eon,

theeof a

elfle

Journal of Experimental Social Psychology37, 228–243 (2001)doi:10.1006/jesp.2000.1446, available online at http://www.idealibrary.com on

A Self-Standards Model of Cognitive Dissonance

Jeff Stone

University of Arizona

and

Joel Cooper

Princeton University

Received June 3, 1999; revised July 10, 2000; accepted July 10, 2000; published online February 28, 2001

This article presents a new model for understanding the role of the self in cognitive dissonance processes. We focus on thcontroversies among three major theories of how cognitions about the self mediate dissonance processes: Self-consistency (Arons1992), Self-affirmation (Steele, 1988), and the New Look perspective (Cooper & Fazio, 1984). It is argued that each of thesecontemporary revisions of dissonance theory assumes that dissonance begins when people commit a behavior and then assessmeaning of the behavior against a standard for judgment. However, each approach makes different predictions for how self-knowledgmediates dissonance because each assumes different self-attributes and standards are used to assess the psychological meaninggiven behavior. The proposed model suggests that the basis of dissonance motivation and the role played by cognitions about the sdepend on the type of self-standards made accessible in the context of discrepant behavior. By examining the ways in which peop

use self-standards to assess the social appropriateness or personal quality of their behavior and use self-attributes to reduce theirdiscomfort, the proposed model can predict the conditions under which each of the contemporary views of the self in dissonance isthe most accurate explanation of the process of dissonance arousal and reduction.© 2001 Academic Press

lf inThise th

bythatper

vioritionat fith,reognanc

son,ug-the

per &e is

bytionson &can,con-levantitive

d ton the9).ro-used

ce be-

wasIMHudenes,

What role, if any, is played by cognitions about the sethe arousal and reduction of cognitive dissonance?question has been debated for over 40 years sincoriginal theory of cognitive dissonance was publishedFestinger in 1957 (see Harmon-Jones & Mills, 1999). Intime, researchers have offered at least three differentspectives on how the self influences responses to behadiscrepancies. One perspective maintains that cognabout the self represent standards or expectancies thcilitate dissonance arousal (e.g., Aronson & Carlsm1962; Aronson, 1968; Duval & Wicklund, 1972). Morecent theory and research, however, proposes that ctions about the self function as resources for disson

Previous versions of this article were written while the first authora postdoctoral fellow at Princeton University and was funded by NPostodoctoral Grant F32MH11025. We thank Elliot Aronson, ClaSteele, Mark Zanna, Dale Miller, Adam Galinsky, Eddie Harmon-Joand Jeff Greenberg for their incisive comments on previous drafts.

Address correspondence and reprint requests to Jeff Stone, PsychologDepartment, University of Arizona, Tucson, AZ 85721. E-mail:jeffs@u.arizona.edu.

228

0022-1031/01 $35.00Copyright © 2001 by Academic PressAll rights of reproduction in any form reserved.

e

-alsa-

i-e

reduction (e.g., Steele, Spencer, & Lynch, 1993; AronCohen, & Nail, 1999). Still other theory and research sgest that cognitions about the self are irrelevant toprocess of dissonance arousal and reduction (e.g., CooDuncan, 1971; Cooper & Fazio, 1984). Indeed, therpublished empirical support for the predictions madeeach perspective concerning the role played by cogniabout the self in the dissonance process (e.g., AronsCarlsmith, 1962; Steele et al., 1993; Cooper & Dun1971). It is fair to suggest, however, that no generalsensus exists among researchers about how self-rethought mediates the arousal and reduction of cogndissonance.

This article introduces a new model that was designeforge a synthesis among the different perspectives orole of the self in dissonance (Berkowitz & Devine, 198Toward this end, the proposed model will highlight a pcess that we believe is central to how dissonance is aroand subsequently reduced. We propose that dissonan

gins when people commit a behavior and then assess thebehavior against some meaningful criterion of judgment.

y

eriono ths thself-(cfork

rolepecy ofer-ointpreandbe-

“per-By

utesty o

anerateredice aseopderfeeldernce

ndetion) ornd,

ingtive,whyance

theks tohreepre-

role ofily onisso-mongmaderon-

al.,oper.

e asthe

n &ron-oplethat

ling2).

pancyttitu-g aper-utesce isralityeau

thatss be-nt orfor

es aonanusin

searcr. Tha-

theeon

oun-son,x &harmmostnanced by88) o, theffectstions

onantingerheo-

tationsssionf these(e.g.,

DS I

Only when an action is measured against a relevant critdoes it take on a mantle that is aversive or threatening tself. In the current model, the criterion used to assesmeaning of behavior is represented in memory asattributes and self-standards or guides for behaviorHiggins, 1987, 1990). We refer to the proposed framewas theSelf-standards Model of Cognitive Dissonance.1

The Self-standards Model (SSM) maintains that theof self-relevant thought described by the various perstives on dissonance is a function of the accessibilitdifferent types of self-attributes. Most important for undstanding the differences between the competing viewpis the distinction between discrepancies that involve resentations of actual self-attributes and self-standardswithin the dimension of self-standards, the distinctiontween discrepancies that involve what we define assonal” versus “normative” self-standards for behavior.focusing on how the accessibility of relevant self-attriband self-standards influence perceptions of the qualiappropriateness of behavior, the SSM can specify howwhen self-concept differences like self-esteem moddissonance processes. Specifically, the model can p(1) the conditions under which self-attributes will operatan expectancy and cause more dissonance among pwith high versus low self-esteem, (2) the conditions unwhich most people regardless of their self-esteem willthe “bite” of dissonance arousal, (3) the conditions unwhich most people will be motivated to reduce dissonavia self-justification processes, and (4) the conditions uwhich self-esteem differences will moderate the reducof dissonance arousal using a direct (i.e., justificationindirect (i.e., affirmation) strategy (e.g., Stone, Wiega

1 The emphasis in the current article on behavioral discrepancicentral to dissonance processes is based on the way in which dissresearch is conducted. While the breadth of cognitions capable of cadissonance was considerable in the original theory, experimental reon dissonance has traditionally focused on cognitions about behaviouse of “induced behavior” became thesine qua nonof dissonance pardigms, as most of the now-classic experiments successfully testedhypotheses by inducing such overt behavioral actions as lying to som(Festinger & Carlsmith, 1959; Cooper & Worchel, 1970), making a cterattitudinal advocacy (e.g., Cohen, 1962; Nel, Helmreich, & Aron1969), making a difficult and irrevocable decision (Brehm, 1956; KnoInkster, 1968), eating grasshoppers (e.g., Zimbardo et al., 1965), oring innocent victims (e.g., Glass, 1964). As this short list illustrates,of the original experiments that tested predictions derived from dissotheory tended to focus on behavior that was likely to be construeparticipants as threatening to the self (e.g., Aronson, 1968; Steele, 19in some way aversive (e.g., Cooper & Fazio, 1984). As a resultinduction of behavior as the mechanism for studying dissonance eproduced revisions of the original theory that relied heavily on assumpabout how people make sense of their behavior. So, rather than dissbeing a function of just any inconsistent set of cognitions as Fesoriginally posited, dissonance is primarily viewed by contemporary t

SELF-STANDAR

rists as a function of cognitions concerning behavior—and in our currentview, about the preexisting cognitions that help determine what the behavior could have or should have reflected, achieved, or produced.

ee

.

-

s-,

rd

t

le

r

Cooper, & Aronson, 1997). We believe that by highlightthe processing implicit to each contemporary perspecwe can propose a comprehensive model of how andcognitions about the self influence the process of dissonarousal and reduction.

To establish the context for the SSM and to examineseemingly contradictory predictions that the model seeaccommodate, we present a brief overview of the tcontemporary theories of the self in dissonance beforesenting the tenets of our new process model.2

COMPETING THEORIES OF THE ROLE OF THE SELFIN DISSONANCE

The debate among dissonance researchers over thethe self and self-esteem in dissonance centers primarthe nature of the self-relevant thought that influences dnance arousal and reduction. Much of the dissension aresearchers has centered on the different predictionsby the self-consistency (Aronson, 1968; Thibodeau & Ason, 1992), self-affirmation (Steele, 1988; Steele et1993), and “New Look” or aversive consequences (Co& Fazio, 1984; Cooper, 1992) theoretical perspectives

The Self as an Expectancy in Dissonance Processes

The perspective that cognitions about the self servexpectancies for behavior was initially advanced byself-consistency model of dissonance (e.g., AronsoCarlsmith, 1962; Aronson, 1968; see Thibodeau & Ason, 1992). According to self-consistency theory, pehold expectancies for competent and moral behaviorthey derive from “the conventional morals and prevaivalues of society” (Thibodeau & Aronson, 1992, p. 59Dissonance is aroused when people perceive a discrebetween their behavior, such as advocating a counteradinal belief (e.g., Festinger & Carlsmith, 1959) or makinquestionable decision (e.g., Brehm, 1956), and their “sonal standards” or self-expectancies for the self-attribof competence and morality. The reduction of dissonanaimed at maintaining a sense of competence and mothrough justification of the discrepant behavior (Thibod& Aronson, 1992).

The self-consistency perspective further proposesself-esteem moderates the dissonance arousal procecause the perception of what constitutes an incompeteimmoral act is a function of the expectations people hold

sceghe

ire

-

r

ce

2 The proposed model does not address at least two other interpreof dissonance, specifically, the self-perception (Bem, 1972) or impremanagement (Tedeschi, 1981) revisions. This is because neither otwo revisions assume a role for preexisting cognitions or arousal

229N DISSONANCE

-Fazio, Zanna, & Cooper, 1977), each of which is central to the argumentspresented here. A more thorough discussion of how these revisions fit theproposed model is beyond the scope of the current article.

erylowg aherhig

vatesis-cultun-ilabl

witlowmildnotdis

(sehus,tionseoral

asromals

tencusedthe

ctiveof ac, ththe

y fo-por-le &ell,selfith-e e

s ththeourcuceg tog awitthei

encevan

ed aopleca-t al.,tive,

ng a

t theancedel

ce ision ofThee fitlieveav-

per-d theyThecribee is

tivel ofavioretentdel,ancenanceiated

itivewhyisso-maket inf ex-

anceachcon-

, theisso-focus-one

ND

their behavior. For example, if a person did not hold vhigh expectations for competent or moral behavior (i.e.,self-esteem), dissonance would not be aroused followinincompetent or immoral act, whereas people with higexpectancies for competent and moral character (i.e.,self-esteem) would perceive a discrepancy and be motito seek self-justification. Despite the fact that the contency effect for those with low self-esteem has been diffito replicate (e.g., Ward & Sandvold, 1963; Cooper & Dcan, 1971; see Shrauger, 1975, for a review), the avaevidence suggests that under some conditions, peoplenegative expectancies (Aronson & Carlsmith, 1962),self-esteem (Glass, 1964; Maracek & Mettee, 1972), ordepression (e.g., Rhodewalt & Agustsdottir, 1986) doexperience dissonance arousal when their behavior iscrepant from socially accepted standards for conductSwann, 1990, for a similar theoretical perspective). Tself-consistency theory assumes that positive cogniabout the self cause people to bemore vulnerable to tharousal of dissonance following incompetent or immconduct (Thibodeau & Aronson, 1992).

The Self as a Resource in Dissonance Processes

The perspective that cognitions about the self serveresource for dissonance reduction derives primarily fresearch on self-affirmation theory (Steele, 1988; seeSpencer, Josephs, & Steele, 1993). Like self-consistheory, self-affirmation posits that dissonance is arowhen people engage in actions that pose a threat toself-concept. But in contrast to the consistency perspeself-affirmation theory maintains that the primary goaldissonance reduction strategy isnot to rescue the specifiself-image threatened by discrepant behavior. Insteadgoal is to restore the moral and adaptive integrity ofoverall self-system. This goal can be accomplished bcusing on other positive aspects of the self that hold imtance for global self-worth (e.g., important values, SteeLiu, 1983; positive social comparisons, Tesser & Corn1991). The accessibility of positive cognitions about theallow people to reduce their psychological discomfort wout addressing directly the discrepant cognitions (Stonal., 1997; Simon, Greenberg, & Brehm, 1995).

In the self-resource model, self-esteem representdispositional availability of positive cognitions aboutself that can be used for dissonance reduction. The resmodel maintains that for self-relevant thought to redpsychological discomfort, people must be able to brinmind more positive than negative self-attributes followindiscrepant act. Based on the assumption that peoplehigh self-esteem possess more positive attributes inself-concept than people with low self-esteem (see Spet al., 1993), the resource model predicts that self-rele

230 STONE A

thought can provide more affirmational resources to peoplewith high self-esteem relative to people with low self-

n

hd

eh

-e

a

oy

ir,

e

t

e

e

hrrt

esteem. Research shows that when people are allowmoment of self-reflection before a discrepant act, pewith high self-esteem report significantly less self-justifition compared to people with low self-esteem (Steele e1993). Thus, according to the self-resource perspecpositive cognitions about the self cause people to belessvulnerable to dissonance arousal and reduction followidiscrepant behavior.

The Self as Irrelevant to Dissonance Processes

A third theory of cognitive dissonance suggests thaself and self-esteem concerns are irrelevant to dissonarousal and reduction. According to the New Look mo(e.g., Cooper & Fazio, 1984; Cooper, 1992), dissonanaroused when a discrepant behavior represents a violatsocietal or normative standards for behavioral conduct.assumption is that people learn as children to monitor thbetween their actions and what their parents or peers beis appropriate behavior—the perceived “norms” for behior (Cooper, 1999). When behavior falls short of theceived normative standards, dissonance is aroused anare motivated to justify the behavioral discrepancy.New Look model assumes that since most people substo similar societal norms for behavior, then dissonancnot predicated on the availability or accessibility of posiself-attributes. Most people, regardless of their leveself-esteem, should feel uncomfortable when their behviolates the internalized normative standards for compand moral conduct. According to the New Look mocognitions about the self have no special role in dissonprocesses and most people are vulnerable to dissoarousal when they perceive that their behavior has devfrom the relevant norms.

In sum, each major perspective on the self in cogndissonance makes specific predictions for how andcognitions related to the self mediate the process of dnance arousal and reduction. However, because theyopposing predictions for the role of self-relevant thoughthe dissonance process, no single theory is capable oplaining all of the available data.

TOWARD A UNIFYING MODEL: THE ROLE OF SELF-STANDARDS IN DISSONANCE PROCESSES

From our perspective, the current state of the dissonliterature resembles a giant jigsaw puzzle of which etheoretical perspective on the self in dissonance hasstructed one substantial section. At this time, howeverdebate over the most parsimonious explanation of dnance processes has obscured the overall picture bying on only one piece of the puzzle. We believe there is

COOPER

comprehensive picture and that it can be constructed fromthe various parts that already exist. For example, it is quite

eope or-conionsrsiv

Newns,

ts ofhavjor

implach

isso-utiome-

uct an oritykesiranctheher

theirus

vior.hicheliesolvrioure otivees b

usenot avio

in alish,ni-aviojudgn ovior

ationim-heldtan-f thrpreine

thetify

ce toeemitionsd toump-nceelf-t of a

pro-gicall dis-isso-e

l, or

lyne’siledeem.self-beit is

thisualelf-

oraredlievepleisso--

eitherdio-l ex-tion.the

winglf arey willtheir. 2).ces-the

ades-av-

DS I

reasonable to suppose that, under some conditions, pare motivated by a desire to maintain specific positivnegative beliefs they have about themselves (i.e., selfsistency); it is equally tenable that, under some conditmost people are motivated to reduce the perceived aveconsequences of their unwanted behavior (i.e., the “Look”); it is also the case that under some conditiopeople would prefer to think about other positive aspecthemselves rather than face the implications of their beior (i.e., self-affirmation). Put this way, the three maperspectives are viewed neither as synonymous nor slinguistic translations of one another. Rather, they edescribe a distinct and important piece of the overall dnance process and, in doing so, make a unique contribto our understanding of how cognitions about the selfdiate cognitive dissonance arousal and reduction.

To link the various perspectives together and constrcomprehensive model, however, requires examinatiowhat they share in common. We believe the key similais that each perspective on the self in dissonance maspecific assumption abouthow people think about thebehavior. That is, each contemporary model of dissonbegins with the assumption that people behave andattempt to make sense out of what they have done. Wthe various perspectives appear to differ most is inassumptions about the self-relevant information peopleto determine the psychological meaning of their behaWe propose that by focusing on the process by wpeople detect a discrepancy between behavior and band then the processes by which people attempt to rethe discrepancy, it becomes possible to connect the vaparts of the puzzle together and provide an overall pictuthe role played by cognitions about the self in cognidissonance. The SSM attempts to integrate the theoridelineating the potential processes by which peopletheir self-knowledge as an expectancy, a resource, orall when they think, feel, and react to a discrepant beha(Stone, 1999; Cooper, 1999; Stone, in press).

According to the proposed SSM, people engagemultistep process to determine if their behavior is fooimmoral, or otherwise discrepant from important cogtions. Once they have acted, people evaluate their behagainst a standard of judgment, and that standard ofment may or may not relate to a cognitive representatiothe self. The SSM predicts that the evaluation of behamay be based on generally shared, normative considerof what is good or bad, foolish or sensible, moral ormoral or it may be based on personal, idiographicallyconsiderations of what is bad, foolish, or immoral—sdards that are connected to individual representations oself. The model maintains that the standard used to inteand evaluate the meaning of a given behavior determ

SELF-STANDAR

the role of the self and self-esteem in the process of dissonance arousal. Furthermore, once dissonance is aroused, t

le

-,e

-

e

n

f

a

ene

e

f,esf

y

tr

r-f

s

ets

SSM predicts that bringing to mind certain aspects ofself can influence the degree to which people will justheir behavior or use their self-knowledge as a resourreduce their discomfort. The moderating role of self-estin dissonance reduction depends on whether the cognabout the self are positive, self-descriptive, and relatethe behavioral discrepancy. Thus, a fundamental asstion of the SSM is that the role of the self in dissonaarousal and reduction is a function of the type of sattributes and standards made accessible in the contexdiscrepant behavior.

The SSM offers a schematic for understanding thecesses that lead from behavior to unpleasant psycholoarousal and the subsequent reduction of psychologicacomfort. As seen in Fig. 1, the model proposes that dnance arousal takes the following steps: (1)people behavand (2) people interpret and evaluate their behavior.Inessence, they ask, is my behavior foolish, immoraotherwise undesirable?

If personal standardsare situationally or chronicalaccessible in memory, the behavior is compared to oown, idiosyncratic expectancies for behavior. As is detabelow, self-expectancies are directly related to self-estPath 1 of Fig. 1 suggests that the higher a person’sesteem, the more likely it is that the behavior willperceived as foolish and immoral, and the more likelyto lead to dissonance arousal. Dissonance arousal incase is “idiographic” and will be moderated by individdifferences in the content of self-knowledge (e.g., sesteem).

If normative standardsof judgment are situationallychronically accessible in memory, the behavior is compto the perception of what most people in the culture beis foolish or immoral. As seen in Path 2 of Fig. 1, if peoperceive a discrepancy from the normative standard, dnance arousal will be “nomothetic” andwill not be moderated by self-esteem.

The SSM proposes that once dissonance is arousednomothetically (i.e., using normative standards) or igraphically (i.e., using personal standards), people wilperience discomfort and be motivated to seek its reducHow they reduce their discomfort also depends uponcognitions about the self that are made accessible follodissonance arousal. If no other cognitions about the sebrought to mind, the SSM assumes that the discrepancremain accessible and people will seek justification ofbehavior to reduce their discomfort (see Path 1 of FigHowever, if new cognitions about the self are made acsible in the context, then dissonance reduction will takefollowing steps: (1) If the self-attributes that are maccessible arepositiveandrelevantto the behavior in quetion, then they will increase the motivation to justify beh

231N DISSONANCE

-heior (e.g., attitude change). As seen in Path 2 of Fig. 2, ifpositive relevant attributes make self-expectancies accessi-

(2)

rvef Figivedhowith

gni-tain. In

beeruct

t al.,balsteem

ann,rchden-tivetiveertain93),their

ins,tify

RE

ble, people with high self-esteem will showmore justifica-tion of their behavior than people with low self-esteem.If the self-attributes that are made accessible arepositivebutnot relevantto the behavior in question, then they will seas a resource for the reduction of dissonance. Path 3 o2 indicates that if positive irrelevant attributes are perceas self-descriptive, people with high self-esteem will slessjustification of their behavior compared to people wlow self-esteem.

The SSM derives its predictions about the role of cotions about the self in dissonance in part from cerassumptions about the cognitive nature of self-esteemthe cognitive dissonance literature self-esteem hastreated primarily as a cognitive psychological const

FIGU

232 STONE A

FIGURE

.

n

(Aronson & Carlsmith, 1962; Aronson, 1968; Spencer e1993). Cognitive models generally maintain that glomeasures of self-esteem such as the Rosenberg self-escale (Rosenberg, 1979) capture in part how peoplethinkabout themselves (e.g., Rosenberg, 1979; Pelham & Sw1989; Blaine & Crocker, 1993; Brown, 1998). Reseashows, for example, that people with high self-esteem itify positive traits as more self-descriptive than negatraits (Brown, 1986; Campbell, 1990), they rate posiattributes as important to possess and are relatively cthat they possess them (e.g., Campbell & Lavallee, 19and they report smaller chronic discrepancies betweenactual self and ideal self-standards (Moretti & Higg1990). Alternatively, people with low self-esteem iden

1

COOPER

ND

2

-de-een

sitivlowThupos

teeme o

&

ernsin tharkundted

nec-andkingcesoralnds

whaxpet hase.”el,self, buor

penelf-theih to

inelate, isy be

cane pras aanc

arlyd binds957

aus

Al-de-

ostlywas

couldt anlievehave

l

92)rds”

vior.ind,

eirduct.itively toeir

d towhocon-

ingtheirxpec-

n onof at therallyiven, p.rallythey

ocialllec-seetionrep-

d byeirwith

ions

arestan-f theelf-may,ch”al”

DS I

both positive AND negative traits as moderately selfscriptive and report larger chronic discrepancies betwtheir actual and ideal self-standards. They also rate poattributes as important to possess, but people withself-esteem are less certain that they possess them.high self-esteem is characterized by an abundance oftive and confident self-knowledge, whereas low self-esis characterized by a relatively less confident balancpositive and negative self-knowledge (e.g., BlaineCrocker, 1993).

Another important assumption of the model conchow the cognitive aspects of self-esteem may be usedassessment of behavior. Several scholars such as M(e.g., Markus & Kunda, 1986; Markus & Wurf, 1987) aHiggins (1996) have argued that the self is a multifacedynamic organization of knowledge. To make the contion between the complex structure of self-knowledgebehavior, Markus and colleagues proposed the “worself-concept”—the subset of self-knowledge that is acsible in memory at any given moment. In any behavicontext, the content of the working self-concept depe“on what subset of selves was active just before, onhas been invoked by the individual as a result of an erience, event or situation, and most importantly, on whabeen elicited by the social situation at the given tim(Markus & Kunda, 1986, p. 859). In the current modself-esteem differences in the structure and content ofknowledge can play a role in the dissonance processwhether self-knowledge operates as an expectancyresource or is irrelevant to dissonance processes deupon the type of self-relevant cognitions or “working sconcept” people use when interpreting and evaluatingbehavior and when searching for a strategy by whicreduce their discomfort.

The role of specific working self-concept informationthe process of dissonance arousal and reduction, as it rto the major perspectives on the self in dissonancdescribed in the following sections. Whereas there maa number of ways in which working self-knowledgeinfluence dissonance processes, we focus here on thcesses by which cognitions about the self can operateexpectancy, a resource, or be irrelevant to dissonarousal and reduction.

THE ROLE OF SELF-STANDARDSIN DISSONANCE AROUSAL

We begin by noting that Festinger provided an einsight into the process of dissonance arousal specifiethe present model. When speculating on the different kof inconsistencies that lead to dissonance, Festinger (1wrote of at least two kinds: One kind was dissonant bec

SELF-STANDAR

the individual perceived two cognitions as psychologicallyinconsistent and another was dissonant because “culture o

e

s,i-

f

es

,

-

t-

-tads

r

es

o-n

e

y

)e

group standards dictate that they do not fit” (p. 13).though Festinger was not definitive about how peopletermine the presence of an inconsistency (he was minterested in what happened once an inconsistencyperceived), it appears that Festinger believed peopleuse idiographic or shared information to determine thainconsistency was present. As is detailed below, we bethe three major perspectives on the self in dissonancenot strayed much from Festinger’s initial speculation.

Self-Standards as Expectancies in Dissonance Arousa

Self-consistency theory (Thibodeau & Aronson, 19maintains that people rely on their own “personal standafor competence and morality to evaluate their behaWhen personal standards for behavior are brought to mpeople will evaluate their behavior in terms of its fit to thown self-expectations for competent and prudent conMoreover, people with high self-esteem, who hold posexpectations for competent and moral conduct, are likeperceive foolish or immoral behavior as falling short of thexpectancies and consequently feel highly motivaterationalize their actions. People with low self-esteem,hold less positive expectations for competent or moralduct, are likely to perceive the same behavior as confirmtheir expectations and are less motivated to rationalizeactions. Thus, cognitions about the self operate as etancies that mediate the arousal of dissonance.

The self-consistency prediction appears to depend oa critical assumption about the cognitive representationself-expectancy. Self-consistency theory maintains thapersonal standards to which people subscribe are “cultuderived, and largely shared, by most people within a gsociety or subculture” (Thibodeau & Aronson, 1992596). In the SSM we define standards that are “cultuderived and largely shared” as normative becauserepresent the norms and rules held by important sgroups that preserve the integrity and welfare of the cotives to which an individual belongs (Higgins, 1990;Miller & Prentice, 1997). This raises an important quesabout the self-consistency model: If personal standardsresent the perceived behavioral norms that are followemost who live within a particular subculture, how can thuse as a measuring stick for behavior cause peopledifferent levels of self-esteem to reach different conclusabout the same behavior?

One possibility is that when “personal” standardsaccessible, people not only think about the normativedards for behavior, but they also bring to mind aspects oidiographic self-knowledge that underlies their level of sesteem (e.g., Brown, 1986; Baumgardner, 1990). Theyfor example, think about how well they personally “matthe normative standard by bringing to mind their “actu

233N DISSONANCE

rself for a given attribute (Higgins, 1989). According to theproposed model, in order for the use of personal standards

dgeardsog-

pec-ion

of

enteelf-nts ace opre

thess fo

self-theibe-men

in-ting

ss thr heral”)sen

tivet anan

dis-ronirsoembe-thebut

whe“I amormof athed itanc

teemmu

selfas

evane isonicrma

ncyers

to theowl-stione oficallyord-

r hisually

nsionone

rms,smallt anddent.t the

nor-tionbe-

stentally

tly in

s aoom.s so-he ishisgen-

classncy,the

mpe-lt, Almoststionanceof his—issis-

that

ighboutelf-

whatistin-heirronicativePeo-ccep-vior

ND

to activate self-expectancies, idiographic self-knowlemust be used in combination with the normative standduring the interpretation and evaluation of behavior. Cnitions about the self are only likely to operate as extancies in dissonance when people use a representatidiographic self-knowledge to determine the meaningwhat they have done.

The SSM holds that self-expectancies can be represin memory by two self-relevant cognitions: (1) the “sconcept,” which is an actual self-attribute that represeperson’s chronic standing on dimensions like competenmorality, and (2) the normative social standards that resent where a person should or ought to stand ondimensions. People may derive idiographic expectanciebehavior from the relationship between the actualattribute and normative standard in that people expectbehavior to confirm or verify the chronic relationshiptween the actual self and the normative standard for disions like competence or morality. For example, whenduced to tell someone a boring task is really interes(e.g., Festinger & Carlsmith, 1959), a person may assecompetence and morality of the act by comparing his obehavior (e.g., “I just said something stupid and immoragainst his or her self-expectancy or idiosyncratic repretation of how well he or she typically upholds the normastandards for these attributes (e.g., “I am usually a smardecent person—relative to the norms for competencemorality”). In this high self-esteem example, when acrepancy is detected between the behavior and chself-expectancy, dissonance will be aroused. But a pewith a less positive self-expectancy (e.g., low self-estemight not perceive a discrepancy following the samehavior. A person with low self-esteem might conductsame act (“I just said something stupid and immoral”),conclude that it does not represent a discrepancycompared against a less positive self-expectancy (e.g.,not alwaysa smart and decent person—relative to the nfor competence and morality”). In this case, the useself-expectancy to interpret behavior—the activation ofchronic discrepancy between an actual self attribute anrelevant normative standard—does not lead to dissonarousal. Thus, the SSM predicts that in order for self-esto moderate the dissonance arousal process, peopleassess the meaning of their behavior using their ownexpectancies, which can be represented in memorychronic discrepancy between the actual self and the relnormative standard (cf. Higgins, 1990). Dissonancaroused when a given behavior fails to confirm the chrrelationship between the actual self-concept and its notive self-guide.

An example may make this view of self-consisteprocesses clear. Imagine George, a student, who answ

234 STONE A

question incorrectly in his history class. He immediatelyrealizes, with help from the teacher’s scowl and his class-

of

d

r-er

r

-

e

-

dd

cn)

n

se

st-at

-

a

mates’ giggles, that his behavior has not measured upnormative standard for academic competence—the knedge that one should be correct when answering a quein class. In addition, George perceives himself to be onthe best students in class, based on the fact that he typanswers correctly when called upon by the teacher. Accing to our model, George held positive expectancies fobehavior because his self-concept for competence is usperceived to match the normative standard on the dimeof competence—part and parcel of what it means to beof the best students in class. In self-consistency teGeorge’s expectation for competence is based on adiscrepancy between his chronic actual self as a studenthe shared normative standards for being a good stuAlthough he knows he is not perfect, he expects thadiscrepancy between his classroom behavior and themative standard will be small. After answering the quesincorrectly, George will indeed experience dissonancecause his behavior—the incorrect answer—is inconsiwith his self-expectancy—the knowledge that he usumatches the normative standard for answering correcclass.

Now contrast George with his classmate Al who hanegative self-expectancy for competence in the classrWhereas Al understands that being correct in class icially desirable, past experience has taught him thatlikely to fall short of the standard for perfection. Unlikefriend George, Al often misses questions in class anderally does not perform as well as George on mostassignments. In this case, Al’s chronic self-expectawhich is the difference between his actual self andnormative standard on the dimension of academic cotence, is comparatively larger than George’s. As a resudoes not expect to answer questions correctly in classof the time. If he too misses the answer to the same quethat George missed, Al does not experience dissonbecause his negative expectancy—the representationchronic actual self relative to the normative standardconfirmed by his behavior. In the language of self-contency, Al’s behavior is consistent with his expectationhe is not a good student.

The SSM proposes that in order for people with hversus low self-esteem to reach different conclusions athe same behavior, they must bring to mind different sknowledge when they attempt to interpret and evaluatethey have done. One representation of the self that dguishes people with high from low self-esteem is tself-expectancies, defined above as the size of the chdiscrepancy between their actual self and the normstandards for attributes like competence and morality.ple may use their self-expectancies as a latitude of atance/rejection for determining whether a given beha

COOPER

represents a discrepancy (cf. Fazio, Zanna, & Cooper, 1977;Rhodewalt, 1998). As illustrated by path 1 of Fig. 1, we

e acs to

eoplncethat-iseon,sten

be-pt asheseoplcallyardfor

m-tionsbe-ep-ude

anceeopthei

ss oingsis-werif helati-zio erecthisntsy bes a

be

highr beteems foconintonicthe

ssinty of

the&

anc

allyn thepre-n theio-causen ofaticSSMss of

al

operd theogni-stify

theoralec-valu-haredents

ouldheirt theynsid-ught

eticaleredards

s arelture.ecu-ildrenrtaingoesearnativecur-

re-hil-are

ownThatlture,ned,nted,of the

anceisely

DS I

predict that when personal standards for behavior arcessible, people will use their chronic self-expectancieassess the meaning of their behavior. As a result, pwith high self-esteem will experience more dissonacompared to people with low self-esteem following actsinvolve deviations from the norms for “. . . lying, advocating a position contrary to their own beliefs, or otherwacting against one’s principles” (Thibodeau & Arons1992, p. 592). In our view, these actions are inconsiwith the expectancies of people with high self-esteemcause they fall outside of the latitude of what they accean accurate reflection of their self-concept. In contrast, tactions are more consistent with the expectancies of pwith low self-esteem because their actual self is chronimore discrepant from the perceived normative standwhich provides them with a wider latitude of acceptancetheir behavior (Rhodewalt & Agustsdottir, 1986; Baugardner, 1990). A given act can promote different reacfrom people with high and low self-esteem when thehavior is perceived to fall outside of the latitude of acctance for those with high self-esteem, but within the latitof acceptance for those with low self-esteem.

The use of self-expectancies as latitudes of acceptrejection in the assessment of behavior may also lead pwith low self-esteem to experience dissonance whenbehavior is highly positive (e.g., an unexpected succea difficult task, see Aronson & Carlsmith, 1962). Returnto our student Al for a moment, according to self-contency theory (e.g., Aronson, 1992), if Al were to anscorrectly in class, he could experience dissonanceperceived his behavior as falling outside of the uppertude of acceptance for his academic competence (Faal., 1977; see Rhodewalt & Agustsdottir, 1986). His coranswer to a difficult question would be discrepant fromself-expectancy that he is not a good student; it represeself-view that contradicts previous experience and madifficult to maintain in the future (cf. Swann, 1990). Aresult, dissonance would be aroused, and Al wouldmotivated to reduce his discomfort.

Note that the SSM does not suggest that people withand low self-esteem have different standards for theihavior; to the contrary, we assume that both self-esgroups strive to uphold the same normative standardtheir behavior. Instead, we are hypothesizing that thetext in which a discrepant behavior occurs can bringworking memory different cognitive aspects of the chroself-knowledge that underlies self-esteem. The key touse of chronic self-expectancies in the process of assebehavior, however, lies in the simultaneous accessibiliidiographic self-attributes and normative standards incontext of a given act (cf. McGregor, Newby-Clarke,Zanna, 1999). For self-esteem to moderate disson

SELF-STANDAR

arousal, the normative standard cannot be theonly criterionby which people assess the meaning of their behavior. In

-

e

t

ee

,

/lern

t

a

-

r-

g

e

addition to the norms, there has to be an idiographicbased, unique conception of the self-concept involved ievaluation of behavior (e.g., actual self-attributes). Wedict that making personal self-standards accessible icontext of a discrepant act will bring to mind both idgraphic self-attributes and normative self-standards bepeople are focused on their own, personal interpretatiowhat they have done. If they do not use idiosyncrself-attributes in the assessment of their behavior, thepredicts that self-esteem will not moderate the procedissonance arousal.

Self-Standards as Shared Norms in Dissonance Arous

From the New Look perspective on dissonance (Co& Fazio, 1984; Cooper, 1999), personal standards anself-expectancies they represent are not what cause ctive dissonance arousal. Instead, the motivation to judiscrepant behavior stems from the need to upholdsocially accepted norms that govern competent and mbehavior. The New Look (Cooper & Fazio, 1984) persptive on dissonance tacitly assumed that the hedonic eation of behavioral consequences was substantially sby members of the culture. For example, when studvolunteer to convince a Dean’s committee that it shadopt a policy that is at variance with their own and tcolleagues’ opinions, they can be reasonably sure thahave brought about a consequence that their society coers to be unwanted. Thus, when Cooper and Fazio brothe concept of aversive consequences to the theortable, they were considering the kinds of events considto be unwanted by most people in the culture. The standthat are invoked to determine that the consequenceaversive are the shared, normative standards of the cu

Fazio and Cooper (1983; Cooper & Fazio, 1984) splated that the ontogeny of dissonance was based on chlearning, via the reactions of their caregivers, that cebehaviors have painful consequences. A child whoabout the house breaking toys, glass, and furniture will lthat the caregivers respond to such behaviors with negsanctions. The possibility of unique or pathological ocrences notwithstanding, most caregivers in our culturespond similarly to certain types of behaviors in their cdren. Children learn that certain types of behaviorsnegatively sanctioned, whether they occur in theirhouse, in their neighbor’s house, or on the playground.our behaviors have generally shared meaning in the cuthat certain types of behaviors are negatively sanctioand that we share common understanding of unwaaversive events are assumptions that lie at the heartNew Look model of dissonance (Cooper,1999).

From our current vantage point, we consider dissonto be a nomothetic process in the New Look model prec

235N DISSONANCE

because it assumes that the assessment of behavior is basedon standards that represent rules or norms of an important

inderrdleey

n asativeh oanddualusal

iesr ase anainpron erowshisith aAl,arityhashisself-y at

toringdoest byst 3

of Aself-owl-ctuarow

ecteisof atheoot

e-metheviorullyausestiny fore is

, I’m

ent ofayeris

ect it,his

ex-ative

thethat

ouldwin

okwhen

nst aaviorartshicess-

e thatidio-ge,

dardfors in

us ontheirself-isso-971;

sed

nceh in-elf-rma-

bilitytheingstan-d in-venessin-lf-at-

anddel,nedd thebe-

ND

group, such as parents, “friends like mine,” or “adultssociety” (Higgins, 1990). According to the SSM, in orfor most people to experience dissonance arousal regaof the idiographic content of their self-knowledge, thmust focus on the relevant normative standards whesessing the meaning of their behavior. When the normstandards for behavior are accessible, people with higlow self-esteem will use the same criterion to interpretevaluate their actions. As a result, we predict that individifferences in self-knowledge will not mediate the aroof cognitive dissonance.

To illustrate how people with different self-expectancfor behavior could come to the same conclusion aftesessing their actions, we return to our students GeorgAl. This time, however, George and Al are in a domdepicted by Aronson (1992) to describe dissonancecesses—basketball. On the school team, George is aceptional shooter; he tends to make 80% of his free thduring practice. Although his goal is to make 100% ofshots, 80% is a high average and it provides George wpositive expectation for his free-throw shooting ability.on the other hand, only shoots about 50% from the chstripe. He too would like to shoot 100%, but practicetaught Al that he can only expect to hit about half ofattempts. Consequently, Al has a relatively negativeexpectancy for his free-throw shooting ability. One dapractice, Al and George both shoot 50%. Accordingself-consistency (Aronson, 1992), missing 5 of 10 dupractice confirms Al’s negative self-expectancy and henot feel dissonance. George, however, is very upsehitting only 5 of 10 because he expected to sink at leamore buckets. Note that what mediates the reactionsand George in this situation is the use of a chronicexpectancy to evaluate their performance, i.e., the knedge of where each usually stands (their chronic aselves) relative to the normative standard for free-thshooting. George is only upset if he realizes that he expto hit 80%; Al is only comforted if he realizes that 50%his usual performance. The simultaneous accessibilitychronic actual self-attribute, the idiographic element ofself-expectancy, and the relevant norm for free-throw shing has determined for whom dissonance is aroused.

Now place Al and George in a slightly different frethrow shooting context: at the free-throw line in a gawith their school behind by 1 point and no time left onclock. In this situation, the normative standard for behais also 100% perfection—players in this situation are faware that they have to sink the last free throw becmissing the shot means losing the game. The interequestion here is for Al, who has a negative expectanchis free-throw shooting ability. If he misses and the gamlost, does he feel dissonance? Or does he think, “Well

236 STONE A

not usually good from the line, so I guess missing the showill not upset me.” We predict that in this situation, the

ss

-

r

-d

-x-

l

l

d

-

g

normative standard carries the weight in the assessmthe missed shot. That is, Al knows that a basketball plshould hit the winning shot if given the opportunity. Hteammates expect it, his parents in the audience expand the other team fears it! The point is that despitenegative expectancy for free-throw shooting, Al wouldperience dissonance because, in this situation, the normstandard for behavior—hitting the winning shot—ismost accessible standard for his behavior. We imagineAl and George, and any other player in that situation, wexperience dissonance if they missed the opportunity tothe game.

In sum, as illustrated in path 2 of Fig. 1, the New Lorevision of dissonance assumes that dissonance occurspeople assess the outcomes of their behavior agaishared or normative representation of what their behshould or ought to have produced. The New Look depfrom self-consistency in its assumption that idiograpself-attributes do not enter systematically into this assment process. Based on that assumption, we proposmost people will experience dissonance, regardless ofsyncratic differences in the content of their self-knowledwhen they are primarily focused on the normative stanrelevant to the situation. Put simply, when the basisjudgment is the same for everyone, individual differenceself-esteem will not moderate dissonance arousal. Foca normative standard to determine the meaning ofbehavior may help to explain why, in some studies,esteem did not moderate justification of a classic dnance-arousing behavior (e.g., Cooper & Duncan, 1Steele et al., 1993; Stone, 1999).

Sociocognitive Factors That Determine the Standard Ufor Assessing Behavior

In the above examples of George and Al, the differebetween using personal standards for behavior, whicclude the simultaneous accessibility of idiographic sknowledge and normative standards, and the use of notive standards alone, which do not include the accessiof idiographic self-knowledge, is in part a function ofperspectivethey were induced to take when interprettheir behavior. The SSM proposes that when personaldards for behavior are accessible, attention is focuseward on one’s own, idiosyncratic interpretation of a giact. It is similar to a state of objective self-awaren(Wicklund, 1975), during which introspection initiatesterpretation and evaluation of behavior using the setributes most relevant to the context (e.g., competencemorality; see Aronson, 1992). According to the mointrospection can bring to mind self-expectancies, defias the chronic relationship between one’s actual self anrelevant normative standards for behavior. However,

COOPER

tcause people with high and low self-esteem hold differentexpectancies for their behavior, a state of self-awareness

eac, theducedisso.

stans thmeaan-ss,ent,the

. Thele tavioeerow-hate odgeteadudetivefrombo

oti-

ce ithefo

ougducle, inory

alityecifieed-tic”;bacpa-

emain-highse to-es-thanelf-usebe

ir insymhigh

wasr in-

&

peri-uctedicatedeforeeo-at theDun-high,ss ofim-isso-

sibilityoun-feed-thatigm;out

(pro-kingmayauses ofativech

a stifi-c antsc saysr

e useo r byp itteda In onee durem 993).F s as ownp com-p d tow medD thep ed tot theiro their“ then ed tot ietalp s thatm wing

DS I

may cause people with high versus low self-esteem to rdifferent conclusions about what they have done. Thusaccessibility of personal standards for behavior may ina state similar to self-awareness that causes morenance among people with high versus low self-esteem

In comparison, when normative rather than personaldards are accessible in the context, the SSM proposepeople take the perspective of others in assessing theing of their behavior. The accessibility of normative stdards engages a state similar to subjective self-awarenewhich attention is focused outward toward the environmbut in this case, attention is directed specifically towardattitudes, values, and opinions held by important othersaccessibility of the normative standards causes peoptake the perspective of others and assess their behbased on how important people, such as parents or pwould interpret and evaluate what they have done. Hever, because attention is focused on a perspective texternal to the self, when people take the perspectivothers, we hypothesize that idiosyncratic self-knowleand expectancies do not necessarily come to mind. Inspeople assess how their behavior deviates from the attitvalues, or opinions held by others—the relevant normastandards. If the behavior is perceived to be discrepantthe perceived norms, then dissonance is aroused andhigh- and low-self-esteem individuals will become mvated to justify their behavior.

Early research on the role of self-esteem in dissonanconsistent with these assumptions. One way in whichaccessibility of personal versus normative standardsbehavior was influenced in the early research was thrthe cover stories or feedback procedures used to inself-esteem and counterattitudinal behavior. For exampa study often cited as support for self-consistency theGlass (1964) provided participants with false personfeedback designed to raise or lower self-esteem. Spcally, participants who received positive self-esteem fback were told they were “considerate and sympatheparticipants who received negative self-esteem feedwere told they were “inconsiderate” and “lacking the cability for leadership (i.e., conformist).” After the self-estemanipulation, participants were induced to distribute pful electric shocks to a confederate under conditions ofor low choice. The results showed that when they chodeliver the painful shocks, participants with high selfteem derogated the victim of their aggression moreparticipants with low self-esteem. According to the sconsistency interpretation, harming the confederate camore dissonance for participants with high self-esteemcause their aggressive act was inconsistent with theduced positive self-expectancies for considerate andpathetic behavior. To reduce dissonance, those with

SELF-STANDAR

self-esteem justified their aggression by derogating the victim. In contrast, participants with low self-esteem did not

tc

h

-

-atn-

in

or

s,

isf

,s,

th

s

rhe

,

-

k

d---

derogate the confederate because harming the victimconsistent with their induced negative expectancies foconsiderate and conformist behavior (cf. ThibodeauAronson, 1992).

Compare the effect for self-esteem in the Glass exment to another relevant self-esteem experiment condby Cooper and Duncan (1971). These researchers replthe manipulation of self-esteem used by Glass (1964) binducing participants, under high or low choice, to vidtape a counterattitudinal message. Despite evidence thself-esteem manipulation was successful, Cooper andcan found that when choice to make the advocacy wasall participants justified the discrepant essay regardletheir level of self-esteem. How could this occur when silar procedures produced self-esteem moderation of dnance processes in the research by Glass? One posconcerns how participants were induced to write their cterattitudinal essay after they received the self-esteemback. Specifically, the investigators used instructionsare quite common in the induced compliance paradthey said, “We think we know how most students feel abthis issue” and “We have enough essays favoring theattitudinal) position; what we need now are essays tathe other position.” Differences in induced self-esteemhave had no effect on attitude change in this study becthese statements, in effect, shifted participants’ focuattention from their personal self-standards to the normstandards for behavior:moststudents would not write su

n essay, but you, the participant, did. As a result, juation was a function of perceived choice and participhanged their attitudes toward the content of their esegardless of their induced level of self-esteem.

Our recent research has attempted to investigate thf personal versus normative standards for behavioriming one or the other after participants have commclassic dissonance-arousing act (e.g., Stone, 1999).xperiment, participants completed a free choice proceodeled after Brehm (1956; see also Steele et al., 1ollowing the decision but before they rerated the itemecond time, participants were primed for either theirersonal standards or for the normative standards foretent behavior. Specifically, participants were askerite a short description of a fictional target person naonald who exemplified the attribute of competence. Inersonal standards condition, participants were instruct

hink about the standards for competent behavior fromwn perspective and then to describe Donald usingown personal standards” to define competence. Inormative standards condition, participants were direct

hink about the standards for competence from a socerspective and to describe Donald using “the standardost people would use” to describe competence. Follo

237N DISSONANCE

-he priming task, all participants were then asked to rate thehoice alternatives a second time. The attitude change data

ardsheirelf-sim

n-een

ativesimdatuchdard

promaageteem

thel wade

ughtancedardbe-ometes,

c ?the

d swepta t arei thert ed tot ccurw romt alin-s ow-e , thea nre-l com-f

siona spec-t n eta ncea tive“ be-h note s ap romo d fora al.,1 in thec thet -l m-sa sso-n mptt 23–2 thes utesa ll re-m r andm usti-fi bem ofs a nor-m f theira

oesp s ford self-a , thel heird havea y tou com-p owl-e . Asn itive

in theotherscribvein,er ans, an

), idio, percrep-theticy alsteemativel anof a

sonaln thetweeide,s jusuide

ND

showed that while thinking about personal self-standcaused participants with high self-esteem to justify tchoices significantly more than participants with low sesteem, thinking about normative self-standards led toilar levels of justification inboth self-esteem groups. Aother experiment replicated this interaction pattern betwself-esteem and the priming of personal versus normself-standards following a counterattitudinal essay taskilar to that used by Cooper and Duncan (1971). Theseprovide some preliminary evidence for our model inasmas they indicate that making accessible personal stanfor behavior appears to engage self-expectancies andmote self-esteem moderation of dissonance, whereasing accessible normative standards for behavior engsimilar dissonance processes in high- and low-self-esindividuals. Regardless of what actual behavior beganprocess, self-esteem moderation of dissonance arousaa function of the type of self-attributes and standards maccessible following the discrepant act.3

THE ROLE OF SELF-STANDARDS IN DISSONANCEREDUCTION PROCESSES

To this point, we have argued that self-relevant thocan be an important factor in the process of dissonarousal when people use self-relevant criteria (i.e., stanthat apply to them) to determine whether or not theirhavior represents a threat to the self or an aversive outcHow then can the accessibility of positive self-attribualso central to the self-affirmation process,reducethe dis-omfort and motivation to justify a discrepant behaviorIn our view, self-affirmation occurs when focus on

3 The distinction between personal and normative self-standardscurrent model is similar to the distinction between the own versusperspective people can take on ideal and ought self-discrepancies deby Self-discrepancy Theory (e.g., Higgins, 1989, 1990). In a similarHiggins (1990) suggested that dissonance paradigms like FestingCarlsmith (1959) may make accessible specific matches, mismatchenonmatches between normative standards (e.g., the helpful subjectgraphic self-guides (e.g., morality), and social context guides (e.g.ception of responsibility). However, whereas Higgins used self-disancy theory to describe one specific example of what we call nomodissonance, the current model expands greatly on this analysis bdescribing the type of discrepancy that is likely to cause self-esdifferences in dissonance arousal. Moreover, given the role of normstandards in the present model, the distinction between personanormative discrepancies may represent two very specific examples“ought self-discrepancy.” That is, when people think about their perstandards for behavior, they are focused on a discrepancy betweebehavior and what Higgins (1990) has defined as the relationship bea chronic “own-actual-self” and “own-ought-normative” self-guwhereas focus on the normative standard in the current model involvea comparison between behavior and an “own-ought-normative” self-g

238 STONE A

st

We see the tenets of self-discrepancy theory as having broad implicationfor understanding dissonance processes and we discuss some of theissues under Implications for the Model under Other Dissonance Processe

-

-a

s-

k-s

as

s

.

iscrepancy between behavior and self-standards isside by the accessibility of positive self-attributes tha

rrelevant to the behavioral discrepancy in question. Rahan reducing dissonance by changing cognitions relathe discrepant behavior, self-affirmation processes ohen new cognitions about the self distract people f

hinking about the dissonant act (Stone et al., 1997; Gky, Stone, & Cooper, 2000). A crucial assumption, hver, is that for self-affirmation processes to operateffirming actual self-attributes must be irrelevant or u

ated to the behavioral discrepancy that caused the disort in the first place.

To make this prediction clear requires some discusbout how dissonance arousal is viewed from the per

ive of self-affirmation theory (Steele, 1988; J. Aronsol., 1999). Self-affirmation maintains that dissonarousal is a function of a threat to the globally posiintegrity” or cohesiveness of the self. Precisely howavior constitutes a threat to self-integrity, however, isxplicitly defined in the theory; anything that threatenositive self-image, including negative evaluations fthers or negative life events, can also motivate a neeffirmation of the self (Steele, 1988; J. Aronson et999). Nonetheless, Spencer et al. (1993) suggest thatlassic dissonance paradigms, if attention is drawn tohreat inherent in a given behavior, “. . . all people (regardess of their level of self-esteem) will try to affirm theelves by deflecting or diminishing the threat . . . if peoplere not focused on their self-resources following (diance), both high- and low-self-esteem people will atte

o rationalize or dismiss the (dissonance) per se” (pp.4). We believe these observations imply that fromelf-affirmation perspective, unless positive self-attribre made accessible in the situation, most people wiain focused on the threat represented by their behavioost will be motivated to reduce dissonance via self-j

cation. From our perspective, this prediction can onlyade if we assume that people with different levels

elf-esteem come to the same conclusion, based onative self-standard, when they assess the meaning octions.Our interpretation notwithstanding, self-affirmation d

redict that self-relevant thought can provide resourceissonance reduction. Specifically, the more positivettributes that are accessible in working self-knowledge

ess people will have to use justification to reduce tiscomfort. Because people with high self-esteemvailable more positive attributes, they are more likelse a self-relevant thought for dissonance reductionared to people with low self-esteem, whose self-kndge presumably contains fewer positive self-attributesoted above, this assumption about the function of pos

ed

dd-

-

o

dn

irn

t.

COOPER

elf-attributes in dissonance is substantially different fromhe assumption made by the self-consistency perspective,

sses.

toeyn as(e.gioralle in

weuldthe

telyorgeht’s

hisat-

—At rever,

tteris?

lpougadeprethe

self-e’sstifys int artainstify

thatvantuseffectctionhen

ecifinalcomtify. Inthey thmosus,ards

toy fo

highas alow

howneivewn,eemde-s a

ourceut ifople

ionitivermi-nanceeempantakes

sthe

p be-t sitives ron-s per,S r ther theJ ed ina on-d itys ped.P ther ken.P thatd hicht t ofp theird wereu high-d reada thed utesw t toldt s—t priatet ationt itives

ther ma-n col-

DS I

which assumes that positive self-knowledge is likelyincreasethe need for self-justification. We believe the kto understanding how positive self-attributes can functioa resource (e.g., self-affirmation) or as an expectancyself-consistency) in dissonance begins with the behavrelevancy of the positive self-attributes made accessibthe context of a given behavior.

To illustrate the importance of behavioral-relevancy,return to our two students, George and Al. How woself-affirmation be possible after George answersteacher’s question incorrectly? Imagine that immediafollowing his academic embarrassment, Al reminds Geabout his winning free throw during the previous nigbasketball game. We suggest that by bringing to mindclutch performance on the basketball court—a positivetribute presumably unrelated to academic competencehas provided George with a positive self-resource thaduces the need to justify his incorrect answer. Howewhat if Al’s comment went right to the heart of the maby reminding George what a good student he usuallyWill this highly relevant and positive information heGeorge reduce his dissonance? We propose that, althAl’s reminder about George’s academic ability has mpositive self-attributes accessible, its relevance to thevailing discrepancy between his incorrect answer andnormative standard for classroom behavior will not beaffirming. Rather, Al’s comment will only make Georggaffe in class seem worse, and George will need to juhis behavior in order to effectively reduce dissonance. Athis case, making positive self-attributes accessible tharelated to the behavioral discrepancy are likely to mainor increase dissonance arousal and the motivation to jubehavior.

As illustrated by Path 2 of Fig. 2, the SSM predictsthe accessibility of positive self-attributes that are releto the discrepant act should increase discomfort becainvokes the standards for behavior. Furthermore, the eof relevant positive self-attributes on dissonance redushould be moderated by self-esteem when, and only wthey activate personal self-standards for behavior. Spcally, if relevant positive attributes bring to mind persostandards for behavior, then self-expectancies may beaccessible, and people with high self-esteem will justheir behavior more than people with low self-esteemcontrast, if relevant positive attributes bring to mindnormative standards for behavior, then, as predicted bSSM, self-esteem will not moderate the process andpeople will be motivated to justify their behavior. Thpositive attributes that bring to mind the self-standrelevant to the behavioral discrepancy are not likelyprovide resources for dissonance reduction, especiallpeople with high self-esteem.

SELF-STANDAR

In contrast, as illustrated by Path 3 of Fig. 2, if thesituation brings to mind positive attributes that are unrelated

lb

.,-

l-

h

-

e

its

,-

e

et

r

to the discrepant act, the SSM predicts that people withself-esteem should be able to use this informationresource against their discomfort, whereas people withself-esteem will not. As noted above, research has sthat people with high and low self-esteem do not percpositive traits to be equally self-descriptive (e.g., Bro1986; Baumgardner, 1990); those with high self-estshould perceive irrelevant positive attributes as morescriptive of the self than people with low self-esteem. Aresult, irrelevant positive attributes can be used as a resto reduce discomfort by people with high self-esteem, bthe traits are not perceived as self-descriptive by pewith low self-esteem, they will have to rely on justificatto reduce their discomfort. Thus, the relevance of posself-attributes to a discrepant behavior is a critical detenant of the use of self-esteem as a resource for dissoreduction. The SSM predicts that the more self-estpeople have, the less they will need to justify a discreact (e.g., by attitude change) when the situation maccessible positive self-attributes that areunrelatedto thetandards people use to evaluate their behavior.Although there is currently no direct evidence for

redictions sketched above concerning the interplayween self-esteem and the behavioral relevancy of poelf-attributes on dissonance reduction, studies by J. Aon, Blanton, and Cooper (1995) and Blanton, Cookurnik, and J. Aronson (1997) do provide evidence fo

ole of behavioral relevancy in dissonance reduction. In. Aronson et al. (1995) experiment, participants engagcounterattitudinal essay-writing task in which, under c

itions of high choice, they wrote that their univershould reduce spending on facilities for the handicaparticipants were then asked if they would like to see

esults of a personality test that they had previously taarticipants were told that they could read paragraphsescribed their standing on positive self-domains in w

hey had scored particularly high. Included in the lisaragraphs were those that were directly related toissonant act (e.g., compassionate) and those thatnrelated (e.g., creative). The results showed that inissonance conditions, participants did not want tobout their positive self-attributes that were relevant toiscrepant behavior. Despite the fact that these attribere positive, participants chose to read paragraphs tha

hem how wonderful they were on irrelevant attributehose that were not related to the self-standards approo that behavior. Participants appeared to select informhat could take their focus away from the relevant poself-attribute.In the subsequent study by Blanton et al. (1997),

elevancy of the positive feedback was systematicallyipulated and attitudes toward the essay topic were

239N DISSONANCE

ected as the measure of dissonance reduction. As suggestedy the J. Aronson et al. (1995) experiment, when the posi-

act,(i.e.se toack

e diswrit

rastviorcom

bys rethe

itivesitive, no(see

pliedo ntheianceelf-

s prpecthe

ir avelThisokeFory beancor

ajorm inn hocrepdis-

urrely oea

vanttheyn th

oren-ene

and., ansue;lson,rpreteter-ssing5) orosedpen-heAn

n ofthe

step

onalTheandcol-

erentivelyl. Fort dis-ed asfromemo-that

nor-l stateingrep-eri-

ch aslabelivelysub-isso-

s be-tion,

theirec-

therheired toer &so-sal isple tohav-and

ND

tive feedback was relevant to the counterattitudinalparticipants showed significantly more attitude changedissonance reduction) relative to participants who chowrite the counterattitudinal essay but received no feedbThis suggests that the positive feedback increased thsonance arousal beyond what was experienced by justing the essay under conditions of high choice. In contwhen the positive feedback was irrelevant to the behaparticipants showed significantly less attitude changepared to high choice control participants. As predictedself-affirmation theory (Steele, 1988), dissonance waduced by positive self-information, but, as predicted bySSM, participants were only self-affirmed when the posfeedback was irrelevant to the discrepant behavior. Pofeedback relevant to the discrepant act caused moreless, need to confront the implications of the behavioralso Galinsky et al., 2000).

In sum, the SSM suggests that self-affirmation, as apto a dissonance situation, is a process whereby people dresolve the discrepancy between their behavior andself-standards (Stone et al., 1997). Normally, dissonreduction would occur in a manner dictated by the sstandard that was made accessible in the situation, adicted by the self-consistency and the New Look perstives. But when positive features of the self unrelated todiscrepancy become accessible, people may alter thetention from the implications of their actions and reinstead in some of their other outstanding attributes.may reduce negative affect (Galinsky et al., 2000) or invtrivialization of the discrepant act (Simon et al., 1995).people with high self-esteem, affirmation of the self maan efficient and satisfactory means of reducing dissonwithout going though the work of changing cognitionsbehaviors related to the source of arousal.

IMPLICATIONS OF THE MODEL FOR OTHERDISSONANCE PROCESSES

The SSM was designed to integrate the three mperspectives on the role of the self and self-esteedissonance processes. The emphasis in the model opeople determine that their behavior represents a disancy may have broader implications for understandingsonance phenomena. For example, whereas the cmodel describes the processes by which people may repersonal versus normative standards to determine the ming of their behavior, these may not be the only relecognitions people use to interpret and evaluate whathave done. Under some conditions, people may focus ofit between their behavior and an important attitudeparticular belief—a more intermediate level of action idtification (Vallacher, 1992; see Stone, in press). This t

240 STONE A

of the SSM implies that it is possible for dissonance to be afunction of a perceived discrepancy between behavior and

,

.--,,-

-

t

otr

e--

t-

e

w-

ntnn-

e

t

important cognitions that range from highly abstractglobal (e.g., the self) to highly concrete and specific (e.gattitude toward an object or a specific belief about an issee Harmon-Jones, Brehm, Greenberg, Simon, & Ne1996). Furthermore, the cognitions people use to inteand evaluate the meaning of their behavior may be dmined by the contextual goals or constraints on proceset forth in the situation (see Swann & Schroeder, 199by the chronic goals and constraints on processing impby variables located within the person, such as the prosity for self-monitoring (Snyder & Tanke, 1976) or tstability of self-esteem (see Kernis & Waschul, 1995).important direction for research concerns examinatiothe various mechanisms by which people determinemeaning and significance of their behavior—the firsttoward dissonance arousal.

Another direction for research concerns the emotiexperience of dissonance (Elliot & Devine, 1994).relationship between our description of self-attributesself-standards and those described by Higgins and hisleagues in Self-discrepancy Theory suggests that difftypes of behavioral discrepancies may lead to qualitatdifferent emotional labels applied to dissonance arousaexample, Higgins and his colleagues have shown thacrepancies from ought self-standards are experiencagitation-related emotions, whereas discrepanciesideal self-standards are experience as dejection-relatedtions (e.g., Higgins, 1989). This leads to the predictionthe dissonance arousal following a violation of ought-mative standards may be experienced as an emotionasuch as anxiety or guilt, whereas dissonance followviolation of personal self-standards, to the degree theyresent an “ideal self-standard,“ may be specifically expenced as dejection- or frustration-related emotions sushame or embarrassment. Furthermore, the emotionalapplied to dissonance arousal may delineate qualitatdifferent states of dissonance motivation, which maysequently influence the strategies people choose for dnance reduction (Stone et al., 1997). The connectiontween specific behavioral discrepancies and emoincluding how people label dissonance arousal and howlabels influence dissonance reduction, are exciting dtions for future research.

The labeling of dissonance arousal points to anoimportant question: How can people misattribute tarousal to an external source if self-standards are usassess behavior (Zanna & Cooper, 1974; see CoopFazio, 1984)? According to the New Look model of disnance (Cooper & Fazio, 1984), after dissonance aroulabeled as negative, a situational cue can cause peomisattribute their arousal to a source other than their beior. This proposition was derived from the Schacter

COOPER

Singer (1962) model of emotion, which specified thatarousal can be undifferentiated when its current source is

speissoisatancencyare

ir usson,givense.urc

assivandselfan

tive

sely a

ss thpro-weesituancdarof thctions o

sych, wesso-dard

eauownwhicWeworctuaThessibself-r, aniveas

oreom-

tan-

opletions84;

anc-pledel,dardusal

tiveancethes re-rrele-hightiontheessi-so-dis-ffectof

arch.ichehav-over

nom-ions, wes. Bynter-ancy,for itsove

tant

bleson.

A makes

A as aal

A dis-

A : An

DS I

ambiguous. According to our model, the salience of acific self-standard may disambiguate the source of dnance arousal, which may then reduce the ability to mtribute the arousal or select a strategy for dissonreduction that does nothing to resolve the discrepaHowever, it is also possible that when self-standardsaccessible, people may not be consciously aware of thein the assessment of their behavior (e.g., Nisbett & Wil1977). Under some conditions, the assessment of abehavior may be very quick, like a conditioned respoWhen this is the case, people may not be know the soof arousal because the processing is automatic and p(e.g., Wegner & Bargh, 1998). The role of automaticityunconscious processing in how people use differentattributes for the interpretation of behavior representsother intriguing direction for future research on cognidissonance.

SUMMARY AND CONCLUSIONS

We have argued that the contemporary views of thein dissonance have at least one common bond—themake important assumptions about how people assemeaning and significance of their behavior. We haveposed that this process involves a comparison betbehavior and a self-standard made accessible by thetion in which the behavior was enacted. When a discrepis detected between behavior and a relevant self-standissonance is aroused. However, we suggest that eachmajor perspectives on the self makes a unique predithat derives from different assumptions about the typeself-attributes and self-standards used to assess the plogical significance of a given behavior. Consequentlypropose that the role of cognitions about the self in dinance depends on the type of self-attributes and stanpeople use to understand and evaluate their actions.

According to the SSM, self-consistency (e.g., Thibod& Aronson, 1992) emphasizes that people use theirpersonal standards when assessing their behavior,brings to mind chronic, idiographic self-expectancies.suggest that self-expectancies are represented in theing self-concept as the relationship between chronic aself-attributes and normative standards for behavior.model predicts that when personal standards are accein a behavioral context, people use their idiographicexpectancies to assess the meaning of their behaviopeople with high self-esteem will be more likely to percebehavior in one of the classic dissonance paradigmssignificant discrepancy. As a result, they will be mmotivated to reduce dissonance via self-justification cpared to people with low self-esteem.

In contrast, the SSM predicts that when normative s

SELF-STANDAR

dards are primarily accessible, individual differences inself-esteem will not moderate dissonance processes becauB

---

.

e

n

ee

--

flle

na-yd,e

fo-

s

h

k-l

le

d

a

the shared evaluation of the behavior will cause most peto perceive a discrepancy. According to the assumpdiscussed in the New Look model (Cooper & Fazio, 19Fazio & Cooper, 1983), normative standards imply stions from others that will be perceived similarly by peowith different self-expectancies. However, in the mopeople must be primarily focused on the normative stanfor the shared evaluation to carry the weight in the aroprocess.

Finally, the SSM predicts that the accessibility of posiself-attributes will operate as a resource for dissonreduction only if the self-attributes are irrelevant todiscrepancy between behavior and the self-standardsponsible for causing the dissonance arousal. When ivant positive attributes are accessible, people withself-esteem will benefit more and show less self-justificacompared to people with low self-esteem. Some ofmechanics of this prediction, such as whether the accbility of irrelevant positive self-attributes reduces disnance by shifting attention away from the behavioralcrepancy (Blanton et al., 1997), reducing negative a(Galinsky et al., 2000), or by trivializing the importancethe discrepancy (Simon et al., 1995), await future rese

It is our hope that by illuminating a process by whpeople assess the meaning and significance of their bior, it may be possible to go beyond the current debatewhich revision is the best account of dissonance pheena. Although the debate over the underlying motivatfor dissonance has been productive for the disciplinefavor more synthesis among all the various perspectivefocusing on the cognitive processes by which people ipret their behavior, detect the presence of a discrepexperience and label their arousal, and seek a strategyreduction, we believe that dissonance theory can mforward in new directions that continue to present imporinsights into human social behavior.

REFERENCES

Aronson, E. (1968). Dissonance theory: Progress and problems. In Aet al. (Eds.),Theories of cognitive consistency: A sourcebook(pp. 5–27)Chicago: Rand McNally.

ronson, E. (1992). The return of the repressed: Dissonance theorya comeback.Psychological Inquiry,3(4), 303–311.

ronson, E., & Carlsmith, J. M. (1962). Performance expectancydeterminant of actual performance.Journal of Abnormal and SociPsychology,65, 178–182.

ronson, J., Blanton, H., & Cooper, J. (1995). From dissonance toidentification: Selectivity in the self-affirmation process.Journal ofPersonality and Social Psychology,68(6), 986–996.

ronson, J., Cohen, G. L., & Nail, P. R. (1999). Self-affirmation theoryupdate and appraisal. In E. Harmon-Jones & J. Mills (Eds.),Cognitivedissonance: Progress on a pivotal theory in social psychology(pp.

241N DISSONANCE

se127–148). Washington, DC: American Psychological Association.

aumgardner, A. H. (1990). To know oneself is to like oneself: Self-

gy,

s

, ande the

es inR. F.:

ingsation

tives

men

ofh lowow

pantds.),

-

of thills

andical

on of

In L.

es in

ss.

of

. In J

.

self-n of.

:

nt ofons.

nitive-

D. E.create

gy,

ony-

fect.

self-n

ic

ationage. In

elf-

andsic

time.

as aty for

elf-.

ial-ds.),al

ber-nitive

& J.in

ho-

totings.

theifica-al

rbal

s to

ND

certainty and self-affect.Journal of Personality and Social Psycholo58(6), 1062–1072.

Bem, D. J. (1972). Self-perception theory. In L. Berkowitz (Ed.),Advancein experimental social psychology(pp. 2–62). New York: Erlbaum.

Berkowitz, L., & Devine, P. G. (1989). Research traditions, analysissynthesis in social psychological theories: The case of dissonancory. Personality and Social Psychology Bulletin,15(4), 493–507.

Blaine, B., & Crocker, J. (1993). Self-esteem and self-serving biasreactions to positive and negative events: An integrative review. InBaumeister (Ed.),The puzzle of low self-regard(pp. 55–85). New YorkPlenum.

Blanton, H., Cooper, J., Skurnik, I., & Aronson, J. (1997). When bad thhappen to good feedback: Exacerbating the need for self-justificwith self-affirmations.Personality and Social Psychology Bulletin,23,684–692.

Brehm, J. (1956). Postdecision changes in the desirability of alternaJournal of Abnormal and Social Psychology,52, 384–389.

Brown, J. D. (1986). Evaluations of the self and others: Self-enhancebiases in social judgements.Social Cognition,4, 353–376.

Brown, J. B. (1998).The self.New York: McGraw-Hill.

Campbell, J. D. (1990). Self-esteem and clarity of the self-concept.Journalof Personality and Social Psychology,59, 538–549.

Campbell, J. D., & Lavallee, L. F. (1993). Who am I? The roleself-concept confusion in understanding the behavior of people witself-esteem. In R. F. Baumeister (Ed.),Self-esteem: The puzzle of lself-regard(pp. 3–20). New York: Plenum.

Cohen, A. R. (1962). An experiment on small rewards for discrecompliance and attitude change. In J. W. Brehm & A. R. Cohen (EExplorations in cognitive dissonance(pp. 78–81). New York: Wiley.

Cooper, J. (1992). Dissonance and the return of the self-concept.Psychological Inquiry, 3(4), 320–323.

Cooper, J. (1999). Unwanted consequences and the self: In searchmotivation for dissonance reduction. In E. Harmon-Jones & J. M(Eds.),Cognitive dissonance 40 years later: Revival with revisionscontroversies(pp. 149–174). Washington, DC: American PsychologAssociation.

Cooper, J., & Duncan, B. L. (1971). Cognitive dissonance as a functiself-esteem and logical inconsistency.Journal of Personality,39, 289–302.

Cooper, J., & Fazio, R. H. (1984). A new look at dissonance theory.Berkowitz (Ed.),Advances in experimental social psychology(pp. 229–262). Hillsdale, NJ: Erlbaum.

Cooper, J., & Worchel, S. (1970). Role of undesired consequencarousing dissonance.Journal of Personality and Social Psychology,16,199–206.

Duval, S., & Wicklund, R. A. (1972).A theory of objective self-awareneNew York: Academic Press.

Elliot, A. J., & Devine, P. G. (1994). On the motivational naturecognitive dissonance: Dissonance as psychological discomfort.Journalof Personality and Social Psychology,67(3), 382–394.

Fazio, R. H., & Cooper, J. (1983). Arousal in the dissonance processCacioppo & R. E. Petty (Eds.),Social psychophysiology(pp. 122–152)New York: Guilford.

Fazio, R. H., Zanna, M. P., & Cooper, J. (1977). Dissonance andperception: An integrative view of each theory’s proper domaiapplication.Journal of Experimental Social Psychology,13, 464–479

Festinger, L. (1957).A theory of cognitive dissonance.Stanford, CAStanford Univ. Press.

242 STONE A

Festinger, L., & Carlsmith, J. M. (1959). Cognitive consequences of forcedcompliance.Journal of Abnormal and Social Psychology,58, 203–210.

-

.

t

e

.

Galinsky, A. D., Stone, J., & Cooper, J. (2000). The reinstatemedissonance and psychological discomfort following failed affirmatiEuropean Journal of Social Psychology,30, 123–147.

Glass, D. (1964). Changes in liking as a means of reducing cogdiscrepancies between self-esteem and aggression.Journal of Personality, 32, 531–549.

Harmon-Jones, E., Brehm, J. W., Greenberg, J., Simon, L., & Nelson,(1996). Is the production of aversive consequences necessary tocognitive dissonance?Journal of Personality and Social Psycholo70(1), 5–16.

Harmon-Jones, E., & Mills, J. (1999).Cognitive dissonance: Progressa pivotal theory in social psychology.Washington, DC: American Pschological Association.

Higgins, E. T. (1987). Self-discrepancy: A theory relating self and afPsychological Review,94(3), 319–340.

Higgins, E. T. (1989). Self-discrepancy theory: What patterns ofbeliefs cause people to suffer? In L. Berkowitz (Ed.),Advances iexperimental social psychology(pp. 93–136). San Diego, CA: AcademPress.

Higgins, E. T. (1990). Personality, social psychology, and person-siturelations: Standards and knowledge activation as a common languL. A. Pervin (Eds.),Handbook of personality theory and research(pp.301–338). New York: Guilford.

Higgins, E. T. (1996). The “Self-Digest”: Self-knowledge serving sregulatory functions.Journal of Personality and Social Psychology,71,1062–1083.

Kernis, M. H., & Waschull, S. B. (1995). The interactive roles of stabilitylevel of self-esteem: Research and theory. In M. R. Zanna (Ed.),Advancein experimental social psychology(pp. 93–141). San Diego, CA: AcademPress.

Knox, R. E., & Inkster, J. A. (1968). Postdecision dissonance at postJournal of Personality and Social Psychology,8, 319–323.

Maracek, J., & Mettee, D. (1972). Avoidance of continued successfunction of self-esteem, level of esteem certainty, and responsibilisuccess.Journal of Personality and Social Psychology,22, 98–107.

Markus, H., & Kunda, Z. (1986). Stability and malleability of the sconcept.Journal of Personality and Social Psychology,51(4), 858–866

Markus, H., & Wurf, E. (1987). The dynamic self-concept: A socpsychological perspective. In M. R. Rosenzweig & L. W. Porter (EAnnual review of psychology(pp. 299–337). Palo Alto, CA.: AnnuReviews, Inc.

McGregor, I., Newby-Clarke, I. R., & Zanna, M. P. (1999). “Rememing” dissonance: Simultaneous accessibility of inconsistent cogelements moderates epistemic discomfort. In E. Harmon-JonesMills (Eds.), Cognitive dissonance: Progress on a pivotal theorysocial psychology(pp. 325–355). Washington, DC: American Psyclogical Association.

Moretti, M. M., & Higgins, E. T. (1990). Relating self-discrepancyself-esteem: The contribution of discrepancy beyond actual-self raJournal of Experimental Social Psychology,26, 108–123.

Nel, E., Helmreich, R., & Aronson, E. (1969). Opinion change inadvocate as a function of the persuasibility of his audience: A clartion of the meaning of dissonance.Journal of Personality and SociPsychology,12, 117–124.

Nisbett, R., & Wilson, T. (1977). Telling more than we can know: Vereports on mental processes.Psychological Review,84, 231–259.

Pelham, B. W., & Swann, W. B., Jr. (1989). From self-conceptionself-worth: On the sources and structure of global self-esteem.Journalof Personality and Social Psychology,57(4), 672–680.

COOPER

Rhodewalt, F. (1998). Self-presentation and the phenomenal self: The“carryover effect” revisited. In J. M. Darley & J. Cooper (Eds.),Attri-

.

n ongy,

gical

S initia

S tteny-

S le are

S teem

S thetal.

S ation

S and

S in cohe

(pp.

S ibute

S hen

S rplays &n-

S y and-

T cho-

T sses.

T g thel

as a

cialf

d.),:

ution

65).

DS I

bution and social interaction: The legacy of Edward E. Jones(pp.373–421). Washington, DC: American Psychological Association

Rhodewalt, F., & Agustsdottir, S. (1986). Effects of self-presentatiothe phenomenal self.Journal of Personality and Social Psycholo50(1), 47–55.

Rosenberg, M. (1979).Conceiving the self.New York: Basic Books.

Schacter, S., & Singer, J. E. (1962). Cognitive, social and physiolodeterminants of emotional state.Psychological Review,69, 379–399.

hrauger, J. S. (1975). Responses to evaluation as a function ofself-perceptions.Psychological Bulletin,82(4), 581–596.

imon, L., Greenberg, J., & Brehm, J. (1995). Trivialization: The forgomode of dissonance reduction.Journal of Personality and Social Pschology,68(2), 247–260.

nyder, M., & Tanke, E. D. (1976). Behavior and attitude: Some peopmore consistent than others.Journal of Personality,44, 510–517.

pencer, S. J., Josephs, R. A., & Steele, C. M. (1993). Low self-esThe uphill struggle for self-integrity. In R. Baumeister (Ed.),Self-esteem: The puzzle of low self-regard.New York: Plenum.

teele, C. M. (1988). The psychology of self-affirmation: Sustainingintegrity of the self. In L. Berkowitz (Ed.),Advances in experimensocial psychology(pp. 261–302). San Diego, CA: Academic Press

teele, C. M., & Liu, T. J. (1983). Dissonance processes as self-affirmJournal of Personality and Social Psychology,45, 5–19.

teele, C. M., Spencer, S. J., & Lynch, M. (1993). Dissonanceaffirmational resources: Resilience against self-image threats.Journal ofPersonality and Social Psychology,64(6), 885–896.

tone, J. (in press). Behavioral discrepancies and construal processesnitive dissonance. In G. Moskowitz (Ed.),Cognitive social psychology: TPrinceton symposium on the legacy and future of social cognition41–58).Hillsdale, NJ: Erlbaum.

tone, J. (1999). What exactly have I done? The role of self-attraccessibility in dissonance. In E. Harmon-Jones & J. Mills (E

SELF-STANDAR

ds.),Cognitive dissonance: Progress on a pivotal theory in social psychology(pp. 175–200). Washington, DC: American Psychological Association.

l

:

.

g-

tone, J., Wiegand, A. W., Cooper, J., & Aronson, E. (1997). Wexemplification fails: Hypocrisy and the motive for self-integrity.Jour-nal of Personality and Social Psychology,72(1), 54–65.

wann, W. B., Jr. (1990). To be adored or to be known?: The intebetween self-enhancement and self-verification. In E. T. HigginR. M. Sorrentino (Eds.),Handbook of motivation and cognition: Foudations of social behavior(pp. 408–448). New York: Guilford.

wann, W. B. J., & Schroeder, D. G. (1995). The search for beauttruth: A framework for understanding reactions to evaluations.Personality and Social Psychology Bulletin,21(12), 1307–1318.

edeschi, J. T. (1981).Impression management theory and social psylogical research.New York: Academic Press.

esser, A., & Cornell, D. P. (1991). On the confluence of self proceJournal of Experimental Social Psychology,27, 501–526.

hibodeau, R., & Aronson, E. (1992). Taking a closer look: Reassertinrole of the self-concept in dissonance theory.Personality and SociaPsychology Bulletin,18(5), 591–602.

Vallacher, R. R. (1992). Come back to the future.Psychological Inquiry,3,348–350.

Ward, W. D., & Sandvold, K. D. (1963). Performance expectancydeterminant of actual performance: A partial replication.Journal ofAbnormal and Social Psychology,67, 293–295.

Wegner, D. M., & Bargh, J. A. (1998). Control and automaticity in solife. In D. Gilbert, S. T. Fiske, & G. Lindzey (Eds.),The handbook osocial psychology(pp. 446–496). New York: McGraw–Hill.

Wicklund, R. A. (1975). Objective self-awareness. In L. Berkowitz (EAdvances in experimental social psychology(pp. 233–277). New YorkAcademic Press.

Zanna, M. P., & Cooper, J. (1974). Dissonance and the pill: An attribapproach to studying the arousal properties of dissonance.Journal ofPersonality and Social Psychology,29, 703–709.

Zimbardo, P. G., Weisenberg, M., Firestone, I. J., & Levy, B. (19

243N DISSONANCE

Communicator effectiveness in producing public conformity and privateattitude change.Journal of Personality,33, 233–255.