention, because individuals with high levels of the trait may be particularly likely to be adverse ly affected by fear appeals (e.g .. Boster & Mongeau, 1984: Dabbs & Leventhal. 1966: Dziokonsk & Weber. 1977; Goldstein. 1959; Hill & Gardner. 1980; Jepson & Chaiken. 1990: Wine & Morrison, 2000). Research to date reports th:.ll audiences with different levels of trait anx ie ty respond differently to fear appeals. However, the pattern of results remains inconsistent: Some studies have found that people with high rather than low levels of trait anxiety are persuaded by strong fear appeals. However. other research has shown that it is lowrather than high-anxiety persons who change their atti tud es when exposed to strong fear appeals (Witte & Morrison. 2000). acteri~tic
228
Part III • Seekillg and Resistillg Compliance: SrlYlre!;ies (lnd Tactics
One notable factor explains the inconsistency. Specifically, the terms used to describe individuals with high versus low trait anxiety have been mi slabeled (Wi tte & Morrison, 2000). For instance. the terms that have been used to label th e construct have included individuals high and low in trait anxiety. repressors/sensitizers, and cope rs/ avoiders. ' Accordi ng to Witte and Morrison. researchers have long mistakenly gro uped individuals high in trait anxiety with repressors/avoiders and indi viduals low in trait anxiety with sensitizers/copers (sec Witte & Morrison. 2000, p. 8 for a full review). As a result, with this categorization. the pattern of results appears consistent: individuals high in trait anxiety and repressors/avoiders are less persuaded. whereas indi vid uals low in anxiety individuals or sensitizers/copers arc more persuaded as perceived fear increases. However. Witte and Morrison (2000) pointed out that according to the scales used to measure these constructs. high-anxiety individuals are in fact se nsitizers/copers and that low-anxiety individuals are repressors/avoiders. With this clarifi cation, it became c lear that there is no consistent pattern regarding trait anxiety and fear appeals' effec ts. Overall, two major hypotheses have been tested as explanat ions for the influence of fear appeals on high- versus low-anxiety individuals. First. meta-analyses of fear-appeal studies (Boster & Mongeau. 1984; Mongeau, 1994) suggested that an interaction between fear and trait anxiety influences attitude. intentions. and behavior. to th e exte nt that hi ghanxiety individuals were not influenced by stron g fear appeals, whereas low-anxiety indi viduals were. Second, Witte and Morrison (2000), using the EP PM. hypothesized that trait anxiety (I) directly intluences threat and efficacy perceptions and (2) indirectly influences fear-appeal outcomes as mediated by perceptions of threat and efficacy. Studies have been relatively unsupportive of any of these explanations. For example. the results of Witte and Morrison's (2000) study indicated that trait anxiety is positively associated with perceived threat and efficacy. However. the fact that trait anxiety did not corre late with attitude. intentions. or behavior was consiste nt with th e results of earl ier research (e.g .. DLiokonski & Weber. 1977: Goldstein. 1959: Wheatley & Oshikawa. 1970). Nor did trait anxiety intluence fear-control responses of message derogation and perceived manipulation. Witte and Morrison (2000) a lso tested the other two proposed relationships described above (direct effects and interaction effects) and found no support for either of these hypotheses. Overall, it appears that the designers o f fearappeal messages do not have to worry about particularly anxious or particularly repressive audiences. as these variables appear to have little inlluence on how fear appeals affecl people.
Fatalism The possible role of fatalism in processing and responding to fear-appeal messages was first noted by Casey ( 1995). According to Casey. fatalism refers to "the belief that death is imminent and unavoidable. a fulfillment of a socia ll y-constructed reality" (p. 20). In oth er words, we're all going to die someday and we can't do mllch about it. Although the effects of fear appeals on audiences with a fatalistic perspective have not been specifical ly investigated, other evidence suggests that this trait may well be a limiting factor. For example. fatalism has been positively associated with attitudes and behaviors that increased the risk of occupat ional hazards (e.g .. Such man, 1967). traffic accidents
Chapler 13 • A Rl'\";t'\\ oj Fear-Appeal tl/t'ch
229
(e.g .. Kouabenan. 1998 ). and contri.lc tion of HIV/AIDS (c.g .. Hardeman. Pierro. & Mannetti . 1997; Kalichman, Kelly. Morgan. & Rompa. 1997). Some rescarch ha."> shown that ce rtain ethnic groups may hold grea ter fatalistic tendencies than OIhcrs and thus may be Ic!-.:--. likely to engage in danger control procer.,:-,c!-o for threats such a:--. cance r (e.g .. Domino. Fmgoso. & Morcilo. 199 1: Straughan & Scow. 1995). Indeed. jf a person considers such dangers are imminent and unavoidable. why bothcr trying to control them? Research is needed to as ...e:--.s whether fatalism moderates response!-o to fear appeals. Studies to date suggest that understanding the ro le of perceived response efficacy may be important in persuading fatalistic audiences wit h ft.!ar appeals. For example. in focus group:--. of Chine:--.c women regarding their attitudt.!:--. toward mammography. the results revealed that "faith in medicine" Illay motivate them to i.Idopt mammography. while fatalism works as a barrier to sllch adoption (S traug han & Scow. 1995). Special messages that pro\ide hope and ... trong self-efficacy perceptions 111a) also be needed to counteract the fatalistic pcn,pcclive. Finally. !-oocielal and MrUl:tural inequities mUM he addre..,~ed to prevent fatalistic respon!o!cs. For example . if COlldOI11~ an.:: promoted in rural Africa, they mu~t bc readily avai lable at a rea:-.onahlc cos I. Similarly. if immuniallions an! promoted in poor. cri me- ridden areas. l ow-co~t clinics mu~t he made :--.afe ly available to clicnt~.
Reactallce PSycllO/O}:iclll reactance (Brehm. 1966: Brehm & Brt.!hm. 1981). a proce ...... that occurs when individuals perceive threats to freedom and freedom of choice. typically results in a boomerang effect whereby people react in a manner oppo:--. il c to the me ... sagc·~ rt.!conllnendatioll\\. For example, a parent might try to scare a child into riding more slowly o n her/his bicycle by say in g. "You're going to knock your teeth out." This might. however, ca use the c hild to ride even faster. Reactance wa.., originally COl1ccpltll.l li Led as a state. but in recent research it is treated both a ... a Lrait and a state (Brehm & Brehm. 1981: Beutler. 1979: Fran~. Jac~"lI1 - Walker. Mar~'. Van Egeren . Loop. & Ol,on . 1998: Jahn & Lichstein, 1980: Rohrbaugh. Tennen. Pre". & White. 198 I). Rc:--.ean.:hers have found that trait reactance is significantly associated with a host of per!-.onality variables that may predict res istance to persuasion (e.g., Dowd, Milne, & Wise. 1991: Dowd & Sander,. 1994: Dowd. Wallbrown. & Yesenosky. 1994). Specifica ll y. trait reactance was "po:--.itively associa ted with \uch personality variables as autonomy. dominance. and independence. and negatively as\ociated with such variables as affiliation. tolerance. inter~st in making a favorable impression, and nurturancevariable; with implications for noncompliance" (Seibel & Dowd. 1999. p. 374). The perceived threal construct in fear appea ls is different from the threat Con\Lruct in Reactam:e Theory. The former refers to perceived \\ul,ccptib ility and severi ty, whereas the latter refers to perceived threats to freedom. However. rear appeals may still be perceived as a form of a threat 10 behavioral freedom. Fear appeab commonly include behavioral recommendations :--. uch a ... dental hygiene practice . . (c.g .. Jani:--. & Fe~hbach, 1953). safe ~cx (e.g .. Witte. 1992b). sl11o"-ing cessalion (c.g .. Kleinot & Rogers. 1982: Rogers & Mev. born. 1976). and so on. all of \\ hich. to a \'arying degree. limit individuals' freedom (Cho, 20(0).
230
Part III • Seeking and Resistillg Compliallce: Stratexies and TaClics
Cho (2000) examined the effects of fear appeals in persuading both high- and lowreactance individuals to engage in behavior to prevent skin cancer. The results suggested that compared with low-reactance individuals. high-reactance individuals evaluate fearappeal messages unfavorably by derogating them as a distortion of truth or mere manipulation. However. while high trait reactance may promote unfavorable evaluations. it does not appear to significantly affect compliance with fear-appeal recommendations in terms of attitudes, intentions. and behavior. In short, people high in trait reactance may not like fear appeals. but they arc persuaded by them nonetheless. Cho (2000) conjectured that there may be two possible reasons for these results. First, because behavioral recommendations are a key feature of fear appeals, fear appeals have thc potcntial to elicit the perception of threatening the audience's behavioral freedom. However. according to Brehm (1966), one factor that determines the boomerang effects of reactance is the importance of the threatened freedom. For example. while a fear-appeal message recommending the use of sunscreen before going out in the sun in essence asks people to compromise their freedom. the reduction in freedom may be considered less important than the corresponding reduction in the risk of skin cancer. Or. the compromised freedom may be considered less serious than the risk of developing skin cancer. A second premise of Reactance Theory is that boomerang effects of persuasion occur when individuals perceive that they have choices (Brehm. 1966). If a fear-appeal message clearly conveys to an audience that ~kin cancer is a serious threat and that a particular audience is vulnerable to it, members of that audience may no longer really consider going out in the sun without putting on sunscreen as a choice. Therefore. highly reactant individuals may not exhibit counterproductive attitudes, intentions, or behavior in response to the fear appeal. A similar line of reasoning and conclusion was drawn by Bushman (1998). In his study examining the effects of warning labels on full-, reduced-, and non-fat food prodlIctS. Bushman found that although subjects preferred full-fat cream cheese, they actually chose to eat reduced-fm cream cheese, because they believed that fatty foods increase a host of health risks. More research needs to be done on the issue of reactance. because the two studies reviewed above may not have involved a great loss of behavioral freedom and therefore may have failed to produce a great deal of reactance. It eould be important to examine highly reactant individuals' response~ to fear appeals advocating the cessation of addictive behavior such as drinking or smoking.
Sensation Seeking Perhaps the individual difference variable with the most significant implications for the effects of fear appeals is the trait of sensation seeking. One key aspect of sensation seeking is defiance of and a positive predisposition toward risky behaviors. Specifically, sensation seeking involves a willingness to experience consequences in order to experience sensation (Zuckerman. 1988. 1994). And sensation-arousing activities, such as drinking and drug use, tend to involve risk (Donohew, Lorch, & Palmgreen. 1991, 1998). Indeed. research to date indicates that the sensation seeker poses a special challenge to persuasion attempts using fear appeals.
Chapter 13 • A Review of Fear-Appeal Effecis
231
\Vill~ and Morrison (1995a) firM inveMigated the effectiveness of fear appeals on high- and low-sensation !o.cekers. The re!o,uirs of the study, advocating safe sex among high I.,chool and juvenile detention youth, indicated that high-sensation seekers were not persuaded by either high- or low-threat messages. whereas low-sensation seekers were per,uaded by both high- and low-threat messages. Wille and Morrison (1995.) suspected that although high-threat message!>. may have captured high-sensation seekers' attention, they failed to persuade them. It appeared that adolescents' perceptions of invulnerability to risk might have worked as a barrier to persuasion. In contrast. Berkowitz', (1998) study. in which she advocated dental hygiene practices among high- and low-sensation . . eekcrs who were college students, found that highruther than low-threat message!o, were more effective for both high- and low-sensation seckers.' How can we explain the different results? Wille and Morrison's (1995a) study indicated that fear appeals have limited utility for persuading high-sensation seekers. However. Berkowitz's (1998) study sugges" that the audience's sensation-seeking level may not influcnce the effects of a fear appeal. The kind of risky and recommended behaviors presented in a fear-appeal message may playa role. Berkowitz noted. "Sensation seeking may affect attitudes, intentions, and behavior!>. only when the activity ha!>. high sensation value" (p. 71). Additional research is needed on topics with a high sensation value like sex, drugs. and driving fast to assess if and whcn fear appeals can effectively persuade high-sensation seekers. Current research ~uggest!o. that without offering an equally appealing alternative to the risky behavior. the fear appeal may fail.
Stages of Behavior Challge Fem appeal~ have been used in a wide range of health communication contexts to bring about behavioral changes associated with risk prevention (Freimuth, Hammond, Edgar, & Monahan. 1<)<)0). Most recently. researchers noted that behavioral change is not a one time event. but rutller a process involving multiple stages (Prochaska, DiClemente, & Norcross, 1992: Weill'tein. 1988). Specifically, Procha,ka and colleagues (1992) asserted that individuals typically go through five stages before they fully adopt and maintain a healthy behavior: precontemplation. contemplation, preparation. action, and maintenance. For example. a smoker who has no plan to quit smoking in the near future is in the precontemplat ion stage. However. a smoker who i~ thinking about quitting smoking within the next one 10 :-.ix months i~ in the contemplation/preparation stage. At the action stage, a person ha:-. refrained from smoking for more than a month, but less than six months. A person who has not ~moked for more than six months is in the maintenance stage. Tht; re!o.ult~ of research on fear appeals and risk communication indicate that an audience's behavioral status may be an important factor to consider in fear-appeal message design. because individuals with different behavioral statuses will respond to a fear-appeal message differently. For example, increa~ing susceptibility for individuals who have already been engaging in ri~ky behaviors may produce deleterious effects. A longitudinal study found that some homosexual men exhibited "obsessive/compulsive behavior, social role il11painnenl. and more imru!>.ivc worries about AIDS" (p. 231) as their perceived susceptibility to AIDS increased (Joseph. Montgomery. Emmons, Kirscht, Kessler, Ostrow, Wortman. O·Brien. Eller, & Eshleman, 1987). In particular, it was suspected that for
232
Pan III • Seeking alld Resisting Compliallce: S1r(lfegie.~ and Tacfics
individuals who have been engaging in risky behaviors. "a sense of pcr~onal risk of the development of HIV infection of AIDS is not beneficially influencing behavior" (Montgomery. Joseph. Becker. Ostrow. Kessler. & Kirscht. 1989. p. 321). Similarly. adults who had low-quality diets were more likely to respond to the threat of food safety with fear-control responses such as defensive avoidance and fatalism than with danger-control response" of rational problem solving (Schafer. Schafer, Bultena. & Hoiberg. 1993). Somewhat similarly, in response to a message presenting heart di~ease risk and the role of exercise in heart health, adolescents who had engaged in exercise indicated greater danger-control responses than those who had not engaged in exercise (Fru in. Pratt. & Owen. 1991). Thus far, a pattern of results has emerged: For the audience who has engaged in the recommended response. perceived threat may create danger-control responses. Conversely. for the audience who has not engaged in the recommended response but has engaged in the risk behavior. an increased perceived threat Illay produce fear-control responses. From the EPPM standpoint. individuals with different behavioral statuses respond to fear-appeal messages differently because their perceived efficacy levels are different (Cho. 1999). Specifically. individuals in later stages of behavioral change would have strong selfefficacy perceptions because they have "ctually performed preventive behavior and therefore do not doubt their ability to do ~o. Those in earlier stages of change. however, would not be so sure of their ability to perform the behavior. By the same token, those in the later stages of change would have strong response efficacy perceptions. because by e ngaging in preventive behavior they have experienced benefits such as improved health. Those in earlier stages. however. have yet 10 experience this benetit. What can be suspected regarding Montgomery and colleagues' (1989) study of an at-risk population is that their increasing susceptibility level was not matched by an increasing efficacy level. Because they had engaged in risky behaviors. the individuals may have believed that preventive behavior was too late to be effective (i.e .. low perceived response efficacy) or that they were unable to change their long-held behavioral habit (i.e .. low perceived self-efficacy). With this premise. Cho (1999) examined unintended. adverse responses to fear appeals among individuals in different stage~ of behavior change delineated by Prochaska and colleagues (1992). Specilically. three types of audience were exposed to fear-appeal messages: those in the precontemplation stage (do not intend to change). those in the contemplation/preparation stage (intend to change). and those in the action/maintenance stage (engaging in preventive behavior). The results showed that individuals in the earlier stages of change exhibited greater likelihood of fear-control responses. such as defen~ive avoidance and fatalism. than those in th e later stages of change. In addition, the magnitude of danger-control responses was greater among individuals who had engaged in preventive behaviors than among those who thought about changing behaviors. who in turn displayed greater danger-control responses than those who had no intention of changing their behaviors. The results illustrate the re lati ve effectiveness of persuasion in reinforcing responses rather than in creating new responses (Klapper, 1960: Stiff. 1994). Also, the results indicated the importance of designing and delivering fear appeals tailored for an audience's unique behavioral status. because individuals who did not intend to change their behavior clearly emphasized fear control over danger control.
Chapter 13 • A Review of Fear-Appeal Effecls
233
Considering the studies reviewed above, future fear-appeal research may need to incorporate behavioral status as an important audience variable. In so doing. fear-appeal research should address the following question: How can fear appeals better persuade those who have not engaged in message recommendations to comply without creating fear-control responses?
Contexts of Exposure Over 40 years ago. Hovland (1959) observed that an examination of the results of experiments and surveys revealed "a marked difference in the picture of communication effects obtained from each" (p. 8). One of the major factors causing such differences, according to Hovland, might be the contexts of exposure to communications. Attempts to understand the impact of contexts of exposure on fear-appeal effects have been made by Horowitz (1969. 1972; Horowitz & Gumenik. 1970) and Berkowitz (\998). In particular, Horowitz's research focused on the role of volunteerism on fearappeal effects. Overall. the results indicated thai individuals who chose to read a fearappeal message had more positive attitudes toward message recommendations than those who did not choose to read one. However. Berkowitz (1998) pointed Ollt that although Horowitz investigated the impact of choice. the choice was whether or not to participate in an experiment, not the kind of message (high or low threat). Berkowitz gave participants the option of choosing either high- or low-threat messages. The results showed that all members of the audience chose to view a low- rather than a high-threat message. However, despite the fact that their choice was unanimously in favor of the low-threat message, participants in Berkowitz's study were still more persuaded by high- rather than low-threat messages. The implications of these studies are twofold. First. in natural settings, fear appeals, perhaps bccause of their arousal of strong, unpleasant emotions, may not be the message of choice from the audience's standpoint. Berkowitz's finding on message choice is even more significant considering that high-sensation-seeking participants of her study chose to view Jow- rather than high-threat messages. which were defined as vivid, intense, and graphic. Sccond. however, the fact remains that the high-threat message was more effective than the low-threat message, despite not being the message of choice. Therefore, feaIappeal researchers may need to develop ways to bring fear appeals to their audience's attClllion in natural contexts. Are effective fear appeals, as a Chinese saying goes, "a good medicine which is bitter to the tongue but effective in curing a disease?" If so, research nceds to focus on ways to inlroJuce selective exposure to fear appeals in naturalistic settings. Future fear-appeal research should also incorporate the factors raised by Hovland (1959), including the impact of competing messages, delayed measurement reflecting post-communication interaction. and fear appeals without sponsorship.
Culture Culture refers to "a set of guidelines (both explicit and implicit) which individuals inherit as members of a particular society"' (Hclman, 1990. p. 2). Therefore, culture determines how members of a community share meanings of verbal and nonverbal messages in their
234
Part III • Seeking and Resisting Compliance: S{({Itegies amI Tactics
interaction (Witte & Morrison, 1995b). Consequently, what is scary in one culture may not necessarily be so in another. For example. Connors (1992) found that for intravenous drug users, being arrested for using drugs might be a more immediate and important threat than contracting HIV/AIDS. As a result. the IV drug user~ may continue to share needles. nOlwithstandiog the risk of HIV infection. For fear appeals to be effective, J11essage~ should address such sociocultural contexts. Witte's (1997) study on teen pregnancy found thai teen mothers' attitlldc~ toward the social consequences of pregnancy may not be all negative. For inner-city teen Illother~. having a baby was viewed positively, but fccling "fat and bloated:' "like they were all alone and that they had no one they could talk to." and "like they lost their friends and their social life and felt extremely isolated" were seen as ~ignificantly negative consequences of getting pregnant (p. 148). Economic threats or other future threats of being a teen mother did not emerge as significant threats to teen mothers. This study suggests that researchers shou ld abandon their preconceived notions of what constitutes a health danger, and find out what a specific culture group actually finds threatening about a health issue. Thus. these findings underscore the importance of clearly understanding an audience's perceived susceptibi lity, an element of the EPPM introduced earlier in thi:-. chapter. A cross-cultural study of fear appeals found that AIDS may e licil a range of threat perceptions. depending on the culture to which the target audience belongs (Wille et al.. 2000). Specifically, audiences with collectivi~tic orientations (group-oriented) werc more influenced by fear appeals threatening the co llect ive such as the family. wh erea~ audience~ with individualistic orientations (self-oriented) were more affected by fear appeal..; threatening the individual. Similarly, a study on attitudes toward smoking revealed that familyrelated consequences such as "harming the health of their chi ldren and family criticism" were more important for Hispanic smokers than for non-Hi:-.panics. because of Hi~panic~' strong family orientation and collectivism (Marin. Marin. Perez-Stable. Otero-Sabogal. & Sabogal, 1990. p. 490). A caveat emerged from Wille and colleagues' (2000) research. which found that ethnicity does not necessarily overlap with cultural orientation. Contrary to the aS~lImp tion that Taiwanese college students arc more collectivbtic than U.S. college student~. Taiwanese college students emerged as more individualistic.
Conclusion The overall results of recent fear-appeal research suggest that high-threat mes~ages work better than low-threat messages-as long as perceived response and self-efficacy are high-regardless of age, trait variab les. culture. and stage of change. However. individuals must truly believe they can effective ly carry out a recommended response before they will do so. Similarly. it is imperative to discover what a f,pecitic target audience perceive~ as threatening about a specific heallh issue. especially in terms of young high-sen:-.ation seekers and members of different cultures. The variables of fatalism and reactance deserve more research, which shou ld provide greater insight ilHo the effects of fear-appeal messages.
With (hc emergence of rcmarJ...<.Iblc (echnological .Il.lv31H.:e:-. that enable lI:-. to dcvelop more and morc tailored message .... i( b important (0 cO(l'l,idcr these diver...c audience variab les when dc\'eloping effecti\t:: fear appeab-.. The designers of fear-appcalmc:-.sages have an unprecedented opportunity to de ... ign. de li ver. and engage in effecthe. audicncccentered peN,a,ion (Buller. Borland. & Burgoon. 1998).
Notes' ________________________________________________________ I. Rcpre" .. op.. tend 10 avoid potentially threatening though",. emotions. and c'
Rejerellces,_____________________________ Bcrkowit/, J. M. ( 1998). The inj1lwnn' of \l'tI\lIliOIi wi'kill.1.: WUlllll'\I(lgl' elwin' 011 rt'If}(J11 If' \ lo,ft'ar lIPpt'tll\. Unpublishcd doctoral di,"crtation. Michig;1Il St;ltc Uni\CNty. Berlo. D. K. ( 1960). The process OICOmJl1Wlinll;oll. Nc\\ Yorio;; Iioit. Rinehan. & Win'ton Beutler. L. E. (1979). TO\Nard "peelfic p..)'cholngicallheraple... for \llCcific condition:-. )ollnw! of CO//IU/tillg (/1/(1 C/illiw! P~y('ho!ogr. oJ7. 882-H97. Bo~ter. F 1.. & ~tongeau. P. (1984). Fcar-arou .. lIlg pc ..... u.l .. i\c mc"ageo,. In R. BO ... tro111 (Ed.). ('omllllll/;('lIt/rm rt'liriJool\. mf. Blpp. JJO ·J75).l\c\I,bur) Park CA. Sage. Brehm, J. W . (1966). A ,heory ojlH,\'C'h%giw/ rt!lIeImla. Ne\N Yorio;. Academic Pre, ... Brchm, S. S .. & Brchm, J. \V. (Il)H I t. P\\'d/(}/o~il'tll rt'tlc/(I/fI'l': A III('OI'.\' offrt'('dolll {Ind {"(III/mi Nc\\ Yorio;; Academic Press. Buller. O. B., Borland. R.. & Burgoon. M. ( I (98). Impact uf bel,.1\ ioral intention on effect,lcnc" 01 rlIC" sage featurcs; Evidence from the tumily "un .. lIfelY proJect. 1I11111l/11 COIlllllllllinllioll Rl'\('(/I"('h, !oJ.
433-153. Bu\hman. B. J. ( 1998). Effect~ of warning and Inf0n11ation label, on con ... urnption 0 1 full -fat. reduccd·fal. and no-fat products. )ollnlll! oj APl'lil'd P\rdlOlogy, 83. 97- 101 . B)mc, D. (1964). Rcpression-..cnliitiwtlon a., a dllllcn",on of pcr..onallty. In B. Malo;cr (Ed.). Progrt'\,\ ill experimelltal persmllllily rt'.H'anh (PP. 169-220). New YorL Academic Press. Ca~ey. M. K. ( 1995). Faw!ism wltl Iht' modijicmi(J11 of the I:..~tellded Parallel Process Modl'l. Paper prco,ented at the annual meeting of the Speech COlllllluniCtltion A"ociation. San AntonIO. TX Cho. H. ( 1999). Ullin/ended effec/,\' offi'll" appea/J: Tile role of J/(lJW\ of change. {"real. wul ('{(intcr. Unpubli .. hcd doctoral di<.,\enation. \tIichigiin State Unl\cr...JI) Cho, H. (2000). Pnc/IO/ogica/ reaL'ltllln' am! fi'or appl:a/\ III "t'ltlt" ('ommllllinilioll: /)0 pl·t/plt' d(l Ille Oppmi{l' H'IIe" feared? Paper prc'cnled at the annual mccung of the :-.lationa1 C0l1l111llnicauon A .. o,OC/allon. Seallle, \VA . Connor.... M. M. (1992). Risk perception. ri<.,k lalo;ing and ri .. 1o; mam'gclllclll among intra\ crHlll' drug u,cr,: Impl ication .. for A IDS prc\ cntion. Soda/ Scil'lll't' and MediC/lit!. 3oJ. 591 ....60 I . Dabbs. J. M .. & Leventhal. H. (1966). Effec\\ of varying the rccomlllcndation~ in a fear-arnll,ing communication. )oumal oj Per.wnwlil\, lind Social P.\Yc!lOlo1-:Y, oJ. ~25-5J I. Domino. G .. Fragoso. A .. & Moreno. II. ( 1991). Cro,,~-cliitural il1\c .. tigations of the imagery of canl'er in Mcx iean nat ionals. Hispanic lOIll'll"/ 0/ B(,ll(lI'ioral Sciel/ce.l. 13, 422---1-35. Do.mhew. L.. Lorch. E. P .. & Pahngrecn. P. (1991). Sc:n ... ation .. eclo;ing and targeting of tclc\i'cd anti-drug PSA ... In L Donohew. H. E. Syphcr. & W. J. Bulo;o<.,lo;r (l::d,.), Pl'r\lw~iI'e ('OIl/lIIlmu'cllimlllll" drug ohme prel'l'mioll (pp. 209-226). Hill ,dalc. NJ: Erlhaum. Donohc\N. L .. Lorch, E. P .. & Palmgreen, P. (1998). App!rc;II10n ... of tl theoretic model o j inlurmalion c'<po .. urc to health imervcnl.on .. , fllIlIIllII COll//l/Il11lWfi(1/I ReH'{/rc!,. 1.J.433-453.
236
Part II I • Seekillg alld Resi.~till~ Comp/hll/a: Srmre,,
DO\~-d.
E. T .. & Sander.... D. (199..\.). Resi"lance. reaCl of exerL'i . . e. jmlHlal ('IApp/it}d Sodal p~Y('''ologL 22.55-69. Goldstein. M . J. (]959). The re lation . . hip betwecn cuping and avoiding behavior and re . . pon'>e to feararou ... ing propaganda. Jot/mal oJ A}mol"l/wl alld Social Psychology. 58. 247-151. Halt:. J. L .. & Dillard. J. P. (1995). Fear appeal . . in health promotion t.:ampa igns: Too lll11(.:h. too little. or ju ... t right? In E. Maihadl & R. L. Parrott (&is.). Desigl/ing /iNtltil messages: Approaches from ('(Jmll1ll11imtion f"eor\' (/1/(1 public lIea/IllpracTice (pp. 65-80). Thou ... and Oah. CA: Sage Han ...en. W. B .. & Malone. K. (1986). Perceived personal immunity: Tile de\elopment of belief. . abom ,>u-.ccptibilily to the con ... equcnces of . . moking . Prel'elllil'e Mt'didl/e. 15 .."\63-372. Hardeman. W.. Pierro. A.. & Mannetli. L. (1997). Determinanh of intcnt ion . . to prac tice . . afe ... ex among 16-15-year-old\. jOllrnal of COlllmwlin' & Applied Social PITdlOlogv. 7. 3..\.5-360. Helman. C. G. (1990). CII/fllre. helllrll. alld jlllll'.I .I. London: Wright. Hi ll. D .. & Gardner. G. (19S01. Repression-\en"itilation and yielding to threatening hea lth cOlllmunica1ion ..... A IIstrafjall jOll/"llal oI Psychology. 32. 183-193. Horo\\ it/. I. A. (1969). Effect . . of volunteering. rear arousal. and Ilumhcr of communication . . on altitude change. Journal oJ Perw11/ality al/d Soda/ Pnclw/ogy. I. 3-4---37. Horowitz. L A. (1972). Attitude change a . . a func tion of pcrcei\"cd mg communication,>. The JOUr/W/ of Abllormal and Social Psychology. -IR. 78-92. Jani!.. I. L.. & Feshbach. S. (llJ54). Personality difft:rence ... aSloooc iatcd with I"t.! ... rom. ivene . . . . to fear-arou),ing (.:ommunication .... jOllrl/ol o/" Per.wl/ality. 23. 15-4---166. Jepson. C .. & Chaiken. S. ( 1990). Chronic i ... loouc-.. . pecific fea r inhibit.... ... y.'.lcmatic pr~e ....... ing of per. . uasivc (.:ollllllunication .... j()lImal oI Social Be/WI·jor WId PersollalilY. 5.61-84. Jo"eph. J. G .. Montgomery. S. B .. Emmon-.. C. A .. Kes!o. ler. R. C .. O ... trow. D. G .. Wortman. C. B .. O·Brien. K.. Eller. M .. & Eshleman. S. (1987). Mag nitude and determinan ts of behavioral rilook reduction: Longitudinal analysis of a cohort at ri~k for Al OS. P.\\"clwlogy (flld Hefllth, I, 73-96. Kalichman. S. C. Kel ly. J. A .. Morgan. M .. & Romp:t. D. ( 1997). Fa tali"'lll. current life . . atbfaction. and ri~k for HI V infection among gay and bisexual men. jOlln/(// ojColI.wltillg fllld Clilliwl Psychology. 65, 542-546 . Klapper. J. T. (1960). The ('ffi'cIS OJllltH.\' l'Omlllllllinll;oll. Ne\~ York: Free Preloo'>. Kleinot. M. C. & Roger ..... R. W. ( 19S2). Identifying effective component'> of alcohol mi . . u . . c prevcntion programs. Jvurl/ol of Sllldie~ Oil Alcohol. -13. 802-H I I.
l Chapter 13 • A Rel'iclI' of Fear-Appeal Effects
Kouabenan. D. R. (199g).
Belicf~
237
and the perception of ri:-.k:-. and accident:;.. Ri.\·k AI/u/rsis. 18. 2-1-3-
252. Marin. B. Y .. Marin. G .. Perel-Stable. E. J .. Otero-Sabogal. R.. & Sahogat. F. (1990). Cultural difference~ in attitude ... toward ... moking: Developing mes~age ... u~ing the theory of rea:-.oncd action. Joumal of Applied Social Psychology. 20. -I-n-493. McLeod. J. M .. & Becker. L. B. ( 1974). Testing the validity of gratification measure:-. through political effects analysis. In J. G. Blumler & E. Katz (Eds.). The I/se.~ of II/(IS.~ cm1/lllimiclltioll.\·: Cilrrel/t per.\pectil'es Oil grariji("(ltiolls re.H'llrcli (pp. 137-164). Beverly Hill". CA: Sage. McMahan. S., Witte, K .. & Meyer. J. (1998). The perception of ri:-.k mes:-.age regarding electromagnetic rields: Extending the extended parallel proces:-. mode! to an unknown ri:;.k. Health COl1llllllllic(Ilioll.
/0.247 - 259. Mongeau. P. A. (1998). Another look at fear-arou ... ing peNlilsivc appeal\. In M. Allen & R. W. Preis:-. (Eds.). Persua.l"ioll: Adl'(lIIce.l' thmugh lIIeta wllllysis (pp. 53-58). Crev,kill. NJ: Hampton Pre~s. Montgomery. S. B .. Jo:-.eph. J. G.. Becker. M. 1-1 .• Ostrow. D. G .. Ke:o.:o.ler. R. C. & Kirsch!. J. P. (1989). The health belief model in understanding comp[i~nce with preventive recolllmendation~ for AIDS: How u:-.efu[? AIDS Edu("{llioll and Pr(!l'l'Ilfio/l. I. 303- 323. Pro(;haska. J. 0 .. DiClemente. C C. & Norcro ... s. J. C. (1992). [n search of how people I.:hange: Appli(;ations to addictive behaviors. AmeriCl/1I P.\"ychologi~·t. 47. [ 102- [ I 14. Rogers. R. W. ( [975). A protection motivation theory of fear appea[~ and attitutle change. Journal (~r PlY' cllOlagy. 9/, 93- I 14. Rogers. R. W. (1983). Cognitive and phy~iological proce~:-.cs in fear appeals and attitude (;hange: A revi~ed theory of protection motivation. In J. Cacioppo & R. E. Petty (Eds.). Social psychophysiology (pr. [53- 176). New York: Gui[ford Press. Rogers. R. W .. & Mewborn. C R. (1976). Fear appeals and altitude change: Effects of a threat' ... noxiou<;nes:-.. probability of occurrence. and the efficacy of coping responses. jOllm(l1 of Per.wH/aliIY lIlId Social Psychology. 3-1. 54-61. Rohrbuugh. M .. Tennen. H.. Press. S .. & White. L. ([98 I). Compliance. defiance. and therupeutic paradox. Alllerican JOIi/'/!al ofOrlhopsychimry. 51. 45-1---467. Schafer. R. B .. Schafer. E.. Bultena. G .. & Hoiberg. E. ([993). Coping with a health threat: A study of food safety. JOl/rnal of Applied P.lyclwlogy. 23, 386-394. Schocnbach[er. D. D .. & Whittier. T. E. (1996). Adolescent proce~\ing of social and physical threat communications. JOll/"l/al ofAdl'erti.\·ing. 25. 37-54. Seibel. C. A.. & Dowu. E. T. ([999). Reactance and therapeutic noncompliance. Cogllilh'c Therapy & Research. 23. 373-379. Shepherd. G. J. (1999). Advances in communication theory: A critical rev iew. Jourl/al of COl1llJliillimliol1. ~9.
156-164.
Slllaicc. J. ([996). BI/limia illfen'ellfions I'i(l illlerper.lUl/lIl il/j7uCl/ce: TIl(' role of thrcal (Illd ctJiC(lcy ill persl/adillg bulimics to seek help. Paper pre~entetl at the annual meeting of the Speech Communication Association. San Diego. CA. Stephenson. M. T .. & Witte. K. (1998). Fear. threat. and p~rception:-. of efficacy from frightening skin cancer messages. Puhlic Health Rel·iews. 26. 147-[ 74. Straughan. P. T .. & Scow. A. (1995). Barriers to mammography among Chinese women in Singapore: A focus group approach. Health Edlluuioll Re s('arch. / O. 43 1--1.41 . Sti ff. J. B. (199-1-). Per.Hwsi I'e COlll1ll1lllicatiol1. New York: Gu i Iford Pre ... :-.. Suchman. E. A. (1967). Preventive health behavior: A model for research on community health campaigm•. ./ouI"I/al of Heallli alld Social Behavior. 8. 197-209. Sutton. S. R. ([ 982). Fear·arousing communications: A critical examination of theory and research. In J. R. Eiser (Ed.). Social psyc'l!ology ([lid behtll'ioml fl1l'dicille. (pp. 303- 337). London. Wiley. Weinstein. N. D. ([ 980). Unrealistic optimism about future life events . .Io/l/"lwl of Per.wmalilY ami Social Psychology. 39. 806-820. Wein:o.tein. N. D. (1982). Unrealistic optimi~m about susceptibility to health problems. Journal of Belwl'iom/ MNlicille. 5. 4-l.1--I.60. Weinstein. N. D. (1988). Precaution adoption prm::e:-. .... lleallh Psychology. 7.355- 386.
238
Part III • Seek;"J.: alld R('si~lillJ.: COI1lI,/iallce: Stmu'~ie,~ (Jlld Tau;n
Wheatley. J, J .. & O ... hi~a\\ n. S. (1970). The rclilllon~hip hetween an'tiCly and po,iti\ e and ncgatiH ad\er· tl ... lng. uppea]..,. JOIln/al of MarJ..er/lIg Ht'wanA 7. 85-89. Wille. K, (1991a). The role of threat and efficacy inA IDS pre\ enlion. lutemal;O/wl QUlIrterly of Cowmll11/" fh'alt" Edl/wt/oll. 12.115-148. Witte. K (1992b). PUlling the fear bad InlO fear appeal,: The cl[tcndcd parallel procc\:~, model. Cmllllllllljnlfioll MOllogrllpln, 59, :nO-149, Wille. K. (1997). Preventing teen pregnam:y through pcr... ua ... l\-e commun ication!.: Realitie!.. myth .... and the hard-fact truth~, Journal (~l COif/mill/in' Health, 22. 137- 154. Wllte. K, (1998). Fear a ... mOli\a\or. fear a ... inhibitor: U... ing the E:xtendcd Parallel Procev., M,xiclto cxplalll fear appeal ... ucce ...ses and failure\. In P. A. Ande~en &. L K. Guerrero (Ed .... ). The Htillc/bool.. of ("OIl/IIlUII;Clllioll alill emot;oll: Ht',l l'lI rc" , Iltl'orl', llppliclIIiolll, lllld COllfe.HI
25.571-585. Willc. K.. Cameron. K. A .. McKeon. J K .. & Bcr~()will. J. M. (19%). Predicting ri ... k behavior.. : De\cI · opmcnt and \ alidation of a diagno ... tic ...calc. j(l/m/(lll~f flell!lh Com""lIIieatiol1. /. 3 17-.'41. Wllte. K , Meyer. G .. & Martell. D, (1(X)]). I:.ffi.'uiw' !It'll/th ri ..k ,1Il',I\(fge .... A step-br·.\tt'1' gllide. Newhul) Park, fA: Sage. Wille. K .. Meyer. G .. & Manell. D. (111 pre .. ,). f)~'\"('lol'illg ej)ecril'(' health r/lk "wnage~ thllt Imrk. Newbury Park. CA: Sage. Wille, K .. & Morri ... on. K. (I 995a), Thc u... e of 'Cllrc lactic ... in AIDS pre\cntion: The ca ... c of ju\cnile detention and high ...chool youth, jOllflla! of Applied COIl1"III";Wt;(U/ Reuarch. 12. 128- 141, Wille. K, & Mom<;on. K. (1995b). Intercultural and cro...... ·cultural health communkation: Undc«;tandlllg people and motivating healthy hehavion.. ""efllaliolla/ wultllierell/fllmi Commllll;carioll AI/I/I/(/I, 19. 216-246. Willc. K" & Morri'on. K, (2000). Examining the inllllcnce of trait :l nxiety/reprc<; ... ion-!.en<;iti7ation on in · di\ iduab' reaction' \0 fear appeal .... We.\lem JOllrnal of COmllflll/;WI;(}II. rH. 1-29 . Wille. K.. Murray·Jolln<;on. L.. Hubhcll. A. P.. Lill. W. Y.. Samp... on, 1.. & Morri<,on. K. (1()(X». Addre ...... · ing cultural orientation ... in fear appeal:-..: Promoting A ID S~protective behavior... among I-II~panic immigrant and Africa n· American adole~cenl~. and American and Taiwanc~e college ... IlHjcnt .... )ollmal of Ht'tI/th Commulliclltion, 6. 1-23, Wille. K .. Peterson. T R .. Va llabhan. S .. Stephcn,on. M. T .. Plugge. C. D .. Given .... V. K.. Todd. J. D .. Bedtold. M G .. ll yde. M. K" & Jarrett. R (1993). Prevcnting tractor-relaled injuries and deilth ... in ruml populatIOn ... : U.;;ing a Pcr,ua ... i\c Health Mc ...... age (PIIM ) framewor~ In fomlati\e evaluallon n: ...carch, InlemarimUlI Qllllrler!" of COIl1IlIIlIIil\' /-Iealth EduClIliol/. /3. 119-251 , Zuderman. M. (1988). Beha\ ior and hlology: Rc\carch Oil <;cn\allon ...ccking and reaction' to the media. In L. Donohc\\'. H E. Sypher. & E. 1'. Higgln ... (Ed .... I. COmmllllil'lltioll . .Ioe/a/ ('or{lIilioll, mltl (Im!et (pp. 173-194). l-lill...dale. NJ. Er1haum Zuckcrman. M. (1994). Behcn'ioral ('~l're.H;mH (Illd /1/010(';(11 h(/\('I of .~ell.\lItIO/, l·eeJ..;".!:. Cambridge: Camhridge University Pre, ... ,
14 Interpersonal Deception Theory Judee K. Burgoon and David B. Buller
Introduction Newspapers and television daily call our attention to all manner of deceptions : :-.pie\ creating false identities and spinning false tales. politicians lying about lheir priYlltc relationships, business executives covering up fraudulent deal\. foreign governmcnt\ creating disinformation campaigns. But deceit is nO( just the Mufr of !o.cnsational headlines. It is all around us, every day and in every relationship. In fact. even the most publicized cases of deceit comprise endless interpersonal encounter~ in which lies. exaggerations, misrepresentat ions and the like are created and perpetuated. An understanding of deception. then, is best realized when grounded in the interpep;.onal interactions that give deceit its sustenance. IlIIerpersollal Deception Theory (IDT) arose out of just this concern that deception should be examined within the nexus of interpersonal encounters. It was formulated to contex tualize an explanation of deceptive communication in what we know about conversation. T his approach stands in contrast to more psychological explanatio ns of deceptive communication. It also draws altenlion 1O the dynamic nature of display5. of deception and to the mutual influence between sender and receiver th aI occurs in all conversations. This c hapter outlines the assumptions on which IDT is built and discu!-oses several key proposi ti ons of the theory. In formulating lOT. we 'ynthesi£ed a broad range of evidence and conceptual perspectives on co nversational behavior. interper..,onal influence. nonverba l communication. normative expectations. and source credibility. The most notab le proge nitors for IDT are the first author' s research into conversat ional expectations a nd behavioral adaptation ( Burgoo n, 1978. 1993; Burgoon. Stern. & Dillman. 1995). the seco nd author's researc h on verbal and nonverbal social influence (e.g .. Buller. 1986. 1987; Buller & Aune, 1988, 1992; Bulle r & Burgoon. 1986), ou r combi ned functional approach to nonverbal communication (Bu rgoon, Buller. & Woodall. 1996). and decades of research on verbal and non ve rbal factors in source credibility (see. e.g .. Buller & Burgoon. 1986; Burgoon. 1976; Burgoon & Hoobler. in press). Given the broad net we cast. IDT
239
240
Part III • Seeking (llld Re\'isting Compliance: Strategies (llld Tactics
qualifie!o. a~ a mid-range theory that has multiple explanatory mechanisms within its propositions. The remainder of the chapter i!o. devoted to summarizing the results of our experimental te~ts of lOT. We present these in a largely chronological order so as to give readers a sense of how the thinking about, and testing of, lOT evolved.
Assumptiolls About Illterpersonal Communication The mutual influence in normal social interaction arises from the active participation of all parties to the conversation. Communicators are not involved in conversation only when encoding messages; they are dynamically engaged in reception of messages, as well. In fact. it is a misnomer in interpersonal interaction to separate senders from receivers, except in an abstract sense (which we do henceforth). In normal conversations. speakers encoding messages are simultaneously monitoring and decoding the conversational behavior of listeners (e.g .. observing feedback, turn-taking cues. and overt reactions to the message- including emotional reactions). Likewise, listeners usually are not passive message recipients. While listening. (hey provide verbal and nonverbal feedback and tUrtltaking cues. manage their demeanors. and formulate their own turn at talk. All parties to deceptive episodes are likewise concerned with multiple goals such as preserving good interpersonal relationships, masking inappropriate emotions, keeping conversations running smoothly, and appearing credible. In achieving these multiple conversational functions. they muM manage a host of verbal and nonverbal behaviors. Thus. conversations are dynamic. multifunctional, multidimensional. and multimodal events in which participants must perform numerous communication tasks simultaneously in real time. Such juggling requires considerable skill to accomplish effectively. Communicators are also responding to a host of cognitive and behavioral factors that influence deliberate communication acts and in turn produce some unintended and unwitting behaviors. Although conducting social interaction is arguably a cognitively demanding activity, it appears that people are generally good at it because much of normal conversation is fairly routinized. Also. social interaction is made easier by the fact that we have learned to follow culturally prescribed rules and expectations. Some of the most important feature~ of lOT are expectations for truthfulness, for conversational involvement. and reciprocal or matching conversation styles. How one reacts to, and interprets, the fulfilhnent or violation of expectations goes a long way toward determining the outcome of conversations containing deception. These features of interpersonal communication are the context in which we formulated lOT.
Assumptions About Deceptioll In lOT, deception is defined as an intentional act in which senders knowingly transmit messages intended to foster a false belief or interpretation by the receiver (Buller & Burgoon, 1996a; Ekman, 1985; Knapp & Comadena, 1979). To accomplish this. senders engage in three classes of strategic. or deliberate, activity: information, behavior. and im-
Chapter 14 • 11I(£'rpt'rwmal Dl'l'('ptioll Theon
241
:.Ige management. The term "management" implie, that deception i, a motivated beh'I\'ior. undertaken for i.l purpo..,e. U,ually. that purpo"l: j", one that benefit ... the ~ender, although ...enders frequently claim that the) dccci\'c to bcnefit the recei\cr or a third part) to the conver~ation. Iliformarion 'tUlIIlIMemem refers to elTort~ to control the contents of a me ... ..,age and u,ually concern~ verhal fcatures of the mcs~age. Bel/lll'ior IIItlllllxemelll refer~ to t.!ITorts to control accompanying nonvcrbal bdmviors that might be telltale ... igns that one i.., dccci\ ing. It derive~ from tht.! a ... sumption that \erbal and nOIl\erbal me..,sages an: con... tructed <.\ ... a unified \\ hole and that nonverbal behaviors are often inh.!nded to augmcnt i.ll1d extend the meanings convcyed by \"I:rbal contcnt. Image IJumagemellf refer.., 10 more gencral cfforh to maintain crt.!dibility and to protect one's face. evcn if caught. It derive ... from the w'I ... umption that individual.., arc motivated to protect their ..,elf-image and public image. The..,c three classes of ..,tratcgic activity work hand in hand to create an overall believable me"i ... agc and demeHllor. By way or example. a student ... u... pected of cheating might tell her profc..,sor. " 1 did not look at m) neighbor's exam" (information management) while cro ...... ing her arm.., to aVOId ncrvou.., gc ... tures or body movement.., (beha\ior management) and ... miling to appear hone ... t (image management). Thi ... a ...... umption that ...cndcr.., arc active agent<.; \.\ 11O"C bchavior renects planning. rchcar... al. cl1iting. and other con ... ciou ... or scmicon ... cioll'" effort.., at succcs ... rul deceit does not preclude uecei\.er ... from also engaging In what we rcfer to a.., lloll.\lraregic actiol/.\. that i..,. c1a..,,,e ... of heh:.1\ ior that may be involuntary ilnd uncontrolled. NOn"ilfUtegic activity may rc<.;uit in poor. unnatural. or Clllhi.IJTa..,sing communication performance .... A case in point i ... when a child blushes \vhen gi\ ing a Ilontruthful answer to a parent· ... inquiry. The complexity of deceptive messages. and the knowledge that deception violates conver..,ational rules and ... oeial prescriptions again..,t deceit. can alter the mcntal ... tate of ..,ellder". It CUll incrca ...e the cognitive effort necded to formulate thi.., multiraceted conversational behavior. It may al ... o increa..,e arou ... al ~lIld provokc ncgati\c affect. All of thc~e processes may result in lIladvertent ... ignals that something i.., not quitc normal in the conversation. that is. nonstrategic activity. although lOT does 110t ih~lIlllC thm ..,uch signals are neces~arily or univer~ally pre ... t.!I1t. Finally. hcci.lu ...e we <.;ituatc deception in cOIl\.'er\
The IDT Model Summarized With the ... e n..,..,umj1tion" about intcrper..,onal communication and deception as i.I backdrop. we formulated i.I theoretical model of deception containing 18 propo~ition~ (table 14.1). They dc..,crihe all iterative proccs.., of Illutual intlucllcc in which the enactment of deception by one convcr..,ational participant provoke.., a cascade of moves and countermoves by hoth pal1ies to thc conversation. The ...c moves are aimed on the one hand at adapting the dccepti\'e me ... ..,agc in order to maintain its apparcnt truthfllines.., (i.e .. achieving deception ... ucce",..,) and Oil the mher at di ... cerning the credibility of the message and the ,ender and
242
Part III • Seekinx and RCJi.\/iIIX CompliclllCl': S/ro/('~ies alld Taerics TABLE 14.1
Propositiolls ill t"terpenwllal Deception Theory
I. Sender and rccei ver cognition ... and behavior'-. vary systematical ly as deceptive communication contexts vary in (a) access to social cues. (b) immediacy. (e) relational engagement. (d) conversational demands. and (e) spontaneity. 2. During deceptive interchange ....... ender and receiver cognitions and behaviors vary systematically as relationships vary in (a) relational familiarity (including information and behavioral familiarity) unci (b) relmional valence. 3. Compared with truth teilers. deceivers (a) engage in greater strategic activity designed to manage information. behavior. and image and (b) di~play more nonstrategic arousal cues. negative and dampened affect. noninvolvement. and performance decrements. 4. Context interactivity moderates initial d~ccption displays so that deception in increasingly interactive context, results in (a) greater strategic activity (information. behavior. and image managemcnt) and (b) reduced nonstrategic activity (arousa l, negative or dampened affect. and performance decrements) over time relative to noninteractive contextS. 5. Scnders' and receivers' initial expectations ofhone~ty are positively related to degree of context interactivity and positivity of relationship between sender and receiver. From Buller. D. B.. & Burgoon. J. K. (1996) Interpersonal deception theory. Communication Theory. 6. 203- 242. 6. Deceivers' initial detection apprehension and associated strategic ac ti vity are inversely related to cxpectations of honesty (which are them ...elves a function of context interactivity and relationships positivity). 7. Goals and motivations moderate strategic and nonstrategic behavior displays so that (a) senders deceiving for \clf-gain exhibit more strategic activity and nonstrategic leakage than senders deceiving for other benefits and (b) receive rs' initial behavior patterns are a function or (i) their priorities among instrumental. relational, and identity object ives and (ii) their initial intent to uncover deceit. 8. As receivers' informational. behavioral. and relational familiarity increases. deceivers not only (a) experience more detection apprehension and (b) exhibit more strategic information, behavior. and image managcment but also (c) engage in more nonstrategic leakage behavior. 9. Skilled senders better convey a truthful demeanor by engaging in more strategic behavior and less n()n~trategic leakage than unskilled ones. 10. Initial and ongoing receiver judgments of sender credibility are positively related to (a) receiver truth bi;.lse .... (b) context interactivity. and (c) sender encoding skil ls: they are invcrsely rclated to (d) deviations of sender communication fro111 expected partern~. 11. Initial and ongoing receiver delection accuracy are inversely related to (a) receiver truth bia!o.cs. (b) context interactivit) . and (c) sender encodi ng skil Is: they are positively related to (d) informational and behavioral familiarity. (e) receiver dccoding skills. and (f) deviations of sender cOlllmunication from expected patterns. 12. Receiver suspicion i~ manifested through a combination of strategic and nonstrategic bchavior. 13. Senders perceive suspicion when it is present. Dcviation~ from expected receiver behavior increase perceptions of sllspicion. Rcceivcr behavior signa lin g disbelief. uncertainty, or the need for additional information increases sender perceptions of sllspicion. 14. Suspicion (perceived or actual) increases sender ... · (a) stratcgic and (b) nonstrategic behavior. 15. Deception and su~picion displays change over time.
Chapler 14 • Ill terpersollal Deceptitm Theory
TABLE I·U
243
('omi"ued
16. Rcciprm:Jly
j ... the predomin
ultimately reaching an interpretation of th e meaning (i. e .. achieving detection success). Because mo.." people know how to carryon a conversatio n, th ese actiol1 l-1 may be "nm off" with rela ti\ c ease and at a low level of co n..,ciouliness rather than requiring significant cogniti ve or physical effo rt. The s ubtl ety of deceptive processes is o ll e reaso n that detection is ... uch a cha ll e nge. as we shall sec. This process and its outcomes arc determined by several factors di..,cu ... sed throughout this chapt er. These include co nt ex tual factors. such as the degree or imcractivity pos~ible, sender~' and receivers' preinteraction charac teri stics slich a.., ..,ocial .. kills. preexi~ting knowledge (called information und behavioral familiarity), th e po..,iti\e or negati",! vale nce of the re lationship between conversationa l partners, and initi a l l!.'I(pccta ti o ns fo r hone..,ty wit hin the excha nge (see figure 14 . 1 for a simplified depi cti on of th e interact iyc process of interpersonal deception). All of Ihe . . e factors sho uld influence whethe r sende rs or receivers hold a re lati ve advan tage during deceptive episodes. IDT has been tested in a program of experi me nt s that address the nature of deception displays in interactive c ircumsta nces (i.e .. ones in which perpetrators o f deceit interac t in rea l tim e with the sa me receive rs w ho re nd er judgme nt s of se nd er truthfulnes s). To tes t IDT properly. the experiment"> had to I11cd a number of othe r c rit eria. They needed to be of .. uf'licient length to captu re the dynamic . . of interacti on. inc luding the potential for a broad range of pos..,ible strategic and nonstrategic actions to e me rge. Our co mmitme nt to an intc.:raclive research design and to ac hievi ng ge nerali zability to a broader range of discOlIl·. . e abo led us to emp loy C0 l111110n rorl11~ of discourse. such a . . interviews or di sc ussions of a partic ular topic. that not o nl y situated the deception wi th in normal conversational routine ... but a lso required far more exte nded ta lk than typ ical deception experiment s u... ing very brief. ... ometime~ sin gle-se nte nce, utterances. Additionally. in stead of the usua l practice of relying on college stude nt samples. many st udi es enro ll ed non student partici pants who we re rec ruited from th e jury assembly room at the coun ty co unhou se. from c ivic organit:.llion'-l. from employ me nt centers (wi th trainin g offered in exc hange for participa ti o n). and from nonprofit group ... (w ho earned money for their o rga ni lation by partici patin g). In a ll cases. we lio ugh t equa l numberl-l of mal e\ and fema les in ou r sample so th at our re~u h .. would be applicable to bot h sexes. In some cases. we ah o inc luded bo th friend ... and strangers so as to furt her increa~e ge nera li zab ility and to uncover any relation!o.h ip dilTerellccl-I. Finally. to compare pal1icipants' percepti o ns wi th o ne another and with objccti\c behavioral data. we a . . ked partil:ipa11l~ themselves to report o n the ir own and their partncr's communicat ion and then s ubj ected videotaped inte racti o ns to extensive
244
Pan III • Seeking
(1"'[
Resi'lring Compliance: Srrlllegie!J and Ta clies
CONTEXT AND RELATIONSHIP
Interaction
Pre-Interaction CognitionsiAHect
-
a:
w
0
zw
- - - --?
• Expectations • Goals • Familiarity • Detection Apprehension
(/)
>-
a. '"
f--7
?!
'" <5
'" "
• Skills
ro ./
~c:
Cognitions/Affect
a:
w ~
w w
U
• Expectations • Goals • Familiarity • Prior suspicion
a:
Behavior
-
'\..
- -
Behavior
• • • •
~
0
.c
• Routines
CognitionsiAHect • Perceived suspicion
"5
Behavior
-
Post-Interaction
information management Image management Behavior management Nonstrategic behaviors
DECEPTION SUCCESS
\
~
CognltionsiAHect ---? • Credibility judgments Behavior • Suspicion display • Uncertainty management • Nonstrategic behaviors
DECEPTION DETECTION ACCURACY
• Routines • Skills
FIGURE 14.1
Depictioll oflhe illteracth'e Process of Illterpersollal Deception
codin g by trained raters. Some s tudie~ abo included third -party observers to test for differcnce~ due to perspective (part icipant ver~us ob"erver). The re,t of this chapter is devoted to reviewing evidence from this program of re:-.carch in ... upport of the propo:-.ilions of IDT.
II/put Factors COlltext Factors A major prcmise of lOT setting it apart from other models of deception is that it hold!<. that deception di!olplays and processes differ according to whether or not the sender of deceptive messages interacts in real time with the intended recipients of those messages. This is the concept of illleraClil'iry. If. for example. a third party observes a videotaped illlerview at a later time. that third party will not have the same responses 10. or asse~.'-.ment~ of. any deception by the interviewee that are availablc to lhe interviewer who asked the que ~ tions .
Chapter 14 • hllerpl'r.H)/W! O('ceplioll Theory
245
The observer is not "interactive" with the sender. Similarly. if two friends communicate electronically. doing so via instant messaging is more "interactive" than exchanging serial e-mail messages over the course of a week. Interactivity is thought to make a difference in how deception plays out. One reason is that people in interpersonal interactions usually expect others to be truthful (often referred to in the deception literature as a "truth bias"). and these expectations are considered to become more pronounced as interactivity increases. Three of our studies examined this premise by comparing judgments of receivers who participated in the conversation (high illleractivily) and those who merely observed it (low interactivity). In each case, participants rated senders more honest and trustworthy than observer~ did (Buller. Strzyzewski. & Hunsaker, 1991: Buller & Hunsaker, 1995; Dunbar. Ramirez. & Burgoon, in press). Participant-receivers also judged sender communication more favorably (rating it as more involved. pleasant, and expected) and evaluated sender credibility more leniently Uudging senders as more competent. dominant. and higher in character) than observers did. Together. these studie.<, clearly s.how that interactivity confers an expectation for honesty. and a bias toward leniellljudgmellts. in receivers. IDT suggests that interactivity should also affect senders' deception displays. Senders should increase strategic activity (e.g .. briefer utterances) and decrease nonstrategic behavior (e.g .. fewer pauses) in highly irlleractive circumstances because (I) senders have the opportunity to repair and adapt their communication as they obtain feedback from receivers on their apparent believability and (2) the rapport and trust engendered by conversational participation produces a more coordinated, smooth interaction thai helps senders behave naturally in conversation (Buller & Burgoon, 1986: Burgoon. Buller. Floyd. & Grandpre. 1996). The experiments to be reviewed shortly attest to the impact of interactivity on senders' strategic and nonstrategic activity. Finally. interactivity should affect receivers' ability to distinguish deceit from truth. Compared to observers. receivers embedded in conversation~ face a complex set of conversational and cognitive demands. They must interpret messages in real time while simultaneously formulating their own conversational turns at talk. providing feedback to senders. and engaging in turn management. These tasks can distract them from cues to deception. At the same time. receivers are enmeshed in relationships with senders. As such, they may commit to their expectations for honesty at the outset of the conversations and not adjust them later. These expectations can be further reinforced by the patterns of reciprocity and the maintenance of self and others' face needs during conversations (Burgoon el al.. 1996). Four out of five studies we have conducted have supported the impact of interactivity on detection accuracy (the exception being Buller & Hunsaker. 1995). Buller and colleagues (1991) showed that participants did not differentiate between truthful and deceptive messages but that ob~ervers did. A study by Burgoon and colleagues (1996) comparing sender, receiver, and observer per. . pectives found that observers were Illore attuned to behavioral differences between truthful and deceptive cOlllmunications than were participant-receivers, but that neither set of receivers labeled deceptive performances as deceit. In a study comparing receivers in a dialogue (two-way communication) and monologue (one-way communication). receivers initially were less accurate at detecting deception in a dialogue Lhan in a monologue; as conversations wore on, however. receivers in
246
Part III • Sf'eking (lnd ResisTillg COlllpli(lI/cc: Slrtilegil's
(I1/(/
T(lClie's
both dialogue and monologue were less accurate at detecting deception (Burgoon, Buller, & Floyd, 200 I). Fi nally, a recent companion study that collected observer data to compare to that of dialoguing receivers found that particip:mt-receivers were less accurate in detecting deception (Oullbaret aI., in press). These studies demonstrate that interac ti vity reduces receivers' ability to detect deception. Two questions that arise are ( I ) whethe r lo..,ses in detection accuracy are only short-term and (2) whether any gains receivers might make in their detection acuity are off~et by senders' own gains from adapting to receiver feedback. Studies summarized below partially speak to these issues.
Relatiollship Factors In lOT, the nature of the rdationship between sender and receiver is another input factor that influences the process and outcomes of deception (Buller & Burgoon. 1996: see also McCornack & Park" 1986: Stiff. Kim, & Ramc,h, 1989), The most important relationship features are relational familiarity and relational va lence.
Relatiollal Familiarity.
Relmional familiarity can be both informational and behavioral. Compared with strangers, people in relation~hip\ clearl y know more about one another and are morc familiar with each other's behavior. Such informational and behavioral familiarity could improve receiyer~' ability to detect deception. Military intelligence specialists and law enforcement personnel routinely judge the likely validity of re!o.pollses from prisoners of war or criminal ~uspects by asking questions to which they already know the truthful an~wer 41\ a way of determining how truthful th e respondent i~ being. The same process cou ld be u~ed by relational partners. eithe r deliberately. when their suspicion" are arou~ed. or inadvertently. when information in a deceptive message does not square with their shared history. A shared history also may make relational partner~ more cogni/ant of how each usually behaves in conversation. thus helping them beller recognize deviations from normal conversational paHerns. In o ne of our st udi es. novi ce interviewer.., improved their detection accuracy when asking acquaintances an unexpected que~tion but not ",hen asking it of a stranger (Burgoon, Buller. Ebesu, & Rockwell, 1994). Their familiarity with the \ource'~ behavior probably improved their ability to detect abnormal responses. Another reason detection accuracy may improve over time is that senders interacting with acquaintances. friends. and family may become worried that their deceit wi ll be detected, precisely because of the information and behavioral familiarity these receivers POS\css. But this i~ a complex issue. On the one hane!. any such concerns could be revealed by the presence of nonstrategic anxiety cues and negative affect. These te lltale indicators of deception should Illl.lkc detecting of deception easier for the receiver. On the other hand. anxiety could also motivate \enders to engage in Illore strateg ic move~ to hide their deceit. making detection more difficult because there would be fewer cues for receivers to rely on. Such adjustments in turn cou ld lead to overcontrol. reducing senders' conversational involvement and producing other performance decrements such as nonOuent speech. These addi ti onal nonstrategic behavior\ should make detection easier. Thus. arousal and anxiety can have very mixed effecls on deception display~ and their detectability when interactants know one another.
Chapler J 4 • IlItNI'erV/Hw/ Deceptioll Theon'
247
Our re!o.earch findings reflect just such a mixed bag of re.!o.ulls. By way of illustration, in a study conducted prior to fonnulating IDT (Buller & Aune, 1987), inrima tes and friends differed from st ran gers in their use of eyc contact. forward lean. nervous tou ches to th e face and head. voca l ncrvous ness, vocal pleusantness. and gestural expressivity. A recent study revealed that friend.!o. may benefit from high interactivity (dialogue) but not low interactivity (mono logue). Scndcr~ dialoguing with friends were judged as managing their behavior and image better than senders dialoguing with Mranger.!o.. but friends did wor!\.e than strangers when delivcring a monologuc (Burgoon et al.. 2001). Other Mud ies reviewed below further confirm th at the nature of the interpersonal relationship between interactan ts influences the deception process.
Relatiollal Valellce.
The degree of positivity and trust associated with a relationship may be a powerful cause of truth bias. leading receivers 10 overlook. discount. or misinterpret signals that the communication is not what it ...eem .... For example. in a recent study (Burgoon et al.. 200 I ). receiver... felt that interactions with friends produced more rapport. trust. and simi larity than interactions with strangers did. In another study (Burgoon et al.. 1994), receive rs also co nsiste ntl y overestimated the honesty or acquain tances. Co nve r.;;ely. when led LO be s uspic ioll", (i.e .. when receiver!>. were induced to place a negative va lence on the relationship). novice receiver~ (a!>. opposed to military intelligence specialists) greatly undere!o.timated ... tranger~· truthfulness. In !>,Uln. features of communication contexts and of interpersonal relationships frame deceptive encou nters and must be taken into account 10 achieve any precision in predicting and explain ing deception. AS"iociated belier!>. and expecta ti on"i innuence the initial jlldglllent~ of !>.ender credibility and ea rl y act ions by deceptive sender... with in the conver~at i on. This is just the beginning of the story of interper!\onal deceptio n. Quickly. sende r a nd receiver begi n to exert mutual influence on each other' s cogn itions. reelings. and behavior so that conversational patterns change over ~lIcce~sive conversational turns. With Iho~e adju~tlllent~ come changes in ~e nd er credibi lit y and deception success. In th e next secti o n. we consider the propositions in IDT that describe the interaction processes in interpersonal deception.
Interaction Processes DeceptiQII Displays The prevailing wisdom prior to IDT was lhat the act of deception is accompanied by a number of im'oluntary reactions con ... titllting telltale signs thai a person is lying. Deceivers were expected to experience anxiety and negative emotions such as guilt or fear of detecti on, to find their cognit ive workload increasing as they tried to concoct plausible li es. and to make (usually un s u ccc~sful ) efforts to s upprcs~ th e signs of nervousness, discomfort. and cognitive effort-all of which would result in behavioral cues that the savvy observer could usc to detect deception. Ekman and Friesen (1969) called this the leakage hypothesis. Like Sigmund Freud. they thought that people would inadvertently reveal or " Ieak" their deceptive intent and that these leakagc cue ... were uncontrollable, so they would
248
Part III • SeekillK
(1/1(/
Resi.\/illl{ Compliance: StrOlexies alld Tl/crin
appear unbidden. All that an astute observer needed (Q do was rune in to the right ones to have some surefire clues that deception was occurring. Decades of research have been dedicated to discovering the anticipated verbal and nonverbal signs of deception (see, e.g .. Ekman, 1985: Zuckerman & Driver. 1985). Yet research has failed to yield many consistently reliable indicators. Pinocchio's nose is not evidem in real life. Coupled with the persi~tent evidence that people's accuracy in detecting deception hinges on chance at best. this raised the que~lion of whether a profile of valid indicators could be identified. An equally important question lurking in the background was whether any of the research findings would generalize to conditions in which deceivers actually interact with the inrcnded recipienrs of their deceit. This is where IDT entered the scene. Our fundamental premise that interactive and nonintcractive deception differ led us to expect that deception displays would be transitory. We also predicted that deceivers (or at least, skil1ed ones) would make use of the communication interplay between sender and receiver to repair their performances over time. thus making detection even more difficult as time pa\sed. We further reasoned that if deceptive performances comprise both strategic and nonstrategic elements, researchers needed to examine a broader range of associated verbal and nonverbal behaviors to see if the presence of intentional (strategic) behaviors would make deceit even less detectable. Our conviction that deceit is an active. goal-oriented activity also raised the prospect of senders employing multiple <.;trategies for enacting deceit. each one replete with its own profile of verbal and nonverbal behaviors needing investigation. Also. our conviction that receivers are themselves active agents raised the prospect that their own communication, including overt indications of suspicion. might alter decei vcrs' displays. We summarize next. in chronological order, those IDT studies that explicitly exalllined deception displays. In all. eight original experiments. plus secondary analyses of several of Bavelas and colleagues' experiments and follow~lIp analyse~ on our own videotaped interactions. were undertaken explicitly to test I DT propositions and hypotheses. Because experiment I focused on receiver suspicion and is discussed later, we begin by summarizing experiment 2. which was the first to address deception displays.
Experimellt 2.
This experiment wa~ the fif'..,t to assess a wide range of percei ved and coded nonverbal (Burgoon & Buller. 1994) and language behaviors (see Buller. Burgoon, Buslig. & Roiger, 1996). It built upon earlier investigations by Buller and colleagues (Buller & Aune. 1987: Buller. Comstock. Aune. & Strzyzewski.1989l showing that deceptive performances included a mix of nonstrategic cues related to arollsal and negative affect and <;trategic activity related to information. behavior. and image management. We hypothesized that compared to truth tellers. deceivers would manage information by obfuscating their verbal responses with vagueness. uncertainty, reticence. and nonimmediacy (avoidance language) and that this information management would be accompanied by nonverbal nonimmediacy (e.g .. less gaze, greater distance) and inexpressiveness. We also hypothe~ized that initially. deceivers would be more nervous. negative, and nonOuent than truth tellers but that the dynamics of interactive deception would allow them to improve behavior management and image protection over time. Our experimental methods were intended to elicit natural interactions and selfpresentational concerns that would be common among friends and strangers alike. since our sample included both. Adapting procedures first used by Toris and DePaulo (1985)
Chapter l-t • bllerperwmaf Deception Theorv
249
and Buller and Aline (1987). we framed the study as related (Q how people prc",ent themselves to others during an interview and how well interviewers can detect misrepresentations of true feelings and actions. Half of the participants were asked to lie after the first five questions. The first five que~tions thus provided a truthful baseline during which participants could acclimate to the task and familiarize themselves with the interviewees' communication style. This approach also approximated real-world circumstance~ in which deception and truth are intermixed. Afterward, both participant . . rated interviewee behavior. and trained coders rated assorted nonverbal behaviors from the videotaped sessions. As hypothesized. deceivers and truth tellcrs behaved differently. Consistent with our contention that deceivers strategically manage what information they reveal, deceivers' statements were characterized by brevity. vagueness. uncertainty. nonimillcdiacy. and nonspecificity (e.g .. "everybody went drinking" verMI~ "1 went drinking"). This had the effect of minimizing the amount of concrete and verifiable detail that deceivers supplied and of disassociating deceivers from what they were saying. Other linguistic patterns were contrary to predictions but also had the effect of making deceptive answers more pallid and less personal than truthful ones. Nonverbally. deceiver~' initial behavioral patterns were as predicted. They displayed some strategic moves- less nonverbal immedil.lcy, less dominance, and more formality than truth tellers- that collectively should have curtailed conversation and/or made them seem passive. polite. composed. and nonmanipulative. But their behavior also included some nonstrategic ingredients - nervousne~s, unpleasantness, and gaL.e avoidance- that together with the overall reduction in comersational involvement created an unexpected and undesirable impression. Contrary to lOT. then. deceivers did not project a more favorable image. Had these patterns persisted throughout the course of the interaction. we might have concluded that regardless of any deliberate efforts to the contrary, deceptive performances lack credibility. However, in support of our position that deceivers actively attempt to control and repair their performance~. deceivers' body language became increasingly relaxed and pleasant over the course of these brief five-minute interactions. With more time, we might have seen even more Mrategic repairs. Even without major repair~. it was likely that the subdued demeanor of deceiver~ was effective in evading detection because. as we shall see in later investigations. deficiencies in deceptive performances need not give the deceiver away. Additionally, other results discussed shortly supported IDT's premise that deception varies by such factors as relational familiarity, partner's communication style, and suspicion.
Experiment 3.
The next experiment lengthened the interview time and replaced novice interviewers with trained interviewers who maintained a con..,i..,tent interaction "'tyle across interviewees. Unacquainted participants first completed a truthful interview with a samesex interviewer and then completed a deceptive interview with a second same-sex interviewer. To learn whether there might be multiple deception profiles, deceivers were instructed to use one of three types of deception- fabrication (outright lying). equivocation (being vague and al11biguou~). or concealment (withholding relevant information)or, in a "general" condition, to use whatever forms of deception they wished. As predicted. deceptive responses were seen as les.1'.i conversationally complete (i.e .. failed to meet usual conversational obligations). direct. relevant, clear, personalized, and veridical (honest) than truthful re'ponse, (Burgoon. Buller. Guerrero. Aflfi. & Feldman.
250
Part III • Seekil1~ and Resisrin,:: Compliance: Srrare,::ies and Tacrics
1996). Those results related to information management confirmed that deceivers have a number of features at their disposal to effectuate deception. Senders also reported being far less truthful when fabricating answers than when concealing or equivocating, but observers failed to see differences among the three types. Linguistically, deceivers again used fewer self-references and tended to use fewer group references than truth tellers (Buller, Burgoon, Buslig, & Roiger, 1994). This stripped deceptive responses of so me of their concreteness and personalization. However, as other linguistic analyses proved problematic. we made several methodological changes before the next test of deception types.
Experimellt 4.
Reported in Buller, Burgoon, White, and Ebesu (1994), this next experiment again employed interviews and compared three types of deceivers: falsifiers. equivocators, and concealers. To see how suspicion alters both sender and receiver behavior, we reverted to novice rather than trained interviewers. half of whom were made suspicious. Participants were civilians and military personnel from two locations and included strangers and acquaintances. The resultant complex findings confirmed our conjecture that deception dispJays are highly respo nsive to such factors as the type of deception, degree of relational familiarity. and partner's sus picion. Information management went largely as predicted. Truthful responses were seen as more complete, veridical, direct/relevant, clear, and personalized than deceptive ones. Of the three deceptive types. falsifiers were the least truthful but also the least reticent; their answers were less vague. hesitant, and brief than those of equivocators or concealers. Equivocators were the least clear, direct, and relevant. Additionally, as predicted, deceivers tended to include di sta ncing and ambiguous language (e.g., levelers, modifiers, and group references). but contrary to predictions they also used more present tense verbs and fewer past tense verbs. (After the fact, we conjectured that past ten se verbs are more likely to sound definitive and therefore deceivers might have been expected to use present rather than past tense verbs.) As for nonverbal patterns. they failed to support a clean strategic-nonstrategic distinction. as many behaviors were opposite to predictions. For example, deceivers were expected to be less expressive than truth tellers; instead, they were more so (and yet less involved). Deceivers were hypothesized to be more formal; in stead. the y were less so. Further. only strangers showed the expected pattern of nondominance when deceiving; acquaintances instead became more dominant than their truth-telling counterparts. The overall conclusion we drew from these data is that in interactive contexts. very few display s of uniform deception are likely due to deceivers adapting to audience and context and employing an array of discourse form s, each with different accompanying nonverbal cues. Still. some suggestive patterns emerged. Deceivers often seemed to opt for greater verbal reticence. withdrawal, and vagueness, which has the advantage of reduc ing their chances of making contradictory or implausible statements; to adopt an informal. nondominant demeanor. which might discourage others from probing too deeply and disassociates them from respon sibility for their statements; to offset verbal nonimmediacy and depersonalization through more ex pressive body language; and to hide arousal by suppressing some, but not all, other physical activity. The net result of all these behaviors was an apparent general reduction in involvement, which was emerging as a strong telltale sign of deception. at least at the outset of conversations.
, 1
I
I Chapter 1-1 • IlIIer"e,wlI(l1 Dece"lioll Theon'
251
Secolldary Allalysis of the Equivocatio1l Experiments.
The next move in our research program (Buller el al.. 1994) took us not to yet another experiment but rather to further analysis of research undertaken by Bavelas, Chovil, Black, and Mulletl (1990),
who had been conducting numerous experiments on equivocation. Janet Bavelas graciously made available the audiotapes and videolapc!o. from six studies so lhal we could code the utterances on the same verbal and nonverbal behaviors we had measured previously. (II ~hould be noted that 8avelas does nO( cOI1\ider equivocation a form of deception, but her characterization of it is still quite similar to ours. namely. a form of discourse that includes truthful but indirect. irrelevant. ambiguous. or evasive information to create an impression that deviates from "the truth, the whole truth, and nothing but thc truth.") Her program of research offered an excellent complcment to ours becausc. unlike our direct instructions to subject~ to be equivocal. her experiments induced equivocation by placing people in avoid-avoid conflicts (i.e., situation~ in which both telling the truth and lying would have negative con~equellccs) that naturally encouraged them to equivocate. For example. in one study. people were a!o.ked 10 imagine that a friend had performed poorly on a task and was asking for feedback on how he or ~he did. Telling the truth would hun the friend's feeling!o.; yct people are oftcn loath to lie outright. hence the likelihood of resorting to equivocation. Most of the experiments contrasted a connict condition with a no-conflict condition. Results <;;howed that equivocators were nondominant. which would be consistent with a withdrawal response. but they were also linguistically immediate. which is an approach response. They were Ics ... expressive vocally but more expressive kinesically. They also "leaked"' some tension and unpleasantness when equivocating. These combined patterns pointed to equivocation including channel discrepancies. which makes sense given that people arc likely to equivocate when they arc feeling ambivalent.
Further Vocal Analyses from Experiment 4.
Our own and others' research had already repeatedly established that vocal behavior can be a major separator between deceptive and truthful speaker!o.. The next investigations (Rockwell. Buller, & Burgoon. 19973, 1997b) therefore sought to identify reliable indicalOrs of deceit using acoustic and perceptual measures. Vocal features generally fall into one of three categories of measures (I) those related to duration or speed of utterance. (2) those related to frequency or pitch. and (3) those related to intensity. Based on the expectation that deceivers would be morc reticent. withdrawn, and uncertain than truth tellers and would exercise greater control over nonverbal behaviors. we hypothesized that deceivers would exhibit shorter overall message durarioll. slower tempos. less nuency. and longer response latencies than truth tellers. At [he same time. the voice is not as easily controlled a.., the body, and previous findings have shown Lhat arousal can be leaked through higher pitch. We therefore also predicted that deceptive voices would be higher pitched. Finally, on the supposition that deceivers would try to "shrink" from discovery with a submissive demeanor. we thought they would speak more softly than truth tellers. Trained coders rated such features as nuency. response latencies. internal pauses, pitch. vocal quality, articulatory precision, and pleasantness. Acoustic software then analY7ed the recordings of the interviews. measuring such features as number of segments of sOllnd and silence. re ... ponse latency. fundamental frequency mean and variance. and
252
Part III • Seekil/g (111(1 Re,~isrif1g COlllplia/1ce: Sfr{m'gies (l1Il1 Taeries
intensity mcan and variance. Re~ults revealed thm deceivt;!r~ constructed shorter mes~age~, spoke more ~Iowly, were less Iluent. and had longer response latencies than truth teliers did. Unlike many other deception ... tudie~. deceivers did not show elevated pitch. but the acou"'lic analy ... is revealed that they di"playcd more pitch variety and a wider intensity (loudness) range than truth teller~. Additionally, deceivers' voices were rated less plea",ant than those of truth teller~. There are several plausible explanations for these findings. For example. deceivers may have adopted these pallerns strategically so as to restrict the amount of information conveycd, to reduce their responsibility for the receiver's interpretation of what is said by being vague and uncertain. or to cover any verbal insufficiencies with an expres~ive voice. Alternatively, behavior pallerns may have been unint ended. Deception may have required greater cognitive cl'fort that prevented rapid responses and long messages. It may have heightened arousal in a manner that impaired efficient functioning of the articulators and prevented deceivcrs from controlling tluclUations in intensity. The indeterminacy of whether these vocal patterns bcst fit ~ strategic or nonstrategic interpretation eventually led to the design of Experiment 6.
Experiment 5. Meantimc. this next ~llIdy. reponed in Burgoon. Buller, White. Afifi. and Bu .. lig ( 1999). was underlakcn to exmnine further the extelll to which senders adapt their deceptive displays over time and to different receiver comlllunication styles. Several hypothe~es were tested: (I) Interactive deception displays dilTer from truthful ones only at the outset of interactions and approximate truthful displays over time: (2) deceivers adapt to receiver communication: (3) receiver involvement affects sender displays: (4) receiver judgments of truthfulness correlate dirt.!ctly with sender displays: and (5) senders' social skill~ affect their deceptive performance. We discuss the last hypothesis later in the chapter. We again recruited a highly reprcsentative community sample to conduct intcrviews. Interviewees were instructed to alternate between telling the truth and deceiving acros~ four blocks of three ljuestions. Half of the interviewees started with truth. so their truth (T) and deception (D) sequence was 'ITTDDDTTTDDD. The othcr half followed a deception-fiN order (DDDTTTDDDTTT). Subsequently. trained coders rat ed the recorded sessions on verbal and nonverbal involvement. Result ... were supportive of JOT generally and the hypothese,', specifically. Tnllh tellers were initially more lalkative than deceivers. but decei\er~ became more verbose O\'i.~r time and spent as much time talking b) the close of the inten-iew. Involvement followed a "imilar pattern. Deceivcrs were le~s involvcd than truth tellers initially bUI achieved COI11parable levels of involvement by the end of the illlerview. They also showed increases in involvement each lime they shifted fromtruthflll responses to deceptive ones. dcmonstrating that they were able to adjll ... t their level of involvement upward despite the normal inclination to become subdued when deceiving. Moreover. interviewee involvement closely followed interviewer involvement levels. indicating a strong tendency for truth tellers and deceivers to adapt in a reciprocal fashion to the interviewer's co mmuni cation style. The exccption~ were lhal. as predicted, deceivers and truth tellers compensa ted for the interviewer's low involvement by becoming more involved. But unexpectedly. interviewce~ who began the interview~ by deceiving also compensated for the interviewer's high involvement by becoming le!o.s involved too. possibly because they thought
Chapter 14 • /Ilterpew",a/ Deceplioll Tileary
253
the high involvement conveyed suspicion and made the interview seem more like an interrogation. By contrast, interviewees who began with truthful answers reciprocated the interviewer's high involvement, probably because the interviewer's behavior remained constant when the interviewee shifted to deception. thereby making it evident that the interviewer's communication was not tied to the interviewee' s behavior. These res ults demonstrated that both verbal and nonverbal behaviors followed I DT predictions.
Experimellt 6.
Up to thi s lime. we had been making assumptions that deception includes many strategic features. and indeed we had classified entire clusters of behavioral cues as strategic. Yet we had not attempted to verify directly that decei vers intend to manage their presentations in the manner we had been hypothesizing. This ncxt small -scale study was therefore intended to learn directly from deceivers what they thought they did to create more successful performances and to use this knowlcdge to refine Ollr '-)ubscquen t experiments. Business students were recruited to participate in or observe soc ial conve rsation s between friends or strangers about stich topics as " responsibility" and "the most unpleasant job" they had ever had to do. Participants designated as "se nders" were a~ked to be truthful on some topics and to deceive on others. Participants designated as "receivers" were merely asked to keep the conversation flowing . Observers watched the convcrsations through a onc- way mirror and rated senders after two target topics. Senders and receivers s ubseq uently rated the videotaped interaction on the same target topics. and senders we re personally interviewed about what they did to appear truthful. Res ults provided clear evidence that deceivers intentionally manage their communication. Senders reported giving less complete information when deceiv ing than when telling the truth and said their deceptive answers were less clear, direct, and rele va nt than their truthful ones. They reported trying to achieve a Ilormal appearance and to shift conversational responsibilities to the receiver. They thought they were as pleasant in deception as in truth. But they al so thought their conversational style was less express ive. less attentive. less smooth, more anxious, and less domina nt when deceiving. Participantreceivers and observers failed to detect many of these apparent changes. Receivers were particularly oblivious, seeing virtually no differences between truthful and deceptive performances. Observers were more sensitive to sender differences. see ing decep tive messages as less complete. less expressive, less other-centered, and more awkward than truthful ones. But discerning these differences made them no less willing to rate se nders as believable. Thus, senders were able to escape negative judgme nts through thei r stra tegic machinations.
Experimellt 7.
A quick perusal of the IDT propositions reveals that interacti vi ty is one of its central features: yet no experiment had tackled its role directl y. This next inveMigation (Burgoon, Buller, & Floyd, 200 I) did. One form of interacti vity follows a dialogue to monologue continuum. Under dialogue (i.e .. high intcractivity). senders and receivers exchange turns frequently and have relatively equal turn s at talk. Under monologue (i.e .. low interactivity), one person hold s forth for lengthy periods while the other lislens. Thi s experiment tested the effects of illleractivity by having underg raduate men and women conduct social conversations while engaged in either a dialogue or a monol ogue. Half of th e
254
Part III • Seeking and Re,\"islillJ,: Compliancc:
Strale~ies (llId Tactics
participants again became instant accomplices who deceived their partner on two of the four topics, using either a TrDD order or a DDTT order across the four topics. Afterwards. participants and trained coders rated sender communication and credibility. Results wcre quite telling. Senders reported better information and speech management in a dialogue than in a monologue, both initially and over time (although these perceptions were altered somewhat by the relationship with the receiver). Behaviorally. senders felt that they appeared more dominant. involved, and pleasant when dialoguing than when monologuing. especially with friends. Coder ratings of behavior and image management behaviors showed that senders displayed more involvement, dominance. pleasantness. and adaptation when dialoguing rather than when monologuing. Finally. receivers detected deceit less accurately when the interaction was a dialogue than when it was a monologue. Overall. these data support the IDT proposition that interactivity in the form of senders engaging in a dialogue with receiver~ enahle~ better strategic management of senders' message content. nonverbal behavior. and overall image.
Experiment 8.
This next investigation entailed two studies intended to continue examination of the dynamics of senders' strategic adjustments to receiver involvement levels (White & Burgoon) 2001). The fIrst study established participants' expectations, motivations, and typical behavior patterns under truth or deception. The second study paired students to engage in conversations similar to experiment 7. during which ~enders were either truthful or deceptive throughout and receivers increased or decreased their involvement level halfway through the conversation. One research question we posed was whether attitudes before the interaction differed between deceivers and truth tellers. They did. Deceivers felt more anxious and more concerned about their self-presentations beforehand. All five hypotheses were also supported. When beginning to deceive. senders showed less involvement than their truthful counterparts. but they increased their involvcment over time up to a level comparable to truth tellers. These involvement patterns are cOllsi~tent with a st rategic behavior management interpretation. in that senders may have initially attempted to mask their deceptiveness by suppressing behavior but eventually were able to adjust to a more normal pattern of interaction. They were thus able to satisfy competing goa ls of evading detection through withdrawal and creating an impression of normalcy through approach behavior. Interaction patterns were also responsive to those of the partner. If the receiver illcrea~ed or decreased involvement. so did truthful senders, thus displaying the kind of reciprocal interaction patterns that characterize normal conversations. Deceivers also malched partner increases, though to a lesser exten t than truth tellers. Deceivers were expected initially to compensate for partners' reduced involvement, that is. to show an offsetting increase in involvement and then to reciprocate. Results showed that they did compensate. but not immediately. and that they generally -"howed reciprocal decreases in involvement. Finally, as predicted, deceivers interpreted receivers' behavior as feedback to thcm about their own performance. Low involvement was interpreted as possible skepticism or suspicion, something that probably motivated deceivers to try harder and con~equently be less inclined to become as uninvolved as the receiver wa~. High involvement was interpreted as a sign of successful deception and probably reinforced deceivers' existing com-
Chapler 14 • Inte''lJenmw! Deception Theory
255
munication patterns, thus again leading to less adaptation by deceptive than truthful senders. We had expected deceivers to show more adjustments over time than senders, and there was definitely evidence or more variabilit y in deceivers' responses when faced with low in volvemcI1I from the partner. In retrospect, however. it makes sense that on ly those deceivers facing "negative feedback" might need to make adjustments. It is also possible that the demands of crafling a credible deceptive performance, combined with monitoring partner feedback and handling all the other usual conversational responsibilities. may have taxed deceivers' cognitive resources and hampered their ability to adapt. These alternative explanations warrant further investigation. because the former interpretation implies that senders had greater capacity to manage their pre~cntations in intentional and deliberate ways. that is. to behave stratcgicully, than does the latter. In sum. the experiments conducted 1O date support many IDT propositions regarding the dynamic and adaptive properties of interpersonal deception, their responsiveness to partner behavior. and the nature of strategic and non!-.tralegic deceptive displays. At the same time. they have produced enough curious findings. and raised enough unresolved questions, to invite further experimentation before lirm conclusions are drawn.
Suspicioll Cues A key tenet of IDT is that both sender and recciver are active participants in deceptive episodes and that se nders are responsive to receiver suspicion. If this is the case, then we should be able to document that receivers exhibit overt behaviors that correlate wit h their suspicio ns. What follows is a brief review of the findings related specifically to receivers' display!-. of suspicion.
Experiment I.
This first t"t or IDT. reported in Bu ller. Strzyzewski, and Comstock (1991). came upon the heels of studies by Buller and colleagues (1989) and Stiff and Miller ( 1986) that had investigated whether use of probing questions. which suspicious receivers might be inclined to use. elicited more detectable deception by senders. Contrary to expectations, previous stud ies had found that probing led receivers to judge senders as more. not less, truthful. Buller and colleagues reasoned that probing might tip senders off to receiver skepticism, leading senders to repair their performances by supprcssing arousal c ucs and increasing immediacy and pleasantness. Thc result would be an honest-appearing demeanor that wou ld defy dctection as deceptive. The current study used an interview format during which senders either lied or lold the truth to friends or strangers. To induce suspicion, half the receivers were told that interviewees are often less than completely candid and truthful when interviewed face to face and that the research assistant would signal them surreptitiously if the sender's responses were not matching a previously completcd version of the questionnaire. This would be their cue to ask follow-up questions about the senders' reasons for their answers to the questions, which were taken from a social desirability questionnaire. Afterward, trained coders ratcd the recorded interviews on a widc array of verbal and nonverbal behaviors. Results showed that probing did elicit behavioral changes. though not all in the hypothesized direction. Sendcr~ who wcre subjected to probing questions gave lo nger
256
Part III • Seekillg alld Resisting
COlllp/i(l//('('.'
Strlllegie,\ alld Tactic,\'
answers, made more pauses and speech error~. u~ed more illustrator gestures and body adaptor gestures (increased head nodding and shaking late in the interview). and displayed le~s pleasant facial expressions. The bul~ of the ...e cues suggeM increased arousal accOIllpanied by efforts to elaborate \'erbal ans\\ers and ~upport them with gestures. However. when senders thought receivers were sll',.picious, thcy used fewer gestures. reduced bodily activity, laughed less, and gave shorter answers-behaviors in keeping with greater reticence and cffoJ1S to mask arousal. Oyerall, thCIl, thesc results showed that deceptive displays are governed not solely by sender intent to be truthful or deceptive but aiM) by how the receiver behaves and whether the receiver is thought to be suspicious.
Experimellt 2.
This next experiment (described earlier) tested suspicion effects directly by creating three levels of sllspicion (low. moderate. or high) based on the stated likelihood that senders misrepresent the truth. We also created separate measures of (I) receiver belief that the sender is untruthful. (2) motivation to detect sender's deceit, and (3) degree of vigilance in observing sender behavior ":iO as to <.I ...... c, ...... uspicion more thoroughly. As Burgoon. Buller, Dillman, and Walther ( 1995) reported. different leve" of ""picion created different beha\ioral profile" for hoth sender... and deceivers. In a sampling of this very mixed set of findings, sender\ were vocally unpleasant and most physically actiyc when facing moderate su\picion but pleasant and inactive when facing high ":illspicion, Conversely. receivers were more fluent and had smoother turn ~witches under moderate than under high suspicion. Deceivcrs thus seemed tlustered under moderate ::,u"'picion. whereas receivers seemed so when they were highly ..,uspiciow.;. These perplexing results led us to speculate that moderate suspicion may create the highest level of uncertainty. which accounts for part of the difference in behavioral patterns. As for the behaviors that provoked su:-,picion among r~ceivers. there were quite a few, Receivers were more suspicious when senders were vague and uncertain. unpleasant vocally and kinesically, nervous and ten ... c vocally and kincsically, nonimmediale (ies ... eye contact, greater physical di"'tance). and generally behaving in ways deemed undesirable. Senders who hid their demeanor allayed receiver suspicions. These findings clearly signal the importance of actual verbal and nonverbal communicative practice in percep· tions of deceit.
Experimellt 4.
In this next suspicion experiment. \\c reverted to two level ... of su~picion (low or high) because of the number of olher faclOrs (three kinds of deceit. experti,c. familiarity. deception. and deception order) in thc de sig n. It will be recalled that ...enders alternated between blocks of truthful and deceptive answer~. which allowed U~ to look for changes across the interview and differences bt.":t\I,.'ecn truthful and deceptive responding. The results, reported in Burgoon, Bullor. Ebcsu. Rockwell. and White (1996). produced a number of very intercsting finding .... In brief. senders' and receivers' behavior differed ~igllificantly depending Oil whcther suspicion ""as present or not. Senders whose partners wcre not suspicious decreased their involvement levels when shifting from truth to deception. showing the typical pattern ... of deceil initially depressing involvement levels. If, however, partners were suspicious, senders were able to maintain a high level of involvement when shifting from trUlh to deception. consistent with lOT's contention Ihat deceiver~ are capable of ... tratcgkally ;.I(lju ... ting inyol\'cment toward Ihe level Iypical of
Chapter 14 • /lIIerper.\OlIa/ Deception Thea,..\'
257
trUlh teller~. An exception was concealers, who likely came acro~s as uninvolved becam,e they say liLLie. Additionally. ~ender~ seemed more submissive when the interviewer was su~picious. consistent with our contention that senders will adopt avoidance and reticence strategies to reduce their cOll\ersational obligations and. hence. their likelihood of being detected. The exception was that. when they were telling olltright lies. they took a more dominam tad... It seems plausible that when trying to "sell" someone on a lie. deceivers would become more assertive. This came at a price, however. Senders fabricating a lie in the face of ~uspicion wcre rated le~~ nuent than those committing other kinds of deceit. Other behavioral changes varied according to the type of deception and relational familiarity. As for receivers, su:-,picion caused them to remain more involved during deception. to be more expressive voca ll y but less expressive kinesically. and to be more informal. Other receiver behaviors varied depending on relational fumiliarity, expertise. and type of deception. These behavioral patterns reveal that suspicion creates potentially noticeable change~ in receiver..;' nonverbal behavior. change . . that can be used to tip off sender. . that they need to make adjuslIllent ..... even when receivers' verbal behavior does not give away th eir suspicion. lOT posits that receivers often telegraph their suspicion" to "enders and that it i!\ this cycling between deceivers' and receivers' 11100es and countermoves that ulti mately allows deceivers to gain the upper hand through strategic adaptation of their behavior. The remaining experiment~ examined suspicion indirectly by manipulating receive rs' interviewing behavior. A~ noted previou~ly. senders were responsive to receiver communication styles and read those . . tyles as feedback about their own performance. Thus. whether . . uspicion was induced directly or "created" by having receivers act the way suspiciolls receivers aCI. deceiver...
Experimellts 5 Through 8.
Outcollles of Illterpersollal Deceptioll II/itial al/d Ollgoillg Credibility Judgmel/ts We have already discus . . ed receiver . . ' judgments during interpersonal deception as it relate" to context and relational factors . Hml,c\'cr. in lOT. judgments of senders' credibility
258
ParI III • Seeking and Nesisril1g Compliaflce: Slrarl'gies and Taclic.\'
(i.e .. hone"ty) arc considered to be dynamic. Context and relational faclors have their greateM impact on initial credibi lity judgments. As conversati ons unfold , recei vers' judgments o r the honesty or senders and their messages become increasingly influenced by se nder performance. To th e ex tent that senders ' behavior departs from what is expected, receivers sho uld alter th e ir judgments accordingly. However. because senders are monitoring receivers for feedback on their success at deception in order to adjust and repair their commu nica ti on. detection ab ilit y is like ly tb decline over time in co nve rsations (B ull er & Burgoon. 1996b: Buller. Burgoon. Afifi, White, & Busli g, 2002). Our research has shown co nsistentl y thar receivers recog nize deception when it is present. However. th ey usually tend to judge deceptive messages as less truthful but not ac tuall y deceptive (i.e .. judgments of both deceptive and truthful message s fall within the truthful portion of the con tinuum) (B urgoo n et aI., 1994). In one study. as deceivers shifted from truth to deceptio n. receivers' j udgmen ts of honesty also shifted upward and downward correspondi ng ly (Bu ller et al.. 2002), indicating receivers' awareness of the shift ing truth value of senders' responses. Among the factors influencing accuracy that we have co nfirm ed are type o f deception (falsification, equivocation , concealment), suspicion. and expertise (Burgoon et aI., 1994). Receivers judge equivocal answe rs as least hone~t. concealments as most honest. and ou tri ght falsifications somewhere in between. Suspicion not only often fail!-. to improve accuracy, it often hinders it. tn one study, expert deteClor~ (m i Iita ry imc ll ige nce special ists) were least accurate when suspic ious. Recei vers were especially harsh on eq ui voca l answers when suspic ious. but suspicion undermined th ei r ability to detect co ncealme nts. Finally, suspicion led receivers to consider acquaintances more honest and st rangers less hon esl.
Effects of Skill alld Motivatioll alld Detectioll Accuracy
011
Credibility Judgmellts
Given the number and complcxi ty of tasks facing communicators in conversation. lDT holds that the ability to persuade others of one's truthfuln ess or to make accurate evaluati ons of se nde r credibi lity are ~ killed behaviors. Our researc h has bolstered others' findings that skilled sende rs are beller able to c raft an honest-a ppearing demea nor and so to escapc detection (Bu rgoo n. Bull er, & Guerrero. 1995 : Burgoon, Buller, Guerrero, & Feldman. 1994). We havc been unable to determine, however, that receivers who have better social sk ills are better able to recogni ze and interpret deceit, leaving open the question of whether native abilities affect de tecti on accuracy. Research by others has documented that with the right kind of training. receivers can improve their accuracy. which s ll gge~ t s that receivers may need to acquire skills through experience and training. Sender motivation to deceive SlIcccf.,sfull y is another factor that can influence receivcr~' abilities to make accurate asscssments of sender c red ibility. An early co nception of the motivation-accuracy relationship he ld that motivated senders were actually less successfu l at deception. This motivation impai rm ent hypothesis he ld that motivation impaired nonverbal performa nce during dece ptio n (prod ucin g more nonstrategic behavior), even though it improved ve rba l pe rform ance (De Paulo & Kirke ndol. 1989). This theory is similar to the proverbial "choki ng under pressure" notion but applied only to nonverbal behaviors and 10 deception (Bu rgoon & Floyd, 2(00).
B) conlra ... t. lOT prm Itlt: ... i.I tllIlen.:nt \ ie\\ of mOl i\ation: IL i-. predicted to facilitate deception ... uccc ...... (Buller & Burgooll. 19960: Burgoon & Floyd. 2(00). A':. 'Iv'e have rc~ \iewed. Illany non\'eroal beh;j\ior ... are contrnllabk and can be strategically dcployed to impro\e ueccption ... ucce ....... Generi.ill~ "'peaJ...lng. then. Il1mi\';'lIion ... hould increa ...e ... tmtegic beha\ ior and i mprm c cOllllllunll:ator, ' con\ er'lJtional pt:rformance both \\ hen telling the truth anti \\ hen decei\ ing. (Thi, i, not to den) that e\trell1C le\ t.!b of motinllion may inler~ fcre with efft:ctl\'c produdlOn not ju,t ol ... oJ1le I1(Hl\'t.!rhal heha\'iop., but al,o of 'Ollle verbal one ..... ) E\idcncc from ollr n:'can:h program hear, out the ... e IDT prediction .... In one . . tudy. dl!t'ei\cr ... \\'ho reported hcing more motivatcd \"cre 1.11 ... 0 judged more . . ucce ... sful by oh . . efvef':. (Burgoon et al .. 19(5). Another . . tudy found that motivation. mea,ured with multiple.: indiciltor~. enhanced . . emle.:r ... · management of their ht.'!IHl\ ior . .II1d image in'espcc~ ti\e of whctht.!r the) were dccci\ ing or tt.!lling the truth. There wa!-. no evidence that any form of 1110tl\aIl011 harmed an) fi.It.:et of ... end!!r ... ' \croa l and nonverbal ~rronllancc. and thc onl} ad\cr ... c impact on l.'redioilit) occurred among truthful and dcceptin:: . . ender ... who "cre merl) COIl(CfI1I!U \\ ith J...l.'eping their urmhal in checJ... (Burgoon & Floyd. :WOO), Thi ... e\ idenc!! heller fit ... the lOT claillllhat 1ll01l\atHHl facilitate" ... lratcgic di ... play . . in C0I1\Cr':.3lion, rather than the motivatioIl impairment hypothe,,, .
Conc/usion In taJ...lllg a C0l11111llllicatioll per'peCII\e on deceptIOn. v.e h~t\c huilt a far richcr and more complicated \ iew of thi . . all ~ hll)-C(lI1ll11nn form of information managcment. It hu'-I forced u, to recoll,idl.'r 'Ollle uf the fumlul11ental i....... uc . . In c011lmunicatiun . . uch a ... what makes communication Illteracti,c. what lI1ultiplicity of communicator goab anti r~ ... pon ... ibililics . . hape interpcr'onal interaction. and hm\ mutual influence in cOllver ... :'l1inn i~ enacted. The,c que ... tion . . have complex
~
TAHLE 14.2
Experiment.'!, Variables, amI Citations/or Experimellts Testing IlIterpersonal Deception Theory
:>
Experiment prc~
lOT
pre~
lOT
2
3
Independent Variables
Dependent Variables
Source
Relationship (intimate/friend/ stranger) Sender deception (yes/no) Interaction time
Sender nonverbal behavior
Buller & Aune. 1987
Sender dect:!ption (yes/no) Probing (yes/no)
Sender nonverbal behayior
Buller. Comstock. Aune. & Strzyzewski. 1989
Part icipation (partici punt/obserycr) Sender nonverbal behavior
Receiver honesty Judgments Receiver accuracy
Buller. Str;ryzewsli. & Hun<;aker. 1991
Sender deception (ye~/no) Probing (yc~/no) Suspicion (yes/no)
Sender nonverbal behavior
Buller. Strzyrcwf'ki, & Comstock, 1991
Participation (participant/observer) Sender nonverbal behavior
Receiver honesty Judgments Receiver accuracy
Buller &
Sender deception (yes/no)
Sender nonverbal behavior
Burgoon & Buller. 1994
Receiver suspicion (high! moderate/low)
Sender and receiver Nonverbal behavior
Burgoon. Buller. Dillman. & Walther. 1995
Sender language
Buller. Burgoon. Buslig. & Roiger. 1996
Sender social skills
Sender behavior
Burgoon. Buller. Guerrero. & Feldman. 1994
Sender deception and type (tru th at time I: fabrication/concealment/ equivocation/general deception at time 2)
Sender infomlation Managemcn[
Burgoon. Buller. Guerrero. Feldman. 1996
Language
Buller. Burgoon, Buslig. & Roiger. 1996
Hun~aker.
1995
Ann. &
TAHLE 1-'.2
Continued Independent Variables
Dependellt Variables
Source
Sender deception (yc~/no): Deception type (fabrication! concca lmc nt1 eq ui voc ar ion )
Sender language
Buller. Burgoon. Buslig. & Roiger. 1996
Sender ~OCi
Sender behavior Receiver accuracy
Burgoon. Buller. & Guerrero. 1995
Receiver suspicion (yes/no)
Receiver behavior
Burgoon. Buller. Ebesu. While. & Rockwell. 1996
Receiver accuracy
Burgoon. Buller. Ebesll, & Rockwell. 1994
Sender information Management and language
Buller. Burgoon, Buslig. & Roiger. 1994
Further analysis of acoustic and perceptual vocal features
Sender vocalic behavior
RockwelL Buller. & Burgoon. 1997a. 1997b
Sender deception (yes/no) and order
Verbal and nonverbal invo lvement
Burgoon, Buller. Wh ite. Alitl.. & 8uslig. t999
Receiver involvement
Same
Same
Sender social skills
Same
Same
6
Deception (yes/no)
Sender verbal and nonverbal behavior Sender intention~
Burgoon, Buller. Floyd. & Grandpre. 1996
7
Sender deception and order
Verbal and nonverbal behavior
Burgoon. Buller. & Floyd, 200 1
Sender participation (dialogue/ monologue)
Same
Same
Deception (yes/no)
Verba l and nonverbal behavior Credibi lit y
White & Burgoon, 200 I
Receiver invo lvement
Same
Experiment ~
R eanal)'!)i~
of Bavdas et al.
stud ies Equivocation (yes/no)
5
(TITDDDTlTDDDIDDDn,'DDDTlT)
CrfDDIDD'n)
8
-
l>:
262
Pari III • Seckillg (l1/l1 R/!.\isril1g C(}lIIplillllU': SrrtlreMies and T(luic:J
applicability beyond face-to-face deception. For example. the rise of new communication technologies h;.l\ moved to the foreground i... sues of how interactivity in various modalities (e-mail. audio- and videoconferencing) affech users' ability to discern truthful from deceptive me~sage ... due to the availability or lack of feedback and to opporlunilie!o> for editing and personali/ing content that facilitate greater evasion of deception detection. The number of questions descrving investigation seems endless as we peer into the cybcrfuture Ihat is close upon us. Taking a communication perspective on these and related issues should yield unique insights beyond those obtained from models that reduce deception to intrapersonal phenomena. It is hoped that by centering attention on communicat ion practices themsclve ... , a ... well as on the psychological and social factors that are their antecedent .... our under"ilanding of interpersonal deceit w ill be enriched and c larified.
Referellces_____________________________ BlJ\ela<". J. Ii .. Blad... A. Chovil. K. & YlullclI . J. ( 1990). EqlliL'o("(/1 ("(IlIIlIIlIlIi('(lIioll. Newbury Pi.lr~, CA: Sage Buller. D. B. (19S6). The clrcc\"-, of di .. traCliun during per..,uJ<,ion: A meta-analytic re\ ic"'. COIIIIIIUllil"llrioll M(//wgrapf/\, 53, 91-11 . 1.
Buller. D. B. (19S7l. Commlllllcatioll 3pprchcn<"illll and reaction" 10 rrnxcmic violations. JOllmaf nINonI'('rlml Bl'/tm·iol". J J. I ~-25. Buller. D. B.. & Aune. R. K. (19S7). Nonvcrbal cue~ to dc(.'eption among intimatc:-., friend:-.. and :-.tranger:.. Journal r~f NO//I'i,rlm! Hi'lIm·io r. II. 269-2YO. Buller. D. B.. & Aune. R. K. (1988). The dTech of vocalic" and nonverbal !.cn:.iti\ it)' on compliance: A <"peech accommodation theory explanation. HIIIIIW! ComlllltlliclIrir)// Re.H'lIrch. 1-1.301 -332. Buller. D. IL & Aune. R. K. (1992). The dfech of <"peech rate similarity on (:ompliance: Application of commun;cilt ion accolllmodation theor). IV('.HNII jrJllmal oj COflllll/llli("(lrirm, I. ~7-53. Buller. D. 13 .. & Burgoon, 1. K (1986). Tht: elfel'l" of vocalic.') and nOIl\crlxll <"en<,,;t;,;t)' on compliance: A repl ieat ion and c xtcn ... ion. 1111111011 COfllllllllliclIlioll Research. 13. 126- 144. Buller, D. B .. & Burgoon. J. K. (1996a). Another loo~ a1 information management: A rejoinder to MeCorna.:~. Le\ine. Morrbon. and Lapin~~i . COII/J1IllIIicarirm MOIIOgl,(ll}h.l. 63. 92- 98. Bulkr. D. B.. & Burgoo n. J. K. (1996b). InterpcNHlal Jeeeption theory. COIIIII1I11/i("{lrioll Theory. 6. 20J2~2.
Bull er. D. B.. BlIr~()on. J K., Alifi. w. A.. White. C .. &: Bu ... lig. A. (2001). The dynamite nature of vcraeit) judgment<". ~1i\[lU .. cript in ... uhmi ....,ion. Buller. D. B.. Burgoon . J. K.. Bu<"lig. A. L. S .. & Roiger. J. F. (199~). Interpcr:..onal deception: VIII. Further anal) .. i.. of non . . erbal and \·crhal corn:late\ of equivocation frolll the Bavela:-. et al. (1990) rc.,eareh. Jrml"lla/ of l"llllgilage and Social Psycl/%KI'. 13. 39fl---.l17. Buller. D. B.. Burgo()n. J. K.. Bu:.lig, A .. & Roiger. J. (1996). Te ... ting intcrper<"onal deception theory : The language of intcrpcr:-.nnal deception. Cmlll//llIli("(/rioll Theory. 6. 168-289. Buller. D. B.. Burgoon. J. K.. White. C .. & Eoc:.u. A. S. (1994). Interpersonal deception: VII. Behavioral profile:-. of fal<"ilicalion. equivocation. and concealment. journal of Lallgllage a/ld Socia! Psychology. /3. ~66· -.W6. Buller. D. B.. Com,lOd. 1.. Aune. R. K.. & Slr/) Icw ... ki. K. D. f 1989). The crfect of probing: on decei\ crs and trutIHclkr~. jO/ll"/la! ofNolIl'l'rha! lJelw\';or, /3. 155-169. Buller. D. B .. & Jlun<.,akcr. F. (1995). Interper"lon;tl deception: X Ill. Su<"picion and the truth-bim. of conversational parlicipant ... In J. Aitken (Ed.). IlIlmper.wJllal CO/llII//lllil"(/li()1I processes reader (pp. 239-251). Wc~tland. MI: McNeil. Buller. D. B.. Str/)/cw<"ki, K. D .. & Comstock. J. (1991 l. Intcrper:-.onal deccption: I. Deceiver.. ' reactions to recei\·e r.. ' w~pit:inn<" and probing. COII/lllllllinlliol/ Mmw~ra"h\·. 58, 1 -2 ~.
Chapler 14 • Interpersonal Deception Theory
263
Buller, D. B.. Strzyzewski. K. D .. & Hun ..aker. F. G. (1991). Interpersonal deception : II. The inferiority of conver... ational participants as deception detec tors. Comlmmicariofl Monographs, 58, 25-40. Burgoon, J . K. (1976), The idea l source: A reexamination of source c redi bility measurement. Central SUltl'S Spe£'cll JOllrl/ol, 27. 21.X}-206. Bu rgoon. J. K, (1978). A commu nicatio n model o f personal space violations: Explication and initial test . Hllman Communication Re_~earch. -/, 129- 142, Burgoon, J . K. (1993). Imerpersonal expectations. expectancy violations. and emotional communication. JOllrl/al of Llinguage and Social P,~y('h()l()gy. 12. 13- 21. Burgoon. J. K.. & Buller. D. B. (1994). Interpersonal decepti on: III. Effects of deceit on perceived com· munication and nonve rbal beha vior d y nami c~. )ourl/al of NOllverbal Behavior, 18, 155- 184. Burgoon, J. K.. Buller, D. B .. Dillman. L. , & Walther, J . (1995). Interpersonal deception: IV. Effects of . . uspicion o n perceived commun icati o n ami nonverbal behavior dynamics. HI/mall Commullication Re.\earcll. 22. 163- [96 . Burgoon, J . K .. Bu[l er. D. B.. Ebe~lI. A., & Roc kwe ll . P. ( 1994). Interpersonal deception : V. Accuracy in deception detection. Comm/lnicatioll Monographs, 6/. 303-325. Burgoon. J. K .. Bullcr. D. B.. Ebe~u. A.. Rockwell , P., & White, C. ( 1996). Testing interpersonal decep· tion theory: Effects of su<;picio n on nonverbal behavior and relational messages. Communication Thl'on'. 6. 2--lJ-267. Burgoon, J. K., Bulle r. D. B., & Floyd. K. (2001). Does participation affect deception success? A test of the lnteracti\ity principle . Humall Communication ResearclJ. 27, 503-534. Burgoon, J. K .. Buller. D. B.. Floyd, K .. & Grandpre. J . ( 1996). Deceptive realities: Sender. receiver. and ob-.ener ~r1<.pecti\ics in deceptive conver\ati ons. Communication Research, 23, 724-748. Burgoon. J . K .. Buller. D. B., & Guerrero, L. K. (1995). Interpersonal deceptio n: IX. Effects of social skill and nOll\ erba l commun ication on decept ion success and detection accuracy. Joun/ af of LAnguage Will Social P,I'yc/w/ogy (spec iul iS1<.ue). /4.289- 3 11 . Burgoon, J. K .. Bull er, D. B .. Guerrero. L. K.. Afifi. W ., & Feldman, C. (1996). Interpersonal deception : XII. Infornmtion management dimensions underlying deceptive and truthful messages. Communi· ("(Irion Monographs. 63. 50-69. Burgoon, J . K .. Bull er. 0. B .. Guerre ro. L. K.. & Feldman, C. M. (1994). Interpersonal deception : VI. Viewin g dcception ~ ucce\ .. from deceiver and observer perspectives: Effects of preinteraclional and interactional factor ... Commullica/ion Stlitlie.I·. 45, 263-280. Burgoon. J. K .. Buller. D. B. , White, C. H .. Afifi, W . A .. & Bu slig , A. L. S. (1999). The role of con versa· tional involvement in deceptive inte rpe r~o nal co mmuni cation . Personalit), and Socia/ Psychology 811l1etill. 25. 669-685. Burgoon. J . K .. Buller. D. B.. & Woodall. W . G. (1996). NOl/ve rbal c:omnllmicatioll: The Ill/spoken dia· /()f.:lle (2nd ed.). New York: McGraw·llill. Burgoon, J . K .. & Floyd. K. (2000). Test ing for the motivation impairment effect during deceptive and truthful interaction. Western Journal of Comnllmica/ioll. 64, 243-267. Burgoon. J. K.. & Hoobler, G. (in press). Nonverbal signals. In M . L. Knapp & J. Daly (Ed.), Handbook of i/l/erpenmllli commlmiclll;OIl (pp. 344-390). Thousand Oaks, CA: Sage. Burgoon. J. K .. Stern. L. A .. & Dillman , L. ( 1995). IlIl erpersollaladapltllioll: Dyadic illierac(ioll pal/ern.... New York Cambridge University Press. DePaulo. B. M .. & Kirkendol. S. E. ( 1989). The motivational impainnent effect in the communication of deception. In J. Yuille (Ed.), Credibility (lssessmell1 (pp. 51-70). Deume. Belgium: Kluwer. deTurck. M. A., & Miller. G. R. (1985). Decepti on and arousal: Isolating the behavioral correlates of de· cepti on. HI/Illan Commllllic(l{iOIl Rl'search. 12,181 - 20 1. Dunhar. N. E., Ramirez. A .. Jr .. & Burgoon. J. K. (i n press). The effects of partic ipation on the ability to judge deceit. Commllnication Reports. Ekman , p , (1985). Telling lies. New York: Norto n. Ekman. P .. & Frie ~e n. W . V. ( [969). Nonverballcakuge and clues to deception. PsychiallY. 32. 88-105. Knapp. M. L. . & Comadena. M. E. ( 1979). Telling it like it isn't: A review of theory and research on deceptivc communication. Hllman Cmmmmic;(l{ion Research. 5. 270-285. McCornack. S. A .. & Park s. M. R. ( 1986). Decept ion detection and relationship development: The other
264
Part III • Seeking and Resisting Compliallre: Strategies and Tactics
side of trust. In M. L. McLaughlin (Ed.). Communication Yearbook 9 (pp. 377- 389). Beverly Hills, CA: Sage. Stiff. J. B.. Kim, H. L & Ramesh, C. N, (1989, May). Truth-biases alld aroused suspicion in relational deceprion. Paper presented at the annual meeting of the International Communication Association, San Francisco. Stiff. J. B.. & Miller, G. R. (1986). "Comc to think of it ... ": Interrogative probes, deceptive communication. and deception detection. Humall Communicatio/l Re.\·earch, 12, 339- 357. Toris, D .. & DePaulo, B. M. (1984). Effects of actual deception and suspiciousness of deception on interpersonal perceptions. jOtlmal oj Penonaliry and Social Psychology, 47, 1063-1073. Rockwell, P.. Buller. D. B., & Burgoon. J. K. (1997a). The voice of deceit: Refining and expanding vocal cues to deception. Communication Re.\"earch Reports, 14, (4). 451--459. Rockwell. P., Buller. D. B .. & Burgoon, J. K. (1997b). Measurement of deceptive voices: Comparing acoustic and pen:eptual data. Applied P5ycllOlinglti!;·tics, 18, 471--484. While, C. H .. & Burgoon. J. K. (2001). Adaptation and communicative design: Pattems of interaction in truthful and deceptive conversations. Human Communicatioll Research, 27. 9- 37. Zuckerman. M., & Driver, R. E. (1985). Telling lies: Verbal and nonverbal correlates of deception. In A. W. Siegman & S. Feldstein (Eds.). Mulric/iwlI1el imegratiol1s of 1101ll'erba! behavior (pp. 129~148). Ilillsdale. NJ: Erlbaum.
15 Inoculation and Resistance to Persuasion Michael Pfau and Erin Alison Szabo
Unlike other theories in this book, which foclls on explanations of influence. inoculation is a theory about how to prevent influence. Inoculation. in a nutshell, is a strategy for re~isting influence auempts. It motivates receivers to bolster their beliefs and attitudes. thereby rendering them less susceptible to influence. Since the 1920s both academics, who investigate influence, and practitioners. who usc ii, have expended considerable lime and resources to discover, understand, a nd perfect new and beller means of persuasion. Although it has received much less emphasis than persuasion (McGui re. 1964: Miller & Burgoon. 1973). resistance to persuasion is simply the flip side of the same coin: persuasion is an exercise of inOuence: resistance is abou t protecting people against influence. This chapter wi ll explore one approach to resisting persuasion: McGuire's Inoculation Theory. The theory posits that refiualiOlw/ treatments. which both raise and refute counterarguments to a person's attitude, confer resistance to influence (Pfau, 1997). inoculation Theory embodies an elaborate and systematic approach to resistance. Eagly and Chaiken call inoculation. "the grandparent theory of resistance to attitude change" (1993.
p. 561). This chapter will first trace the origins and describe the nature of Inoculation Theory and then summarize research findings on inoculation to date. Next, it will explore unresolved questions concerning inoculation and the process of resistance. Finally. the chapter will address applications of In oculation Theory in a variety of actual contexts.
Origins and Nature of Inoculation In ocu lation traces its origins to early research on the relative superiority of one- versus two-sided messages. One-sided messages simply reinforce attitudes a person already
265
266
Part III • Seeking and Resisting Compliance: Strmegies and Tacfics holds. For example, a parent might tell a nonsmoking adolescent that smoking causes cancer. Two-sided messages raise arguments contrary to a person's altitudes, called counterarguments, and then offer arguments and ev idence to refute those counterarguments. For example, in addition to the "cancer" argument. an adolescent might be told, "Your friends may tell you that you can quit smo king anytime. but nicotine addiction is extremely difficult to overcome." In a classic experiment on message "sidedness," Lumsdaine and Janis (1953) found that one- and two-sided messages were comparable in influe ncin g people who were not subsequently exposed to an opposing point of view. Two-sided messages. howeve r. were superior in "producing sustained opinion changes" in people who were subsequently exposed to an opposing point of view (p. 311). That is, the two-sided message co nferred lasting resi stance to persuasion. The critical question was. \Vhy? The researchers speculated that two-sided messages may " inoculate" people, thereby making them more resistant to counterpersuasion. It was the aftermath ofthe Korean connict that provided the impetus for Inoculation Theory. During the war, some American POWs had broken. yielding to pressures from their North Korean captors to renounce thei r country (United States Senate. 1956). Congressional hearings following the war raised alarm about the seeming effectiveness of North Korean "brainwashing" techniques. How mi ght such brainwashing be prevcIlIcd? This question was the catalyst for McGuire's interest in ways to instill resistance to propaganda and other form s of influence (McGuire & Papagcorgis. 1961). McGuire devised Inoculation Theory as an explanation for resistance to inlluence. He formulated a biological analogy to ex plain how inoculating messages might confer resistance: "We can develop belief resistance in people as we develop di sease resi stance in a biologically overprotected man or animal; by exposing the person to a weak dose of the attacking material strong enough to stimulate his defenses but not strong enough to overwhelm him" (1970, p. 37). Ju st as injecting a person with a small dose of a flu virus can build up the person 's immunity to the t1u, exposing a person to weak arguments and refutation on an issue can increase the person's resi stance when she or he is later exposed to stronger arguments on that issue. Early applications of Inoculation Theory argued that these effects were particularly true for "overprotected" people. Note that, if a person has been rai sed in a germ-free environment, he or she will be more vulnerable when exposed to germs than someone who grew up less protected. In the same way, some beliefs. which McGuire called "cultural truisms" have never been challenged and may therefore be more vulnerable when attacked. In other words, cultural trui sms are beliefs that are uncontested and therefore taken for granted. Early on in his research, McGuire focused on such trui sms as the benefits of having an annual physical; the merits of penicillin; use of chest x-rays to prevent tuberculosis ; and the need for regular dental hygiene (McGuire, 1970). Although McGuire was tempted to frame Inoculation Theory much more broadly (McGuire, 1961 a), research findings of the period cast doubt on whether inoculation protected beliefs other than cultural trui sm s (Brehm & Cohen, 1962). Thus, McGuire narrowed the boundaries of Inoculation Theory to c ultural truisms (McGuire. 1970). Later, Pryor and Steinfatt (1978) extended the theory be yond cultural truisms. They maintained that McGuire's rationale for restricting inoculation to "germ-free" (e.g .. un-
Chapler 15 •
IlIo('UfllllOllllllll Rt!.,I\lWIC"(' 10 Pt'nl/{I.\IOII
267
contested) beliefs was faulty: that the biological analogy applied to the ... pecific argumen .... that might be raised in an inoculation treatment. not the topic ihelf. Pryor and Steinfatt\, results failed to support the \ iability of inoculation on behalf of Illiddh> or higher-Ie\cl beliefs. However. other researcher ... were able to confirm that re./wmiolla/ preemptioll. one of the key uxiollls of the theory. was effective in fostering rc ... i... tancc to influence. evcn with controversial topics (Adams & Beatty. 1977: Anatol & Mandel. 1972: Burgoon. Burgoon. Ricss. Butler. Montgomery. Stinnctt. Miller, Long. Vaughn. & Caine. 1976; Burgoon & Chase. 1973; Burgoon. Cohen. Miller. & Montgomery. 197H; Burgoon & King. 1974; Crane. 1962; Cronen & LeFleur. 1977; Hun!. 1973; McCroskey. 1970; McCroskey. Young. & SCOII. 1972; Miller & Burgoon. 1979; Sawyer. 1973: Szybillo & Heslin. 1973; Tale & Miller. 1973; Ullman & Hodaken. 1975). Inoculation Theor) posit ... that Iwo components contribute to rl.!si ... talll:c: threat and re/Ittaliollll/ preemption. Thrcut con",ish of warning a pefl'lon that hi ... or her existing attitudes arc lik.ely to be challenged. For thrl.!t1t to work.. these challengt!.., must be.:: ... urficielllly powerful 10 make people accept that their exi:-.ting attitudes may be vulnerable. Thu.s. the threat serves a'" the mOliwlliollOI Irigge,. in the inoculation model. It mOlivale.s the indiv idual to strengthen his or her attitude ........ t!tting in mollon the internal proccs", of rcsi~ tance. The "iecond element. refutational prt!t!mptiol1. il1\olve", the pro(;e:o."i of rai..,ing. and then an ... wering. specific objections. The refutational preemption component function ... much like the two-"iided message: It rai ... c ... oppo"iing argull1ents and tht!n prm ides the arguments and e\idence to refute them. Thc",e integral component ... thn.~at and refutation;]1 preemption. work. in tandem : Threat 1110{J\'u tt!s the indi\ idual to bol .. ter hi", or her attitudt!:-.: refutational preemption offer... "'pecilic content that can be u .. t!u to protect and defend one'", attitude .... Of the two clements. the threat component may be thl! more important, because it i.. the "motivational effect'· thut bllost~ a per"ion·:-. immunity to all form .. of attacks. "If the construct were limited to preemptive refutation. it would afford limited utility ",ince COI11municator", would need to prepare "'pccific preemptive mes\agt!.., cOITe"'ponding to each "nd every anticipated "lIack" (PI'"u & Kenski. 1990. p. 75). lIowever. due 10 the "motivational etTcct:' inoculation unlca",he", a praces ... that o'Tt!r", a broad umbrella of protection. extending beyond the ... pecific arguments raised in the opp<)..,ing message. Inonllation increa",es receivers· resi ... wnce to novel. unfamiliar counterargumen .... a\ \\ell. The best e\idence of this broad umbrella of protection j<.. found in the re:o.ults of studie", that reveal comparable efficacy for both relwatiolllli same and tlWl'rellf inoculation treatment ... Rcfutational "same" trealmcnh :-.y:-.tel11atically rcfute one or mort: "'pecific counlerargumellts that will be encountered, whereas refutational ·'diffen.::nt" l11essage.s art! more generic in nature. For example. during the 1988 pre:-.idential campaign. polling dma indicated that Democrat Michael Dukaki ... wa ... vulnerable to Republkan attack. for being ",oft on crime. prau and hi .. colleagues (1990) prepared a refutational ·· .....lIne .. inoculation me ... sage that rai",ed this counterargument and preemptively refuted it. The l1le.s",age argued that Dukak.is favored tough sen tences. but that tough sentences i.llonc did little to reduce crime. The message also argued that drug ... are res.pon ... ible for l1lo .. t crime and that Dukaki ... fa\ored increased funding for drug education and enforcement. The researchers also prepared a refutational ··different"· message. arguing that in contra ... t 10 his opponent. Michael Dukaki .. believed that it was time for a change: that America call. and mu",t, do
268
ParI III • Seeking ond Resistillg Comp/ialla: Strmegie.\· and Tactics beller to extend the economic expan~ion to those who have been left oul. provide basic heulth insurance for all families, and stop the pollution of America's air and water. The resull~ of the study revealed that. fo llowing inoculation, when Dukakis supporters were confronted with a Bush message attacking Dukaki~ for being soft on crime. they were less persuaded by the attack than comrol participants who were not inoculated. The effectiveness of the refutational '"same" and "different" inoculation messages was comparable. Indeed, the early laboratory research on inoculation (McGuire, 1961 b, 1962, 1966: Papageorgis & McGuire, (961), later "eld research (Prau, 1992: prau & Burgoon, 1988: Pfau, Kenski, Nitz, & Sorenson, 1990: Pfau, Van Bockern, & Kang, (992), and more recent experiments (Pfau, Tusing, Koerner, Lee, Godbold, Penaloza, & Yang, 1997a; prau, Szabo, Anderson, Morrill. Zubric. & Wan. 2000) have all indicated that refutational "same" and "different" treatments arc virtually identical in their capacity to confer resistance to influence. Additional evidence for the crit ical ro le of threat in resistance comes from two areas. First, a number of studies have confirmed a positive relationship between threat and re..,istance, Studies that ovenly manipulated threat conditions (McGuire, 1962, 196-1: McGuire & Papagcorgis, 196 L (962) or the presence of threat levels (Pfau & Burgoon, 1988: Pfau et aI., 1990; Pfau et al.. 1992; Pfau el aI., 1997a: Pfau, Szabo et aI., 2000) round that greater threat enhances resistance. Second. research findings have suggested that communication increases attitude persistence. not so much due to the spec ific conte rll emphasized but by motivating people to generate their own thoughts about the topic (Love & Greenwald, 1978; Petty, (977).
Empirical Support for Inoculation Early research on inoculation compa red the effective ness of supportive (bo lstering) versus re!utafional (inoculation) approaches to rcsistance. The supportive approach seeks to reinforce existing beliefs and attitudes. If a person believes that a ballistic mis~ile defense sy~ tem is desirable, a supportive approach wou ld provide arguments and evidence to bolster this position. By contrast. the refutational or inoculative approach attempts to threaten people's attitudes by warning of possible challenges to altitudes and then raises and preemptively refutes these challenges, The person who suppons a balli"ic missile defense system would be told that she or he can expect to encounter strong arguments oppo~ing her or his position. Then specific objections would be raised and immediately refuted. Studies direct ly comparing the effectiveness of the supportive and refutational approaches have generally indicated that the refutational approach is superior (Anderson & McGuire, 1965; Crane, 1962: McGuire, 1961 a, 1962, 1966: McGuire & Papageorgis, 1961, 1962: Papageorgis & McGuire, 1961; Sueureld & Borrie, 1978: Tannenbaum, Macaulay, & Norris, 1966; Tannenbaum & Norris, (965), although a number or these stud ies revea led that the use of both approaches was better than the use of ei th er one alo nc (McGuire, 1961 a, 1962; Tannenbaum & Norri s, (965), These studies established the viability of inoculation and in the process resolved a nagging caveat aboLit use of inoculation, namely. whether prompting people in regard to arguments that are contrary to attitudes might unwillingly produce attitude change.
Chapter 15 • Illom/arion alld Re,\ist(llU'e to Per.\Uosiol/
269
Early ~ tudi cs also explored the persiMe ncc of inoculation treatme nt s mcr time. revea lin g that although treatme nt s did deteriorate over time. refutati onal "diff(!rcn!" me . . sages decayed less than rerutational '·same" message, (McGuire. 1962. 1964. 1966: Pr)or & Steinfatt, 1978). Also. stud ies ex plored the pOlelllial of reinforcemcnt, or booste r se ... sions. anempting (Q di scover whethe r it was pos . . ible to prolong a rc ... i ... wnce effect. The an,wer was a qu alified ··ye,:· McG uire ( 196 1b) repo rted that ··doub)e deren,e,·· provided additional reinforce ment. but only with refut ational " sa mc" treatments. Tannenbaum and coll eagues ( 1966). however. revealed that "concept-boost" mcssage~ ... tn::ngthcned re ... i... lance. although lhe effec t fell !-.hort of slali"'lical significance. All in all. re~earc h indicates th at inoc ul ation b amazingl) robu~1 in conferring re\i ... lance to influence. Although trea tments need 10 incorporate tht! core elel11en t\ of threa t and refutational pree mption. it see m ... to make very lillie difference how else inoc ul ati on treat ments are de ... igned. Researc h ~uggest s com parabl e efficacy for inocula tion treatment... w hether they are wri tte n <.1\ refutat ional "same" or "diffe rent" (see work ... c ited above). designed 10 elici t central or peripheral message processing (Prau et al.. 1997b). built to emphasize a message's con ten t or source (Freedma n & Scar~. 1965: Stonc. 1969). constructed as cog niti ve or affective positi ve/negati ve (Lee & Pfall. 1998) or affective happyl angry ( Pfau . Swbo e L a l.. 2000). or de li vered via print or \ ideo (Pfau. Holbert. Zubric. Pas ha. & Lin. 2000).
Questiol/s about the Process of Resistance Recen t year~ ha ve w itn e~sed a resurge nce of illl ere~t in inocuiatioll. SOI11I! research re turned to the corc axioms o f inoculation. ~eek in g to un cover further nuance ... abo ut th e process of re~is tan ce and updating the theory on the ba~i\ or the new knowledge about soc ial innuencc that has accumulated since the 1960s. This ... cc tion of the c hapter wi ll focus on these effo rt ~ to under~tand the proce ... s of resistancc.
The Core Compollellts of Resistallce As indicated above, [hreat and refutat ionC:i 1 preempti on arc core elemen t.., of In oculation Theory. Threat \erves a~ the motivational cata l y~t for im'K:u lation. Refutational precmp· ti on provide.., co ntc nt that people might use in order to stre ngthen their alt itudes and afford s thel11 an opportunit y to pructice defending them. Thu .... the proce~~ of re ... istance i.., part moti vati onal and part cog nit ive. Howeve r, much morc needs to be learned about the inne r workings orthe proce", (Eagly & Chaike n. 1993). Rece nt ... tudies have focu ...ed on the role of these and (Hher critical elements in the process of resistance. It i... known. for example. that threat motivates receiver\ to bol ... ter beliefs. but how? Precisely what imenral pl'Oce,'\s does threat trigger? b the proces~ a ll cog nitive. c,llIsing people to co me up with cou!ltcra rgume nts to defend th eir attitudes. or i... it both cog niti ve and arrective '! prau and his colleague, ( 1997a. 2000) tri ed to answer th e~e questions by pinpointing the ro le and impact of threat. refutational preemption. and other potentially cruc ia l element ... in the process of resistance.
270
Part III • Set'king (llId ResisTing COlllplianCl': Strateg;l'l
{lIId
{actin'
The initial \tudy (1997a) examined a variable that Petty and Cacioppo (1979b) called i,\,we illl'o/remellf. People tend to di ... play greater i\sue involvement when the outcome of an i...... ue affects them pcr... onall). The qucqioll examined in the "'lUdy wa ... whether inoculation· ... effectiveness hinged Oil rcceher invohement in an issue. Pfall and his colleague ... ( 1997a) ... peculated that involvemcnt might function a ... a prerequisite for rcsi ... tance. in other word .... that involvcment might dictate the boundary conditions for Inoculation Theory. Previoll\ research had hinted (hilt thi ... might he the case. Studies found. for example. that for inoculation to work. the receiver Illu ... t be involved in the topic or issue (Kamins & A-seal. 1987: Papagcorgi\. 1968: POllY & Cacioppo. 1979a: pr"u. 1992). On the other hand. the results of another study by prall and his colleagues (1997a) revealed that threat incre<.lsed resistance. primarily with I/Jodemle/r irn'olving topics. Specifically. with a moderately involving topic. threat accomplished two things. Fir>;t. it immediately and directly strengthened the altitude>; that were heing. threatened. Second. it caused people 10 comc up with arguments defending their attitudes (i.e .. more counterargument . . ). which also strengthened their rcsi>;tance to att:It':k. Thi ... effect was not nearly as pronounced with low-involving topic .... and it \\'<.1'" nOllexi ... tent \\ ith high -involving topic.;;. Why? prall and hi>; colleague . . rca ...oned that there i... an optimal Ic\el of involvement for inoculation 10 work. If im'olvement i ... too low. inoculation Cl.Illnot generate enough threat: if it is too high. inoculation i>; unable to generate further threat. In other \.\ords. if an individual already care ... about an i... sue. it is difficult to threaten her or him further. The pcr ... on's high level of involvement already ensures that he,;: or ... he i... alert and \igilant to oppo ... ing messages. In addition tn issue invohemcnt and the core clements or Inoculation Thcorythreat and refutational preemption-pratr and colleagues (1997a) also operationali7cd the process of counterarguing in terms of a pcr"':()[1's ability to generate reflltational responses to argumcnts contrary to exbting attitude .... McGuire a ... slImed that threat c li citcd "an un derlying process of covert counterarguing" in inoculated receivers (Eagly & Chaiken. 1993. p. 56-\.). Although he often employed the term cOIlllferargllillg. to explain the process of re ... istanct:. he never operationalilcd it in his O\\.-n research. prau and colic ague.'. . ( 1997a) re ... ults were mixed insofar as counterarguing was concerned. With the moderately involving topic. re ... uhs revealed that both inoculation treatlllen .... and is ... ue im'oh'emcnt elicited threat. which in turn contributed to countcrarguing and thereby to resistance to pcr... uasivc att<.lck~. Howc\'er. the re ... ulh were Ie ...... clear for the role of cOllntcrarguing with either 10\\- or high-involving topic ... . In addition to the role of threat. i.... suc involvement. and receiver counterarguing in the process of rt!sistancc. results revealed that rnoculation treatment<., make a direct. Silable contribution to resistance. Thi>; effect wa ... evident for each of the three issues in th e study. Thi ... finding suggests one of two possihilitie~: Either the refutational preemption component of inoculation treatments cOlllributcs independently to resistance. or there arc as yet unexplained element.;; in the process of resistance. In ... ko ( 1967) was the first to raise the specter of unexplained clements in resi>;tancc. He noted: "Beyond the.'.e defense alerting Ithreat elementl and defen ... t: producing Irefutational preemption element I mechanisms, it is entirely possible that other mechanism ... re ... ult in the creation of defen . . es ... " (p. 3 19 ): "a complctc explanation of re ... istancc to persua ... ion will depend upon the In ocll-
Chapter 15 •
If/ocillation lIml Re.w tllllce to PenlU/\'ioll
271
lalion Theory mechanisms as well as upon , , . additional mech:'lI1i ... ms" (p. 328). The next wave of inoculation studies will attempt to dctermine if attitude accesl..ibility or priming reveal further nuances in our understanding of the proccsl.. of re~istance .
Role of Affect ill Resistance The next investigation to illuminate the process of resi\tancc (prall. S/aba et aI., 2000) examined the role and impact of oJ/eel on resi stance. Specifically. the ~ tudy examined whether making people angry or happy altered their rcsi~tancc..: to i.I mes..,age that threat ened their attitudes. Lee and prau (1998) had initially i nvestigatcd th e relative effectiveness of cognitive and affective (positive and negative) inoculation treatmcnts in conferring resistancc against cognitive and affective attacks. The resultl.. re\ealed th at all three treatmenlS (cognitive. positive affect. and nega tive affect) co nferred re ... i..,tam:e. but their effectivenes~ varied depending on the type of persuasi\'e attack. encou nt ered. Lee and Pfuu ( 1998) reported that cognitive inoculation trcatment~ provided the most resistance overall. T he cognitive treatments were effective with cognitive and affective positive attacks, but not affective-negative attacks. However. results for the afft;:ctive inoculation treatments were obfuscated due to weak affect manipulations. The Pfau and colleagues' (2000) investigation also compared the c flicacy of cognitive and affective (anger and happine s~) inoculation treatments . Howe ver. this study focused mo re broadly on the process of rcsil.,tance. especially the role of elicited affect in resi stance. The researc hers reasoned that affect should play an in~trumcntul role in resi stance because threat functions as the l11oti\-ational catal)~t to re . . i . . tance (Pfau. 1997) and mot ivation is more affective than cognitive in nature (l7i.ud. 1993). The study predicted thaI all three treatment types would foster re ... istance. si nce they featured both threat and refutational precmption. but that th ey might do so in different ways. The res.earchers postulated that cognitive inoculation treatments would be more effective than affective-anger treatments. which in turn would be superior to affeclivehappiness treatments in fOMering coullterarguing. Howcvcr. rcsult~ indicatcd that all three treatment types triggered counterarguing and. thereby. rcsiMancc. The reason was that self-efficacy moderated resi stance outcomes. Self-efficacy. a measure of an individual' s perceived confidence in handling environmental obstacles ( Bandura. 1983). was thought to predict how a person responded to threat. Interesti ngly . affective-happine~s message . . did not work through the co re resi stance mechani sms of e licited threat or receiver counterarguing. The effectiveness of affective-happiness messagcs was also limited to receive rs with low self-efficacy. Affective-a nger inoculation treatments produced the greatest res istance in the high self-efficacy condition. whereas cognitive inoculation treatments produced a curvilinear effect. producing maximum resiMance at mooerate levels of selfefficacy. prau and colleagues (2000) found that iss ue involvement tri ggered receiver counterarguing and contributed to resi stan ce. providing fUl1hcr . . uppOI1 for the re sult s of Pfau et al. (1997a). especially in revealing that inoculation treatment, mak.e direct and indirect contributions to resistance. Inoculation treatments acted directly. independent of internal processes. to foster resis tance. In addition. treatment . . worked indirec tl y to promote resistance. The pattern of res ult, indicated that cog niti ve iJnd affecthc-anger treatments
272
Pan III • See!..;I1!: (lml Refi.HlI1g Camp/iWICt': 5tmtegil'_, ami Tactics
elicited threat. which in turn contributed to recciver counterarguing and elic ited anger. thu:-. indirectly fostering re~i"tance. The role and impact of elicited emotion in the process ofre ... istance was particularly noteworthy. The re ... ull\ revealed that e li ci ted anger and happiness exerted siLable. but opposite. effects: anger facilitating resistance. happiness inhibiling il.
Other COllsideratiolls ill Resistallce Other important issues in resi..,tance include the timing and persistence of inoculation and reinforcement treatments and the role and influence of communication modality in resistance.
Timillg alld Persistellce Questiolls The optimal timing of inoculation treatments has been the focus of a number of "'lUdies. but the issue remains unsettled. Because of the role of threat in resistance. some lag between treatment and attack is required in order to pennil people to generate cOlIIHerarguIllenls (Miller & Buron. 1973). How Illuch delay is unclear. especially in lighl of research indicating that inoculation. like any message stimulus. decays over time (McGuire. 1962: Prau. 1997: prau el al.. 1990: Pfuu & Van Boc~crn. 199.+: Pryor & Sieinfall. 1978). Early inoculation scholar~ reasoned that the type of treatment a per\on receives may impact ho\\< persistent his or her atlitude~ will be. McGuire ( 1964 ) and Manis and Blake ( 1963) found that the resi"itancc produced by in ocu lation "different" as opposed to "same" treatments increased following a modest delay prior to attack. In addition. McGuire (1962, 1966) ond Pryor und Sieinfall (1978) reported Ihal Ihe decay of inoculaliun-induced resistance could be reduced yia refutational "different:' as opposed to "same:' treatments. Lalcr "udics by Pfou ond Burgoon ( 1988) and Pfau el aL ( 1990) indicaled Ihm Ihe resistance conferred via inoculation "different" treatmenl\ persisted longer. at least with character attack~. Finally. research examined the potential of reinforcement. or booMcr ses:-.ions. seeking to determine if the inoculation effect could be prolonged. As noted previously. two early studie . . of bOOMer intlucnce produced weak results (McGuire. 1961 b: Tannenbaum et al.. 1966). Further. more recent ..,tudies by Pfall and hb colleagues featuring reinforcement or booster treatments (1990. 1997a. 2000) failed to clarify the matter. revealing no incremental persistence for reinforcemcnt messages. The reason may involve timing. All ... lUdie.., to date have admini..,tered booster session~ within a brief interval following initial treatment (maximum of four weeks). In order (0 test reinforcement fairly. future research ... hould increase the time interval between the administration of inoculation treatments and boosler messages and shou ld employ multiple reinforcements administered over time.
Role alld Illfluence alld the Commullicatioll Medium Wilh Iwo exceplions (Godbold & prou. 2000: prau el al.. 1992). all Ihe inoculalion research to date has employed print mes ... ages. The reason is that inoculation is ..,upposed to
I Clmptcr 15 • IIIO("I//afirm
{lilt!
Re.~israflce In Perl'llasi{J/1
273
be an active. cognitive process and print is perceived as the optimal medium for triggering this process. Following Medium Theory, prau. Holbel1 and colleagues (2000) reasoned that both print and visual media should function as effective vehicles for the communication of inoculation treatments, but that they might do so in different ways. Because print is considered the more effective medium for systematic processing. that is. actively thinking about a message (Chaiken & Eagly. 1976. 1983: Pelly & Cacioppo. 1986). Pfau, Holbert, and colleagues (2000) predicted Ihal print inoculation treatments would be superior in triggering counterarguing. Because video elevates visual over aural content (see Chesebro. 19R4: Meyrowilz. 1985: Salomon. 19R7). the researchers posited that video inoculation treatments would rely more heavily on source cues and less on counterarguing. These suspicions were confirmed. Both media conferred resistance. but in markedly diffcrcm ways. Video treatmcnt!, worked immediately. whcrcas trcatmcnts using print mcdia took more time. Video treatments fostered resistance based on the source of the message. Viewers displayed more positive perceptions of the source of the inoculating message. This in turn resulted in more negative perceptions of the credibility of the source of the counterattitudinal attacks. In contrast. print inoculation treatments worked via message conteill. requiring more time to foster resistance.
Applications of Inoculation Theory Becausc inoculation is designcd to protect people against influence. it has received increased attention in recent years in a number of applied contexts (Eagly & Chaiken. 1993). This section will focus on the applications of Inoculation Theory and research in the political. commercial. and health campaign contexts.
Political Applications The scope and intensity of attack messages in highly visible U.S. Senate and presidential races have grown significantly in recent years (Ansolabehere & Iyengar, 1995: Jamieson. 1992: Kern. 1989: prau & Kenski, 1990). now comprising nearly half of all political ads (Johnson-Cartee & Copeland, 1997). Political atlack messages. which seek to "create negative images" of an oppo!'ing candidate's position on issues. past record. or character (Gronbeck. 1992) are considered to be a very effective strategy to influence "Ieaners" and undecided volers (Ansolabehere & Iyengar. 1997: Johnson-Cartee & Copeland. 1997: Kaid & Boydston. 1987: Kern. 1989: Kern & Just, 1995: Mann & Ornstein. 1983: Newhagen & Reeves. 1991: Pfau & Kenski. 1990; Pinkleton. 1998: Sabato. 1981. 1983: Tarrance, 1980: Tinkham & Weaver-Lariscy. 1993). The question for political practitioners is. given the near certainty of facing an opponent's attacks. what can be done to deflect their influence? The standard options. refutation or response ads or news media adwalchcs. are after-the-fact remedies and are therefore onen unable to undo the damage inflicted by the attack. By contrast. inoculation is a preemptive strategy. Inoculation seeks to make potential voters resistant to attacks before the attacks occur. The potential of inoculation to deflect the influence of political attacks was investigated in Iwo large tJeld studies. The first centered on Ihe 1986 campaign for the U.S.
274
Pari III • S('('~ing alld Resil1il1g Compliallce: Slmlc'Kie,\ all(/ TlInin Senate from South Dakota involving incumbent Republi(.·an Jame ... Abdnor and Democratic challenger Tom Daschle (Prau & Burgoon. 19881. The other centered on the 1988 presidential campaign involving Republican George Bu~h and Democral Michael Dukaki ... (prau et al .. 1990). The 1986 ... tudy included 733 potential votcr ... and featured an inten ... e campaign. in which the two well known and highl) regarded candidatc!-. made cxten!-.ivc u\e of polillc~tI advertising .... pending $6.6 million. or $22 pcr vote cast in the election (Brokaw. It)R6). The results of the 1986 study indicated that both inoculation "same" and "different" mc ... ... ages conferred resistance to the influence of !-.ubsequcnl political attach (Pfau & Burgoon. 1988). regardle!-.:-. of political party. i...... uc. and character conlent. Overall. inoculation effech were most pronounced among :-.trong party identifiers. although an i nler'lclion revealed that inoculation "same" treatments were better with strong identifil.:rs, whereas "different"' treatments were beller among weak identifiers (Pl'au & Burgoon. 1988). The ... ub!-.equent 1988 ill\-e~tigation featured 31-J. pro... pective YOler'" and \\,11 ... conducted in September and October or the Bu\h-DukaJ..i\ pre\idential campaign. The 1988 study featured a number of new twi ... t~. It employed direct mail to admini ... tcr inoculation treatments. It featured boo!-.ter '\ession!-.. and it allowed for a direct compari:-.on of inoculation versus POq hoc refutarion. Inoculation j, a proccs ... that inrrin\ically imohe ... acknowledging vulnerabilitic .... ba ... ed on the r<1tionale that this is the be ... t way to protcct again\t counterinfluence. If there i", a dO\lvn ... ide to this approach. it can only be expo ... ed in direct comparison to PO!-.I hoc approachc .... Yct lhi.., ha ... "rcceivcd scant attelltion in the extant literaturc" (Pfau et al.. 1990. p. 19). The only previous asscs ... mcnt of pre- and po ... trefutation treatments was conducted by Tanncnbaum and colleague ... and revealcd a .., Ii ght advantage for prerefutation (Tanncnbaum & Norri .... 1965: Tannenbaum et 1.11 .. 1966). Results of the 1988 investigation offered further evidence for the viability of inocu lation in a political campaign context. Results confirmed previous finding .... indicating that both ·'same" and ··different" inoculation treatmcnts increa:-.cd re~istancc 10 suhscqucllI political attach (PfiJU et al.. 1990). Result ... did not indicate efficac) for boo ... ter ... e ...... ion .... perhaps due to the premature administration of the reinforcing material .... Finally. the results indicated that a jJl"l)-wtack approach. operationali7ed as inoculation and inoculation-plus-reinforcement. wa~ ",uperior to pm.t hoc refutation in ... afegu<-lrding voters' original attitude ... about candidates. The effect was most pronounced with 'trong party identifiers and with nonidclltifier .... With weak party identifiers. thc ... upcriority of the pre-attack over the post hoc approaches wa~ confined [0 the character attack condition errall et al .. 1990). The combined finding!-. of the 19H6 ilnd 19X8 studie:-. suggest that inoculation ofler:"<.I viable approach for candidates 10 deflcl't the persuasiveness of political attack me:-.\age," (prau & Kenski. 1990, p. 160). Both >ludies u,ed a sing le inoculation. injecLed during the final weeks of intense campaigns. l'onditions that ... llOUld have weakened rcsi ... tance effects. Inoculation would excrt more impact in less intense campaign .... or. in more intense campaign!-.. if initiated early. prior to opponents' attacks. In today· ... <.Illack oriellted political campaigns. inoculation offer:-. jj viable remedy . A:o. Republican consultant Jim
Chapter 15 • II/oclllation alld Resistance to Persllasion
275
Innocenzi (cited in Ehrenhall. 1985, p. 2563) advises, "Inoculation and preemption are what win campaigns." The next application of inoculation in a political context was as a potential preventive to Noelle-Neumann's "spiral of silence." Noelle-Neumann (1974, 1984) posited a "dark side" of public opinion and democracy. She theorized that many people who hold opinions that deviate from majority sentiment on controversial issues become hesitanteven unwilling-to express those opinions publicly. Their reticence stems from "the fear of sanctions resulting in social isolation" (1984. p. 65). The spiral of silence sets in motion a "snowball effect." in which those in the majority continue to speak out but those in the minority fall increasingly silent. Lin (2000) examined the potential of inoculation to break the spiral of silence in Taiwan. a tledgling democracy. He chose a highly controversial issue: Taiwan's relationship with China. The study tested the potential of inoculation treatments to boost the attitude confidence and willingness to speak out on behalf of people holding minority opinions on this issue. Results of the two-stage survey of 206 randomly sampled adults indicated that inoculation strengthened attitudes. Compared to the control group. those who were inoculated became increasingly confident in their attitudes over time. They showed more attitudinal confidence, greater willingness 10 speak out on behalf of attitudes. and increased likelihood of resisting the opposing positions of others (Lin, 2000). Lin concluded that inoculation can "break" the spiral of silence. potentially invigorating public deliberation of issues. a bedrock of viable democratic political systems.
Commercial Applications Inoculation has been tested in two commercial contexts: as a strategy to mitigate the influence of comparative advertising messages, and as a strategy to protect the public image of corporations in crisis situations. In both instances, preliminary results were promising. Five studies established a foundation for later inoculation research in commercial advertising. Two of them focused all social marketing. First. Bither. Dolich. and Ne ll (1971. p. 60) examined the use of inoculation to foster resistance against attacks espousing movie censorship. The study concluded that inoculation was able to reinforce attitudes. Szybillo and Heslin (1973) inoculated people's belief that airbags should be installed in automobiles. finding that refutaLional treatments were superior to supportive treatments. Two other studies compared the efficacy of refutational and supportive treatments in promoting resistance to Federal Trade Commission (FfC) attacks. Hunt (1973) concluded that refutational treatments were superior. but Gardner, Mitchell, and Staelin (1977) reported 110 differences. A fifth investigation examined the efficacy of refutational and supportive print ads on behalf of five products. Sawyer (1973) found that refutational ads were superior. with limitations based on attitude toward the product and/or the product class. These studies employed a refulationai approach. which is .a necessary. but not sufficient. condition for inoculation. What they did not do was feature threat. Research on inoculation in a commercial advertising context either failed to operationalize threat and/ or failed to measure it. Since threat is one of the two core elements in inoculation,
276
Part III • Seeking and Resisting Complian ce: Strategies and Tactics functioning as a motivational catalys t for rc~istance. th c~e manipulations fail ed to manipulate the inocu lation co nstruct properly. What these studi es have suggested is th e potential for the other core element, refutational pree mption , in foste ring resistance. A more rece nt stud y. which did confirm effectiveness of the threat manipulation. found th at inoc ulation works. but with cel1 a in caveats (Prau, 1992). The caveats arc re ceil'e r illl'oh'emelll in the product class, whi ch is viewed as an " important mediator of CO I1sumer behavior" ( Mitche ll , 1978, p. 195 ). and compararil'e message formlll. whi ch involves the sty le and direc tionality of th e co mpariso n (La mb. Pride, & Pletcher, 1978: Pride. Lamb. & Pletcher. 1977. 1979) and has been found to moderate th e persuas ive ness of comparat ive messages (Lam b. Pletcher. & Pride. 1979). Pfall ( 1992) reported that inoculati on treatments had a direc t effect o n rece ive r brand attitude. but that thc effec t of inoculation in th e product class depended 011 receiver's level of invol vement. Follow-up tests indicated that inoculation is effective in conferring resiMance to co mpariso n ads:. btu o nly for hi ghl y invol ving products. Thi s finding is consistent wit h recent research on In oc ulation Theory, which suggests that i s~ u e involvemenl ma y dictate th e boundary co nditions for Inoc ulation Theo ry (Pfau et al.. 1997a). Scholars or practitioners interes ted in applications of inoculation in advertising should recogn ilc that it may prove effective only with hi ghly involving products (Kamins & Assea!. 1987: prau. 1992). Thi s would limit its potential , since so much adverti sing is on behalf of goods and services that aren't hi ghl y involving. In th e contex t of public relation s, resea rch has foc used on whether inocui. upe rior 10 image promotion. This prediction i ~ consisten t wi th th e ea rly resistance research, indicating that refutational defe nses are superior to su pport ive defenses (A nde rso n & McG uire. 1965 : McGuire. 1961a. 1962. 1966: McGuire & Papageo rg is. 1961. 1962 : Papageorg i' & McG uire. 1961: Tanne nbaum e t al.. 1966: Tannenbaum & Norris. 1965). To t e~ t (hi s notion. Wan (2000) studi ed public attitudes toward a real petroleum company over a one-month period. Partic ipant s received an inoculation "sa me" treatme nt. an inoc ul ati on "d ifferent" message. an im<.lge-enhancing "bolstering" message. a COI11 bined "'refulational-bolstering" message. or no message (co ntrol ). Later. SO me partic ipants
Chapter 15 • Il1oCl//orion and Resistance 10 Persuasion
277
assigned to crisis conditions received a counterattitudinal crisis message. The results indicated that for participants possessing positive initial attitudes toward the organization. all treatment approaches-inoculation "same" and "different." image bolstering and combination-were effective in protecting the image of the organization following exposure to the crisis scenario. but that no one approach worked best (Wan. 2000). With subjects not exposed to the crisis scenario. the bolstering approach was slightly superior to inoculation, suggesting some downside to inoculation. but a lack of power rendered this finding tentative.
Health Campaign Applications Because inoculation is a useful approach in situations in which altitudes are vulnerable if challenged. it is receiving increased attention in the health campaign context. particularly targeting adolescent behaviors. In many areas. children's attitudes are initially formed during preadolescence but are then subject to intense pressure during adolescence. Millman and Botvin (1983) have observed this developmental pattern in a number of adoiescelll controversies. such as tobacco. alcohol, drugs. and sexual intercourse. Jessor and Jessor (1975) refer to these as "transition-marking behaviors." For these behaviors, inoculation would seem to be an ideal strategy because it is designed to make attitudes more resistant to change (Miller & Burgoon. 1973). In this section we will examine research 011 inoculation's ability to reduce the onset of adolescent smoking and drinking.
Smokillg Prevelltion. As a result of the efforts of parents and teachers. most children develop attitudes opposing smoking early in their lives. Typically. these attitudes persist until the transition to the middle grades, when they erode due to physiological changes (Hamburg, 1979) in conjunction with intense peer pressure (Bewley & Bland, 1977: Evans & Raines, 1982; Flay, d'Avernas, Best, Kersell & Ryan, 1983: Foon, 1986; Friedman, Lichtenstein & Biglan, 1985: Goldberg & Gam, 1982: Gottlieb & Baker, 1986: Harken. 1987: Hurd, Johnson, Pechacek, Best, Jacobs, & Luepker, 1980; McAlister, Perry, & Maecoby, 1979; O'Rourke, O'Byrne, & Wilson-Davis, 1983; Pechacek & McAlister, 1980: Pederson & Lercoe, 1982; Rosenberg, 1965; Salomon, Stein, Eisenberg, & Klein, 1984). Just before and during this period, many adolescents' antismoking attitudes soften (Elder & Stern, 1986; Evans & Raines, 1982: Hamburg, 1979: Johnson, 1982; Killen, 1985: Pfau & Van Bockern, 1994; Rokeach, 1987). Consequently, this is a highrisk period for the onset of smoking. McAlister and colleagues emphasize that, "More than half of all current young people who adopt the habit of daily smoking do so before or during their ninth grade school year" (1979, p. 651). Experts agree that a strategy of resistance is needed on or before this critical transition period; a strategy designed to protect against attitude slippage (Allegrante, O'Rourke, & Tuncalp, 1977: Bernstein & McAlister, 1976; Evans, Rozelle, & Mittlemark, 1978: Harken, 1987; McCaul, Glasgow, O'Neill, Freeborn, & Rump, 1982). The Centers for Disease Control (1998) recommended schoolbased prevention programs that exert maximum intensity during grades 6 through 8. The most popular antismoking approach has been social inoculation, which combines one facet of Inoculation Theory, refutational preemption, with Bandura's Social Learning Theories (Wallack & Corbett, 1987). Despite its name. however. social
278
Part 111 • Seekill}.: (lnd Resi.\lillg COl1lpli(ll1a: 51rl1regil s (111(/ Tactics J
inoculat ion is not inoculation. II doesn 't operationalize the threat component, whi ch provides the internal moti vatio n for people lO resist subseq uent innuence (McGuire. 1962). Instead. it features a potpourri of tactics, including teac her- and/or peer-led di sc ussion sessions. slide and vi deo presentations, peer mode ling. sc hoolwide smoking prevention campaigns, and ot hers. Thi~ smo rgasbord of methods makes replication diffi c ult. which undermin es ib utilit y. Also. it makes it impossible to i ~o l a te what elemcllIs are responsible for outcomes. Foo n questioned. " If th c~e programs are working, what special features arC
working and how?"" ( 1986. p, 1025 ). Flay ob,erved that. ", , , we really know ve ry little at thi s time about which of these program co mpone nts are necessary for program effecti vene" , , ," ( 1985, p, 378), The potential of inoculmion. including both threa t and refutational pree mplion e lements. has been the focus of rece nt studi es. prau and colleag ues ( 1992) co nduc ted a longitudinal field study of stude nts making the transition from ele men tary school to junior high sc hool in Sioux Fall~, South DakOla, startin g in fall 1990. Inocu lation was accomplished via professionally prepared 12- to 25-m inute videos. Students were studied for two years. Results of the first year indicated that inoculation vi deos instill ed resistance to smoki ng onset. but only among adolesce nts with low self-estee m (Pfall ct al.. 1992). This interaction patlern of inoculation and receiver 'ielf-esteem dissipated by the end of the second
year of the study, At thi s point. 84 weeks after the inoculation treatments. participants in the treatment group he ld less favorab le attitudes toward smokin g and smokers (Pfall & Van Bockern, 1994), The next study of inoc ul ation and smoking pre ven tion by Szabo (2000) exami ned its potential with fifth- and sixth-grade nonsmokers in both a metropolitan area ( 163 predominantl y African American stude nt s) and a rural setting ( 157 overw helm in gly whi te s tudents in Iowa). Szabo'!., study incorporated additional e leme nts. warra nted by recent research documen ting a large and unexplained upsurge in adolesce nt smokin g and by recent research on Inoculation Theory. First, Szabo posited that the surge in adolesce nt smoki ng during the I 990s, even in the face of perva<.;ive antismoking campai gns, mi ght in part reflec t a " boomerang effect"' in response to the antismoking ca mpai gns. Using assumptions of Brehm 's Psyc ho logical Reactance Theory. Szabo designed nonnative inoculation appea ls. whi ch stressed peer disapproval of smokin g, plus morc tradit ional hea lth-based messages. She
posited that both normati ve and health-based treatments. which included threat and refutational preemption e lements. would inst ill resistance ill most adolescents but that the normati ve appeals wou ld be l es~ likely to tri gger psychological reactance ill some ado lescents. Second. to shed further li ght on the ro le of affect in resistance. Szabo designed two normat ive messages, one to e lici t anger and one to tri gger happiness. In addition. she mcasured affect e licited by inocu lation treatments. Third, Szabo inc luded c riti cal indi vi dual differe ncc va riables in the design. such as self-efficac y and self-esteem. to determine their roles in the process of resistance. Szabo's res ults (2000) \uggc~ t ed that a fine linc se parates antismoking messages that foster resistance from those that tri gger reactance. Co ntrary to the predic ti on th at all inocu lati on treatment~ would en hance resistance to smokin g. the res ults indicated that inoculation effects were contin ge nt on the subpopulation of interes t. In oc ulation elicited resistan ce. but onl y on behalf of two of the four gro ups targeted. The normative-a nger and
Chapter 15 • Inom/miml
(/1/(/
Rt'sisf(lIIce 10 P(Jr,wasioll
279
traditional health-based appeals achieved attitudc resistancc in rural \ixth grader\. while the normative-cognitive and he~llth-ba~ed messages conferred attitude resistance in urban fifth grade". Neither rural fifth graders nor urban sixth graders showed any signs of resistance: ruther. they exhibitcd psychological rcactance to all the inoculation attempts. In addition. there was some evidence of reactancc in all subpopulations. For example. the health-ba\cd appeal produced reactance in rural fifth graders. the cognitivc messagc elicited reactance in rural sixth graders. and the normative-anger treatmcnt triggered reactance in all urban students. This study suggested that antismoking mcssage, produce both resistance and reactance in children, depending on the message type and the subpopulution of interest. Finally. Szabo's (2000) results revealed that receiver self-efficacy was strongly associated with resistance to smoking for all ~lUdents in all conditions but that self-esteem w<.\s not.
Drinking Prevention.
For the "iame reasons thai inoculation i"i a promising approach for smoking prevcntion. it offers potential for reducing the onset of adolescent alcohol use. School-based programs designcd to fo~ter resistance to alcohol usc have become increasingly prevalent (Elder et aI., 1987). The,e programs feature life ski lis training and arc patterned after the social inoculation approach dcst:ribed previously. Researchers maintain that prevention is a more effective approach than trying to persuade adolescents to stop drinking once they·ve begun (Hansen. Graham. Wolkenstein. & Rohrbach. 1991: Krelltter. Gewirtz. Davenny. & Love. 1991: Webb. Baer. & McKelvey. 1995). To date. however. re~iSlancc efforts have yielded mixed rcsult~ (Foxcroft. Lister-Sharp. & Lowe. 1997: Slater, Beauvais. Rouner. Van Lellven. Murphy. & Domenech-Rodrigue7, 1996: Tobler. 1986). The rcason could he "the limited application of a theoretical basis for predicting when pecr-based resistance programs are cffective" (Godbold & Pfall, 2000. p. 413). As with antis-moking social inoculation. it is il1lpo:-.~ible to pinpoint the specific clements Ihat are responsible for outcomes with "Illulticomponent" alcohol resistance programs (Krelltter et al.. 1991). Godbold and Pfau (2000) admini'tered an inoclilation study involving 417 nondrinking sixth graders. Inoculation was accomplished via threc-minute informative or normative videos. de!o.igncd to resemble public "en icc announcements. Participants were exposed to attack messages either immediately after inoculation or two weeks later. AItacks consiMed of actual beer commercials embedded in a series of television advel1i\ements. made to look like a television break. Godbold and Pfau (2000) predicted that both informative and normative inoculation upproaches would be effecti\'c in conferring resh.tance to drinking initiation but that normative message:-. would be ~lIperior because research indicates that they arc better ~uited to "judgmental ta~h" (Kaplan, 1989). The results of the study indicated that normative messages produced the lowest e~timates of peer acceptance of drinking. However, normative inoculation messages were no better than informative mcssages in instilling resistance to the beer commercials. and in fact both message strategies exerted limited innucnce on attitudes and behavioral intentions. The results also revealed that immediate attacks resulted in more resistant attitude and behavioral intentions than did delayed attacks. Overall. Godbold and Pfall (2000) reponed relatively weak effect' involving either the normative or informative inoculation strategy. Instead. the pattern of result:-. revealed
280
Part III • Sed.illg alld R£'.\isrillg Complialln': Stmu'gie.\ lIlId TaClic.\·
that the key factor in resistance is adolescent perception of peer approval of drinking. "hich ,ubsequcntly leads to threat and. finally. to resi"ance to alcohol use. The result, ~uggt.!~ted that practitioners should employ normative messages in order to lower the e~ti mates of peer acceptance and peer pres~ure to drink in order to enhance resistance to adole;ccnt alcohol use. Godbold and prau placed the blame for the weak inoculation erfect> on the experimental design. Results revealed that the threat manipulation was not adequate.
Conclusioll Inoculation appears to offer an effective means of bolstering receivers' resistance to opposing messages. Inoculation is particularly u\eful bet:ause it increases resistancc not only 10 the "'pecifie arguments included in the inoculation treatment but also to novel argument":. on the same topic or issue. Two key components of Inoculation Theory are threat and refutational preemption. Threat i\ the motivational trigger that promph cOllJltcrarguing. When receivers feel threatened. they are more likely 10 bolster their attitudes against impending attach. Refutational preemption involves raj",ing objections and then refuting them within an inoculating message. This approach tends to increase receivers' resistant:e not only to the specific objections rai,cd but also to other. novel objections. Although a number of caveats and qualifications regarding the use of inoculation remain. the theory has been show n 10 work in a variety of persua~ive settings.
Referellces_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __
c..
Adall"". w. & Beatt y. M . 1. (1977). Dogm
7.237-248. Anam!. K. W. E .• & Mandcl. J . E. (1972). Strategic ... of reSI':Itance to pcrsua ... ion : New ... uhjccl mallcr for the tcacher of spect'h t'onununicalion . Cel/lral Stalr.f Spet'ch Jour/wi, 23. 11 - 17. Andcr\on. L. R .. & McGuire. \\. J (1965). Prior rc: ....... ur.... ncc or group con\cnsu, 11' a raclor in producing rc .... 'tancc to per,ua . . ion. SocIOmetry. 28. 44-56. A nsolahcherc. S .. & Iyengar. S. ( 1995). Goill g lU'gMi\·e: fI(m polilical lIl/l"erlisellll'III.\ shri"k and po/ari:r Ilu· eit'UorCUt'. Ne\~ York: The Free Pre ...... Bandur:l. A. ( 1983). Selr-efficacy determinanh of antil.:ipaletl fear... and c;'llamilie .... JOIln/al of PenOtllllit)' and Sodal Psych% !!.\" . .J5. 464-469 Bern~lclll. D. A .. & McA lj 'ler. A. L. ( 1976). The mo<.iilicalion of ... moking behavior. AddictiH' Be/tlll'iors, I. 89-t()~. Bew lcy. B. R .. & Bland. J. M. ( 1977). Acatlcmic performance anti social raclor~ related to c igarette r..moking by \choolchildrcn. 8rili.lh JOIlrt/al of Pre\·eTllllti\'e alld Social Medicine. J I. 18- 24 . Bither. S. W .. Dolich. I. J .. & Nell. E. B. ( 1971). The applicat10n or attitude immunization technique ... in m(1rkcting. JOll/"llal of Markeling R,'search, IN. 56-61 . BOivin. G. J. (1982). Broadening the focus or '>moking prcvcntion strategies. In T. J. Coates. A. C. Peter~n. & c. PelT) (Etk). PromOll1lX ado/elcell! /tea/I": A dialogue 011 research "TId practiu (pp. 137- 147). New Yor": Academic Press. Brehm. J. W ., & Cohen. A. R. ( 1962). Expl()r(llion~ il/ cogllili\·,' dinollllllce. New York: Wiley.
Chapler 15 • Inocu/a/ioll allli ReJiSlat/ce 10 Persllasioll
281
Brokaw. C. (1986. December 10). Abdnor·Daschle race seh new .. laIc spending record . Sioux Falh ArRlIs
Leader. p. 2. Burgoon. M .. Burgoon, J. K., Rie~!-.. M .. Butler. J. , Momgomery. C. L.. Stinnett. W. D .. Miller. M .. Long. M .. Vaughn. D .. & Caine. B. (1976). Propensity of persuasive attack and inten<.,ity of pretreatment mev;agcs as predictors of resistance to persuas ion. JOllmal oj P,~.w ·hol()gy. 92. 123- 129. Burgoon. M .. & Chase. L. J. (1973). The effects of differential linguistic pattern!> in me ......ages attempting to induce resistance to pen.ua!age comprehensibility. JOlIl'lla/ oj PersOIllllitv and Social PsycJlOlngy. 34. 605-614. Chaiken. S .. & Eagl). A. H. ( 1983). Commun ication modality as a determinant of per<.,u:l<:.iun: The role of communicator sal ience. Journal oj PerJollali/.v alld Social Psrdwlogy. 45, 24 1-256. Chesebro. J. W . ( 1984). The media reality : Epistemolog ical functioll'l of media in cultural <.,YMenh. Criti· cal Stlldies ill MaH COIllI1H1l1iccuio/l. 111-130. Coombs. W. T. (1998). An analytic framework for crisis si lllations : Better responsc'I frolll a better under~I:lnding of the situation. JO/lI"I/{II oj PI/blic Relati(JIIs Research. 10. 177- 191. Cmne. E. (1962). Immuni 7ation-w ith and without usc of cou lltemrgumc nl s. J(Jllmali.\"I1I Quarterly. 39.
445-450. Cronen. V, E .. & LaFleur. G. (1977). Inoculation aguinst persuasive attacks: A test of alternati .... e ex plana. lion'l . Journal of Social P.n-c/wlo}p'. 102.255-265. Eag ly. A. 1-1 .. & Chaiken. S. ( 199]). The pSyc/lOlogr of mlll/uies. Orlando. FL Harcourt Brace Jo . . anovich. Ehrenhalt, A. (1985). Technology .... trategy bring new campaign en\. COl/gressimwl Quarter/v Week'" Re· port. 43, 2559-2565. Elder. J . P .. & Stem, R. A. ( 1986). The ABCs of adolescent ... making prevention: An Environment and <:.ki ll s model. Health EdllC(/fioll Qut/rterly. 13. 18 I - 191. Elder. J. P .. Stern. R. A .. Ander .. on. M .. Hovel!. M. F .. Melgaard. C. A., & Seidman. R. L. (1987). Contingency·bu...ed strategies for prevent in g alcohol. drug. and tobacco use: Mi~<;ing or unwanted components of adolescent health prommion. £dtlCllIioll alld TreCltmelll oj Cltildrt'll. 10. 33-H. Evans. R. I.. & Raines. B. E. (1982). Control and prevention of smoking in adole<.;cenl<.;: A p<;ychosocial perspective. In T . J. Coate<." A. C. Petersen. & c. Perry (Eds.), PromnrillR adoh'.H·1!1II healIlt: A dialoglle 011 research alld practice (pp. 101-136). New York: Academic Pre~s . Evans, R. I.. Rozelle. R. M .. & Miltlemiuk. M. B. (1978). Deterring the onset of ... moking in children. JOIII'II£// oj Applied Social Psychology. R. 126-135. Fink.. S. ( 1986). Crisis /lllIl/ageme/lI; Pla/lllillg fiu' the ille\·illlble. New York: American Management A ... sociation. Flay. B. R. (1985). Prosocial approaches to sllloking prevcntion: A re\ it!w of finding<.,. Health P.lyclwlugy. ~.
449-488 .
Flay. B. R.. d' A,ernalo.. R. 1.. Beq. J. A .. Kersel1. M. W .. & Ryan . K. B. (1983). Cigarette <:. Illoki ng: Why young people do it and way<" of preventi ng it. In P. J . McGrath & P. Fire~tone (Ed ... ), Pediatric and at/o/t'Jct!1II "e"adoral medicille (pp. 132- 183). Ncw York: Springer. Foon. A. E. ( 1986). Smoki ng prevcntion programs for adole<.;cents: The value of social p~ycho l ogical approache~. The IlIIernariollal jOIlf'll(l1 oj the Addicriolls. 21. 1017-1029. Foxcroft. D. R .. Lister·Sharp. 0 .. & Lowe. G. (1997). Alcohol mi~use prevention for young people: A <.,y~tcillalic review reveals methodological conccrns and lack of reliable evidencc of effect ivc ness. Addic/ion. 92.53 1-537.
Freedman. J . L.. & Sears. D. O. (1965). Warning. di . . traction. and rc .. istunce .w)Il(lli/y (lnd Social Psyc/wlogr. I. 262-266.
10
influence. 10111'/1(11
oJ Per-
282
Part III • Seeking and Resisting COlllpliaflu: Strategies and Tactics Friedman, L. S., Lichtenstein. E.. & Biglan, A. (1985). Smoking onset among teens: An empirical analysis of initial situations. AddictiFe Behavior.\". 10, 1-13. Gardner. M .. Mitchell. A., & Staelin, R. (1977). The effects of attacks and inoculations in a public policy context. In B. A. Greenberg & D. N. Bellenger (Eds.), Contemporary marketing thouX}If: 1977 edumIurs' proceedings (pp. 292-297). Chicago, IL: American Marketing Association. Goldberg, M. E., & Garn. G. J. (1982). Increasing the involvement of teenage cigaretle smokers in anti~ smoking campaigns. journal of CommulIication. 32( I). 75-86. Godbold, L. c., & Pfau, M. (2000). Conferring resistance to peer pressure among adolescents: Using inoculation theory to discourage alcohol use. Communication Rt'search. 27,411-437. Gottlieb. N., & Baker, J. (1986). The relative intluence of health beliefs, parental and peer behavior~ and exercise program participation on smoking, alcohol use and physical activity. Social Science alld Medicine, 22. 915-927. Gronbeck, B. E. (1992). Negative narratives in 1988 presidential campaign ads. Quarterly Journal of Speech. 78,333-346. Hamburg, D. A. (1979). Disease prevention: The challenge of the future. American Journal of Public Heal/h. 6Y. 1026-1 OJ3. Hansen, W. B., Graham. J. W., Wolkcnstein, B. 1-1., & Rohrbach. L. A. (199 1). Program integrity as a moderator of prevention program effectiveness: Results for the fifth-grade students in the adolescent alcohol prevention trial. JOl/mal a/Studies 01/ Alcohol. 52, 568-579. Harken. L. S. (1987). The prevention of adolescent smOking: A public health priority. Evaluatioll & the Health Professions. /0,373-393. Heath, R. L. (1997). Slrategic is.\"Ia!.I· mal/agement: Orgalli;:.ations and public policy challenges. Thousand Oaks. CA: Sage. Hunt. H. K. (1973). Effects of corrective advertising. JOltl"lla/ of Advertising Research. 13. 15-22 . Hurd , P. D., John son. C. A., Pechacek. T .. Bast. L. P.. Jacobs, D. R.. & Luepker, R. v. ( 1980). Prevention of cigarette smoking in seventh grade <;rudcnts. journal of Behavioral Medicine. 3. 15-28. Insko. C. A. (1967). n,coric.\· of attitude change. Englewood Cliffs, NJ: Prentice-Hall. ILard. C. E. ( 1993). Four systems of emotion activation: Cognitive and noncognitive processes. Psycho108;("(11 R('I'ie\\', 100.68-90. Jamieson, K . H. (1992). Dirty politics: Deceprion, distraction. alld democracy. New York: Oxford University Press. Jessor, R. , & Jessor. S. L. (1975). Adolescent development and the onset of drinking: A longitudinal study. Journal of Sflldics 011 Alcohol. 36, 27-51. Johnson, C. A. (1982). Untested and erroneous assumptions underlying antismoking programs. In T. J. Coates, A. C. Petersen, & c. Perry (Eds.). Promoting adolescellf health: A dialogue 011 research aud practice (pp. 149-165). New York: Academic Press. Johnson-Cartee. K. S" & Copeland. G. A. (1997). Manipulating the American voter: Po/itiml campaign ,o/llmercials. New York: Pracger. Kaid, L. L., & Boydston, J. (1987). An experimental sllldy of the effectiveness of negative political advertisements. COIIIIIHlllicmion Quarterly, 35. 193-20 I. Kamins. M. A., & Asseal, II. (1987). Two-sided versus one-sided appeals: A cognitive perspective on argumentation. source derogation. and the effect of disconfirming trial on belief change. JOllrnal of Marketillg Research. 24,29-39. Kern, M. (1989). Political adl1ertising in the eighties. New York: Praeger. Kern, Moo & Just, M. (1995). The focus group method, political advertising, campaign news. and the construction of candidate images. Political Communication, 12, 127- 145. Killen, J. D. (1985). Prevention of adolescent tobacco smoking: The social pressure resistance training approach. Journal of Child Psychology alld Psychiatry, 26.7-15. Kreutter, K. 1.. Gewirtz, H., Davenny, J. E.. & Love. C. (1991). Drug and alcohol prevention project for sixth graders: First-year findings. Adolescence, 26, 2R7- 293. Lamb, C. W .. Pletcher. B. A .. & Pride, W. M. (1979). Print readers' perceptions of various advertising formats. jmlr1wli.\"/1I Quarterly. 56, 328-335. Lamb, C. W .. Pride, W. M .. & Pletcher. B. A. (1978). A taxonomy for comparative advertising research. JOllrnal of A{h·errisillg, 7,43-47.
Chapter 15 • IIIOm/alion (lml Resis{(lllce to Persuasion
283
Lee. W .. & pfau. M. (1998. July). The effl'<"/il'(,lIess o/cof,:lIilil'e lind IIUeClil'c inoculaliollappeals ill COf/Jerr;f/X resis/(mce aga;nst cognit;I'e allllllfft'Clil'e [Illacks. Paper prescnlcd at the annual meeting of the Intemational Communication A<;,sociation. Jcrusalem. I<;rae l. Lin. W .-K . (2000). Use of il/ocll/lIIin" 10 cOllllm{ Ihe \piral nJsi/e"ee: A study of public op;nioll ill democrac.\'. Unpublished doclOral dissertation. UI1I\cr.. ity of Wi<.;{.'Onsin-Madison. Lovc, R. E.. & Grccnwald. A. C. (1978). Cognitivc re"ponse, 10 pcr<;uasion as mediators of opinion change. journal of Social P.5ydlOlogy. J(N. 231-241. Lumsdaine. A. A .. & Janis. I. L. (1953). Rc~i~tance to "l:Ounterpropaganda" produccd by one-sided and two-.. ided "propaganda" prcscntations. Puhlic Opinioll Quarrerly. J7. 311-318. Manis. M .. & Blake. J. B. (1963). Interpretation of persuasive Ille~\age .. a., a function of prior immunization. journal oj'AIJ11orllllll alld Social P.\'ycl/O/o~y. 66. 225-230. Mann. T. E" & Orn~tein. N. J. (1983). S~nuing a mes ..agc: Voter., and Congress in 1982. In T. E. Mann & N. J. Orn:-.tein (Eds.), Tilt' AlII('I'icfllI eil'cliol/ (!f 19N2 (pp. 133-152). Washington. DC: American Enterprise In sti tute. McAli'iter. A. L.. Pcrry. C" & Maccoby. N. (1979). Adolescent smoking: On<.;et and prevention. Pediatrics. 63. 65Q-.658 . MCC.lUl. K. D., Glasgow. R .. O·Neill. H. K.. Fre~born. V .. & Rump. B. S. ( 1982). Predicting adolesccnt :-.moking. journal oJScllool Health. 52(6). 3-'2-346. McCroskey. J . C (1970). The effects of evidence a<.; ,10 inhibitor of countcr~pep.;ua.,ion. Speech MOIIO gmphf\. 37. 188-19-'. McCro.. key. J. C. Young. T . 1.. & Scott. M. D. (1972). The effects or message sidedncss- anu evidence on inoculation against countcrpcr'>ua'>ion in .,mall group communication. Speech MOllographs. 34.
205-212. McGuire. W. J. (196Ia). Thc errecti\ene:-.<.; of .,upportive and rcfutational dcfenses in illlllluni/ing and restoring be lief., against pen,u",:-.jon. Sndomt'fry. N. 184-197. McGuire. W. J . (196Ib). Res-i...rance to pcr:-.ualo.ion conferrcd by active and passive prior rerutation of the same and altcrnative counterargument,. jOlll'l1l1/ (!f AIJllormllllllld Social P.~ycholo8Y. 63. 326-
332. McGuire. W. J. (1962). Persi:-.lence of the resistance to per.,ua<;ion inuuced by various types of prior belief defcnse:-.. journal oj AhI/orilla/ WId Social P.\)'('hology. 64. 24 [-2-'8. McGuire. W. J. (1964). Indu cing re!-.i<;tance to pcrsua'iion. Some contemporary approaches. In L. Berkowitz (Ed.). Ad\'(/f/('e~;11 expaimellwl social psy('hology (vol. J. pp. 191 - 229). New York: Academic Press. McGuirc. W. J. ( 1966). Per<;istcnce of the re\i:-.tance to persua<.;ion induced by various type,> of prior belief dcfen<;es. In C. W. Backman & P. F. Secord (Eds.). Pro!JIelllJ in .wcfal psychology (pp. 128-135). New York: McGraw-Hili . McGuire. W . J. ( 1970. February). A vaccine for brainwa ... h. P_~ydwlog.r Today. 3. 36-39. 63-64. McGuire. W. J .. & Papageorgi:-.. D. ( 1961). The relative efficacy of various types of prior belief-deren .. c in producing immunity against persuasion. Journal oj Abllormal a"d Social Psychology. 62. 327-337. McGuire, W. 1.. & Papageorgi!.. D. (1962). Effecli\'enes., of forewarning in dcveloping resistance to persuasion. Public 01';";011 Quarterly. 26. 2+-34. Mcyrowit7. J. (1985). No lell~e of place. Ncw York : Oxford Univcrsit) Pre~!. . Millcr. G. R .. & Burgoon. M . (1973). New recllniques o!per.wl{/sioll. New York: Harper & Row . Miller, M. D .• & Burgoon. M . (1979). The relation'ihip bel ween violations of expectations and the induction of resistance 10 pcrsuasion . HI/mall COllll1lllllic(lIioll Research. 5. 301-313. Miller. N .. & Baron. R. S. (1973). On mea .. urillg countcrarguing. JOIlrlwl Jor rhe Theory oj Soci(lIBel/a\';01'.
I. 101 - 118.
Millman, R. B .. & BOIvin, G. J. (1983). Sub<.tance u<.;c. misuse. and dependence. In M. D. Levine. W. B. Carey, A. C. Crocker. & R. T . Gro~<; (Eds.). D('I'e/opmellral-beh(ll'joml petiilllrics. Philadelphia: W. B. Saunders. Mitchell. A. A. (1978). In \'olvement: A poten tially important mediator of consumer behavior. In W. L Wilkie (Ed.). Adl'allces ill {,O/ISlIlIIt'r re.I'e{lrcli: Proceedillg~ Jor rhe Anocia/ioll for C01l\'umer Re.\'(!(Irch nimh [lilli/wI cOflferem'e M;ami HemA P!orida. October 1978 (vol. 6. pp. 191 - 196). Urbana. I L: Association ror Consumer Re<.;carch .
284
Pan III • Seekillg alld Re,~ist;lIg COlllpli{l//{'e: SI({fIl'1-:il'\ allli TU('/in Nc\\hagcn. J E .. & Rcc\c .... B. (1991). Emotion and mcmo!) rc ... pcctrum. Pcder,on. L. L.. & Lcfcot!. N. M. (1982). MultivariJtc anal)"i ... of variables related to cigarette <.,llloJ...ing among children ill grade ... four to "ix. C(/fUUlillll jOllrlllll of Pllhlie HealTh. 73. 172-175. Petty. R, E. (1977). The importance 01 cogniti\c n:",poIN:'" in per<.,ua\ion. Adl"{lll('('s ill Coml/11/('r Re.\('(/nA .J. 357··362 Petl). R. E.. & Cacioppo. J. T. (1979a). Effect, 01 I'orewilrning of pcr\uasi\ c intent and im oh emel1l on cognili\c re~pon\c ..... nd pef\ua .... on, Persollalm' (llId Sorilll PI\'dlOlogy Bllllelill. 5. 173-176. PClly. R. E.. & Cacioppo. J. T. ( 1979b). 1''''lIe in\ oh CIllCI\I can increa,e or dccrea ...e per'iua .. ion by cnhancing mc ..... age-relc\ al1l cogniti\c rc\pon ... e~. jOllfl/(// of Pef'lonalil\· alld Social P"yclwlo,f.:v. 37. 1915 - 1926. Petl). R. E.. & Cacioppo. J. T. (19861. Cmllllllll1ic{lli(}II {l1U1 per,wlIIi(}II: Central ami peripheral rOIlIl',1 to (lttlf/lde ehallxe. NI.'\-\< York: Springer-Verlag. Pfall. M . (1l)l)O). A channel approach to telcvi ... ion innllcncc. jO/ll"lllll oj lJrolldc{/I·fillg & Electronic Ml'diu. 3.1. 195-2 1....
Prau. M. (1992). The potential of inoculatiun in promoting. re ... i... tance 10 the effccti vcne ... s of comparative ad\crll ... ing me<., ... agc .... COIl/mulI/car;on Quarfl'rh. .JO. 26-4-1. Pfau. M . (1997), The inocu lation model of rcsi ... tance 10 innuencl.!. In F. J. Bo .. tt.!r & G. Barnet (Ed\.), /J rown .,· of Cf!lI/mlllliCllfiOI1~nell("e, (vol. I J. pp. 133- 171). Norwood. NJ: Ablex PlIbll~hing Curporation. Pfall. M . & Burgoon. ~1 . (1988). Inoculation in political campaign commun ication. III1I1lUII C0I1/I1/Imicalioll Rcw(m-h. 15.91-111 Pfall. M.. lIolbert. R. L.. Zubrie. S. J. Pa,ha. N. II ., & Lin. W.-K . (2(X)O). Role and influencc of communicalion modality in the proce ... <., of re ... i~tancc to pcr,utl,ion. Ml'dill P.\yclwlogy. 2. 1-33. Pfall. M .. & Kem,h H. C. (1990). AIIlll'{ poliricJ: SIf(lIt'gr and deJl'IIH'. New VorL Pracger. Pfall, M. Ken ... J...i. II. C. NilZ. M.. & Soren<.,on. J. (]990). Efficac) of inoculation :o.tratcgie .. 111 promotlllg re"'l ... tance to political altacJ... mc ...... ;tge ... : Application to direct mail. Coml1l1l1/lmtiol/ MmlONrtlphL 57. 1 12. Pfau. M.. S/abo. E. A.. Ander<.,on. J.. Morrill. J.. Zuhric. J.. & Wan. 11 .-11. (2000). The roll' lIlIlI impllcl (1· affect ill the pl"O("('J,I· of re,li.HlIIICl', Paper prc~cnted at thc annual mecting of thl.! International Com· Illunication A......ociation. Acapulco. Mexico. Pfall. t\L TlI~lng. K. J.. Koerner. A. F-" •• Lce. W.• Godbold. L. c.. Penalol.a. L. J.. Yang. V. 5 .. & lI ong. Y. (1997a) . Enriching the inoculation con~truel. Thl.! ro lc of CritICa l componcnh in thc procc<., ... of re\I ... tancc. Hllmall Communicatioll Re.H'arch. 2.J. IH7- 2t5. Prall , M., Tu:-.ing. K. J. , Lee. W.. Godbold. L. C. KOt.!l"Iler. A .. Pcnalo/a. ilong. Y.. & Yang. V. S. (1997b). Nuance)" in inoculation: The role of inoculation approach. ego .. involvcment. and rne ... sagc procc ... , 1Ilg. di<.,pmilion in rc<.,i<;tance. COllllllllllicalioll Qllarlal\·. .15. -161-48 L Pfau. M. & Van BocJ...crn. S. ( 199-1). Thc pcr... i... ten(:c (If inoculiltion in conferring rcsi ... tancc to ... 1110J...1I1g initiation among adolescents: The ...ccond year. /lilli/till COII1IlIlllliClltinll Rl'Jl'lIrch. 20. -I13- BO. Pfau. M.. V;m Bockern. S .. & Kang. J. G. (1992). U~t.: of inoculation to promote rC<"i'tance to \lllok1llg initIation among ado lc ~cenh. Commlil/icafion MOllogrtll'h~. 59. 213-230.
Chapter 15 • IlIoculatiulI mill Re'iistllllC'e
to Pl'n llfHirm
285
Pfau, M " & Wan. H.- H. (in pres~). Pcr.ua .. ion : An intrinsic function of public relation ... In C. H. BOlan & V. JlOlIclion. Jr. (Eds.). P"Mic rellitiolH thl'ory 1I . lIill~dalc, NJ: Erlbaum. Pinklcton. B. E. (1998). Effects of print comparativc advenising. on politu.:al de::ci'iion-Illal..ing and panicipation. JOIln/al oj Commllllimtioll, .JR, 24-36. Pride, W. M .. Lamb. C. W .. & Pletcher, B. A. (1977). Are comparative adverti ..elllcnt<.; more infommlive for owner... of the mentioncd competing brand than for nonowner';? In B. A. Gn:l!nbcrg & O. N. Bclll!ngcr (Ech.), Conreml'0rary markerill}: "wI/gilt: 1977 educator\' pmccedillX\ (<.;c ries 41. pp. 198-30 I ). ChiCCd lllechani"'llls. Hllman Cmlllllllllic{/f;,m He.H'arch. .J. 217-2JO. Rol..cach, M. (1987). Health I'(lILU'.\. Parer presented to the In"titutlon for i-Ic'llth l'romOli()n and Di'>Case Prt!\cntiun. Pasadena. CA. Ro ... enbcrg. M . (1965). SocielY will at/ole\,·etl/ Sl'/.fimfl.t:l'. Princeton. NJ: Princeton Uni\er... lI) Prc ~s. Sabato. I. J. ( 198 I) . The rise oj politiwl cml,HlIIlIllH: Nell It'("'.~ of II'inlling elntiom. Nev... York: Ba ... ic Bool.. .... Sah~lto. L. J. (1983). Partie>;. PACs. and independent group,.,. In T. E. Mann & N. J. Orn\telll (Ed,>.), The Alllnialll ('Iection oj 1982 (pp. 71-110). Wa\hington. DC: Amcrican Entt!rpri .. 1.! In ... titutc. Salomon, G .. Stein. Y .. Eisenberg. S.. & Klein, L. (1984). Adolescent !-.l1lol..er~ and non"'llloker!-.: Profile!-. and their changing strllcturl.!. Prel'l'lIIit'l' Medicine. /3,446-461 . Salomon. G. (1987). IlIferactirm .\· of mNlia. ('ognitioll, [lnd leol'll;lIg: All explanatioll of hal\' ,\'.","/Jo/ic Iorm,\ Cliltil'llre mell/al ,~kil/.\ mltl affect kn(JIl'/edge (lcqlli,\ilioll. San Franci'icu: Jos,.,ey-B;'I\s. S~I\\Yl.!r, A G. ( 197J). The effeeh of repetition of refutational and \upponi\c adveni'iing appc.lJ... jOllr/lal of Markl'1itlg ResellTch. 10. :2.'\-33. Slater. M 0 .. Beauvai\. F.. Rouner, 0 .. Van Lcu\cn, L Murph)'. K .. & Dmnenech-ROliriguez. M . M. ( 1996). Adole~cent counterarguing of TV beer ad\ert"cments: E\ idencc for cffectl\cnes\ of alcohol t!ducation and critical \ iewing di'\Cu ..... ion\. journal of Drug EdllWlio", 26. 14,1-158. Stulle. V. A. ( 19(9). Indi\ idual diffcrt!nct!' and Inoculation again ... t pcr.ua\lon, jOltrlwll\11I Quarterly, .J6, 167 173. Sucdfdd, P.. & Borrie. R. A. (1978), Sensory deprivation. altitude change. and defense a£ain .. t pcNta... inn . Callat/iall jOllrnal (~r Bl'hfl\ 'ioral Sciellce, 10. 16-27. S/ahu. E. A. (2000). IlI0clllalioll. lIormatit'(' appt'(lIs lind emolio" lU\·tratl'gie.\ /(J promote rt'sisllIlIce 10 mloit'.H'('1It \l1wkillg. Unpubli ... hed doctoml dis'>ertation. Univer ... ity of Wi ...con ... in·Madl'on. S/yhillo. G. L & HC'ilin. R. (1973). Rc .....tancc to pcrsua ... ion : Inoculation theory in a marl..ctlllg context. j Olll"l/al of Markellllg Reseanh. 10.396---403. Tannt!nhuulll. P. II.. Macaulay. J. R.. & Norm. E. L. (1966). Principle of congnlit)' and reduction in per"lIi1'ion. jmmllll oj Persol/alin' wul Socillll'sH'llOlogy. 2. 123-:!38. Tannenhaum. P. 1-1 .. & Norri ... , E. L. (1965). Effcct'i of combining congruil) principle ... tr:tlt!gie~ for the rt!duction of pcr:-.uasion. Sociomelr\', 28. 145-157. Tarrance. V L.. Jr. ( 1980). Negafil'e campaiWH fwd neglllil'e \'OIes: The 1980 ('It'l·tim/.{. Wa ... hington. DC Frt!c Congre\s Research and Education Foundation. Tate. E, & Miller. G. R. (1973, April). ResisllIlI('t! to perslI(u;o" /ol/oll'illg nuwreml1illldinlll at/mclIe.\': Smlll' preliminary fhollgh/.\. Paper presented at the annU<11 mceting of the International Communication A ......ocialion, Montreal . Tinl..h~trll. S. F.. & Weaver-Lariscy, R. A. (1993). A diagnostic approach 10 a\scs\ing thc impact of ncgativc roliticallelevision cOlllmcrci:tI~. journal of Broadcasf;lIg & electronic: Media. 37, 277-399. Tobler. N. S. (1986). Meta-analy'>i~ of 143 adolescent dnlg prcvcnlion program ... : Quantitative outcome rc!-ult ... of program panicip:tnt~ compared to a control or compari\on group. jOI//"1wf of Drug Isslles, 16.5.17-567. Ullman. W. R .. & Bodaken. E. M. (1975). Inducing resistance to pcr'\ua<.;ivc attack: A w\! o f two 'itrategie~ of communication. Weslem JOIln/al of Speech Commullication. 39, 240-2-l8. United State ... Senate (1956). Committee on Go\'ernment Operation". Permanent Subcommittee on Investigillinu ... (84th Congress, 2nd Session). CommlilliJl i1l1erroga/ioll. imJot'trillll/u}f/ and exploillltioll oj Amt',-inlll military lind poliliwll'riwmer.\'. Wa'ihington. DC: U.S. Go\'ernment Printing Office,
286
Part III • Sel'/';inf,: and Re,~;.HiI/X Compliance: Slnttegje.\ lind 1(/('1;0 Wallack.. L.. & Corbel!. K. 1191-17). Alcohol. tobacco and nKlnJu~IIl'1 u ...e among )outh. An oycnlc,", 01 epidemIOlogICal. program and policy trcnd~. lIeallll l:.'tl/lCClliOi/ Quarterly. N. 223-2·W. Wan.I-IA·1. (2000). illocu/w;o/llind priming illihe COIlIl'xl ofcri\'/I' c(JI1/lI/l/llicali(JI/. Unpuhli <.;hed doclOral di ......ertation. Univer... ity of Wi ...consin- Madi ... on, Webb, J, A .. Bacr. P. E.. & M(,·Kchcy. R. S. (1995). Development of a ri~k profile for IIltcntiom, to u . . c alcohol among fifth and \ixlh grader.... JOIII'tUl/ of the AIII('I"/«(11/ Academy of Child ami Adult'IH'tll Psychiatry. 3-1. 772-77H.
Part
IV Contexts for Persuasion In chapter 3 of Ihis volume. Daniel O'Keefe discussed what he believed were three broad recent developments in the study of persuasion. social illiluence. and complia nce gai nin g . Among thc ...c developments wa!o. the increasing amount of research focusi ng o n specific conlext", for per\uasioll. While ...OI11C pcr ... uasive strategics arc field invariant or co ntext free. most arc not. Even sli ch generic .., trategies as fear appea ls. for examp le. must be tai lored to the specific audience. topic, and setting. Tlms. a fear appea l that an utlorney mi g ht
use in
<-I
clo:-.ing argument
10
a jury (" 'I' you leI the defendant go free. he will kill aga in ,")
wou ld likely bl.! different from a fear appeal that a ca r sa lespe rso n mi ght li se in trying to clo ... e a ... a le ("Do you really wallt to risk your life ill a car without side ai r bags?"). Similarly. a po liti ca l consultant who wa~ de ... ign in g campaign ad ... that included fear appeals ("Our opponent \I.'ants to rai~e lUxes and cut social ... ccurity.") would like ly use a different approach than a pol ice officer as\igncd to th e D.A.R.E. program ("D ru gs aren't just physi-
cally addictive. they are psychologically addictive as welL"). The same app lies to other persuasive stratr.:gics, principles, and processes. Persuasive strategies rarely come in a "onc size fits va ri ety. There probably neve r wi ll be a single, unified theory o f persuasion capab le of encompassing a ll persuasive phenomena in every con tex t. Thus, the need to study how per ... uasion operates in "'pecifie contexts is vi tal. The devil. as they say. is in the detaib. and the ~ub Ll e nuances that ca n make or break a persuasive attempt lie in the particular strategies and tacti!:s that are unique lO each persuas ive contex t. For this reason we have included this final section. which ill ustrates that the study and practice of social influence have many "nooks and crannies:' In other words. contex tual or situat iona l factors such a~ the lime and place of the persuasive encounte r. or the nature of th e relationship between the persuader and the persuadee. affec t not o nly th e types of influence strategies people use but a lso how dfcetive such st ratcgies might be. Chapter 16. by Renee Klinglc. for example. explore!'. comp li ance gai ning ill med ica l context:-.. In it. you \I. ill \ce how theories and models of per~l1as i on have been u~ed to identify effective and ineffective strategies in physician-patient encounters. If a husband fails to convince.! his wife that they ... hould buy a big screen lclevi ... ioll. the conseque nces for
287
288
Part IV •
CO/lfe.usjor PerSIUlSioll
failure aren't too dire. If a phy-,ician i:-:. unable to per~lIade a paticnt to lose weight. \IOP <..,moking. or lower hi<.., or her blood pres~urc. however. the patient· ... life may hang in the balance. Chapter 17. hy Lc.,lic Baxter and CamHi Bylund, examlne~ persuasion in clo~e relationships by juxtapo.,ing three unique perspectives: the traditional, the social-meaning. and the dialogit: communication approaches. The chapter illustrates not only that the nalure of relationship.., affects the process of persua<..,ion but al"io that the way in whi<.:h we cOllceptualiLe and study per ... uasion affecl"i our undcrstanding of it. Finally, the laM two chapters in thi"i volume explore ,ocial inlluence in traditionul organiLational context .... Chapter 18. by Randy Hirokawa and Amy Wagner, examine ... per..,ua~ion a~ it occur~ between superiors. <..,ubordinate"i. and coworker"i within an organi/ation. Specifically. it examines the nature of upward and downward inOuence attemp ts in organizational sellings. Chapter 19. hy John Seiter and Michael Cody. examines persuasion in retail sales contexts. Specifically. thi"i chapter focuses on the nature of pcr"illa~ion in "elling contexts by exploring how buycl""i' and sellers' charactcri"itics. tactic<.." uno goab influence succes<..,rul sale~ and buyer-seller relationships. These are. of course. only a fc\\ of the myriad contexts in ",hich persuasion occurs. By gaining a bettcr under"tanding of ho\\ persuasion operutes in these four specitic contexts. you should develop a greater appreciation for the unique features of persuasion. "iOcial influence. and compliance gaining in other contexts and settings as well.
16 Compliance Gaining in Medical Contexts Renee Storm Klingle
Introduction to Medical Compliance Although medical adherence is generally in the best interest of the patient, patients who
,eek expert medical advice often fail
10
follow through with the pre,cribed or ,uggested
tremment regimens (Eraker, Kir ... cht. & Becker. 1984: Pitt ....
J 991:
Roter. Hall. Merisca.
Nordstrom. Cretin. & Svarstad. 1998). In 1996. I argued that "ignoring medical adv ice seems to be as common as the common cold and equally difficult to cure" (Klingie. 1996. p. 206). Since that time. remedies have been advanced that are quite effective in reducing the ,everilY and duration of the common cold. The verdict on how be . . , to eradicate patient noncompliance. however.
i... 't ill out. Even the act of defining compliance. often known as
medical adherence. has been open to debate (Henson. 1997). In the not 100 di"'Ia!l1 pa",t, the term medical compliance simply referred to the degree to which patients yielded to the suggestions. orders. or recommendations given by their health care provider (Karo ly. 1993). Failure to do so could take a number of forms. ranging from inability to comply. refusHI to comply. overcompliance. and partial compli ance wit h what the physician specifica ll y recommended (Fletcher. 1989). The recent l11oven1elll toward patient-centered care and away from paternalistic care (Ba llard-Reisch, 1990: Jones & Phillips. 1988). however. puts a different spin on the definition of medical compliance. a,>; well as the appropriate terminology. Consistent with this trend. authorityladen term'" such as compliance and adherence ha ve quickly begun to be replaced by words that highlight the more reciprocal nature of the physician-patient relationship such as mutuality. cOllcordance. cooperation. and therapellf;c alliance (Kyngas. Duffy, & Kroll, 2000). In turn. definitions of the phenomenon began to contain elements relating to the patient's collaboration with the health care provider and the patient's responsibility for se lf-care (Kyngas. 2(00). Medical compliance came to mean adhering to what the physician and patient mutually agreed upon a,>; the appropriate course of medical action. 289
290
Part IV • COIl/eX!.1 for Pcrsllmiol1
The more politically correct name:-. and definitions. hO\vever. could not change the fact that compliance rates , especially in the long term. often fall below 50 percent for patients of;)11 agc:-.. :-.ocial classes. and intellectual levels, and they remain low regardless of the severity of the :-'Ylllptoms or the life-threatening nature of the disease (Adams, Pill, & Jones, 1997; Klopovich & Thruewonhy. 1<)85: Kyngas. 20(0). The consequences of noncompliance or failed cooperation arc often quite serious: unnecessary or dangel'Ous diagnost ic and treatment procedure~ (Bec,,"er & Maiman. 19HO; Norell. 1(80), exacerbation of the medical condition and progression of the ailment (Stewart & Clulf. 1972), inaccurate assessment regarding the value of prescribed medicines or treatment regimens (Wilson. 1973), and additional costs to the patient and the medical community (Berg. Dischler. Wagner. Raia. & Palmer-Shevlin. 1<)93: Weinstein, 2000). In :-.um. a patient's Jack of adherence to medical recommendation:-. that arc either sct by the phy~ician or mutually agreed upon by the physician and patient is nothing to sneeLe at. Unlike the common cold. noncompliance has the potential 10 be life-threatening for the patient and an economic disaster for the medical community. For decade~. ~()cial scientists have conducted studic,- aimed at assessing interventions linked to improved medic.:d compliance rates. Studies in the medical field have typically focused on educational intervention.'- such a:-. verbal. audiovisual, and written instructions and reminders: behavioral pattern shaping and skill building: and appeals to emotions and social relationships (Roter ct al.. 1(98). In each advocated intervention, comIllunication between health care providers and paticnh is cited as playing an integral role in adherence lO medical prescriptions. Medical article~ (e.g .. Dube. O'Donnel. & Novack, 2000: Meryn. 1998) also reiterate the need for better physician commun icalion sk ills to improve adherence rates. The exact content and nature of the physician's compliancegaining <.IlIempl. however. is genemlly unspcci fied (KI ingle. I <)96; Ph iII ips & Jones. 1991). Rather, advice tends to be overly hroad or general in nature. Most problematic is the fact that the vast majority of thc compliance-gaining recommendat ions offered have been atheoretical. After an extensive 1"C\ie\\-' of the medical compliance literature from 1970 to 1989, Dunbar. Dunning. and Dwyer (1993) concluded that "as the absolute number of studic~ has increased. the proportion of studies that have been theory driven have decreased" (pp. 36-:n). Lillic has changed since. particularly in the area of interpersonal inlluence messages and medical adherence (Klinglc. 19<)6). Given the lack of theoretical guidance, it is no wonder that complialll:c-gaining advice is broadly stated and often limited to methods of increasing patient compreh(:n~ion. This chapter deline,lIes eITecti\(: medical compliance-gaining strategies by reviewing per~uasion models and theories thaL uniquely address medical adherence. rathe r than reviewing the plethora of atheoretical research from the medical field. Additionally. the focus of this chapler is on communication as opposed to other factors thaI may affect compliance sllch as patient personality or access to medical care. An examination of three models that focus on health beliefs. Lhe HealLh Belief Model. the Extended Parallel Processing Model. and the Theory of Reasoned Action and Planned Behavior. provides the starting point. These more traditional belief-adj ustment models have frequently been used lo guide and evaluate the "contcnt" of health prevention campaigns and have recently been used by interpcrsonal researchers in their qu(:~t to develop interpersonal messages directed at incrcasing adherence rates. These belief change models. however. tend to
Chapter 16 • Compliance Gaining ill Medical COl1fe\1.\
291
ignore the role of nonverbal coml11unic[ltion and the relational meanings attached to met-.· sage",. The second part of thi", chapter [lddrc::-.ses Reinforccment Expectancy Theory (RET) which addrcs ...es thc relational \lyle of the cOl11pliance·gaining message rather than its content. The final part of this chapter brielly examines the role or patient participation in the health care decision-making process.
Health Belief Perspectives and Patient Compliance Medical compliance is often regarded as 11 consensual process, involving collaboration bet\.\cen the health care provider and patient as to medical regimens and therapeutic expectations (Anderson & Kirk. 1982: Linden. 1981). Corresponding with this viewpoint i ... the development of numerous compliance models espou ... ing the importance of education and attitudinal adjustmcnt in order to "~dign" expectation ... (Becker. 1974: Hciby & Carl· ,on. 19R6). According to a meta-analy,i, by Roter et al. (1998). the bulk of medical compliance-gaining studies publi:-.hed between 1977 and 1994 focu:-.ed either exclusively on educational interventions or on combinations of both educational and behavioral inter· vent ions (e.g ..... kill building and behavioral modeling). Educational methods for eliciting compliance also "eem lO be the ... trategy of choice among phy~ician~. According to empirical findings by Burgoon and colleaguc, (M. Burgoon. Parrott. J. K. Burgoon, Birk et al.. 1990: M. Burgoon. Parrott. J. K. Burgoon. Coker et al.. 1990) both patient' and physician, report that health carc pro\'ider ... are mo~t likely to emplo} the lISC of expert power by ghillg simple direction ... or appealing to the knowledge or experti:-.e of the physician. rather than using threats or prosocial reinforcing ~trategic:-.. Schneider and Beaubien'~ (1996) l1aturali~tic investigation of compliance-gaining stratcgie~ employed by doclOrs in medical interviews also found expertise and liking strategic ... to be umong the most widely used compliance-gaining MrHtegie .... However. educational programs mu~t involve more than expert advice and instructions. since merely under~Landing what to do and how to do it have nOf been ~trongly associated with compliance (cf. Fulmer et al.. 1999). The phy,ician must make appropriate attitudinal adju,tments to patient" belief 'y'tem, (Becker & Maiman. 1980: Webb. 1980). Commonly used belief models and theories in the health arena ~uch as the Heallh Belief Model. Extended Parallel Proce"ing Model. the Theory of Rea,oned AClion. and the Theory of Planned Behavior provide recolllmendations for phy ... icians by addressing ~pe· cific belief~ that should form the core of educational intervention directed at improving patient compliance. It is beyond the scope of this chapter to articulate fully the complete formulation ... of each of these 1110dels. Instead. a summary of the central aspects of the Jl10deb and theories that arc 1110~t germane to the topic of message content is offered. followed by a discussion of each critical variable that should guide the content of the compliance-gaining message.
rhe Health Belief Model The Health Belief Model (HBM) was originally developed to e'pl"in low compliance with preventive health beha\'iors (Bt:cker & Maiman. 1975. Janl & Becker. 1984:
292
Part IV • COIl/ex/sf"r Persuasio/l Ro,en!o.tock. 1974) and has been used as an organizing framework for developing and explaining a variel) of preventive mcssages. from wearing bicycle helmel~ (Witte. Stokols. Ituartc. & Schneider. 1993) to ha\ing mammograms (Hyman. Baker. Ephraim. Moadel. & Phillip. 1994) lO managing eating disorder~ (Grodner. 1991: Smalcc & Klingle. 2000) to practicing safer sex (Matt,on. 1999). According to the HBM model, patients will adhere to the physician's recommendations when the physician's inlluencc me,sage addresses several components. The first component physicians should address includes perceptions or severity and susceptibility. According to the HBM. threatening messages must communicate to the recipient both lhal the effects of not complying with a health recommendation are serious (perceil 'ed se\'erilY) and that the recipient is pcrsonably vulnerable 10 the negative consequence (percei,'ed s/lsceptibility) (Becker. 1974). Health risk belief,. or Ihreals. are considered by many lO bc a vital clement in persuading others to comply with health recommendations. bccause without threat. the danger goes unnoticed, and no aClion is laken (Lave. 1987). Perceived ,everity rellect, the individual's beliefs Ihal nol following health recommendations will result in dangerous. life threatening. and serious outcomes (Becker. 1974: Grodncr. 1991). The more severe the threat appears. the more the individual will be intere"ed in avoiding Ihe harmful con,equences (Rogers. 1983: S'lIIon. 1982). For instance. Klohn and Rogers (1991) found that women who were given information regarding the severe effects or osteoporosis. such a, ,,>erious disfigurement. reported significantly higher perceptions of Ihe sevcrity and greater intention to comply with the recommended responses than did those who received only general information or no information at all. Thus. phy ... icians attempling 10 gain patient cooperation must inform patients of the potentially seriolls consequences associated wilh not following through on the mutually agreed recommcndat ions. Paticnt, mu,t also feel personally at risk for contracting or developing seriolls effects. As an illustration of the importance of perceived susceptibility, Thurman and Franklin (1990) found that although 60 percent of the students in their study feared a c<1mpuswidc spread of AIDS. les ... then 25 percellt of them felt personally susceptible to contracting Ihe disease and less than 50 percent changed their behaviors. Increasing paticnt ... · pcrcepliol1s of susceptihility can be a major obstacle for physicians. since researchers ha\e found Ihal few people believe Ihal the worsl will happen 10 Ihem (Perloff. 1983: Wci'Ncin. 1983: WeinSlein & Lachendro. 1982). Wille (1992,,) poinl' OUI Ihal adolescents may bc the most challenging audience. ,incc young people tend to believe that they are invincihle. In addition 10 perceptions of threat. the HBM suggests that sc({-e.tJicacy is an important component for physicians to address in their influence messages. In relation to health problems, self-efficacy refers to patients' own expectations that they can easily alter their own actions to enhance their well-being (Hertog, Finnegan. Rooney, Viswanath. & Potter. 1993). In the area of notoriollsly hard-to-change behaviors. perceptions of self-efficacy have proved panicularly important in gelling patients 10 consider a~ well as actually take approprialc aClions (Herlog el al.. 1993). Perceplions of self-efficacy have predicled alcohol abstinence (DiClemente. Carbonari. Montgomery. & Hughes. 1993: Rolnick &
Chapler 16 • COl1lpliance Gaining ill Medical COl/texts
293
Heather. 1982). smoking cessation (DiClemente, 1981). relapse of addictive behaviors in
general (Marlatt & Gordon. 1980), preventive oral health behaviors (Tedesco. Keffer. & Fleck-Kandath. 1991). changes in diet to prevent cancer risks (Hertog et aI., 1993). adolescents' ability to resist pressure to smoke (DeVries, Kok. & Dijkstra. 1990), and whether bulimics will seek medical help (Smalec & Klingle, 2000). Although many studies have merely looked at the relationship between perceived efficacy and compliance, several studic!o. have shown that messages can be effectively manipulated to communicate efficacy. which in turn can facilitate compliance. Rippetoe and
Rogers (1987). for example. found that women who received high-self-efficacy literature on the ease of performing breast self-exams were significalllly more likely to conduct selfexams than women who were given low- self-efficacy literature. Prentice-Dunn, Jones. and Floyd (1997) manipUlated self-efficacy in messages regarding skin cancer prevention and found that subjects who read the high-efficacy message believed that wearing sunscreen and reducing sun exposure time would be effective. in contrast to subjects who read
the low-efficacy message. Smalec and Klingle (2000) demonstrated that individuals with eating disorders were more likely to seek help for their problems if someone in their social network communicated to themlhat getting help would be simple and relatively effortless.
Witte (1992b) manipulated the efficacy of an AIDS prevention message and found that subjects receiving the high-efficacy message as opposed to the low-efficacy message believed more strongly that they could use condoms (0 prevent AlDS. developed more favorable aftitudes toward condoms. and were more likely to report using condoms as long as they perceived a significant AIDS threat in l1ddition to eflicilcy.
According to the HBM. the third component that should be addressed in physicians' influence messages is perceptions a/benefits and costs. Specifically. the HBM states that the benefits of complying must outweigh the costs or barriers associated with not comply-
ing. Perceived benefits in the HBM have been defined by Brown, DiClemente, and Reynolds (1991) as "the individual's beliefs regarding the effectiveness of strategies designed to decrease vulnerability or reduce the threat of illness" (p. 51). Perceived barriers are described as obstacles in the path of performing the recommended response (Leven-
thal & Cameron, 1987). Jan7 and Becker (1984) found that perceived barriers accounted for more behavior change than any of the other HBM components. Spector (2000) argues that compliance can be improved by communicating to patients the potential baniers to compliance such as side effects to medication. Barriers can be more than just the side effects, however. They can be physical, psychological, or I1nancial. Barriers related to quilting smoking, for instance, might include weight gain and irritability, fear of losing friends who still smoke, and financial constraints associated with buying the patch or attending a smoking cessation clinic. According to the HBM. if the barriers outweigh the benefits, noncompliance is more likely; if the
benefits outweigh the barriers. compliance is more likely (Hayes, 1991; Janz & Becker, 1984). The final component discussed by the HBM is cues to action. When attempting to influence threat and efficacy perceptions, the HBM states that "cues to action," such as advice from a physician. are needed to create awareness of the threat (Janz & Becker,
1984). These cues may be either internal. such as medical symptoms experienced by the
294
Part IV • CO/llexts for PerSl/asio/l
patient or external, such as advice fmm a physician. a reminder letter, or a gruesome picture. Both internal and external cues have a direct effect on threat and an indirect effect on health-compromising behaviors. From a physician's standpoint, cues to action are the physician's messages that help bring the threat to the forefront for the patient and persuade him or her to carry out the recommended response. According to Manson (1999). communication cues to action should be at the center of the HBM, since health beliefs, perceptions, and behavioral decisions are "socially constructed and contingent upon interaction with others" (p. 258). Applied to thc medical setting and physician compliance-gaining strategies specifically, the HBM postulates that patients will be more likely to comply if the physician communicates to them (e.g .. provides "communication cues to action") that (1) the outcomcs of noncompliance are severe. (2) they are vulnerable to these outcomes. and (3) the advocated health recommendation will produce positive results that outweigh the difficulties or barriers associated with noncompliance. At the same Lime the physician must convince patients that they have the skills to carry out the recommended action effectively. Although the HBM delineates the importance of critical message variables in shaping or altering patient health behavior. the relationships between the HBM variables are not theoretically specitied in the model. The Extended Parallel Processing Model (EPPM) reconciles this problem and illustrates the relationship between threat and eflicacy.
Extellded Parallel Processillg Model Witte's (1992a, 1992b) EPPM (see also chapter 13 of this volume) improves upon the HBM by explaining "how" to construct fear-appeal messages efficaciously lIsing the right combination of threat and efficacy. Although fear appeals and threats have been a primary foclis of many persuasion theories, health communication scholars in the medical compliance arena have predominantly focused on interpersonal communication that produces satisfying relationships. arguing that compliance is facilitated by patient satisfaction (Cousins, 1985; DiMatteo, Prince, & Taranta, 1979; Hanson. 1986; Pendleton, 1983). Research evidence associating satisfaction with compliance. however. is often negligible (M. Burgoon, 1991: Ley, 1988). Findings that threatening communication by the physician leads to patient compliance (e.g .. M. Burgoon, Birk, & Hall. 1991; Kaplan, Greenfield, & Ware, 1989; Robberson & Rogers. 1988) also appear to contradict the presumed effect of a provider's friendly bedside manner on adherence. Kaplan and colleagues (1989) examined physician communication style in a longillidinal study and found that it was physicians' negative affect, nor their positive affect. that related to positive health status. Perhaps the "fear" of lIsing fear appeals stems from the fact that threats have actually been known to backfire. Witte (I 992a) developed the EPPM to address how to use threats without fear of a boomerang effect. The EPPM illustrates the interdependent relationship between threat and efficacy and proposes that messages communicating threat (severity and susceptibility) and effi· caey (self- and response efficacy) influence perceptions of threat and efficacy, which interact to produce either health-promoting actions or defensive actions. According to Witte's EPPM (1992a, 1998). delivery of a threatening message activates a two-appraisal
Chi.lpter 16 • Compliance Gaining ill Medical COllfexts
295
process in the receiver. First. the individual evaluates the influence message to determine the degree of threat possible if the recommended response is not adopted. Similar to the HBM, the (wo underlying dimensions of a threatening message are susceptibility and sevcrity. If the message doesn't adequately convey to the individual that he or she is susceptible to a severe threat, the message will not be appraised further. People are only motivated to continue message processing when a threat is relevant and serious. If the message does adequately convey threat, the individual goes on to the second appraisal process, which involves the evaluation of efficacy. Efficacy, like threat, is a twodimensional variable con!ooisting of respollse efficacy and self·efficacy. Self-efficacy is defined in a manner consistent with the HBM's self-efficacy component. while response efficacy is defined in a manner consistent with HBM's perceived benefit component. Specifically. response efficacy is a person's belief that the advocated response will produce the desired outcome; perceived self-efficacy is a person's belief that he or she has the ability to execute the desired response. According to EPPM, perceived efficacy interacts with threat to determine whether an individual will go into danger control (message acceptance) or fear control (message rejection). As long as efficacy perceptions are higher than threat perceptions, the individual will go into danger control, which is an adaptive response involving message acceptance and taking the recommcnded action to prevent the danger from happening. However. if efficacy perceptions are lower than threat perceptions, the individual will go into fear control. which is a maladaptive response involving message rejection. Rather than accepting the message, the individual who has high threat and low efficacy then attcmpts lO control the level of fear associated with not being able to avert the threat through slich methods as rationaliling away vulnerability. Wine (1994) clearly pointed out, and has shown through a recent meta-analysis (Wine & Allen, 2000). that allhough threats are needed to motivate people, if perceived as too high, they can create slich insurmountable levels of fear that people will respond by expending energy to rationalile away the fear (e.g., ''I'm not going to get lung cancer if I !oomoke cigarcttes for a few more years. I can quit later") rather than by taking action to avoid the danger (e.g., ''I'm going (Q quit smoking as my doctor recommends"). Back in 1977. Bandura also theorized that high arousal states can be debilitating to performance and that individuals tend to avoid threatening situations that exceed their coping skills. Several ~tudies addressing adherence rates in patients with chronic disease have shown thar when the treatment prescribed is challenging, threat is negatively related to compliance (e.g., Bond. Aiken, & Somerville, 1992; Hartman & Becker, 1978). Applied to the medical seUing, EPPM postulates that patients will be more likely to comply with advocated health advice if the physician communicates to the patient that the outcomes of noncompliance are severe and that the patient is vulnerable to these outcomes. At the same time, the physician must clearly communicate that the patient can enact the behavior recommendations and that they will be effective. It is of critical importance, according 10 EPPM. that threatening messages not be used in isolation. Compliance-gaining attempts by a physician must do more than scare the patient into action. The compliancegaining attempt must also include arguments demonstrating that the advocated response is effective at eliminating the threat and that the patient can carry out the recommended behavior.
296
Part IV • COlltexts jor Persllasion
The Theory of Reasoned Action and the Theory of Planned Behavior Missing from both the HBM and the EPPM is the concept of social normative beliefs, or an individual's beliefs regarding what other people think he or she should or should not do. The component of social normative beliefs. when compared to other cognitive components such as health beliefs, has heen shown to be the stronger predictor of health behaviors (Seibold & Roper. 1979). Classic persuasion perspectives sllch as Kelman's (1961) influence processes and French and Raven's (1959) power bases provided the original foundation for the claim that individuals in our environment serve as frames of reference for making judgments about whether or not to carry out behavioral recommendations. According to Kelman's identification process, compliance is often the result of a recipient'S desire to maintain a self-defined relationship with the source. Similarly. French and Raven's reference power indicated that compliance can result from a recipient's identification with the source and desire to do things to please the source. Regardless of whether one presumes that social normative beliefs influence compliance because of individuals' innate need for social approval or because of their need to maintain satisfying relationships, the power of normative beliefs in influencing compliance decisions should not be underestimated or overlooked. Two of the most prominent theoretical models that address normative beliefs are the Theory of Reasoned Action (Ajzen & Fishbein, 1980: Fishbein & Ajzen, 1975) and the Theory of Planned Behavior (Ajzen , 1985, 1988). Both theories argue that intention to behave has a direct influence on actual behavior and is a function of the person's attitude toward the be/wI'ior and the subjective norm with regard to the behavior. Attitude toward the behavior is a function of the personal consequences expected from enacting a behavior and the affeclil'e value placed on those consequences. Personal consequences include the costs and benefits associated with the behavior. Thus, personal consequences encompass notions of benefits, barriers, response efficacy, susceptibility, and severity addressed in the HBM and EPPM. Subjective norm is often equated with social pressure (Friedman. Lichtenstein, & Biglan, 1985) and is defined as beliefs about what specific individuals or groups think they should do and their motivation to comply with or desire to please each of these referents. The Theory of Planned Behavior extended the Theory of Reasoned Action by claiming that intention toward a behavior is a function of behavioral control in addition to atti tude toward the behavior and subjective norm (Ajzen, 1985, 1988). This modification was necessary for the theory 10 be applied to situations in which the behavior in question was not under complete volitional control. For instance, practicing safer sex is affected by the actions of one' s sexual partner and one's own ability to practice safer sex, in addition to one's attitude toward practicing safer sex and normative beliefs. Petraitis, Flay, and Miller (1995) argued that behavioral control is a concept similar to self-efficacy in that both terms are related to "perceptions of control over the sliccessful completion of a particular behavior" (p. 69). According to the Theory of Planned Behavior, self-efficacy (i.e., behavioral control) has a direct influence on both intention and behavior. Thus, an indi vidual may have the correct attitudes and nonnative beliefs but still fail to carry out the behavior because of low self-efficacy. This claim is similar to Witte's (1992a, 1994) prediction that
Chapter 16 • CompliCll/ce Gail/111M ill MedicClI CmUt:,lts
297
threat beliefs are influential only to the extent thai the individual believes he or she can carry out the recommended actions necessary to avert the threat. The Theory of Planned Behavior, however. adds the construct of subjective norm to the compliance equation. Both theories have proved useful in predicting a variety of health related behaviors (Kashima. Gallais. & McCamish. 1992; Petraitis et al.. 1995). and the addition of the subjective norln has turned out to be quite valuable. For instance, in adolescents, one of the best predictors of sub~tance use among high school students is social normative beliefs regarding classmates' substance use (Johnston. O·Malicy. & Bachman. 2000; Klingle & Miller, 1999). College students' decisions 10 use substances such as marijuana have been strongly associated with social normative beliefs regarding marijuana use (Ajzen. Timko, & White. 1982). In regards to patient compliance, Tedesco and collcagues (1991) found that intention to brush and floss was significantly a. . sociated with attitude toward :he behavior and subjective norm. Applied to the medical setting. the Theory of Reasoned Action and the Theory of Planned Behavior suggest that a physician needs to construct messages aimed at altering (I) attitudes of perceived severity, susceptibility, and efficacy toward the health behavior, (2) subjective nonn ..... and (3) efficacy or behavioral control. Subjective norms are often equated with loved ones such as family members and friends. The physician. however, is also an important referent, and studies have found that a physician's recommendation is among the most influentiul factors in a patient's health care decisions (Dube et al.. 2000). Kyngas (2000) showed that paticnts who received support from their physician were more likely to comply than patients who did not receive . . uch ~lIpport. As an important referent, the physician can modify his or her verbal and nonverbal language choice.s to communicate approval or disapproval of the patient's actions. which in turn . . hould influence a patient's future behaviors. Thi . . is a central notion of Reinforcement Expectancy Theory.
Reillforcement Expectallcy Theory Although attitudinal adjustment and skill building are certainly integralLO a patient's ability to follow prescribed treatment regimens, they do nOI guarantee that a patient will act appropriately. As Gross (1987) Slates, ··knowing what 10 do and how to do it in no way in ... ures cooperation" (p. 10). A substantial amount of research (e.g., Cummings, Becker, Kirscht. & Levin. 1982; Dunbar & Angras, 1980; Kirscht & Rosenstock, 1979; Mazzuca, 1982; Podshadley & Schweikle, 1970; Webb. 1980) has demonstrated that patients who hecome more knowledgeable are not necessarily more compliant. According to Altman and King (1986). these noncOirpliant patients who comprehend and concur with the necessity of the medical regimen are often the nonn. rather than the exception, in chronic disease trcatment and prcvention programs. Montgomcry el al. (1989) stated that chronic. lifelong diseases such as AIDS necessitate the development of more adequate theoretical frameworks than those offered by morc traditional models such as the HBM. Where the belief change approaches fall short is in flot recognizing that in many situations patients have the appropriate belief structure but are confronted with a number of obstacles that affect subsequent behavior. Patients who are chronically ill or who are seeking lifestyle changes often fail to comply because of motivational shortfalls rather
298
Part IV • Contexts/or Persuasion
than as a result of misaligned belief st ructure s. These patients visit heahh care professionals on a rcgular basis and arc often asked to comply with complex managcment regime ns. The adherence rates for such lifelong changes are generally lower than for short-term medical regime ns. and the se rates decrease dramatically with time (Bloom. Ccrkoney. & Hart, 1980; Epstein & eluss, 1982). Although the number of patients aUempting difficult lifestyle chan ges is increasing. most strategic suggestions to improve chronic conditions and to alter patients' lifesty les are generalized from studies associated with relatively brief regimens (Klingle. 1993; Turk, Salovey, & Lin, 1986). This approach is somewhat suspec t because it fail s to recognize the uniqueness of long-term compliance. Long-term compliance by definition co ntains a time element that greatly affects the meanings attached to commu nication tran sac tion ~ . individuals' expectations for future interactions (J. K. Burgoon & Lc Poire, 1991 ). and the acceptability of communicalion respo nses (Le vi nger & Huesman. 1980). Additionally. situations invol vi ng long-term compliance are more likely to require time for positive consequences to manifest them selves. which decreases a patient's initial moti vation to comply (Gross, 1987). Grounded in the notion that medical noncompliance is often related to patients' motivational difficulties in adhering to long-term management programs, Klingl e (1993 , 1996) developed Reinforcement Expectancy Theory (RET) to address how physicians can use verbal and nonverbal compliance-gaining strategies to motivate patients in initial encounters as well as to increase behavioral persistence. The following sec tiun s examine this theory.
Overview of RET The logic underlying RET rests on the premise that human behavior is driven by th e need to gain rewarding stimuli and eliminate aversive stimuli. To address influe nce messages. RET classifies verbal and nonverbal messages as e ither rewarding. neutral. or aversive based on the degree to which th ey com municate approval or disappro va l for a patient; physician approval cues are assumed to be innately rewarding. and phys ician disapproval cues are assumed to be innately aversive to patients. The theory illustrates how approval and disapproval messages can shape patients' communication reinforcement expectations and in turn motivate and guide patients' prese nt and future behaviors (Klin g lc, 1993; 1996). To articulate the theory fully requires describing the co mpliance-guinin g conceptualization scheme used by RET; explicating RET' s predictions for initial encounters with a physician; comparing those predictions with those offered by Language Expectancy Theory (LET), which makes identical predictions but for different reasons; and, finally. delineating RET's unique predictions for repeated influence attempts by th e physic ian .
Conceptualization of Verbal alld NOllverbal Physician Illfluence Strategies Application of reinforcement principles to human illlc ractions must begin by conce ptually defining compliance-gaining messages based on thei r rei nforci ng properties. Based on the
Chapter 16 • Complilll1Cl' Guill/Ill: III Medical COI1It'.\f\
299
notion that humans have a ~trol1g need for approval from other~ (Harre. 1980). RET classifies verbal and nonverbal ~trategies based on thc type of approval cOlllmunicated by the mcssage (Klingle. 1993). Communication choiccs that signal positivc fcelings towards an individual or that individual's actions are considered positi\le regard straret;ies. Convcrsely. communication choice~ that indicate negative feeling~ toward an individual andl or that individual's action»' arc labeled "ega/iI'e regard stta/egie.f. Comlllunication choice~ that are ~imple directives orjuMifications are llelllral regard strategies that. used in isolation from other strategies. do not indicate approval or disapproval for an indi\idual andlor an individual'~ actions. The RET framework use~ the relational l11e~:-.age approach for categoril.:ing compliance-gaining messages because of the enorll1l}u:-, importance hUlllan:-. attach to the relational meanings of messages and to approval cue ... in gcncral (1. Burgoon & Le Poire. 1991). In the health care context. researchers have suggested that a health care provider's relational message is much morc inlluential than the content of the mes\ugc (Buller & Buller. 1987: Street & Wiemann. 1987). Although a variety of relational mes\age~ differentiate compliance-gaining strategies (e.g .. 1. Burgoon & Hale. 1987). RET rOCUSe\ on messages that ~ignal (he degree of approval because approval cues are uniycrsal rcinforcers and arc thus ideal for the applicalion of reinforcement principles. Conccptualiling compliam;e-gaining strategies based all their reinforcing properties is consistent with several communication scholars' classifications of innllencc attempts (Miller. 1983. Roloff & Barnicotl. 1979: M. Burgoon. Parrotl. 1. Burgoon, Coker el al.. 1990). Missing from most past conceptualizations are innucnce aile mph that arc neither rewarding nor punishing and nonverbal intluence strategies. Neutral strategie~ used in isolation during single-episode encounters would be neither reinforcing nor aversive and would tllll!o. have little motivationailltility. However, a~ discussed later. when used in combination with reinforcing or avcr:-.ive strategie . . , neutral :-.trategies call becomc potential influence attempts because they represent the removal of either a positive or negative state of atTair:-., A second problem with pre\ious compliance-gaining conceptualiwtions is thaI most measures of inlluence are ba~ed entirely on verbal utterance ..... which Berger ( 1985) states constnlin researchers to ~tudying "the tip of a very large iceberg" (p. 483). Nonverbal immediacy behaviors serve as comparable communicative indices to the proposed verbal categorization scheme, because they are associated with the degree of interpersonal warmth or approval for another. According to Mehrabian (1969), immediacy helul\'hu's are approach behaviors that signal interest in. involvement with, and affect for another. Conversely. /wl/immediacy belim'iors signal hostility toward. exclusion of. and a lack of sensory engagement with another. Numerous nonverbal cues associated with immediacy have been cited in the literature, the most common being close conversational di\tance. direct body and facial orientation. positive reinforcers such as smiling and pleasant facial expressiveness. touching. forward body leaning. and a high degree of eye contact. Opposite behaviors. including scowls. negative facial expressions. and cold vocal tones. communicate Ilonimmediacy (Anderscn, 1985), A substantial amount of rc~carch has shown that nonimmediacy behavior~ signal disapproval or negative regard and act as aversive stimulation. whereas immediacy behaviors signal approval or po-.itivc regard (J. K. Burgoon. Buller. Hale. & deTurk. 1984: Coker & 1. K. Burgoon. 1987).
300
ParI IV • Cmuextsfor Persuasioll Table 16.1 dj~plays somc cxamples of the three types of verbal and nonverbal compliance-gaining strategies used by RET: positive regard. neutral regard, and negative regard. Classifying messages based on the degree to which they communicate approval or disapproval for the patient allows a number of additional physician influence atlcmpts lO be generated that are not specifically depicted in the conceptualization scheme. The initial test of the conceptualization scheme provided empirical evidence that positive, neutral, and negative regard strategies are differentially evaluated: positive regard strategies used by physicians show the greatest amount of approval and are rewarding to patients. wherea!o. negative regard strategies used by physicians ~how the least amount of approval and arc aversive to patients (Klingle & Burgoon. 1995).
Initia/lnfluence Attempts Numerous research efforts have attempted to conceptualize and categorize compliancegaining strategies (e.g .. Boster & Stiff. 1984; M. Burgoon. Dillard. Koper. & Doran. 1984; deTurck. 1985). Such efforts. however. have rarely addressed effective strategy use. Even rarer are research efforts that have looked at effective strategy use over time. RET was specifically developed to explain strategy effectiveness in both initial and ongoing physician-patient encounters. The predictions advanced by RET for initial physician-patient encounters coincided with those advanced by Language Expectancy Theory (LET) and were based in part on some of the basic tenets of LET. As outlined in chapter 9. LET maintains that message persuasiveness is a function of the receiver's language expectations that are grounded in sociocultural norms or standards for variolls speakers. According to LET. message acceptance occurs when a speaker engages in a positive violation of expectations (i.c .. uses strategies that are both appropriate and more desirable than expected); message rejection occurs when a speaker engages in a negative violation of expectations (i.e .. uses strategies that are socially inappropriate given what is expected for a certain communicator) (M. Burgoon. 1990; M. Burgoon & Klingle. 1998). One characteristic that has been shown to influence receivers' judgments regarding expected language b the sex of the communicator. Numerous studies have shown that women are expected to be more emotional. less verbally aggressive. less assertive. more affiliativc. and 1110re nurturing than men in the way they communicate (Bell. 1981; M. Burgoon. Dillard. Koper et aI., 1984; Fitzpatrick & Winke. 1979; Weisman & Teitelbaum. 1989). Studies tCSling LET (e.g .. M. Burgoon. Dillard. & Doran. 1984; M. Burgoon. Dillard, Koper et al.. 1984) have supported the claim that females in ini tial encounters are most persuasive when using affiliative or non-intense strategies, as opposed to aversive or intense strategies. On the other hand. males are most persuasive using either affiliative or aven.,i ve strategies as opposed to neutral strategies. M. Burgoon and colleagues ( 1991 ) put their theory to the test in the medical encounter and demonstrated that male physicians were most persuasive using affiliative or aggressive compliance-gaining strategies. as opposed to the neutral expertise strategies that physicians typically use. Female physicians, on the other hand. were most persuasive using affiliative. compliance-gaining strategies and least persuasive using neutral expertise strategies, followed by aggressive strategies. One problem with applying LET to medical compliance is that LET focuses on single-episode encounters and is unable to adequately address physicians' efforts to gain
Ch~lptcr 16 • COlllpliwKt! Gajlljllg
T f\B LE 16. 1
jll
Ml'llic(I/ Cmllt!.\l\
30]
Strategy Defiuitiom; lIlId £xllI"ples
Ne utra l Regard Stratcg) Definition: Communication requc~t\ that are "Iimplc dircctive~ or ju,tifications. These \erbal \tratcgie, ... ignal neither approval nor di,ap-
prmal for the patient or the paticnt\, act lOll' ;:U1d would be complemented with moderate nOll verbal immediacy behavior, (moderate eye contact. no bod) lean. 1;:1t.:'" of ~l11i l illg and p leasant facial exprc:-.:-.ivenc!o.:-'). 7)'1)('.\ ami EtWllples Dirl'cr RtJ{jJl(Jsr: ReqLle~t' that td l the patient what to do.
"Thc;re arc ~cveral dietary change ... I would like you to ma"'c"· "You need to change your c;:lIing habit,:· }wrijicat;oll Ba'\t!d (HI E.tpe,.ri,\e: Rcquc,t.., ba ... cd on experti ...e or re"'c'lfI.:h.
·· In my opinion. you ... houldn't put off the te,,"""·
'·Since re ... earch indicate.., that diet i... "'t!), plca ...c log \\ohat
YOll
cat:'
}u,Hi(icliriol/ BWelloll Palient Condilion: Reljue""" made bcc~lU ...e of the pa tient· ... particular illness. ··Seeing a dietitian is the be"t advice' can give for your situation:· '" know from treating ... imilar C~I"'C'" that the 'Ie change" u ... uall) ,ol\c the prohlcm."
l)ositi H Rega rd Stra tegy Definiti on: Communication reque"'b that arc ",uppol1i\c. cOI1\":y lIndcr:-.tanding. or stress conct:rn for the p.tticnt. The. . e verbal ... trategic ... "gnal approval of the patient and/or thc patient· ... actions and would be complemented w ith high nOll\erbal immediaC) behavior ... (llirect hody/facial orientation .... miling. plc'l:-.anl fadal cxpre . . sivencss. forward body Ic,lning, and a high degrcc of eye conIaCI) and al(c rcentrism (vocal warmth/in tercq). TVjJt!., {lml
Ell/lllph'.\
SlIjJporrinJ Reqllest~: Requc'l" that reinforce. rca ...... ure. complimcnt. or promi ...e
bendit ... for compliance. ··You·1I feci "10 much bellcr about your..,elfbeeau ...e you·1I know you arc doing what it takes to prc\cnt problcm ... in the future:·
·· 1can lell you've been Irying really hard- no\\ ju ... t lake that extra qep ;;lI1d eliminalc all the food.., we di..,cu.., ...ed." Validmioll Reql/est.,\: Reqllc ... " that :1cknO\\ Icdgc the difficuhy of the compliance act and ind icate confidence in thc patient following the request.
·' 1 know that changing one· ... eating Iwbil ... i'l very difficult, but you'rc the kind of pcr... on who can do il." " If YOli make these change'l- and I "'now you call----cvcryon~ will be ... o proud of you because we .. 11 know how diflicult it is:· COllllflOlllllifY
oJGoals: Requc..,,, that ... tre..,,, mutual concern. "ffeet. or
"\,,·c"ne,,·,. ··We both wanl you to get beller .... o p1t:a"c cat right evcry d<1Y:· "Wc both \\ant to find oul what could be l'ausing you to feci ... 0 run down . ..,0 ple,l',e make the appointment to haV'L~ thc tC'lt ... :· (amtlllllf'd)
302
Pan IV •
COIllt!.\f.\.for Persuasion
TABLE 16.1
Continued
Negathe Regard Strategy Definition: Communication requeM!o. thal :.mack or critic;/!! the p:'ltienl'~ pa~l bcha\ior~ or potential future behavior. or reqlle"l~ lhal ;:tttribule primary rc",pom~i · bility to the p:.uient for feeling ill. Thc~c verbal ... tralcgies signal di"';'lpproval for the patient and/or the patient's action~ and would be complemented with an unplea\tlnt and dominant nonverbal intcr:.Iction style that included egocenlri~m (backward lean. cold vocal tones. rcduced eye con· tact) and negative feed hack (neutral or negativc facial exprc,>,')iom. an occa~ional direct look). 1\pe.\ amI
EXlIl1If1fl, \
NOII.HlPl'onil 't! Rl' qllt!J/s: Reqlles(~
that ~ugge . . t the ~implicity of the request andlor indic:.Ite di . . belief in the patient's willillgne ... ~ to m;'lke the change~ . .. It· ... not going to take Ihat much of your time to
~cc
"Jf thi ... i. . diabetes. the ... olution i. . genera ll y quitc rC~1 of your life wi ... hing you had."
a dietitian:'
~iT11p l e-stick
to your diet or spcnd the
11I1'lI/idalio" R(' quesl.\: Reque~ts that critici/e or :.ttI.lck the patient· ... self-concept andlor
indicate di ...appointment in the patient· . . previou ... actions. "You can't keep fooling around \... ilh your dil!t-:.I nO\\ i~ Ihe time to take charge." "You have to see by now that
it'~
re~pon!o.ible
ah . . olutely irrational not
10
pcr<:.on would know that
make the
changc~
we
di~cu . . ~ed."
Negatil'e CmHl'qlll'flCt!,\:
negative
Reque,,",s that ",uggc,,", that noncompliant actions will lead to
con~cqucnces.
"Jf you won't follow this advice. you're going 10 continue to feci run dO\\n and that . . implc:·
tired - it'~
"Your irregular eating habits arc bound to make you overeat and gain weight."
ongoing compliance. RET extends LET by applying reinforcement principles. in conjunc· lion wilh LET'~ claim ... regarding gender differences and language acceptabi lity , The application of reinforcement principles allow ... both initial and long· term medical adherence
to be addre"ed. RET', predictions regarding effective strategy usage by male, and females in initial encounters are identical to those advanced by LET. The rationale. however, i ... different. According to RET. strategy effectivcness in a patient's initial encounte r with a phy... ician i~ based on a {u'o-appraisal process in which the patient judges first the appropri· ale"eJS of Ihe illj111eIlCl' messa/{e and then its morinltiOlllll l·allle. Corresponding with LET. if the mc!-.sage is judged inappropriate by the recipient. il will be immedi:Jtcly rejected. Like LET. RET presumes that Illc!o.sage appropriateness in initial encollnter~ i!o. influenced by phY!o.ician ge nde r in such a way that aggressive communication ( i.e .. negative regard strategies) by females is rejectcd ou tri ght and aggressivc co mmuni cat ion by males is tolerated. Although acceptable communication is necessary for an influen ce mcssage to be successful. RET acknowledges that it i... not a sufficien t condition for increasing adherence rates for people who need motivation.
Chapter 16 • COlllplillllC£' Gaillillg ill Medical COlltexl,~'
303
If the message is judged appropriate by the recipient. RET argues that the reci pient will next evaluatc it\., reward pOlel1liu/ or 1I/()lil'(lli01w/l'(l/ue (i.c .. the dcgree to which the message is rewarding or i.!vcr ... ive). Although much of the medical adherence literature has focused on material rewanJs such as tohens or money (see Chesney. 1984), RET focuses on the symbolic rewards in (he form of verbal anti nonverbal approval cues (hat were discussed in the previous section on strategy conceplUali7.ation. Messages that are either aversive (i.e .. negative regartl strategics showing disapproval) or rewarding (i.e .. positive regard . . trategies showing approval) arc stored in working memory called reil!lorcel1lent expectatiol1s and guide the patient's future behavior. Significantly. approval cues have been shown to be more effective than material rcwards in motivating humans (Greenbaum, Turt1(::L Cook, & Melamed, 19(0), and te ... ts of mutual influence models (e.g .. Andersen. 1985: J. Burgoon & Hale, 1(88) have demonstrated that receivers adjust their behavior to avoid unpleasant exchanges or to maintain pleasant ones. Thu\, RET's cen tral claim i\ that human behavior is driven by the need to avoid aversive commun icat ion cncouillers and to maintain rewarding communication encouiller .... In sum. RET propose~ that during initi::11 intluencc altempts, verbal and nonve rbal messages fooignaling either approval or disapproval for the patient and/or the patient's actions can be used by the physician to shape a patient',,> reinforcement expectations. These reinforcement expectations guide the patient's future actions. The :-,trategy chosen by the physician. however. must be viewed by thc patient as socially appropriate in order for the patient to pay attention to the influence attempt. Because of sociological norms and the motivational properties of reinforcement expectation\." male physicians can increase compliance in initial encounters by using either negative or positive regard strategies, whereas female physician ... can increase compliance in initial cncounters by u. . ing only positive regard strategies (sec figure 16.1). RET predictions for initial encounters were tested by Klingle and Burgoon (1995) w•. ing a multiple message design in which numerous diffcrcnt medical episodes were put lO the test using positive, neutral. or negative regard strategies. Thc study supported the claim that message appropriatenes ... in initial encounters is influenced by the gender of the physician: negative regard strategie\., were viewed by patients us more appropriate when (hey were lIsed by male physicians than by female phy . . icians. The study also clearly demonstrated the predicted physician gender by '>trategy type interaction on message persuasiveness. Specifically. in initial encollnter\" patients were lllort! persuaded when the male physician used positive or negative regard strategies than when he used neutral regard ~trategies. With female physicians, on the other haJ1(\. patient", were most persuaded by positive regard strategies and lea ... t persuaded hy negative regard stratcgie~.
Seqllelltiallllj1l1ellce Attempts Patiellt~
who vi.sit health care providers on a regular basi~ have the opportunity to co ntinually ob.serve the reinforcement behavior of the health care provider and formulate more elaborate reinforcement expectations (Klinglc. 1(93). Although the physician's frequent use of rewarding positive regard or aversive negative regard qrategies would seem to establish the strongest patient expectancy for future rewarding or avcrsive comlllunication. the occasional usc of nonrcinforcing exchange\., is needed to develop motivating reinforce-
304
Part IV • COlIIl'x1.\· for Per,nw.\i(m
Physician Message:
Message Evaluation Process:
PhysiCian Uses Positive Regard
Physician Uses Neutral Regard
J.
J.
Viewed as appropriate for male and female physicians; triggers second appraisal.
Viewed as appropriate for male and female physicians; triggers second appraisal.
J.
J.
!Motivated to Keep II
Not Motivating
J.
Physician Uses Negative Regard
"
Viewed as appropriate for male physician; triggers second appraisal.
\, Viewed as inappropnate for female phys,cians; message rejected
J. I Motivated to Eliminate I
I
Reinforcement Expectation Developed Reinforcement Expectation:
(Patients beheve they will be reinforced for complYing by genlng a similar rewarding message or avoiding an aversive one.)
Motivation to Comply:
FIGURE 16.1
l
MOTIVATION TO COMPLY FACILITATED BY PH YSICIAN COMMUNICATION
J MOTIVATION TO COMPLY HINDERED BY PHYSICIAN COMMUNICATION
Reinforcement Expectancy PredicliOiIS for I"itial Physician-Patient Encolllllers
me n! expect<Jlion..,. RET predict~ thut palient ... develop motivating reinfon.:emcnt expcclati o n ... when the physician use ... both rewa rdin g and non-rewarding '\ tral eg i e~. wherca~ nonmotivating reinforcement c'(pectati()n~ are developed whe n the phy~i(:ian use ... on l) one type of \trategy. The theory indicate.., that repetitive comm unication patterns by the physician (i.e .. all positive regard Mrategie.." all negative regard strateg i e~. or a ll neutral regard '\trategies) cause patie nt s to believe that their be ha vior does not and cannot impact the phys i cian'~ communicati on behavior. Thu~. nonmotivating reinforcement expectations dc\'elop following repealed u~e of one \trategy. becau\e patients begin to expect the same type of ~trategy re gard l es~ of th e ir behavioral adjustment. Ph ys icia ns who u..,e nonre petiti ve co mmuni cation patterns (e.g .. the use of posi ti ve regard with negative regard, the use of positive regard with ne utral rega rd , or the use of neutral regard with negative re ga rd ) assist patients in developing motivating re inforcement expectations, or expec tati ons that the ph ys ic ian 's behavior is a respo nse to the patient 's health-re lated acti ons. Greater motivation to comply and hi gher behavioral maintenance are predic ted to occur following motivating reinforcement expectations than following nOll moti vating reinforcement expectation s (Klingie. 1996). Klin gle (1994) found that motivating reinforcement ex pectatio ns following th e physician's uo;e of nonrepetitive influence ~trategies developed even in case.., where the
Charl~r I() • COIII/I/Wlln' (jaillillg ill Medica/ ComeXIJ
phy~ician's
305
use of reinforcement and nonrcinforccmcnt was not actually contingent on the patient's compliant action and where the ordering of the strategies (e.g .. whether the order ended \vith a r~\\:1rding or :1\,cr... lvc ... trategy) "ccmcd to be irrelevant. Specifically. in an initialte ... t of RET. Klingle (199..t) used four "equcntial physician-patient session~ to create three conceptually different nonrepctitive <.;trategy combinations (positive and neutral regard: pm,iti\'e and negative regard: and negati\'c and neutral regard) and three repetitive strateg) cOl11hinations (all po"iti\e: all neutral: and all negative). These epi ...odc" were the ... ame epi ... ode ... 1I',cd to test the cllcctivene ... s of ... trategie ... used in the initial. one-shot encounter~ with the phy"ician that wcre.! discussed earlier. The ordering of the strategy COI11hination ... wa ... varied so that some ended with a reinforcing strategy and other... with a nonreinforcing strategy: none of the strategies was directly tied to the patient's actual compliance. In all case ... the nonrcpetiti\C~ \tratcgy combinations. regardle\s of order. produced morc motivating reinforccment expectation" and greater patient motivation than did the repetiti\c communication pattern .... The fact that phY"iician" were able to motivme pallen .... \\.ithout having to determine whether the p<.Hicnt complied is of critical importance given that phY"iiciam. are not able to Judge accurately the level of patienl compliance (Norell. 19X1: ROlhert. 19X~). When RET \\ a~ originally advanced in 19Y4. the a~sumption wa ... made that effee· tl\C strategy combination ... \voulcJ he Inlluenced by ... ociological language norm .... Thu .... although nonrepctill\e ... tratcgy combll1alJon~ wcrc predicted to produce the most moti\·atIng rt!inforccmcnt expectation". fClllale physician ... were not predicled to be effective If u"ing combination" involying negative regard strategic,,_ RET predictcd that female physician ... would he 1110 ... t persuasive in ongoing exchanges with paticnts if they used combination ... or po"itive.! and ncutral regard ... triltegie .... whcrea ... male physicians would be 1110 ... t persllCl\i\e lI ... ing any of the Ilonrcpct iti ve "itratcgy combinations (Kl ingle. 199..t: Klingle & Burgoon. 1995). An inltialle'l of Ihe Ihcor) (Klingk. 1994: Klingle & Burgoon. 1995) ,howed Ihat. a" predicted by RET. nonrepetitivc strate.!gy combinations motivated patients to comply with physician rccomll1endation~ more than repctitive ... trategy combinations did. Howe\'er. the predicted gender differences did not ... celll to extend to establi"hed relmion"hip .... 'Ince both male and female ph) ... icians were equally successful when using each of the nonrepetitl\c strah:gy combination ... _ The rc"earcllcr.. concluded that established relJtion... hips have the potcntial to o\"crrid~ .. ex-role stercotypes present in initial relation",hip ... (Khnglc. 199..t: Kltnglc & Burgoon. IYY5). This helief coincide~ with other re ...earcher,,· claims that gctting to knmv a pcr"on cause ......ex-role ... tereotypcs to playa less significant role and that some di ... sati ... rying comlllunication cxchanges are to be expected by both ... cxt!" ;:" relational familiarit) incrc~I'e.!" (Crockcr. h ... ke. & Taylor. I 98..t: Deaux & Lewis. 19X4: Roloff. I 'IX?), Although not yet empirically tc ... ted. RET al ...o argues that nonrepctitive communication pallcrns will facilitate behavior per~i ... tence.!. By applying AI1l~ers (1967)[rusfrarirm /lrl}()fliesi.L RET statcs that individuals who arc exposed to occasional nonreinforcemcnt becomc conditioned to expcl.:t nunreinforcemcnt in the future. whereas individuals exposed to continual reinforcement ... oon become fru ... trated ",hen the reinforcement is taken away and quickly re\erllo old hab", (Klinglc. 1993). Klinglc (1996) explained Ihal when paticnt", occ3 ... iollall) recei\e nonrcwarding stimuli lrom their health care prO\'ider. they
306
Part IV •
COIl/n/.\ .fi)r Pen//(nio/l
begin to build up a frustration to lerance for not being rewarded. In reinforccmcnt expc<.:l<.1tion tcrnl''''. patients reali7e that Illedi<.:al adherence will not alway') result in .,ocial prai.,c . A ... a result. when they enter Ihelr natural environment. \\ohere ... ocial reinforcers mu) \\cll be infrequcnl. they conti nul.': to adhere to the mcdical recol11l1lendatiom. rather than becoming fru~tratcd by the lack of reinfo rceme nt Ihey get from their friend ... or family. In sum, RET argues that long-term compliance and behavioral pCP.. i~tence depend on the development of motivating reinforcement expectation .... which arc formed when ;.1 physiciml u.,e ... a comb inat ion of reinforci ng and nonre inforcing comp li ance- ga inin g mes.,aget-.. These ll10livating reinforcement expectations guide the paticnt' ... future actio ns b) in still ing in that patient a ... en ... e of hope that adhe rin g 10 long-term health maintenance wi ll. at ",ollle po int. result in p leasant communicat ion exchange., with the phy ... ician or other., in his or her en\'ironment.
Patiellt Participation: A Compliance-Gaining Strategy or Just Another Compliallce Problem? One comp li an<.:e-gaining technique not ye t addres~ed in thi ... chapte r but widel) advo<.:ated in the ClIITent literature as an cffecti\e tool for enlisting and maintaining patient ad herence is thc use of active patient participation (e.g .. Ka ... sirer. 19H.1: Kim. Klingle. Sha rkey. ParJ.. .
Sm ith . & Cai. 2000: Steward. 1995: Yo un g & Klingle. 1996). In fact. cu rren t research elTorts havc nearly abandoned thc Iradilional line of re ... earch on phy!-.ician comp li anccgaining ... trategic ... that \verl.': abundant from the late 1970 ... to the earl) 199(h and h,we shifted the hurden of compliance to patient ... (Phillips & Jone .... 19(1). arguing that the) IllU ... t share re ... ponsibility for and in the mcdical deci ... ion-making proce ... s (Bc i!-.ecker &
Bei,cd,er. 1990). According to Kaplan (1991 J. activc patient participation build ... a ... en ... e of per:-.onal control over heallh mallcrs and ha ... a po ... iti\e influence 0 11 both p;'Hi e lll com pli ance and health ... WIlI ... . The logic behind the claim that pUlient participation increa ... es compl iance is genera ll y based on Festinger·:-. (1957) Cognilive Disson<.ll1ce Theory. which staled that people desire 10 appea r con ... i... tcn t with \\ hat th ey have :-.aid and done . Decade ... ago. per... uu ... ion research documented that acti\c ly constructing argument ... and communicating one· ... intention., in front of others lead ... to greater beha\'ioral change and per ... i... tence than passive me ...... agc reception (Greenwald & Albert. 1968: Jani ... & King. 1954: Slamed.a & Gmt'. 1978). Compliance. howc\er. is not th e on ly outCOlll1! en ha nced by active participati on in the medical encounter. The health ... talus of the patient a]..,o improve .... because the ve ry ac t of participating c reate ... feeling ... of personal control that boo ... t the immune sy ... tem and ullim and arc Ie ...... likely to give in \\he n the act of compliance re.,ults in negative outcome ... (Wagener &
Taylor. 1986). Patient panicipmion i... indi ... pulah ly an \!\cellcnt tool for facilitaling p;.nient compliance. Unfortunately. the act of patient participation i~ a lot liJ..e the act of medical adherence ihCIf. Specilically. C\'1!11 though patients J..now that acti\e ill\ohement i... in their best
Chapter 16 • Compliance Gail/ing in Medica! COl/tex!s
307
interest and tend to prefer lllorC egalitarian interactions (Bcisccker & Beisecker, 1990: Ende. Kazis, Ash. & Moskowitz. 1989: Kim et al.. 2000: Street & Wiemann. 1987), patient participation rates are alarmingly low (Braddock, Fign, Levinson, Jonsen, & Pearlman, 1997). Thus, it is important to remember that the physician must first of all be able to persuade the patient to comply with the request to participate actively in the deci~ion-making proces~. This act in itself involves addressing barriers to patient participation and using motivational communication style. Several researchers have advocated the use of patient-centered interviewing to increase patient participation (Levenstein, Brown, Weston, Stewart, McCracken. & McWhinney. 1989: Marshall. 1993). The physician's goal during patient-centered interviewing is to elicit patients' feelings. thoughts. and expectations through a more openended interviewing style. as opposed to asking traditional, closed-ended questions. Proponents of this style suggest that it leads to a morc accurate exchange of information, which results in establishing the right diagnosis and treatment and enables the patient to advocate actively for his or her choice of treatment regimens. In addition to changing the physician's interviewing style, the physician may also need to communicate in a manner that dismantles some of the barriers to patient participation. Several recent studies (e.g., Kim et al.. 2000: Young & Klingle. 1996) have looked at barriers to patient participation in the hopes of improving participation rates and have shown that variables such as selfefficacy (i.e .. patient believes he or she has the skills to participate) and response efficacy (i.e .. patient believes participating will result in positive outcomes) can have an effect 011 patient participation rates. Thus, many of the theories previously discussed could be applied to address how the physician should communicate to increase patients' willingness to participate.
Conclusion Physicians al!empting to alter patient health behaviors must construct influence messages and argumcnts that attempt to alter patients ' threat, efficacy, and normative beliefs. whi le at the same time adapting their message style (e.g., using both reinforcing and nonreinforcing messages) to help motivate patients to enact the behaviors that correspond to those beliefs. Recent research efforts in the communication field have focused on patient participation as a viable tool for enlisting compliance and seem to have left physicianimplcmented inlluence messages by the wayside. Although patient participation is a proven motivational tool for cnlisting compliance. physicians need to communicate in a manner that will convince patients La pat1icipatc actively in the exchange of medical information and ideas. Rather than abandoning traditional compliance-gaining approaches, a comprehensive theory should be devcloped that dictates the message content and style of the physicians' influence al!empts. These attempts must effectively encourage patient involvement in the mcdical exchange and motivate patients to adhere to mutually agreed medical regimens. Each of the theories and models reviewed in this chapter serves as a starting point. Following are spccific recommendations for health care providers who wish to increase medical adherence:
308
Pan IV • Cf)llfl'.llS for Persuasion ~hould develop mes~age~ that encourage patient participation by indicating that not participating may result in serious health errors that will directly affect the patient. At the same timc. the health care provider must convey hm\ ea~y it i~ to participate in the interaction (i.e., build self-efficacy) and communicate that patients are expected to participate because participation will help the mcdi<.:al process (i.e .. build response efficacy). 2. Phy\ician intlucnce messages directed at enlisting adherence to mutually agreed recommendations must make the health outcome appear se rious (i.e .. perceived ~ever ity) and personall y relevant to the patient (i.e .. perceived susceptibility ). while at the same time communicati ng lO the patient that he or she can effectively overcome barriers and achieve the benefits (i.e .. efficacy). Unf'0I1unate ly. Illany of the benefib ure not immediate. Thus. the style of the physician'~ message is of cri tical importance. 3. Normative beliefs influence health care decisions and can be used by the physiciull to motivate patients. One method of u~ing normative beliefs would be to adapt me!>'sage conte nt to include arguments suggcMing that loved ones want the patient to engage in appropriate health behaviors. A second method involves considering the health care provider a~ an important referent group on which patients base their deci\ions. Health care provider~ can alter the Mylc of their communication to mothate patients. since pmiems are driven by the need to receive social approval from important referents. including their physician. In initial encounte rs. female health care providers should use verbal and nonverbal mes~ages cOlllmunicali ng thai they approve of the patient and the patient'S action~ by u~ing nonverbal immediacy behavior\ and po~itive regard strategies that involve statements like, ." know you are the type of person who can follow through with the agreed upon regimen" (i.e .. U'ie po~itive regard strategies). Male health ca re providers can use similar messages but a lso have the flexibility to use messages that com municate disapproval for the patient and the patient's inability 10 comply (i.e., use negative regard strategies). Thus, the male physician cou ld use nonimmedime nonverbal behaviors and stern ly say, "You have to see that it would be absolutely irrational not to follow through w ith the agreed regimen." When commu ni cating with patients over time, both male <.Ind female physicians must vary their language sty le and incorporate a variety of regard strategies to help motivate patients to stick to the agreed regime n. Physicians shou ld use nonrepelitive strategy combinations (i.e .. combinations of different regard strategies) rather than repetitive strategy combinations to keep patients motivuted and promote behavioral persi\tence.
1. Health care providers
In sum. compliance is facilitated by communication that correct ly modifies a patient's belief ~ystem and motivates patients to fo ll ow through with actions that are consi'itent with their beliefs. The conten t of the physician's message should include argu ments related to severity. susceptibility. self-efficacy. response efficacy, and normative beliefs. The sty le of the message must involve more than just reinforcing strat egies and Illust involve the patient in the decision-making process. Motivation to comply and behavioral persistence are enhanced through patient participation und throu gh the physician's u-.e of nonrepetitive strategy combinations that inc lude messages of approval and disapproval . A comprehensive approach to medical compliance invol ves a combi nation of care-
Chaptl!r 16 • Compliance Gail/illg iI/ Medical Come.\(\
309
fully constructed arguments by a physician that are delivered in a motivational manner and involve the patient in the decision-making process.
Referellces_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __ Adam .. , S .. Pill. R.. & Jones, A. (1997). Medication. chronic illnc\\. and idcntity : The pcr\pcctivc of people with .... thma. Social Science alltl Ml'c1idlle, 45.189-201 . AJ/en. I. (1985). From deci,ioll'" to ac..:tions: A theory of planned beha\ lOr. In J. Kuhl & J . Bed.mann (Ed ... ). Action-amtro/: From coglliflOtllO hellm'ior (pp. 11 -]9). Ney. York: Sprlngcr. Aj/en. I. ( 1988). Altitlldes. perwmalitv. alld helll/dor. Ilomcwood. I L: Dorl-cy Pre ...... Ajzen. 1.. & Fi .. hbcin. M. (1980). UmlaSIWul;/11-: fluillides lIIId prt'dicll1I8 social beIUII'ior. Englewood Cliff\. NJ: Prcnlice Hall . Ajlen. I.. Tim~o. C .. & White. J. B. (1982). Self-Illonltoring and Ihe ~llIitudc-behavior rdation. JOIl,."a/ oj PenOlw/u-" lind Social PH'cllO/OKY. -12. 426-435. Altman. D. G .. & KlIlg. A. C. (1986). Approachc\ 10 compliance III prinmry prevention . The jourl/a/ oj Comp/;allet'/I/ Hell/lh Cart'. 1.55-73. Am ...e l. A. ( 1967), Partial reinfon:cmcnt effecl~ on vigor and persistcnce. In K. W. Spence & J. T. Spence (Ed ... ). Thl' plyclw/ogy oI/earl/illK and /I/O/il'll/ion : I (pp. 1-65). New York: AC:ldcmic Press. Andcrsen. P. A. ( 1985). NOll\crhal imJllediacy in inlerpcr,onal CIlIlllllunication. In A. W. Siegman & S. Feldstcin (Ed)..). MuJ/ic/uu",t'I ",tl'gmtWf/\ oj mH/\'erhallu,h{II'ior (pp. 1-36). Hillsdale. NJ: Erlbaum. Andcr..on. R. J .. & Kir~. L. M. (1982). Mcthods of IInpro\ing patient compliance in chrom(: di).ea!>e Mate ... Arcll;\'('.1 oj IlIlanal Medici/If'. 1-12. 1673- 1675. Ballard-Reisch. D. S. ( 1990). A model of participative decision making for physici:m ~pallcnt interaction . Health Comlllunicalion. 2. 91-104. Bandura. A. (1977). Social learl/illg theory. Englewood Cliff... NJ: Prcrllice Hall . Becker. M. II. (t974). The heahh belief model and pcr... onal health behavior. Hettl,I! ttluwlioll Mmw graphs. 2, 324-473. Becker. M. H .. & Maiman. L. A. (1975). Sociobchavioral delerminant!>. of compliance wilh health and medical CHre recommend"tiolll,. Medical Cart'. 13. 10-24. Beder. M. II .. & Maiman. L. A. (1980). Stratcgic!> for cnhancing patient compliance. JO//rl/a/ of Comlll//nilY Hea/th. 6.
113~135.
Bel ...ecker. A. E .. & Beisecker. T.D. (1990). Pmient infonnation -seck lllg behaviors when communicating wilh doctor\. Medica/ Care. 28( I). 19~28. Bell . R. R. (198 1). Friend .. hip of women and of mcn . P,\'ychology of Women Quarterly. 5. 408-417. Berg. J. Di .. chl er. 1.. Wagner. D. Raia. J. J.. & Palmer-Shevlin. N. (1993). Medication compliance: A hcalthcilrc problcm. AmHfol PlwrmaCOlht'rtlP.\·. 27. S I ~S24. Berger. C. R. (1985). Social power and interper-.onal communication In M. L. Knapp & G. R. Miller (Ed ... ). Handbook oJimer"t'rwmal commllllicmioll (pp. 439-499). Bcverly Hill .... CA. Sage. Bloom. N .. Cer~oney . K .. & Hart. L. (1980). The relallonship between Ihe heahh belief... model and compliance of pcr ..ons with diabetc\ mellitus. Diabett's Care. 33. 490-500. Bond. G. G .. Aiken. L. S .. & Somerville, S. C. (1992). The health be li ef model and adolc:-,ccnI"l with il1 .. Ulin-dependcnt diabetc\ Illcllilu ... Health Psvclwlog\'. I J. 190-198. Bmter. F. 1.. & Sllff. J. B. (1984). Compliance gaining me ...sage ,election beh,,\ior. Hllman CommllniC'{Ilioll Reflt'arch. /0. 539~556. Braddock. C. H .. 1-lhn. S. 0 .. Levin'lol1. W .. Jon ..en. A. R.. & Pearlman. R. A. (1997). How doctors and patient!'> di"cllss routine clinical decisions: Informcd dcci<.ion making in the ou tpat ient sctting. jOllmal aIGl'nero/lllll'rnal M('{Jicille. /2.339-345. Brown. L. K .. DiClemcnte. R. J.. & Reynolds. L. A. (1991). HI V prcvenllon for ado1c')ccnh: Utility of the health belief model. AIDS till/clIlio" {lnd Pre,·el1tioll. 3. 50-59. Buller. M. K.. & Buller. D. B. (1987). Physician ... · communication ... Iyle and patienl ... ali\faction. JOimw/ of Hellith (II/(/ Sodal Bl'hm'ior, 2M. 375-388.
3J 0
Part IV • Cf}l/teXI:,jor Persuasioll
Burgoon. J. K .. Buller. D. B.. Hale. J. L.. & deTurck. M. A. (l98.t). Relational mer.;~agei. a<;!.ociated with non\erbal tx:ha,iurs. HWlIllll Comlllllll/nllloll R('U(lrch. 10.351-378. Burgoon. J K.. & lIale. J. L. ( 1987). Valid:uinn and mea!.urement of the fundamental theme~ of relational communication. Communication Monof,:mpln. 54. 19~ I Burgoon. J. K. & Halt!. J. L. (1988). Nonverbal expectancy ,iolation .. : Model clabor.stion and applicalion to Immediacy tx:ha\'iors. Communication MOIIO).:raIJhr. 55. 58-79. Burgoon. J. K.. & Lc Poire. B. A. ( 1991. NO .... l:rnbcr). l:.1/ecls of comlllUllicm;OIl e.\pec/(lIIcie.{. actual ('(",,mWI;clllio". alld e xpectallcy di.\·cOII{imwlhm 0/1 el'a/u(uioll of c:ommllll;ClIlOrJ (lfld their commuTI;('(I(i(1I/ lu.JICll'ior. Paper presented .. I the annual meeting or the Speec h Communication A"Mx:iatioll. At lant a. Burgoon. M. (1990). Social influence. In II. Gi le<; & P. Robinson (Eds.). floudbook of /aflf,:IWf,:e lind .w(';{/I'H\'dlOl(J~v (pp. 51 -72 ). London: John Wiley. Burgoon. M. ( 1991. May). Stranger.\·;11 II ,\·trll/lxe lalld: The behlll';oral.lc;l'Ilfist ill the world oJ the medical doc/or. Paper presented at the annun l meeting of the Internat ional Commun icatio n A!.5ocimion Convcnt ion. Burgoon. M .. Blrk. T. S .. & Hall. J . R (1991). Complmncc and sat i~facti()n with phy<;ician-palicnt communication: An expectancy theory IIltcrpretation or gender dirrerences. HllffWlI Commullicatioll Re~e(ln". IX. 177-208. Burgoon. M .. Dillard. J. R.. & Doran. N. (1984). Friendly or unrricndly per<;ua... ion: The effects of violation .. or expectation .. by malc<; and fema le:.. Humall Commullicatioll Research. /0.283-294. Burgoon. M .. Dillard. J. P .. Koper. R.• & Doran. N. (1984). The impact o r commun icati on context and pcr\uadcr gender on pe r~ua\i\c mc:'~'lgc .. eJection. Women's Smd;e,\' ;11 CommUII;m/;oll. 7. 1- 12. Burgoon. M .. & Klinglc. R. S. (1998). Gender differences in being inn ucntia l and/or innue nced: A cha llenge to prior cx.planations. In D. Camlry & K. Dindia (Ed ... ). Sex differellces (flld .\ ;milari,ie:, in commll1li('(/fiOll (pp. 257-285). Hill .. d:.llc. NJ: Erlbaum. Burgoon. M .. Parrott. R.. Burgoon. J. K .. Birk. T .. Pfau. M .. & Coker. R. (1990). Primal') care phY5ician',' <;election or \crbal compliance-gaining "Ir
Che .. ney. M. A. (1984). Behavior modific
222. DiClemente. C. C. ( 1981). Self-efficacy and .. moking ce!.<;ation maintenance: A preliminary repon. Cogni,il't, [hempy Research. 5. 175-187. DiClcmente. C. C .. Carbonari . J. P .. Montgomery. R. P. G .. & I-Iu ghe ... S. O. (1993). The alcohol absti nence \c lr-e fficacy scale. Journal ofSlllclie.\ 011 Alcollol. 55. 141 - 148. DiMatteo. M. R.. Prince. L. M .. & Taranta. A. (1979). Patient ... · perception<; of physicians' beha, lor: Determinanh or patient commitmcnt to the therapeutic relationship. Journal oj Comml/lI;ty Heldth. 4.
28()··290.
Ch
311
Dube. C. E.. O·Donne\. J. F .. & No\ad. D. II. (2000). Communication .. kills for prcventivc intervellliom,. Acat/CIIH' of Medi!'ille. 75 (7). 545 ·54. Dunbar. J. M .. & Angra... W. S. (19KO). Compliance with medical rcgimen ... In J. M. Fergu>;on & B. T;lylor (Ed ... ). Comprehemin' /llllldhooJ.. (II h('hCII'iorallll('t/icille (pp. 115 ~1·tlij). Nc\' York : Spectrum. Dunbar. J.. Dunning. E. J.. & Dwyer. K. (1993). Compliance re ..earch m pediatric and adolc ..ecnt population ... : Two dccade .. of research . In N_ A Kra ... negor. L. Ep .. tein. S. 8. Johnson. & S. J. Yaffe (Ed ... ). D(!I't'/op/1Il'lIfal (Hpecrs of IIealrh ('ompliallce be/ulI'ior (pp. 29~51). Iii Ilsdalc. NJ: Erlbaum. Endc. J .. Kali:... L.. A ... h. A .. & Mmkowill. M. A. (1989). Mca .. uring patient...' desire for autonomy: Deci ~ion making and information ..ceking preference:.. anl()ng rnedic:11 patient .... jourt/al of Gt'nerollllremal Met/icine. -I. 23~30. Epqein. L. II .. & Clu ... s. P. A. (1982). A bchavioralmcdicinc pcr .. pcctive Oil adherence to long-Ienn mcdical regimen ... joumal of COlI.wltil/g lIIId Clil/ical P.Jyc/wlogl". 50. 950-971. Eraker, S. A .. Kirsch!. J . P .. & Becker. M. H. (1984). Under.. tandlng and impro\"ing patient comp liance. AI/I/"/\ of IlIlenw/ Medicine. I{)(). 258 -268. Fco;;linger. L. A . ( 1957). A I"eo/,\, of {'o~"ilil'l' tiimmallce. Stanford. C A: Stanford Unl\er ... lty Prc>;' Fi .. hbcin. M .. & Aj7en. I. (1975>. B{'lit~r. (Iltillltie, illfe"fiol/ ami ht'hm'ior: All illlrodfiCIUII1 /0 Iheol'\" lind reH'arcir. Rcading. MA : Addi,on-We,lc). Fit/patrick, M. A .. & Winke. J. (1979). You alway ... hun thc onl! you love : Strateg.ics and tactic ... 111 interpcn-ona l conllic!. CommfllliCllfioll QUflrtl'f'/\" 27. 3-11. Fletcher. R H. (1989). Patient compliance with therapcutic ad\ ICI!: A modern view. Mo/illi Sillol jOfln/al of Met/icine. 56 (6). 453-158 French. J. R. P .. Jr.. & Ra\en. B. H ( 1959). The ba~e!. of ..ocial Jxw,cr. In D. CamHight (Ed.). Swt/ieJ ill w('/a/ pOll-a (pp. 150-167). Ann Arbor. MI : t.:ni\"er':oll) of \tichigan. Institute of Social Rc-.earch. Friedman. L. S .. Lichtenstein. E .. & Bighill. A. (19M5). SmokIng un':oet among teen~: An empiric .. 1 anal,!.i!. of initial ... ituations. Addinil'e Bellm'ion. 10. I ~Ll Fulmer. T. To, Feldman. P. H .. Kim. T . S .. Cany. B.. Beer>;. M .. Molina. M .. & Putnam. M. (1999). An inlt.:rvcntion .. lUdy in enhanced Illedu.:ation complIance in cmnmunic:uion-dwclling elderly indi vidual .... jOI/rt/al of Gerrmtol()~ical NlIn·ill~. 25. 6-14. Greenbaum. P. E .. Turner. Co, Cook. E. W .. & Melamed. B. G. (IWO). Denli~ts' voi<.:e control: Effc:cts on children' ... di .. ruptive and affectiVe! behavior. Hellllll Plrdw/ogr. 5. 546-558. Greenwald. A . G .. & Alben. R. D. (1968). Acceptance and recall of improvised argument,. journal of £rpt'rimell/(ll Psychology: /lumllll lL(/n/ing ",/(1 /'vIemorr , oj. 592-6()..t. Grodncr, M (1991). U!oing the health belief mollel for bulimia pre\ention. College Healrh. 40. 107- 112. Gro~~. A. M . ( 1987). A behavioral approach to the compliance prohlem!. of young diabetics. The journal ofComplhlllCl';1/ Healrh Carl'. 2. 7· 21. lIan ... on. R. W . (1986). Phy!.ician -paticllt communication and compliance. In K. E. Gerber & A . M. NehcmJ...i .. (Ed .... ). COlllplilllll'e: nu' di/t'IIWIfI of Ihe c/mmicallv ill (pp. I 83~212). New York: Springcr. Harre. R. (1980). Social being. Totowa. NJ: Adam!.. Hariman. P. E .. & Beder. M. H. ( 1978). Noncompliance with prc'iCribed regimen among chronic hemodi aly!.i ... patient,: A melhod of predlcllon and cducational diagno ... i... Dialysis (J1Il1 Trall\plllllllllion. 7.
c.
978-989 Haye~.
J, A. ( 1991). P!oychosocial barrier... to behavior change in preventing human immunodeficiency viru ... Cl-lIY) infection. COllI/w'lill.': P,'ivc/wlo.,:i.H. 19.585....602. Heiby, E. M .. & Carlson. J. G. (1986). The hcalth compliance mO<.le!. journal ofColllpli(lll((J ill Ilea/III Care. I, 1J5- 152. Hen . . on. R. II . ( 1997). Analy .. is of the conccpt of mUlUality . [1I/(IMt!· The journal of N/ming Scho/ar.lhip, 2V,77-81. Hertog. J . K .. Finnegan, J. R .• Rooney. 11.. Vi"wanath. K.. & PUller, J . (1993). Self-effic:lcy a ... a target population <.,egmentation strategy in a diel and cancer ri .. J.. reduction campaign. H{'(i1lh COII/lllullic(lliot/, 5. 21~·tO. Hyman. R. B .. Baker. 5 .. Ephraim. R .. Moadcl, A .. & Phillip. J (1994). Hcahh belief model \:lriable ... a!. prcdiclOr<- of screening mammography lit ilinltion . JOIll'llal (~/lJe"{I\ 'iom! Mt'llil"illl'. 17. 391-106.
3] 2
PJrt IV • COll/ert.\ for Persuasioll Jani .... I. L.. & King. B. T. (1954). The innuenl.:c of role-playing on opinion change. JOllnlal of Ahnormal wul Social Pndwlo!:y. -19. 211-218 Janl. N, K.. & Becker. M. H. (1984). The health belief models: A decade later. Hea"', Eduwtion Qllar· ferly, II. I-H John ... lon. L. D .• O·Malley. p, M .. & Bachman. J. G, (Dec, 2000). Nlltiollal Sun·e., ReSlllH 011 DruK UH' from the Mo"iwriflg fhe FUlllre Swd\', 1975-Z000. Ann Arbor: Ul1lver~ily of Michigan Nev", lind Information Services. A vailable online: www.moniloringthefulUrc.org: accessed 03114/0 I. Jone,. J. A .. & Phillips. G. M. (1988). CtlIllfllllllic(lIillg lI'itll your doctor: Rx for good mediclI! ('(1ft'. Carbondale. IL: Southern 11Iino i... Univer ... ity Pre ....... Kaplan. R. M. ( 1991). Health-re lated quality of hfe 111 patien t deci ... ion maJ...ing. Joum,,! of Social 1\\11('\. -17. 69-90. Kaplan. S. H .. Grcenlicld. S .. & Ware. J E. (1989). A ...... e~ ... jng the effect... of phy ... ician-patient intcraction, on the outcomc!, of ch ronic di~ea ...e. Ml'lliml Care. 27. 110-126. Karoly. P. ( 1993). Enlarging the 'icope of the compliance conc.;truct: Toward developmental and mo tiva tional rclevance. In N. A. Kra'>llcgor. L. Ep~ t cin. S. B. JohiNIn. & S. J. Yarfe (Ed~.). Dl'I'l'iopmc,,tulll.\IUYtJ of II1'11lrll complia"ce be/wI'ior (pp. I 1-27). H ill c.;dalc. NJ: Erlbaum. Ka~h1ll1a. Y .. Galloi .... C. & McCami sh. M. ( 1992). Predicting the u<,e of condoms: Past behavior. nOrln.... and the ...c'(ual partner. In T. Edgar. M . A. Fitzpatrick. & V. S. Frt"i muth (Ed .... ). AIDS: A u)!lIIfIImi· cari(J11 penpeoin' (pp. 21--46). Hill ...dale. NJ: Erlbaum. Kav,irer. J. P. (1983). Adding in'iult to Injury: U~urping patienh' prerogati ... e~. Nell Ellglllmi Jmlnllli oj Medicim'. 308(15). 898-901. Kelman. H. (1961 ). Proce.;,ses of opinion change. Public Opiniol/ QUlIrteriy. 25. 57-78. Kim . M. S .. Kling le. R. S .. Sharkey. W . F .. Park. H. S,. Smith. D. II .• & Cai. D. (2000). A le"'t ora cultural model of paticnt"" motivation for vcrbtll commun ication in patient-doctor int crac ti on~. Commllllicario" Mmwj:raphs. 67. 262~283. Kir ...cht. J .. & Ro\en~tod. I, (1979). Patienl<,' problcm~ in following recommendations of health c .'(pert .... In G. C. Slone. F. Cohen. & N. E. Adler (Ed~.). Nellirh IH" C'h%g\': A ha"dbook. San Fnm c i~c o: J O"''iey-Ba~<,
Klingle. R. S. (199.1). Bringing lime into phY'lcian compliance ga ll1ing re~earch: Toward a reinfor~e lllent expectancy theory of strategy effectivcnc ... 'i. flt,(i/fh Comfllllllicotiol/. 5. 283-308. K lin gle. R. S. ( 1994). Pmiellt colllpiialll'e and mti.VfI('tioll with pin'Jiciml inflllellce alfempfs .. A rt'ill/tJr('C'ment e\peclllIIC\' approach 10 complillllcl'-lertation. Klingle. R. S. (1996). Physician communication a~ a nlotiv.ilional tool for long-term patient complhmcc: Rcinforcement expecta ncy theory , COlllllllllli('(lfioll SllIdie.\·. 47. 206-217. Klingle. R. 5 .. & Burgoon. M. ( 1995). Paticnl com pliance and <:;ati sfaction with phy ~ician influence attempt ... : A rcinforcemcllI expectancy ;.pproach to compliancc ga ining o\cr time . CO/1//1/llIIicmiot/ Re,<;earch.22. 148-187. Klinglc. R. S .. & Miller. M. (1999). /99X Nmmii Hlldem alcohol lUUJ dmg IHe stlld\' (/99/-/99/'1): /-Ia · wmi l/dole\cl'lII rreatmt'lIIlll'ec/.ion~. ScienCt!. 23tS. 29 1-
295. Levenstei n. J. H.. Brown. J . B.. Weston. W. W .. Stcwafl . M .. McCracken. E. C .. & McW hinney. I. ( 1989). P:'lIient-ccnlcred clinical illlcrviewing. In M. Stewart & D. Roter (Eds.). COlllllllmicmill!-: u'i,,, medical patit'IIH (pp.I07-120). Newbury ParJ.... CA: Sage.
Chapter 16 • Complillllce Gaini/IK ill Medical COl/texts
313
Leventh"l, H.. & Cameron, L. ( 1987). Behavioral theoric:-. and the problem of compliance. Patienl £du('(/tiO// Counselillg. 10. 117-138. Levinger. G .. & Hue~mann. L. R. (1980). An "incremcntal exchange" perspective on the pair: Interpersonal reward and level of involvement. In K. J. Gergen. M. S. Greenberg.. & R. J . Willt~ (Eds.). Socilll eJ.ch(lIIge: Adm",:e... ill theory (I/Id rew!tlrd, (pp. 165·188). New York: Plenum. Ley. P. (1988). Cmn/Illmiauillg \\'itll p"tiell/s. London: Crool11 Helm. Linden. M. (1981 ). Definition of compliance. JOIln/al of TherapJ (//ul Toxicology. /9. 86-90. Marlatt. G. A .. & Gordon. J. R. (1980). Determinant ') of relapsc: Implications for the maintenancc of behavior change. In Davidson and Davidson (Ed .... ). 8l.J/CII'ioral medi('ille: Cha/lgillR health lijl'l1y/e,\. New York : Brunner-Maze\. 410-452. Marshall. A. A. (1993). Whose agenua i!' it anyway? Training medical residents in patient-cen tered interviewing techniques. In E. Berlin -Ray (Ed.). Caw SlUdif!.\' ill heelltil C:Olllllllll1icalioll (pp. 15-30). Hill ..dale. NJ: Erlbaum. Mattson. M. (1999). Toward a reconceplUulilation of communication cues to action in the health belief model : HIV test counseling. COlllllwllinllioll MOllographs. 66. 240-265. Mazzuca. S. A. (1982). Does patient education in chronic disea"e have therapeutic value? JOIln/al (!f Chronic Disease, 35. 521-529. Mchrabian. A. (1969). Significance of po~ture and position in the communication of attitude and statu" relationships. Psychological 8"l/etill. 71. 359-312, Meryn. S. ( 1998). Impro\ing doctor-patielll communication: NO( an option. but a nece,sity. Brili\h Medical JOllnw/. 3/6. 1922. Miller. G, R. (1983). On various ways of ski nning ,ymbolic cats: Recent research on pcrsua .. i\e me:o."<:lge strategies. Joumal of umgllage lind Social Psychology. 2. 123-140. Montgomery. S. B.. Joseph. J. G., Becker. M.II .. Ostrow. D. G .. Kes~ler. R. C. & Kirscht. J. P. (1989). The health belief modcl in underManding compliance with preventative recommendations for AIDS: How useful? AIDS Educarioll (//ul Prel't!lIIioll. 1(4).303-323. Norell. S. E. (1980). Medication behavior: A ~tudy of outpatients treated with piocarpine eye drops for primary open angle glaucoma. Acta Oplilalm%gica. 14. 1-3. Norell. S. E. (1981). Accuracy of patient interviews and estimates by clinical staff in dctcnnining medH;ation compliance. Social Sciel/ce Medicine. /5. 57-61. Pendleton. D. (1983). Doctor-patient communication: A review. In D. Pendleton & J. lIa ... ler (Ed:..,). DoclOr-parielll commllnicatioll (pp, 5-56). New York : Academic Press. Perlofr. L. S. ( 1983). Perceptions of vulnerability to victimiLation. )olll'l/al of Socia/ h'-Sill'S. 39. 41-61. Phillips. G. M .. & Jones. J. A. (1991 ). Medical compliancc: Patient or physician responsibility? American Bellal'iora/ Sciellti.\'I. 34(6). 756-767. Petraitis. J.. Flay. 8. R .. & Miller. T. Q. ( 1995), Rev iewing theoric . . of adolescent ,>ubstance use: Organi/ing piece .. in the puzzle. Psw..'lwlogi('(l/ BIII/elill. /17.67-86. Pills. M. (1991). The medical consultation. In M. Pills & K. Phillips (Ed~.). The /JSydw/og" ofheCllth: An introdllction (pp. 49-63). London : Routledge. Po(hhadlcy. A. G .. & Schweikle. E. S. (1970). The cffectivenes:.. of two educational program ... in changing the perfonnancc of oral hygiene by elementary school childre n. JOIl/7/(/1 of Public Heallh Dl'lIlistry.
30. 17-20. Prentice-Dunn. S .. Jone .... J. L.. & Floyd. D. L. (1997). Persua.<.ivc appeal, and the reduction of ..,kin cancer risk: The roles of appeamnce concern. perceived benefits of a tan. and efficacy infonnation. JOllr//al of Applied Social PsychO/OR.\'. 27. 1041 - 1047. Rippetoe. P. A .. & Rogers. R. W. (1987). Effects of components of protection motivation theory Oil adaptive and maladaptive coping with a he'llth threat. )olll'llal of PersO/wlity alld Social P\'y('/wlogv. 52. 59(\...;;04. Rohberl,on. M. R .. & Rogers. R. W. (1988), Beyond fear appeals: Negative and positive per.. uasive appeals to health and self-e..,tecm. )Ol//'I/{/I of Applied Social Psychology. 18, 277-287. Rogers. R. W. (1983). Cognitive and physiological proces:-.cs in fear appeal ... and altitude chunge: A revised theory of protection motivation. In J. Cacioppo and R. Petty (Eds.). Social p\.\'chopllniolo1-:Y· New York: Guilford Press.
3J4
Part IV • Contexts/or Persuasion
Rolnick. S .. & Heather. N. (1982). The application of Bandura':-, sclr~cfficacy theory to ah,tincnceoriented alcoholism treatment. Addiclil'e Belwl'ion. 7. 243- 250. Roloff. M. E. (1987). Communicalion and reciprocity within intimate relationships. In M. E. Roloff. & G. R. Miller (&is.). IllIerpenollal processes: Nell' direclioll.\· ill commlwication research (pp. 11 - 38). Bc\crly lIills. CA: Sage. Roloff. M. E.. & Barnicon. E. (1979). Thc influence of dogmatbm on the situational u~e of pro- and anti~ocial compliance-gaining strategies. Southem Speech Communicatioll Journal. -/5. 37- S-L Rosenstock. I. M. (1974). The heahh belief modd and preventalive health behavior. Health Educatioll MOlloXmph,\·. 2. 354--386. Roter. D. L., lIalL J. A .. Merlsca. R.. Nordstrom. B.. Cretin. D .. & Svardstad. B. (1998). Effectivenes\ of Interventions to improve patient comp li ance: A meta-analysis. Medical Cure. 36. 1138-1161. Rothert. M. L. (1982). Physicians' and patients' judgmcnts of comp liance with a hypcrtensive regimen. Medical De('isioJl Makinx, 2. 179- 195. Schneider. D. E., & Beaubien. R. A. ( 1996). A naturalistic investigation of compliance-gaining stratcgies employed by doctors in medical interviews. Southem Communicatioll journal. 332- 341. Seibold. D. R .. & Roper. R. E. (1979). Psycho~ocial determimulI:-' of health care intention'>: Test of the Triandi, and Fi,hbein models. In D. Nimmo (Ed.). Commullica/ion rem'book 3 (pp. 627- (43). New Brunswick, NJ: International Communication Association. Siamecka. N. 1.. & Graf. P. (1978). The generation effect: Delineation of a phenomenon. Joumal of Experimelllllf Psychology: Humall Leami"~ (lnd Memory. 4, 592--6O-l. SmaJcc. J. L. & Klingle. R. S. (2000). Bulimia intervention via interpersona l in fluence: The role of threat and efficacy in persuading bulimics to '>eek help. journal of Behm'ioral Medicin£'. 23. 37- 57. Spector. S. (2000). Noncompliance with a~thma therapy-Arc there so luti on~? JOllrtla/ oj Aw/mlll, 3. 38 1-
388. Steward. M. A. (1995). Effective phy,ieian-patient COllununiC'ltion and health outcome~: A review. Ccmadial/ Medical AS.fOcjat;oll Journal. 152. 1423- 1433. Stewart, R. B.. & Cluff. L E. (1972). A rcview of medication errors and compliance in ambulant patient!!. Clillical Pharmacology alld Therapeutic.~, 13. 463--468. Street. R. L.. & Wiemann. J. M. (1987). Patient "atisfaction with phy,icians' intcrpersonal involvemcnt. cxpre\sivene,s. and dominance. In M. L. McLaughlin (Ed.). Communication rearhook 7(pp. 59 1612). Beverly Hills. CA: Sagc. Sulton. S. R. (1982). Fear-arousing commtJnicatio n ~: A critic(ll examination of theory and re"earch. In 1. R. Ei,er (Ed.). Social P.\ycholoRY alld Behm'ioral Medicine (pp. 303-337). London: Wiley. Tede\c(J. L. A .. Keffer. M. A., & Fleck-Kandath. C. (1991). Self-efficacy. reasoned aClion. and oral health heha\ ior reports: A social cognitive approach 10 compliance. Joul'llal of Behavioral Medicine. /4.
341 - 353. Thurman. Q. C .. & Franklin, K. M. (1990). AIDS and collegt! health: Knowledge. threat. and prevention at a northeastem university. journal of American Collegl' Health. 38. 179-184. Turk. D. C .. Snlo\·e). P .. & Litt. M. D. ( 1986). Adherence: A cognitive-behavioral perspective. In K. Gerber & A. M. Nchemkis (Eds.). Compliance: The dilemma of Ihe chronically ill. New York: Springer. Wagener. J. 1.. & Taylor. S. E. (1986). What else could I have done'! Patients' responses to failed treatment decision~. Heal/h P.<;ydlO/ogy, 5. 481-496. Webb. P. A. (1980). Effeclivene~s of patient education and psychosocial counse ling in promoting compliance and control among hypertensive patient ... JOllmal of Family Practice. 10. 1047-1055. Weinstein, A. G. (2000). Asthma treatment and noncompliance. Del Medical Journal. 72.209-213. Wein,tein, N. D. (1983). Reducing unreali"t!c optimism about illnes~ susceptibi lity. Health Psychology. 2,
11 - 20. Wein:-.tcin. N. D .. & Lachendro, E. (1982). Egocentrism as a source of unrealistic optimism. Persol/ality ami Social PJ.w·llOlogy HII/letill. 8. 195- 200, Weisman. C. S .. & Teitelbaum. M. A. ( 1989). Womcn and health care co mmunication . Patient Educafion alld CO/II/selil/g. 13, 183-199. Wil ..on. J. T. (1973). Compliance with instructions in the cva luation of therapeutic efficacy: A cOlllmon but frequently unrecognized major variable. Clillical PediarriCJ. /2. 333.
Chapter [6 • Compliallce GaillillK ill Medical Contexts
315
Witte. K. (1992a). Putting the fear back into fear appcal~: The extended parallel process model. Communi· cation MOI/()/{raphs. 59.331 - 349. Wine. K. ( 1992b). The role of threat and efficacy in AIDS prevention. IlIIemationlll Quarterly oj Comml/' 1I;("{Itioll Heal,h Education. 12.225- 2"9. Witte. K. (1994). Fear control and danger control: A te,t of the extended parallel process model (EPPM). CommlmicaT;oll MOllograph~. 61. 113- 13".
Witte. K. (1998). Fear as motivator. fear a ... inhibitor: U.,ing the EPPM to explain fear appeal "ucce~ ...e .. and failures. In P. A. Andersen & L. K. Guerrero (Ed .... ). The handbook of commlillic(tliOIl and emo· tioll (pp. 423-450). New York: Academic Press. Witte. K.. & Allen. M. (2000). A meta-analysis of fear appeals: Implications for effective public health campaigns. HealTh (/lui Educatioll Bellm'ior. 27. 591 - 615. Witte. K., Stokols, D .. llUane. P.. & Schneider, M. (1993). Testing the health belief model in a field study to promote bicycle safety helmcts. C()I1/l1llll1inllioll Research. 20, 564-586. Young. M .. & Klingle. R. S. (1996). Silent partncrs in medical carc: A cross·cultural study of patient participation. Hcalth COII/mUllicoThm. 8. 29- 53.
17 Social Influence in Close Relationships Leslie A. Baxter and Carma L. Bylund
Ke lley and his colleagues ( 1983) have defi ned a re lati onship as close "if the amoun t of mutu a l impaCI two people have o n eac h other is great or, in o ther words, i f th ere is high interde pende ncc" (p. 13). In lig ht of thi s de fi niti o n, all of th e co mmuni cative exchanges between c lose fr ie nds. ro mantic part ners, spo uses, and fam il y me mbe rs could be considered instances of soc ial innuencc or pers uas io n. S uch an a ll -e ncompass ing approach is certai nl y compatib le with Gas; and Seiter's (2003) s uggcstion that any and all communi c~lI ivc events can be exa mined fo r the ir persuasi ve qu alit ies. However, an ex haus tive review of th e research and th eory surro unding all face ts of commun icHl ion in close re lati o nships is beyond O Uf scope-a nd o ur page limi t. Ins tead. we are goi ng to foc us mo re narrowly o n three specific perspect ives o n socia l influe nce between part ies in close re lati o nships. The fi rst perspec ti ve is more or less accepted as the c!a,\'sic, o r traditional, approach in th e socia l sC iCllli fic resea rch literature: A pcrsuader has a goal o f seekin g be hav iora l o r ,Ht itudin al c ha nge in th e part ner (the "target") and strateg icall y deploys a message desig ned to accompl ish the des ired goa l. The focus of attention is o n the persuader's stra teg ic choice of a tacti cal message. This app roach ha ils from many disciplinary t rad iti oll~. in c ludin g cO llllllunication, marketin g, psychology, and chi ld develo pment. It is generall y c harac teri zed by q uant itati ve methods in whi ch research subjects arc asked to construct a :-.in gle strategic message for so me spec ified goa l. o r to choose a strategy amo ng several cho ices prese nted by the researcher. Thi s perspecti ve has dom inated the study of socia l innu ence in c lose re latio nships for over th irt y years. and it has prod uced a large bod y o f resea rc h fi nd ings, By contrast, th e second and third perspec tives di sc ussed in thi s c hapter arc re lat ivel y rece nt arri va ls o n the social innuence sce ne. We present them as points of contrast with the traditiona l perspective in o rder to pose alternative understandings o f soc ia l innue nce beyond the domin ant strategies approac h.
317
318
ParI IV • COl/texts.for Persulision The seco nd perspective, one \\c labcl the social·melll/illg approach, al ... o rOCUSe!; on a per~uader who has a goal of cha nging thc targct in so me way. However. from the socia lmean in g perspective, the meanin g of th e per ... ua ... ive e nte rprise and it-. exc(;ution are negotiated between th e two partie ... to th e relationship. Ce ntral to thi ... meaning- makin g perspective are th e identities of the parties and th e rights and obl igat ions thaI gu ide action in the ir relalion:\hip. From the soc ial- mea ning perspective, th e target is not a passive rec ipient of the persuad er's strat egic message but an active CoCom-MU Clcr of meaning. From thi s second perspective. then, the foc us of attention is on th e co mmunicative exchange between pe rsuader and target. This seco nd perspective. rooted in the traditions of anthropology, e thn ography of com muni catio n and conversation analysis. emp hasi zes the qualitative study of the mi cro-details of naturally occurring interaction between persuader and target. The third per ... pective. dialogic c()If/II//OliclIIioll. sus pends di,tinct persuader-larget roles and focuses instead on those co mllluni cative ep i sode~. ca lled dia logic 1110111el1l~, in whi ch two relationship parties jointly change each o th er wit hout prior inten tion to do so. The focus of attention is on these occasions of dialogic cO lllllluni(;a ti o n, or ge nuine di alogue, between relation ship parties. This thi rd perspective. with root ~ in dialogism th eory, feminist theory, and social cons tru ct ioni sm. features qualitative and interp retive methods of inquiry.
The Traditiollal Perspective The first sec ti o n of this chapter provides an ove rview of litera turc:: regarding soc ial influence s trat egic~ in two types of close relationship~: intimate rela tion '-lh ips (including friendship and romantic re lation ships) and family re lati o nship s hpousa l and pare nt -c hild relationships). Examining inlluc nce ~Iratcgies in re lational co nt ex ts is important beca use "the relationship presc ribes which strat eg ies are proper and usual" (Poppe. va n del' Kloot. & Valke nbe rg. 1999. p. 456). We established two guiding parameters for this ~ec ti o n . First. the litera ture included had to be spec ific to o ne o r more of the types of close relationship, we identified. We present the lite rature separatel y in eac h re lational co ntex t. although '-lo me , Iud ies ha ve examined how persuasion strat egies vary across different types of int ima te and no nintimate relationships (e.g .. Buss. 1992: Cody. Mt.:Laughlin. & Schneider. 1981: Dunn & Cowan. 1993: Poppe et al.. 1999). Second. th e idea of a strategy had to be present in the stud y to be in cl ud ed here . Terms such as "pcr'-luasive strategy," "(;ompliance-gai ning strategy," " influence strategy." "control strat egy," "power strategy," and "manipulation tacti c" are pe rvas ive throughout thi s literature . We ha ve adopted the term persuasive strategy to e ncompass a ll of th ese. We do not. howcver. include the literature on co mpliance res istan ce (e.g., Abc & Ilard. 1999: Ife rt. 2000: Mem. Cupach. & Imahori. 1992: White. Pea rson. & Flint. 1989). In the~c "udies, re~eaJ'chcrs re ly 011 different taxonomie", some developing their own for a pmticular co ntext. A" noted by Kellerman and Cole (199-1-), different classificati o n sys lem ~ ha ve led to di!'.order and confu~ion. It is difficult to make I.,enl.,e of this Iitera-
Chapter 17 • Social h!f7l1ellce ill Close Relariomhips
319
lUre because of thi~ cOl1fu~ion. We do not explain each taxonomy here. I nor do we critique the various taxonolllie~ u"ed in ~tudying relational persuasion strategies: instead. we at· tempt to describe and explain findings using thelll. We have tried to make the names of these persuasive strategics clear and to give definitions where possible. although we direct the reader to the original \tudies for more clarification.
Persuasive Strategies il/ [I/fimate Relatiol/ships Int imate relationo;hips between friend~ and romantic partner\ have been a we ll-studied context for persuasive strategies. These are important contexts to consider. because compliance-gaining effectiveness is related to the intimacy of the persuader and target (Bo>ler. Rodriguez. Cru/. & Marshall. 1995: Cody et al.. 1981: Guerin. 1995: Mallalieu. 1999: Miller. Bo,tcr. Roloff. & Seibold. 1977: Roloff. Janiszewski. McGrath. Burns. & Manrai. 198H). HO\vevcr. 1110st of this research has been limited to college-age students and young adults (see Jones. 1995. for a study on young children's use of persuasive ... tnnegics in frienct\hips). As Kellerman and Cole (1994) pointed out. two guiding foci of compliance·gaining re~earch have been the types of compliance-gaining "Irategies available and the times when those strategies are used. We adopt th~ laller as a consideratio n for this section and add another: the effectivenes ... of per\ua ... ivc :-.trategies.
COllditiolls of Strategy Use.
Strategy usc has been studied in two ways: the frequency of strategy use and the type of strategy employed. Condition ... of strategy use have been opcrationali7ed as characteri~tic ... of the individual persuader or target (including sex. power. culluml background. and per\onality characterio.;tics). characteristics of the relationship. and type of reque ... !. Frequency of stratcgy use has received much less research attention than lype of strategy. In examining frequency of\trategy lise. one study reported that sex alone was not the best predictor of frequcncy of per\uasive strategy use: however, among females. interpersonal orientation (10) predicted frequency of persuasive strategy lise. with high 10 females using per..,uasivc strntegies more often than 10w· 10 females (Cataldi & Reardon. 19%). The linding~ on sex as a predictor of type of strategy use are mixed. In one study, some sex differences were correlated with type of strategy use. but none was consistent across four conditions (e.g .. se lf·report and other report in two conditions each) (Buss, Game .... Higgins. & Lauterbach. 1987). The sex of the persuader was not shown to be a predictor of u\ing indirect or direct ... trategies (Steil & Hillman. 1993) with American, Korean. or Japanese college ... tudents. In heterosexual romantic couples, however, men were marc likely than women to report using bilateral and direct strategies (Fa lbo & Peplau, 1980). Additionally. in the context of ,-;exual inllucnce strategies, males have been found to be more liJ..ely than females to u... e pressure or manipulation as a persuasive strategy (Chri' toph cr & Frand,en. 1990). The relative power of the persuader ha~ also been linked to the type of st rategy used. Those with more power in the relation"hip have reported a greater like lihood of using bi· lateral and direct strategies when compared to those with les~ power (Falbo & PepJau. 19RO). In a ~tudy on married heterosexual couples and cohabiting homosexual couples,
320
Part IV • COflleXIJ!Or Per.HW.lirm
having a position of weakness in the relationship was shown to be ast-.ociatcd with the use of weak persuasive strategies (e.g .. manipulation). while a position of strength in the relationship was associated with the use of strong persuasive strategies (e.g., bullying) (Howard et aI., 1986). In a study of two t-.pecific genres of persuasive strategies. synthetic benevolence (e.g .. flattery) and synthetic malevolence (e.g .. threat), participants reported being more likely to direct synthetic benevolence at more powerful targets and synthetic malevolence at less powerful targcts (Fung. Kipnis. & Rot-.now. 1987). Another way of examining power in intillli1te re1Jtiont-.hips is to look at the number of people the persuader is trying to persuade. In the friendship relationship, assertiveness and exchange strategies have becn shown to be used more in relationships where the power was balanced (OIlC person influencing one friend). whereas coalition and upward appeal strategies were used more in situations in which power was not balanced (one person trying to persuade three people) (Mallaheu & Faure, 1998). Miller (1982) found that the power of the target correlated positively with the probability of the lise of five of Marwell and Schmitt's (1967) strategies: debt, moral appeal. self-feeling. altercasting (positive) and altercasting (negative). We will return to the issue of power when we discuss facework in the next section. Cullural background also appeart-. to playa role in predicting strategy use. Korean and Japanese students report using less confrontational strategies than American students (Steil & Hillman. 1993). Further. Japane:-.e women are more likely to use strong persuasive strategies (e.g .. bullying) with their male friends than are American women, who more frequently use weak strategies (e.g., manipulation) with male friends (Dunn & Cowan, 1993). In this study, no difference was found among female friends. Buss and colleagues (1987) also reported some association between personality characteristics and persuasive strategies. For example. participants who rated high on neuroticism were more likely to usc regression (e.g., pouting or sulking until the partner complies) and the silent treatment than were those who rated low. FUI1hcr, they also reported that the type of persuasive strategy used varied according to whether a target was trying to elicit behavior or terminate a behavior. Specifically. charm was used more frequently to elicit behavior. whereas silent treatment and coercion were used morc frequently to tcrminate behavior. Rescarchert-. have also examined characteristics of the relationship as predictors of strategy use. In looking at those in satisfied versus dissatisfied relationships, Bui, Raven, & Schwarzwald (1994) found that both men and women reponed that men were likely to lise strong or controlling tactics (such as coercion or obligation) in dissatisfying rather than in satisfying relationships. Women reported an equally low likclihood of using strong tactics in either satisfying or dissatisfying relationships. Romantic couples whom interviewers judged to be less similar and less well matched were more likely to usc the silent treatment, debasement. and reason as persuasive strategies than were those couples judged to be more similar and better matched (Buss et al.. 1987).
Effectiveness of Strategy Type.
The effectiveness of persuasive strategies has not received much attention in the literature. although a few studies have examined this topic. One study reported that direct requests are more effective than indirect requests at gaining verbal compliance when combined with high levels of relational intimacy (Jordan &
Chapter 17 • Social Illfluence ill Close Refar;OIuhips
321
Roloff. 1990). Funher, the persuasive strategies of pregiving and direct request have been shown, among friends, not to differ in effectiveness (Boster et aI., 1995). The examination of persuasive appeals regarding condom use has shown effective persuasive strategies to be appeals to caring, pleasure. and responsibility, although in one study three other strategies (heallh, threat. and fear) were effective but less well liked (Sheer, 1995). Additionally, the strategics of pleasure. fear, threat. and heallh were more effective with high- than with low-sensation seekers. Males and high-sensation seekers raled pleasure strategies as most appealing while low-sensation seekers and females rated caring and responsibility as most appealing (Sheer, 1995). Another study demonstrated that males' ratings of nine appeals for using a condom showed a strategy's persuasiveness and the likelihood of lise from that strategy to be consistent. Females, however, did not rate these consistently. Although this study found only little resistance 10 condom use, bOlh males and females preferred strategies that focused on health concerns (Edgar, FreimUlh. Hammond. McDonald, & Fink, 1992). Researchers have also investigated premarital sexual influence strategies. Using the Sexual Influence Tactics Scale, which they created for this research. Christopher and Frandsen (1990) determined the existence of four general influence strategies in this type of contex!: (I) antisocial acts, (2) pressure and manipulation, (3) emotional and physical closeness, and (4) logic and reason. The latter two were the only ones related to sexual behavior: emotional and physical closeness positively. and logic and reason negatively. Although social influence occurs frequently in friendships and romantic relationships, opportunities also abound for social innuence between different family members.
Persuasive Strategies ill Families Family Systems Theory explains that families are both made up of subsystems and nested in higher-order systems (Galvin & Brommel, 2000). The literature has focused on persuasive strategies in two of these subsystems: spouse-spouse and parent-child. We use these two sub!-,ystems as our organizing framework for this section. However, we do not focus on research examining influence strategies used in families with a history of violence (Frieze & McHugh, 1992; Oldershaw, Wallers, & Hall, 1986).
Spousal.
The interdependence of spouses' behaviors and lives makes the marital relationship a rich context in which to examine social influence (Witteman & Fit£patrick. 1986).' Merging Kellerman and Cole's (1994) considerations and our organizational schema in the previous section. we examine the types of persuasive strategies that are used, explore how characteristics of the persuader and of the relationship affect the use of strategies, and determine which strategies are most effective.
Types of Persuasive Strategies.
A basic consideration is what types of persuasive strategies spouses use. One approach to this has been to test developed taxonomies on married couples. For example, marital partners rated 12 of 16 compliance-gaining strategies (Marwell & Schmill. 1967) as more likely to be used when trying to influence a spouse rather than a stranger; threat was the only strategy raled as more likely to be used with a
322
Part IV •
COl/ln/.lfo/" PNI//ll.li/1J/
stranger than i.I -;rouse (Sil1ars. 1980). Additionally. in a ... tudy examining only dual career couples. <.;POll-;L''1 reported u<.;ing more direct ... tfategies than indirect strategies \vhcn using Falbo and Pcplau' . . (10XO) di-;llnction ... betv,.·een the:-.e '1lrateg) lypc-; (Steil & Weltman. 1992). U'ling cultural COlhcnSLIS modeling. and based {Ul an a priori sel ofper,ua ... ive strategic ..... Mexican immigr;..I11l men and Vd.JJllCn \\cJ"e <.;hO\I,.·n to h,l\e a ... hared belief sy ... tel11 about the type ... of per ... ua ... ive strategic<.; hoth <.;cxe'" usc \I.. ith their "'pou<.;es or pmtncn. (Beckman. Harvey. Satre. & Walker. 1999). A second approach invohc'l u'ling indlKt!\C h:::chniqUj....... to create taxonomic ... of pcrsuasi\"c :-.tratcgie...... perifie to the spou ... al context. Re . . earch extll1lining spollsal per'-iui.lsive strategics u<.;ed in the context of joint purrha ... ing deci ... ion ... re ... ulted in a taxonolll) of IX persua-;i\·e strah:gie,. \\ ith bargaining and reason con . . i-;tcntly rcported to be the most fre quently used qralcgic~ (Kirehlcr, 1990, 1993). Ncv. 'lO n <.Ind Burgoon (1990) used previou ... research and literature to create a categorical -;y"'cm for \erbal influence tacti(~ used by ~pou~e.., or partners in rc . . ol\ing di..,agrecment .... rinall). Tucker and Mueller (2000) dc\ eloped a taxonomy of ... trategie . . Lhcd by hu . . band ... and wive . . to modify thcir spou~cs' health beha\ ior:-.: tho-;e mo~t u . . cd ,",crc engaging in health behm ior together. cngaging in facilitati\"e bchm ior. di . . cll ...... ing the health i...... uc \\·ith the ... pou . . e, alld rcque~tillg that the partner engage in the health behavior.
DaTering Uses (~r Pers//(/sil"t' .)·frlllegi(!\. A . . ('cond an.!a III thi.., literature invc~tigale~ differing u... e ... of per. . ua . . i\c ... trategic ... b) indi\ idual characlcri ... tie ... of the . . pou ... e . . and by Chi.lractcri\tic\ of the ... pOu~C\ a-; a marital couple. Many rc ... earchcr ... have exaillined \\ hethe!" men and wOlllcn use different types of ... trategie\. For the J110~t parI. ... e\ doc ... not . . cem to be a ... trong predictor of the type ~tratcgy u ... cd. v. ith a 11..'\\ C\CcptIOll\ (Aida & I. .·albo, IlJl) I: Beckman el a!.. 19l)9: Bu ... :-.. 1992: Dillard & Fitzpatrick. IlJXS; Sagre ... tano. Chri'>lcn ... cn. & Heavey. 199~L Tucker & Mueller. :WOO: Z\·onk(l\ic. Schmicge. & Hall. 19l)..1-: \CC Kirchler. 1993. for a de ... cription of hu\bands' and wive": differellt u... e ... of per... Lw ... ive ... tr~lIcgie ... in purcha ... ing dcci:-.iOlh). A ... an alternative to ... c'<. as thl..' predictor of type of pcr ... ua ... ive ... trategy use. S;Jgre~tallo. Chri:-.ten ...cn. and He,l\ey (199X) dellHHl\trated that il i-; actually the . . ocial role of the spouse in thc cOIl\er\<.Itioll (which spou:-.c i. . sce~ing change) that predicts the type of per... ua:-.ive . . tn.ltcgy u<.;cd. Other individual characteristic ... that ha\e been found to relate to type of ... trategy u\ed Include 10\\ confidence and \\i\(~-;' Ilurtllntnce (Steil & Weltman. 1992) a~ well as wivc,,' masculinity (SI..''<.ton & Perlman. 1981)). In addition. BlI ... :-' ( 1992 ) dClllon ... trated that the Big Fi\"c pcrsonallty f~\(.:tor ... (Surgency. Agreeableness. COJl\cicnliousnes . . , Emotional Stabilit). and Intcl1cct-Opennes:-.) related to choice:-. of \pou~al persuasive :-.trategic .... Finally. 1\\0 ... tudie:-. have dCll1on<.;lratr..'d that husbamb' and wives' frequency of u... ing per-;u,hi\e ... Irategir..' . . arc correlated (Dillard & Fit7patrick. 1985: SexIon & Periman. 19!:l9). Specifically. in looking at pcr ... uu\i\e :-.trategics regarding health behm ior .... wive . . havr..' not been -;hown to u... e the<.;e more freljuciltly tiwil hu ... banJs (although the sample \Va ... relati\{~I) young) (Tucker & Ander.... 20(1). Fitlpatriek· . . three couple type-; have bl..'cn a ... sociatcd with variou ... type:-. of perslIa~ive ~trategies in "'poLl ... al influence (Witteman & Fit/patrick. 19!:l6). FirM. Traditional couples (high on traditional ideology and ... haring ,md hm on conflict avoidance) rely on messages about po ... itive or ncgatlvc outCOI11C<.; of the dcebioJ1. Separate cOllplc'-I (h igh 011
or
Chapter 17 • Sociallllj7I1('11Ct' ill Close Relatio1/ships
323
traditional ideology and contlict avoidance and Iowan sharing) use constri1ll11ng mes'"ges. The third type. Independent couple, (low on traditional ideology and contlict avoidance and high on sharing) use a variety of me~sages, relying all more power bases than the mher couple types do. In purchasing decisions, the choice of persuasive Mrategy is affected by the type of connict. relationship characteristics (marital satisfaction, power pattern,. and duration of the relation,hip). and gender (Kirchler. 1993). Researchers have also been interested in how the employment status of spouses affects relational persuasion. One study found that career orientation (whether just the husband or both spouses work) did not predict the types of ' trategies used and that perceived equity did not correlate with uses of means control or credibility influence strategies by either single- or dual -ca reer couples (Sexton & Perlman. 1989). However. results about which type of career-oriented couple uses more persuasive Mratcgies overall seem to be mixed. Although Sexton and Perlman (1989) found that dual-career spouses use persuasive strategies more often than single-career spouses. Aida and Falbo (199)) found the opposite: equal partners (those who report it is both spouses' duty to provide income for the family) use fewer overall sirategies than traditional partners (those who repon it is the husband's duty 10 provide income for the family). The literature reports a consistently negative relation ~ hip between marital satisfaclion and use of indirect persuasive strategies. One study reported that more satisfied s pouse~ were less likely to use indirect strategies than dissati~lied ~po u ses (Aida & Falbo, 1991). and a second ~tudy reported Ihal dissatisfied spouses more frequently lIsed emotional influence. which the authors deemed to be the most indirec t ~ trategy in their study (Zvonkovic et al.. 1994). EJjectil'elles.'I (~lPersu{lsi\'e Strategies. Only a few ~ tudie!o, have examined which types of persuasive strategies used by ~pouses are morc or less effective, Generally, spouses who gain compliance rely more on messages focused on the activity that is being requested rather than on power or control in the relationship. They also rely on direct statements, direct reque sts. and question, (Witteman & Fitzpatrick, 1986). Additionally, prosocial strategies such as content va lidation (e.g .. agreement. explanation, problem solving) and other-support (strategies confirming or reinforcing the relationship or the other) are effective in gaining a partner' s compliance. although there are some variations by sex (Newton & Burgoon. 1990). Direct requests of wive, by husbands (Dillard & Fitzpatrick, 1985) have also been found to be effective. Different Iypes of emotional di sclosures as part of a persuasive strategy have not been found to influence the objective effectiveness of requests (i.e., whether or not the ..,pous.e complies). However. these types of e motional disclosures do affect the subjective effectiveness of requests (i.c" the spou se's respon se to the compliance request, the spouse's attilude toward his or her spouse, the spouse's attitude toward compliance, and the spouse's attitude IOward se ll) (Shimanoff. 1987). Finally. the effectivencss of persuasive strategies for spouses' control of health behaviors has also been examined (Tucker & Mueller, 2000). Some strategies seem to be more effective than others for both husbands and wives. including e ngaging in the health behaviors with a spouse, having a spouse engage in facilitative behavior. and having a spouse provide emotional support.
324
ParI IV • COII/nf,\.for Persuasion Parelll~Child.
The inherent imbalance of control between parent and child make!; thil., familial ~ub~)~tem interesting to examine. though it is significantly different from ~p()lI ... al influence. The development of the chi ld through the stage ... of infancy. todd ler. sc hoo l age. and adolesccnce brings different con,;"idcrmiono; to bear in cX
l're.\c!tool. The majority of thc literuture on the persuasive qrategie~ parent,;" use with preschool children is based on observing or recording playtime interact ions bctwee n a parent and child. Rcsearch indicates thai in play situations. parents use a majority of directives (telling the child what to do) rathcr than prohibitory statements (te llin g the child what not to do) in trying to get thcir chi ldren to comply. Parents also tend to usc more action control ... (trying to influence the child's immediate physical behavior) than attention wnl
attentioll-perl.,lIasive
s(rategie~
than with
actioll-per~lIa~ive
strategies (McLaughlin. 1983)
and with strategically timed per~uasi\'c strategies that attended to the child'~ current toy involvement (Schaeffer & Crook. 1980). Although the compliance of the child was not dependent on the se>.. of the parent or the sex of the child. the effectivene..,s of strategie~ did vary b) the child's age (McLaughlin. 1983). In a crol.,s-cultllral study. American mothers had more sliccess in getting their children to comply in picking up tOY1- during pl:.lY than Japancse mothers did. when both sets of mothers gave direct command .... Perhaps thi ... is because America n mOl hers are more lik ely 10 usc their authority and give direc t comm;'lI1d .... so America n chi ld ren arc more accu~ tomed to this (Ahe & Izard. 1999). However. during mother-child playtime. directiveness stra tegies did not predict compliance in either the American or Japan ese mother-child dyads (Abe & liard. 1999).
Chapter 17 • Socitlllllj1l1el1ce ill Close Relariol1ships
325
Ado/esceJIIs. The study of persuasion strategies used by parents on their adolescent children and vice versa relies on self-report data, a notable methodological difference from the studies with preschool children. Early research on parental persuasive strategies used with adolescents conceptualized these strategies into three categories: coercion. induction (obtaining voluntary compliance), and love withdrawal (deTurck & Miller, 1983). Subsequent research has shown more variance in the persuasive strategies that adolescents perceived parents using. Perceptions of strategy have been shown to be associated with adolescent gender, age, parental gender, and context (deTurck & Miller. 1983). while adolescents' reported strategy use with mothers has been associated with adolescent gender (Cowan & Avants. 1988). There is also some evidence that adolescents' use of persuasive strategies that anticipate noncompliance correlate with their mothers' use of these strategies (Cowan & Avants. 1988). Effective strategies for gaining adolescent compliance include. for mothers, frequent maternal praise and moderate levels of attempted control: for fathers, they include moderate or high levels of attempted control. For both parents. the use of command techniques positively correlates with adolescent compliance (Smith. 1983). One body of literature examined the consumer socialilation of children (see John, 1999, for a review). including adolescents' lise of persuasive strategies on their parents during consumer decision making. Unlike younger children (aged 3 to II). who tend just {Q ask for products (Isler, Popper, & Ward. 1987). adolescents use a more varied range of persuasive strategies. A typology of adolescent persuasive strategies used to influence consumer decisions and parental response strategies was recently developed by Palan and Wilkes (1997), This typology includes seven types of strategies used by parents and/or adolescents. Reasoning. a substrategy of the bargaining type of strategy. was the most effective adolescent persuu!->ive strategy named by adolescents. mothers, and fathers. The traditional perspective has produced a sizable body of research literature for friendship. romantic. spousal. and parent-child relationships. Strategy use. particularly the type of strategy employed by the persuader, has recei\-ed more research attention than strategy effectiveness. Efforts to identify patterns in findings across swdies are hampered by the lack of a common strategy taxonomy.
The Social-Meaning Perspective Two primary social-meaning approaches can be identified in existing literature: the illcrel1lenta/~i11lerac(ive approach articulated by Sanders and Fitch (200 I) and the face~ work approach grounded in theorizing by Goffman (1967) and by Brown and Levinson (1987). Although these two approaches are different in ways discussed below, they share a common assumption: that all interaction, including social influence, is a socia/undertaking, by which we mean that the parties' actions are guided by rules of appropriate conduct and hold implications for their identities and the nature of their relationship. Sanders and Fitch (200 I) have argued that social influence is of necessity an incre~ mental and interactive activity in which the parties jointly negotiate the social meaning of solicitation (e.g .. requesting, compliance seeking) and response (target compliance or
326
Pan IV • C()flleXI.~ for Persllasiofl
resistance). Per. . uader and target engage in a subtle and complex interactional "dance" in which they jointly determine whether a persuasive attempt is necessary and allowable, how solicitations are interpreted as influence attempts rather than alternative things we can do with words, whether and how the persuader's inducements to change are appropriate and sufficient. how responses are interpreted as acts of compliance or resistance, the appropriateness of the target's responses, and the implications of the solicitation and response for the parties' identities and their relationship. Some research from the traditional perspective has examined sequential influence attempts by the persuader when the target initially resists compliance (e,g,. deTurek, 1985; Lim, 1990b; Wilson, Cruz, Marshall. & Rao. 1993). However, several crucial differences can be identified between this body of work and the incremental-interactive approach. By way of discussing these differences, we'll work from a concrete example that fortuitously occurred between one of us and her six-year-old daughter as this chapter was being written. The conversation was prompted by the daughter's placement of her chair about six inches in front of the television screen, blocking others' view of the image: Mother: "That suunds like an interesting television program you're watching. What is it?" 2
Daughter; "1 dun no,"
3
Mother: 'Td like to \\'atch it too."
4
Daughter: "With me? OK I"
5
Mother: "It would be a lot easier for me to see the TV if your chair wasn't blocking the screen."
6
Daughter: "Let's both move to the couch to watch."
7
Mother: "Please move your chair back away from the screen so we can both see the program from the couch. Remember, nobody can see through the back of a chair. It's important to be "I Care" and think of others when you act."
8
Daughter: "OK," [chair is moved backJ
This conversation features a particular kind of speech act known as a directive (Searle, I 976}- an utterance attempting to get someone to do something they otherwise would not do. The mother initially expresses the directive indirectly through hinting, first in utterance 1, then again in utterance 3, and yet again in utterance 5. In utterance 7, the directive has been expressed explicitly. along with reasons for the request. The traditional research on sequential influence attempts presupposes thaI successive attempts are necessary because of target resistance; that is, the persuader-target exchange is framed as an adversarial one. However. our illustrative conversation does not appear to be characterized by target resistance. Although close-relationship parties can be engaged in an adversarial battle in which the persuader seeks to overcome target resis-
Chapter 17 • Socia/I"fluence ifl Close Relationships
327
tance. particularly in time, of connict. Sanders and Fitch (2001) have argued that everyday soc ial innuence in close relationships is far more likely to be characterized by a spirit of cooperation and agreeableness in which both parties arc motivated to accommodate the needs of each other. A similar point has been made by Roloff and his colleagues (1988). [n our example, the daughtcr does not resist the mother's request; she simply fails to understano the indirect directives of utterances I and 3, hearing them instead as an invitation to do something together. Although she does not take the hinted action of utterance 5, she responds in ulteran ce 6 with a cooperative and agreeable solution to her mother's viewing difficulty-mov ing to a different location-a solution that also advances the child's social goal of doing so mething together. The mothers explicit directive of utterance 7 has acco mmodated to the child's desire to sit together on the couch. The child's utterance 8 acco mmodatel-. to the rcquel-.t to move the chair. Whe reas the traditional approach views sequ ential attempts as necessary only under co nditions of target rc si!o.ta nce. the incremental -interactive approach presumes that all social influence is of necessity an unfolding dynamic that is enacted across multiple utterunces between the persuader and the target. Sanders and Fitch (200 I) have advanced two theoretical real-.o ns why soc ial influence is an incremental and interactive process. The first reaso n, developed by Sanders (1987,1997), is that a given utterance is under~tandable only as pan of the larger interaction stream of preceding and subsequent utterances. Put s imply, utteran ces do not derive their meaning in isolation. Interacting parties are busily engaged in the business of nego tiating what utterances mean in, and through. th ei r !o. uccessive utterances. For example, when the mother's utterances I and 3 are linked together, they can be heard as a desire to see as well as hear the television. However. it i ~ clear from the daughter's utterance 4 that the child hears this utterance st ream differe ntly-as an invitation to a joint viewing of television . In short, meaning emerges out of the "dance" between utterances. Whether and how an influence attempt is hea rd unfold ~ ~I CroSS the interaction. The meaning of the target's response similarly un· fo lds as persuader and target interact. Thus, the daughter' s utterance 6 can be heard as a mcre continuation of her agenda to achieve a joint viewing event, but it can also be heard as an attempt to respond cooperatively to her mother's viewing plight, as expressed in utterance 5. The seco nd theoreti cal reason social influence must of necessity be incremental and interactive is grounded in Fitch's (1994. 1998. in press) work in the cultural premises of persuasion . "Culture" refers to the premises about personhood, relationships, and communi cation that arc shared by a community. whether that community is a society, a social network. a family. or a couple. Relationship parties are embedded in larger cultural systems and si multaneously construct their own unique culture of two, or dyadic culture (B
328
Pan IV • COnfeXl,
enacted. The range of actions and beliefo;; about which people may be persuaded con~titutc whal Filch (in press) calls persuadable". Each culiUTe has a bandwidlh of acceplable persuadables. bounded on one end of the continuul11 with actions and beliefs that are so ingrained into the social world that they require no persuasion. and on the other with actions and beliefs that are persuasion "taboos:' Through their interaction, pers.uader and target negotiate whether the solicitation is legitimi7ed as a persuadable. The righh and obligations of the parties that characteriL.e their relationship are obviously important in defining whether a given solicitation b a persuadable. In the context of our 1110therdaughter example. it is acceptable in the societal culture and in this dyad's culture for the parental figure to make this specitic request. However, we can well imagine several other kinds of encounters in which the child might challenge the parent's right to inlluence. particularly as the child reaches adolescence-for example. clothing preferences, friend~hip choices. and so on. In any social-inllucnce interaction, the persuader and the target negotiate whether the request will be understood as a persuadable. Cultural premises also provide a backdrop of symbolic resources available to persuader and target as they negotiate a social-innuence encounter. For example, in utterance 8 above. the mother invokes the symbolic resource of "1 Care." a program at the daughter's school in which children are encouraged to be thoughtful of others. Through their interaction. persuader and target jointly detennine which of these symbolic resources will hold currency. This is not necessarily an ea,y task, as competing cultural prcmiscf.. can be introduced. In our mother-daughter exchange. for example. the child could have introduced individualism and her right to "do her own thing" but chose not to do so. Last. cultural premises guide the pragmatic enactment of a social influcnce attempt- for example. rules of turn taking. listening to the other's position, and so forth. In our example. the mother and daughter operate on the prcmise that it is important to give reasons to accompany requests rather than simply ordering commands based on authority. Fitch (in press) argues that face work is one important aspect that guides the pragmatic enactment of a social-inlluence encounter. The facework approach presupposes that social-inlluellce attempts are socially meaningful in that the parties' images or identities are at stake-what Goffmall (1967) referred to a~ "face:' Brown and Levinson', (1987) Politeness Theory has been the most influential theoretical framework in understanding the implications of a persuader'f.. re quest on the target's face. although several of its claims have been 1110dified by two decades of research aClivilY. Much of Ihis research has employed quamilalive melhods in Ihe strategies tradition. although Brown and Levinson grounded their theory in the microdetails of naturally occurring talk between interactants. Brown and Levinson (1987) identified two kinds of face. Negative/ace is the desire to maintain one's own autonomy, that is. not having one's privacy invaded. one's resources spent, and one's actions constrained. Positi\'e face is the desire to have one's attributes and actions approved of by significant others. Subsequent research has subdivided positive face into two components: fellowship face. the desire to be liked and included by others; and competence face. the desire to have one's abilities and actions respected and valued (Lim, I 990a; Lim & Bowers. 1991). According La Brown and Levinson (1987). directives intrinsically threaten a target's negative face because an attempt is being made to constrain his or her actions in some
Chapter 17 • Sncia/IIlj7l1ellC(' ill C/me Re/atlOllSltip5
329
way. However, not all directives are equally face-threatening , Brown and Levin ... on argued that the amount of face threat created by a directive i... a function of three variables: relational distance (the more distant the partie!o., the greater the face threat 10 the target): power (the greater the power of the target relative to the per... uader. the greater the face threat); and the culturally defined ranking of how much imposition i ... implicated in the directive (e.g .. a requc\t to borrow $1 00 i~ widely accepted 3\ a greater imposition than a request to borrow $5). The reque" by the mother in our example would thus be regarded by Brown and Levin~on a\ relatively low in it\ face threat to the daughter: Their relation~hip i~ close, the child hl!\ le ... s power than the mother. and a request to move a chair is not widely regarded as involving much imposition. Brown and Levinson (1987) articulated a hierarchy of five slrategie~ available {Q persuader\ who face the prospect of threatening the target'\ face. The lowest-level strategy i... to make the reque~( directly without any effort~ to ma ...... age (he target's face through politenes\ c!Torts. Second. they could express the reque ... t directly with redressive actions intended to enhance the target'i\ positive face- for example. assurances that the target is liked and appreciated. Third. they could express the reque,t directly with redre"ive actions intended 10 enhance the target' ... negative face- for example, an apology for making an imposition. Fourth. they could decide to exprc ... s the request indirectly rather than directly. affording the target freedom to hear the utterance as a directive or not. Fifth, at the highest level of the hierarchy. they could decide not to . . eek change in the target. backing off from the face-threatening directive. According to Brown and Levinson. persuaders decide to use higher-level strategies in proportion to the amount of face threat implicated in the directive. Two decade ... of politeness research suggest six primary moditications or extension~ to Brown and Levinson's (1987) Politeness Theory. First. a number of ...cholars have argued that face threat is not a unidimen ... ional phenomenon: instead. it i., important to take into account qualitative differences in the type of face that i.., threatened (Craig, Tracy, Spi,a~. 1986: Leighty & Applegate. 199 I: Lim & Bower<. 1991). Fellowship. competence. and autonomy face need ... are qualitatively diMine!. and a given directive could threaten these ~ind, of face to different degrees. Second. and relatedly. inOuence acts are not all alike with respect to their face-threat implications. For example, Wilson and his colleagues (Wilson, Aleman. & Leatham. 1998: Wil,,,n & Kunkel. 2000) have compared three different kinds of influence attempts-giving advice. requesting a favor. and enforcing an obligation-and have found that po~itive face threat and negative face threat vary acro~\ these types. In giving advice. for in\tancc. the persuader may be implying a threat to the target's positive face. perhaps doubting. however indirectly. the target's competence to act. Such a threat is not implicated in the influence attempt of requesting a favor. Third. in light of differences in types of face threat. the hierarchy of politene" strategies i.!l problematic in plncing negative-face redress at a higher level than positive-face redrcs!o. (Melts, 2000). In~tead. positive-face redre~~ is viewed as respon<;ive to positive face threats and negative-face redress is viewed as re~pon\ive to negative face threats, with no higher-lower hierarchical placement of the ... e two politeness strategies. Founh. the absolute level of face imposition captured in the ranking variable has not appeared to be a very good predictor of the overall amount of face threat implicated in an
330
Part IV • Contexts/or Persllmioll act; instead, the persuader's right to make the reque~t appear.l. to be a more important faclOr (Baxter, 1984; Craig et al.. 1986; Lilll & Bowers. 1991). The right to makc a reque>! of another is consistent with the research finding that the more dominant the persuader. thc less he or she redresses the face needs of the target (Baxter. 1984: Dillard. Wilson. Tusing. & Kinney, 1997; Leighty & Applegate, 1991; Lim & Bowers, 1991). Fifth, Brown and Levinson's (1987) claim with re\pect to relational distance nced.l. to be Illodified (Baxter. 1984: Lim. 1990,,: Lim & Bowers. 1991: Leighty & Applegate. 1991: Roloff & Janiszewski. 1989: Roloff et al.. 1988). When the directive involves a modest imposition on the target. the Brown and Levin:-.on prediction appears to hold: Intimacy correlates negatively with face-redre ... sive action. However. when the imposition is great, the opposite pattern appears to hold: Intimacy corre lates positively with faceredressive action. In close relationships. the parties recogni7e that their interdependence and commitment will naturally compromi"ie their autonomy needs: tim .... modest impo ... itions are expected and perhaps even function as an index of intimacy. Hml,.'ever. when an imposition exceeds what would normally be expectcd in a cOlllmittcd relation:-.hip. thc face threat is substantial and requires face-redressive action by the per.l.uadcr. Sixth. and last, indirectness is not nccessarily rcgarded a~ morc polite. or face redressive, than directness. Although :-.omc cultures view indirectne\:-. a\ more politc than directness, indirect hinting is regarded as rude in other cultures (e.g .. Fitch & Sande". 1994). Further, in our closest relationships. whcre opcnl1c.l.s and honcsty are ideali/cd. indirectness can be interpreted as a form of face threat nuhcr than face redress (e.g., Dillard et aI., 1997). Despite their differences, the incremental-interactive and faccwork approaches share the view that social influence in close rclatiol1.1.hip\ is a social-meaning-making enterprise in which parties' rights. obligation.l.. and identitics feature prominently.
The Dialogic Communication Perspective It is in our closest relationships that we O1ost change and grow as individuals- VI. hat Aron and Aron (2000) refer to as "elf-expansion. In fact, Baxter and Montgomery ( 1996) have argued that we regard our relationships as close to thc cxtcnt that they expand the parties' selves. The delicate interplay bctween similarities and differences betwcen the parties provides the scaffolding for such changes in self (Aron & Aron. 2000: Baxter & Montgomery, 1996; Baxter & West, 2002: Wood. Dendy. Dordek. Germany. & Varallo. 1994). Similarities between partics arc obvioLl.l.ly important in providing the common ground upon which bonds of intimacy can be built and sustaincd. But equally nccessary arc differences. As the dialogic theorist. Mikhail Bakhtin. ob.l.crvcd over seventy years ago. "What do I gain by having the other fuse with me'! He will J..now and see but \\ hat I know and see .... Let him rather stay on the outside v::lI1tage point. ~1I1d he can thll~ enrich essentially the event of my life" (Bakhti n, as quoted in Todorov. 1984. p. 108). But what is the communicative process of influence by which partics change or expand? The first two perspectives would answer this ql1c ... tion by pointing to change in the target as an outcome of intentional. goal-dirccted activity by thc pcr\uader. However, from the third perspective, the persuader-target distinction frame.l. thc parties' relationship
Chapter 17 • Social Iflj1I1('ffC(' ill Clow Relatiof/.\/lfp_(
331
as one of power and dominance in which the per<"tuader functions to devalue the pcr ... pective of the other person. In!..tcad. the third perspecti ve di ..... olve!"t the per!..uadcr-targct di stinction and eXclll1ines the change that can spontaneou ... ly result when relational parties engage in genuine dialogue who ... e purpo..,e i.., underqanding rather than influence. Such change is "a re,uit of new understanding and in .. ight"l gained in the exchange of idea ..... a:-. the parties "allow diverse po..,ition .. to be compared in a proce ...... of di,covery and qu e ... tioning that may lead to transrormation ror themselves" (Foss & Griffin. 1995. p. 6). The proces!.. by which this happen!.. goes by variou ... labels reflecting the different intellectual roOb of the perspective. including "dialogue" with root~ in dialogic theory (e.g .. BaJ...htin. 1981. 1984; Bohm. 1999; Buber. 1958; for u more complete bibliography. see Ci"",. & Anderson. 1998). "relational respon ... ibility" with roots in !oIocial l:onstrlll:tioni ... m and fCI11ini~t theory (e.g .. McNamee & Gergen. 1999: Shepherd. 1992 ). and " in vi tational rhetoric" "ith rooh in fernini" theory (e.g .• Fo" & Griffin. 1995 ). CiS!oIn3 and Anderson (1994. pp. 13-15) have provided a comprehen..,i\c dbcll..,sion of the characteri ... tics that are present in the dialogic communication envbioned by the third per~pective . The partie~ do not focus on strategic goal ... or outcome ... but rather participate in a !-.pol1tancous. unrehear:-.ed exchange char::u.:tcri/ed by improvi ... alion and creativity. They refuse 10 assume that they already know one another's thought ... and feelings. instead di~playillg a willingness to rccognile their "strange ()thcrne~\." The partil!!.. c.:ngage one another in a spirit of authenticity and ge nuinellc.: .... : the parties are \pcaJ...ing from a base of hOllc~ty. not sLrategy. They ha ve a collaborati\e orientation in \\'hich th ey ... hare their pers pel:tive!'>- with pn~sion and perhap~ even heated argument-not with a goa l of "winning over" the other but in a !'>pirit of sharing their views. Foss and Griffin (1995) refer to thi ... a ... the di~course of offering. " the gi\ ing of expre~~ion to a per!'>pcl:ti\'c without advocating it~ support or sceking its acceptance" (p. 7). Parties ... ustai n a vulnerability to one another's views. an openness to be changed. Dialogic encounters are unique. fleeting moments that punctuate the everyday. mundane. ta ~k-o riented exchanges o f relationship partie .... But they do ex i... t in our friend .. hips. romantic relation :-.hips. and families und are vi\ idl y rem embered by participa nts us emotionally intcn,e and deeply meanin gful in a variety of ways (Baxter & DeGooyer. 200 I). Dialogic encounters are those occa ... ion ... in which the interaction between the parties has created the opportunity for their selves to become. The panie~ are deeply influenced by the dialogue between them.
COllclusioll In a chapter of thi ... length. we can do no more than provide H crude map of the fore~t of social influence in close relationship!,> and examine a few specific trees along the way. We have identified three per~pectives. which conceive of the influence proce\s in radically different ways. Table 17.1 summarilcs the key fealures of these perspective~ as we have discussed them . Rather than viewing these three perspectives as either-or options by which to understand social influencc between close relation ship partner.... \\c prefer to view them as complementary. Tnken together, they afford u ... a more complete view of ~ocial influence than any single perspective can provide alone.
332
Part IV • COlllextsJor Persllasion
TABLE 17.1 Summary o/Three Perspectives on Sociallnfluellce ill Close Relatiollships
Perspective Traditional
Social-Meaning
Dialogic
distinct persuader and target roles for the relational parties
yes
yes
no
intentional goal of seeking change in the other
yes
yes
no
communication of interest
the persuader's strategic choice of a persuasive message
the exchange between persuader and target
dialogic moments of mutuality between the parties
primary focus
target compliancega in ing
the social meaning of the influence attempt for the parties' identities
the expansion of the parties' selves
role of persuader
proactive
proactive
both parties proact ive
role of target
reactive
proactive
coparticipants
Feature
Notes ___________________________________________________________ 1. In the section on persuasion strategies in the family, in cases in which a taxonomy was created specific to a certain relational colllexL we do briefly explain it. 2. In two of the studies summarized here, both unmarried and married coup les were examined as one sample (Beckman et aI., 1999: Newton & Burgoon, 1990).
References_______________________________________________________ Abc, J. A. A., & Izard. C. E. (1999). Compliance, noncompliance strategies, and the correlates of compliance in 5'year-old Japanese and American children. Social Det'eiopmellf, R, 1-20. Aida. Y.. & Falbo, T. ( 1991). Relationships between marital sati sfaction, resources, and power strategies. Sex Roles. 24,43- 56. Aron, A., & Aron, E. (2000). Self-expansion motivation and including other in the self. In W. Ickes & S. Duck (Eds.). The social psychology of persollal relariollships (pp. 109- 128). New York: John Wiley. Bakhtin, M. M. (198 1). The dialogic imagination: FOUl" e5.HlY.\· by M. M. Sak/l/in (M. Holquisl. Ed.: C. Emerson & M. Holqui sl. Trans.). Austin, TX: University of Texas Pre~s.
Chapter 17 • Sociallllj1uellcl! ill Close Relatio/lships
333
Bakhtin. M. M. (198-1.). Pmblell1.~ of Oo.\·WC\·st... y·.\· pf/ctiCl (c. Emcr"on. cd. and tram•. ). Minneapoli:.: Univer:.it), of Minnesota Pre"s. (Original work publi:-.hed 1929.) Baxtcr. L. A. (1984). A n invest igation of com pi iance-gai ni ng a:-. pol itellcss. HUll/ali COII/II1/lIIic(/fioll Re.\wlrch. 10. -1.27--456. Baxter. L. A. ( 1987). Symbols of relation:-.hip identity in relation~hip cultures. JOllrnal of Social and Pc/"mllal Relatio/lships. -I. 261-2~W. Baxter. L. A.. & DcGooyer. D. H. (2001). Perceived ae ... thetic characteristic ... of interpersonal cOllver~a tions. SOl/them COII/IIII/Ilica/ioll .tournaI. 67, 1-18. Baxter. L. A., & Montgomery. B. M. (1996). RelatillR: DialoRlles & dialeclin. New York: Guilford Press. Baxter. L. A .. & West. L. (2002). Couple percepliol/.\· of lheir .~illljlaritie.\· 1I11d differn/ce.\".· i\ dialectical perspeclil·e. Manuscript under review. Beckman, L. J., Harvey. S. M., Satre. S. J .. & Walker. M. A. (1999). Cultural belief... about social influence strategies of Mexican illlmigrant womcn and their heterosexual partner.... Se.1 Roles, -10. 871 892. Bohm. D. (1996). Oil dialoglle (L. Nichol. Ed.). New York: Routledge. Boster. F. L Rodriguez. J. 1.. Cruz, M. G., & Marshall. L. (1995). The relative effectivcnc:-.<;, of a direct relluest Illc~sage and a pregiving message on friends and stmngcr~. Commullication Research. 22.
475-484. Brown. P., & Levin~on. s. C. (1987). Polirelle.~.\·: Some ullil·er.HlI.\· ill lallgl/(lgl' IIsage. Camhridge: Cambridge University Pr~ss. Buber, M. (1958). I alld thou (R. G. Smith, Trans.). New York: Scribner·s. Bui, K. T., Raven. B. H .. & Schwarlwald, J. (1994). Influence ~Irategics in dating relation~hip~: The effects of relationship satisfaction, gender. and perspccti\'e. Journal ofSociaf Behm'ior (llId Per.l"Ol/ality. 9. 429--442. Buss, D. M. (1992). Manipulation in close relationships: Five personality factor ... in interactional context. Journal of Personality, 60.477--499. Bus~. D. M .. Gomes, M .. '·Iiggins. D. S .. & Lauterbach. K. (1987). Tactics of manipulation. JUlln/(/1 of Persollality (lnd Social P.\yclwlogy. 52. 1219- 1229. Cataldi, A. E., & Reardon. R. (1996). Gender, interpersonal orientation. and manipulation tactic u... e in close relation:-.hip .... Sex Roles, 35. 205-218. Christopher. F. 5 .. & Frandsen, M. M. (1990). Strategies of inlluenl.:e in :-.ex and dating. journal afSocial alld Per.Wllol RelariOlHhips. 7, 89-105. Cis~na, K. N., & Andersoll. R. (1994). Communication and the ground ofdialoguc.ln R. Anderson, K. N. Cissna, & R. C. Arnett (Eds.), 711e reach of dialogue: CO!ljirmatioll. I'oiee. (llId ("OIlIlIIllIIi/)" (pp. 930). Cresskill. NJ: Hampton Press. Cissna. K. N .. & Anderson. R. (1998). Theorizing about dialogic moment:-.: The Buber-Rogers position and postlllodern themes. Commullicalioll TheOl)", X. 63- 10-1.. Cody. M. J., McLaughlin, M. L.. & Schneider, M. J. (1981). The impaci of relational con~eqllence ... and iJ1limacy 011 the selection of interpersonal persuasion tactics: A reanalysi". Commullicatioll Quarterly. 29,91-106. Cowall. G .. & Avants, S. K. (1988). Children's influence strategie!o.: Structure. :-.ex diffcrencc~. and bilateral mother-child innuence. Child Delle/oplllelll. 59,1303-1313. Craig. R. T., Tracy, K., & Spisak. F. (1986). The discourse of requests: Assc~smenl of a politeness approach. HUIIIWI COl1ll1lllllicmioll Research, /2,437-468. de Turck. M. (1985). A transactional analysis of compliance-gainlllg behavior. Humall COllll1llll1icmio/l Research, 12.54-78. deTurck. M. A .. & Miller. G. R. (1983). Adolescent perceptions of parental persuasive message strategies. Journal of Marriage and the Family. 45, 543-552. Dillard. J. P .. & Fitzpatrick, M. A. (1985). Compliance-gaining in marital interaction. Personality (llId Social Psychology Bllllerill, II, 419-433. Dillard, J. P .. Wilson, S. R., Tusing. K. J .• & Kinney, T. A. (1997). Politeness judgments in personal relationships. Jotlmal of ulIIgul/ge alld Social Psychology. 16.297-325. Dunn. K. F .. & Cowan, G. (1993). Social inl1ucnce strategies among Japanese and American college women. P.\"ycholo~y of Women Qllorrer!y, 17, 39- 52.
334
Part I Y • COllfext,\ for Persuasion
Edgar, T" hClIlluth. y, S .. Hammond. S_ L.. Md)ollaJd. D. A.. & Fin"- E. L. (1992). Stratcgic ..e\uJI communication: Condom u'\c re ... i..,tance and rc .. pon!'.c. lIe,,!t" con/mllllicarioll. -I. 8)-IO·t Falbo. T. & Peplau. L. A. (1980). Po\\ er 'tratcgies in intimate rclation",hlp\. Jouma! of PersOfwlilY (III(/ SOl/a! Pn-dlO/ogy. 38. 618--628. Fitch. K. L. (1994). A cross-cultural .. tud), of dircctl\e <'cquence" and "orne implic;.lIion .. for compliancegallilng rc!'.Carch. Commu"icmirm Mmwgnlf1h,·, 61, 185-209. Fitch. K L. (1998). Speaking relariOlwll\': eli/flirt', COlllllllmicmioll, wul illterpenOl/(/1 ("olllleetioll. New Yor~: Guilford. Fitch. K. L. (in pre .... ). Cultural per~uadable ... COllllllllllimtioll Theory. Fitch, K. L.. & Sanders. R. E. (1994). Culture, communication, and preferences for dircclne .... in cxpn: .. . . ion of directives. Commullicmion Theon', -I. 219-245, Fo<; .. , S. K" & Griffin. C L. (1995). Beyond pcr,ua .. ioll; A proposal for an invitational rhetoric. Conl/llll11i· catiOIl MOllo?,mphs. 62. 2-18. Fnc/c, I. H.. & Mcllugh. M. C. (1992). PO\\er and lntluencc '>tnltcgie .. in \iolcnt and nomiolelll marnage,. P,ydwlogy of Wom('tI Quanah. 16, 449~65. Fung, S. S_ K.. Kipni .... D .. & Ro ... now. R L. (1987). Synthetic benevolence and m;'llc"olence a~ .. Irategic .. of relational compliance-gaining. Jrmmll/ of Social am/ Penmwl RelarimHIli!1.\", -I. 129-141. Gal\in. K M. & Bromme!. B. J. r2()(X». Famil\' COlllllllllliCCIlioll: Colle_\iof/ alld Change (5th cd.). Nc", Yor~: Addi,on-Wesley Longman. Gn, ... R II , & Seiter. S. J . (2003). Pl'/"llll/lioll,H}c/al illlll/l'Itct', tlf1(/ cOlllpliclf1ce-gtlllli1lg (2nd cd.). So,· tun: Allyn & Bacon . Goffman. E. (1967)./IItertlcti(J11 I"illlol: E\\un (mj(l("e-I()~r(lce bellm·ior. New Yor~: Anchor Book<.. Guerin. B. (1995). Social inOuencc in one-tn-one and group ~itualiom.: Predicting influence tactin rrom ba .. ic group proce ... ses. JOlll"lwl of SOl'illl P,nch(llog\'. 135, 371-385. lIoward .1. A. Blumstein. P .. & Sch"'t1rtl, P (1986). Se\. po",cr. and 1I10uence tactic ... in intimate rclaIlon,hlp\ JOllnwl of Persollality am/ SOCial Pn'dlOlo8Y. 5/, 102-109. Ifert. D. E. 12(X)Q). Resistance to lllterpcr ..onal rcquc ... h; A ... ummary and critique or recent re\earch. COlli1I1111tinll101t YearboOk. 23. 125-16 1 hkr. L.. POPI>cr. T.. & Ward. S. (1987). Children' .. purcha ..e rcqllesh anll parental respon\e ... Jourl/al oj Adl'tlnilill?, Rnearclt. 27, 28- .19. John, O. R. (199lJ). Consumer socialilation of children: A retrospective loo~ at twenty-five years of rl'\l'arch. JOUr/wI of C011SUIIIl'l" Rt'watch. 26. 183-2 1:t Jone~, D. C. (1985). Persuasive appt.!;lJ .. and rc .. pon ..c.. tn appeal .. among friend" and acquaintancc:-,. Child Ot'l't'IOf)II1C'III, 56, 757-76:'. Jordan. J M .. & Roloff. M. E. (1990). Acquiring a..... i .. lance from other .. : The effect of indirect requc .. ", and relational intimacy on \erbal compli~lIlcc. lIumoli Communication Rewardl, 16, 5 J9-555. Kellerman. K. & Cole. T. (1994), Cla .... ifying compliance gaining mes ..age .. : Taxonomic di .. ordcr and \trategic confusion. Commlllliwtioll 'J11l'Of'\'. -I, J--60. Kelley, Ii, Ii, Bcr\chcid. E .. Chri .. len\Cn, A .. Haf'\t:)'. J. H .. Hu.,lOn, T. L .. Lc\inger. G .. McClintoc~. E .. Pepl;lu, L. A .. & Pelerson. D. R. (1983), C/o<;e relmiomhlp_~, Nc" York: W. H. Freeman. Klrchler, E. (J 990). Spou"e,,' influence .. tr'•.1Iegie\ in purcha'>c deci ... ion .. a\ dependent on conflIct t)-pc and rch.lion .. lup characteristic<". Journal of 1::("{lI1omic P.\"W·/IO!Og\" II, 101 -118. Kin.:h lcr. E. (1993). Spouse,,' joint purcha ..e deci .. ion,> : Dctenllinanh of intlucnce tactic .. for muddling through tht.! procc'\s. Journal of E't'OlUmll<' P.\rdlOlogy, 14. 405~:'8. Ln)" K. L .. Watcr... E .. & Park, K. A. (19R9). Maternal rer..pon .. ivene"" and cbild compliance: The role of mood .... a mediator. Child Del'tdoplltel1l, 60, 1405- 1411. Leighty , G .. & Applegate, J. L. (1991). Social-cognitive and ,>itllational inl1ul'nce .. on the u~e of face":1\-1I1g per"uasi ve stralegie ... 1f1ll1U1II Comllllll1i('(l/iO/l Rt'_~e{lrch. 17. 451-484. Lim. T. S. (199Oa). Politeness behavior in .. oeial influence 'iituation ... In J. P. Dillard (Ed.), SeekillS COII/plialler: The production of il1lerpuI(ntlll inj7IIl'lIce me.I·,Wlge,\ (pp. 75-86). Scomdale. AZ: Gor,>uch Scari'\bric~.
Lim, T. S. (1990b). The influence of recel\cr'" re,j .. t;1nce on per"utldcr.. ' \-erbal aggre''''I\ ene~s. Commu· nicatioll Quarterly, 38. 170-188.
Chaptl'r 17 • Sm;tlllllfiu{'IICl'
11/
('low RdOliomJIII'\
335
LIEIl, T S., &. I·hml'r.... J. \\ I IlJlJl I. Llcl'\\urk SuIHbrtl). approh;ulon. and [;\1.:1. Hllmall ("OI/lI/I/III;m/;(1I/ HI'II'W'C", I '. ~ 15 -tiO. :Vlallalil'll,1.. (191)1)),\11 t:\alllina[lnl1 01 inll'qll'l ...onal intlul'IH:t: in CtllNllllPlioll and nnrH.'on",ulllption domain ... \dL'{II/('('\ 1/1 (""II.\UII1("/" /(("\I'flll II, :!6. IlJ6-.:!02. ~l;dl;tJil'u.I .. &. 1·>Iure. (" (1l)l)X). I m\;lld.1Il umkr .. tandln!! "flhl' dltJlI.:e ollllt1ucnCl' Wltic". The Impact 01 Po\\ el At/nll/l'I'1 /1/ COli \ 11111("1 /(1'.\I·Wl 11. 25. ~()7 -t 14 ~ lam d I. G, &. S~:hm ill. 1) R (1967 J. I )IIlII,·Il ... ion .. 01 t:ol1lpJiant'e-!!'lin ing heh;!, inr' An empi rkal ;lnilly"'1" ..\fllltIll/(·IIT, 3fi . .':\)0- .164 Md.'lu!!hlrn. Ii (19X3), Clilid compllanl'c 10 p;lrenwll'olllrol tcchnlquc .... Ot'l'dO/JlIII'IIWI P\\'choIOf,p, 19. 667- 673 1\ It.: ~ .lInt'c. S , L\:. Gergen. " J (1:1."'-), t 1l)l)lJ) RI 100io/ll/i rt',\/}(JII \ ",ilily. Ri',lown'l .lor ,1// I la;l/aMi' dialog/II' 'I hllu ... aml Oak ... C\ Sa!!e i\kth. S. (2000), race ,lIld bCl·\\mJ...: Ill1pilt:;!tillll'" for lhl' ... wd) 01 pl'N1nal rdatiol1 ... hlp .... In K, Dindia &. S, Dud (Llk), ("Ollllllllllimliall fllld !'t'nfl/wl rt'ltlll(1/1\11I1H (pp. 77-l)~1. Nl'\\ y(lr~ John \\ ilc) I\ku .... S .. ('up.ldl. \\ FL L\: 1Ill'lhun.', T II'}I).:!). P('rn:plIl1Ih ul "l'\u,ll t:nmpliance·n: .. I .. ting I1lc""age, in thrl'c I) pc .. of lft " ... -...C\ rl'lalitHl ... hip... \\ '1'\lUII JO//I'I/(/I "I ('(lII/III/minlliol/, 56. I 17 \ldll'r. (i. Bn ... tcr. I· .. Rohll!. 1\1.. (\: St.'lholJ. I), (1977J. COlllpli;lIh:l'·g;unlllg IUl' ..... a!:!t.' .. trall'git.' ... : A t}polog) mOlhl'r-t:hlld d)iuk ,\11 nt"'l'l"\aIHlIJ;11 ... tud) (""tid /J("I"t"iOJlIllt"llI. 57. 7~~ 7J::! Palan. K \I. L\: \\lIke .... R E. (1997), :\dolt· ...cl'llI·parl'nt IIHl'radinli mlamil) dl'ci ... ion maJ...1Ilg. jOllmal of COI/IIIlI/a Rnnm h, 2-1. I'll) 169 PMpal. \1. &. \Ial"tlh). I E. f 19K"), \1.lll'rn;t! rt."p0I1 ... 1\t.'nl' ..... ;tnd ~ub ... t.'llut.'nt child t:umplwllcl'. Child Ondll//lI/i'III, 56. 1-'26- I.t~~ POPPl:, M" 1;111 lIer Kloul. \\" & ValJ...cnhl'rg. II. (19'-)9). The implll'il ... lrLlClUrl' innul'l1ee ... tralt.:gie ... and "'o('ial rd;lllnn .. lup", journal oISOIialll/1(/ PUWJ1111 Relati(ll/lilipl. J6. +-I-J-----t5X. Rnioll \1 I L\: Jani"'l'\\ ... J...i. C \ (19X91, (herCOlll1ll!! (lh... tal"ll'~ In intl'rrlCr... onalt'omplianl'(,: 1\ principk' 01 Illl''''''ilg.l' ulIl"'lructioll IIII/mlll ('1II1/I1/I1I/;nl/;(1/1 Rnj'{mlt. 16.3.':\ 61. Ruloil. ~ I I.. Jamvt.'\\ ... kl. C :\, \h;(iritth. \1 .\. Burn .... ( S .. &. ~1anrai. LA. (19XX). Ar.:qllinng re"!HIreI.''' lrom int1lnate": Whl'lI Ilhli~;ttlllll .. uh... IIIUh.' .. for ]'lCr... ua ... ioll, "'//1u/ll (,(l/Ill11l11liclIl/O/i Re-
nr
wun-h. 1-1. 1(H .llJ6. Sa~rl' ... tan().
I \1 .. Chri .. tl'n~cn. ;\,. & Ilt.:a\l')-. (' I !1l)lJX). SIX."ial intlul'l1cl' technique, dunng Ilwri[al connie!. PI'fllll/al Ht,ltl/illl/l!,,/)I. 5.75 !oIt). S;mdl'r... R. I (19M7), ('ol[lIilil I' ((l/f//fiflliu/I I fll nll( '1/1(/1('" ,ll't'I'("Ii. ('ollfmlllllg Ill1tiu,lllIliding \ i/l ('mll'erI(/Ii(ll/ 11111/1'1'1"1/11/,\1011. Alh;LIl)-. 1\)' Sl:I\Y Prl' ...... Salldl'r... R, I:. (19971. Ih(' prodllt:IHHl \11 ~)1l1h(llle ohJl'l·l ... " ... t'Olllponenl:- of larger wholc .... In J, O. Grl'l'lll' (hi,), \II"\\(/,I:( /,,.,,.1//(1;/11/ \'/1(/11('1'\ ill ("(I/II1/I/Ulinlll(}llllil'Or'l (pp. 245-277). Mahwah, ~J hihallill Sandl'r.... R I, .. &. Filt:h. t\ I (2()O I). rhe i(t:Hlal prat:til'l' 01 cmnplhtncc ...cdlng. Cmmlllmicalirm Theory, 11,26.l 2MIJ. Sdladft.'r. JI R .. L\:: CrooJ.... C " 119}10J, Chtld l.:ompli'lIll:e .5-l--61 Sl'ark,.I. (11)76). A dl ...... ificatinn nr illlll..'lIlioll;ll) at'I"', I . wlgmlge ill SoC"il'l.\, 5, 1-23. St'\ ton. C S .. & Periman. D. S, ( I 'lSI) ), (\Iu]lk'~' t'arct.'r orientation. gender role orienlal ion, and pcret.'i\'ed CqUH~ ,I" dl'\crminanh 01 mantal PO\\ l'f journal (lrM(lrri(/~(' IIm/I/1l' FWII;h. 51. 9.U-l)-l1 She(,r, V (' (1995). St'lhation "'L'dll1~ prl'di"'po.. lIi(m ... and ... u. .ecpllhilil} In a . .exual partner' ... appeal ... for l'lllldOIll U"'l' JOlin/til o/"\/)/,liol ('(111111/1111;('(1//011 Rl'H'lIIrlt, 23, ::! I ~-·~29.
336
Part IV •
COllfe:rfsjor Perslt{lJiOIl
Shepherd. G. J. (1992). Communication a~ innuence: Definitional excl usion . Communication Studies. 43. 203-219. Shimanoff. S. B. ( 1987). Types of emotional di ...closures and request compl iance between ~pouse~. Com11I1mication Monographs, 54. 85- 100. Sill ars. A. L. (1980). The stranger and the spouse a~ target persons for compliance-gaining strategics: A !:iubjective expected utility model. Hilma" Communication Research, 6, 265-279. Smith. T. E. (1983). Adolescem reactions to attempted pare ntal con trol and influe nce tec hniques. Jot/mal of Marriage and the Family, 45. 533-542. Steil. J. M .. & Hillman . J. L. (1993). The perceived va lue of direct and indirect influence stratcg ic,,: A cross cultural comparison. P.~ycl/Ology of Women Quarterly, 17.457-462. Steil. J . M .. & Weltman. K. ( 1992). Influence strategies at home and at work: A study of sixty dual-career couples. JOllrnal of Social alld Persollal Relationships, 9. 65-88. Todorov. T. ( 1984). Mikhail Bak!!ri,,: The dialogiC pril/ciple (W. Godzich. Trans. ). Minneapo li s: Universi ty of Minnesoia Press. (Origina l work published 1981). Tucker. J. S .. & Anders. S. L. (2001). Social cont rol of health behaviors in marriage . JOIIl'llal of Applied Social Psychology. 31,467-485. Tucker. J. S .. & Mueller. J. S. (2000). Spouses' socia l control of health behaviors: Use and effec ti venes ... of specific strategie1-.. Per.'iOlllllity lind Social P~'ycl/O logy Bulletin, 26. 1120--11 30. While. K. D .. Pe a r~on. J. C .. & Flint. L ( 1989). Adolescents' compliance resistance-Effects parent,' compli ance \lrategy and gender. Adole~allce. 24.595-62 1. Wilson. S. R.. Aleman. C. G., & Leatham. G. B. (1998). Iden tity impli cat ions of influence go.a l<:>: A re\i,>ed analysis of face-t hreatening act ... and application to seeki ng compli a nce with o;ame-sex friend ... Hlllllall COn/mlilliauioll Re'iearch. 25.64-96. Wilson. S. R.. Cruz. M .. Mar~hall. L.. & Rao. N. ( 1993). An allributional analysis of compliance-gaining interactioll"'. COmllllllli('{ltioll Monographs. 60. 352-372. Wil~on. S. R .. & Kunkel. A. W. (2000). Iden tity implications of influence goals: Similarities in perceived f'lce threats and face work across o;,:ex and close re lationships. JOllrna l of Langllage muJ Social PsycllOloKl'. 19. 195-221. Willemnn. H.. & Fitzpatrick. M. A. (1986). Co rnplian ce~gain in g in marital interaction: Power bases. processes. and outcomes. Communication MOllogl"llp/u, 53. 130- 143. Wood, J. T .. Dendy , L. L.. Dordck . E.. Gerrnany. M" & Varallo. S. M. ( 1994). Dialecti c of diffe rence: A thematic analysis of intimates' meanings for difference. In K. Carter & M. Prisnell (Eds.), Inrerprer;I'e approaches To illlerpersoflal communication (pp. 115- J 36). New York : SUNY Press. Zvonkovic, A. M .. Schmiege. C. J .• & Ha ll , L. D. ( 199-l). Influence strategies used when couples mak e work-family decisions and their importance for marital sat isfaction. Family Relatiol/s. 43. 182-
or
188.
18 Superior-Subordinate Influence in Organizations Randy Y. Hirokawa and Amy E. Wagner
Social innucncc i~ an ever-present aspect of organizational life. From formal board meetings to informal employee il1tt.!raclions. individuals seek (0 contro l the opinions and activities of others in the pursuit of personal and orgallil3tional goa ls. It is not surprising. then. that scholars from a variety of academic di..,ciplines have pursued the study of how individua ls and social units in an organizational context use verbal and nonverbal messages to modify the cognitions. beliefs. attitudes. values. and behaviors of others (Barry & Watson, 1996), This chapter takes stock of what we know abollt superior-subordinate influence in organizational sellings and in doing so assesses the strengths and limitations of this re\earch and \ugge ... h direction ... for future investigations.
Definitioll of Social InflueJlce The term "social influence" ha~ been used rather loosely in the organizational literature. Some authors treat .;;ocial influence and power more or les~ interchangeably (French & Raven. 1959: MintLberg. 1983: Salanick & Pfeffer. 1977). Others equate social influence with the exercise of interpersonal control (Kipnis. Schmidt. & Wilkinson. 1980). Still other ... use the term synonymou!o.ly with persuasion (Hirokawa & Miyahara. 1986). Because writers use the terlll ,,"ocial influence in different ways. it seems prudent to begin this chapter with a clear definition of what we mean by this term. As used in this chapler. social influence refers to the modification of elll individual's lJel/lJvior(s) through the \·erba! al/d/or 1l001\'erbaf symbolic actions of allotlier individual. " can be thought of. and seen a..,. the communicative exercise of power and social control in the organization (Kipni!o.. Schmidt. & Wilkinson. 1980). The essence of social influence is belun'iural challge-that i!o.. the !-.ymbolic action(s) of the influencer (agent) causes the
337
338
Part IV •
eml/('H,\ ./In' PerslIl/\ioll
inilucncee (larget) to cngage 111 bchavior(~) different from what he or !'! he wou ld ha ve oth erwise pnxluced. For example. an cmployee report s late to work on a regular basis. Hi!'! ~upcrvi~or tell . . him that he mu"t "tart co ming to work on time or she will fire him. Social innuence b a\\umed to tal-.e pl:.lce if the supe r vi~o r' s warn in g or threat causes th e e mployec to change his behav ior and :.,tart reporting to work on time.
Strategies Versus Tactics Social innuCllct; in organiwtions can bc differentiated on the basis of strat eg ies and tact ics (sec. e.g .. " arper & Hirobwa. 1988; Hi rakawa. Kod ama. & Harper. 1990; Hirokawa. Mickey, & Miura. 1991; Hirakawa & Miyahara. 1986; Kipnis, Schmidt, & Wilkinson. 19XO; Mowday. 197~; Sch il it & Locke, 1982; Yuki & Falbc, 1990). Strategies are blueprint" for goal achievement: they represent :.,ystematic plan:., that organizational members follo"v' to influence the beha\ ior... and actions of others in the organization (Berger, 1986). Fro ... t (1987) identifies fi\'c ...ocial influence :.,trategies that arc com mo nl y used in organi zational selling... :
I. Rell.\ollillg- the usc of facts and daw to ,uppon the development of a logica l argument (c.g .. "These charts and figures clearly indicate that X is the right thing to do"). 2. Ingratiation- the use of impres"ion management. nauery. and the creat ion o f goodwill (e.g .... , know I can count o n you to do X because you' re one of o ur best employee,"). 3. Assl'rlil'('I/{'.\.';'-the usc of a direct and forceful approuc h (e.g .. "Do X right away"). 4. Sanctiol1s- the u"e of orga nizationall y derived rewards and punishme nb (e.g .. " If you dOIl't do X. your promotion w ill be in jeopardy"), 5. Altrui.\I1/- appea lin g to the goodw ill of o th ers (e.g .. "For the sa ke of the company ~.II1d your colleagues. please do X"). In contra ... t. fllC1ic'I arc ... een a ... "instantiations" of strategies: that i.... they represent the ... pccific \erbal or nomcrbal :.,ymboli c actions (messages) produced by the agent to carry OUI hi ... or her strategy. Any strategy will have a number of differen t tactics associ· ated with it. For example, in using all "a ltrui sm" "trategy. an o rganizational member might choO',e to u"e i.l lact ic such a ... a "favor" ("A:-. a personal favor to me. co uld you please start corning to work on timeT) or "du ty" ("You owe it to your coworkers to report to work o n time"). Similarl). in employing a ":-.anctions" :-.trategy, an orga nizati ona l member may choo:.,e be(\\iccn a "promise" (" If yo u start comi ng to work on time. I will recommend you for the prolllotion you have been asking for") and a "warn in g" {"'Unl ess you stop being late for \\ork. your future with this company wi ll be in ~eriou~ jeopard y").
Identifying Strategies and Tactics In vestigation ... of social influence in organ iza tions have emp loyed two con tras tin g meth(xh- thc ... o-called det/uctil'e and inc/ucth'l' approaches.
Chapter 1X • SIII'''l"i(1/' SlIlwnlil/{lIl' 11I{711t"IC!' ill Org(llli::lIrioll\
339
Dedllclh'e approach Some re ...ean.:her... hi.l\~ dra\\n un Intapn:tiltion ... of c\i..,tlllg theoric ... of ..,ocial pO\l.er and intcr~r ... onal relation..,hip'" to generate a priori Ii ... t" (or ill\-ellloric ... ) of inllllence-"ec~ing bclul\ ior, ll"CU b) organizational mcmhcr.... Ba ... cd on the pioneering \\ orJ.. of Man\ ell &
Sehillilt (1967). the typical ,tudy (,cc. e.g .. AIl,ari. 1989: Mowda). 1978. 1979: Riehmom!. Da\ i..,. Saylor. & McCro"J..e). I \JX4: Vccch io & Su.., ... mi.ln. 1(89) prc..,enlcd rc ... pondenh with a prcdctennineu Ii,] of ,tratcgic bchuvior, <.Iml a..,J..ed thcm to indicatc which one ... they typicall) cmploy in intlucllcc-"ieeJ..ing "itualion .... Rkhmond and colleague ... (19X4). fur examp le. ll ... ed a uedUClin! approach to examine ... ubordinmt! pt!rccption ... of their ()wn and th eir MIPCn. i,or,,' u,e of ,ocial-intlucnce tuetic ... (\vhat they c<.t ll ed ·'beha\· ior alternatiOIl technique ... ·· or .. BAT... ··). Thc panicipant ... in their'tlldy \\crc pre,cnled \\ith 18 dliTerenl Ille".. agc ... (e.g .. "Your group nect"" )OU to do it:' "If you don·l. other, will tx: hurt:' "You promi"cd to UO It") a, ...ocialcd \~Ilh 18 diffal.!nt bcha~i()r alternation techI1Ique ... (c.g .. "<.iul):· "guilt:· "ueht"). They wcrl.! a ... J..ed to rate on a !I\c-point ...cale ho\\ frequcntl) Ihe) u"ed each of the Illc",agc ... 10 get thcir ..,llpervi,or to change 1m. or her behavior (5;;; vcry ortcn. 4;;; often . .1 ;;; llcca ... ionally. ~;;; 'cItJOtn. I ;;; ncver). U... ing the ... al11c I11c",age" and Ihe-point .;,ca le,. Ihe particlpan .... were then a"J..cd to indicate hO\Io,' often Iheir "upervi ... or lI:-..ed each Illc, ... agc to change their beha~ ior. The re:-..ul .... or Richlnond and colleaguc ... · ... tuLly indicated that 'llbordinatl.!'" "eldlllll u ... e mo"t ollhe 18 BAT... in their IIllerat.:tlon ... \~ ith ,upef\ i"or". but \~ hen the) do. the)' fi.l\Of the u'c of "expert" and ··...elfc ... tccm" approachc .... Supen 1,0r... were found to u ... e me ..."age ... a ...... fx... ialcd with BAT ... 11.1bekd "cXpeI1:' ·· ...elf·e ... teem:· "rc~~ aru from hdulvior:' "'eglli nwte-highcr authority:' and "pcr'onai rc:-..pon,ibility'· (p. 85). Critic ... or the dcdllcti\c approach ha\e argllcl1 that Ihi ... Illl.!thod i, of tJlIc"'lionahle \allll.! in idelllifying the actual inrIucncc-... ecJ..ing bdm,·ior) of organi/
1IIllllclive approaclz A numbcr 01 rc ... earcher... havc argued thal a beller way to idenlify the Infillellce-!o.ccking bell;'1\ ior... of urganiLational mcmber, i'i 10 u,c:.tn inductive methou (,cc. c.:.g .. Kipni!o. ct al.. 1980: \Vi,eman & Sl:hcllcJ..-Hamlin. 19M1). Here the re ... earchcr pr..:,cnh ... ubjecb with a h)potiletical influence ... ituation and then a,b them tn indieale \\Iwt me ...... age{ ... ) they would producc to inllucnce the turgc.:1. The re ... earcher .,ub~equclltly analY/c", the con· ,trucled mc.: ...... age ... to idenlif) the Infiuc.:llcc- ... ceJ..ing tactic ... di"iphlyed in them. For cxample. lIiroJ..i.lwu. Kodama. and lIarper (1990) prc ... ented manager ... wilh the follo'...·ing h) pothctical ,cenario:
340
Part I V •
CmUt' Xl.\!Or Per.\/UI.\iOIl
One of your ,>ubortlinaw!>. ha .. been reporting 10 wor'" late on a regular ha~is. In mOM casc!>.. he is never more than 15 minute!>. late. but hi .. regular tardincss is becoming an annoyance to other people in the ofticc. What wou ld you say to thi~ emp loyee 10 convince him 10 report 10 work on time?
The researchers analyzed the written responses of the manager!o. and identified four general types of inlluence-'-.eeking messages: "reward" strategies. involving va lued resources or outcomes (c.g., "If you come to work 011 time. others in the office will have greater respect for you"); "punishment" strateg ies. in vo lvi ng negative sanc ti ons or outcomes (e.g .. " If you don't start coming to work on time. I will have no choi ce but to fire you"): "altruism" strategies. relying on the goodwill of the manager or su bordinate (e.g .. " Please do mc a big favor and start coming to work on time"): and "rationale" strategies. involving the u.. e of explanation or juc.,tilication (c.g .. "You need to come to work on lime because other .. depend on you to perform their jobs"). Although the inductive approach appear .. to be favored over lhe deductive approach. it is not without critics. Several \Hiters havc notcd that the effectiveness of this method depends on the inherent realism of the hypothetical scenario or si tuation presented to organiwtional members. That is. in order for respondelll ... to produce influe nce-seeking messages that arc consiste nt with their actual behaviors. th ey must be presented with scenarios that coincide with the actua l or likely situat ioll!o. they face in th e organization (sec. e.g .. Buric,on. Waltman. Goering. EI). & Whaley. 1988; Canary & Spitzberg. 1987; Cody & McLaughlin, 1980). If the ...cenario i-. too far removed from the respondent'~ range of experiences. the intluence-.. ee"ing me ...... age th e respondcn t constructs i.. like ly to ha ve no correlation \\ ith his or her actual behavior (Miller. BO;ler. Roloff. & Seibold. 1987).
Alltecedel/ts of Illfluellce Behaviors A number of authors have notcd that the range of influence behaviors organizational members select and use is a 'TPO thi ng:' that is, it depend ... on lime, place, and occasioll (Cody & McLaughlin. 1980). In their comprehensive review of the lite ratu re. Barry and Watson (1996) organi/ed situational determinants of influence allcmpts into four categori es: (I) nature of re lationship between superior and subo rdinate. (2) organizational charactcri ... tics. (3) goal(s) of the influencing age nt. and (4) individual attribut es of the influencing agcnt. We discu .... cach of these categories in turn.
Nature of Relatiollships Relational Closelless.
Within an orga nizatio n. degree of personal liking will in pal1 determine the "ind of influence .. trategy and tactic used by the agen t on the target regardless of relative status. Several studi es have show n that the level or relational c lose ness (liking) between the influencer (agent) and the person being influenced (target) affects the type of influence behavior used by the agent (e.g .. Cody. McLaughlin. & Schneider. 1981; FitLpatrick & Winke. 1979; Miller. Boster. Roloff & Seibold. 1977). In general. the more the agent likes the target, the marc likely s.he or he is to use a positive strategy such as a
Chapter IR •
SllperiO/·~S/{b()rdillllte 11Ij7I1l!llce ill Orglllli::lItiom
341
"debt" ("I wi ll owe you big time if you do X for me") or a "favor" (,"Could you do me a favor and do X for me?"). On the other hand. the more the agent dislikes the target. the more like ly she or he is to lise a negative strategy such a~ a " threat" (",The next time yo u fail to do X. I will report you to management") or a ··negative moral appeal"' ("Only an irresponsible employee would fail to do X").
Relatiollal Power. The kinds of influence tactics onc uses in an organiLational setting often depcnd on th e balance of power in the relation:-.hip. Influence attempts can occur in three direction;: (I) upward (superior/target and subordinate/agent). (2) downward (subordinate/target and superior/agent), and (3) lateral (th e agent and target occupy the sa me space in the organizational hierarchy). Research has shown that the perceived level of subordinate power is the primary determinant of the tactic a superior will use in downward inlluence attempts (Tjosvold. Andrews & Struthers. 1992). For instance. a superior/agent who Pos!\csses more power than the subordinate/target will tend to use what Kipnis (1976) labeled ··directive" or ··power-over tactics" (e.g .. ··You will be fired if you do not get this report done in time··). However. when the superior and subordinate have rclatively equal power the superior/agent will lise ··collaborative tactics" (e.g., ··Let"s work to get this report done on time") (Kanter. 1977). Upward intluence attempts are gene rall y marked by tec hniqu es in vo lvi ng rational persuasion such a~ logical prescntations (Sch ilit & Locke. 1982) or rcason (Chacko, 1990). It is important to notc, howevcr. that other studies have shown th at rational persuasion is the tactic of choice rcgard less of thc dircction of inlluence (Barry & Bateman. 1992: Yuk i & Falbe. 1990) and that even when differences in directional intluence tactics are found. they tend to bc small. indicating that relational power may not be a major determinant of strategy selection (YukI. Falbe. & Youn. 1993). Leadership Style.
Upward inlluence attempts vary acco rdin g to the type of leadersh ip a superior employs. Studies show that participative leaders. those who empower their employees. are the recipients of more influence aucmpts (Cobb. 1986) than their aUlhoritarian coun t crpart~ and that these intl uence attempt~ are more direct in nature (Krone. 1992). In general. the tactics used by subordinates to influence participative leaders can be c haracterized as direct. rational. and overt (Ansari & Kapoor. 1987: Krone, 1992). Conversely. upward influence tactics used on less participatory leaders tend to be assertive. threatening, and politically motivated (Krone. 1992). Such ncgative upward influe nce tactics can also be expected when a leade r i ... perceived as ineffective by his or her employees (Chacko. 1990).
Subordinate Competencies.
Little research has been done on the relationship between subordinate communication style. performance. and downward influence attempts. The available literature shows that superiors use assertive strategies, coalition-b uildin g strategies. appeals to higher authority. threats of sanctio ns. and reason with subordinates possessi ng an unattractive comll1unication sty lc (i nattent ive, unfriendly. and unrelaxed) but use friendliness and reason with subordinates possessing a n attractive commun ica ti on style (attent ive. friendly. relaxed) (Garko. 1992). As might be expec ted, superio rs use positive influe nce tactics slich as ··reward·· and ·'exchange" (e.g .. ·'If you SlaY late and work on this report I wi!! pay YOll overtime'·) with subordi nates who perform well bUI use
342
Pari IV •
COllfexf.I·.Ii)r Per.H/wjoll
negative influence tactic~ ~uch as "as~ertjon" and "~anctions" (e.g., "I do not want 10 hear your excuses for Ilot having the report done") with poor performers (Ansari. 1989).
Target Resistallce.
The agent's tactical choice for an inlluence attempt is abo mediated by how much re~istance is expected from the target. Wilson. Cruz. Marshall. and Ran (1993) found that agents adapt their tactics depending on the target's reason for noncompliance. Their experimental study yielded two major findings: (I) agents are more likely to use antisocial tactics (e.g., "warnings") when they believe that the target's reasons for noncompliance arc within that person's control. and (2) agents are more persistent and use tactics that take i<.,sue with the target's reason for noncompliance when the target is in COI1trol of the situation but is doing something inconsistent with what the agent wanb done.
Orgallizatiollal Characteristics Scholars have investigated whether organizational-level constructs such a<., the size, norms. climate. unionization. or type of institution influence how persons within the institution go about their selection and execution of influence attempts. Overwhelmingly, these organizational-level constructs have heen shown 10 have little or no effect on influence tactic selection. a lth ough Schilit and Locke (1982) reported that results of much of the existing research might be confounded due to researchers' lack of vigi lance in controlling for these constructs. Their analysis revealed a difference between influence tactics used in small or private organizations and those used in large or public ones, with more informal influence methods being used in the former than in the latter. Additionally, Krone (1992) found that influence agents in institutions with decentraliled authority employed more open and direct tactics (e.g .. '"1 think that we ~hould develop a contingency plan before dedicating all our resources to a single project") than those who worked in institutions with centralized authority. Despite the lack of evidence that situational variables have an effect on an agent's choice of influence tactics. scholars do not unilaterally accept that these variables have no effect. Instead, some scholars point to problems in the methodology and conceptual ization of variables in existing studies (Burleson et aI., 1988: Cody, Greene, Marston, O·Hair. Baaske & Schneider. 1986; Jackson & Backus. 1982).
Goals of the Influencillg Agent The choice of influence tactic varies according to what the inlluencer wishes to achieve (Dillard, 11)90). For instance. an agent whose intentions have to do with the quality orthe relationship or interactions with another party is likely 10 li se co mpliance-gaining st rategies that remind the target of the costs of noncompliance (e.g .. ""If you don't teillhe boss you agree with me 011 this matter. she's going to think our department is disorganiled"), wherea~ the agent is likely to usc reason and evidence if the desired action is deemed especially imporlam (e.g .. '"Based on last year's reports, I think we should make a decision immediately"). An agent who is trying to achieve an objective individually will use ingratiation tactics, but agents who have organizational objectives in mind will use a multitude of tactics- including upward appeals. blocking, and rational persuasion (Ansari &
Chapter 18 • Superior-Suhordinate Influence ill
Orglllli';lIfiolls
343
Kapoor, 1987). Furthermore, studies have co nsi stently fOllnd that agent~ wishing to promote new ideas and precipitate change employ intluence strategies of reason and coal ition building (Howell & Higgins. 1990: Schmidt & Kipni s. 198-1) but re ly on ingratiation and assertiveness when they want (0 change the behavior of the larget (Schmidt & Kipni :-.. 1984).
Individual Attributes oj the Illfluencillg Agellt Sex.
Evidence regarding whether the sex of the agent affects the type.., of influence strategies likely to be used is mixed. Thus, it is uncl ea r whether men and women employ different intluence tactics. Yet consideration of any suc h differences is important in ordc.:r to understand superior-subordinate interaction in organizational sett ings. The 1110\t con"i\tent finding in support of sex differe nces indicates that mal es choose inlluence tuclics that are direct and involve power. whereas females choose tactics that arc indirect and collaborative (Gruber & White. 1986: Offerman & Schrier. 1985 ). Add itionally . the goal of the agent and the reaction of his or her target may differentially impact whe n men and wome n choose to reward the targe t (White, 1988). Harper and Hirokawa (1988) found that mal e managers reported using more punishment-oriented ~ trategie ~ , whe rea!-! their female counterparts used more rational and altruistic methods. Specifically, 64 percent of th e male but only 37 percent of the female managers surveyed in th e stud y indicated that they would rely on puni!'.hment-based tactics such as "ultimatum" (",Shape up or find your\elf another job"). "warning" (" If you don ' t ~ hapc up, you won't be with thi ~ com pan y very much longer"). "threat" ("The next time you show up late, I will start dock in g your pay") and "negative esteem" ("Unless you are punctual. others will not view you as crc.:d ible and tru stworthy"). In comparison. 30 percent of the women but only IJ percent of the men indicated that they would rely on altruistic tactics such as "counsel" (" Is there anything I can do to help youT) and "duty" (.. It is your obligation to report on time for work"). The remainder of the female manager ... reported that they would rely on rationale-based strategies like "direct request" ("I would like you to make a special effort 10 start co ming to work on time") and "explanation" ("You need to report to work on time because ... ") (p. 164). Notably, however. a comparable amount of empirical research suggests that th ere are no significant sex differences when it co mes to the target 's choice of innuence ~tra t egy (Vecchio & Suss mann. 1991: Yuki & Falbe. 1990). Researchers have arg ued that the \e, d ifferences that have been found co uld be better explained by si tuati onal and individual difference factors such a:-. the legitimacy o f the request (Hirakawa, Mickey & Miura, 1991). personal power of the individual (Hirakawa. Kodama & Harper. 1990). or the power associated with one's position (Howard. Blumstein & Schwart/. 1986, Mainicro. 1986).
Culture. Research shows that cultural differences affect the selection of intluence tactics depending upon the va lues of the cuhure and the kind of rcqu c\t being j ... sucd. Fitch (1994) conducted an analysis of pre vious studies and found that the likelihood of a target using directive tactics was contingent upon whether directness was valued by that t argc t '~ culture. Other resea rch notes a relationship between how effective an inlluencc tactic i\
344
Part IV • COlllexrs for Persuasion perceived to be and the appropriateness of that tactic in a particular culture (Kim & Wilson, 1994). Hirokawa and Miyahara's (1986) comparison of American and Japanese managers revealed that when a request is obligatory (c.g .. pcrforming onc's assigned tasks), Americans employ punishment-based strategies. whereas Japanese employ altruism or rationale. Conversely, when a request is not obligatory (e.g .. staying at work late) Amcrican managers prefer rationale or reward strategies. but Japanese managers prefer altruism. Overall it appears that reason-oriented strategies are used most often in American. British. Australian (Kipnis & Schmidt, 1983), and Japanese (Sullivan & Taylor. 1991) culLures when attempting to influence a subordinate. Hirokawa and Miyahara (1986) accounted for these findings in terms of three cultural imperatives that differentiate Japanese and American managers. I. Japanese and American I1w/wgcrs share difJtrenr as.'l//II/IJliolls regardillg rhe mosr eF feclil'l~ way 10 influellce others hI the orgal1i:;.orioll. In Japane~e organizations, manag-
ers appear to believe that the most effective way to influence employees is to take into account the circumstances involved and appeal to the personal motivations of the elllployee. American managers, on the other hand, appear to operate under the assumption that rhe most effective way to influence an employee is to lise one's ability to mediate or control rewards and puni . . hlllents for that individual. 2. Japanese alld American managers rely on d~fferelll JJOI\'er hases 10 if~/7I1ell('e their employees. In Japanese organizations, managers generally rely on organiLational identification to hring about changes in emp loyees' behavior. whereas American managers rely on their ability to control or mediate organizational resources to influence their employees. That is, the intluence tactics used by Japanese managers depend on the fact that Japanese organiLationai members have embraced the goals and values of their organization. The intluence tactics used by American managers. in contrast, do not reflect this fundamental assumption. Rather. they are based on the belief that employees' behaviors are tied directly to organitational resources and that manipulating those re. . ources can therefore bring abOUI changes in behavior. 3. Japanese mal/agers place a greater emphasis Oil ('OI/,ol'([le IIlliry Iltall Allleric(l1I I1lOfIagel'S do. Japanese managers place a greater emphasis on corporate participation and cooperation than their American counterparts. Whereas Japanese managers attempt to bring about change by getting their employees to view their role within the general scheme of the organization, American managers typically do not attempt to foster this corporate identification. Simply stated. Japanese managers appear to deal with employees in a "holistic" way (i.e .. as part of the organization as a whole), whereas American managers tend to treat employees as individuals within the organization. (pp. 262- 263)
Persollality Variables.
Scholars have sought to understand whether certain aspects of an individual's personality will help to predict the kind of innuencc tactic that individual is likely to use. The personality variables that have received notable research attention include ambition. Machiavellianism, self-monitoring, locus of control. verbal aggressiveness, and dogmarism. Each of these will be considered in turn. The ambition variable as it is defined here generally refers to an individual"s combined desire for power and achievement. Highly ambitious individuals have been shown to be largely effective in their influence attempts. especially where upward influence is involved (Schilit, 1986). They tend to exercise influence attempts frequently and use in-
i
I
Chapter 18 • Superior-SII/JordillllTe Influence ill Orgal/i::ariof/s
345
nuence tactics of reason and coalition building (Chacko. 1990) and manipulation and persuasion (Mowday, 1979). A person who is said to be Machiavelliall is willing to use power and deceptive or manipulative methods to achieve his or her goals (Christie & Gei~, 1970). Findings regarding Machiavellianism and innuence tactic selection have been mixed. O' Hair and Cody (1987) attribute this to the fact that the trait most likely represents more than one construct. High self-monitors are concerned with how others perceive them and modify their behavior to the dictates of a given situation (Snyder, 1994). Studies reveal that high selfmonitors tend to be more expressive when trying to persuade a target about an issue that is emotional or relational and use more ingratiation tactics overall (Farmer, Fedor, Goodman & Maslyn, 1993). In addition, high-self-monitoring males reported using significantly more compromise. emotional appeals, coercion. and referent influence than did their lowself-monitoring male counterparts (Smith, Cody, Lovette & Canary, 1990). Locus oj control can be defined as the extent to which individuals feel they control events in their environment. "Internals" believe they are responsible for the events that happen to them (e.g .. "I get promoted at my job because I work hard"). whereas "externals" believe events that happen are outside their control (e.g., "( can't get promoted because this organization doesn't recognize hard work") (Rotter, 1966). Internal influencers have been found to rely on rationality, the manipulation of positive feelings, and relational ties in order to gain their target's compliance (Canary. Cody & Marston, 1986). Conversely, external agents avoid rational persuasion and rely on soft strategies such as requests (Farmer et aI., 1993). Research has neglected to address locus of control possessed by the target: however, Wheeless. Barraclough, and Stewart (1983) hypothesize that external targets will respond more favorably to influence auempts that involve relational appeals than will internals, since the fOrlner group has a greater need for belonging. Verbal aggressiveness (also see chapter 7) is the willingness to engage in communication that attacks or injures the self-concept of another. Agents who are highly verbally aggressive use innuence tactics that include teasing, swearing, attacking the other's competence, and nonverbal expressions. Interestingly, aggressive communicators believe their behaviors are less hurtful than those of nonaggressive communicators. One study suggested that verbal aggressiveness might be a response chosen after a target has displayed unfriendly resistance to the initial innuence attempt (Lim, 1990).
Tactical COllsideratiolls In addition to the personality variables addressed above. scholars have also examined innuence tactics to determine how effective they are, in what combination they are likely to be used, and what outcomes they yield. Research has shown that the effectiveness of an influence attempt depends on the nature of the relationship between the agent and target. For instance, in lateral dyads (two people belonging to the same place in an organizational hierarchy) rational tactics are effective (Barry & Bateman. 1992). Ingratiation and exchange tactics have proved successful when influence attempts are lateral or downward, yet these same strategies have been ineffective for upward innuence attempts. Strategies that have been seen to be effective regardless of the direction of the innuence include
I
I
346
Part IV • COl1lexts!or Persuasioll
"consultation" (e.g., "We should screen the clients together before admitting them"), "inspirational appeals" (e.g., "This is the best idea I've had in years and it is sure to increase our profit margin"), and "rational persuasion" (e.g., "Based on the merits of the case we should sue" (Yuki & Tracey, 1992). On the other hand, "repetition" and "exaggeration" are ineffective in all directions (Barry & Bateman. 1992). Barry and Watson (1996) noted that the existing research on influence effectiveness is far from comprehensive and is limited in part due to methodological problems such as the difficulty of executing experimental manipulations and obtaining data from both the agent and target. A complex question that has received recent research attention is how the order and combination of influence tactics affect the compliance-seeking encounter. Research shows that strategies such as "legitimizing" (validating the target's point of view), consultation (asking for the target's point of view), "ingratiation" (purposely getting into the target's good graces), and "inspirational appeals" (persuading through enthusiasm) tend to be used in combination with one another, but rational persuasion is used alone as often as it is in combination with other strategies (Yuki et ai., 1993). Maslyn, Fedor, and Farmer (1994) found that the influence tactics chosen in initial and subsequent influence attempts vary with the personality of the agent and elements of the situation. In a similar vein, Kipnis and Schmidt (1988) developed a typology of agents and their associated inlluence strategies. The four types are (l) "shotgun" agents, who use a variety of tactics; (2) "ingratiators," who primarily rely on friendliness; (3) "tacticians," who favor rational persuasion; and (4) "bystanders," who engage in little influence-seeking behavior. Earlier work by Perreault and Miles (1978) also demonstrated that certain personalities are more likely to choose particular influence tactics. The five types they identified based on strategy use are (I) noninlluencers, (2) expert inlluencers. (3) referent influencers, (4) multiple strategy intluencers, and (5) position power wielders. The preceding discussion of research findings on the tactics superiors and subordinates use to gain compliance would not be complete without consideration of the outcomes these tactics yield. Job satisfaction has been found to be affected by influence tactic selection. Specifically, Vecchio and Sussmann's (1989) study revealed that subordinates are dissatisfied and have a lower quality relationship with their superior when the superior's choice of influence tactics does not match the subordinate's preference. Conversely, superiors who use coercive techniques tend to view their subordinates more negatively than those who use rational techniques. leading the scholars who conducted the study to conclude that employers may devalue employees they can easily control (O'Neal, Kipnis, & Craig. 1994). Research has shown that the influencer's strategy choice is also related to the target's job satisfaction (Roach, 1991) and satisfaction with supervision (Richmond. McCroskey, & Davis. 1986), as well as how the target evaluates the agent's communication competence (Johnson, 1992). Other real-world outcomes are associated with influence tactic selection. For instance. Judge and Brentz (1994) found that subordinates who employed tactics with the objective of getting Lheir supervisor to like them more were more satisfied with their jobs and received more promotions. Conversely, subordinates who used influence tactics aimed at getting their supervisors to view them as competent were less satisfied and received fewer job-related benefits. Dreher, Dougherty, and Whitely's (1989) analysis of M.B.A. salaries revealed a positive correlation between use of upward influence strategies
Chapter I g • SlIl'erior-Suhortiillllll' Itl}711f!1ICt! ill ()rglllli~atium
347
and pay. This finding is sOlllewhat at odds with research by Kipnis and Schmidt (1988), \\ hich concluded that male . . ubordinate . . who adopted a shotgun influence ... Iyle (characterized by frequent usage or a variety of influence tactics) had lower incomes and were viewed Ie ... !'! favorably by . . upervisor!'! than were !'!ubordinate~ who adopted the tactician ~tyle (emphasi/ing rational persuasion).
COl/clusion Nearly four decades of research on :..upcrior-Mlbordinate communication in organizational settings have yielded con:..idcrable insights into the kind:.. of messages used by superiors and subordinaws to influence and persuade one another. a ... well as the factors that affect their selection of persuasive messages. These accomplishmcnts notwithstanding. three problem ... exist conccrning our under:..landing or superior-~ubordinate influence in organizational contexts. First. few studies of superior-subordinate influence have actually studied the cOlllmunication belmvior of superiors and subordinates in organizational context ..... Virtually all of the studies we rcvicwed in thi:.. chaptcr identify the influence tactics of superiors and subordinates in one of two way!'!: (I) asking the subject:.. (Q select from a predetermined list the behaviors they typically employ in influence-seeking situations or (2) presenting a hypOlhetical compliance-gaining situation to the subjects and asking them to indicate how thcy would attempt to influence the person in the scenario. Notably absent are ~tudies that examine how superior. . and ~ubordinatcs aClually attcmpt to influence one another in real organizational . . ituatinns. Naturalistic studies of this kind are crucial in the future. because there is good reason to believe that the way we attempt to gain the compliance of sOllleone in a hypothetical situation could differ greatly from how we actually attempt do so ill a real situation. Likewise. the influence tactics we think we would use in an organizational context Illay differ greatly from the Lactics we actually use in that context or situation. A second problem with current superior-MJbordin~lte influence research is that it has generally overlooked the interactive nature of social innucllce in the organizational context. Rarely, if ever. does social inlluence occur through the lise of a single message tactic. To the contrary. we often encounter resistance to our initial in/luence attempts, and we must rc\pond to that resistancc with an alternative influence lactic. To date, very few studie~ haye examined how superiors and subordinates adjust or adapt their influence tactics to the re:..istallce they encounter from those they arc trying to influence. Future superiorsubordinate in/luence studies need to look more closely at the interactive nature of socialintlucnce proce~ses in the orguninllional context. A third problem with ~uperior-sllbordinate influence studies concerns our under:..tanding of the effeclivenes ... of various t)pes of tactic') and strategies. In large pan because we have neglect cd to cxamine actual influence processes and have failed to take into account the interactive nature of those processes, we currently do not know much about what really works. For example. we do not know whether "hard'" tactics like "threats" or '·warnings" achieve compliance more effectively than "soft" tactics like "reasoning·' or ··ingratiation:' Future rc\carch \hould thus pay closer attention to the so-called bottom line, that is. which tactics actually . . uccccd in bringing about desired intluence.
348
Pan IV • Context.flor PerJllmioll
Referellces_____________________________ An .. ari. M. A. ( 1989). Effects of leader ~ex. s ubordinate sex. and .. ubordinate performance on the use of innuence .. tralegies. Sex Role.\, 20. 283-293. An"'~lri. M . A .. & Kapoor. A. (1987). Organilalional con text and upward innuence tactics. Orga"i:..mimwl 8f'IUIl'ior al1d Huma" D ecision PrrJ('e<;."es. 40. 39-49. Barry. B.. & Bateman. T. S. ( 1992). Perception ... of influence in managerial dyads: The role of hi erarchy. med .... and tactics. Huma" Relatio".\·. 65. 555-574. Barry. B.. & Watson. M. R. (1996). Communication aspect ... of dyadic ...ocial in flucn ce in organizations: A review and integration of conceptua l and empi rical development .... In B. Burleson (Ed.). CommunicOlioll Yearbook /9 (pp. 269-~ 17) Thou ... and Oaks. CA: S'lgc. Berger. C. R. (1986). Social power ami intcrperson al com munication . In M. L. Knapp & G. R. Miller (Ed ... ). /Ialldbook oJimerperw}//(/! cmwlllmication (pp. 439-499). Bcverl y Hill s. CA: Sage. Burleson. B. R .. Wilson. S. R.. Waltman. M . S .. Goering. E. M .. Ely. T. K .. & Whaley. B. B. ( 1988). Item dcs irabllity effects in compliance-g~Hning researc h: Scven sludie~ documenting art ifacts in the .. Irategy :>.clection procedure. Hwmm CmmllImicll1iol1 Rese(lrch. I.J. ~29-t86. Canary. D. 1.. Cody. M . 1.. & Mar.. lon. P. 1. ( 1986). Goal types. compliance-gaining. and locu\ of control. Joun/al of Ltmgua1:e alld Social Pn-chologr. 5. 249-269. Canary. D. 1.. & Spitzberg. B. ( 1987). Appropriateness and effec ti vc ne .. ~ perceptions of con nict strategies. Hllman Communication Rt'\wlrch, 14.93- 118. Ch acko. II . E. ( 1990). Methods of upward innuence. moti vational need:-. and admini'itrators' perception of their ~upe rvi sors' Icader.. hip "Iylc". Group alld Organi:..alioll SlIIdie.\·, 15.253-265. Ch ri ... tie. R.. & Gei ... F. L. (1970). SllIdie.\ ill Ma chiavellianism. New York: Acadc mi c Prclis. Cohb. A. T (1986). Informal influe nce in th e formal organi7ation: Psyc hological and situational correlate ... Group and Orgalli:.llIimlfll S1I4die:J II. 229-253. Cody. M . J.. & McLaughlin. M. L. ( 1980). Perception:>. of compliance-gaini ng MlUations: A dimensional anal)' ... :-.. CommunicatiOIl MOllographs. -17. 132-1 ~8. Cod), t\·1 J.. McLaughlin. M . L.. & Jordan. W. 1. (1980). A multidimen .. ional .. caling of three sets of compliance-gaini ng <.,Iraleg .e .... COl1lllllflltcmioll Quarterly. 28, 34---46. Cody. M. 1.. MCLaughlin. M . L. . & Schncider. M. J. (1981). The impact of rclational co n ~c quen ccs and intimacy on the selection of inlcrper"onal persuasion tactics: A reanalysis. Communication QuorIt'rI\'. 29, 9 1- J()6. Cody. M . J.. Greene. J. 0 .. Mar.. ton. P. J.. O·I-Iair. H. D .. Baaskc. K. T. . & Schne ider. M . 1. ( 1986). Situation percept ion and message fo.tratcgy ..e lect ion . In M . L. Mc Laughlin (Ed.). Comllllmicmio/l Year· hoof.. 9 (pp. 390-420). Bc\erl y Hill ... CA: Sage. Dillard. J. P. (1990). A goal-dri vcn model of interpersonal in nuence. In J. P. Dillard (Ed.). Seeking com· pliwue: A production of illlerperwnal influence mes.wger (pp. 41 -56). Scottsdale. AZ: Gorsuch Scari ... hrick. Dreher. G. F.. Dougherty. T. W .• & Whitely. w. ( 1989). Influence tactics and .!)alary attainment: A gender· ... pcc ific analysis. Sex Roles, 20. 535-550. Farmer. S. M .. Fedor. D. B.. Goodman. J. S.• & Maslyn, J. M. ( 1993). Factors affecting the /He oJllp'H'ard influC'nce n/'ategier. Paper pn: ...e nlcd at the 53rd Annual Meeting of the Academy of Management. Atlanta. GA. Fitch. K. L. ( 1 99~). A cro:-.~-c ultural .. tudy of directive :o.equences and some implication s for compliancega ining re:-.earch. Comf1//U/ic(ltiof{ Monographs. 61. 185-209. Filtpatrid.. M. A.. & Winke. J. ( 1979). You always hUll the onc you lovc: Stratcgie~ and tactics in interpcr...onal confli ct. Communication Quarterly. 27, 3- 11. Frcnch. J .. & Raven. B. H. ( 1959). The bases of soc ial power. In D. Cartwri ght (Ed.). SlIIdies i" social power (pp. 150-167). Ann Arhor. M I: In stitute for Social Research. Fro .. t. P. J. (1987). Powe r. politics. and inlluence. In F. M. Jablin . L. L. Putnam. K. H. Roberts. & L. W. Porter (&h.). HlllldboQA oj org(lll i:"llIiOl/al commun ication (pp. 503-548). Newbury Park. CA: Sag.e. Garko. M. G. (1992). Persuading ... ubordinatcs who communicate in attractive and unattractive styles. MlUlt1gemem Communication Qu{/rtl'rly. 5. 289-315.
Chapter 18 • Superior-Subordinate Inflilence
11/
OrXlllli:ationJ
349
Gruber. K. J .. & White. 1. W. (1986). G\!nder difference,\ in the percept ion of self ... and ot hers' use of power !.trateg ies. Sex Roles. 15. 109- 11 8. Harper. N. L.. & Hirokawa. R. Y. ( 1988). A comparison of persuasl\e 'trategie ... u,ed hy female and mal e manager, I: An examination of downward inlluence. Comllllmicariol/ QlllIrlt'r/r, 36. 157- 168. Hirokawa. R. Y.. Kodama. R. A.. & Harper. N. L. (1990). Impact of managerial power on perl,uasivc ,tratcgy "e lection by female and male manager, . Mal/ageml'flf Commll1lic(lIIOII Quarterh'. 4, 30-
50. Hirol-.awa . R. Y.. Mickey. J .. & Miura. S. (1991). Effect'\ of rcq uc ... t legitllnacy 011 the compliance-ga ining lactics of male and female managers. Commllnication MOII()g raph.~. 58, 42 1-436. Hirakawa. R. Y.. & Miyahara. A. ( 1986). A com pari ~on of inlluence Mrateg ic, utlh l.cd by manager\ in American and Japanese organization'). Commllllicmioll QlllIrterly. 3.J, 250--265. Howard. J. A. . Blumstei n. P.. & Schwartz. P. ( 1986). Sex. power and innllcncc tactIc", 111 Inllmate relation!.hip'. JOllrnal of Personalit\, (lnd Social Psychology. 51, IO::!- J09. Howell . J, M.. & HI ggi ns. C. A. ( 1990). Lcader\hip behaviors. inl1uence tactics. and career experiences of champion" of technologica l innovation. Leadership Quarterly. I. 249-264 Jack'\on. S .. & Badus. D. (1982). Are compliance-gai ning '\Iralegies deJX!ndent on .... lUalional \-ariables? Central SllItes Speech Joltnw/, 33. 469-479. Joh n..on. G. M . ( 1992). Subordinate perception, of ,,;upenor' s COlllmumcallon compe tence and t.. sl-. allraction related to superi or'\; u,>e of comp liance-gai ning lactic, Wl'.\tem JOllnta/ of Commllnication, 56. 54-67. Judge. T, A .. & Brentz. R. D. (1994). Political inllue nce behavior and ca reer 'lIcce,". jOllmal (~r M(If/(/!?elIIelll, 20, 43-65, Kanter. R. M. ( 1977). Mell and \\'(}mefl of the corpol'(I{ioll. New York: Basic B oo~ ... KellenmlO . K.. & Cole. T. (1994). Cia ... ,ifying comp liancc-gaiOlng mc"'age ... : Ta'(onomic di,order and ,trategic confusion, Commullication Theon, -I. 3-60. Kelman. H. C .. & Hami lton. V. L. ( 1989). Crimes of ()betliellce: TOImrd (I .wdal/H\'dwlog\ of alltllOrm alUl re~fJo/l.fibiIiTY. lew '·Ia\en. CT: Yale Univen.i ty Pre.,s, Kim . M .. & Wilson. S. R. (199-1 ). A cro",-cultural comparison of impliCit theories of rcquc"IIO£. Commll"inOiOtI MOllographs. 61, 2 10-235. Kipni .... D. ( 1976), The powerlwlder~ . Chic
350
Pari IV • CO/ltcxtsIor Pers/lasioll
M inti' berg. H. ( I ()Rl). PO\\"er ill alld around orMwli::.atiol1.l, Englewood eli ff~. NJ: Prentice-I-In II. Mowday. R. T. (1978). The exerci~e of upward influence in orgnnil.ntions. Admilli.llmlil'e Science Quarterly, 23. 117-[56. Mowdny. R. T. ( 1979). Leader chnracteri~tics. ~elf-('onfidence. and methods of upward inlluence in organizational decision situation~. A('adcmy of ManaMell/enl Jour/wi. 22. 709-725. Offerman. L. R., & Schrier, P. E. (19RS). Social influence strategie<;: The impact of~ex, role and attillldcs tuward power. Penonality and Social P.lyclwlogr BIII/l'till. II, 286-300. O'Hair, D .. & Cody, M. J. ( 19R7). Mnchiavellian beliefs and <;ncial inlluence. Wewern jOllmal (~f.Si}ecch Commullicaliol/, 51. 279-303. O'Neal. E. c.. Kipnis, D .. & Craig. K. M. (IY9.t). Effects on the persuader of employing a coercive inlluence technique. Basic and Applied Social PsycholoMY. 15.225-238. Perreault, W . D .. & Miles. R. H. (1978). Innuence strategy mixes in complex organi/.ations. Beh(ll'ioml Sciellce. 23. 86--98. Richmond. V. P .. Davis, L. M .. Saylor, K.. & McCroskey. J. C. (1984). Power ~trategie~ in organinllion~: Communicat ion techniqlle~ nnd messages. Hllmall COIwllllllic{/fioll R('.~earch, II, 85- 108. Richmond. V. P., McCroskey. J. C .. & Davis. L. M. (1986). The n.:lationship of :-.upervi~or lI~e of power and affinity-seeking ~trategie~ with subordinate <;ati~faction. COllllllllllimtioll Quarterly. 34. 178193. Roach. K. D. ( 1991 ). University department chairs' LIse of compliance-gaining ~trategies. COll1l1l11nic(l/ioll Quarterly. 39, 75- 90. Roller. J. B. ( 1966). Generaliled e)..pectancics for internal versus external conlrol of reinforcemelll. Poly· c/wloXiclIl MOIwMraph.l. 80 ( I. Whole No. 609). Salanick. G. R .. & Pfeffer. J. (1977). Who gets power- and hDW they hold on to it. O/'gal1i~a'iol1al Dynamics. 5. 3-21. Schilit, W. K. (1986). An examination of individual differences a~ moderators of upward influence activity in ~trategic decisions. HIlIlIllIl Reiatiun.l, 39, 933-953. Schilil. W. K.. & Locke. E. A. (1982). A \lu(ly of upward inlluence in organiJ:ation~. Adlllini.ltmlil'e Sden('(! Quarterly, 27. 30+-316. Schmidt. S. M .. & Kipni~. D. (198'"') Manager~' pUf<;lJit of individua l and organii'ational goa l ~. Hlllllan Relatiolls. 37, 781 - 794. Seibold. D. R.. Canlrill. J. G .. & Meyer~. R. A. (199 .1-). Communication and interpersonal influence. In M. L. Knapp & G. R. Miller (Ed~.). Handbook of iflterper.lo/wl commllllic(l!ioll (2nd cd .. pp. 542588). Thousand Oaks. CA: Sage. Seibo ld. D. R .. Cantrill. J. G .. & Mcyer~. R. A. (1985). Communication and interper~onal inlluencc. ln M. L. Knapp & G . R. Miller (Eds.). Handhook of illrerjler.\onal comlllilnication (pp. :')51 - 611). Beverly Hills , CA: Sage. Smith, S. W .. Cody. M. J.. Lovette. S .. & Canary. D. J. (1990). Self-monitoring. gender and compliancegnining goals. In M. J. Cody & M. L. McLaughlin (Eos.), The psychology (~rl{/cli("(/l coml1Jllllicofirm (pp . 91-134). Clevedon, England: Multilingual Matters. Snyder. M. (1974). Self-monitoring of exprev..,ive behavior. journal of f-'ersOIllIlity and Social Psycho/oMY. 30, 526-537. Sullivan. J.. & Taylor. S. ( 1991 ). A cross-eu Itural te~t of compliance-gaining theory. Manage/l/ellt CommUllication Quarterly. 5. 220-2.19. Tjosvold. D .. Andrews. I. R.. & Struthers. J. T. (1992). Leader:-.hip influence: Goal interdependence and power. jOllrnal ofS()cial Plyc!wlog\', /32. 39- 50. Vecchio. R. P .. & Sussman, M. (1989). Preference~ for forms of supervisory socinl influence. Journal of Or~{/lIi::.alional Bellm'ior. 10. 135- 143. Vecchio. R. P .. & Slls~rnan. M. ( 1991). Choice of innucnce tactics : Indi\idual and organizational determi· nant~ . jOllrl/a/ (~l Orgalli::,atiolla/ Bellm'ior. 12. 73-80. Wheeless. L. R .. Barraclough. R .. & Stewart. R. (1983). Compliance-gaining and power in persuasion. In R. N. Bo~trom (Ed.). COlllllllll1icarioll Year/wok 7 (pp. 105- [45). Beverly Hills. CA: Sage. White. J. W. (1988). Inlluence tactic<; as 11 function of gender. insult. and goal. Sex Rolf!.\·, 18,433--448.
Chapter 18 • Superior-Subordillllte Influence ill Organizatiofls
351
W il son, S. R.• Cruz, M . G .. Marshall , L. J., & Rao. N. ( 1993). A n anribUl io nal analysis of compliancegaining interactio ns. Communication Monographs. 60. 352-372. Wi seman, R. L.. & Sche nck- Hamlin . W. ( 198 1). A multid ime nsio na l scaling val idation of an inducti velyderi ved set of compli ance-gain ing strategies. Communication MOllographs, 48, 25 1- 270. YukI, G., & Falbe , C. M . ( 1990). Innuence tactics and objectives in upward . downward and lateral innue nce atte mpts. Journal of Applied Psychology. 75, 132- 140. YukI. G .. Fa lbe, C. M .. & Yo un , J. Y. ( 1993). Patte rns of inOue nce behavior for managers. Group and Orgalli ..atiol/ Studies. /8.5- 28. Yu kI. G., & Tracey, J. B. ( 1992). Conseq uences of innue nce laclic!"' u~ed with subordinates. peers, and the boss. Journal oj Applied Psychology. 77.525- 535.
19 Social Influence in Selling Contexts John S. Seiter and Michael J. Cody
Successfu l influence often email ... knowing how
(0
your audience. Some influence
agents may routinely try to be "Iikable:' while others try to be ··expert.'· These people probably lend to be more influential than those who seem unlikable or who are lacking in expertise. However. there is evidence that adapting to one's audience pays the best dividends. This point was vividly discussed in Ciald ini·s (2001. pp. 198- 199) observations of "VinccllI:' it waiter at an upscale restaurant who made the most in lips. When serving a
family, Vincent was clever, friendly. even clownish with children. keeping everyone happy. When serving a couplc. he would become more formal, recolllmending expensive
items. acting even slightly "impcriou\." When "erving an older married couple, he retained formality. acted less \upe rior, and showed respect. When serving a large dinner parry. he recommended what was best prepared that evening. and he was an expert in wines. Vincent knew whar hi\ various types of ci iems valued, and he gave them all what the) wanted; he knew how to ensure that he would either increase the amount of the bill (on which the tip was calculated) or increase th e amou nt tipped--or both. This chapter sllggesl~ that such an :.Ibil ity to adapt to particular audiences is a key :.I"pect of 'illcces,ful inlluence attempts. in restaurant settings and beyond. Specifically, it focuses on adaptation and other factors that lead to success in one of the most common of <111 persuasive encounters: the buying and selling of merchandise in retail stores. In doing so, the chapter draws upon work in a number of disciplines, including psychology, business. marJ..:eting. cOIllJ11unication. and personal sell ing. Considering that volumes have been written on this topic. we must narrow our focus to a few key issues. We begin by exploring the nature of goals and how they affect selling encounters. Second, we examine the type!'. of resources retail salespeople need in order to interact with customers successfully. Finally, we discuss the resuits of two of our own studies illustrating the implementa· tion and effectivcnc~s of several influence tactics used by different types of salesclerks in re tail stores.
353
354
Part IV • CO/l/('xHjor PerW(lS;OIl
Multiple Goals ill Sellillg EnCollllters Chapter II in this volume presents a detailed account of cOlllmunication as a goal-directed activity. When producing a message. goals lead peoplc to make plans aimcd al achieving those goals, and the plans in turn are used to selcct and guide behaviors ror carrying them out (see also Dillard & Solomon. 2000: Greene. 2000). Selling merchandise. like most communication encounters, invol ves the same proce:..s; ... ale:-.people forlllularc plans for making sales. and these plans innuence the strategies, tactics. and message,", that they then communicate to custolllers. Although at tirst glance. thi s process Illay seem simple. it is not. People often pursue multiple goals (see chapter II). and their goals may change during the course of an interaction (Greene. 2000). For Ihal reason, we suggesl that effective salespeople arc those who have the ability to juggle multiple goals. For exan1ple. Spitzberg and Cupach (1989) suggested that competent communicators are both effective and appropriate. We believe the sa me can be said of salespeople. To be effective. salesclerks must meet their instrumental goals by making sales. but if they U"iC inappropriate or unethical tactic ~ slIch as deception or intimidation. they may not succeed in cultivating "return" cu . . tomcrs or may be perceived as manipulative. In short. . . alesc lerks must walk a thin line between goals that have the potential of competing with one another. Prev ious research has sugges ted that people pursue three general type" of goals through communication : instrumental. self-pre . . entation. and relational (Clark & Delia. 1979: Dillard, 1990: O'Keefe & McCornack, 1987: Tracy. Craig, Smith, & Spisak, 1984: Wilson, 1990). First we turn to a discussion of self-presentation and relational goals in selling contexts. Later in the chaptcr we di sc uss salcspersons' instrumental goa].., and tactics for achieving them.
Self-Preselltatioll Goals ill Sales Illteractiolls According to Impression Manage ment Theory. most people want to be percei\ed in a positive light and therefore communicate in order to create desired impres ... ions of themselves (see Goffman, 1959). This i~ also true of salcspeople. According to Leathers (1988). because impression management focuses all how people sell themselves, it is not surprising that impression managcment skills are especially important to salespcople. Previous litcrature has suggcsted that sales training is now a big business and that a large number of training courses focus on teaching salespeople how to project positive images (Leather;, 1988 ). Jones and Pittman (1981) di scus . . tive common self-presentation styles (sec also Canary, Cod). & Manusov, 2000). First. the illgraliaJor \.\ants to be liked and therefore engages in behaviors such us praise and appearing friendly. kind, helpful. and positive. Second. the il/limit/afOl" desires to be seen as dangerous and tough and may use strategies such as threats and displays of <:Inger. Third. the selj~prol1lO1er wants to be perceived as competent. effective. and successful and therefore may try to boost his or her credibility by attempting to appear smart or talented. Foullh. the exelllplijier tries to appear dedicated, committed. and self-sacrificing. Finally, the slipplicalor wants to be perceived as helpless. unfortunate, and in need of nurturing.
Chapler 19 • Socia/Influence in Seiling Contexts
355
Clearly. not all of the,e ,ell~prc,cntational "yles will benefit salespeople who want to be effecti,;e. What cu",wmer. for example. wants a helpless salesclerk assisting him or her? Moreover. it is hard to imagine that threatening tactics would work well with too many customers. Having ..,aid that. we need to determine what images are most important for a salesperson to project. Though previous literature ha ... identified a large number of characteristic!-. de ... irable for ..,ulc!-.pcP'Ions (e.g .. a-.!-.eniveness. attractiveness. interestingncss). perhaps the most important and well documented are credibility and likability. We discus.., these next.
Credibility.
According to Gass and Seiter (2003), although credibility may be made up of ..,everal dimcn ... ions. scholars generally agree that competence and trustworthiness are the two th:1I are almo ... t alway ... rele,"anl to the evaluation of sources (see also chapter 6). Research and theory on personal selling have confirmed that these dimensions are important for customers intcracting with salespeople. For example. previolls research has shown that experti,e is beneficial both for establishing cu;tolller trust (Busch & Wilson. 1976) and for producing the intended behi]\ ioral re ... ponse in the customer (Bu~ch & Wilson. 1976: Jones. Moore. Stanaland. & Wyatt. 199H). Comstock and Higgins (1997) reported that tru..,' ilnd rapport are unequivocally the moq important lheme!-. during the sales proce,,~ and that cu~toll1er ... prefer salel"peoplc who arc trust\l"orthy oyer those who are similar to thcm. Other re ...carch has I"hown that when cll"tomers tru\! salespeople. they are more likely to engage in open and free-Ilowing communication (Chow & Holden. 1997). are more loyal to the company for which the salesperson work' (Chow & Holden. 1997: Garbarino & John!-.on. 1999). and are more cooperative with the salesclerk (Schurr. & a/anne. 1985). Finally. in a mctu-analysi!-.. Swun. Bower". and Richardson (1999) concluded that although trust ha" a moderate influence on the development of cu ... tol11er 3uitude~. intentions and behavior..,. its influence is beneficial. Givenlhi..,. what characteriMic .. lead to perceptions of expertise and trustworthinc~s? Retail ... alcsperson.., might demonMrate experti..,e by claiming personal experience of a product or by being kno\\ kdgeablc about fashions. designers. material~. care. and so forth. Trustworthine ... s might be establi..,hed through non1l1anipulutive tactics or by pointing out po ...... ible drawbad.s to cenain products. Nonverbally. salespeople might project a credible image by appearing relaxed. u"ing appropriate eye contact and vocal cues. smiling at appropriate times. and speaking nuently (Leathers. 1988). Finally. a study by Ramsey and Sohi (1997) found that when customers perceived that ..,ale"'peoplc were listening carefully (i.c .. \cn ... ing. evaluating. and rc\ponding to what they ... aid). they not only trusted the I"ale\people. they expected future interaction.
Likability.
In addition to projecting a credible image. effective "ale ... cJerks tend to be likable. Sales clerk' may bolster their likability in ,everal ways. First. people tend to reciprocate liking. A" "'lIch. ~alc\people who demonstrate liking for their customers by being friendly or by doing favor ... arc more liJ...cd them..,elvcs. According to Leathers (1988). likable ... alespeople tend to ... mile more. u\e a lot of eye contact. engage in affirmative head nodding. and establi ... h an open body position. Moreover. clerks who demonstrate such positive emotion ... may be more persua ... ive. For example. Sharma (1999) found that when ... alc!-.pcople dCll1on..,tratcd po ... itive emotions toward their customers. their cu ... tomers
356
Part IV • COllfexrsfor P('r~lIaS;OIl
listened (Q sales pitches more carefully and were more easily persuaded. Jones et al. (1998) also found that salesperson likability positively influenced customers' purchases. Second. customers respond more positively to salesclerks when they perceive them as similar to themselves (e.g., Boles, Johnson. & Barksdale, 20(0). Among other things, clerks can appear more similar by matching their customer's nonverbal behaviors (Leath-
ers, 1988) Hnd by conforming opinions (e.g., "I Hgree, that blouse looks great on you.") (Cody & Seiter, 200 I). Studies have shown that cu>tomers tend to trust, be more cooperative with. follow the advice given by, and buy more from salespeople with \\'hom they perceive they share demographics, experiences. appearances, attitudes, communication
styles, and personality traits (e.g., Busch & Wilson. 1976; Dion, Easterling, & Miller. 1995; Evans. 1963; Fine & Gardial. 1990; Gadel. 1964). Finally, salesclerks who praise their customers may be liked more than those who do nol. An analysis of several studies by Gordon (1996) suggested that ingratiation is an effective tactic. But what if the ingratiator's ulterior motives are transparent? For example. what if customers suspect that salespeople are praising them just to make a sale? According 10 Burgoon (1994), ingratiation is most effective when the ingratiator's motives are
concealed. Even so, Cialdini (200 I) noted that false flattery leads to almost as much liking for the ingratiator as sincere flattery does. Not surprisingly. then, research suggests that
salesclerks who praise Ihcir cuslOmers are more effective then Ihose who do not. though Strutton. Pelton and Lumpkin (1995) suggested that if salespeople use ingratiat ion. they ~hould
do
M)
only to a moderate degree to reduce the risk of mistrust on the part of
cuSlOmer~.
Relatiollal Goals ill Sales Illteractiolls White salespeople pursue instrumental (making sales) and self-presentational (creating a
credible and likable image) goals. they also must consider relational goals. Traditionally. building a relationship with customers has not been seen as an important goal for retail salespeople. whose interactions with customers have traditionally been characterized as one-lime-only and brief. The past decade. however. has witnessed a major change in theory and practice related to personal selling and marketing. For example. Wortuba
(1991) and Weitz and Bradford (1999) suggested that the nature of personal selling has evolved through four "eras"-production, sales, marketing, and partnering-and that only the last era stresses the importance of interpersonal communication and of building and maintaining long-term relationships with customers. Nowadays. buyer-seller relationships are acknowledged as potentially important in alllYpes of selling conrexts. Indeed, recent work, including the Commitment-Trust Theory
of relationship marketing (Morgan & Hunt. 1994). has suggested that although some customers have low relational orientations and are simply interested in satisfaction with a product or service, other customers have high relational orientations and as such are strongly intluenced by the amount of trust or commitment they have in a company and its
salespeople (see Dwyer. Schurr, & 011, 1987; Garbarino & Johnson, 1999). For these people, salespersons are often expected to play the role of "relationship managers"
(Crosby, Evans, & Cowles, 1990) between themselves and customers. Nordstrom's. for example. emphasizes not only customer-salesperson relationships but suggests that the
Chapter 19 • Socia/lnfluellce ill Selling ComeXfS
357
pcr"ionali7cd "icrvice that charactcri/c"i thel.,e interactionl., helps to build customer-company relation .... In \hort. rcgardlc ... :-. of the \clIing COIllCXL. developing buyer-scller relationships can be important. Leigh and McGraw (1989) found that in industrial sales, successful salespeople reported thm their "iecond. third. and fourth 1110st frequent goals in initial sales calls were related to rclatiol1",hip dc\-e lopmcnt. Reynold"i and Beatty (1999a) deli ned a l:u ... tol1lcr-sa le ... per",on rclationllhip as existing "when there i... an ongoing ...erie"i of interal:tion"i bctwecn a "ialer..person and a customer and the pi.ll1it::-. know each other"(p. 12). Previolls literaturc indicated that such relationships Gill have nUl1lerou~ advantagc"i for l:ustol1lcrs. Thcl.,c advantage, include functional benefit"i (c.g .. :-.:1\ ing timc. convcniclll:c. fashion advice. beller purchase deci ... ions) and social benelit"i (e.g .. enjoying another pcr"ion'\ company. enjoying time SpCIlI with a salespCNHl) (Beatty. Mayer. Coleman. Reynolds. & Lee. 1996: Gwinner. Gremler. & Bitner. 1998: Reynold ... & Beall}, 1999a). In addition. I.,uch relatiol1"ihip ... can have advantages for bu ... ine ... "ie .... Such advantage ... include cu ... (Omcr "iati ... faction with ..,ale"'per...olll., and companiel.,. cu..,tomer loyalty. fa\orablc \\ord of mouth. and increa~ed pUrcl1
Differellt Types of Clerks Based Oil Differellt Goals PlIrslled As \\'c have already di"icu ......ed. "ialc"iclt.:rb have instrumental. ",elf-prc"ienlational. and relalIonal goal "i. Thi.., is not to ..,ugge ... t that all <.:Ierk"i pur!'>ue all of these goals. For example. though Ford (1999) noted that many ... ale ... people de\elop a combination of service styles that include courteous. per... onali/l!d. and manipulative communication behaviors. previau ... rC"'L'~In: h ha ...... uggested that "iOllle salcl.,people \tick to one ~tyle of selling. Different ..,alescler\... ... may thu", be charactcri/cd by different configurations of goab-.. npproaching buyer-seller interaction::. in a \'aricty of way .... A good deal of previous re~earch has attempted to l:U1cgoriLc ",ale ... peoplc bu!>.ed on their approaches to selling (e.g .. see Busch & Wilson. 1976: E""" 1963: Williallls & Spiro. 1985), One or our ;lUdies (Cody & Seiter. 100 I ) idcnt ified four different ... alc"'pcrsoll style ... based on ob\crvation~ of buyer-seller interaction"i in rl!tail "itore .... Many of thc..,c I.,{} les correspond to the !o.elf-presentational styles we di"icu"'''ied earlier in thi ... chapter. Fir"it. ingmliarioll-sly/i' clcrk ... \'vcrc characterized by a reliance on tactics that helped them culti\ate an image that wn ... lil-..ablc and fricndl}. Second. wsk-oriellled clerks focll"ied on inl.,lrumental goal\ and tactic!>. aimed primarily at "making lhe sale," efficiency, and controlling the l:uslOmer. Third. dil'llI-oril l1Il'd clerks u\ed a variety of tactic .... tried to J
358
Pari IV • COlllexlsjor Per.HllIsioll
understand customers' needs, and adapted accordingly. Finally, passil'e-illactive c lerks were characterized as apathetic. lIsing few selling tactics. approaching customers infrequently, and performing "mechanical" tasks (e.g .. running the cash regbter). As might be expected. each of these styles influenced the outcomes in sales interactions. We examine these resulLs later in this chapter. Now. however, we turn to a discussion of salesperson resources.
Resources and Selling Comstock and Higgins (1997) argued that effective salespeople are competent comm uni cators. But what makes a communicator competent? Past literature has suggested thai competclll communicators have the knowledge. motivation. and skills necessary to be effective and appropriate in a given context (Spitzberg & Cupach, 1989). How do salespeople come by such resources'? One approach argues that certain trails characterize successful salespeople. For example. in a review of literature. Anselmi and Zemanek (1999) noted that no one personality profile exists for the perfect sa lesperson but that slIccessful salespeople tend to be we ll mannered. verbally skilled. punctual. enthusiastic. social, and competi ti ve. However1 research attempting to identify universally effective sellin g traits has also been inconsistent and equivocal (Sprowl. Cm·veth, & Senk. 1994; Weitz. 1978). Perhaps this is because se iling encounter~ are interactive and thereby influenced by bOlh buyer and sell er characteristics (Sprowl et aI., 1994). For example, Fine and Schuman n (1992) found that the personality of both the salesperson and the CuMomer influenced the outcome of sales encounters. Therefore, when specific salesperson traits influence successfu l sales, it may be because the trait~ fos ter involvement and interaction with customers (Boorom. Goolsby. & Ramsey, 19~8). Though we believe traits may playa ro le in effective sales, our position is that successful selling is primarily the result of learning. This notion is consisten t with theories of message production and cognitive selling paradigms. which foclIs on linkin g behaviors (e.g .. smiling wh ile showing merchandise) to underlying knowledge that indi vid ual s learn with experience (see Dillard, 1990: Gengler, Howard, & Zolner, 1995: Greene, 2000: Leigh & McGraw, 1989: Macintosh. Anglin, Szymanski, & Gentry, 1992: Meyer. 2000: Weitz. Sujan. & Sujan, 1986). The general idea is that to maximize career effectiveness, over time sale ... people abandon ullsuccessful approaches while developing. refining. and remembering successful ones (Weitz et al.. 1986). Meyer (2000) suggested that thi s process occurs unconscious ly as the result of imp li cit learning. perhaps because of fa ilures. For example. upon starting a job as a suit salesperson. a clerk might deal with each customer using the same approach-a simple "may I help youT-that may often lead to watching the cuMomer for a few minutes. commenting on the quality of the merchandise, and then waving goodbye without a sale. Eventually. though, that salespe rson is likely to learn that certain customers like to be left a lone and that hovering too much causes that type of customer to leave the store quickly. By experimenting with different approaches, the salesperson then learns to recognize other types of shoppers and to develop a repertoire of strategies for selling to them.
Chapter 19 • Soci(ll II/fluence in Selling Contexts
359
Previous research supports the view that selling skills are learned over time. For instance. a study of automobile salespeople by Gengler and colleagues (1995) found that. compared to tess experienced salespeople. those with more experience used more adaptive behaviors and differentiated morc between customers. VandeWalle. Brown. Cron. and Slocum (1999) found that salespeople with a learning goal orielltation (i.e .. those who would presumably seek to develop more detailed knowledge structures) were more successfulthan salespeople with a peiformallce goa/ orientalioll (i.e .. those who viewed challenging tasks as a threat and therefore failed to adapt) (see also Sujan. Weitz & Kumar. 1994). Thus. retailers shou ld expect new salespeople to improve their performance over time. To facilitate this process, retailers should motivate and help salespeople develop detailed knowledge structures about important features in selling encounters, With that in mind. what types of knowledge are necessary for successful selling? Previous research and theory (e.g .• Sujan. Sujan. & Bellman. 1988: Weitz et aI., 1986) suggests that to be successful salespeople need two types of information--declaralive and procedural knowledge. Sujan and colleagues (1988) explained: Declarative knowledge is the set of facts used to describe the category. whert.!pect of declarative knowledge is knowledge of traits. motives. and behaviors of the different types of clistolller~ encountered .... For sale ... people. procedural knowledge corresponds to knowledge of sales strategies to be used with each type of Cll!-.lOlllcr. (p. 82)
In other words, to be effective. salespeople require knowledge not only of the selling strategies but also of the various types of customers. We discuss these issues next.
Types of ClIstomers Weilz ( 1978) argued that the process of adaptive selling occurs through several stages. First. the salesperson forms an impression of a customer. Second, the salesperson selects a goal and plans a message to achieve that goal. Third. the salesperson communicates the message. Next. the salesperson evaluates the effectiveness of the message. Finally, either the goal is achieved or the first three stages are adjusted and the process starts again. Through a ll of this, it is clear thut forming an impression of the customer is crucial to the process of selling. To be sure. Weitl and his colleagues (Weitz, 1978, 1981; Weitz. Sujan, & Sujan, 1986) suggested that adaptive selling depends on salespeople's ability to organize their knowledge of and expcrience~ with customers. Specifically. through experience, salespeople develop "scripts" or stereo types about what typically unfolds during a selling cncounter. These scripts contain knowledge about different types of customcrs. As such. thcse scripts or knowlcdgc structures guide their behaviors in intcractions. Morcover. the morc accurate and detailed these sc ript s. the more successful salespeop le are. For example. research has found that the accuracy with which a salespcrson perceives customers and th e number of ways in which a salesperson differentiatcs customers arc related to that salesperson's effectivenes!>. (Lambert. Marmornstein. & Shanna. 1990: Sujan ct al.. 1988). Not surprbingly, then. in an efforl to help sale;people under;tand the types or customers
360
Pan IV • COlllexlljor Penllwion
they might encounter, a considerable amount of research has focused on developing different typologies of customers. The Illost basic typologies are based on demographic characteristic:... For example.
Goff. Bcllenger. and Stojack ( 1994) round that shoppe'" age. ge nder. and whet her or not they u,ed a "purchase pal" (a shopping companion) innuenced the degree to which they were per>uaded by salespeople. Specifically. males were ""ceptible to salespeople who used relational message~. older customer ... were ,usceptible to informational and relational messages. and customers who used a purcha\c pal were moM ,usceptible to recommenda-
tional wategies (Goff et al.. 1994). One of the fi"t typologies of reta il customers was developed by Stone (1954). who identified four categories of shoppers. Ecol/omic cOl/sumer.\ arc cautious shoppers, paying close attention to the quality and price of merchandise. Per.wm{lli~illg cOllsumers are intereqed in establishing relationships with ... alespeople. Ethical CIl_flOmers place moral va lu e ... above economic ones (e.g .. they would rather help the "little guy" lhan save money in a big department store). Finally. the
aparheric ('(m.wlIler doc:..n'( enjoy shopping and does it
only out of necessity.
Like Stone's (1954) typology. most of those that followed cmegoriLed custome" on the basi' or their motivation ror shopping (e.g" see Dawson. Bloch. & Ridgway. 1990: Gorr & Walters. 1995). According 10 Oa" son and colleagues (1990). all of these motivation ... can be classified into one of three general categorielo: product-oriented. experiential. or a combinarion of both product and experiential. In the fir~t case. a .,tore vi .. it i~ Illotivilled by purcha ...c need .. or the desire to acqllire prod· uct information. For in~tance. a COn\llll1Cr'S necd to find an anniversary gift within the next two hours wi ll produce ... trong product Illotivc. The second class of moti,e" in the typology ha., a hedonic or recrcillional orientation . ... Here. the attention is on store or mail vi ... ilS made for the plC<.ISllre inherent in the vi ... it itself. ... The last motive category combine .. product and expericntial elements, and occurs when the store visitor seeks to sa tisfy a purcha .. e need as well as have a pleasurable recreational experience in the oUllet. For example, a pen.,on Illay visit an outdoor equipmenllotore in order to purchase hiking gear. but also to panicipate in enjoyable conver.,ation ::Ibout hiking experiences with a ... ale ... pcr,on or other patron . (p. 410) Clearly. each of these motives affcch the ways shopper... respond to salespersons and innuence tactics. For instance. Goff and Walters ( 1995) found that recrea tional shop· pen., arc especially su,ceptible to salc..,per"oll influence, and Dawson and co lleagues
(1990) found that shopper> with product motives were signilicantly more likely to make a purcha ...c.
While each or the above-ment ioned typologies was developed by observi ng shoppers or asking shoppers about themselves. we find a final typology offered by Sharma and Levy (1995) especially interesting . because it was developed by going straight to the salc speople. These researchers asked 229 retail sa le speopl e to describe the dimensions
rh ey use to catego ri ze shoppers. Their responc"es fell into eig ht categories: (I) price! promotion-conscious shoppers (shoppers interested in buying products on sale). (2) need! product-based shoppers (who buy products for their own use and may need assistance),
(3) gift buyers. (4) browsers (shoppers who are "just locking" and do not want to be dis-
Chapter 19 • Socia/lnflLlence ill Selling Co/HeXf.\'
361
turbed by salespeople). (5) shoppers who need and seek sales help. (6) negatively labeled shoppers (a category that contains negative customer descriptions). (7) knowledgeable shoppers (who know what they want and don't need help). and (8) decision-style shoppers (who have their own way of making decisions).
Kllowledge oj Influellce Tactics In addition to developing a thorough knowledge of different types of customers. effective salespeople also have a large repertoire of influence tactics they can use to persuade their customers. The study of compliance gaining focuses on strategies and tactics aimed at getting others to do something or to act in a particular way. Traditional research in this area
sought 10 develop typologie> of compliance-gaining strategies (e.g .. threat. promise. deception. and so forth) by relying on theory (French & Raven. 1959), or by asking research participants to describe the tactics they might use in a variety of situations. Whatever the method. an enorlllouo;, number of tactics and typologies have been developed . Kellerman
and Cole (1994). for example, identified 74 typologies of compliance-gaining messages that they integrated into a "super" typology of 64 distinct strategies We have argued elsewhere that an alternative to this approach b to observe influence tactics that are used in actual encounters involving face-to-face interactions and to c~ltegorize the tactics on the basis of a finite set of po;,ychological principle ... underlying
why people comply with requests (see Cody & Seiter. 200 I). One viable set of processes was presented by Cialdini (200 I: \Oe also chapter 12 in this volume). Many of these processes were aJ...o covered earlier in this chapter. particularty the importance of credibility and likabi lilY. which can be established by creating images of tru<.,t and experti~e and w'Iing tactics such as appearing friendly or similar to the Cll!o.tomer, conforming opinion~. praio;,ing [he customer, and rendering favors. In addition to these processe<.,. this section brieny di<.,cus\es five others: (jcarcity. social proof. contrast. commitment. and reciprocity .
Scarcity.
People are often surpri!o.cd to learn that merchandise shortages are sometimes planned by retail storcs. Indeed, retailers know what research has shown for a long time: Scarce object~ are typically perceived a~ more de<.,irable or unique. and as a result people
are willing to p"y higher prices for thel11 (sec Lynn, 1991; Verh.llen & Robben, 1995). Moreover. if a product is percei\'ed a... scarce (e.g.. "II's the last one in your ~ile.'·). people may be more eager to buy it for fear of not having the freedom to do so in the future . Cialdini (2001) refers to this as the prillciple o/scarcity. In addition to shortages. the principle of scarcity can be used in other ways. For examplc. retailers often use restrictions that can assume several forms, which include limited time offers. limits on the quantity of a product that may be purchased. or a store purchafote of a certain minimulll dollar amountlO qualify for a good price (Inman, Peter, & Raghubir,
1997). Whatever the case. when confronted with the principle of scarcity, people may perceive the product more favorably and/or purchase it in an effort to ··relieve" themselves of the restriction ... placed upon them. Previous research sugge~led that scarcity is an effective sales tactic. For example. Inman and colleagues (1997. study 1) found that sales were twice as high when a restriction was in place than when it was not.
362
Part IY •
CmUCXIJ'/or Per.Hi{lS;OI/
When people are trying 10 decide what clothes they ~ hould wear or, for that matter. what behaviors are appropri ate. they often look at what o th er people are doing or weari ng. C ia ldin i (200 I ) labe led this form of innuence social proof"
Social Proof.
The tcndency 10 see an action a . . more appropriate when ot hers arc doing it normally works quite wel l. As a rule. we will make fewer mistakes by acting in accord wi th social evidence than cont rary to it. Usually whc n a lot of people are doing :-.omethin g. it is the right thing to do. This feature of the principle of soc ial proof is si nlllitan eollsly its major strength und its major weakness. Like the other weapons of influence. it provides a conve· nicnt shortcut for determining how to behave. but. at the same time. makes one who uses the ... hortclit vu ln erable to the attacks of profiteers who li e in wai t along its path. (p. 1(0) In selli ng contex ts, social proof b a commo n intluence tactic. "Best-sellers" and "top ten" lists are examples of ways in w hi ch rctailers try to se ll products by showing that other pcople arc also using them. Some salespeople are tra ined to let c ustome rs know whe n a particular product is "the season'~ hottest color," "what everyo ne is buy ing:' or "the most popular.·· In sho rt, social proof s.e ll s me rchandi se by claiming that whatever is popular is good. so custo mers should buy what's popular.
COlltrast.
Chapter 12 of this book discusses sequential influe nce tacti cs that include the contrasT effect. We briefly rev isit the concept here to talk about how thi s e ffect operates in sell ing COJltext~. The basic idea behind this e ffect is that when people are ex posed to so me standard amount of violence. beauty. prices o n co mmoditi es. temperature. happiness. and so forth. they become adjusted o r adapted to that "standard"' level. After this level is establi:-.hed. a con tra ~t effec t occ urs whe n something is judged again!>.1 th e sta ndard. For example. a lukewarm swimming poo l mi ght feel cold to someo ne who has just come from a hOI Jacu.lZi but hot to so meo ne e lse who had just stepped out of a snows torm. The co ntras t e ffect can be implemented in selling cont ex ts in a coupl e of ways. One approach. known as the door-;II-rhe-/ace tactic, involves makin g a large request and fol low ing it w ith a smalle r requ c!>.t. or course. the small er request is what the persuader wa nted all along. For example. a salespc r~on may know that a particular c ustome r wou ld never conside r buying a $ 1.000 eve ning gown. Even so. th e sa l e~ p e rso n might beg in by showi ng th e clistomer such a gown and as king if she woul d like to purchase it. Whe n the customer declines. th e salespe rson ca n the n sho w a 5500 gow n and ask for a sale. A second approach i~ know n a~ the top-dowll .m /n tactic (Donoho & Swenson. 1996). The idea is th e same as before. on ly thb time requests are not mad e. Instead. salespeople begin by ~ h ow in g the top-of-t he-l in e or highest-pri ced items as a reference point and then proceed by stepping down \0 lower- and lower-priced items. Regardle ss of the approac h lIsed, the contra st effect is th e underlying princ iple; $500-dollar items see m much less expensive when compared to $1.000 items. Compared to $ 1OO-dollar items, however. they may seem expens ive. The effect iveness o f these con trast tacti cs has been well documented o utside sales co ntexts, though according to Donoho and Swenson ( 1996), 1110" support for the effectiveness of the top-down approac h w ith in sales contexts is anccdotal. In thei r study. how-
Chapter 19 • Social h!/ll1(!IICe
itl
Selling
Coru€XlS
363
ever, Donoho and Swenson found that the top-down sales tactic was more effective than the bottom-up sales tactic (discussed below). The commitment principle sllggest~ thaI the more a person is com mitted to a group. cam,e. or idea. the more likely he or she is 10 stick with it. For example, once a couple make a commitment to gel married, buy an engagement ring. and send out wedding invitations. it becomes difficult for them to change their minds about gelting married-too much cognitive dissonuncc. The same is true in selling contexts. Once customers become com milled to the idea of making a purchm,c. it is difficult for thel11 to back out. Two commonly researched strategies that rely on the commitment principle are the !oot·;,,·the-t!oor taNic (a person agrees to a small reque ... t. which creatcs a comm itment 10 an idea. and then i" more likely to <.Igrec to a second. larger re(luc ... t) and the 1()I\'-/}1I11 taer;e (u per... on ugrees to a reque ... t or commits to a decision and Imer learns there were hidden costs involved in making such a decision) (for more on these tactics, scc chapter 12). We have suggested cl ...e\.\here that in selling context ... commitment call work in a variety of other ways (Cody & Seiter. 200 I). For eX3mple. once a cu ... tomer becomes committed to the ide<.l of buying a product. he or ,he can be shown more and more expensive items. Donoho and Swen,on (1996) called thi-. the holtom-llp sales tactic. Second. once a customer becomes committed to buying a product, the salcspcr,on might innuence that cus tome r 10 "complctc" th e purchasc with add-on itcms. For example. a shopper who originally came to purchase a "iuil might be pero;uaded to complete thc ensemble with the proper 'hirt', tie .... shoes, and so fOrlh.
Comlllitmellt.
Reciprocity. the principle
The expre ... sion "You scratch my back. and I'll scratch yours" epitomi7es reciprocity. Stated simply. this principle suggests that "we ... hould try to repay. in kind. what another PCN'II has provided u'·(Cialdilli. 200 I. p. 20). In other word .... if someone docs a faH)r or gi\cs a gift. the beneficiary feels indebted to him or her. As such. the beneficiary is morc likely to comply with the gift giver's requests, This principle i... at work when "free stuff' i\ given to L'u<.;t0111crs. For example. Hickory Farms and Sec's Candies frequentl) give customers samples of their products. Cosmetic departments allow cu'-.tomcrs to test perfumes and often hire cmployees to administer the sample for added pre;sure. Whatever the gift. such behavior may make shoppers feel indebted. Greenberg (1980) argued that the,e feelings make people feel uncomfortable by threatening their sense of independence. A ... a resuit, they become motivated to repay the debt. (~r
Adaptillg Tactics to Cllstomers A study by Sprowl and colleague, (1994) found that none of the four compliance gaining strategies reportedly used by real estate salesl>cople (denigration. aggrandizement, intimidation. and nnional appeal) led to ... ales sllcce~s. The researchers suggested that a possible explana ti on for thi ... is that successful salespeople not only know about different strategies but also how to adapt their strategy choice to different situations. Indeed, so far we ha ve argued that successful ,ales people have detailed knowledge structures regarding types of
364
Part IV •
CO/ltCX1J for Per.m(Jsiol/
shoppers and the types of strategies they mighl use to influence them. However. much research argues that detailed knowledge about customers and strategies may not be the only information necessary to be a successful salesperson. For instance WeitL and colleagues ( 1986) noted: We suggest thut, to practice adaptive se lling effectively, salespeop le need an elabonlle knowledge stru ctu re or sales silUations. sales behaviors. and contingencies that link specific behaviors to situations. To utilize thi s knowledge. sa lespeople need to be skillrul in collecting information about (:ustol11ers so that they ('an relate knowledge acquired in pre vious sales situations to the interaction in which they are c urrently e ngaged. (p. 176)
In other words, successful salespeople must be able to join knowledge about customers and strategies in order to adapt to specific selling situations (S ujan et 31.. 1988). For this reason, sc holars have argued that communication is essential to the selling process (e.g .. Sprowl et aI.. 1994: Williams & Spiro, 1985). Through comillunication, salespeople are not only able to develop the types of knowledge di sc ussed earlier, they are able to elicit the needs and wants of specific customers and adapt accordingly. A large body of research and theory supports the idea that successful sa lespeople are those who are able to recognize different types of customers and change their behaviors to match the situation (e.g., Levy & Sharma, 1994: Sharma & Levy. 1995; Sujan et aI., 1988: Weitz et aI., 1986: Will iams & Spiro. 1985). In light of thi s, a useful approach to understanding successful selling in retail con texts would be to examine the types of influence tactics that are effective for particular Iypes of customers. In the next section, we revi ew two studies that examine thi s and other i ss u e~.
Bringing It All Together: A Review of Two Studies In order to develop a better understanding of influence processes, Cialdini (1980) advocared a strategy called '"full-cycle socia l psychology." A central concern of this strategy is the identification of inlluencc methods used by everyday practitioners who are undoubtedly motivated to develop and refine intluence tactics in order to maximize career success. With this in mind, we published two studics (Cody & Seiter. 2001; Cody. Seiter, & Montagn e-Miller, 1995) based on observations of "real life" encounters between 416 shoppers and 416 commission sa lesc lerks in retail department stores. BOlh of the studies examined the use and effectiveness of Cialdini's principles of compliance discussed earlier (i.e., liking, credibility. reciprocity, cOlllmitment, contrast. social proof. and scarcity). We review these studie s here because they illustratc and extend many of the issues already discllssed in this chapter.
Study Olle As already noted, effective salespeople not only need to know about different types of customers and the different Iypes of inJlucllce tactics they might use but also about what types of tactics work best on which customers. Our first study (Cody et al., 1995) examined
Chapter 19 • SOcillllllflllellce ill Sellillg COIl/('UJ
365
this issue by having trained students observe and record the "itrategic~ salcsclerb used and how these strategies influenced different types of shoppers (for more detail on methodology. see Cody et al.. 1995). Shoppers were categorized in two ways. Fir\!. we examined whether malc.\ and females were influenced differently. Second. we looked at whether a customer's moti ve for shopping influenced thc way he or she responded to salesclcrks. The three types of shoppers we included were focI/sed shoppe" (who knew what they wanted to buy), recreational shoppers (who were "just browsing"). and gift buyer.\. Table 19.1 illustrates how various tactics influenced males and female\. The results compare purchase amounts depending on whether each tactic was used or not. As can bc :-.ecn, males spent more than females but were significantly influenccd by fewer tactics. Although women spent significantly more in re~ponse to virtually all of the tactic~. mcn were most influenced by reciprocity. commitment. and ingratiation tactics. Table 19.2 illustrates how various tactics influence recreational \hoppers. focused shoppers. and gift buyers. A~ can be seen. focused shoppers spent the mO!o.t money. fol lowed by gift buyers and then recreational shoppers. Moreover, although recrcalional ~ h oppers werc influenced by almost all the tactics, focused shoppcr\ were not significantly influenced by opinion conformity. similarity, helpfulnes~. contraM. and scarcity tactics. Gift buyers were significantly influenced only by opinion conformity and social proof. In general. the~e results indicate that the most effective tactics are tho..,e that mcet cus tomcrs' goals. For example. the typical male wants to buy objects as quickly and efficien tl y as possible. so the best approach is simply to Ill'lp him lind the desired object. praise his selection, and then recommend an accessory that goc!o. with his choice. On the other hand. gift buyers want to find an object that !o.omeone besides themselves will likc.
TABLE 19.1 A.'erage Amoullt (Dollars) Spellt by Male and Female Shoppers When Tactics Were and Were Not Used
Male Shoppers Tactics Li k i ng/i ngrat i'lt ion Prai-..e Opinion conformity Render favors Fricndline~ .. Similarity Credibility Helping/reci procity Commitment Contrast Social proof Scarcity ' Indicate~
SO/lree:
Female Shoppers
Not Used
Used
Not Used
Used
63.85 58.56 85.14 81.82 72.67 93.77 68.62 64 .09 74.05 96.64 81.80 90.55
127.32 156.84 141.32 127.35 122.83 117.34 116.70 132.49' 161.67' 113.89 139.21 139.04
45.95 44.71 65.23 58.70 57.82 65.24 45.82 45 .67 51.98 66.47 56.02 64.51
107.39' 122.34 109.70' t07.83 86.35 110.77' 10.1.16' 91.82 m .32' 115.71 110.53 105.55'
that the tactic was significantl), related to increased ,ales. Adapted from Cody. Seller. & Montagne-Miller. 1995.
366
Pan IV • COl/texIS for PerJ/{{Hio" Average Amolllll (Dollars) Spellt by Differellt Types of Shoppers Wlren Tactics Were and Were Not Used
TABLE 19.2
Tactics
Recreational Shoppers (N=105)
Focused Shoppers (N=256)
Gift Buyers (N=54)
Not Used
Used
Not Used
Used
Not Used
Used
9.21 18.14 19.06 16.68 22.92
109.42 I 32.s:1 101.96 42.37 106.43
62.46
141.~9
92.26 84.18 73.76 89.65
114.~6
121.46' 125.45 137.24
75.47 67.98 74.96 76.46 87.13
91.75 135.01 87.92 81.82 57.73
19.95 26.73
73.34 96.27'
76.35 87,77
115.90 140,25
69.07 70,70
91.79 108,97
4.40 6,92 19,04 26,78 17,60 19,71
31.28 78,91 124.75· 95.29"t99,81' 103.69'
98,82 7503 73.38
96,25 143,76 16 UX) 133,17 121.83 122,87
75.43
79,69 94,51 101.26 75,36 113.33 83.22
Li ki ng/i ngratiat ion
Praise Opinion conformity Render faVOf\ Friendlinc\~
Similarity Crcdihility Expertise Trust Hel pi ng/rec iprocit y
Helpfulness Leading help Commitment
Contra", Social proof Scarcity
92.12 84.43 90,60
6X.62
68.10 79,97 60.68 78,57
Indicate ... that the tactic .... as .. ignificantly rt.'!latl,!d In increa . . ed .. ale . . _ Source: Adapted from Cody. Seiler. & MOlllagnc-Miller. 1995.
Thus, the two tactics thai indicate a product is liked by others (opinion conformity and social proof) seem to be the mOM effective for thi~ type of cuMomer.
Study Two While our first study focused on types of customer:. and larategies. it said nothing about the ways in which different types of salesclerks affect buyer-seller interactions. Our second ;Judy (Cody & Seiter. 200 I) examined this i"lIc, Recall that earlier in thi~ chapter we described the four different styles of salesperson identified in this second study: ingratiation-style clerks. task-oriented clerks. clientoriented clerks. and passive-inactive derks. Once again. based on observations of real interactions between commission sale..,people and their customers. we sought to identify the types of strategies used by clerks and how effective each clerk was. Table 19.3 illustrates our results. As can be seen. reciprocity tactics were used 1ll0~t often by all types of salesclerks. followed by liking or ingratiation attempt .... The table also shows that client-oriented clerh used significantly more tactics than the other clerks and that ingratiation-oriented clerk, were characterized primarily by the
367
Chapter 19 • SOcilll l ll/llll'lIce ill Sdlillg COlllt'.\!'
TABLE 19.3
Sales Tactics Usel/ by Four Types oJ Clerks
Tactics Li kinglingr:.ttiat ion Crcdibilit) Helping/reci procil)' COlllmitment Contra ... Social proof Scarcity PurCh:hC ,!lnounl SO/mI'
Clie"tOriellted (N=20) 8.45 3.70 10.35 1.05 .lJ2 1.(Xl .45 $21.1.82
Illgratiator (N=68) 3AR
1.38 5 . 9~
.46 .13 AI)
.28 51~1.07
TaskOriellted (N=167)
Passive/ Illactive (N=161)
1.96 .90 3.40
.66 .28 1.81 .09 .03
.~9
.13 .34 .14 588.86
. 12
.10 $2058
Adapted from Cod} & Seill'r. 10()1
use of friendly and helpful tactics. Intereqing ly. these IWO active ..,ty les of sellin g. which a1 ... 0 resulted in the large..,t purchase amounts, made up only about 21 percent of all th e buyt!r-~eller imeraclion.., we ob..,erved (5 percent c lient-oriented and 16 percenl ingrmiation-oriented). Task-oriented clerks. who relied on high levels of helping tactics and moderate level ... of liking tactics. made up 40 percent of our ~ample: passive-inactive clerk .... who u.:;ed minimal helping tactic ... and no Olher I) pe of tactic. made up 39 percent of our ... ample. A:.. might be expected. ciienl-orienlcd clerk ...... old significantly more Illerchantli ... c than ingratialion-oriented c lerb-. ingratiarion-oriented clerks ... o ld ~ignifi cantly morc merchandise lhan task-oriented c lcrk:-" and ta ... k-oricnted clerks sold ~ig nificalllly more merchandi~e than pas~ive-inactive c1l!rk .....
Implications Whal arc the implicalion:.. of the ... e studic ... ? For tho\e intere\lcd in impro\'ing sales, our project ... \ugge"tthal ... ucces ... ful ... elling depends on clerks' ability to help shoppers achieve their goaI:-.. adapt to shoppers' I1l.!cds, and foslcr an image of likability. If a customer is :..hopping for a gift. the be~1 <.ipproach is 10 use ~ocial proof and opi ni o n conformity tactics to make a ... ale. If the c ustomer i... a focused or recreational shopper. the best advice is 10 be custol1lcf-orienlcd. u . . ing a wide variety of tactics in order 10 adapt to the customer's needs (Cody & Seiter. 200 I).
Con elusion The pril11~lry theme in thi ... rcvie\\ of socia l innuence and retail sa les is lhat in order 10 be effective .... ale:..people musl be competent communicators who are able to adapl to a variety of ... clling ... ituations. To this end ... ale~people must learn to juggle an assonment of goal .... which range frol11 making the . . ale in an appropriate manner to cretlt ing a c red ible and likahlc image. 10 building relation ... hip ... with cu ... tomers. We have sugge ... led Ihal much
368
Part IV • COf/tex(sjvr Persuasion
of thi~ ability is learned with experience over time and that with training. ~alesclerks benefit from enhanced knowledge struclures. These structures contain information about different innuence strategies and Iypcs of cllstomers that salespeople encounter. A key ingredient of effective sales is knowing which of these strategies work best with which type of customer. We concluded this chapter with a description of our own studies that have helped to address this issue. Although we have covered many topics here. we freely admit Ihal we have merely scratched the surface of the body of literature on selling. For example, most of Ihe selling tactics we discussed (e.g .. scarcity, liking, ~ocial proof) operate heuristically. innucncing customers without causing them to think Illuch about things like product quality or arguments regarding why they should or shouldn't purchase a product. Stafford (1996) suggested that many of the taclies discussed in this chapter are cffective but do not represent the full range of effects that can take place in response to sale..,clerk< selling attempts. For example. the scarcity tactic may have a negative effect if used on shoppers with a high need for cognition (i.e .. people who need 10 think things through carefully) (see Inman et al.. 1997). Moreover, Whittier (1994) ... ugge~ted that if customer ... interpret these lacties as manipUlative seiling ploys, negative effects may occur. and Stafford (1996) argued that when a buyer becomes aware that these taclies arc being used, it may cause him or her to scrutinile the salesclerk's messages. Whatever the case, such research highlights the importance of ethical and appropriate behavior on the part of salesclerks and. once again. the necessity of adaptability on the part of salespeople. Indeed. effective salespeople should be able to adapt to all types of customers. even those who do not base lheir purchase decisions on simple heuristics.
Referellces_ _ _ _ _ _ _ _ _ _ __ _ __ _ _ _ _ _ _ __ _ _ _ __ An'>t.!lmi. K .. & Zemanek. J. E. (1997). Relati()n~hip ,>clling: /low pcr~onal charactcri~lic" of ~alespeoplc ,tffcl.:l huyer '>atisfaclion. }olll"lla/ Socia! 1fl'1/(/ dOl" (lnd Pl,,..wl/(/liIY, 12. S39- S50. BcallY. S. E., Maycr. M. L.. Coleman. J. E.. Reynold<,. K. E.. & Lee. J. (1996). Cu:-.tol1lcr-sales a~socialc retail relation .. hip<,. Journal of Rl,tlIilill~. 72, 223- 247. Berry. L. L., & Para<,uraman. A. ( 1991 ). MtlrJ;.(' rillg\(' n ·ices: CO/llpi' rillg rhmll~h ljlUllin·. New Yor~: Fn.:c Pre",.;. Boorom. M. L.. Goobb). J. R .. & R,umcy, R. P. (199H). Relational communication trait" and their effect on adapti,cne<,<, and :-.ale ... performance . ./olll"l/a/ (~f the Academv (~rMarJ..erll/g Scil'Iln'. 26. 16-30. Bole'>. J. S.• Johnson. J. T.. & Bark ... dak:. II . C. (.2000) . HO\~ <,ale"f.lCnple huild qu:tlit) relation~hips: A rcpliGuion and extension. JOflrnal (~f /J'Hil/i'.B Rt'.Il'(m-li. 48. 7S-R I. Burgoon, 1. K. (1994). Nonverb~l ~ignaJ<,.ln M . Knapp & G. M. Miller tEd".). Uwullwok (~IiI1fI'IJ}(>f"\"(m(l1 COII/IIIIII/icarinll (2nd ed., pp. 2292RS) . Tho\J~and Oaks. CA: Sage. Bu"ch, P., & Wilson, D. T. (1976). An experimental ~nalysi,> ofa " expert and rererent na<,e .. of ~ocial power in the buyer <,eller dyad. JOlll"llal (~rMarJ;.elillg, 13,3- 11. Canary, D. J., Cody, M. J., & Manusov. V. L. (2000). IlIferper.w}//a/ ("0/1I/1llI11inlli()lI: A gnals-ha.\"('(J lIt>prO(wh (2nd cd.). BO'>lon: Bedford/51. Martin'". Chow. 5 .. & Iioiden. R. ( 1997). Toward an under"tamJing of loyalty : The moderating rok of trust. JOl/rnal of M£IIw~eriall.'iSljes. 9. 27S- 298. Cialdini, R. 8. (1980). Full-cycle social p"ychology . Applied Social P~ychol(J~y Allllllal, J. 21-...45. Cialdini. R. B. (200 I). Illflflence: Scit'flce amI practice 14th cd.). 80<,1on: Allyn & 8;I\.:on. Clark. R. A.. & Delia. J. (1979). Topoi and rhetorical competence. Quarterh JOllrnal (~r S,wech. 65. 187-
or
206.
Chapter 19 • Socialllffluence ill Selling Contexts
369
Cody. M. J .. & Seiter. J. S. (2001). Compliance principle.:; in retail sa le" In (he United Statr.:,,_ In W. Wosinska. R. B. Cialdini. D. W. Barrell. & J. Reykow\ki (Eds.). The practice of,meia! illfluence III multiple Cllltures (pp. 325-341)_ Mahway. NJ: Erlbaum. Cody. M. J .• Seiter. J. S .. & Montagne-Miller. Y. (1995). Men and women in the mark!!t place. In P. J. Kalbfleisch and M . J. Cody (Ed\.). Gel/der. pOl\'er. amI comltllwicllIioll ill imerper.fOlloi rdOI/OIlships (pp. 305-329). Mahway. NJ. Erlbaum. Cormtock.1.. & I-liggins. G. (1997). Appropriate relational mc\sage~ in direct selling interaction: Should !>alespeople adapt to buyer,' communicator style? Joumal of Bus;IIeJS Commlllliwtioll. 3-1.401418. Cro ..by. L. A .. Evan ... K. R.• & Cowles. D. ( 1990). Relati on"hip quality in lien'ice .. selli ng : An interpcr...onal influence per<.;pcctl\e. Joumal of Marketing. 54.68-81. Dawson. S .. Bloch. P. H .. & Ridgway. N. M . (1990). Shopping motive". emotional state ... and retail outcome,!;. JOlll'llal of Rt!wiIiIlR. 66. 408-427. Dillard. J. P. (1990). Primary and secondary goals in inlerpersonal influence. In M. J. Cody & M. L. Mclaughlin (Eds.). The psychology of /(Ictleal cOl1llmmicat;on (pp. 70--90). Clc"edon, England: Multilingual Mauers. Dillard. J. P.. & Solomon. D. H. (2000). Conceptualizing context in me"sage production research. ComImmicatioll Theory. 10. 167- 175 . Dion. P.. EaMeriing. D .. & Miller. S. J. (1995). What is really necessary in ~ucces.. ful buyer/sellcr relationships? Journal of MarkelmR Mallagemem. 24. 1-9. Donoho. C. L.. & Swenson. M. J. (1996). Top·down versus bottom-up ~ales tactics effects on the presen tation of a product line . JOllnwl of Business Research. 17. 51-61. Dwyer. R. R., Schurr. P. H .. & Oh. S. (1987). Developing buyer-~eller relatioll"hip .... JOllrnal of Markelillg, 51.11-27. Evan!\. F. ( 1963). Selling a<; a dyadic relation"hip: A new approach. Amerinm Belumoral ScielltiH. 6. 7(r 79. Fine. S. H.. & Gardia!. S. F. (1990). The effcct~ of self-monitoring and !>imilarity on ~alc~person inference proccs!>. Jotlmal of Per.wmal Sellillg and Sale!) Mwwgement. /0. 7-16. Fine, L. M .. & St:humann. D. w. (1992). The nature and role of sa l e~pcN>n perception ... : The Interactive effects of sale'lper.. onlcustomer personalities. JOImwl of Confllmer P.~ycll(ll(Jgy. I, 285-296. Ford. W. S. Z. (1999). Communication and customer service . In M. E. Roloff (Ed.). COII/llllmicalioll yt'arbook 22 (pp. 341-376). Thousand Oak. .... CA: Sage. French. J. R. P.. & Raven. B. (1959). The ba:.es of ..ocial power. In D. Can wright (Ed.). Smdies ill social ponter (pp. 150--167). Ann Arbor: University of Michigan. Imlilule for Social Research. Gade!. M. S. (1964). Concentr~Hion by sa lcsmen on congenial prospects. Journal of Marketing. 28. 6466. Garbarino, E.. & John\on. M. S. ( 1999). The different role ... of satisfacLion. trust. and commitment in customer relationship.... Journal of Marketillg. 63. 70--87. Ga ... ". R. H .. & Seiler. J. S. (200:\). Persullsioll. weilll illfluellce. and compliance gail/il/g (2nd cd.). 80"Ion: Allyn & Bacon. Gengler. C. E .. lIoward. D. 1.. & Zolner. K . (1995). A per~onal construct analysis of adaptive ...elting and ...ales eXJXricncc. P,fydwloRY (Illtl Marketillg. 12. 287-304. Goff. B. G .. & Walters. O. L . (1995). Susceplibility to ~alespersons' influence and eonliumers' <;hopping oriental Ion ". PsydwloglclIl Reports. 76.9 15-928. Goff. B. G., Bellenger. D. N.. & Stojack. C. (1994). Cues to consumer su ...ceptibility to salespe~on influence: Implication.. for adilptive retail '>Clling. Journal of Persollal Sellillg III1tI Sales Managemem. I~. 25-39. Goffman. E. ( 1959). The presellllltioll of self ill e\'erwlay life. Garden City. NY: Doubleday. Gordon. R. A. (1996). Impact of IngraliaLion on judgments and evaluation .. : A meta-analytic investigation. JOIIl'llal of Pt!/,w)//a!iry and Sodal Psychology, 7 J (I). 54-70. Greenberg. M_ S. (1980). A theory of indebtedness. In K. Gergen. M. S. Greenberg, & R. Willis (Eds.). Sodal exchange: Adwil/ces ill theory alld research (pp. 3-26). New York: Plenum . Greene. J O. (20CKl). Evane~cent mentation: an amelioralive conceptual foundation for re"carch and theory on message production . CommlllliclIIioll Theory. 10. 139- 155 .
370
Part IV • C()lITe.II,I.!Or Per.I/lwioll G rilTin. J . (1995). Ct/\{Olllcr lowl/I H (m- ffI ('am il. 1/01\-10 f.;l'('fI il. t\e ..... York Lexington Boob. Gwinner. K. P .. Gn.::ll1kr. D. D .. & Bitner. M. J. (1998). Rclmiolwl bencfil:.. in :..cr\'icc . . IndLl~tric _.. : Thc cll~tolllcr·~ pcr"pcl.:li\c. j()lIl"IIal oil"e "cadell/I' orMarkelim: Sciem·l'. 2(,. 101-11-1 Inman. J. L Peter. A. C. & Ra~hllhir. P 119(7). Framing the lI('al' The role of re~lrictiOll~ III accenlllilllllg deal value. journal 01 Cmlll/Iller Rl'Ii'arch. 2-1. 6H-79. Jonc:... E.. Moore. J. N.. Stanl:tml. i\, J. S .. & Wyatt. R. A. J. (19lJX). Salc~pcr'.(ln racc and gender and Ihe aet:e ..... and lcgllHllcr... nn n:lali()n~hip-; III retailing . hl/lnllli (!/Rl'flIiling. 75. 11-.12. Re)nold-;. K. E.. & Be.llt). S . l:. (llJl)l)b). ,\ n:~lat l (IIl:-.llIP l.'U,[Orllcr 1)J)!.)log~ Journal 01 R(,/clilj,l.~. 75. 50lJ-5:!1. Schurr. P. J I.. & Ol
or
71.71-81. Spi tzbcrg. B. II .. & Curacll. \\.]{, (1989). II{lmlhook ollllll'rflenO/lul ("OlI/p<'t/'llf'(' rl'H'lIl"ch_ Ne\\ York : Springer- Verlag . Sprowl. J . P .. Carvl'th. R .. & Senk. M. (19')4). TIIC clTec! ofcolllpliance-!,!aining ... tralCg) (;I\{)icc and (;0111IllUni(
Chapte r 19 • Sociall llfluen ce in Selling Corltexts
371
tion on the social psychology of city Stone. G. P. (1954). City shopper:-. and urban identification: ob~erva life. Americ all jOllrl/al of Sociolo gy, 60. 36-45. ces in ingratiatory behavio r: An investiStrutton . D.. PelIOn. L. E.. & Lumpk in. J. R. (1995). Sex difreren of Busines s Research, 34, 35gation of influcn ce tactics in the ..ale~pcrson-cu"tomer dyad. Journal 45. e dirreren ces between more effective Sujan. II .. Sujan. M .. & Bellman. J. R. (198R). Knowlc dge structur 81-86. h. and Ics.. crfecti\ e ..ale .. pcopl~ . JOl/rnal of Market ing Researc 25. smart, and erfective selling. working Sujan. H.. WeitL B. A .. & Kumar. N. (1994). Learnin g orientat ion. jOllrl/al ofMark etillg. 58. 39··52. tru<.;t in the salesperson: An integrative Swan. J. E.. Bowers . M. R.. & Richardc;;on. L. D. (1999). Cu .. tomcr sr Research. 44.93-107. review and meta-an alysis or the empiric al literature. jOl/rnal of BII.
369- 387.
ance and undcr..tanding custom er deciWeitl. B. A. ( 1978). The rclation .. hip between 'iale~person perrorm :-.ion making . journa l (1 MarketillM Re,\ellrch, 15, 501 - 516. n-::y frameworK. Journa l of Marketillg. Weitl. B. A. ( 1981 ). Effectiv enes .. in sales interactions: A continge
45.85- 103.
management: A relalion!-.hip marketing Weit/. B. A .. & Bradfor d. K. D. (1999). Personal ~elling and ~alc~ Sciell("('. 27. 241 - 2."n. per~pecti\"e. jOllrna l of the Academ y ojMark elillg motivation. and adaptiv~ hehavio r: A framedge. Knowle . Weill. B. A .. Sujan. H.. & Sujan. M. (1986) llg. 50. 17+-191 . Markeli uf JOIll'll,,1 work for improv ing .. clling effectivenc!<.s. rcpre\;entatives' per3uasioll strategie~. Sale~ : .. tic .. heuri Whitlh:r. T. E. (199-1). Elicitin g cunsum cr choice JUllmal of Pt'/'\'OIW ! Sellillg alld Sales Managem(,lII. 14. -II - 53. ..ale<;pcr<,on custom er dyad. journa l of William s. K. C .. & Spiro. R. L. (1985). Commu nication .,tyle in the Market ing Researc h. 22. 43-1-+t 2. model of Interaction goal ... CommtlllicaWilson . S. R. (1990). De\elo pment and test of a cogniti ve rule~ liOI/ 1\-I(H1ogrtlf1h!>. 57. 81 - 103. Wortub a. T . (1991). The evoluti on of personal II/efl/. II. I- I:!.
~cJling.
JOllmal of Pu.wJl/a/ SellillM alld Sale.\ Manage -
Author Index Abe, J. A. A ...11 S. 324 Abend. T. A., Adam ... S .. 290 Adam." W. c.. "267 Afin. W.o 249. "252. 258. 260.261 Aho. L. In Aida. Y.. 321. 3D Aiken. L. S .. 295 Aj/en. L 33. -t6. 47. 53. oX.
n
79.117.296.297 Albarracin. D.. 50..1 Albert. S .. 173. 174, J06 Aleman. c., 19 I. 32lJ Allegrante. J. P .. 277
Allen. M.. 3K 226. 295 Altemeyer. R. A .. l.JS Allman. D. G .. 297 Alvaro, E .. 159 Amir.Y .. 49 Am ..el. A .. 305 Analol. K. W. E.. 267 Ander .. , S. L.. 322 Andl'r~en. J. F.o 170. In.
.len Ander1'oen. J.W .. 4~ Ander~en. K .. 95 Ander ..en. P.A .. 13.26. 116. 165.166.167. 16S. 172.
m
Ander<;on. C. M .. 125
A nder ..on. J .. 268 Ander-.on. L. R.o 268. 276
Ander<,on. R.. 331 Ander.,on. R. B.. 34 Anderson. R. 1.. 291 AnJrev...,. I. R.o 341 Andrews. J. C .. g 1,99 Anglin. K, A .. 358 Angras. W. S .. 297 Ansari. M. A .. 339. ;\42 Anselmi. K.. ;\57. 35~ Ansolabehere. S .. 273 Apple. W .. 169. 176 Applegate. J. L.. 3~9. 330 Areni. C. S .. 39. 79 Argyle, M .• 6 Aristotle, 4. 65. 95 Aron. A .. 330 Aron, E .. 330
372
Arom.on. E.. 3. 103. 152 Aron~on. J.. 50. 52 A~canli. K .. 21M A ... h. A .. 307 A,hcr, T .. 216 A .. ~ca1. H .. 270. 276 J.W .. 116 Aune. R. K .. 169. 175. 176. Atkin~on.
239.
247.2~8.~49.260
Au~tin.
L 102 Au,tin. J. T. 189 A\
A vtgi~.
S. K .. 325 T.A .. I 15. 124
Baa~ke.
K. T .. 342 Babbitt. P.. 35 Bachman. J. G .. 297 Backu\. 0 .. 3~2 Badger. E.E .. 124. 125. 127 Baer. P. E .. 279 Baker. L 277 Baker. S .. 292 Baker. W. E .. 49 Bakhtlll. M. M .. 330. 331 Ballard-Reisch. D. S.. 289 Banduril, A .. 19. 33. 271. 295 B;trb.dale. H. C .. 356 Barnicott. E.. 21,19 Barraclough. R .. 3~5 Barrett. D.W .. 50 Barron. K.E .. 52 Barry. B .. 337. 340. 341.
J-I5. J-I6 B,,\eheart. J. R.. 92 Ba\il. M .. 35 Ba~\ett. R.. 210, 218 B'I'.\ili. J. N .. 32 Bateman. T S .. 341. 345. 346 Ilauku\. R.A .. IIR. lit) Ba\'eJa~. J. B.. 13. 24K 251 Baxter. L. A .. 327. 330, 33 I Beaber. R.. 169 Beach. B .. 158 Beach. R. I.. 154 Beach. W. A .. 13 Beaman. A. L.. 213 Bearden. L.. 21. 123. 124 Beany, MJ .. 115.267
Be.my, S. E .. 357 Beaubien, R. A .. 291 Be.lUvai .... F.. 279 Bcau\
Scm. D. 1.. 68 Benoit. W. L.. 95. 99.100. 101. I ().I. 105. 106
Berg. J., 290 Bergen. A. . 158 Berger. A. A .. 6 Berger, C. R .. 45. 185, 189. 192. 194. 196.299.331'1 Berger. J.. \35. 136. 138 Bcrkowill. 1. M .. 226. 231. 233, 235 BerkowitL. N. N .. 95 Berlo, D. K .. 223 Bcrn~tein. D. A.. 277 Berry. L. L.. 357 Ber<;chcid. E .. 95 Be~t, D. L.. 134. 136 Be<;t, J. A .. 277 BctlJnghau!.. E. P.. [4. 19 BCllman. J. R .. 359 Beutler. L. E.. 229 Bewley. B. R .. 277 Be/jian-Avery. A .. ~O Bhatia. S .. 40 Bi(:kman. L.. 172 Biernat, M .. 135 Biglan. A.. 277, 296 Binik. Y .. 54 Bird ...ell. D, S .. 39 Bir"- '1'.. 141. 157. 158. 169, ]70. 291. 294
81"'3117. G.L.186 Bither. S. W .. 275 Bitner. M. J .. 357 Biter. G. Y.. 67. 83s
Black. A .. 25 I Blair, 1. A .. 39 Blake, J. B.. 272 Bland. J. M .. 277 Blank. A .• 193 Ble"'. H ..
n
Bloch. P. H.. 360 Block. C,J .. 134 Bloom. N.. '298 Blum-Kulka. S .. 192 BIUlmtein. P.. 343 Bock. D .. 102 Bochner, $ .. 102. 103
Bodaken, E. M .. 267 Bohm. D.. 331 Bohner. G., 72 Bob, J. S.. .156 Bond. G. G .. 295 Boorom. M. L .. 358 Booth-Butterfield. $ .. 58 Borcher<.. T. A.. 6. 14 Borgida. E.. 134. 137 Borland, R.. 157.235 Born ... tein. R.E. 48. 49 Sorrie. R. A .. 268 O()\ter. F. 1.. 16. 17. 57. 59. 77,79. 1[3.120.122, 123.177.227.228.300. J 19. 320. 340 Bo\trom. R. N .. 14,21.92 Oot\'in. G. L 277 Bnwen. S. L.. 158 Bower. G .. 74 Bowc~.
J. W.. 328. 329.
.130
Bowers. M. R .. 355 Bo)'d ... tol1.1.. 273 Boynton, G. R.. 39 Braddock. C. H., 307 Bradford, K. D .. 356 Bmdlcy, P. 1-1 .. 134. [39 Brady. R. M .. 175 Bra ... her .... D. E.. 38 Bratman. M. E.. 192 Breda. C .. 139 Brehm. J. W .. 229. 230. 266.
278. 279 Brehm, S. $ .. 229 Brekke. N .. 74 Brentl, R. 0 .. 346
Author Index Brickman. P.. 49 Brifiol. P.. 73. 8~
Bu:-.~.
Brock. 1'.
Butler. D.. 136 Butle r. L 267 Butner. 1.. 50 Buttne r. E. H.. 139 Byrne, D .. 235
c.. 67. 73. 8 1. 96.
98.99. 101 Brockner. L 170 Brockricde. W.. 5 Broka w. c.. 274 Bromer. P.. 32 Brommcl. B. J .. 321 Broned. K .. 159 Brook,. R. D .. 150 Broverll1an. D. M.. 13-' Brovernw n, I. K .. I ~4. 136 Brown, C. E.. 141 Brown. J. B.. 307 Brown . L. K., 293 Brown, P .. 188, 190. 199, 325.328.329 Brown. S. P.. 359 Brownlow, S., 6 Buber. M .. 331 Buddebcrg-Fischc r. B.. 133 Bui. K. T .. 320 Bukov:lc. J. L.,-l0 Bull, R., 167 Buller. D. B.. 20, 157.158.
169.173.17.1.175.176. 235. 239. 2-l0. 242. 245. 246. 247. 248. 249. 250. 251.252. 253. 255. 256. 258.259.260,261.299 Buller. M . K .. 158.299 Bulten ... G .. 232 Burger. J. M .. 21 1. 214. 215.
216.220 Burgc,s. D.. I ~7 Burgoon. J. K.. 20. 56. 166,
167.169.170. 173. 174, 175. 176.239.240. 2.1 2, 245,246. 2-l7. 248, 249, 250. 25 1. 252. 253, 254. 256. 258. 259, 260. 261, 267.29 1. 298, 299, 303. 322. 323. 332. 356 Burgoon, M.. 2, 6.14. 16.
19.20.36,45,56.122. 139.141. 149. 150. 15 1. 152. 153. 155. 157. 15S. 159, 163.235.265.267. 268, 272. 274. 277, 29 1, 29.1 , 299. 300, 303. 305 Bur]..c. K.. 16 BurIC'lon. B. R.. )40, 342 Bur", . C. S .. 11}3. 319 Burri " c.T.. 52 Burrough<;. N.F., 2 1 Busch. P .. 355. 356. 357
Bu shman. B. L 173. 230 Bu slig. A. L. S .. 248. 250. 252.258.260.26 1
D. M .. 3 18. 3 19. 320.
322
Cabm. 1.. 67 Cacioppo, J.. 15. 19.36.57.
58.66.68.70.73.74.76. 77. 7S. 79. SU2. 83. 95. 96.97.98.99.100. 10 1. 103. IW. 121. 175. 176, 177,210,2 18.270.273 CaL D .. 306 Caine. B .. 267 Calder. B .. 40. 66 Caldwell. D. F. 2 14 Calli<;tcr. M. A .. 156 Cameron. K. A .. 226 Camcron. L .. 293 Cmnp lin. G.. IR8 Canary, DJ .. 123. 12-l. 186.
340.345.354 Cantril!. J. G .. 188.339 Carbon ari. J. P.. 292 Carli. L. L.. 134. 135. 136.
137.138. 139. "0. 141 . 142.143 Carl smith. J. M.. 51, 52. 68,
103 Carlson. J. G.. 81. 29 I C •• ruso. C .. 54 Carveth. R.. 358 Ca . . ey. M. K .. 22~ Ca . . ey. S .. 70 Cashi n. J. R.. 33 Cataldi. A. E.. 319 Cegala. D.L 14. 19,21 Cejk:l. M . A .. 137 Cerko ney. K .. 298 Chacko. H. E.. 341. 345 Chaike n. S .. 32. 46, 57. 58,
66. 6~. 71. 74. 76. 8 1. 85. 95.96. 100. 10 1. 104. 134.172. 227. 265. 269. 270.273 I Chano .... liz. B.. 193 C hase. L. 1.. 150, lSI. 152. 267 Chebat. J-C .. 104 Chen. S .. 58. 59 Chesebro, 1. W.. 273 Chesney. M. A .. 303
Cho. H.. 229. 230. 232 Choo. T.. 103 Chovi l. N.. 25 1 C how. S .. 355 Chri slensen. A .. 322
Christie, R .. 345 Christopher, F. S .. 319. 32 1 Cialdini. R. B.. 5. 33. 5 I.
141. 166. 177.207. 208. 210.211. 215. 2 16. 217. 2 18.220.353. 356. 36 1. 362.363.364 Ciarella. 1.. 54 Cicero. 95 C i ~'na.
K. N.. 33 1
Clark . R. A.. 354 Clarkson. F. E .. 134 Clevenger. T.. 13.95 C lo re. G. L.. 72. 75 Clufr. L E.. 290 Clu'ls. P. A.. 298 Cobb. A. T .. 34 1 Cochmn . S.D .. 54 Cody. M. LI4 . 19.21. 123. 124.186. 187.191.318.
3 19.339.3' 0.342.345. 354.356.357.36 1. 363. 364.365.366.367 Cohen. A. R.. 266 Cohen. C. E.. 188 Cohen. G .. 50 Cohen. M .. 267 Coker. D. A .. 166.299 Coker. R. A.. 166.29 1. 299 Cole, C. M .. 213 Cole. T.. 181. 318. 3 19. 321.
36 1 Cole man. J, E.. 357 Comlldcna, M. E.. 240 Compton. R.L 49 Comstock, 1.. 248, 255, 26. 355, 357. 358 Congalton, K. J .• 18 Conley. T. M.. 31 Conno r, J. M .. 137 Connors. M . M .. 234
Cook. E. W.. 303 Cook. K .. 138 Cook. T. D.. 101 Coomb.... w. T.. 276 Cooper. L 50, 51. 52, 172 Cooper. M .. 14, 19.2 1. 93 Cooper. M. D.. 16 Copeland. C. L.. 141 Copeland, G. A .. 273
Corbett . E. P.L 2 Corbett . K.. 277 Corby, N. 1-1 .. 35 Corli<;s. R.. 55 Corlle liu.'). T .. 211 Cornell. K. R.. 49 C()~ta. P.T .. 11 5 Co,tello. S .. 172 Cotton. J.L .. 52 Coupland, J.. 199
373
Coupland. N .. 189, 199 Cousins. N .. 294 Cowan, G.• 318. 320. 325 Cowles, D., 356 Co,;, K. C .• 39 Craig. K. M.. 346 Craig. R. 1'.. 329, 330, 354 Crandall, R.• 49 Crane, E.• 267 Cretin. D., 289 Crocker. J.• 305 Cron. W. L. . 359 Cronen, V. E., 267 Cronkhite, G.. 13 Crook, C. K .. 324 Cro~by, L. A., 356 Cru!'Co, A. H .. 168 Cruz. M. G .. 199.3 19.326,
342 Cummi ngs. K. M. , 297 Cunningham, E.M., 124 Cupach. W. R.. 5, 3 18.354,
358 Cup.lch. w. D .. 167 Currie. 1.. 188 Currier. L 54 Dabb~,J.M .. 173, 174.227 Dabu!. A. J.. 141 Dallinger, J. M .. 120, 121. 197 Dance. F.E.X .. 13, 16 Danckert, 1.. 188 Danielson. K.. I 17 Darley. J. M .. 172 Da\cnny, J. E.. 279 d'Avemas. R. J.• 277 Da "i~. L. M.. 339, 346 Dawson, E .. 6 Dawson. E.J . 13, 19 Dawf\on. S .. 360 Dean. R.. 102 Deault. K .. 134, 136, 305 Debono, K . G .. 95 DeCesare. K .. 220 DcGooyer. D. H.. 33 1 Dejong. W .. 2 13 Delia. J.. 35-l Demaine. L. J.. 216 Dendy. L. L.. 330 Denning, V.. 149. 150. 155 Denton, R. E .. 6. 16. 19 DePaulo, B. M .• 248. 258 DerJega, V. J .• 172 Dc!lhaies. P.• 54 DeStena, D., 7 \. 72. 8 1 deTurk. M . A., 299. 300.
325.326 Devine. P.G .. 52 DeVries. H .. 293
374
Author h,de.\
Dholakia. R .. 66.100.102 DIBcranJini-;. J.. 138 Dir.:bon. P. R .. 40 DiCkll1ellle. C. C .. 35. 231, 292,293 Djjk~tra, J. L-IO Dljk... ml, M .. 293 Dillard, J. P., 13. 16. 19.20, 21.46,1-'1. 157.185, 186. 187. 1~8, IM9, 190. 191. 192, 194. 196, 197. 19K 199.201,203.223. 300. 322, 323. 330. 342, 354.358 Dillehay, R.C .. 47 Dillman, L.. 239. 256, 260 DiMaueo, M. R.. 171,294 Dion. K. K .. 140 Dion. I>.• 356. 357 Di-.chler. L 290 Dobo-;, J.A .. 115 Dolich. I. L 275 Domenech-Rodrigue/. M M.,279 Domino. G .. 229 DOIlOhcw, L.. 230 Donoho. C. L.. 362, 363 Doran. N., 157,300 Doran, N. E .. 141 DordeL E.. 330 Dorne'>, B .. 105 Dougherty, T. W .. 346 Duvidio, J. F, 141 Dowd. E. T.. 229 Doylc. A .. 140 Dreher. G. F .. 346 Drew. S .. 105 Dri-;kcll.J. E .. 134. 141 Driver, J.. 188 Dri\er. R. E .. 248 Dube, C. E.. 290. 297 Duffy, M., 289 Dunbar, J. M .. 290, 297 Dunbar, N. E .. 245, 246 Dunn, D .. 6 Dunn. K. F.. 31 H. 320 DUllning. E. 1.. 190 Dupont. L.. 6 Dwyer. K .. 290 Dwyer. R. R .. 356 Dziokon~ki. W .. 227. 128 Eag\y. A. H .• 32. 46. 57, 58, 81.96,9R, 134, 135, 116. 137,138, 142,143.265, 269, 270, 273 Easterling. D .. 356 Ebesu. A .. 246. 250. 256. 261 Edgar. T., 231,321 Edward~, J. D. ISO. 152
Ehn:nhalt. A.. 174 Ei,>cnberg. S .. 277 Ekman. P.. 240. 247. 248 Elder, J. P., 277. 279 Ellt,r. M .. 23 I Elliot. A.J .. 52 Elly~on. S. L.. 141 Elman. \-1., 140 Elny.S. V .. 169 Ely. T. K .. 140 Enll110n~. C A.. 231 Ende, J.. 307 Ender. R. A, ]J7 Endler, N.S .. 116 Eng~lrom. C. 167 Ephraim. R.. 292 Ep'>tein. L H., 298 Ep... lein. S .. 116 Eraker, S. A" 289 Ericbon. B .. 134 bhleman. S .. 23 I E"'>e ... , V. M .. 74 E\'an,. D. A" 185 E\'an~. F .. 196.357 E .... an~. K. R .. 356 E\'an". L. Moo 68 E\'an'. R. L. 277 Fabrigar. L. R" 46, 72. 79.
RI,85 Fagol. B. L 142 Falbe, eM .. 338, 3-11. 343 Falbo,T" 138, 139.319, 322,323 Farmer, S. M .. 345. 346 Faure, C, 320 F:llio, R.H" 48. 51 Fedor. D. 8.. 3-15. 346 Fcclc). TH .. 18 l~cing()ld. A .. 172 Fcldm:lIl. C. 249. 258, 260 Feldman. P. H .. Feldman-Summer.... $" 139 I:cnnell. M. L. 136 Fe<,hbach. S., 154, 229 Fc"tinger, L., 3. 5 I, 52. 6~, 306 Fether'itonhaugh. D .• 70 Figueiras. Moo 40 Fign. S. D .. 307 Filiatrauh, P" I ~ Fine, L. Moo 358 Fine. S. H" 356 Fink, E. 1... 321 Fink. S" 276 Finnegan. J. R" 292 Finucane. Moo 115 Fire'tone, I. 1.. 167 Fi-;ck. M. H" 135 Fi ... hbein, M .. 33, 46. 47, 5:1, 54.68.79,117,296
Fi ... hcr, D" 81. 177 Fisher, R, J" 47 Fi~ke, S. T, 134. 136, 305 Filch, K. L.. 194.325,327, 328,330.343 Fil.lpatrick. M. A" 300, 321. 3:!:!. 323. 340 Flay, B. R" 277. 278, 296 Fleck-Kandath, C, 293 Fleming, D .. 39 Fletcher. R. H" 289 Flint. L. 3 18 Flora, J. A" 34 Floyd. D. L. 293 Floyd. K .. 245, 24(i, 25:1, 258.259. 261 Foan. A. E .. 277. 278 Ford. R. S .. 135, 138 Ford, T. W" 135. 138 Ford, W. S. Z .. 357 Forga.... , J. p" 74 Fo,>chi, M" 135 Fos~. S. K.. 2, D 1 Foxcroft. D. R" 271.) Fragoso, A" 229 Frandsen. M. M .. 319. 321 Frank, C , 159 Frank, S. 1.. 229 Franklin. K. M .. 292 Fra
93.168.117.155
Ga.'ltil. L 54 Geis. F. L . 136, 345 Gengler, C. E., 358, 359 Gentry, J. W .. 358 Gerbner. G.. 13, 16 Gergen. K. 1.. 331 Germany, 1\1" 330 Gerrard. M" 139 Gewirtz. H .. 279 Giacalone, R. A .• 141 Gibbon .... F. X .• 139 Gib.<.oll-Robill',on. E" 167 Giles, H .. 175. 199 G ill ig, P. Moo 101 Givens. V.K" 226 Gla ... gow. R.. 277 Godbold. L. C .. 268, 272, 279.280 Godin. G .. 54 Goering, E. M., 3·W
Goff. B. G.. 160 Goffman. E., 325, 328. 354 Goldberg, M. E., 277 Golden. B .. 152 Goldman. M .. 168 Goldman. R" 76, HI. 98. I ()(). 10 I. I {).l Gold ... tein. M. 1.. 227, 228,
235 Gomes, M., 3 19 Goodman. C. 3 J Goodman, J. S .. 3-15 Goolsby, J. R.. 358 Gorassini. D. R" 213 Gorden. W. I., 125. 126 Gordon , J. R.. 293 Gordon, R. A., 356 Gomik-Duro-;e, M., 50 Gottlieb. N .. 277 Gouilloux, F.. 212 Gouldner, A. W., 217 GraL P., 306 Graham, J. W .. 279 Granberg, 0., 54 Grandpre, J.. 159, 245. 26 1 Green, J.. 72 Greenbaum. P. E.. 303 Greenberg, B. S" 100 Greenberg. J" 19 Greenberg. M . S., 363 Greene, J. 0 .. 103. 342, 354,
158 Greene. K.. 54 Greene, R.T .. 54 Greenfield. S .. 294 Greenwald, A. C. 268 Greenwald, A G" 50, 66. 67.96, 10 1.106 Gremler, D. D., ~57 Grether. L 167 Griffin. C L.. 2, 331
Author Index
Grillin. J .. 357 Gruarl..e, I ... J9
GroJner. M.. 2Y2 Gronbed... B. E.. 273 Gru".. , A. M .. 297. 298 Gruhcr, K. J.. J..JJ Guadagno. R. E.. 21 .... 216 Guerin. B.o 319 Guerrero. L. K .• 2-W. 25M. 260. 261
Gumenil.. W. E.. 2.1:\ Gulklll. T. B.. lt Guy. L 101 Gwinner. K. P.. 357 lIacka. W .. 187 Iladdod.. G .. 3S Iialer, C L.. 73 Ilagan. R" 142 Hagen. K. M .. 34 H.une,. M .. 3.'\ 1Ii11e, J.1... 54,122.16(-,. 22.'\.299. 303 Hak. J, S .. 18 J-I;tlc. S. I ... 3H lIaleta, L. L .. 92 Iiall. C.. 159 Ililll, D. K.. .'\2 I H,tll, J. A.. 143.289.294 Hilll,J. R.. 141. 157. 158 1t.11I. L. 0 .. .'\22 Hamburg. D. A" 277 Hamid. P.f' .. 49 lI ami[lOn. M.A .. 57.163 Hamill, H .. 167 I lammonLl. S. L.. 23 I ..'\21 Hample. D .. 120, 121. 197 H,In,cn, R. A .. 216 Han,cn, W. B.. 227. 279 1I.III,on. R W .. 294 Har..lI. H .. 219 Hardeman. W .. 229 Harden. L [9. 20 1I 'lnJcn. J. M .. INH Il arkCIl. L. S .. 277 Il arkll". S. G .. 73. 99 lI arkne ..... C. D .. [99 Hilrmon-Jonc~. Eo, 3. 51. 52. 53.74 Hanmh. R.. 95 Harper. :-J. L. 33H. .U9, 343 H.IITC. R .. 299 IlJrriman. S .. 172 I larri,. AM .. [35 11arri,. M. B .. 172 I larri'on. A. A .. 49 lI art. L., 298 lIanhcuck. A .. 105 Hartman. P. E.. 295 Hartwick. 1.. 53 IIM\(!}.1.. 100
375
lI arvey. S. M .. 322 Ha.,." R. G.. 95. 101 lI allield. E.. 190 Haugtvcdt. C. p" 57. 67. 76. 7H.HI.H3
Howell. J. M., 343 Huesmann, L. R.• 298 Hughes. S. 0., 141, 292 Hunsacker, F.G., 19
Johnson, H. H.. 100. 101.
HUfl'laker. F.. 14. 158.245.
Hau .. I..ncchl, D .. 77
260 Hunt. H. K .. 267. 275 Hunt. S. D .• 356 Hunler. J. E .. 47.53,57.
Johnson, R.T .. 124 Johnson-Cartee, K. S., 273
HawJ..in\. K. W.o 136 lIaye .... J. A .. 293 Hale!. H. 14 H,I/chon, V .. 167 H'llen. fl.!. D.. 13M. 139 l leath. R. L.. 27Cl Heather. ~ .. 29.'\ Ileavey. C. L.. 322 Hedge .... L. V...'\~ lI ee~ackcr. M. H .. 77 Hciby. E. M .. 291 Heider. F.. SO Heilrn'lIl. M.E.. 1.'\4 lIei ..cl. A.D.. liS IIci ...cr, R,A .. 52 lIehJrnan. C. 35 lIel1er. T .. 136 Ikller. W., 49 lIelman. C. G .. 23J lI eh\eg-Lar~en, M .. RI lIen<,1cy. W.E.. 117 lIen .. nn. R. H.. 2H9 lIenwuod, K.. 199 lIerington, A. D .. 125 Ilertog. J. K .. 292. 293 Ile,lin. R.. 267.27.'\ IIc ....c. F. \\ .. 39 IIcwgill. M.A .. 154, [55 lIeyLluk. R. G .. 49 Iliggin ... C. A .. 343 lIigglll'. D. S .. 319 lIigglll". G .. 355. 357, J5N 11111. D .. 227 Jtillman. J L..319.J20 lIimlllelfarb. S .. 66 lIiroka\\a. R. Y.. 337. 3.~8. 339.343.344 Hoh!"h. J. R .. J 85 lI oihcrg. E.. 232 Iiolbert. R. L.. 269. 272 I/oldcn. R .. 355 Iiolmberg. S .. 54 lIoltgra\e\. T., 134 lIomer. P. M .• 95 Honcycutl. J. M. 18H Iloobler. G .. 239 lIoro\-\-it7. I. A .. 233 Horvath. c.L.. I [9. 124 lIo .. ey, K.. 219 IloughlOl1. G .. I H8 HO\land. C. L. 31. 65. M, 76.91. 223. 233 Ilo\-\-ard. D. 1.. 35M Ho"ard. J. A .. :\20, :\43 Hlmcl1. C .. 81
122.163
102.103
Johnson. 1. T.. 356
Johnson, M. S., 355. 356 Johnston. O. D.• 14. 2 1 Johnston. L. 0 .. 297
Jones. A.. 290 Jones. D. c., 319
Hurd. P. D .. 277 Hu<,ek. T. R.. 100 Hyde. M.K .. 226 Hyman. R. B .. 292
Jones. E .. 354. 355. 356 JO"es. J. A., 289. 290. 306 Jones. J. L., 293 Jones. K.. 138 Jones, S. B., 139. 152. 153.
Iacobucci. D" 40 Ifert. D. E.. 21. 123. 194. 318 Ilardi. B. C, 136 Im:lhori. T. T . Jig Infante. D. A .. 10. 113, 114. IIS.116.117.IIH.119. [20. 124. 125. 126. 127 Inman. J. J .. 36 1. 36R Inr.ko. CA.. 66. 102. IOJ, 270 "Ier. L .. J25 lsocrate,. 95 huane. P .. 292 Iyengar. S .. 273 Izard,C. E.. 271. 3IM . .'\24 Ino, J .. 126
173 lonsen. A. R .. 307
Jacklin. C. N .. 138. 140 Jad\on, 5 .. 37. 3R. 342 Jachon·Walkcr. S .. 229 Jacobo;. D. R.. 277 Jawbs. 5 .. 37 Jacob}. LL. 49 Jahn. D. L.. 229 Jamico;on. K. H.. 133. 273 Jamner, M. S .. 35 Jrum. I. L. 3 I. 65. 91. 154. 223.229.266,306 Jani .. Lcw .. ki. C. A" [93, .119.330 J:UlL, N. K.. 291. 293 Jarrell, R.. 226 Janis. W. B. G .. 68 J:I\'omi<;ky, G .. JJ9 Jen..cn. A.D .. 172 Jcp<,on. C .. 227 Je .. <,or. R.. 277 Jc .. ~or. 5. L. 277 Ji:mg. J. 1.. 40 John. D. R" 325 Johnson. B. C .. 134 Johnson. 8. T. 54. 98, 143 Johnson. C. A .• 277 Johnson. D.W .• 124 Johnson. G. M .. 346
Jordan, F.F.. 125. 126 Jordan, J. M .. 320 Jordan. W. L 339 Jorgen<,en, C .. 167. 169 Jo-.eph. J. G .. 231. 232 Joule, R. V .. 51. 52. 212 Judge. T. A .. 346 Ju ... t. M .. 273 Kable. J.. 15 Kahle. L. R .. 95 Kahncman. D.. 78. 177 Kaid. L. L .. 273 Kale. A. R., 49 Kalichman. S. C .. 229 Kallgrcn. C. A .. 33, 37. 73. 81.134 KamlO ... M. A .. 270. 276 Kang, G. 1.. 13 Kang. 1. G .. 268 Kanter. R. M .. 34 I Kanwi .. her. N.. 188 Kao. c.. 76 Kao, F. C .. 12 1 Kaplan, R. M .. 100.279. 294.306 Kapoor, A .. 343 Karau. S. J .• 136 Karoly. P.. 289 Karlen. 5. 1.. 135, 136 Kashi ma. Y .. 297 Kasmer. J. A .. 78 Kasmcr, J.E .. 57 Kasprz) k, 0 .. 139 Kas<'irer, J. P .. 306 Kaufman. 0 .. 167 Kazeolas. D" 2 1. 123, 124 Kazi<,. L.. 307 Kearney. P.. 2 1 Keascy, C. B .. 172 Keffer. M. A .. 293 Keller, J., 167 Kellerman. K .• 5. 181, 1M9. 190.318.319.321.361
376
Allfhor Illdex
Kelley. C. M_. 49 Kelley,H_1I.31.65.91. 223. J 17 Ke ll y. G.A .. 152 Ke ll y. J. A.. 229 Ke ll y. P.. 188
Kol.. G_. 293
Kelman . H.C.. 47. 69. 296
Krapfc1. R. E .. 173 Krau"". R. M .. 169 Krau ... ~. S. J .. 47 Krcutlcr, K. J .. 279 Kroll. T.. 289 Kronc. K. J .. 34 1. 342 Kron ... hcrg. S .. 142 Kro~nick. J. A .. 32 Kruglan<.,ki. A. W., 59. 73. 78.1':1.82 Kumar, N.. 359 Kunkel. A. W .. 329 Kynga .... B.. 289. 290. 297
Ken nedy. K. A .. 101. 104 K en~ki. 1-1 . 11 36. 267. 268. 27~. 274 Kern. M .. 273 Ker>eli. M. w .. 277 Ke
c..
Kim . M.. 344 Kim. M. S .. 47. 53. 306. 307 Kim. M·S .. 190 King. A.
c.. 297
King. B. T .. 306 King. L. B.. 267 Kinne). T. A.. 115. 122. 192.198, 199. )30
Kipni!.. D.• 187. 199,320. 337. ))8. ))9. ).1. )43. 344.346,)47 Kirch lcr. E.. 322. 323 Kirk . L. M_. 291 Kirkcndol. S_ E.. 258 Kirkland, S. L.. 19 Ki rM:hl. J. P., 23 1. 232, 289, 297
Ki shorc. S .. 142 Kitc.M .. 134,136 Kiyohara.O .. 168 Klapper. J. T .. 232 Kle in. D_ J .• 71 Klein, G .. 40 Klein. L.. 277 Kleinke, C L.. 166, 167. 170.172 Kleinol. M C. 229 Klinger. E.. 185 KHng1c. R. S .. 56. 158.289. 290.293.298.2~).300.
303. 304. 305. )06. 307 Klahn. L. 5 .. 292 Klonsky. B. G .. U8. 142
Klopovich. PM .. 290 Knapp. M. L .. 240 Knower. F. H .. 31 Kobrynowicl, D.. 135 Kock. C .. 167 Kodama, R . A .• 338. 339.
343 Koerner. A. F.. 268 Kohn. M .. 186
Koper. R. 300 Ko~berg. R. L.. 115. 118, 119.124
Kouabenan. D. R.. 229 Koulad. D.. 102
Lachendro. E .• 292 LaclIliak. C. 81 LaAeur. S. J .. 137. 100m. 141. 143 Lake. R. A .. 39 Lamb. C. W .• 276 Lambcn. D. M .. 359 Lambkin. D .. 188 Langer. E. J .• 18. 193 Larouche. M .. I~ Lar-.on. C. U .• 14.21. 92. 93 LaNIIl. C. E.. 13 La ... ky. B.. 134 Lau. R. R.. 35 Lauterbach. K .. 319 Lave. L. B.. 292 Lay. K. L.. 124 I.e Poirc. B. A .. 169. 2lJ8. 299 Leap. C.J .. 117 Leatham. G .. 191. 329 Lcathe ..... D. G. 354. 355. )56 Leanlt. C.. 66. 100. 102 Lee. A Y .. 49 Lee. C .. 54 Lee. J .• 357 Lec. WOo. 268. 269. 271 Leech. G .. 199 Lcfcoe. N. M .. 277 LcFcHe. R.. 168 LcFleur. G .. 267 le ich lite r. J. 5 .• 33 Leigh. T. W.. 357. 358 Leighty. G .. 329. J30 Lcinb:lch. M. D.. 142 Lenihan. M. G .. 168 Lculhe~se r . L.. 357 Le\cn~lein. J. H .. 307 Le\cnlhal. H .. 154.227.293
Le\-m. N W .. 297 Lt'\mc. N . 158 Lninc. TR .. 54. 120. 123. 124.125.127 LC\lnger. G.. 298 Lev in ...on. 5 .. 188.190. 199. )~5. )28 Lcvin<.,on. W .• 307 I n itl . 5 .. 138
Levy. M.. 360. 36-1 Le .... i .... l .. L.. 305 I.e). P.. 294 Li. H .40
Lil"lcrman. A .. 58. 81 Lich,tclll. K .. 229 Lichlcn"'lein. E.. 277. 296 Llebrand . W. B. G .. 40 Lim , T .. 117. 125. 326. 32S. 329. HO. 345 LIIl , W.-K_. 269. 275 Lind. E. A.. 134 Lmden. f\'1.. 291 LmlJqul\I. B.. 158 Linimon. D.. 138. 139 LlIlkey. H. E.. 167 Li ....... B_. 167. 169 Li \tcr-Sharp. D .. 279 Lill . M D.. 298 LlI1 lcJohn . S. W .. 10.95 Lockc. E. A.. 338. 341. 342 LuckhcelJ. M. E.. 135. 13H Locher. C 137. 140. 142.143 I.ong, M.. 267
c..
Loop. K.. 229 Lurch. I: . P.. 230 Lord . K. R., 40 Love, c., 279 LO\c, R. I: .. 268 I _m-ette. 5 .. 345 Lov.e. G_. 279 LOMc)". T M .. 81 Lucpl.cr. R. V .. 277 1_umpkin. J. R.• 356 Lum ...daine. A. A. 65. 266 Lund . F. H .. 3 1 LUll . R. J .• 79
Lynn. M.• 361 Macaui:IY. J R..268 M:lccoby. E. E.. 13K 140. l24 Mac(.·oby. N .. 277 Machado. L.. 188
Macbl cr. E.. 187 Macinto~h.
G., 358
Mad.ie. D.M .. 58. 59 Mad" lchlan. J., 169. 176 MacNeill. L.. 172 Magnu ......on. D .. 116
Mahe.,v.aran. D .. 58. W. 76, 104 Mahoney. J. M .. 168 ~laibach. E .. 34 Maiman. L. A .. 290. 291 Mainiero. L.. 343 Maio. G. R.. 74 Makhijani. M. G .. I.lR. 142 Mallalicu. L. 3 19, .n o Mallon. D.. 54 Malone. K.. 169 Malotte. K .. 227 Mandel. J. E.. 267 Manlre. S. H.. 141 Mani ... M .. 272 Mann . T. E .. 273 Manneui. L.. 229 Manrai. L. A .. 193. J II) Manuso .... Y .. 166,354 Marin. B. Y .. 233 Marin. G .. 233 Marks. M.. 229 Marlatt. G. A .. 293 Mamlom~tein. H_. 359 Marshall. A. A .. 307 Marshall. L.. 3 19. 326. 342 Man-ton. P. J .. 342. 345 Marleau. T. M .. 40 Manell. D.. 223 Mancil. R.• 134 Manlll. L. L. 72 Martin. M. M .. 125 Maruer. P.. 188 Maruyama. G.. 134. 169 Man.~· eJI. G .. 181. 320. 321,
339 Mas lyn. J. M .. 345. 346 Massey. J. E.. 3K Mal'iChiner. M .. 140 Mattson. M .. 292. 21)4 Mayer. M. L.. 357 May'i. Y.M .. 54 Manuca. S.A .. 297
McAli.,lCr. A. K.. 277 McCami-.h. M .. 297 McCaul. K. D.. J5. 277 McCornack. S. A .. 246. 354 McCracken. E. (' .. 307 McCrae. R. R .. 11 5 McCro~key.
J. C ..
~.
96.
115,124.267. 3W. 346 McCro~ke). L.L., 124 McDonald. D. A .. 32 J McEnall). M.. 139 McGi nley. A. 16H McGi nley. H .. 16K McGinni cs. E.. 100. 10J McGrath. M. A.. 193.319 McGralh. M. P.. 141 McGra\\. P. F.. 357. "~5K
Aurhor Index
McGuire. W. J.. 165. 266, 268. 270, 272. 276. 277 McHugh. M. C .. 32 1 Mc Kclvey. R. S .. 279 McKeon. J. K.. 216 McLaugh lin. B .. 324 McLaughlin. M. L.. 3 18.
))9. )40 McLeod. D. M .. 141 McLeod. J. M.. 223 McMahan. S .. 226 McMahon. A. M.. 136 McNamee. S .. 33 I McPeek. R. W .. 150. [52 McWhinney. I.. 307 Meeker. B. F.. 136. 137 Mehamed. B. G. Mehrabi:lII. A .. [68.299 Meilman. P. W .. 33 Mei<;chke. H., 190 Melamed. B. G .. 303 Melcher. C. 158 Mcri<;ca. R .. 289 Mertz. T .J .. 171 Meryn. S .. 290 Me ....~ari'i. P.. 39 Mel«;, S .. 3 1R. 329 MClLler. A.. 10 1 Mewborn. C. R .. 229 Meyer, G .. 213 Meyer. J.. 226 Meyer. J. A .. 189. 194 Meyer, J. R.. 358 Meyers. R. A .. 339 Meyrowitz. J .. 273 Mickey. J.. 3:'17. 343 Mtle~. R. H.. 346 Milgram. S. 173 Miller. A .. 54 Miller. c.. 159 Miller. D. T.. 33 Miller.G.R .. 2.6. 13. 15. 22.92.93.100. 113. 149.
[50. [54. [55. [H2. [85. 255.165.267.272.277. 299.3 19.325.340 Miller. J. A .. 2 10. 218 Milkr. M .. 267. 297 Miller. M. 0 .. 320 Miller. N .. 169. 176 Miller. S. J .• 356 Miller. T. Q .. 296 Millm:m. R. B .. 277 Milb. J.. 3. 51. 52. 53. 74. [ ()()
Milne. A.A .. 9 Milne. C. R .. 229 Mineo. P.. 166 Miniard. P. W .. 40 MintLberg. fl .. 337
Mitchell. A. 275.176 Mittlemark. M. B.. 277 Miura. S .. 338. 343 Miyahara. A .. 337. 33R. 344 Mladinic. A .. 134. 136. 137 Moadel. A.. 292 Mohr. D .. 219 Monahan. J. L.. 231 Mongcau. C .. 54 Mongeau. P.A .. 15.57.59. 79. 121. 122. 177.227.
22H Munin. B.. :B Montagne·Miller. Y .. 364.
365.366.367 Montano. D. E .. 139 Montgomery. B. M .. 330 Montgomery. C. L. 267 Montgomery. R. P. G .. 292 Montgomery. S. B.. 231. 232.232.297 Moon. Y .. 40. 17 1 Moo re. D. L.. 77 Moore. J. N .. 355 MonHl . P. 167 Moreland. R.L.. 48 Moreno, H. 229 Morgan. M .. 229 Morgan. R. M .. 356 Morrill. J.. 268 Morri~. D .. 168 Morri~. K. 1.. 73. 81 Morri~on. J.. 40 Morri~on, K.. 227. 22~. 231. 234 Mo~kowil/. M. A .. 307 Motley. M.T .. 1J Mowd .. y. R. T .. 338. 339.
.145 Moyer. R. J.. 95 Mueller. J. S .. 322. 323 Muellerleile. P.A .. 54 Mullen. B .. 171 Mullett. J .• 25 I Mura ... hima. F.. 102 Murnen. S. K.. 140 Murphy. K.. 279 Murphy. S.G .. 49 Mycr~. D. G .. 134 Naigles. I.. R. 141 N
Newcomb. T.M .. 50 Newel l. S.E .. 187 Newhagen. J. E .. 273 Newsom.J. T .. 2 15 Newton. D. A.. 322. 323.
332 Nicotera. A.M .. 124 Niles. P .. 154 Nitl. M .. 6. 13.36.268 Noe lle-Neumann. E.. 274 Norcro!o.~. J. c.. 231 Nordstrom. B.. 189 Norell. S. E .. 290. 305 Norman. R.. 66 Norman. R. Z .. 135 Norri .... E. L .. 268. 274.276 Nonon. R. W .. 125 Noth~tine. W. L.. 14. 19.21.
93 Novack. D. H.. 290
O·Barr. W. M .. 134 O·Brian. K., 2J I O' Byrne. D. 1.. 277 O·Connell. E.J .. 49 O'Donnel. J. F., 290 O·Hair. D. J .. 21. 345 O·Hair. H. D .. 21. 342 O' Keefe. B. 1.. 189,354 O'Keefe, D. J.. IS, 17. 19.
21.36.37.38.79.81. [00 O'Malley. P. M .. 297
O·Ne'l. E.
c.. 277. 346
O;ltS. R. G., 34 Obenynski. M. A .. 73 O·Donne li . V .. IS Oe'>termeier. U.. 39 Offerm .. n. L. R .. 343 Oh. S .. 356 Older-.haw. L.. 32 1 Olrn-Shipm:lI1. C .. 136. 142.
[43 Olm:o.tead. B.. 134 Ohon. J. M .. 2 13 Olson. K .. 229 Onyekwere. E.O .. 11 6. 11 8 Ornstein. N. J.. 273 O' Rourke. A. H.. 277 O· R ou r~e. T. W .. 277 O~g()("ld. C.E .. 50. 102 O:-.hikawa. S .. 228 Oskamp. S .. 49 Osterhouse. R. A .. 98 O~lrom. T. M .. 67 O~ lru W. D. G .. 231. 232 Otero-Sabogal. R.. 233 Otto. S .. 137 Owen. N .. 54. 232 Oianne. J. L.. 355
377
Palan. K. M.. 325 Palmer. M. T .. 198 Pahncr-Shevlin. N., 290 P:llmgreen. P.. 230 Papageorgi~. 0 .. 266. 269. 270. 276 Par",>uraman. A .. 357 Park . H. S .. 54. 306 Park. K. A .. 324 Park~. M. R.. 246 Parp.. 1. M .. 324 Parrott. R .. 158.29 1, 299 Pa:.ha. N. H .. 269 P
Perot. R.. 15.21 Perreault. W. D .. 346 Perrien. J .• 104 277 Perry. Persc. E. M .. 141 Peter. A. C .. 361 Pe terson. P. D., 102 Peterson. T.R .. 226 Petraitis. J.. 296. 297 Petty. R. E .. 15. 19.32.36.
c..
37. 46.57.58.66.67.68. 69.7 1. 72. 73. 74. 75. 76. 77. 77. 78. 79. 80. 8 1. 82. 83.84.85.95.96.97.98. 99. [00. [01. [03. [04. [21. [75. [76.177. 2[ I. 270.273 Pfannenstei1. D. A .. 168 Pfau. M .. 13, 15.2 1,36.
158. [69. [70.265.267. 268.269.270.275.276. 277.278. 279. 280 Pfeffer, J.. 337 Phillip. J.. 292 Phillip:.. D .. 172 Phillips. G. M.. 289. 290.
306
378
Author /ndex
Pickering. B , A .. 39 Pie rro. A .. 229 Pilkoni.!., P. A.. 66 Pill, R. , 290 Pinkleton. B. E.. 273 Pillman. T. S.. 354 Pius. M .. 289 Plato. 2 Plax. T.G .. 21 Pletcher. B. A .. 276 Plugge. CD .. 226 Podshadley. A. G .. 297 Poppe. M .. 3 18 Popper. T .. 325 POller. J .. 292 Powe ll. F.A .. 154 Powell. J. L.. 168 Power. T. G .. 141 Pratt. c., 232 Prei sler. R. M .. 37. 73. 81 Prei ss. R. W .. 38 Prentice. D. A. . 33 Prentice-Dunn. S., 293 Presley, C. A .. 33 Press, 5 .. 229 Pressman. B ., 170 PresIo n, M.. 213 Pribram, K . H., 185 Price, H., 40 Pride. W . M ., 276 Priester. J. R .. 73. 79, 8 1 Prince, L. M .. 294 Prochaska, J. 0 .. 35. 231. 232 Propp. K. M .. 138 Pryor. B. , 266, 267,17 1 Puckett. J .• 81. 177 Pugh. M. D .. 139 Pyszczynski. To, 19 Quinlan. K. B .. 35 Raden, D .. 32 Raghubir. P.. 361 Raia. J. J.. 290 Raines. B. E.. 277 Rajecki. D. W.. 49 Ramage. K., 138 Ramesh, C. N .. 246 Ramirez, A. , 245 Ramsey, R. P.. 355, 358 Raneer, A. S., 10. 11 3, 114. 115.116.11 8.1 19.1 20. 124, 125.1 26 Rao, N., 326, 342 Rauner, S., 220 Raven, B, H.t 296, 320, 337. 361 Raymond. B . J., 172
Reardon. R .. 319 Redfield. J.. 49 Reed. M .. 220
Ro~enthaL R.. 38 Roser. C. 9& Ro<;kos-Ewoldsen. D. R.. 32 RO'ino\\. R. L.. 3'20 RO'i"itt~r. C. M.. 91 ROIcr. D. L.. 189. 290. 291
Reed. Y.. 133 Reeves. B .. 35. 273
Rothert. M. L.. 305 Rothman. A. J.. 35
Re ic hert. T .. 156 Reno. R. R.. 33
Rotter. J. B .. 345 Rouncr. D., 279 RO/c ll e. R. M. , 277
Razran. G . H. S .. 72 Reardon. K. Ko, 15. 19.92.
123
Rey nold~. K. E .. 73. 357 Rey nolds. L. A .. 293 Reynolds. R.A., 123 Rhi ne, R. J.. 100. 103 Rhoades. M. J. R.. 140 Rhode. D. L. 133 R ic h a rd ~on. L D.. 355 Richman . S. A .. 72 Richmond. V. P.. 12"'.339. 346 Riddle . B.L. 119 Riddoch. c.. 35 Ridgeway, C. L , 1J6. 138. 139. 142 Ridgway, N . M .. 360 Riess, M .. 267 Riordan. C. A .. 141 Rippetoe. P. A .. 293 Roach. K . D .. 116. 117. 118. 346 Robben. H. S. J.. 361 Robberson. M. R.. 294 Roberto. AJ .. 115 Roberts. L. 149. 150. 155 Robinson. L. M .. 2 16 Rockwell. P.. 246. 25 1, 256. 261 Rodri guez. J. I.. 3 19 Rodriguez. R .. 76. 12 1 Rogers. R. W.. 66. 223. 229. 292.293.294 Rohrbach. L. A .. 277 Rohrbaugh. M .. 229 Roiger. J. F.. 248. 250. 260. 261 Rokeach. M .. 277 Rolnick. S .. 292 Roloff. M. E.. 15. 18.21. 93. 113.193,194.199.299. 305.319.32 1. 327, 330. 340 Rompa. D.. 229 Roni<;. D. L. . 50. 66 Rooney. B., 292 Roper. R. E .. 296 Ro rbeck. L.. 167 Rosenberg. M.. 277 Rosenkrantz. P. S .. 134 Rosen<;lock. I. , 292. 297
R O/ili~.
J.. 220
Ru bin. D ,L .. 54 Rubin. J.Z .. 121 Ruom. R.B .. 116. 118 Ruble . T. L.. 134. 136 Rucker. D. D .. 7 1. 72. 84 Rudd . J.E .. 115 Rudman. L. A .. 141 Ruiqing. D .. 18 Rule. B. G .. 186. 187 Rump. B. S.. 277 RlI ~c h er. J. B .. 134. 136 Ryan. K. B. , 277
Sabato, L. 1.. 273 Sahugal. F.. 233 Sacki. M .. 189 Sagre~[an(l, L M .. 321 Saine. T .. 102 Sa iyadain. M .. 102 Salan ic k. G. R.. :n7 Sala'i. E.. 134. l-t 1 Salomon. G .. 273. 277 Salovey. p .. 298 Samp. J.A .. 190. 193 San d er~. 0 .. 229 Sandcr'i. J. A .. 18 Sanden. R. E., 194. 325.
327.330 Satrc. S. J.. 322 Sawyer. A. G .. 267. 275 Sa) lor. K .. 339 S(·haefrcr. II. R.. 324 Schafer. E.. 2.n Schafer. R. B.. 132 Schand•. R. C. 31 Schein. V. E.. 134. 138 Sd lCn<.:k- Harnlin . W. J.. 192.
339 Scher. S.J .. 50. 51 Schili!. W. K .. 338. 34\.
342. 344 Schmidt. S. M.. 187.3:\7 338. 339.343. 3-1-1. 34(). 347 Schmiege. C J.. 322 Schmitl. D.R .. 181 .. 321. 3:'9
no.
Schm iu . N .. 134 St:hnc ider. D. E .. 291 Schneider. J.. 138 Schneider. M .. 292 Schneider. M. J .. 318. 34f), 342 Schocnbachkr. D. D.. 227 SdlOo lcr. C. 35 Schrader. D. c.. 188. 190. 191.199.200.201.203 Schrier. P. E.. 343 Schumann. D. W .. 71. 72. 7H.98. 100. 103. 104. 358 Schurr. P. 1-1 .• 355. 356 S(·hw<Jrt/. P...~4J Schwar/. N .. 72. 75. 83 Sch .... ar.lwa ld. J.. 320 Sd' .... ciklc. E. S .. 297 Scilcppi . J. A .. 100. 101. 10.1 Scott. L. M .. 39 Scott. M. D.. 267 Se;Ir1c. J.. 326 Scars. D.O .. 269 Sedikitlc'i. C .. 72 Seed~. DE. I 17 Segrin. 6.19.20. 122. 165.166.167. 168. 169. lB. 174. 176. 177. 188 Se ine !. C. A.. 229 Seibuk1. D. R.. IIJ. 296. 339.340 Seller. J. S .. 4. 6. 14. 18.19. 93 . 168. 3 17. 355. 356. 357. 361. 363. 36-1. 365. .~ 66. 3(,7 Sen L M .. 358 Scow. A .. 229 Serbin. L.. 137. 140 Severance. L. J .. 103 Sc'(ton. C. 5 .. 322. 323 Sh'lckelfurd. S .. 135. 139.
c..
141 Sharkey. W. F.. 54. 306 Sharma. A .. 355. 359. ]60. .164 Shavitt. S .. 81 Shechter. D .. 59 Sheer. V. C .. 321 Shcflie ld . F. D .. 65 Shepherd. G. J.. 13. 2(l. 189.
201.223.331 Shepherd. P.E.. 117 Sheppard. H.H .. 53 Sherif. C. W .• 54 Sheri f. M .. 54 Sherrel l. D. L 96. 102 Shimanofr. S. B .. 323 Sh imp. T . A .. 81. 99
379
Author Illdex Slcgcl. J.T.. 158 Sigelm,m. L.. 6, 35 Sillar-.. A L.. 322 Sil\'e~tri. V.N. 114 Simmond ... G .. 35 Simon ... H. W .. 2. 6.15.17.
95 Singer. D. A.. 166 Siamccka. N. J.. 306 Sim, .... L..]]3 Shller. M. D.. 279 Slocum. J. W.. 359 Siovic. P.• 177 Smalec. J. L.. 226. 292. 293 Smi l o~ III. M .. 124 Smith. D. H .. 306 Smi th. E.R .. 48 Smith. S. M.. 67. 72. 74. 76. ~I
Smith. S. W .. 123. 186.345 Smith. T E.. 325 Sn)der. M .. 47.167. 345 Sohl. R. S .. 355 Soley. L. C. 37. Solomon. D. H.. 188. 190.
193.354 Somcn-illc. S. C. 295 Son. L.. 134 SorclNIIl. J .. 11 ..l6. 26M Spear. S F.. 33 Spector. S .. 29.l Spiegel. S.. 59 Spiro. R. L.. 357. 364 Spi~al<... 1< 329. 154 SpitLbcrg. B.. 5. 340. 354.
358
c..
Spraf"l.lI1. 14() Sprowl. J. P.. 35M. 363. 3M Staab.. A. W .. 6(1. 68 Staat.... C K.. 66. 68 Slad .... D. W.. 17.l. 174. 177 Staclin. R.. 275 Stafford. T.......l68 Stang. D. J.. 49 St'lIlland. A. J S ..l55 Stcad. B. A .. 1n Steele. CM .. 52 Steen. D.M .. 54 Stencn. V. 1.. 1:\7 Steffian. G .. 33 Steil. J M.. 319. 322 Stcin. R T .. 136 Stein, S.. 140 Stein, Y .. 277 Steiner, V .. 187 Steinfml. T. M.. 266, 267. 272 Step. M.M .. 124 Stephen. R .. 16M
no.
Stephen ...on. M.T.. 226 Stem, L. A .. 239 Stern. R. A .. 277 Sternthal. B.. 66. 76. 100.
102 Steward. M. A .. 306 Stewan. D .. 139. 152. 153 Stcwan. M .. 307 Stewan. R.. 345 Stewan. R. A. 116, 117. 11M Stcwan. R. B.. 290 Stiff, J. B.. 15, HI. 19.22. 57.59.77.78.79. 80. ~I. 10 1. 102.177. IH2. 232. 246,255.300 Stinnett. W. D .. 267 Stnj:tel<... C. 360 Stok.n l ~. D.. 292 Stone. G. P.. 360 Stone. 1.. 52 Stone. V A .• 269 Strathm:lIl. A. J.. 72. 81 Str:mgh'lII. P. 1'.. 229 Street. R L.. 175. 299. 3()7 Streeter. L. A .. 169 Struther.. J. T.. 341 Slrul1on. D.. 356 Strt)l..eW ... k.I. K. 0 .. 245. 248. 255. 260 SlUtman. R. K .. [87 Sm:hman, E. A .. 228 Sue!. GJ .. 50 Sueda. K .. 18 Suedrcld, P.. 268 Suja n. II. . 358. 359.164 Sujan. M . 358. 359 Sulli\an.1. 344 Su .... man. M .. 339 ..l43. J~6 Sutton. S. R.. 35. 227. 2l)2 S, arrhtad. B.. 289 Swan. J. E.. 355 S",an. S .. 81 S",ap. W.C. 121 S"'en"on. M. J .. 36'1. 363 Sv.-im.1.. [34 S/abo. E. A .. 268. 269. 27\. 178 S,ybiIJo. G. 1.. 267. 275 S,yman,l<..i. D. M . 358 Tannenbaum. P. H .. 50. [DO. 102.268.272.274.176 Taranta. A.. 294 Tarpley. W.R" 52 Tarrance, V. L.. 273 Tate. E" 267 Tauer. J.M., 52 Taylor. A. M .. 15!,! Tay lor. S .. J·W
Taylor. S. E.. 305 . 306 Tedc'>Co. L. A.. 293. 297 TeJlelbaum, M. A .. 300 Tennen. H .. 229 Te\cn. J. 1.. 96 Thamodaran. K.. 77 1lIOmp>;on. E. P.. 5lJ, 78, 8 [.
X2 Thomp ..on, \1.. 9H Thrueworth). R. C .. 290 Thurman. Q. C .. 292 Timko, C .. 297 Timminga. E.. 40 Tinkham. S. F.. 273 Tipper. S. P.. 188 Tju... , old. I) .. 34 [ Tnblt:r. N. S. 279 Tlxld. J.D .. 226 Tlxlnro\. 1'.. 330 Tnllllirholl-Kca ..ey. (' . 172 ToneaUo. 1 . 54 Tori .. , D.. 24H Tnnnal:l. / L .. H3 Tmce}. J. B .. 346 Tracy. K.. 18H. 1~9 ..l~9.
354 Tracy. S. 1.. INS Trafilllo\\,. I)., 54 Traupm;ulII. J .. 190 Trebing. JI), 117 Trenholm. S .. 16 Trope. Y., 57. 85 Tro'l. HR .. 171. 207.215 Tudcr, J. 5 .. 322, .l23 Tumlin. S.A. . I 19 TUlicalp. S .. 277 Turbl:l". A A .. J~ Turk. D. C.. 2lJ8
Turner. C ..'03 Turner. J. A .. 103 Tul1oll. P...l5 Tu,ing. K. L 192.268.330 r\Cf\k.). A.. 177 T",engc. J. M. 13K Vllman. \\0 R.. 167 l·nger. R K .. 172 Llnn:lva. HR .. 40 Lltne. M. K .. I90
V:llencil·. K.M. 115 Valentinc. ~1. E.. 167 Valkcnhcrg, I I., 318 Vall:rbhan, S.. 226 Val[erand, RJ., 54 Van Bocl<..crn. $ .. 13, 26H. 272. 277. 278 \an der Kloo!. W .. JI8 V;1!1
Egcrcn. L. A.. 119
Van Leuven. J. 279 Vance. K .. 52 Vancou\·er. J. B.. 189 VandeWalle. D., 359 Vamllo. S. M.. 330 Vaughn. D .• 267 Vecchio. R. P.. 339, 343. t
346 Vedder. R. G.. 40 Veliclllan. G .. 35 Verhallen. T. M. M .. 361 Ve\ca.1. L .. 38 Vi~wanath. K .. 292 Vogel. D .. 4() Vogel. S. R.. 134!. von Cranach. M.. [87 Vrij. A.. [8,22
Wagener. J J .. 306 Waggen ... p.. tL B.M. [17 Wagner. B., 1.'9 W:t/lncr. D. 29() Wagner. D G., US. 138 \\ahmlan. R.• 139 Waldron. V R.. [89. 194. .:!()[ \\all<..cr. H A .. 136 W,,[l<..cr. M. [67. [71 Wallo.cr. ~1. A.. 322 Wall. C.H .. 117 W'IIt;Ll·k. L.. 277 Walthnmn.1 . .:!29 '" al ... tcr. E.. 95 Wa[ter.... D. I .. 360 Wa[ther.J.. 256. 260 W,,[ullan. ~ 1 S .. 340 W:II1, H.· H .. 168. 276. 277 Wanke. M .. 81 Ward. CD .. 100. 103 Ward. S.. 325 Ware. J. E.. 294 \\aro.hav.. P R .. 53 Waler... E.. 321. 324 Wat ...on. C. 104 Wahon. ~1. R.. :n7 ..l40.
346 Wall .... W .. \02 Wca\cr. S .. [8S We:l\er. R \1.26 WCil\"cr-L.. ri~c). R A .. 273 Webb. J. A .. 179 Webh. P. A.. 191. 297 Weber. F.. 212 Weber. S. J.. 227. 228 Wegellcr. D. T .. 37, 46. 68, 71.72,73.74.75.78.79.
81. 85 Weimann. O. [40
380
Allthor J"dex
Wein·aein. N, 0 .. 35, 227. 231 , 290. 292 Weisman, C. 5 .. 300 Wei!.s. W .• 65. 66. 76 Wei!.!.C. C. S .. 34 Weitl. S , A .. 356. 358. 359.
364 Weitzel -O·Ne ll. P. A .. 136.
137<; Well~ . G. L.. 73. 98 Wells. W. T .. 105 Weltman. K., 322 We!>t, L.. 330 WCMon. W. W .. 307 Wetzel. C. G .. 168 Whaley. B. B .. 340 Wheatley. J. 1.. 228 Wheeler, S. C .. 67. 83 Whecles<;. L. R.. 345 Whetstone-Dian, R.. 141 Whicker. M. L. 6 White. c.. 250. 252. 254.
256.258.261 White. J. 8..297 White. J. W .. 343 White. K. 0 .. 318
White. L. 229 White, P.. 74 Whitccap. V.G .. 115 Whitely. W .. 346 Whitney. D. J .. 34 Whinier. T. E.. 227. 368 Wicker. A.W .. 47 Wiemann, J. M .• 299. 307 Wigley. C. 1.. 115. 117 Wilke~. R. E.. 325 Wilkinson. I.. 187.337.338 William!>. J. E.. 134. 136 357. 364 Williams. K. Williams. K.D .. 74 Williams. M. R, 16 Williamson, S .. 49 Willi!>. F.. 167 Wilson. B., 172 Wilson, C. P .. 34 Wilson. D. T. 355. 356. 357 Wilson. E. 1.. 96. 102 Wihon. E. V .. 188. 196. 197 Wilson. J. T .. 290 Wil<.on. S. R .. 16. 189. 190. 191 ,192. 19-t. 203. 326. 329. 330. 342. 344. 354
c..
Wilson. T. D. 74 Wibon-Davi ... K.. 277 Winke.J.. 300. 340 Wi!.e. S. L.. 229 Wi')cman. R. L.. 18. 192.
339 Witte. K .. 13.223. 224. 226.
227.228.229.231.234. 292.293.294.295.296 Wiuem:Hl, H .. 321. 322. 323 Wolibki. R. 1.. 35 Wolken stein. B. H .. 277 Womack. D. F.. 10. 113 Wood. J. T. 330 Wood. W .. 37, 46. 73. 74. 81.134.135.136.139.
141 Woodall. W. G.. 174. 176.
239 Woodward. G. C.. 6. 16. 19 Worchel. 5 .. 135.139.141 Wortman. C. B.. 231 WOrluba. T. 356 Wosinska. W .. 51. 141 Wyatt. R. A. J .. 355
Wygand. A.W .. 52 Yandell. B.. 66 Yang. V. 5 .. 268 Ye!>eno!>ky. J. M .. 229 Yoder. 1.. 172 Youn. J. Y., 341 Young. M .. 306. 307 Young. M.J .. 18 Young. T 1.. 267 YukI. G .. 338. 341.343.346 Zajonc. R. B .. 47. 48 Zanna. M. P.. 35. 48. 66 Zarate, M.A .. 48 Zelditch. M .. 135 Zemanek. J. E .. 357. 358 Zigur!>, I.. 188. 196. 197 Zinkhan. G. M .. 173 Zolner. K .• 358 Zubric.J.. 268. 269 Zuckerman. M. 230. 248 Z'·onko\'ic. A. M .. 322. 323 Zweigenhafl. R. L.. 167
Subject Index abili ty 10 proce!->~. and elaboration likelihood model. pp. 67. 70.%-98 accidental persuasion, pp. 19- 2 1. 23 active p,tradigm. and Lang uage Expectancy Theory.
p. 150 additivity hypothesis. and He uri<"lic-Syslcmal ic Model. pp. 58-59 adolesccllIs, p. 325 affirming sty le. and arguing, pp. 125- 126 agenlic qualities of mal e~. pp. 134 anchor point. und Social Judgment Theory. p. 55 antagoni .. m. and arguing. p. 119
appearance clle~. pp. 171- 173 argumen t quality versus qu an tit y. pp. 70. 79
argumentativeness. see eh. 7 gene rall y. pp. 92 asse rti veness. and arguing. pp. 11.f assimilation effect, and Social Judgment Theory. p. S6
attack ads. p. 35 attenuation hypothesis. and Heuristic-Systematic Model. p. 59
allilUde. pp. 32.
~6--l 7.
72.
83-8~
and La nguage Expectancy Theory. p. 153 allilUde-bchavior relationship. p. 47 altitud e ~trength postulate. and Elaboration Like lihood
Mode l. pp. 75-76
commi tment. pp. 209. 363 Commitmcn t-trust Theory. p. 356 communa l behuvior. and fcmale pcr'iuadcrs. pp. 1..1 1-
143 Co mmun il.:.Hi on Accolllmodation Theory (CAT). PI'.
175- 176 communic~Hion
competence. Pl'. 199-200 competenct! (cxpcI1i'ic). p. 95 and gcndt.!r. pp. 139- 141 co mpli ance gaining and com puter mediated communication (C MCl. pp.
39-40 deductive "'lrHtcgie!o.. p. 339 definition or. pp. 20-21 a nd erfect<., emphasis. p. 21 and fal11ilic .... pp. inductive 'itrategies. Pl'. 339- 340 and intim ates. pp. 319-321 rC'iearch .tbou!. PI'. 16-17 Congru ity Theory. p. 5 1 Consistency Theory. pp. 50-53. 209 contrast erfect and ... ales. p. 362 and Social Judgme nt Theory. p. 55 correctnt!ss postu late. and E laboration Likelihood
Model. pp. 68-69 bait and sw itch tactic. pp. 207. 2 10. 2 11-212 belief dillconfirmation. and cogni ti ve dissonance. p. 52 bias hypothesis. and H curi~ ti c-Sy'itcmatic Model.
p.5Y biased processing. and Elabora ti on Likelihood Model.
pp.7+-75.84 body language (see kinesics) borderline persuasion. pp. 17- 1H central route. p. 67. 96. 176-177 classical conditioning. pp . 47-4R. 68. 72 client-oricllIed ba lespeopJc. p. 357 clothing (!-ice appearance cues) Cognitive Disllonance Theory (COT). pp. 3. 7n. 5 1-53 cognitive edi ting of arguments. p. 121
counterargui ng. p. 270 credibi lity ....ce Ch. 6 genera ll y, pp. 76-77. 92 and appeara nces. pp. 171-173 and deception. pp. 257-259 and Language Expectancy Theory. pp. 154-1:'i5 repair or restoration of. pp. 104-107 and rctail sa les. p. 355 cultivat ion. and arguing. p. 119
cultural fac lors. pp. 18. 158. 233- 234. 32~. 327- 328. 343- 344 c ultura l trlli~ms. and inoculation. pp. 266-267 c ustomer t ypc~. pp. 359- 360. 365- 367 danger control. and fear appeal .... PI'. 224--2.25. 295 decept ion. sec C h. 14 generally. pp. 182- 183
381
382
SlIhjecr Illdill
deception (('(JIlfjIllWd) and credibility. pp. 257-259 definition of. p. 2·m and Interpcr..onal Deccption Theory (IDT )...ec Ch.8 as a form of pcr .. uasion, p. 22 deci .. ion rules. p. 58 descriptive gender stereotypc .... p. 137 dialogic perspective. pp. 3 18. 330-33 1 diffu\e stalU ... characteristic. and gender. p. 135 Direct Effect.., Model of Immed iacy. p. 166 directive specch acl. p. 326 discrepancy. PI'. 102- 103 distraction. pp. 174-175 door in the facc ... trmcgy (DlTf). pp. 38. 217- 219. 362 dual.proces .. theories. pp. 57-59 effech cri terion. pp. 2 1 -:~2 effort justification paradigm. and cognilive di .....o· nance.pp.52-53 elabormion continuum . pp. 67. 70 Elaboration Lil..elihood Mode l (ELM). \CC Ch . .5 gcnera ll ). PI'. 36-37. 57-59. 96- 100. 176-177 ethic .... and persuasion. pp. 1-7 etho....... ce Ch. 6 gencrally. pp. 65-66 e"idcnce and argumcntativcness. p. 12J and Uni modc l of per .. ua\ion, p. 82 Expectancy Vu) lations Theory (EVT). pp. 56. In. 173- 174 experti"e. p. 95 Extended Oarallcl Process Model (EEPM) ...ec Ch. 8 generall y. pp. 294-295 eyecontac t.pp.141. 166-167. 170. 17'" facework. pp. 325. 328-330 fear appeals. "ce Ch. 13 generally. pp. 154---155. 156157 and age. p. 227 and anxicty leve l. pp. 227-228 and fatalism. pp. 228-229 and reactance. pp. 229-230 and sensation ~eking. pp. 230-23 1 fear control, pp. 225-226. 295 fluency. and mere exposure. p. 49 food servers, pp. 168. 353 foot in the door strategy (FITD ). pp. 210. 212-216 forced compliance (see induced compliance) forewarning. p. 74 (see also, Inocu lation Theory) free choice paradigm. and cogni tive dissonance. p. 52
full·cydc "ncial p:-y<..' hology. p. 364 gaLe.pp. 166-167. 170 gcnder differencc\ ...ec Cil. H genera ll y agentic <'Iualiti es of male persuader... pp. 134- 135 cOTllmu nal qualities of female persuaders. pp. 13613K. 141 - 143 a" a diffu'ie stalu .. characteristic. p. 135 .Uld Language Expec tancy Theory. pp. 153- 154 and phy .. ician-patient influcnce. \cc Ch. 16 gcncrally and organizational inlluence. p. 343 and .. tatus difference:-. PI'. 135- 136 and "",COI)pes. PI'. 13-1-135. 136-138 Gml1 \- Pl arl'..·Act ion (G PA ) model of imerper<;onal intlucncc. sec eh. II generall y goa l struct ure co mpl ex ity. pp. 190-1 92. 199-200 goal>. pro 186-192. 354 Goldberg paradigm. p. 134 hnloeffcct.pp.171 - 172 haptic:-. pp. 167- 168. 170 Ilea lth Be li ef Modcl. pp. 2<)0. 291 -294 heuri.<,tic cues. p. 58 hcuri ... tic proces!-.ing, p. 58 II curiqic-Systematlc Model (IISM). pp. 58-59 ho... tility. and argui ng, pp. 11 5- 116 idcntity goals. p. I HR image repair. pp. 104- 107 Illllllctlr;'IC)'. pp. 166. 169. 170-171. 299 incremcnta l· intcractl\.:e approac h. p. 325 induced co mpli ance. anti cog niti ve di ... so nancc, p. 52 in tluencc. p. 20 Inoculation Theory ...ee Ch. 15 genera ll y. p. 36 and political cnmpa ign .... pp. 273-27 5 ingratiation. pp. 345. p. 357 intent critcrion. pp. 19-2 1. nn Int ent ion, and Theory of Reasoned Action, p. 53 interac tionist pefl.pcctivc. pp. 116-11 8 Intcrper...ona l Deception Theory (IDT) ....ee C h. I'" gencmlly. p. 20 intimatc'i. 'tee Ch. 17 generally involvcmcnt and cn.:di bility , pp. 101 - 103 and the Elabonllion Likelihood Model. 'oee Ch . 8 generally and inoculation. pp. 270-27 1 kincsics. pp. 168- 169
SlIhjecT Index
language and persua'\ion. p. 91 Language Expectancy Theory (LET) . .,ee Ch. \l gcncrall) latitude of acceptance. Pl'. 55- 56 latitude of noncornmitmcnl. pp. 55-56 latitude of rejection. Pl'. 55- 56 Lay Epistcmic Theory (LET). p. Xl leadership :-.tyle. p. 3..J.I Icgilirni;.o;ing :-.tratcgy. p. .146 liking. p. :t;;S-356 logm. Pl'. 65- 66 low-hall tactic. pp. 20t)-21 0 Machiavellianisill. p. 345 medical comp liancc gaining. "ee Ch. 16 generally medium of communication. pp. 172- 27.1 merc expmure (ME). Pl'. 47-50 rneta-analy"is. pp. 38.167 moderating \ariable\. Pl'. 36. l) I mood and Elaboration Likelihood Model. p. 72 and Inoculation Theory. pp. 171-272 motivation. and In oculation Theory. pp. 267 motivation to comply. and Theory of Reasoned Action. p. 54 motivation to process. and Elaboration Likelihood Mudel. PI'. 67. 70. 96- 98 multiple me'\sagc ..,wdies. p. JR multiple role" postulate. and Elaboration Likelihood Model. PI'. 7 1- 73. n. so need for cognition. p. 121 negative incentive effect. p. 51 nonallirming ..,tyle. and arguing, pp. 125- 126 nonstrategic hehaviors. and deception. l'. 241 nOIl\erbal inlluence. see Ch. 10 generally
383
pathn-.. pp. 65- 66 perceived efficacy. and fear appeals, pp. 224-225 perceived thrcat. and fear appeab-. pp. 224-225 peripheral route. p. 67. 96. 99-100.176-177 pcrsi"tcncc of persuasion. pp. 67.75- 76.272 personality factor~. pp. 34..J..-.345 per"uasion borderline per\ua<;ion. pp. 17-IS. 23 definition of- scc Ch. 3 in gene ral. pp. 92-93 ethic" and. pp. 4-5 fcminist per~pective of. pp. 2. 7n negative connotation" of. pp. 1- 2 cOlllmunication competence and. p. 5 inllucncc versus. p. 20 linear view of. pp. 20. 22 pervasivcnc ... " of. PI'. 6. 31 proce ... " vcr... m outcome. p. 11 pure persuasion. pp. 17-18.23 phy~ician-patienl compliance gaini ng. see eh. 16 generally pluns. pp. 192- I95 Politeness Theory. p. 219 preference forcon:-.i"tency. pp. 215- 216 prescriptive gender "tereotypcs. pp. 137. 142 primary goal". pp. 187- 190.201-202 pSYChological reactance. pp. 229- 230 public ..,ervice announccment" (PSA:-.), Pl'. 156--159.
277-2HO pure per"u
J
384
Suhj{>ct hu/e.l'
selective expo~ure. p. 52 self-efficacy. pp. 33~34, 271. 292. 295 self~l1loniloring. pp. 48. 345 Self Perception Theory, and foot in the door. pp. 2 13215 self-validation processes, and the Elaboration Likelihood Model. pp. 83-84 sen~alion seeking. and fear appea ls. 230 "icqllcntial influence <;tralegie~. see Ch. 12 generally, pp. 303~306 sex differencc"i. ~ee Ch. 8 generally simu ltaneous proce~sing. pp. 57-58 sing l e~process model of per~uasion (Unimodel). pp. 81-82 smiling. p. 169 social desirability bia!:'. p. 48 soc ial influence, pp. 20-21. 27n definition of. p. 337 gender differences. pp. 13R-141 Social Ju dgment Theory. pp. 54-56 social meaning pcrspwive. pp. 166,318.325-330 socia l proof. p. 362 stage lJlodel~ of persua!:'ion. p. 35 stereotypes and gender. pp. 134-135 !:, lIbject ive norms, and the Theory of Reasoned Action. p.54 sufficiency principle. and Heuristic~SyMematic Model. p.58
superior~subordinate
influence. see Ch. 18 generally suspicion. and deception. pp. 255-257 systematic processing. p. 58 that's not all technique, pp. 219-220 theory. definition of. p. 10 Theory of Planned Behavior. pp. 296-297 Theory of Reasoned Action (TRAJ. pp. 33.53-54. I 17-1 18, 290. 296-297 IOp~down sales tactic, p. 362 touch. pp. 167- 168, 170 trad e~off postulate. and the Elaboration Like lihood Model. p. 75 traditional perspective. and close relationships. pp. 3 17~319
transact ional perspective. p. 120 and deception. pp. 240, 244 Transtheoretical Model. p. 35 tru'itworthiness. pp. 95. 103 Unimodel of persuasion. pp. 8 1-82 unintended reccivers. pp. 19-20 verbal aggressivene~~. see Ch. 7. gene rally. pp. 92. 345 visua l persuasion, pp. 39-40 vocal cues. pp. 169. 25 1-252