Interpersonal Evaluations Following Threats to Self: Role of Self-Esteem

Embed Size (px)

Citation preview

  • 7/28/2019 Interpersonal Evaluations Following Threats to Self: Role of Self-Esteem

    1/12

    PERSONALITY PROCESSES AND INDIVIDUAL DIFFERENCES

    Interpersonal Evaluations Following Threats toSelf: Role of Self-Esteem

    Todd F. Heatherton and Kathleen D. VohsDartmouth College

    In 2 studies, the authors used dyadic interactions to assess the influence of ego threat on likability as afunction of self-esteem. In both studies, 2 naive participants engaged in a structured conversation; in halfof the dyads, I participant received an ego threat prior to the interaction. In the 1st study, threatened highself-esteem participants were rated as less likable than were threatened low self-esteem participants. The2nd study confirmed that ego threats are associated with decreased liking for those with high self-esteemand with increased liking for those with low self-esteem. A mediational analysis demonstrated thatdecreased liking among high self-esteem participants w as due to being perceived as antagonistic. Study 2also indicated that the findings could not be explained by trait levels of narcissism. These patterns areinterpreted in terms of differential sensitivity to potential interpersonal rejection.

    The construct of self-esteem has rarely been examined in a trueinterpersonal context. The vast majority of research on the topichas examined the influence of self-esteem on motivational, behav-ioral, affective, and cognitive mechanisms within individuals withhigh or low self-esteem (e.g., Banaji & Prentice, 1994; Baumeister,1993, 1998; Brockner, 1983; Brown, 1993; Campbell, 1990;Crocker & Major, 1989; Greenberg et al., 1992; Greenwald &Banaji, 1995; Higgins, 1996; Pelham, 1995; Swann, 1996; Tesser,Millar, & Moore, 1988). Although several important theories ofself-esteem are interpersonal in nature (Cooley, 1902; Leary, Tam-bor, Terdal, & Downs, 1995; Mead, 1934; Shrauger & Schoene-man, 1979), people's assessments of those with high and lowself-esteem are nearly absent from the literature. These variousinterpersonal theories of self-esteem share a common view thatpeople's self-evaluations reflect, in part, beliefs about how they areperceived and valued by significant others. In the current research,we examine the validity of self-perceptions by examining inter-personal judgments of likability.

    The extent to which people are viewed as likable and as desir-able associates has a number of important mental health conse-quences. Those who feel ostracized or rejected experience negativereactions, including physical illness, emotional problems, and n eg-ative affective states (Bowlby, 1969; Downey & Feldman, 1996;Rutter, 1979; Williams, 1997). Furthermore, social support is

    Todd F. Heatherton and Kathleen D. Vohs, Department of Psycholog-ical and Brain Sciences, Dartmouth College.

    This research was supported in part by a grant from the RockefellerCenter for the Social Sciences at Dartmouth College. We thank PeterScheidt and Michael Stern for their assistance in this research and Jay Hullfor his comments and suggestions.

    Correspondence concerning this article should be addressed to Todd F.Heatherton, Department of Psychological and Brain Sciences, DartmouthCollege, 6207 Moore Hall, Hanover, New Hampshire 03755-3549. Elec-tronic mail may be sent to [email protected].

    known to be an important contributor to positive mental andphysical health (Cohen & Wills, 1985), and people who are dis-liked, antagonistic, or emotionally distressed are Less likely toreceive support and assistance from others (Bolger, Foster, Vino-kur, & Ng, 1996). Baumeister and Leary (1995) proposed thatpeople have a fundamental need to belong, such that they arestrongly motivated to seek out positive social interactions andavoid interactions that are conflicted or that contain negativeaffect. Their work outlines the considerable psychological tollexacted from an absence of positive and meaningful interpersonalrelationships.

    One key ingredient to the formation of strong so cial bonds is theextent to which people have likable personal characteristics, suchas being attractive (Berscheid & Hatfield, 1969) or possessingdesired personality traits (such as warmth and competence, seeLydon, Jamieson, & Zanna, 1988). Although self-reports of posi-tive personality traits are often correlated with self-esteem, rela-tively little is known about the extent to which perceivers endorsethe positive self-views of those with high self-esteem.

    Interpersonal Liking and Self-Esteem

    It has long been recognized that people like those who areattractive, competent, warm, and morally sound (Anderson, 1968;Lydon et al., 1988; Taylor, Peplau, & Sears, 1997). Of relevanceto the current research is the extent to which self-esteem is asso-ciated with personality traits of positive or negative valence. Interms of self-report, those with high self-esteem rate themselves asattractive (Diener, Wolsic,& Fujita, 1995), intelligent (Gabriel,Critelli, & Ee, 1994), socially outgoing (Briggs, Cheek, & Buss,1980;Schmutte &Ryff, 1997), socially skilled (Jones, Freemon, &Goswick,1981;Riggio, Throckmorton, & DePaola, 1990),unself-ish (Brockner, O'Malley, Hite, & Davies, 1987), emotionallystable (i.e., not neurotic, Scheier, Carver, & Bridges, 1994), andmorally sound (Dickstein & Hardy, 1979). Of course, highself-

    Joumal of Personality and Social Psychology, 2000, Vol. 78, No. 4, 725-736Copyright 2000 by the Ame rican Psycholo gical Assoc iation, Inc. O022-3514rtW$5.0O DO I: 10.1037/AM22-3514.78.4.725

    725

  • 7/28/2019 Interpersonal Evaluations Following Threats to Self: Role of Self-Esteem

    2/12

    726 HEATHERTON AND VOHS

    esteem people generally like themselvesand believe favorablethings about themselves,and thereforeit is not surprising that theyrate themselves highlyon most positive personality traits (Brown,1993). Unfortunately,the evidence failsto support a strong linkbetween self-esteemand objective appraisals (Swann, 1996).Forexample, although people with high self-esteem claimto be moreintelligent than those withlow self-esteem, objective data suggestthe association is weak at best (Gabrielet al., 1994). Similarly,self-esteem is more strongly relatedto self-perceived attractive-ness than objectively rated attractiveness (Dieneret al., 1995;Jovanovic, Lemer,& Lemer, 1989). Indeed, evenin terms ofjudgments of self-esteem thereis little correspondence betweenself-ratingsand observer ratings (Shrauger& Schoeneman, 1979).

    Our review of the literature uncovereda surprising lackofresearchon the interpersonal appraisalsof those with highand lowself-esteem.One of the few studies that directly assessed interper-sonal evaluationsof people as a function of self-esteem found thatthere were no overall differencesin terms of how much peopleliked those with highand low self-esteem (Brockner& Lloyd,

    1986). However, people with high self-esteem were more likelythan were people w ithlow self-esteemto believe that others likedthem. Even among children,the typical findingis that those withhigh self-esteem are more likely than those withlow self-esteemtobelieve that others like them (Bohrnstedt& Felson, 1983),al-though actual sociometric status often doesnot differ as a functionof self-esteem (Bishop& Inderbitzen, 1995). Thus,it does notappear that people generally like those with high self-esteem morethan they like those withlow self-esteem. Indeed, some evidencesuggests the possibility that people with high self-esteemmay bedisliked in certain situations, namely when they have experiencedthreats to self.

    Extremely positive self-appraisals have been linkedto poorinterpersonal outcomes, especially when such self-appraisalsarechallenged or discredited. Baumeister, Smart,and Boden (1996)examined the literature linking self-esteemto interpersonalvio-lence.In contrast to widely held assumptions thatlow self-esteemis associated with violent actions, they founda consistent patternin which thosewho thought highlyof themselvesbut encounteredsome sortof threat or challenge to their positive self-view weremore likely to engage in hostile and violent ways. BushmanandBaumeister (1998) have recently demonstrated that narcissists,wh o by definition view themselvesin positive terms, become moreaggressive following threat.

    Other evidence suggests that those with highly positiveself-views may exhibit poor interpersonal skills when challenged.Colvin, Block,and Funder (1995) examined individuals withap-

    parently inflated self-views,as indicatedby the difference betweenself and other ratings. Duringa dyadic interaction, they foundsuch individualsto be viewedas hostile and unlikable. This patternis most likely to occur when high self-esteem individuals feelpersonally challenged.For instance, Schlenker, Weigold,and Hal-lam (1990) exposed highand low self-esteem participantstocontexts in which they were motivatedto make a positive im pres-sion on a critical or supportive audience. They found that highself-esteem participants became egotistical when evaluative pres-sures were greatest. They concluded that "people with highself-esteem become more boastfulas the social stakes increase"(p.861). Similar findings were reportedby Schneider and Turkat(1975) who found that high self-esteem participantswho also

    expressed high needsfor social approval presented themselvesmuch more positively following negative feedback than followingpositive feedback. Perhaps ironically, high self-esteem participantswho receive negative feedback about their intellectual abilitiesclaim to have especially good social skills (Brown & Smart, 1991).However, the available evidence doesnot appear to support theseclaims.

    Self-Esteem and Response to Threa t

    Why might people with high self-esteem respondto ego threatsin a way that leadsto negative interpersonal evaluations?We startwith the assumption thatall people havea fundamental needtobelong that may well be rooted in our evolutionary history(Baumeister & Leary, 1995). However, thereare individual dif-ferencesin the extent to which people view themselvesas possibletargets of rejection. Mark Learyand his colleagues have recentlydeveloped a theory linking self-esteemto interpersonal rejection(Leary & Downs, 1995; Learyet al., 1995). This theory proposes

    that self-esteem functionsas a monitorof the likelihoodof socialexclusion. When people feel that theyare targets of social exclu-sion or rejection, they experiencea reduction in state self-esteemthat servesas a warning function.

    Accordingto the sociometer theory, high self-esteem individu-als generally expect othersto like and include them (Walster,1965),and therefore they maybe less concerned with interpersonalevaluation than people with m oderateor low self-esteem (Leary&Downs, 1995).As stated by Leary et al. (1995), "peoplewhoalready feel included, accepted,and socially integrated neednot beas concerned with fittingin as people who feel less so" (p. 520).Although people with high self-esteem apparentlydo experienceareduction in feelings of state self-esteem when excludedor ostra-cized, their self-esteemmay not drop to a level that suggests theyare in imminent dangerof being rejected. Situations that threatenself-esteembut do not explicitly indicate relationship devaluationma y not raise fearsof rejection for those with high self-esteem.

    By contrast, people withlow self-esteem possessa sense ofcontingent acceptance, believing that others will reject themif theyfail. For instance, Baldwinand Sinclair (1996) found that peoplewith low self-esteem linked personal failure withthe expectationthat they wouldbe rejected by others whereas high self-esteemparticipantsdid not show sucha contingency. Thus, although bothgroups experience negative self-feelings when they feel rejected,people withlow self-esteem expectto be rejected in failure situ-ations where people with high self-esteemdo not. Following anoninterpersonal ego threat, therefore, moderateor low self-

    esteem individuals m ay experiencea sufficient dropin self-esteemto activate the sociometer systemand heighten fearsof rejection.Although both highand low self-esteem individuals experienceareductionin state self-esteem followingego threats (Heatherton&Polivy, 1991), those withlow self-esteemmay be especially likelyto have their fearsof social exclusion activatedin ego-threateningsituations.

    Because the need to belong is a fundamental human motive,wepropose that people who feel theyare in imminent dangerof beingrejected engagein amiable, reparative,or prosocial behaviorstogain approval from othersand affirm social bonds.We thereforeexpect that low self-esteem individuals respondto ego threat bybeing friendly and courteous and by attending to their social

  • 7/28/2019 Interpersonal Evaluations Following Threats to Self: Role of Self-Esteem

    3/12

    SELF-ESTEEM AND EGO THREAT 727

    partners. By contrast, when those with high self-esteem receive anoninterpersonal ego threat, we propose that they do not experi-ence concerns about possible social rejection, and therefore, theydo not engage in efforts to affirm social relationships. Indeed, it ispossible that they use self-repair strategies such as self-aggrandizement and downward social comparisons to affirm their

    sense of self; if so, these strategies may have negative interper-sonal consequences (Leary, Bednarski, Hammon, & Duncan,1997).

    We therefore predicted that ego threats would have oppositeeffects on the interpersonal evaluations of those with high and lowself-esteem. Specifically, we predicted that low self-esteem indi-viduals would be evaluated more positively (more likable) andhigh self-esteem individuals will be evaluated more negatively(less likable) following ego threat. To test these predictions, wehad high and low self-esteem individuals engage in a structuredconversation that was designed to increase feelings of intimacy.Prior to the conversation, one participant in half of the dyadsreceived an ego threat. We used the conversation partner's ratingsto assess how ego threats are related to evaluations of those withhigh and low self-esteem.

    Study 1

    Method

    Participants

    Eighty-four male undergraduates were participantsin a study of inter-personal interactions. Participants received either extra course creditor apayment of $5 .

    ProcedurePrior to arriving at the experiment participants completeda standard

    measure of self-esteem (Fleming& Courtney, 1984,on the basis of Jam's& Field, 1959), whichwas used to divide participants into highor lowself-esteem groupson the basis of a median split(Mdn = 132, range =97-167).

    Participants signedup to participate in a study of interpersonal interac-tions, and the sign-up sheethad two spaces availablefor each time slot.Thus, two individuals arrivedat the lab at the same tim e. After arrivingatthe experiment,the experimenter asked the two participants whether theyknew each other. After assurance that they did not, one memberof the dyadwas randomly assignedto be the targetof evaluation (we hereinafter referto this person as the "target" and their partneras the "rater"). The targetwas then randomly assignedto be in either the control group or the

    ego-threat group. Thus, in each dyad there wasa target anda rater, andthetarget was either threatenedor not.

    Participants then completedthe Remote Associates Test (RAT;Med-nick, 1968). The RA T asks participantsto find one word that connects setsof three seemingly unrelated w ords. For example, the words "lick," "sprin-kle," and "mines" are linkedby the word "salt." Ego threat was createdinthree ways:by the description given about the implicationsof the test,thetest difficulty, and the performance feedback givento the participant.Ego-threatened participants were toldto complete the RAT in 4 min, thatthe RAT was a strong indicatorof academic achievement, and that the RATwas predictiveof future earning potential. Moreover,the RAT items wereextremely difficult,and pilot testing indicated that participants wouldbeunlikely to solve more than 3 or 4of the 12 problems. Non-ego-threatenedparticipants were told thatthe test was being piloted for a different

    experimenter and were asked to "giveit a try for a couple of minutes." TheRAT given to participants in the control condition was moderately easy.

    For ego-threatened targetsthe experimenter returned after4 min withared pen with whichto score their test. The experimenter looked surprisedas he scored the test and then excused himselfto check on the otherparticipant, leaving the answer key on the table. The answ er key containedthe correct answersbut also false statistics indicating the average scoreforDartmouth students(9.0 of 12.0 correct)and for college students nation-wide (7.2out of 12.0). Non-ego-threat targetsand raters did not receivefeedback on their performance (underthe guise that it was simplya pilotto see what participants thoughtof the test).

    Next both participants completedthe State Self-Esteem Scale (SSES;Heatherton & Polivy, 1991) and a 24-item mood scaleto assess theeffectivenessof the manipulation. The m ood scale consisted of four factorsdeterminedby earlier researchin our laboratory: positive affect, anxiety,dysphoria,and hostility.

    The participants were then brought togetherto engage in a structuredconversation. Becausethe feedback was given privately to those in theego-threat condition, theyhad no reason to believe that the other personknew how they had performed. We useda procedure devised by Sedikidesand his colleaguesto create intimacyin a short periodof time (Sedikides,

    Campbell, Reeder,& Elliot, 1998). This method involves participantsasking each othera series of questions that begin with ordinaryandimpersonal questions and continues throughto very personal and intimatequestions. Participants were instructedto spend a pre-established amountof time on each section: 2 min were given to discuss low intimacyquestions suchas "How old are you?" and "What is your major?";6 minwere givenfor moderately intimate items suchas "What is one embarrass-ing thing that has happenedto you this year?";and 10 min were givenforintimate topics such as feelings of loneliness and describing earlymemories.

    After the interaction, participants were separated again and givena set ofquestionnaires. The first question was "Based on what you know ab out thisperson so far, how much do you like him ?" and was rated on a 7-point scale(1 = not at all; 7 = very much).They then rated their partner on 22 bipolartraits (e.g.,unintelligentintelligent, recklesscautious, lazy-hardworking,

    shy-outgoing, calm-anxious). The bipolar traits were selected to representa broad arrayof personality dimensions(on the basis of the Big Five)aswell as to indicate traits knownto be related to liking (e.g., genuine-fake,honest-deceptive, cheerful-gloomy, modest-arrogant). Participants ratedwhich of the two opposing bipolar traits were more characteristicof theirpartner usinga 7-point scale (e.g.,1 = reckless; 7 = cautious). Finally,allparticipants were fully debriefed aboutthe purpose of the experiment,including assurancesto participants in the ego-threat conditions thatall ofthe RAT scoresand data were false.

    Results

    Manipulation Checks

    A 2 (high vs. low self-esteem) X 2 (ego threat vs. control)analysis of variance (ANOVA) was used to evaluate the effective-ness of the manipulation on the target's mood and state self-esteem. As may be seen in Table 1, we found that ego-threatenedtargets reported increased dysphoria, f ( l , 38) = 6.96, p < .02,decreased positive affect, F(l, 38) = 3.12, p < .09, and increasedhostility, F(l, 38) = 14.46, p < .0005. There were no significanteffects on self-reports of anxiety, F < 1. Self-esteem did notinteract with ego threat on any of the mood measures (/?s > .15)and the only main effect of self-esteem was that low self-esteemparticipants reported greater anxiety than did high self-esteemparticipants, F(l, 38) = 7.51, p < .01 (again, this did not interactwith condition). There were no significant decrements of state

  • 7/28/2019 Interpersonal Evaluations Following Threats to Self: Role of Self-Esteem

    4/12

    728 HEATHERTON AND VOHS

    Table 1Manipulation Checks for Study 1

    13.43.5

    18.37.3

    39.97.4

    11.33.4

    69.48.6

    17.64.8

    20.06.1

    34.311.4

    17.93.7

    65.310.2

    12.13.8

    14.15.4

    39.212.2

    11.24. 2

    78.07.3

    15.15.2

    14.63.2

    33.19.9

    14.14.8

    81.85.2

    Low self-esteem High self-esteemMood and self-esteem

    measures Control Ego threat Control Ego threat

    DysphoriaMSD

    AnxietyMSD

    Positive AffectM

    SD

    HostilityM

    SD

    State Self-EsteemMSD

    self-esteem as a function of ego threat, f(1 , 38) = 2.32,p < .15,although high self-esteem participants did score higher overallthan low self-esteem participants,F(l, 38) = 23.50,/? < .0001 (asis typically found, see Heatherton & Polivy, 1991).

    Primary Analysis

    We predicted that self-esteem would interact with ego threat inits effects on likability. A 2 (high vs. low self-esteem) X 2 (egothreat vs. control) ANOVA revealed the expected interaction,F(l,38) = 3.91,p = .055, as well as a marginal main effect ofself-esteem level,F(\, 38) = 3.19,p < .09.1 As may be seen in

    Figure 1, although there were no significant differences betweenhigh .and low self-esteem participants in the control grou p,((38) < 1, following ego threat those with low self-esteem wereliked more than those with high self-esteem,t(3S) = 2.66,p < .01.In terms of changes within self-esteem category, threatened high

    6.5-,

    Low SE High SECondition

    Figure 2. Ratings of antagonistic behavior by condition in Study 1. SE:

    self-esteem.

    self-esteem men were disliked somewhat more than were non-threatened high self-esteem m en, f(38) = 1.68,p .10, whereasthreatened low self-esteem men were liked somewhat more thannonthreatened low self-esteem me n, ?(38) = 1.35,p < .20.

    To clarify the basis of the likability ratings, we examined theraters' judgments of the target on the 22 bipolar trait dimensions.We used a principal-components analysis to reduce the number oftrait dimensions. This resulted in a four factor solution based oneigenvalues greater than 1. The rotated factor solution led us tolabel the factors as Composed (comprising the traits calm, honest,congenial, intelligent, reasonable, refined, and unassuming), De-pressive (comprising lethargic, gloomy, lazy, shy, timid, lowself-esteem, and yielding), Inhibited (comprising restrained, cautious,and practical), and Antagonistic (comprising arrogant, fake, unco-operative, rude, and unfriendly). The latter factor was of particularinterest since these traits are associated with the potential negativeaspects of high self-esteem (Leary et al., 1997).

    A 2 X 2 ANOVA on antagonism scores obtained a marginallysignificant interaction between self-esteem and ego threat,F(l,38) = 2.74, p < .10, and a main effect of ego threat, F(l,38) = 6.23,p < .02. As may be seen in Figure 2, although thereare no differences between high and low self-esteem targets in thecontrol condition, r(38) < 1, as predicted, high self-esteem targetswere rated higher on the Antagonism factor than low self-esteem

    Low SE High SECondition

    Figure 1. Ratings of likability by condition in Study 1. SE = self-esteem.

    1 An important preliminary analysis is to demonstrate that participants'ratings were independent of each other. That is, we had to ensure thatparticipants' ratings of liking were unrelated to each other in order todemonstrate independence of the data (and therefore not violate one of thecentral assumptions of ANOVA, see Brockner & Lloyd, 1986). That is,self-perceptions of whether the other person liked them should not berelated to their liking for the other person. Indeed, there was no relationbetween participants' m etaperception of their partner's liking for them andthe extent to which they liked their partner,r(4G) = - . 1 7 , p > .25.Moreover, the extent to which targets liked themselves should not correlatewith their liking for their partners , and it did not, r(40) = .03 ,p > .25.Thus, we were justified in using ANOVA to analyze these data.

  • 7/28/2019 Interpersonal Evaluations Following Threats to Self: Role of Self-Esteem

    5/12

    SELF-ESTEEM AND EGO THREAT 729

    targets following ego threat, r(38) = 2.24,p < .05. In addition,high self-esteem participants were rated as being more antagonisticfollowing ego threat than they were in the control condition,r(38) = 2.93,p < .01 . The re were no significant m ain effects orinteractions on any of the other factors.

    Mediational Analysis

    To examine the prediction that changes in antagonistic behav-iors were responsible for the decreased liking of high self-esteemtargets, we conducted a mediational analysis (Baron & Kenny,1986).We tested the effect of the proposed mediator, antagonism,using a regression model that included self-esteem scores, ego-threat condition (a dichotomous variable, ego threat vs. control),their multiplicative interaction, and antagonism ratings to predictlikability. All continuous variables were centered by subtractingtheir means prior to being entered (Aiken & West, 1991). Thisanalysis revealed that the addition of the mediator, antagonism,renders the interaction between self-esteem and ego-threat condi-tion nonsignificant, r(37) = .54,its, b = .01. The main effectsof self-esteem and ego-threat condition were also not significant inthis model, r(37) = .27 and r(37) = .76, both ps =ns . Thus, itappears that antagonism ratings mediated the relation betweenself-esteem, threat, and liking.

    Discussion

    The results of Study 1 provide initial support for our predictionthat ego threat would lead to differential liking as a function ofself-esteem. In the control condition, there were no differences inliking between high and low self-esteem targets, replicating Brock-ner and Lloyd's (1986) earlier findings. However, our ego threatwas associated with decreased liking for those with highself-

    esteem. This supports our hypothesis that high self-esteem indi-viduals react negatively to ego threat and behave in ways that areviewed as unlikable by others. In contrast, as we predicted, lowself-esteem targets were viewed as somewhat more likable follow-ing ego threat. This supports the hypothesis that those with lowself-esteem become concerned with the possibility of rejectionwhen ego threatened.

    One limitation to our first study is that although the interactionbetween self-esteem and ego threat was nearly significant, we didnot obtain clear evidence that high self-esteem participants wereless likable and low self-esteem participants more likable follow-ing threat. It is possible that our ego threat manipulation was notsufficiently powerful to have its intended effect. For instanc e, therewere only modest changes in scores on the SSES and therefore thetargets might not have been adequately threatened by our manip-ulation. Therefore, we decided to boost the threat in Study 2 so thatwe could more confidently examine its effects. We again predictedthat ego threat would lead to decreased liking for those with highself-esteem and increased liking for those with low self-esteem. Inaddition, we examined possible mediation of likability ratings byratings on the Antagonism factor, as was seen in Study 1.

    Many of the behaviors displayed by threatened high self-esteemparticipants in Study 1 suggest the possibility that they wouldscore high in measures of narcissism. Indeed, there is modestoverlap in measures of self-esteem and narcissism (Emmons,1984; Raskin, Novacek, & Hogan, 1991a, 1991b; Raskin & Terry,

    1988; Rhodewalt &Morf, 1995), and studies of narcissism haveobserved correlations between high scores in narcissism and in-terpersonal hostility (Raskin et al., 1991b; Rhodewalt &Morf,1995).Bushman and Baumeister (1998) have recently shown thatnarcissists who received an ego threat responded with increasedinterpersonal aggression, as measured by their willingness to ad-

    minister a blast of noise to an individual who gave them negativefeedback. Rhodewalt and Morf (1998) found that narcissists re-sponded to failure feedback with increased anger, which certainlysuggests that they might be viewed as antagonistic. Thus, itseemed important to establish whether our findings in the firststudy were due to the confounding of narcissism and self-esteem.

    Study 2

    Method

    Participants

    One hundred twenty-six male undergraduates were participants in astudy of interpersonal interactions. Participants received extra course creditin exchange for their participation.

    Procedure

    Prior to arriving at the experiment participants completed a standardmeasure of self-esteem (Fleming & Courtney, 1984, on the basis of Janis& Field, 1959), which was used to divide participants into high or lowself-esteem based on a median split(Mdn = 127;SD = 20.6). Participantsalso completed the Narcissistic Personality Inventory (NPI), a measure oftrait narcissism (Raskin& Terry, 1988). Mean score on the NPI was 18.5(Mdn = 20; SD = 5.9).

    The method used in Study 2 was a m odified version of Study1 designedto increase the intensity of the ego threat.2 Participants signed up for a

    study of interpersonal interactions and the sign-up sheet had two spacesavailable for each time slot. Thus, two individuals arrived at the lab at thesame time. After being assured that the two participants were unac-quainted, they were then randomly assigned to be either the target ofevaluation or the rater. The target was then randomly assigned to be in thecontrol group or the ego-threat group. It was at this point that participantscompleted the RAT , which again served as our manipulation of ego threat.

    Control condition. These participants completed the easy version ofthe RAT and w ere told to "give it a try for a couple of minutes." They werenot given any information about the implications of the test, their perfor-mance, or how others typically perform. Prior to the dyadic interaction,they completed the SSES and a 24-item mood scale to assess the effec-tiveness of the manipulation.

    Ego-threat condition. These participants completed the difficult ver-sion of the RAT. They were told that the RAT reliably predicts academic

    achievement and even future earning potential. Participants were told thatthey had a 4-min time limit, accentuated by the starting of a stopwatch.After four m inutes, the experimenter scored the R AT and looked surprisedat the participants' answers (mode = 1 correct out of 12). We bolstered the

    2 We included a cognitive load condition in order to test whether thebehavior of high self-esteem targets was due to cognitive load. W e used astandard manipulation for cognitive load of having subjects memorize a9-digit number (Gilbert & Hixon, 1991). On further reflection, and someprodding from the reviewers, it became clear that this was not an appro-priate task to test the hypothesis. Moreover, the putative load condition didnot differ in any fashion from the control condition. Therefore, we excludethat condition from the analysis and discussion.

  • 7/28/2019 Interpersonal Evaluations Following Threats to Self: Role of Self-Esteem

    6/12

    730 HEATHERTON AND VOHS

    ego threat by having the experimenter leave the room to allow participantsto "look at the correct answers." Like many difficult tasks, once a personsees the solution and is able to use top-down processing, the connectionsbetween the word sets are obvious. This showed participants that the RATitems were solvable and also provided false average RAT scores forcollege students at Dartmouth and nationwide (means were listed as 9.2and 7.1, respectively). Next participants completed the SSES and moodscales.

    The target and the rater were then brought together to engage in astructured conversation (Sedikides, Campbell, Reeder, & Elliot, 1998).After the interaction, participants were separated again and given a set ofquestionnaires. The first question was "B ased on what you know about thisperson so far, how much do you like him?" and was rated on a 100-pointscale (1 = not at all; 100 = very much). They then rated their partner onthe same 22 personality traits that had been used in Study 1. Finally, allparticipants were fully debriefed about the purpose of the experiment,including assurances to participants in the eg o threat conditions that all ofthe RAT scores and data were false.

    Results

    Manipulation ChecksW e conducted a series of 2 (high vs. low self-esteem) X 2 (ego

    threat vs. control) ANOVAs on the mood and state self-esteemmeasures. As may be seen in Table 2, ego-threatened participantsreported less positive affect, F(l, 38) = 7.80,p < .01; greaterhostility,F(l, 38) = 11.46,p < .001;and lower state self-esteem,F(l, 38) = 4.52,p = .04. There were no significant effects onmood ratings of anxiety or dysphoria (ps >. 10). Trait self-esteemlevel did not interact with ego threat on any of these measures.These results generally replicate the findings in the first study anddemonstrate that our ego threat had the intended effect.

    Primary Analysis

    Our first prediction was that ego threat would lead to decreasedliking for high self-esteem participants and increased liking forlow self-esteem participants. To test this we conducted a 2 (egothreat vs. control) X 2 (high vs. low self-esteem) ANOVA on theraters' liking for the target. This analysis produced a significant

    Table 2Manipulation Checks for Study 2

    Mood and self-esteemmeasures

    DysphoriaMSD

    AnxietyMSD

    Positive AffectMSD

    HostilityMSD

    State Self-EsteemMSD

    Low self-esteem

    Control

    11.16.2

    29.09.2

    43.811.0

    8.83.3

    71.310.5

    Ego threat

    13.83.5

    29.75.2

    30.89.4

    13.53.8

    61.46.2

    High self-esteem

    Control

    9.23.9

    23.36.6

    45.311.0

    8.24.1

    80.88.7

    Ego threat

    8.83.5

    20.98.7

    40.18.6

    12.14.7

    78.312.3

    LowSE HighSECondition

    Figure 3. Rating s of likability by condition in Study 2. SE = self-esteem.

    interaction, F(1, 38) = 7.99,p < .008. Decomposing the interac-tion (see Figure 3), we found that high self-esteem participantswho received an ego threat were liked considerably less than lowself-esteem participants who were threatened, r(38) = 5.26,p .30. The difference between thesecorrelations is significant (z 1.97, p < .05). There is also asignificant difference betweenthe correlationsof ego threat andlikability among highand low self-esteem targets,r(21) = -.36,p = .10, and r(21) = .49,/? < .03, respectively,z = 2.39,/> < .05.Stated differently, beingin the ego-threat conditionwas positivelyassociated with likability ratingsfor low self-esteem targetsandnegatively associated with likability ratingsfor high self-esteemtargets. These analyses illustratethe differences in perceptionsofhigh and low self-esteem targetsas a function of ego threat.Forhigh self-esteem targets, beingin the ego-threat conditionwasrelated to higher ratingson the antagonism factorand decreasedlikability, whereas beingin the ego-threat conditionfor low self-esteem targetswas unrelated to ratings on the antagonism factorbut significantly relatedto increased likability. Thus, antagonismappears to mediate the relationship betweenthe Self-Esteem XEgo-Threat interactionand likability for high self-esteem targets

    bu t not for low self-esteem targets.

    Discussion

    The resultsof Study 2 provide strong supportfor our predictionthat ego threat changesthe way people with highand low self-esteem are evaluated. Although likingdid not vary as a functionofself-esteem in the control condition, after being threatened, highself-esteem individuals were ratedas more antagonisticand unlik-able whereas low self-esteem individuals were ratedas morelikable. Although both highand low self-esteem participantsex-perienced diminished positive affect, decreased state self-esteem,and increased hostility when threatened, their subsequent behaviorled them to be evaluated quite differently. People with highself-

    esteem were viewedas antagonistic (i.e., rude, uncooperative,fake, unfriendly, and arrogant) when threatened. Taking intoac-count the mediating effectof antagonism reducedthe effect of eg othreat on liking as a function of self-esteem. In contrast to theevaluationsof high self-esteem targets, those withlow self-esteemwere liked more when threatened. Althoughthe general patternssuggested that they becam e nicer,the trait ratings effects werenotsignificant, exceptfor them being ratedas more inhibited.

    High self-esteem participantsin both studies were viewedasantagonistic, whichled them to be perceived as unfriendly andunlikable.The behavioral responseof those with high self-esteemmay reflect the motivational orientationof self-enhancement ratherthan self-protection (Baumeister, Tice,& Hutton, 1989). Firmly

    entrenched within high self-esteemis the belief thatone is valuedand respectedby others, and therefore those with high self-esteemmay generally assume that theyare accepted and included. Peoplewith high self-esteemdo not expect to be rejected simply becausethey failedat a task (Baldwin& Sinclair, 1996),and therefore theirsociometersdo not become triggeredby the types of ego threatused in this research. It is possible that people with highself-esteem expend their effortsat repairing or enhancing theirownself-views following threat, with relatively little concernforwhether other people like them.For instance, people with highself-esteem reactto ego threat by focusing on their strengthsanddownplaying their weaknesses (Dodgson& Wood, 1998).Al-though having high self-esteem helps peopleto compensate forand neutralizethe negative affect inherentin ego threat, the be-havioral correlatesof this compensationmay have negative inter-personal consequences.

    We hasten to add that we are not suggesting that people withhigh self-esteemare motivated to act antagonisticallyin order toenhance their self-esteem. Rather,we are suggesting that followinga noninterpersonalego threat, those with high self-esteem placeapriority on self-reparationat the expense of impression manage-ment. From attachment theory, individuals withan avoidant at-tachment stylewho also possess high self-esteem (Collins&Read, 1990)have been foundto react to threatening situationsby decreasing relianceon others and reasserting their autonomy(e.g., Mikulincer, 1998).It may be, then, that there are linksbetween the behaviors associated with avoidant attachment styleand the behaviorsof threatened high self-esteem individualsin thepresent study.We have no direct evidence regardingthe strategiesthat high self-esteem participants usedto cope withthe ego threat.Whatever theydid, they were evaluated negatively becauseof it.

    Those with low self-esteem link failure with expectationsofinterpersonal rejection (Baldwin& Sinclair, 1996),and thereforetheir sociometersare activated in situations where they fail.Theactivationof their sociometer warns them that theyare at risk forsocial exclusion,and because the need to belong is a fundamentalhuman m otive, theymay try to forestall actual rejection. Accord-ingly, their behavior following threatis viewed as somewhat morefriendly, but also as cautious and restrained. We contend thatpeople withlow self-esteemare viewed as somewhat more inhib-ited following failure because theytry not to do or say anythingthat wouldbe perceived negatively.

  • 7/28/2019 Interpersonal Evaluations Following Threats to Self: Role of Self-Esteem

    9/12

    SELF-ESTEEMAND EGO THREAT 733

    General Discussion

    Relation to Other Variables

    Because it was possible that reduced likingof high self-esteemthreatened targets w as dueto narcissism rather thanto self-esteem,

    we includeda standard measureof narcissism to see whether itpredicted or mediated the interpersonal evaluations.We did notfind that narcissismhad an impact on likability after threat:Theeffect of self-esteem and threat remained significant after control-ling for ego threat; therewas no association between trait levelsofnarcissism and liking among high self-esteem threatened partici-pants; and an analysis of liking as a function of narcissism andthreat did not lead to any significant effects. Although previousreports suggest that narcissists reactto threat by becoming hostileand aggressive (Bushman& Baumeister, 1998; Rhodewalt &Morf, 1998), the current studiesdo not show evidence that theydisplay their emotionsto their dyad partners. Thereis some evi-dence that narcissistic self-aggrandizementis hidden from observ-ers (Robbins& Dupont, 1992),and it has also been reported thatnarcissists moderate their negative responsein public (Morf &Rhodewalt, 1993). Thus,it is possible that although narcissistsreact angrily to threat, and although theyare willing to anony-mously administer more noisea m easureof aggressionto thosewho threaten them (Bushman& Baumeister, 1998), theymaymask these personal reactionsand behaviors in front of others.Paulhus (1998)has shown that people's first reactionto peoplehigh in trait self-enhancementa construct similarto narcis-sismis positive,and that it is only with time that they dislikethem. Hence,it seems plausible that narcissists manageto hidetheir true feelings overthe short-term.

    One possible criticismof our findings is that the patterns ob-tained do not represent genuine high self-esteembut a defensive

    self-esteem orientation.As pointed out by Leary et al. (1997),it iscommonly believed that egotistical behaviors arise froma fragileor insecure senseof self. Baumeisteret al. (1996) proposed thatinterpersonal violence occurs most frequently among those withanunstableor fluctuating senseof self-esteem(see also Kernis' workon unstable self-esteem, &ernis,1993; Kernis, Granneman,&Barclay, 1989). Thus,it is possible that only fluctuatingor defen-sive high self-esteem leadsto reduced interpersonal liking whenchallenged. Nonetheless,a sufficient numberof such individualsexist to produce an overall significant effectof self-esteem onlikability. Thatis, our findings suggest that defensivenessin theface of ego threat is a general pattern amongmen with highself-esteem.

    The core ideaof defensive high self-esteemis that some aspectof low self-esteem underliesthe self-reportsof high self-esteemforsome individuals (i.e., theyare actually insecurebut report likingthemselves for reasons of social desirabilityor self-deception).From this perspective,it is low self-esteem and insecurity thatleads defensive high self-esteem individualsto engage in antago-nistic behaviors. However, note that threatened high self-esteemparticipants did not behave in a similar fashion to those whoreported low self-esteem; thosewho scored low in self-esteemdidnot become obnoxious when challenged,and in fact, they wereviewed as somewhat nicer. Hence,the hidden low self-esteemview doesnot fit the data from these studies.Of course, it is alsoplausible that those with defensive self-esteemare qualitatively

    different from those withlow self-esteem, but it is difficult toknow how to identify such individuals.

    The desire to explain possible negative aspectsof high self-esteem by inferringthe existenceof low self-esteemor some othermaladaptive formof high self-esteemmay be due to the deeplyheld societal beliefs that self-esteemis not only good, but is

    something to be treasured and nurtured in our classroomsandhomes. Although many high self-esteem individualsmay possessthe positive characteristics thatwe expect from them,it seemsplausible that many peoplewho lack these positive characteristicsalso like themselvesa great deal. Thatis, there is a tendency tojudge people's self-esteemby whether or not their actionsfit the.prototypesof high and low self-esteem.

    The notion that thereare observable differences between thosewith genuine and defensive self-esteemhas been an issue inself-esteem researchfor some time (Schneider& Turkat, 1975).For instance, self-esteemis positively correlated with social desir-ability (Raskin, Novacek,& Hogan, 1991a),and some researchershave argued that researchers shouldbe careful in whether theyare

    examining genuineor defensive self-esteem (Smalley& Stake,1996). The development of measures of implicit self-esteem(Greenwald & Banaji, 1995) may allow for differentiation ofhealthy and defensive self-esteem,but the possibility exists thatwhat is healthyfor the self is not particularly beneficialfor smoothand pleasant social interaction. Baumeister (1998)has noted thatmost of the benefitsof high self-esteemgo to individuals, whereasmost of the costs and the consequencesgo to those around them.

    We examinedand rejected the possibility thatour findings weredu e to narcissism rather than self-esteem.We believe, however,that one aspect of high self-esteemis a tendency towards occa-sional narcissism. From this perspective, beliefsby people thatthey are especially capable, competent,and better than averagemay form positive illusions (Taylor& Brown, 1988),and theseillusionsmay allow themto maintain their senseof self-esteeminthe face of hardships, rejections, obstacles,and failures. AsBaumeisteret al. (1996) point out, "Obviously,if high self-esteemis defined in a way that stipulates thatit can only produce positive,desirable characteristics, thenit cannot leadto violence or aggres-sion, but this is circular" (p. 27). It may be that defensiveandantagonistic responsesto ego threats characterize thosewho arerelatively highin both self-esteemand narcissism,but a sufficientnumber of those with high self-esteem reacted antagonisticallywhen challengedto carry the effect for the entire group.

    Future Research

    Our research demonstrates that people with highand low self-esteem are evaluated quite differently followingego threat. Al-though popular culturemay hold the belief that positiveself-conceptionsare associated with positive interpersonal evaluations,the interpersonal correlatesof high and low self-esteemare poorlyunderstood.Our results suggesta numberof possible avenuesforfuture research.For instance, it seems importantto examine thebasis of gender differencesin liking as a function of self-esteem.It is possible that the interpersonal consequencesof self-esteemdiffer for men and women, perhapsas a function of differentialbehaviors displayed by highand low self-esteem m enand women.An important limitationof our study is that we included onlymen

  • 7/28/2019 Interpersonal Evaluations Following Threats to Self: Role of Self-Esteem

    10/12

    734 HEATHERTON AND VOHS

    as participants. Additional research is necessary to examinewhether these patterns obtain equally for women.

    It is also possible that judgments of those with low and highself-esteem evolve or change over time, and understanding suchpatterns is important for understanding long-term relationships andsocial support. Joiner's research (Joiner, Alfano, & Metalsky,1992; Joiner & Metalsky, 1995) has provided clear evidence thatlow self-esteem men engage in negative interpersonal behaviors(e.g., reassurance seeking) that lead to rejection. Conversely,Paulhus (1998) has recently shown that although self-enhancersare initially liked, they become more unlikable over time. Thetemporal dynamics of self-esteem and relationship formation re-quire further empirical consideration, especially when we considerthe deleterious consequences associated with social rejection andostracism (Baumeister & Leary, 1995; Williams, 1997).

    Summary

    Our studies provide compelling evidence that the interpersonal

    consequences of ego threat differ depending on level of self-esteem. These data challenge the belief, held by many, that peopleespecially like those who have high self-esteem. We found that, ina neutral situation, there are no differences in how people evaluatethose with high and low self-esteem. In this regard, our findingsreplicate the earlier research reported by Brockner and Lloyd(1986). Hence, in most circumstances, the positivity of self-evaluations are not reflected in the evaluations made by others.These initial findings are not very surprising. It is not uncommonto meet people who have low self-esteem in spite of great successand the loving attention of friends and family (Heatherton & Vohs,in press). Indeed, that people dislike themselves in spite of objec-tive evidence is the central paradox of low self-esteem (Baumeis-ter, 1993). Conversely, sometimes we can be puzzled by how

    much people like themselves in spite of their numerous failures,obvious shortcomings, and even in the face of blatant rejection andostracism. Our findings provide an important reminder that thecentral premise of interpersonal theories of self-esteem is that it isthe personal belief that one is valued and respected that predictshaving high self-esteem. Whether others actually share these pos-itive beliefs is of secondary importance.

    References

    Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing andinterpreting interactions. Newbury Park, CA: Sage.

    Anderson, N. H. (1968). Likableness ratingsof 555 personality-trait w ords.Journal of Personality and Social Psychology, 9, 272-279.

    Baldwin, M. W., & Sinclair, L. (1996). Self-esteem and " i f . . . t h e n "contingenciesof interpersonal acceptance.Journal of Personality andSocial Psychology, 71, 1130-1141.

    Banaji, M. R., & Prentice,D. (1994).The self in social contexts.AnnualReview of Psychology, 45, 297-332.

    Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variabledistinction in social psychological research: Conceptual, strategic,andstatistical considerations.Journal of Personality and Social Psychol-ogy, 51, 1173-1182.

    Baumeister, R.F. (1993). Self-esteem: The puzzle of lowself-regard. NewYork: Plenum Press.

    Baumeister, R. F. (1998). The self. In D. T. Gilbert, S. T. Fiske, & G.Lindzey (Eds.),The handbook of social psychology (4th ed., pp . 680740). Boston: McGraw-Hill.

    Baumeister,R. F., & Leary, M.R. (1995). The needto belong: Desireforinterpersonal attachmentsas a fundamental human motivation.Psycho-logical Bulletin, 117, 497-529.

    Baumeister, R. F., Smart, L., & Boden, J. M. (1996). Relation of threatenedegotism to violence and aggression:The dark side of high self-esteem.Psychological Review, 103, 5-33.

    Baumeister, R.R, Tice, D. M.,& Hutton, D. G. (1989). Self-presentationalmotivationsand personality differencesin self-esteem. Journal of Per-sonality, 57, 547-579.

    Berscheid, E., & Hatfield, E. (1969). Interpersonal attraction. Reading,MA: Addison-Wesley.

    Bishop, J. A., & Inderbitzen, H. M. (1995). Peer acceptance and friendship:An investigationof their relationto self-esteem. Journal of Early Ado-lescence, 15, 476-489 .

    Bohm stedt, G. W., & Felson, R. B . (1983). Explaining the relations amongchildren's actualand perceived performancesand self-esteem: A com-parison of several causal models.Journal of Personality and SocialPsychology, 45, 43-56.

    Bolger, N., Foster,M., Vinokur, A. D., & Ng, R. (1996). Close relation-ships and adjustmentsto a life crisis: The caseof breast cancer.Journalof Personality and Social Psychology, 70, 283-294.

    Bowlby, J. (1969). Disruptionof affectional bonds and its effects onbehavior. Canada's mental health supplement, 59, 12.

    Briggs, S. R., Cheek, J. M., & Buss, A. H. (1980). An analysis of theSelf-Monitoring Scale.Journal of Personality and Social Psychol-ogy, 38, 679-686.

    Brockner, J. (1983). Low self-esteem and behavioral plasticity: Someimplicationsfor personality and social psychology.In L. Wheeler (Ed.),Review of personality and social psychology (Vol. 4, pp. 237-271).

    * Beverly Hills, CA : Sage.Brockner, J., & Lloyd, K. (1986). Self-esteemand likability: Separating

    fact from fantasy.Journal of Research in Personality, 20, 496-508.Brockner, J., O'Malley, M . N., Hite,T., & Davies, D.K. (1987). Reward

    allocation and self-esteem: The rolesof modeling and equity restoration.Journal of Personality and Social Psychology, 52, 844-850.

    Brown, J. D . (1993). Motivational conflict and theself: The double-bind oflow self-esteem.In R. Baumeister (Ed.),Self-esteem: The puzzle of lowself-regard (pp. 117-130).New York: Plenum Press.

    Brown,J. D., & Smart, S. A. (1991). The self and social conduct: Linkingself-representations to prosocial behavior.Journal of Personality andSocial Psychology, 60, 368-375.

    Bushman,B. J., & Baumeister,R. F. (1998). Threatened egotism, narcis-sism, self-esteem, and direct and displaced aggression: Does self-loveorself-hate leadto violence? Journal of Personality and Social Psychol-ogy, 75, 219-229.

    Camp bell, J. D. (1990). Self-esteem and clarityof the self-concept.Journalof Personality and Social Psychology, 59, 538-549.

    Cohen, S., & W ills,T. A. (1985). Stress, social support,and the bufferinghypothesis.Psychological B ulletin, 98, 310-357.

    Collins, N. L., & Read, S. J. (1990). Adult attachment, working models,and relationship qualityin dating couples.Journal of Personality andSocial Psychology, 58, 644-663 .

    Colvin, C. R., Block, J., & Funder, D. C. (1995). Overly positiveself-evaluations and personality: Negative implicationsfor mental health.Journal of Personality and Social Psychology, 68, 1152-1162.

    Cooley, C. H. (1902). Human nature and the social order. New York:Scribner.

    Crocker, J., & Major, B. (1989). Social stigmaand self-esteem:Theself-protective propertiesof stigma. Psychological Review, 96, 608-630.

    Dickstein, E. B., & Hardy, B. W. (1979). Self-esteem, autonomy,andmoral behavior in college men and women.Journal of Genetic Psychol-ogy, 134, 51-55.

    Diener, E., Wolsic, B., & Fujita, F. (1995). Physical attractivenessand

  • 7/28/2019 Interpersonal Evaluations Following Threats to Self: Role of Self-Esteem

    11/12

    SELF-ESTEEM AND EGO THREAT 735

    subjective well-being.Journal of Personality and Social Psychology, 69,120-129.

    Dodgson, P. G., & Wood, J. V. (1998). Self-esteem and the cognitiveaccessibility of strengths and weaknesses after failure.Journal of Per-sonality and Social Psychology, 75, 178-197.

    Dow ney, G., & Feldman, S. I. (199 6). Implications of rejection sensitivity

    for intimate relationships.Journal of Personality and Social Psychol-ogy, 70, 1327-1343.Emmons, R. A. (1984). Factor analysis and construct validity of the

    Narcissistic Personality Inventory.Journal of Personality Assess-ment, 48, 291-300.

    Fleming, J. S., & Courtney, B. E. (1984). The dimensionality ofself-esteem: Hierarchical facet model for revised measurement scales.Jour-nal of Personality and Social Psychology, 46, 4 0 4 - 4 2 1 .

    Gabriel, M. T., Critelli, J. W., & Ee, J. S. (1994). Narcissistic illusions inself-evaluations of intelligence and attractiveness.Journal of Personal-ity, 62. 143-155.

    Gilbert, D. T., & Hixon, J. G. (1991). The trouble of thinking: Activationand application of stereotypic beliefs.Journal of Personality and SocialPsychology, 60, 509-517.

    Greenberg, J., Solomon, S., Pyszczynski, T., Rosenblatt, A., Burling, J.,

    Lyon, D., Simon, L., & Pinel, E. (1992). Why do people needself-esteem? Converging evidence that self-esteem serves an anxiety-buffering function.Journal of Personality and Social Psychology, 63,913-922.

    Greenwald, A. G., & Banaji, M. R. (1995). Implicit social cognition:Attitudes, self-esteem, and stereotypes.Psychological Review, 102,4 - 2 7 .

    Heatherton, T. F., & Polivy, J. (1991). Development and validation of ascale for measuring state self-esteem.Journal of Personality and SocialPsychology, 60, 895-910.

    Heatherton, T. F., & Vohs, K. D. (in press). Self-esteem. In E. Borgatta(Ed.), Encyclopedia of sociology. New York: Macmillan.

    Higgins, E. T. (1996). The "self-digest": Self-knowledge servingself-regulatory functions.Journal of Personality and Social Psychology, 71,1062-1083.

    Janis, I. L., & Field, P. (195 9). Sex differences and personality factorsrelated to persuasibility. In C. H ovland & I. Janis (Eds.),Personality andpersuasibility (pp. 55-68; 300-302). New Haven, CT: Yale UniversityPress.

    Joiner, T. E., Alfano, M. S., & Metalsky, G. I. (1992). When depressionbreeds contempt: Reassurance seeking, self-esteem, and rejection ofdepressed college students by their roommates.Journal of AbnormalPsychology, 101, 165-173.

    Joiner, T. E., & Metalsky, G. I. (1995). A prospective test of an integrativeinterpersonal theory of depression: A naturalistic study of college room-mates. Journal of Personality and Social Psychology, 69, 778-788.

    Jones,W. H., Freemon, J. E.,& Goswick, R. A. (1981). The persistence ofloneliness: Self and other determinants.Journal of Personality, 49,2 7 -4 8 .

    Jovanovic, J., Lerner, R. M., & Lerner, J. V. (1989). Objective andsubjective attractiveness and early adolescent adjustment.Journal ofAdolescence, 12,225-229.

    Kernis, M. H. (1993). The roles of stability and level of self-esteem inpsychological functioning. In R. Baumeister (Ed.),Self-esteem: Thepuzzle of low self-regard (pp. 167-182). New York: Plenum.

    Kemis, M. H., Granneman, B. D., & Barclay, L. C. (1989). Stability andlevel of self-esteem as predictors of anger arousal and hostility.Journalof Personality and Social Psychology, 56, 1013-1022.

    Leary, M. R., Bednarski, R., Hammon, D., & Duncan, T. (1997). Blow-hards, snobs, and narcissists: Interpersonal reactions to excessive ego-tism. In R. M. Kowalski (Ed.),Aversive interpersonal behaviors (pp.111-131). New York: Plenum Press.

    Leary, M. R., & Downs, D. L. (1995). Interpersonal functions of the

    self-esteem motive: The self-esteem system as a sociometer. In M. H.Kemis (Ed.),Efficacy, agency, and self-esteem (pp. 123-144). NewYork: Plenum Press.

    Leary, M. R., Tambor, E. S., Terdal, S. K., & Downs, D. L. (1995).Self-esteem as an interpersonal monitor: The sociometer hypothesis.Journal of Personality and Social Psychology, 68, 518-530.

    Lydon, J. E., Jamieson, D. W., & Zanna, M. P. (1988). Interpersonalsimilarity and the social and intellectual dimensions of first impressions.Social Cognition, 6, 269-286.

    Mead, G. H. (1934).Mind, self, and society. Chicago: University ofChicago Press.

    Mednick, S. A. (1968). The Remote Associates Test.Journal of CreativeBehavior, 2, 213-214.

    Mikulincer, M. (1998). Adult attachment style and affect regulation: Stra-tegic variations in self-appraisals.Journal of Personality and SocialPsychology, 75, 420-435 .

    Morf, C. C , & Rhodewalt, F. (1993). Narcissism and self-evaluationmaintenance: Explorations in object relations.Personality & SocialPsychology Bulletin, 19,668-676.

    Paulhus, D. L . (1998). Interpersonal and intrapsychic adaptiveness of traitself-enhancement: A mixed blessing?Journal of Personality and SocialPsychology, 74, 1197-1208.

    Pelham, B. W. (1995). Self-investment and self-esteem: Evidence for aJamesian model of self-worth.Journal of Personality and Social Psy-chology, 69, 1141-1150.

    Raskin, R., Novacek, J., & Hogan, R . (1991a). Narcissism, self-esteem, anddefensive self-enhancement.Journal of Personality, 59, 19-38.

    Raskin, R., Novacek, J., & Hogan, R. (1991b). Narcissistic self-esteemmanagement.Journal of Personality and Social Psychology, 60, 9 11 -918.

    Raskin, R., & Terry, H. (1988). A principal-components analysis of theNarcissistic Personality Inventory and further evidence of its constructvalidity. Journal of Personality and Social Psychology, 54, 890-902.

    Rhodewalt, F., &Morf, C. C. (1995). Self and interpersonal correlates ofthe Narcissistic Personality Inventory: A review and new findings.

    Journal of Research in Personality; 29 , 1-23.Rhodewalt, F., &Morf, C. C. (1998). On self-aggrandizement and anger:

    A temporal analysis of narcissism and affective reactions to success andfailure. Journal of Personality and Social Psychology, 74, 672-685.

    Riggio, R. E., Throckmorton, B., & DePaola, S. (1990). Social skills andself-esteem.Personality and Individual Differences, 11, 799-804 .

    Robbins, S. B.f & Dupont, P. (1992). Narcissistic needs of the self andperceptions of interpersonal behavior.Journal of Counseling Psychol-ogy, 39, 462-467.

    Rutter, M. (1979). Maternal deprivation, 1972-1978: New findings, newconcepts, new approaches.Child Development, 50,283-305.

    Scheier, M. F., Carver, C. S., & Bridges, M. W. (1994). Distinguishingoptimism from neuroticism (and trait anxiety, self-mastery, andself-esteem): A reevaluation of the Life Orientation Test,Journal of Person-

    ality and Social Psychology, 67, 1063-1078.Scnlenker, B. R., Weigold, M. F., & Hallam, J. R. (1990). Self-servingattributions in social context: Effects of self-esteem and social pressure.Journal of Personality and Social Psychology, 58, 855-863.

    Schmutte, P. S., &Ryff, C. D. (1997). Personality and well-being: Reex-amining methods and meanings.Journal of Personality and SocialPsychology, 73, 549-559.

    Schneider, D. J., & Turkat, D . (1975). Self-presentation following successor failure: Defensive self-esteem models.Journal of Personality, 43,127-135.

    Sedikides, C, Camp bell, K. W., Reeder, G., & Elliot, A. T. (1998). Theself-serving bias in relational context.Journal of Personality and SocialPsychology, 74, 378-386.

    Shrauger, J. S., & Schoeneman, T. J. (197 9). Symbolic interactionist view

  • 7/28/2019 Interpersonal Evaluations Following Threats to Self: Role of Self-Esteem

    12/12

    736 HEATHERTON AND VOHS

    of self-concept: Throughthe looking glass darkly.Psychological Bulle-tin, 86, 549-573.

    Smalley, R. L., & Stake, J. E. (1996). Evaluating sourcesof ego-threatening feedback: Self-esteemand narcissism effects. Journal ofResearch in Personality, 30, 483-495.

    Swann, W. B., Jr. (1996). Self-traps: The elusive quest for higher self-

    esteem. New York: Freeman.Taylor, S, E., & Brown, J. D. (1988). Illusionand well-being:A socialpsychological perspective on mental health.Psychological B ulletin, 103,193-210.

    Taylor, S. E., Peplau, L. A.,& Sears, D.0 . (1997).Social psychology (9thed.). Upper Saddle River,NJ: Prentice-Hall.

    Tesser, A., Millar, M., & Moore, J. (1988). Some affective conse-

    quences of social comparisonand reflection processes:The pain andpleasure of being close.Journal of Personality and Social Psychol-ogy, 54, 4 9 - 6 1 .

    Walster, E. (1965). Th e effectsof self-esteem on romantic liking.Journalof Experimental Social Psychology, 1, 184-197.

    Williams,J. M. G. (1997). Depression.In D. M. d a r k & C. G. Fairburn

    (Eds.), Science and practice of cognitive behaviour therapy (pp. 2 5 9 -283). Oxford, England: Oxford University Press.

    Received March 17, 1999

    Revision received July 11, 1999

    Accepted July 28, 1999

    Correction to Garcia-Marquesand Mackie (1999)

    In the article "The Impact of Stereotype-Incongruent Information on Perceived Group Variability

    and Stereotype Change," by Leonel Garcia-Marques and Diane M. Mackie (Journal of Personality

    and Social Psychology, 1999, Vol. 77, No. 5, pp. 979-990), Table 3 (p. 987) contained an error.

    The row "Number of subgroups" was inadvertently omitted. The corrected table appears below.

    Table 3

    Dispersion and Central Tendency Measures as a Function of Critical Information

    and Cognitive Load Levels (Experiment 3)

    Dependent measureIncongruent/

    no-loadIncongruent/

    load-at-encodingIncongruent/

    load-at-retrieval Control

    Chosen dispersion levelRange spreadProbability of distribution scoreStandard deviationNumber of subgroupsChosen mean levelMean rangeDistribution meanImpression ratings

    2.6714.44

    .881.824.333.39

    11.446.057.05

    2.2512.50

    .851.884.073.69

    11.946.176.94

    1.9312.46

    .831.604.333.87

    12.806.446.47

    1.9611.37

    .821.723.323.93

    11.966.126.89

    Note. N = 75. Higher numbers represent greater stereotypicalityor higher variability.