We are, therefore they arenâ€™t: Ingroup construal as a standard

Journal of Experimental Social Psychology 41 (2005) 515–526

www.elsevier.com/locate/jesp

We are, therefore they aren’t: Ingroup construal as a standard of comparison for outgroup judgments�

Bertram Gawronskia,¤, Galen V. Bodenhausenb, Rainer Bansec

a Department of Psychology, University of Western Ontario, Social Science Centre, London, Ont., Canada N6A 5C2b Department of Psychology, Northwestern University, 2029 Sheridan Road, Evanston, IL 60208-2710, USA

c Department of Psychology, University of York, Heslington, York YO10 5DD, UK

Received 8 January 2004; revised 26 October 2004Available online 10 February 2005

Abstract

Five studies tested the assumptions: (a) that ingroups are habitually used as a standard of comparison for outgroup judgments,and (b) that outgroup judgments are generally contrasted away from a momentary construal of the ingroup. Results generally sup-port these assumptions. Experiments 1 and 2 demonstrated increased activation levels of ingroup knowledge as a result of corre-sponding outgroup judgments. Experiments 3 and 4 showed that outgroup judgments depend not only on cognitively accessibleoutgroup exemplars, but also on accessible ingroup exemplars. Finally, Experiment 5 demonstrated that the impact of accessibleingroup exemplars on outgroup judgments is mediated by changes in the construal of the ingroup, such that: (a) outgroups werejudged lower with regard to a given trait the higher participants perceived their ingroup with regard to that trait, and (b) controllingfor the eVect of ingroup construal attenuated the obtained eVects on outgroup judgments. 2005 Elsevier Inc. All rights reserved.

Social judgments are context dependent in a varietyof ways. For instance, the interpretation of a particularbehavior often depends on the situational context inwhich the behavior is observed (e.g., Snyder & Frankel,1976; Trope, Cohen, & AlWeri, 1991; Trope, Cohen, &Maoz, 1988). In a similar vein, the same behavior may bejudged diVerently as a function of the group membershipof the actor (e.g., Duncan, 1976; Dunning & Sherman,

� This research was partially supported by a grant from the GermanScience Foundation (DFG) to the Wrst author (Ga 747/1-1). Experi-ment 3 was conducted while the Wrst and the third author were aYliat-ed with Humboldt-University Berlin, Germany. We thank KaiEpstude, Thomas Mussweiler, and Katja Rüter for helpful commentson an earlier version of this article; Angela Chiong, Scott Gelman, Ju-lia Grafe, Sara Hart, Iris Petzold, and Kimmy Van for their help in col-lecting the data; and Johanna Zukova for her help in preparing thestimulus material.

¤ Corresponding author.E-mail address: [email protected] (B. Gawronski).

0022-1031/$ - see front matter 2005 Elsevier Inc. All rights reserved.doi:10.1016/j.jesp.2004.11.002

1997; Gawronski, Geschke, & Banse, 2003; Hugenberg& Bodenhausen, 2003; Sagar & SchoWeld, 1980).

A special case of such context eVects on social judg-ments is the inXuence of diVerent comparison standards(Festinger, 1954). Perceivers often tend to judge the sameobject diVerently, depending on the standard of compar-ison they employ. For example, people may judge them-selves as more attractive when they compare themselvesto an unattractive person. However, they may judgethemselves as less attractive when they compare them-selves to an attractive person. Since nearly any kind ofsocial judgment implies a comparison to a standard(Mussweiler, 2003), changing standards can have strongeVects on how people judge a particular target (e.g., Bier-nat & Manis, 1994; Higgins & Lurie, 1983; Kahneman &Miller, 1986).

The main goal of the present research was to investi-gate the role of comparison standards for judgmentsabout outgroups. SpeciWcally, we propose: (a) thatpeople habitually use a construal of their ingroup as a

mailto: [email protected]

mailto: [email protected]

516 B. Gawronski et al. / Journal of Experimental Social Psychology 41 (2005) 515–526

standard of comparison for judgments about outgroups,and (b) that perceivers usually try to diVerentiatebetween outgroups and their ingroup, thus leading tocontrast eVects with respect to ingroup construals andoutgroup judgments. For instance, students of psychol-ogy may judge students of other areas as extraverted ifthey consider psychology students to be introverted.However, psychology students may judge students ofother areas as introverted if they consider their peers tobe extraverted. Most importantly, contextual inXuenceson psychology students’ perceptions of their peers maynot only aVect their judgments about psychology stu-dents in general. Rather, such inXuences may also aVecttheir judgments about students of other areas, in thatthose judgments are habitually based on a contrast withpsychology students.

Self-reference and diVerentiation in social comparison

Our Wrst hypothesis, that ingroups are habitually usedas a standard of comparison for outgroup judgments, isconsistent with a basic notion of social identity theory(Tajfel & Turner, 1986) and self-categorization theory(Turner, 1987). SpeciWcally, these theories assert thatoutgroup judgments should generally trigger an inter-group perspective, whereas ingroup judgments can trig-ger either an intragroup or an intergroup focus.However, even though previous research oVers clear evi-dence for these assumptions (e.g., Haslam, Oakes,Turner, & McGarty, 1995), we are not aware of anystudy that has directly tested their implications for pro-cesses of social comparison.

More direct evidence for self-referential comparisonprocesses comes from studies that investigated the use ofhabitual standards on the individual level. SpeciWcally,these studies have consistently shown that judgmentsabout other individuals often employ the self as a stan-dard of comparison (e.g., Dunning & Hayes, 1996; Holy-oak & Gordon, 1983; Karylowski, 1990; Srull & Gaelik,1983). Dunning and Hayes (1996), for example, foundthat participants provided descriptions about their ownbehavior more quickly when they had previously judgedanother individual with respect to same behavior. Thisresult suggests that people activate correspondingknowledge about themselves when judging other indi-viduals. Applied to present case, one could argue that the

predominance of self-referential comparisons also holdsat the group level, such that people base judgmentsabout outgroups on a comparison to their ingroup (seealso Mummendey & Wenzel, 1999). Accordingly, thesame outgroup may be judged diVerently, depending onpeople’s momentary construal of their ingroup.

Notwithstanding the proposed role of ingroups as astandard of comparison, merely comparing ingroupsand outgroups has uncertain implications for the partic-ular eVect such a comparison will have, because it couldlead either to assimilation or to contrast eVects on out-group judgments (Mussweiler, 2003). Drawing on previ-ous evidence for the accentuation hypothesis (Tajfel &Wilkes, 1963), we argue that, by default, people try todiVerentiate between social groups, and thus betweentheir ingroup and corresponding outgroups (e.g., Krue-ger & Clement, 1994; Krueger, Rothbart, & Sriram,1989). According to Cadinu and Rothbart (1996, p. 662),for example, “judges should, on the basis of what theyknow about one group, tend to infer the opposite aboutanother group, and this tendency should exist for judgeswhether or not they are members of the relevant targetgroups.” Applied to the present case, this reasoningimplies that people should generally contrast their judg-ments about an outgroup away from their momentaryconstrual of the ingroup. For instance, people may judgean outgroup as less favorable when they have a positiveconstrual of their ingroup, but they may judge an out-group as more favorable when they have a negative con-strual of their ingroup.

Overview of the experiments

The proposed sequence of processes in outgroupjudgments is depicted in Fig. 1. SpeciWcally, we arguethat outgroup-related inferential goals (Step 1) willhabitually elicit a judgment about the ingroup (Step 2).Outgroups are then diVerentiated from this spontaneousconstrual of the ingroup (Step 3), which in turn leads toa contrast eVect on outgroup judgments (Step 4). Inorder to provide empirical evidence for these processes,we conducted Wve experiments. Experiments 1 and 2were concerned with the impact of outgroup-relatedinferential goals (Step 1) on the cognitive accessibility ofthe ingroup (Step 2). These experiments tested whethermerely judging an outgroup inXuences the activation

Fig. 1. Proposed sequence of processes involved in outgroup judgments.

B. Gawronski et al. / Journal of Experimental Social Psychology 41 (2005) 515–526 517

level of relevant ingroup knowledge. SpeciWcally, Experi-ments 1 and 2 manipulated outgroup-related inferentialgoals (Step 1) and then measured the activation level ofrelevant ingroup knowledge (Step 2) with a lexical deci-sion task (Experiment 1) and a judgment facilitationtask (Experiment 2). If ingroups are habitually used as astandard of comparison, merely judging an outgroupshould be suYcient to increase the activation level of rel-evant ingroup knowledge. Expanding on these studies,Experiments 3 and 4 were particularly concerned withthe proposed diVerentiation of outgroup judgmentsfrom momentary ingroup construals (Step 3). For thispurpose, these studies manipulated participants’ con-strual of their ingroup (Step 2) and then assessed judg-ments of corresponding outgroups (Step 4). Experiment3 manipulated ingroup construals by providing informa-tion about an ingroup member; Experiment 4 manipu-lated ingroup construals by asking participants togenerate ingroup members with a speciWc trait. If out-group judgments are generally contrasted from amomentary construal of the ingroup, contextuallyinduced changes in the perception of the ingroup shouldlead to diametrical changes in judgments about out-groups. Finally, Experiment 5 addressed the proposedconnection between ingroup construal (Step 2) andingroup-outgroup accentuation (Step 3) more directly.In this study, we manipulated participants’ construal oftheir ingroup (Step 2) and then assessed judgments ofcorresponding outgroups (Step 4). In contrast to Experi-ments 3 and 4, however, we additionally assessed partici-pants’ construal of their ingroup in order to test whetherthe obtained eVects on outgroup judgments are indeedmediated by participants’ construal of their ingroup.

Experiment 1

The main objective of Experiment 1 was to testwhether merely judging an outgroup inXuences the cog-nitive accessibility of the ingroup category. If peoplehabitually use their ingroup as a standard of comparisonfor outgroup judgments, then merely judging an out-group should result in an increased activation level ofthe corresponding ingroup category. In order to test thisassumption, participants were asked to make severaljudgments about a particular outgroup. Immediatelyafterwards, the activation level of corresponding andnon-corresponding ingroup categories was assessed witha lexical decision task.

Method

Participants and designA total of 47 Northwestern students of US American

origin (34 female; 13 male) took part in a battery of twoexperiments on social judgment including the present

study and a second one on impression formation. Orderof the two studies was counterbalanced and did notaVect any of the dependent measures. Subjects receivedcredit for experiment participation requirements. Partici-pants were randomly assigned to judge one of two out-groups (either Germans or Harvard students) and thencompleted measures of the accessibility of ingroup cate-gories concerning both Americans and Northwesternstudents. It was hypothesized that merely judging a par-ticular outgroup would result in increased accessibilityof associates of the corresponding ingroup category butnot of associates of the non-corresponding ingroup cate-gory (e.g., judging “Germans” would activate the cate-gory “Americans” but not “Northwestern students”).

ProcedureOn arrival, participants were welcomed by an experi-

menter and seated in front of a personal computer. Writ-ten instructions explained that the study concernedjudgments about social groups. For this purpose, partici-pants would be asked to indicate their personal opinionsabout one particular social group. Following the instruc-tions, participants were asked to judge either Harvardstudents or Germans (i.e., outgroups) with respect to atotal of four diVerent traits (eYcient, outgoing, studious,and sociable). These judgments were assessed with ratingscales ranging from 1 (very low) to 7 (very high). Imme-diately afterwards, participants completed a lexical deci-sion task designed to assess the activation level of theingroup categories “Northwestern students” and“Americans.” SpeciWcally, participants were presentedeight Northwestern-related words (Ladycats, Wildcard,Evanston, Purple, Kellogg, Norris, Lake, and North-western), eight American-related words (Freedom, Clin-ton, Bush, Capitalism, Flag, Liberty, Power, andAmerica), and 16 corresponding non-words, which wereselected on the basis of pretests. Words and non-wordswere randomly presented in the center of the screen. Par-ticipants’ task was to press a right-hand key (“5” of thenumber pad) when a meaningful English word was pre-sented on the screen, and a left hand key (“A”) when ameaningless non-word was presented on the screen. Par-ticipants were instructed to respond as quickly as possi-ble without making too many errors.

Results and discussion

If participants habitually use their correspondingingroup as a standard of comparison for outgroup judg-ments, then the activation level of the correspondingingroup category should be higher, compared to a non-corresponding ingroup category. Hence, participantsshould be faster in reacting to Northwestern-relatedwords when they had previously judged Harvard stu-dents than when they had previously judged Germans.In contrast, participants should be faster in reacting to


American-related words when they had previouslyjudged Germans than when they had previously judgedHarvard students.

In order to test this assumption, response latencieswere Wrst log-transformed (Fazio, 1990) and then sub-mitted to a 2 (outgroup) £ 2 (ingroup) mixed-modelanalysis of variance (ANOVA) with the Wrst variable asbetween-subjects factor and the second as within-sub-jects factor.1 Consistent with the present predictions, thisanalysis revealed a signiWcant two-way interaction ofingroup category and outgroup judgment,F (1, 45) D 4.00, p D .05, �2 D .082. The interaction patterndisplayed in Fig. 2 indicates that participants were fasterin responding to Northwestern-related words when theyhad judged Harvard students before than when they hadjudged Germans before. In contrast, participants werefaster in responding to American-related words whenthey had judged Germans before than when they hadjudged Harvard students before. In other words, partici-pants were faster in responding to ingroup related wordswhen they had previously judged a corresponding out-group than when they had previously judged a non-cor-responding outgroup. These results are consistent withthe assumption that merely judging an outgroupincreases the cognitive accessibility of the correspondingingroup category.

Even though the present Wndings are consistent withour assumption that people habitually use a construal oftheir ingroup as a standard of comparison for outgroupjudgments, one could object that these results indicateonly a spontaneous activation of the category. However,they do not indicate that participants activate judgment-relevant ingroup knowledge that could be employed fora comparison. In other words, it is not clear whether par-

1 Even though analyses were conducted with log-transformed data,cell means are reported in milliseconds for ease of interpretation.

Fig. 2. Mean response latencies in milliseconds for the identiWcation ofingroup-related words as a function of ingroup category (Northwest-ern students vs. Americans) and prior outgroup judgment (Harvardstudents vs. Germans), Experiment 1.

ticipants only activated their ingroup category withoutusing this category as a standard of comparison, orwhether the increased activation level of the ingroup cat-egory arises in the course of making a judgment aboutthe ingroup. The second experiment addressed this ques-tion more directly.

Experiment 2

The main goal of Experiment 2 was to test whetherjudgments about outgroups inXuence the activation levelof judgment-relevant ingroup knowledge. In order totest this assumption, we employed a judgment facilita-tion paradigm successfully used in previous studies onsocial comparison (e.g., Dunning & Hayes, 1996;Mussweiler & Bodenhausen, 2002) and judgmentalanchoring (e.g., Mussweiler & Strack, 1999; Strack &Mussweiler, 1997). SpeciWcally, we Wrst asked partici-pants to judge an outgroup with respect to a given trait.Immediately afterwards, participants were asked tojudge either a corresponding or a non-correspondingingroup with respect to the same trait. As a dependentmeasure, we assessed participants’ response latency fortheir ingroup judgments. If perceivers habitually usetheir ingroup as a standard of comparison for judgmentsabout outgroups, subsequent ingroup judgments shouldbe faster when ingroup and outgroup correspond to oneanother (i.e., when participants already judged therespective ingroup in the course of making a judgmentabout the outgroup), but not when ingroup and out-group do not align (i.e., when participants did not makea judgment about the respective ingroup before).

Method

Participants and designA total of 75 Northwestern students of US American

origin (46 female; 29 male) took part in a battery of twoexperiments on social judgment, including the presentstudy and a second one on impression formation. Orderof the two studies was counterbalanced and did notaVect any of the dependent measures. Subjects receivedcredit for experiment participation requirements. Partici-pants were randomly assigned to the conditions of a 2(outgroup judgment target: Harvard students orGermans) £ 2 (ingroup judgment target: Northwesternstudents or Americans) between-subjects design.


menter and seated in front of a computer. Writteninstructions explained that the present study was con-cerned with judgments about social groups. For this pur-pose, participants would be asked to indicate theirpersonal opinions about a number of diVerent social


groups. Following the instructions, participants wereWrst asked to judge the eYciency of either Harvard stu-dents or Germans (i.e., outgroups) on a rating scale rang-ing from 1 (very low) to 7 (very high). Immediatelyafterwards, participants were asked to judge theeYciency of either Northwestern students or Americans(i.e., ingroups) on a scale with the same response format.As a dependent measure, participants’ response latencyfor their ingroup judgment was assessed.


Response latencies for ingroup judgments were Wrstlog-transformed (Fazio, 1990) and then submitted to a 2(outgroup) £ 2 (ingroup) ANOVA.2 This analysisrevealed a signiWcant two-way interaction, F (1,70) D7.02, p D .01, �2 D .090 (see Fig. 3). Consistent with ourpredictions, the interaction pattern indicates that partici-pants were faster in judging Northwestern studentswhen they had judged Harvard students before thanwhen they had judged Germans before. In contrast,judgments about Americans were faster when partici-pants judged Germans before than when they judgedHarvard students before. In other words, participantswere faster in judging their ingroup with regard to a spe-ciWc trait when they had previously judged a correspond-ing outgroup with regard to the same trait than whenthey had previously judged a non-corresponding out-group. These results indicate that merely judging an out-group not only leads to a spontaneous activation of theingroup category, but also to an activation of judgment-relevant ingroup knowledge.

However, even though these Wndings are consistentwith our assumption that ingroups are habitually used

2 Even though analyses were conducted with log-transformed data,cell means are reported in milliseconds for ease of interpretation.

Fig. 3. Mean response latencies in milliseconds for ingroup judgmentsas a function of ingroup category (Northwestern students vs. Ameri-cans) and prior outgroup judgment (Harvard students vs. Germans),Experiment 2.

as a standard of comparison for outgroup judgments, itis still an open question whether perceivers actually usethe activated ingroup knowledge when making judg-ments about the outgroup. Moreover, the present resultsare still ambiguous as to whether perceivers actuallycontrast outgroup judgments from their spontaneousconstrual of their ingroup, as we hypothesized. Theseassumptions were addressed in Experiment 3.

Experiment 3

Several models of social judgment suggest that acces-sible exemplars play an important role in judgmentsabout social groups (e.g., Lord & Lepper, 1999; Smith &Zárate, 1992). This assumption is consistent with a num-ber of studies demonstrating that recently encounteredmembers of a given group aVect judgments about thegroup in general (e.g., Bless, Schwarz, Bodenhausen, &Thiel, 2001; Bodenhausen, Schwarz, Bless, & Wänke,1995; Henderson-King & Nisbett, 1996; Sia, Lord, Bles-sum, Thomas, & Lepper, 1999; Wilder, Simon, & Faith,1996). Applied to the present investigation, one couldargue that cognitively accessible ingroup membersshould aVect judgments about the ingroup in general(e.g., Marques, Yzerbyt, & Leyens, 1988) and therebyalso inXuence judgments about corresponding out-groups. Moreover, given the present assumption thatoutgroup judgments are generally contrasted from aconstrual of the ingroup, cognitively accessible ingroupexemplars should lead to contrast eVects in judgmentsabout outgroups, such that positive ingroup exemplarsshould lead to less favorable judgments about the out-group whereas negative ingroup exemplars should leadto more favorable judgments about the outgroup.

In order to test these assumptions, participants werepresented a videotaped interview of either an ingroupmember or an outgroup member who behaved either ina friendly or an unfriendly manner. Afterwards, evalua-tions of the outgroup in general were assessed. Drawingon previous evidence for the impact of cognitively acces-sible exemplars on group judgments (e.g., Bless et al.,2001; Bodenhausen et al., 1995; Henderson-King &Nisbett, 1996; Sia et al., 1999; Wilder et al., 1996), weexpected a positive outgroup member to lead to morepositive evaluations of the outgroup than a negative out-group member. Going beyond previous demonstrations,however, we expected a positive ingroup member to leadto more negative evaluations of the outgroup as com-pared to a negative ingroup member.

Method

Participants and designA total of 122 German psychology students (103

female, 19 male) participated in a study on impression


formation, receiving credit for experiment participationrequirements. Participants were randomly assigned toone of the four experimental conditions of a 2 (target’sethnic origin: German vs. Turkish) £ 2 (behavior:friendly vs. unfriendly) between-subjects design.


menter and informed that they were taking part in astudy on impression formation. The experimenterexplained that a number of interviews had been con-ducted about public transport in Berlin (Germany) andthat each participant would watch one of the videotapedinterviews. One of four videos was then assigned by arandom procedure. The videos consisted of a briefsequence in which a female interviewer approachedeither a typically German- or Turkish-looking malepasserby (a confederate of the experimenters), asking ifhe had some time for a short interview concerning hissatisfaction with the public transport system in Berlin.Questions and answers were scripted and held identicalacross the clips. The Turkish actor had a slight Turkishaccent; the German actor had no foreign accent.Friendly and unfriendly behavior was operationalizedby the target’s nonverbal behavior (e.g., looking inter-ested vs. disinterested) and the tone of his answers (e.g.,happy vs. irritated). After a few questions, the target ter-minated the interview with the excuse to be in a hurry.As a function of the experimental conditions, this expla-nation was accompanied either by a smile (friendlybehavior) or by a frown (unfriendly behavior). Afterwatching the clip, participants were asked to indicatetheir impression of the target’s behavior and to predicthis behavior in a number of hypothetical situations. Thistask was followed by a questionnaire to assess preju-diced beliefs about Turkish people in general. Finally,participants were probed for suspicion, debriefed, andthanked for their participation.

MeasuresPrejudice against Turkish people was assessed with a

German version of Coenders, Scheepers, Sniderman, andVerberk’s (2001) General Prejudice Scale.3 Prejudice wasassessed with rating scales ranging from 1 to 5. In orderto assure that the targets were actually perceived asTurkish or German, participants were asked to rate thetarget with respect to his ethnic origin (i.e., German,

3 Coenders et al.’s (2001) General Prejudice Scale is an adaptationof Pettigrew and Meertens’ (1995) Subtle and Blatant Prejudice Scales.The General Prejudice Scale includes all items of the Subtle and Bla-tant Prejudice Scales, the only exception being the exclusion of allitems assessing the exaggeration of cultural diVerences. For the presentpurpose, Coenders et al.’s adaptation was preferred over Pettigrew andMeertens’ original scale, because the former (in contrast to the latter)does not explicitly require a comparison between ingroup and out-group.

Turkish) on two scales ranging from 1 (very unlikely) to5 (very likely). Participants’ judgments of the target’sbehavior were assessed on six dimensions (e.g., friendly,rude) which were rated on response scales ranging from1 (not true) to 5 (true). In order to promote elaborateprocessing of the available exemplar information, partic-ipants were additionally asked to predict the target’sbehavior in 10 hypothetical situations that could eliciteither positive or negative behavior (Gawronski et al.,2003). The subjective likelihood of negative behaviorwas rated on response scales ranging from 1 (veryunlikely) to 5 (very likely).


Manipulation checksRatings of the target’s behavior were averaged into a

single index of negative behavior by reverse coding posi-tive behavioral dimensions (Cronbach’s �D .89). Submit-ted to a 2 (ethnic origin) £ 2 (behavior) ANOVA, thisindex revealed a theoretically uninteresting main eVectof ethnicity, F (1, 118) D 36.80, p < .001, �2 D .238, indicat-ing that the behavior was rated more negatively whenthe target was German (M D 3.54) than when he wasTurkish (M D 2.87). More importantly, there was a sig-niWcant main eVect of behavior, F (1, 118) D 133.18,p < .001, �2 D .530, indicating that the behavior was ratedmore negatively when the target behaved in anunfriendly manner (M D 3.83) than when he behaved in afriendly manner (M D 2.58). This eVect held for both theGerman target, F (1, 59) D 48.23, p < .001, �2 D .450, andthe Turkish target, F (1,59) D 87.97, p < .001, �2 D .599,and was not qualiWed by higher order interactions. Thesame main eVect of behavior was obtained for thebehavior prediction index (Cronbach’s �D .73), reveal-ing higher scores in the prediction of negative behaviorwhen the target was unfriendly (M D 3.07) than when hewas friendly (M D 2.67), F (1, 118) D 19.44, p < .001,�2 D .141. Again, this eVect held for both the German tar-get, F (1,59) D 13.18, p D .001, �2 D .183, and the Turkishtarget, F (1, 59) D 7.53, p D .008, �2 D .113, and was notqualiWed by any higher order interaction. Moreover, a 2(target ethnic origin: German vs. Turkish) £ 2 (behavior:friendly vs. unfriendly) £ 2 (rating category: Germannessvs. Turkishness) mixed-model ANOVA on category rat-ings revealed a signiWcant two-way interaction betweenethnic origin and rating category, F (1, 118) D 131.11,p < .001, �2 D .526, indicating that the German-lookingtarget was rated as more likely to be German rather thanTurkish (Ms D 4.23 vs. 1.52, respectively), F (1,60)D 246.38, p < .001, �2 D .804. In contrast, the Turkish-looking target was rated more likely to be Turkish ratherthan German (Ms D 2.31 vs. 3.13, respectively),F (1,60) D 10.55, p D .002, �2 D .150. Taken together, theseresults indicate that both experimental manipulationscan be regarded successful.


Prejudice against the outgroupPrejudice ratings were merged into a single index by

calculating mean values (Cronbach’s �D .83). This indexwas submitted to a 2 (ethnic origin) £ 2 (behavior)ANOVA, revealing a signiWcant two-way interaction,F (1, 118) D 5.36, p D .02, �2 D .043 (see Fig. 4). Consistentwith previous research on exemplar eVects on groupjudgments (e.g., Bless et al., 2001; Bodenhausen et al.,1995; Henderson-King & Nisbett, 1996; Sia et al., 1999;Wilder et al., 1996), participants who watched a Turkishtarget showed a non-signiWcant, but recognizable ten-dency to exhibit a lower level of prejudice against Turk-ish people when this target was friendly than when hewas unfriendly, F (1,59) D 1.04, p D .31, �2 D .017. Morenovel, however, was the Wnding that respondents whowatched a German target exhibited a signiWcantly higherlevel of prejudice against Turkish people when this tar-get was friendly than when he was unfriendly,F (1, 59) D 5.31, p D .02, �2 D .083.

These results further corroborate our hypothesis thatpeople habitually use their ingroup as a standard ofcomparison for outgroup judgments. Moreover, thepresent Wndings oVer Wrst evidence for the assumptionthat outgroup judgments are contrasted away from amomentary construal of the ingroup. Consistent withthis assumption, participants showed a higher level ofprejudice against a corresponding outgroup when theysaw a videotaped interview of a positive ingroup exem-plar than when they saw a negative ingroup exemplar.These results indicate that outgroup judgments are notonly aVected by cognitively accessible outgroup mem-bers, as prior research has shown (e.g., Bless et al., 2001;Bodenhausen et al., 1995; Henderson-King & Nisbett,1996; Sia et al., 1999; Wilder et al., 1996), but also byaccessible ingroup exemplars. SpeciWcally, it seems thatingroup exemplars aVect the subjective construal of theingroup, which in turn is used as a standard of compari-son for judgments about the outgroup.

Fig. 4. Mean scores of prejudice against outgroup as a function of tar-get group membership (ingroup vs. outgroup) and target behavior(friendly vs. unfriendly), Experiment 3.

Experiment 4

The main goal of Experiment 4 was to replicate theWndings obtained in Experiment 3 using a diVerentmanipulation of ingroup construal. Rather than provid-ing participants with information about ingroup mem-bers, participants in Experiment 4 were asked togenerate ingroup members with a speciWc characteristic.This manipulation was inspired by previous research onthe ease-of-retrieval eVect, showing that people oftenbase their judgments on meta-cognitive inferences fromthe experienced ease of retrieving information ratherthan on the particular content or amount of retrievedinformation (for a review, see Schwarz, Bless, Wänke, &Winkielman, 2003). In a study by Schwarz et al. (1991),for example, participants were asked to recall either ahigh or a low number of assertive behaviors that theyhad engaged in, and to indicate their general level ofassertiveness afterwards. In contrast to what would beexpected by a mere accessibility eVect (cf. Tversky &Kahneman, 1973), participants rated themselves higherin assertiveness when they had to recall a low number ofassertive behaviors than when they had to recall a highnumber of assertive behaviors. According to Schwarz etal. (1991), these results indicate that participants basedtheir self-judgments on the experienced ease of retrievinginformation rather than on the particular content oramount of the activated information. More precisely,participants seem to have inferred that they are not veryassertive when the recollection of assertive behaviorswas diYcult (high number), but they seem to haveinferred that they are very assertive if the recollection ofassertive behaviors was easy (low number).

Even though the particular processes underlying ease-of-retrieval eVects are still somewhat controversial (e.g.,Schwarz et al., 2003; Tormala, Petty, & Briñol, 2002;Wänke & Bless, 2000), Schwarz et al.’s manipulation ofself-judgments seems also suitable to inXuence partici-pants’ momentary construal of their ingroup, and thusto investigate the impact of ingroup construals on out-group judgments. SpeciWcally, participants may assumethat their ingroup has a high level of a given trait onlywhen it is easy to generate ingroup exemplars with ahigh trait level, but not when it is diYcult to generate rel-evant ingroup exemplars. Accordingly, ingroups shouldbe judged higher with respect to a given trait when theyhad to generate a low number of ingroup members withthis trait (easy) than when they had to generate a highnumber of ingroup members with this trait (diYcult).These momentary construals of the ingroup should inturn aVect judgments about outgroups, such that out-groups are judged higher with regard to the trait in ques-tion when participants perceive their ingroup to have alow level of that trait than when they perceive theiringroup to have high level of that trait. This reasoningimplies that outgroups should be judged higher with


respect to a given trait when participants had to generatea high number of ingroup members with this trait(diYcult D low trait level for ingroup) than when theyhad to generate a low number of ingroup members withthis trait (easy D high trait level for ingroup).

Method

A total of 32 psychology students (25 female; 7 male)were asked to think up either 3 or 10 psychology students(i.e., ingroup members) who are either high in introversionor high in extraversion. Thus, the experiment had 2 (num-ber of ingroup exemplars)£2 (trait dimension) between-subjects design. Assignment to the experimental conditionswas random. After completing the exemplar generationtask, participants were asked to judge students majoring in10 areas other than psychology (computer science, mathe-matics, physics, languages, pedagogy, chemistry, medicalscience, law, economics, and biology) with respect to theirintroversion/extraversion on 7-point scales ranging from¡3 (very introverted) to +3 (very extraverted). The experi-ment was run in a mass testing session in an undergraduatecourse on research methods at the end of the semester.Subjects received a candy bar for participation.


Judgments of students majoring in areas other than psy-chology were averaged into a single index by calculatingmean values (Cronbach’s �D .64). Submitted to a 2 (num-ber of ingroup exemplars)£2 (trait dimension) ANOVA,this index revealed a signiWcant two-way interaction oftrait level and number of exemplars, F(1,28)D10.35,pD .003, �2 D .270 (see Fig. 5). Consistent with the presentpredictions, participants judged non-psychology studentsas more extraverted when they had to generate a low num-ber of introverted psychology students (easy) than when

Fig. 5. Mean scores of judged extraversion of outgroup members as afunction of trait type of generated ingroup exemplars (introverted vs.extraverted) and number of generated ingroup exemplars (3 vs. 10),Experiment 4.

they had to generate a high number of introverted psychol-ogy students (diYcult). In contrast, participants judgednon-psychology students as less extraverted when they hadto generate a low number of extraverted psychology stu-dents (easy) than when they had to generate a high numberof extraverted psychology students (diYcult). Moreover,non-psychology students were judged higher in extraver-sion when participants had to generate a low number ofintroverted psychology students than when they had togenerate a low number of extraverted psychology students.In contrast, participants judged non-psychology studentsas less extraverted when they had to generate a high num-ber of introverted psychology students than when they hadto generate a high number of extraverted psychology stu-dents. Taken together, these results corroborate ourassumption that ingroups are habitually used as a standardof comparison for outgroup judgments, and that changesin the momentary construal of one’s ingroup also aVectjudgments about outgroups.

Experiment 5

In order to keep the proposed comparison betweeningroups and outgroups as spontaneous as possible, wedeliberately decided not to include measures of ingroupconstrual in Experiments 3 and 4. This strategy may beconsidered appropriate given that explicit requests foran ingroup judgment may stimulate social comparisonsthat participants would otherwise not engage in. How-ever, this strategy also confers the disadvantage of pre-cluding strong conclusions about the mediatingprocesses postulated in this paper. SpeciWcally, it is notclear if the eVects on outgroup judgments obtained inExperiments 3 and 4 were indeed mediated by partici-pants’ construal of their ingroup. The main goal ofExperiment 5 was to address this question more directly.In this study, we again employed an ease-of-retrievalmanipulation to inXuence participants’ perceptions oftheir ingroup. In contrast to Experiment 4, however, weadditionally assessed the impact of this manipulation onparticipants’ construal of their ingroup. These ingroupjudgments were then used to determine if the obtainedeVects on outgroup judgments are indeed mediated byparticipants’ construal of their ingroup, such that: (a)outgroups are judged lower with regard to a given traitthe higher participants perceive their ingroup withregard to that trait, and (b) controlling for the eVect ofingroup construal attenuates the obtained eVects on out-group judgments (cf. Baron & Kenny, 1986).

Method

Participants and designA total of 70 students of the University of Western

Ontario (47 female; 23 male) were asked to think up


either 3 or 10 Western Ontario students (i.e., ingroupmembers) who are either high in introversion or high inextraversion. After completing the exemplar generationtask, participants were asked to indicate how introverted/extraverted they think Western Ontario students are ingeneral. Immediately afterwards, participants were askedto judge the students of eight other Canadian universities(i.e., University of Waterloo, University of Toronto, Uni-versity of British Columbia, McGill University, SimonFraser University, Wilfrid Laurier University, YorkUniversity, and Concordia University) with regard tointroversion/extraversion. Ratings of introversion/extra-version were assessed with 7-point scales ranging from¡3 (very introverted) to +3 (very extraverted). As amanipulation check, we additionally asked participantsto indicate how diYcult they experienced the generationof the requested number of Western Ontario students ona scale ranging from 1 (very easy) to 7 (very diYcult). Theexperiment was run in a larger battery of multiple studies,lasting approximately 45 min. Subjects received credittoward experiment participation requirements.


Manipulation checksSupporting the eVectiveness of the present manipula-

tion, a 2 (number of ingroup exemplars) £ 2 (trait dimen-sion) ANOVA revealed that participants experienced thegeneration of 10 Western Ontario students as more diY-cult than the generation of 3 Western Ontario students(Ms D 5.15 vs. 3.78, respectively), F (1, 66) D 11.12,p D .001, �2 D .144. There was also a theoretically lessinteresting main eVect of trait dimension, indicating thatparticipants experienced the generation of introvertedWestern Ontario students as more diYcult than the gen-eration of extraverted Western Ontario students(Ms D 4.89 vs. 4.00, respectively), F (1, 66) D 4.62, p D .04,�2 D .065. There was no signiWcant interaction betweennumber of ingroup exemplars and trait dimension,F (1, 66) D .08, p D .77, �2 D .001.

Ingroup judgmentsA 2 (number of ingroup exemplars) £ 2 (trait dimen-

sion) ANOVA on judgments of Western Ontario stu-dents revealed a signiWcant two-way interaction,F (1, 66) D 5.23, p D .03, �2 D .073 (see Fig. 6). Consistentwith previous demonstrations of the ease-of-retrievaleVect, inspection of means indicated that participantsrated Western Ontario students higher in extraversionwhen they generated 3 extraverted Western Ontario stu-dents (easy) than when they generated 10 extravertedWestern Ontario students (diYcult). In contrast, partici-pants rated Western Ontario students lower in extraver-sion when they generated three introverted WesternOntario students (easy) than when they generated 10introverted Western Ontario students (diYcult).

Outgroup judgmentsJudgments of students of other universities were aver-

aged into a single index by calculating mean values(Cronbach’s � D .78). Submitted to a 2 (number ofingroup exemplars) £ 2 (trait dimension) ANOVA, thisindex revealed a signiWcant two-way interaction of traitdimension and number of exemplars, F (1, 66) D 4.24,p D .04, �2 D .060 (see Fig. 7). Replicating the interactionpattern obtained in Experiment 4, participants judgedstudents of other universities as more extraverted whenthey had to generate a low number of introverted West-ern Ontario students (easy) than when they had to gener-ate a high number of introverted Western Ontariostudents (diYcult). In contrast, participants judged stu-dents of other universities as less extraverted when theyhad to generate a low number of extraverted WesternOntario students (easy) than when they had to generate

Fig. 6. Mean scores of judged extraversion of ingroup members as afunction of trait type of generated ingroup exemplars (introverted vs.extraverted) and number of generated ingroup exemplars (3 vs. 10),Experiment 5.

Fig. 7. Mean scores of judged extraversion of outgroup members as afunction of trait type of generated ingroup exemplars (introverted vs.extraverted) and number of generated ingroup exemplars (3 vs. 10),Experiment 5.


a high number of extraverted Western Ontario students(diYcult).

In order to test whether the obtained eVects on out-group judgments were indeed mediated by participants’construal of their ingroup, we conducted an analysis ofcovariance (ANCOVA) with number of ingroup mem-bers and trait dimension as Wxed factors, ingroup judg-ments as covariate, and outgroup judgments asdependent measure. Consistent with our predictions, thisanalysis revealed a highly signiWcant eVect of the covari-ate, F (1, 66) D 16.47, p < .001, �2 D .202. SpeciWcally, stu-dents of other universities were judged as lessextraverted the more extraverted participants judgedWestern Ontario students (r D ¡.48, p < .001). The inter-action eVect of number of ingroup exemplars and traitdimension on outgroup judgments, however, failed toreach statistical signiWcance after controlling for ingroupjudgments, F (1, 66) D 1.21, p D .27, �2 D .018. A Sobel testindicated a signiWcant mediation eVect of ingroup judg-ments, z D 1.99, p < .05. Taken together, these results cor-roborate our assumptions: (a) that participants employ amomentary construal of their ingroup as a standard ofcomparison for their judgments about outgroups, and(b) that outgroups are contrasted from this momentaryconstrual of the ingroup.

General discussion

The main goal of the present research was to investi-gate the role of ingroup construals in judgments aboutoutgroups. SpeciWcally, we proposed: (a) that peoplehabitually use a construal of their ingroup as a standardof comparison for judgments about outgroups, and (b)that perceivers diVerentiate between outgroups and theiringroup, thus leading to contrast eVects with respect toingroup construal and outgroup judgments (see Fig. 1).More precisely, we argue that outgroup-related inferen-tial goals (Step 1) elicit a judgment about the ingroup(Step 2). Outgroups are then accentuated from this con-strual of the ingroup (Step 3), which in turn leads to acontrast eVect on outgroup judgments (Step 4).

Results from Wve studies oVer converging evidence forthese assumptions. Experiments 1 and 2 manipulatedoutgroup-related inferential goals (Step 1) and thenassessed the activation level of relevant ingroup knowl-edge (Step 2). Consistent with the assumption that par-ticipants habitually activate their ingroup in the courseof making judgments about outgroups, participants inExperiment 1 showed faster responses to ingroup-relatedwords in a lexical decision task when they judged a cor-responding outgroup before than when they judged anon-corresponding outgroup before. These results werecorroborated in Experiment 2, showing that participantswere faster in judging their ingroup when they hadjudged a corresponding outgroup before than when they

had judged a non-corresponding outgroup before.Drawing on these Wndings, Experiments 3 and 4 furtherdemonstrated that situationally induced changes in themomentary construal of an ingroup (Step 2) inXuencecorresponding judgments about outgroups (Step 4), andthat such changes in ingroup perceptions inXuence out-group judgments in a contrastive manner (Step 3). Spe-ciWcally, participants in Experiment 3 exhibited a higherlevel of prejudice against a corresponding outgroupwhen they saw a positive ingroup member than whenthey saw a negative ingroup member. This eVect ofmomentary ingroup construals was conceptually repli-cated in Experiment 4 employing an ease-of-retrievalmanipulation (cf. Schwarz et al., 1991). In this study, par-ticipants judged outgroups higher with respect to a giventrait when they had to generate a high number ofingroup exemplars with that trait (diYcult D low traitlevel for ingroup) than when they had to generate a lownumber of relevant ingroup exemplars (easy D high traitlevel for ingroup). Finally, Experiment 5 demonstratedthat the obtained eVects on outgroup judgments areindeed mediated by participants’ momentary construalof their ingroup (Step 2), such that: (a) outgroups werejudged lower with regard to a given trait the higher par-ticipants perceived their ingroup with regard to thattrait, and (b) controlling for the eVect of ingroup con-strual attenuated the obtained eVects on outgroup judg-ments.

The present studies extend previous evidence for self-referential comparisons in social judgment. A number ofexperiments have shown that people habitually use self-knowledge as a standard of comparison for judgmentsabout other individuals (e.g., Dunning & Hayes, 1996;Holyoak & Gordon, 1983; Karylowski, 1990; Srull &Gaelik, 1983). The present Wndings indicate that the pre-dominance of such self-referential comparisons alsoholds at the group level, such that people habitually basetheir judgments about outgroups on a spontaneous com-parison to their ingroup. However, an open questionrelated to this point is whether such comparisons to theingroup play a privileged role in judgments about out-groups. SpeciWcally, one could argue that even thoughthe ingroup may habitually be employed as a standardof comparison, judgments about a given outgroup couldalso be inXuenced by comparisons to other outgroupsthat are similar to this outgroup. With regard to judg-ments at the individual level, for example, Karylowski,Konarzewski, and Motes (2000) recently demonstratedthat self-referential comparison is only one of severalcomparison processes involved in judgments about otherindividuals. Notwithstanding these Wndings, however, itis important to note that ingroup comparisons may stillplay a pervasive role in outgroup judgments. Forinstance, US Americans may use Americans and theFrench as a standard of comparison when making judg-ments about Germans, and they may use Americans and


Saudi Arabians when making judgments about Iraqis. Inother words, even though judgments about outgroupsmay be inXuenced by comparisons to various (diVerent)outgroups, outgroup judgments may always be inXu-enced by comparisons to the (same) ingroup. Futureresearch comparing the role of diVerent comparisonstandards in outgroup judgments may help to clarify thepredominant role of ingroup related comparisons.

An open question is whether the obtained eVects ofingroup comparisons are more pronounced for stereo-typical or non-stereotypical trait dimensions. Withregard to stereotypical trait dimensions, for example,one could argue that such traits may be more likely toelicit the proposed contrasting between outgroups andthe ingroup. As such, the obtained eVects may be morepronounced for stereotypical as compared to non-ste-reotypical traits. On the other hand, however, one couldalso argue that judgments in stereotypical trait dimen-sions may be directly retrieved from memory withoutrequiring any kind of social comparison process. If so,then the obtained eVects may be more pronounced fornon-stereotypical as compared to stereotypical traitdimensions. Future research may help to clarify the roleof trait stereotypicality for the role of ingroup construalsin judgments about outgroups.

Another interesting question for future research iswhether the obtained eVects of ingroup exemplars onoutgroup evaluations are limited to explicit judgments,or whether similar eVects can be obtained on implicitattitude measures (e.g., Fazio, Jackson, Dunton, & Wil-liams, 1995; Greenwald, McGhee, & Schwartz, 1998;Wittenbrink, Judd, & Park, 1997). Dasgupta and Green-wald (2001), for example, found that that mere exposureto pictures of admired Black or disliked White individu-als can change Caucasian participants’ automatic evalu-ations of African Americans. In a similar vein, Lowery,Hardin, and Sinclair (2001) found that mere interactionwith a Black experimenter (as compared to a Whiteexperimenter) is capable to reduce automatic negativeevaluations of African Americans. In contrast to thepresent studies, however, these experiments did notexplicitly test the diVerential eVects of ingroup and out-group exemplars. In Dasgupta and Greenwald’s (2001)studies, for example, participants received either picturesof admired Blacks and disliked Whites or pictures of dis-liked Blacks and admired Whites, thus confoundingvalence and group membership. Accordingly, it is notclear whether the obtained eVects stem from the expo-sure to positive vs. negative ingroup exemplars or fromthe exposure to positive vs. negative outgroup exemplars(or perhaps both). In a similar vein, Lowery et al. (2001)did not manipulate the valence of Black and Whiteexperimenter’s behavior. Hence it is possible that Low-ery et al.’s results do not stem from mere interaction withan outgroup member in contrast to an ingroup member,but from the (presumably) positive behavior exhibited

by these individuals. If the experimenters had behavedunfriendly, mere interaction with a Black experimenterin contrast to a White experimenter might have beeninsuYcient to reduce the automatic activation of nega-tive attitudes toward African Americans. In contrast,negative behavior by a Black experimenter could havereinforced automatic negative evaluations whereas nega-tive behavior by a White experimenter could havereduced automatic negative evaluations. Future researchmay further investigate the eVects ingroup exemplars onimplicit outgroup evaluations.

In sum, the present Wndings suggest that diVering per-ceptions of an ingroup can have important consequencesfor judgments about outgroups. SpeciWcally, it seems: (a)that perceivers habitually use a momentary construal oftheir ingroup as a standard of comparison for outgroupjudgments, and (b) that perceivers generally accentuatethe diVerences between outgroups and their ingroup,thus leading to contrast eVects between ingroup con-strual and outgroup judgments. In other words, judg-ments about one and the same outgroup can be contextdepend when this context is able to shift people’s con-strual of their ingroup.

References

Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator vari-able distinction in social psychological research: Conceptual, stra-tegic, and statistical considerations. Journal of Personality andSocial Psychology, 51, 1173–1182.

Biernat, M., & Manis, M. (1994). Shifting standards and stereotypebased judgments. Journal of Personality and Social Psychology, 66,5–20.

Bless, H., Schwarz, N., Bodenhausen, G. V., & Thiel, L. (2001). Person-alized versus generalized beneWts of stereotype disconWrmation:TradeoVs in the evaluation of atypical exemplars and their socialgroups. Journal of Experimental Social Psychology, 37, 386–397.

Bodenhausen, G. V., Schwarz, N., Bless, H., & Wänke, M. (1995).EVects of atypical exemplars on racial beliefs: Enlightened racismor generalized appraisal? Journal of Experimental Social Psychol-ogy, 31, 48–63.

Cadinu, M. R., & Rothbart, M. (1996). Self-anchoring and diVerentia-tion processes in the minimal group setting. Journal of Personalityand Social Psychology, 70, 661–677.

Coenders, M., Scheepers, P., Sniderman, P. M., & Verberk, G. (2001).Blatant and subtle prejudice: Dimensions, determinants, and conse-quences. European Journal of Social Psychology, 31, 281–297.

Dasgupta, N., & Greenwald, A. G. (2001). On the malleability of auto-matic attitudes: Combating automatic prejudice with images ofadmired and disliked individuals. Journal of Personality and SocialPsychology, 81, 800–814.

Duncan, B. L. (1976). DiVerential perception and attribution of inter-group violence: Testing the lower limits of stereotyping of Blacks.Journal of Personality and Social Psychology, 34, 590–598.

Dunning, D., & Hayes, A. F. (1996). Evidence for egocentric compari-son in social judgment. Journal of Personality and Social Psychol-ogy, 71, 213–229.

Dunning, D., & Sherman, D. A. (1997). Stereotypes and tacit inference.Journal of Personality and Social Psychology, 73, 459–471.

Fazio, R. H. (1990). A practical guide to the use of response latency insocial psychological research. In C. Hendrick & M. S. Clark (Eds.),


Research methods in personality and social psychology (pp. 74–97).Newbury Park, CA: Sage.

Fazio, R. H., Jackson, J. R., Dunton, B. C., & Williams, C. J. (1995).Variability in automatic activation as an unobtrusive measure ofracial attitudes: A bona Wde pipeline? Journal of Personality andSocial Psychology, 69, 1013–1027.

Festinger, L. (1954). A theory of social comparison processes. HumanRelations, 7, 117–140.

Gawronski, B., Geschke, D., & Banse, R. (2003). Implicit bias inimpression formation: Associations inXuence the construal of indi-viduating information. European Journal of Social Psychology, 33,573–589.

Greenwald, A. G., McGhee, D. E., & Schwartz, J. K. L. (1998). Measuringindividual diVerences in implicit cognition: The Implicit AssociationTest. Journal of Personality and Social Psychology, 74, 1464–1480.

Haslam, S. A., Oakes, P. J., Turner, J. C., & McGarty, C. (1995). Socialcategorization and group homogeneity: Changes is the perceivedapplicability of stereotype content as a function of comparativecontext and trait favorableness. British Journal of Social Psychol-ogy, 34, 139–160.

Henderson-King, E. I., & Nisbett, R. E. (1996). Anti-black prejudice asa function of exposure to negative behavior of a single black per-son. Journal of Personality and Social Psychology, 71, 654–664.

Higgins, E. T., & Lurie, L. (1983). Context, categorization, and mem-ory: The “change-of-standard” eVect. Cognitive Psychology, 15,525–547.

Holyoak, K. J., & Gordon, P. C. (1983). Social reference points. Journalof Personality and Social Psychology, 44, 881–887.

Hugenberg, K., & Bodenhausen, G. V. (2003). Facing prejudice:Implicit prejudice and the perception of facial threat. PsychologicalScience, 14, 640–643.

Kahneman, D., & Miller, D. T. (1986). Norm theory: Comparing real-ity to its alternatives. Psychological Review, 93, 136–153.

Karylowski, J. J. (1990). Social reference points and accessibility oftrait-related information in self-other similarity judgments. Journalof Personality and Social Psychology, 58, 975–983.

Karylowski, J. J., Konarzewski, K., & Motes, M. A. (2000). Recruit-ment of exemplars as reference points in social judgments. Journalof Experimental Social Psychology, 36, 275–303.

Krueger, J., & Clement, R. W. (1994). Memory-based judgments aboutmultiple categories: A revision and extension of Tajfel’s accentua-tion theory. Journal of Personality and Social Psychology, 67, 35–47.

Krueger, J., Rothbart, M., & Sriram, N. (1989). Category learning andchange: DiVerences in sensitivity to information that enhances orreduces intercategory distinctions. Journal of Personality and SocialPsychology, 56, 866–875.

Lord, C. G., & Lepper, M. R. (1999). Attitude representation theory. InM. Zanna (Ed.), Advances in experimental social psychology (Vol.31, pp. 265–343). San Diego, CA: Academic Press.

Lowery, B. S., Hardin, C. D., & Sinclair, S. (2001). Social inXuenceeVects on automatic racial prejudice. Journal of Personality andSocial Psychology, 81, 842–855.

Marques, J. M., Yzerbyt, V. Y., & Leyens, J. P. (1988). The “BlackSheep EVect”: Extremity of judgments towards ingroup membersas a function of group identiWcation. European Journal of SocialPsychology, 18, 1–16.

Mummendey, A., & Wenzel, M. (1999). Social discrimination and tol-erance in intergroup relations: Reactions to intergroup diVerence.Personality and Social Psychology Review, 3, 158–174.

Mussweiler, T. (2003). Comparison processes in social judgment: Mech-anisms and consequences. Psychological Review, 110, 472–489.

Mussweiler, T., & Bodenhausen, G. V. (2002). I know you are, but whatam I. Self-evaluative consequences of judging ingroup and out-group members. Journal of Personality and Social Psychology, 82,19–32.

Mussweiler, T., & Strack, F. (1999). Hypothesis-consistent testing andsemantic priming in the anchoring paradigm: A selective accessi-bility model. Journal of Experimental Social Psychology, 35, 136–164.

Pettigrew, T. F., & Meertens, R. W. (1995). Subtle and blatant prejudicein Western Europe. European Journal of Social Psychology, 25, 57–75.

Sagar, H. A., & SchoWeld, J. W. (1980). Racial and behavioral cues inblack and white children’s perceptions of ambiguously aggressiveacts. Journal of Personality and Social Psychology, 39, 590–598.

Schwarz, N., Bless, H., Strack, F., Klumpp, G., Rittenauer-Schatka, H.,& Simons, A. (1991). Ease of retrieval as information: Another lookat the availability heuristic. Journal of Personality and Social Psy-chology, 61, 195–202.

Schwarz, N., Bless, H., Wänke, M., & Winkielman, P. (2003). Accessi-bility revisited. In G. V. Bodenhausen & A. J. Lambert (Eds.), Foun-dations of social cognition. Mahwah, NJ: Erlbaum.

Sia, T. L., Lord, C. G., Blessum, K. A., Thomas, J. C., & Lepper, M. R.(1999). Activation of exemplars in the process of assessing socialcategory attitudes. Journal of Personality and Social Psychology, 76,517–532.

Smith, E. R., & Zárate, M. A. (1992). Exemplar-based model of socialjudgment. Psychological Review, 99, 3–21.

Snyder, M. L., & Frankel, A. (1976). Observer bias: A stringent test ofbehavior engulWng the Weld. Journal of Personality and Social Psy-chology, 34, 857–864.

Srull, T. K., & Gaelik, L. (1983). General principles and individualdiVerences in the self as a habitual reference point: An examinationof self-other judgments of similarity. Social Cognition, 2, 108–121.

Strack, F., & Mussweiler, T. (1997). Explaining the enigmatic anchor-ing eVect: Mechanisms of selective accessibility. Journal of Person-ality and Social Psychology, 73, 437–446.

Tajfel, H., & Turner, J. C. (1986). The social identity theory of inter-group behavior. In S. Worchel & W. G. Austin (Eds.), Psychology ofintergroup relations (pp. 7–24). Chicago, IL: Nelson-Hall.

Tajfel, H., & Wilkes, A. L. (1963). ClassiWcation and quantitative judg-ment. British Journal of Social Psychology, 54, 101–114.

Tormala, Z. L., Petty, R. E., & Briñol, P. (2002). Ease of retrieval eVectsin persuasion: A self-validation analysis. Personality and SocialPsychology Bulletin, 28, 1700–1712.

Trope, Y., Cohen, O., & AlWeri, T. (1991). Behavior identiWcation as amediator of dispositional inference. Journal of Personality andSocial Psychology, 61, 873–883.

Trope, Y., Cohen, O., & Maoz, Y. (1988). The perceptual and inferen-tial eVects of situational inducements on dispositional attribution.Journal of Personality and Social Psychology, 55, 165–177.

Turner, J. C. (1987). A self-categorization theory. In J. C. Turner, M.Hogg, P. J. Oakes, S. D. Reicher, & M. S. Wetherell (Eds.), Rediscov-ering the social group: A self-categorization theory (pp. 42–67).Oxford, UK: Basil Blackwell.

Tversky, A., & Kahneman, D. (1973). Availability: A heuristic for judg-ing frequency and probability. Cognitive Psychology, 5, 207–232.

Wänke, M., & Bless, H. (2000). The eVects of subjective ease of retrievalon attitudinal judgments: The moderating role of processing moti-vation. In H. Bless & J. P. Forgas (Eds.), The message within: Therole of subjective experience in social cognition and behavior (pp.143–161). Philadelphia, PA: Taylor & Francis.

Wilder, D. A., Simon, A. F., & Faith, M. (1996). Enhancing theimpact of counterstereotypic information: Dispositional attribu-tions for deviance. Journal of Personality and Social Psychology,71, 276–287.

Wittenbrink, B., Judd, C. M., & Park, B. (1997). Evidence for racialprejudice at the implicit level and its relationships with question-naire measures. Journal of Personality and Social Psychology, 72,262–274.

Documents

We are, therefore they arenâ€™t: Ingroup construal as a standard