Group Processes Intergroup Relations-2006-Fielding-199-218.pdf

7/27/2019 Group Processes Intergroup Relations-2006-Fielding-199-218.pdf

http://slidepdf.com/reader/full/group-processes-intergroup-relations-2006-fielding-199-218pdf 1/21

http://gpi.sagepub.com/ Relations

Group Processes & Intergroup

http://gpi.sagepub.com/content/9/2/199The online version of this article can be found at:

DOI: 10.1177/1368430206062077

2006 9: 199Group Processes Intergroup Relations Kelly S. Fielding, Michael A. Hogg and Nicole Annandale

Reactions to Positive Deviance: Social Identity and Attribution Dimensions

Published by:

http://www.sagepublications.com

can be found at:Group Processes & Intergroup Relations Additional services and information for

http://gpi.sagepub.com/cgi/alertsEmail Alerts:

http://gpi.sagepub.com/subscriptionsSubscriptions:

http://www.sagepub.com/journalsReprints.navReprints:

http://www.sagepub.com/journalsPermissions.navPermissions:

http://gpi.sagepub.com/content/9/2/199.refs.htmlCitations:

What is This? - Apr 4, 2006Version of Record>>

at University of Texas Libraries on October 14, 2013gpi.sagepub.comDownloaded from at University of Texas Libraries on October 14, 2013gpi.sagepub.comDownloaded from at University of Texas Libraries on October 14, 2013gpi.sagepub.comDownloaded from at University of Texas Libraries on October 14, 2013gpi.sagepub.comDownloaded from at University of Texas Libraries on October 14, 2013gpi.sagepub.comDownloaded from at University of Texas Libraries on October 14, 2013gpi.sagepub.comDownloaded from at University of Texas Libraries on October 14, 2013gpi.sagepub.comDownloaded from at University of Texas Libraries on October 14, 2013gpi.sagepub.comDownloaded from at University of Texas Libraries on October 14, 2013gpi.sagepub.comDownloaded from at University of Texas Libraries on October 14, 2013gpi.sagepub.comDownloaded from at University of Texas Libraries on October 14, 2013gpi.sagepub.comDownloaded from at University of Texas Libraries on October 14, 2013gpi.sagepub.comDownloaded from at University of Texas Libraries on October 14, 2013gpi.sagepub.comDownloaded from at University of Texas Libraries on October 14, 2013gpi.sagepub.comDownloaded from at University of Texas Libraries on October 14, 2013gpi.sagepub.comDownloaded from at University of Texas Libraries on October 14, 2013gpi.sagepub.comDownloaded from at University of Texas Libraries on October 14, 2013gpi.sagepub.comDownloaded from at University of Texas Libraries on October 14, 2013gpi.sagepub.comDownloaded from at University of Texas Libraries on October 14, 2013gpi.sagepub.comDownloaded from at University of Texas Libraries on October 14, 2013gpi.sagepub.comDownloaded from at University of Texas Libraries on October 14, 2013gpi.sagepub.comDownloaded from

http://gpi.sagepub.com/


http://gpi.sagepub.com/content/9/2/199


http://www.sagepublications.com/


http://gpi.sagepub.com/cgi/alerts


http://gpi.sagepub.com/subscriptions


http://www.sagepub.com/journalsReprints.nav


http://www.sagepub.com/journalsPermissions.nav


http://gpi.sagepub.com/content/9/2/199.refs.html

http://online.sagepub.com/site/sphelp/vorhelp.xhtml


http://gpi.sagepub.com/content/9/2/199.full.pdf

































































http://gpi.sagepub.com/content/9/2/199.full.pdf

http://gpi.sagepub.com/content/9/2/199.refs.html










Group Processes & Intergroup Relations

2006 Vol 9(2) 199–218

GPI

R Reactions to PositiveDeviance: Social Identity and AttributionDimensions

Kelly S. Fielding, Michael A. Hogg and Nicole AnnandaleThe University of Queensland

This research examines whether evaluations of positive deviates (i.e. high achieving groupmembers) are influenced by the attributions they make for their performance. We argue that ingroup positive deviates who make group attributions help enhance the ingroup’s image andthus attract favorable evaluations. In Experiment 1, ingroup positive deviates who made groupattributions were generally evaluated more favorably than ingroup positive deviates who madeindividual attributions. There was also evidence that the positive deviates’ attribution style

influenced group and self-evaluations. Evaluations of outgroup positive deviates were not influenced by their attribution style. In Experiment 2, an ingroup positive deviate who wassuccessful and attributed that success to the group was upgraded relative to an ingroup positivedeviate who made individual attributions. Group evaluations were also higher when the positivedeviate made group attributions. This pattern did not emerge when the positive deviate failed.The results are discussed from a social identity perspective.

keywords attribution, positive deviance, social identity

W HAT do we know about reactions to individuals who are exceptional or outstanding onan important dimension—individuals whoexceed our expectations? On first reflection theanswer to this question might seem self-evident:exceptional or outstanding behavior will attract respect, admiration, and generally favorableperceptions. However, findings from the socialcomparison literature suggest a more complexstory. Comparing oneself with high achievingindividuals can be a threatening experienceunless they are friends who are achieving on aself-irrelevant dimension (e.g. Tesser, 1988),the achievement is readily attainable (e.g.

Lockwood & Kunda, 1997), and crucially, thehigh achievement is by ingroup members in acontextually salient intergroup context (Brewer& Weber, 1994; Schmitt, Silvia, & Branscombe,2000).

Consistent with the social identity perspective(Hogg, 2003; Hogg & Abrams, 1988; Tajfel &

Copyright © 2006 SAGE Publications(London, Thousand Oaks, CA and New Delhi)9:2; 199–218; DOI: 10.1177/1368430206062077

Author’s note

Address correspondence to Kelly Fielding,School of Social Work and Applied HumanSciences, The University of Queensland,Brisbane, QLD 4072, Australia[email: [email protected]]



Turner, 1979; Turner, 1982; Turner, Hogg,Oakes, Reicher, & Wetherell, 1987) this latterfinding points to the importance of identity motivations in our judgments of, and reactions

to, high achieving individuals. Being outper-formed by others can be threatening to our self-concept unless we can somehow benefit froman association with the high achiever. When asocial identity becomes salient, individualsself-categorize as group members, and self-evaluations become dependent on group evalu-ations. In this context high achievement by other ingroup members contributes to a posi-tively distinctive group identity and thus satis-fies self-enhancement motives. Put simply, high

achieving ingroup members make the grouplook good and this in turn allows ingroupmembers to feel good about themselves—they can bask in the reflected glory of the highachieving ingroup members (cf. Cialdini et al.,1976; Cialdini & De Nicholas, 1989; Cialdini &Richardson, 1980; Wann & Branscombe, 1990).

This analysis accords with models that outlineprocesses of reactions to deviates (e.g. subjec-tive group dynamics; Marques, Abrams, Paez, &Martinez-Taboada, 1998; Marques, Abrams, &Serôdio, 2001). Exceptional, high achievingindividuals are extreme group members whoexceed the normal or average level of perform-ance in a group; they are what we term in ourresearch positive deviates . Research within thesubjective group dynamics tradition has shownthat pro-norm deviates, that is, group members

who are extreme but in a direction that supports group norms, are generally evaluatedas favorably as normative group members andmore favorably than anti-normative deviates(Abrams, Marques, Bown, & Dougill, 2002;

Abrams, Marques, Bown, & Henson, 2000). Insubjective group dynamics research, pro-normdeviates were extreme in terms of their attitudi-nal position or group-related personality traitsrather than their achievements or perform-ance. Nevertheless, the research suggests that deviance that validates ingroup norms and con-tributes to a positively distinctive group image

will be met with favorable reactions.Similarly, Hogg, Fielding, and Darley (2005)

argue that when social identity is salient positive

deviates help to satisfy self-enhancement motives and are therefore generally evaluatedpositively. This analysis focuses on the func-tional aspect of deviance; evaluations of

deviates are determined by the direction of their deviance and the extent to which thedeviant behavior helps to satisfy the predomi-nant group motivation.

Group life and group contexts are, however,more complex than this. Group membersmight not only take note of the direction of deviation but also the way in which deviatesaccount for their behavior. Attributions might therefore be important in determining evalu-ations of positive deviates. High achieving

scholars or successful sportspeople might beevaluated differently depending on the way they explain their success. In group contextsthe most important aspect of a positive deviate’sexplanation might be whether it allows thegroup to legitimately claim ownership of thepositive deviate’s success. If the group’s role inthe success is acknowledged, then the success of the positive deviate is more easily reflected ontothe group and its members. Group members

who make group attributions for their successshould therefore be evaluated favorably.

If the success is personalized, however, andthe role of the group is denied, ignored, ordownplayed, positive deviates and their achieve-ments are distanced from the group. In thissituation it may be much more difficult forgroup members to bask in the reflected glory of the positive deviate. Hence a positive deviate

who makes individualized attributions for his orher success may be met with relatively morenegative reactions.

Previous research has investigated the role of attribution processes in group contexts. Forexample, researchers have investigated how group members react to the attributions othersmake for group success or failure. Groupmembers who distance themselves from thegroup by taking little responsibility for groupoutcomes, are evaluated less favorably thangroup members who allocate responsibility equally to group members (Forsyth & Mitchell,1979). Similarly, group members evaluateothers who make self-serving attributions (i.e.

Group Processes & Intergroup Relations 9(2)

200



blaming others for group failure and claimingcredit for group success) more negatively thanthose who make equitable or other-servingattributions (Forsyth, Berger, & Mitchell, 1981).

A study by Bond, Chiu, and Wan (1984) foundthat participants preferred confederate partners

who make group-enhancing attributions ratherthan group-effacing attributions for groupsuccess or failure. In a study of ice hockey teams, Taylor and Tyler (1986) found that members react more favorably toward otherplayers who show group-serving biases (i.e. takeresponsibility for group failure and shareresponsibility for group success) than self-serving biases in their attributions. Taylor and

Tyler argue not only that the attribution style of group members is readily recognized by others,but also that the attributions group membersmake for group success and failure have animpact on person perception.

This research suggests that, in group contexts,social identity motives may play a role in thetype of attributions group members make, andmay influence how these attributions are per-ceived by others. Taylor, Doria, and Tyler(1983) argue that a group-serving bias in attri-butions helps to promote ingroup cohesionbecause group members are involved in actively promoting the group’s image. Although thisresearch supports the argument that attribu-tions for success or failure are important forevaluations of the group and its members, it does not examine how the attributions groupmembers make for their own individual per-formance influences evaluations of them. Fur-thermore, it has focused primarily on smallinteractive group contexts rather than largesocial categories where members are only sub-

jectively interdependent.

The present research

The aim of the current research is to test whether the attributions that positive deviatesmake for their high performance influencesthe responses of other group members. Most research on reactions to high achievement hasbeen conducted in the social comparison tra-dition where evaluations of the self after

exposure to high achievers, rather than evalu-ations of the high achievers themselves, areassessed (see Collins, 1996, for a review).Research that has examined evaluations of high

achievers has for the most part not explicitly tested the importance of social categorizationfor evaluations (e.g. Feather, 1989, 1994). Anexception is research by Schmitt et al. (2000)that has shown that social identity concerns play an important part in how group members react to high achieving others. They found that, in anintergroup context highly identified groupmembers preferred a high to a low performingingroup member and that they were morepositive to high performing ingroup members

than low identified group members. The contri-bution of our research is to go beyond groupmembers’ reactions to achievement, to look at how actions and behaviors that accompany theachievement can influence reactions. In sodoing, we recognize that, in reality, our percep-tions are often influenced by multiple factors.

Our general prediction is that ingrouppositive deviates who make group attributionsfor their success will be evaluated more posi-tively than those who make individual attribu-tions. Moreover, we predict that because groupattributions allow the group and its members tobenefit more easily from an association with theingroup positive deviate, group and self-evalu-ations should also be more positive wheningroup positive deviates make group attribu-tions rather than individual attributions fortheir success.

Experiment 1

The goal of the first study was to test our pre-diction that the type of attributions that ingroup positive deviates make for their success

will influence evaluations of the deviates. Inaddition, we sought, by including measures of group and self-evaluations, to test the argument that ingroup positive deviates who make groupattributions contribute to a positively distinctivesocial identity.

Australian university students were presented with two positive deviates from the same university (ingroup) or from a different university

Fielding et al. positive deviance

201



(outgroup). The inclusion of an outgroup con-dition allowed us to see whether reactions tothe positive deviates were based on identity concerns, as we propose, or more broadly on

expectations generated by wider social norms.The targets were positive deviates by virtue of exceptional academic or sporting achievement.

We assumed that high achievement on both of these domains would be important to groupidentity; academic achievement because it iscentral to a university identity, and sportingachievement because it is normative andimportant in the Australian context (Feather,

Volkmer & McKee, 1991; Haslam, Oakes,Turner, & McGarty, 1995). Including two types

of positive deviates also allowed us to see whether reactions generalized across domainsor whether the centrality of a particular domain(e.g. academic achievement) might influenceresponses. We varied the type of attributionsmade by the positive deviates by having themattribute their success solely to individualfactors or by having them also give credit totheir university. Participants then evaluatedboth of the positive deviates and also evaluatedthe group and themselves.

We predicted that ingroup positive deviates who made group attributions for their success would be evaluated more favorably than those who made individual attributions. What theeffect of the attribution style of the outgrouppositive deviate would be was less clear. In ahighly competitive and conflictual intergroupcontext, outgroup positive deviates who gener-alize their success to the group may be evalu-ated negatively because of the threat they poseto the ingroup’s position. In contrast, whenrelations between groups are relatively stableand the ingroup is assured of its status, theattributions made by outgroup positivedeviates might have little consequence foringroup members and might not affect evalu-ations of the outgroup positive deviates. Thelatter context describes more closely the situ-ation between the ingroup and the outgroupin the present study and so we predicted that evaluations of outgroup positive deviates

would not be influenced by their attributionstyle.

We also predicted that when ingroup positivedeviates made group attributions for theirsuccess, group and self-evaluations would behigher than when the positive deviates made

individual attributions. We included both direct and indirect measures of group perceptions.The indirect measures included perceptions of group homogeneity and ratings of the import-ance of sporting and academic achievement.

Varying perceptions of group homogeneity isone strategy that can be used to maintain a pos-itively distinctive group identity (e.g. Doosje,Spears, & Koomen, 1995; Pickett & Brewer,2001). If, as we argue, group attributions helpto reflect the positive deviates’ success onto the

group, this success can be embraced by empha-sizing the similarities among group members.By increasing perceptions of group homogene-ity group members are saying ‘we’re all alikeand different from you’. Moreover, wheningroup members achieve highly in a particulardomain and reflect that achievement onto thegroup, group members may also seek toembrace the achievement by emphasizing theimportance of that type of achievement, astrategy akin to the social creativity strategiesproposed by social identity theory (e.g. Hinkle,Taylor, Fox-Cardamone, & Ely, 1998; Jackson,Sullivan, Harnish, & Hodge, 1996).

Put simply, the logic for the above hypothesesis that high achieving ingroup members will be

judged more favorably when they help to makethe group look good. If this is the case, then theeffect of attribution style on target evaluationsshould be mediated by group-based evaluations.

Method

Participants and design Participants were 62introductory psychology students at the Uni-

versity of Queensland, who participated inexchange for course credit. There were 34females and 28 males (mean age = 19.08 years,SD = 3.11). Participants were randomly assignedto conditions and tested in one-hour sessionsthat ranged in size from one to nine partici-pants. The experiment involved a 2 (Attribu-tion Style: individual, group) 2 (Targets’Group Membership: ingroup, outgroup) 2(Achievement Dimension: sporting, academic)


202



mixed design with repeated measures on thelast factor.

Procedure Participants volunteered for a

study of the effect of media style on social per-ception. They were told that we were compar-ing perceptions of print versus online sourcesof university media information, and that they

were in the condition assessing print media.They were told that they would read a numberof print articles, and after reading each theirperceptions of the articles would be assessed.

Where the articles presented were about specific people, they were told that the namesand some details had been changed to elimi-

nate the possibility that their perceptions might be affected by previous knowledge of theperson.

The first article presented to participantsserved to make participants’ identity as Uni-

versity of Queensland students salient. Thearticle was constructed to set up competitionbetween the University of Queensland (UQ;the ingroup) and the Queensland University of Technology (QUT; the outgroup). QUT waschosen as the outgroup based on responses toa pilot study showing that it was perceived asthe main competitor for UQ. The articleexplicitly acknowledged the competitionbetween the two universities, both located inthe same city, in terms of their rivalry in attract-ing Australian and overseas students. Afterreading the article, participants responded to aquestionnaire assessing their perceptions of

journalistic style, competition between UQ andQUT, and their identification with UQ. Thesequestions served to emphasize further theintergroup context. Ancillary steps were alsotaken to increase the salience of the intergroupcontext; for example, questionnaires bore theUQ logo, and the target evaluation question-naires asked participants to record the uni-

versity affiliation of the target at the end of thequestionnaire.

Participants were then presented with twoprint articles that described a very high achiev-ing student receiving a prize for his or herperformance: a prize for academic achievement or for sporting achievement. The academic

achievement article described a student who was presented with the ‘Australian Tertiary Students’ Prize’. This prize was ostensibly awarded on the basis of leadership qualities and

academic excellence. The sporting achieve-ment article described a student who had wonthe 100 metres backstroke race at the Aus-tralian University Games. In addition, his or hertime had set a new games record.

The university affiliation of the positivedeviate was used to manipulate the targets’group membership: they were either from theUniversity of Queensland (ingroup target) orfrom the Queensland University of Technology (outgroup target). The attribution style of the

target was also manipulated. In the individualattribution condition the targets alwaysdescribed their achievement using the pronoun‘I’. The targets described how they had workedhard to succeed and they claimed the achieve-ment as a personal victory (swimming article)or that the award gave them great personalsatisfaction (academic article).

In the group attribution condition, thetargets acknowledged that they worked hardand felt a sense of personal achievement (forthe sake of credibility), but they also attributedtheir success to their university. In the academicarticle the target acknowledged that the prizereflected on his or her university. They went onto say that the working and social environment of the university made it possible. In theswimming article the targets acknowledged that they felt proud not just for themselves but alsofor their university. They said that the support received from their university had helped themachieve their goal. The article always describeda target who was the same sex as the participant (Michelle/Michael Reid for the academictarget and Natalie/Nathan Jones for thesporting target).

After reading each article (presentation of the two articles was counterbalanced) partici-pants completed a questionnaire assessing theirevaluations of the target. A final questionnaireassessed group perceptions, self-evaluations,and checks on the manipulations. Participants

were then debriefed and thanked for theirparticipation.


203



Measures The first questionnaire assessedparticipants’ impression of the journalistic styleof the article (included to maintain the credi-bility of the cover story), how much competi-

tion they perceived between UQ and QUT, andgroup identification. The perceived competi-tion item was included as a background checkto ensure that QUT was seen as a relevant outgroup for UQ. Participants rated the compe-tition between UQ and QUT as relatively high(M = 7.03 on a 9-point scale where 1 indicatednot very much competition and 9 indicated

very much competition). Moreover, the mean was significantly higher than the midpoint of the scale (t (61) = 10.79, p < .001). This confirms

that QUT was a relevant outgroup for UQ.Eight group identification items adapted fromHogg and Hains (1996) were used to primegroup identity. Participants were asked, forexample, how much they identified as a UQ student. Overall, identification as a UQ student

was relatively high (M = 6.30). However, level of identification was not of theoretical interest inthis study and cell sizes did not allow us tomeaningfully examine any effects of initialidentification on the dependent measures.

Again, for the sake of credibility, the target evaluation questionnaire asked participantstheir impression of the journalistic style of thearticle. Participants were then asked how wellthe target had performed, how much they respected the target, how pleased they felt about the target’s academic/sporting perform-ance, and how typical the target might be of students at his/her university. These items wereused to tap aspects that theoretically andempirically relate to attraction to individuals asgroup members (e.g. Feather, 1993, 1994;Fielding & Hogg, 1997; Hogg, 1993, 2001).Previous research has shown that group-basedliking is related to the prototypicality of groupmembers and the contribution they can maketo the group (e.g. Hogg, Cooper-Shaw, &Holzworth, 1993; Hogg & Hardie, 1991;Schmidt et al., 2000). Moreover, respect hasbeen shown to relate to people’s instrumentaland expressive contribution to group goals (e.g.Segal, 1979). The average of the four items wastherefore used as a measure of the targets’

group member attractiveness and had reasonablereliability (academic target = .66, sportingtarget = .69). The typicality item was also usedas a single item to check whether there were

any differences in the perceived typicality of theacademic and sporting positive deviate. Partici-pants then rated the target on five semantic dif-ferential scales measuring how rational, sincere,cooperative, honest, and modest the target was.These items were similar to those used inprevious research assessing the personal charac-ter of high achieving others (Feather, 1989).1

The five items were averaged to form a compos-ite index of the targets’ personal character (academic target = .86, sporting target =

.88). Responses to all items were made on9-point scales (e.g. 1 irrational , 9 rational , 1insincere , 9 sincere ). Finally, to check that partici-pants attended to the group membership of thetarget, they were asked to record the target’suniversity affiliation.

After the target evaluations, participantscompleted a final questionnaire. To maintainthe credibility of the cover story, they wereasked three questions about their use of differ-ent media types. Perceptions of ingroup andoutgroup homogeneity were then assessed withan item asking how similar UQ/QUT studentsare to each other. Evaluations of the ingroupand the outgroup were measured by askingparticipants their global impression of UQ andQUT students, and asking them to rate theeducational and social environment of UQand QUT. The mean of the three UQ evalu-ation items was used as an ingroup evaluation scale = .69) and the mean of three QUTevaluation items was used an outgroup evaluation scale = .74).

Achievement dimension importance wasassessed by asking participants how important they thought academic and sporting achieve-ment was. Participants were then asked to ratetheir own academic and sporting ability. Thetwo self-evaluation items did not correlatehighly (r = .32) and were therefore retained asseparate items. Finally, the manipulation of thetarget’s attribution style was checked. Partici-pants were asked how much each of the targetstook personal credit for their success and how


204



much each of the targets attributed theirsuccess to their university. Responses to allgroup and self-evaluation items were made on9-point scales (e.g. 1 not very much , 9 very much ).

Results

Preliminary analyses including participant sexand target presentation order showed that these variables did not qualify any of the pre-dicted interactions. For this reason these vari-ables are not included in the analyses reportedbelow. Means and standard deviations for thetarget evaluation measures are presented inTable 1.

Manipulation and background checks On theitem checking on the typicality of the twotargets, participants judged both the academictarget (M = 4.88) (t (61) = 0.57, ns ), and thesporting target (M = 4.22) (t (61) = 3.72, p <.001), as relatively atypical group members,although only the rating for the sporting target differed significantly from the midpoint of thescale. A 2 (Achievement Dimension) 2(Target’s Group Membership) 2 (Target’s

Attribution Style) mixed analysis of variance(ANOVA) on the typicality item revealed amain effect for achievement dimension ( F (1,55) = 10.09, p = .002). The sporting target was

judged to be significantly less typical than theacademic target.

Target’s group membership In all cases theparticipants correctly identified the university affiliation of the target.

Attribution style As expected, participants ratedthe positive deviates in the individual attribu-tion condition as taking more personal credit for their success than in the group attributioncondition (M s = 8.01 and 6.82 respectively)( F (1, 57) = 20.78, p = .000). Participants alsorated the positive deviates in the group attribu-tion condition as giving more credit to thegroup for their success than in the individualattribution condition (M s = 6.74 and 3.68respectively) ( F (1, 57) = 64.25, p = .000).2 Thus,the manipulation of the target’s attributionstyle was successful. T a

b l e 1 . E x p e r i m e n t 1 : M e a n s a n d s t a

n d a r d d e v i a t i o n s ( i n p a r e n t h e s e s )

f o r t a r g e t e v a l u a t i o n s c a l e s a s a f u n c t i o n o f a c h i e v e m e n t d i m e n s i o n ,

t a r g e t ’ s

g r o

u p m e m b e r s h i p , a n d a t t r i b u t i o n s t y l e

A c a d e m i c t a r g e

t

S p o r t i n g t a r g e t

I n g r o u p

O u t g r o u p

I n g r o u p

O u t g r o u p

I n d i v i d u a l

G r o u p

I n d i v i d u a l

G r o u p

I n d i v i d u

a l

G r o u p

I n d i v i d u a l

G r o u p

a t t r i b u t i o n


a t t r i b

u t i o n






P e r s o n a l c h a r a c t e r

6 . 4 0

7 . 4 9

6 . 7 0

6 . 6 6

6 . 3 5

7 . 0 9

6 . 4 7

5 . 9 9

( 1 . 5 5 )

( 1 . 0 3 )

( 1 . 0 1 )

( 0 . 9 3 )

( 1 . 6 7 )

( 1 . 1 9 )

( 0 . 9 9 )

( 0 . 9 3 )

G r o u p m e m b e r a t t r a c t i v e n e s s

6 . 9 5

7 . 0 0

6 . 7 5

6 . 8 3

6 . 1 4

6 . 8 8

6 . 5 2

6 . 4 1

( 1 . 0 1 )

( 1 . 1 7 )

( 1 . 3 2 )

( 0 . 8 3 )

( 1 . 4 6 )

( 1 . 0 3 )

( 1 . 0 6 )

( 0 . 8 6 )


205



Target evaluations

Group member attractiveness There was a signifi-cant main effect for achievement dimension( F (1, 58) = 13.09, p = .001), such that the

academic target was evaluated more favorably than the sporting target (M s = 6.88 and 6.48respectively). The Target’s Group Membership Attribution Style interaction was not significant ( F < 1), however, the Achievement Dimension Target’s Group Membership

Attribution Style interaction was significant ( F (1, 58) = 4.08, p = .048). Inspection of meansshows that the academic target’s attributionstyle did not affect evaluations of the ingroup orthe outgroup target ( F s < 1). There was a

tendency, however, for the ingroup sportingtarget to be rated as a more attractive groupmember when he or she made group ratherthan individual attributions ( F (1, 58) = 3.34,

p = .073). There was no effect of attributionstyle for the outgroup sporting target ( F < 1).

Personal character The predicted AttributionStyle Targets’ Group Membership interactionemerged on the personal character measure( F (1, 55) = 4.65, p = .035). As shown in Figure

1, ingroup positive deviates who made groupattributions were evaluated more favorably thaningroup positive deviates who made individualattributions ( F (1, 56) = 5.66, p = .021). There

was no difference in evaluations of theoutgroup targets ( F < 1). There was also a maineffect for achievement dimension ( F (1, 55) =4.72, p = .034): academic targets were evaluatedmore favorably than sporting targets (M s = 6.81and 6.48 respectively).

Group perceptions

Group evaluation The means and standarddeviations for group and self-evaluationmeasures are presented in Table 2. A Group(UQ, QUT) Target’s Group Membership

Attribution Style ANOVA with repeated


206

Figure 1. Target’s personal character as a function of target’s group membership and attribution style.

P e

r s o n a l c h a r a c t e r

Ingroup target Outgroup target

Target’s group membership

Individual attribution

Group attribution

7.4

7.2

7

6.8

6.6

6.4

6.2

6

5.8



measures on the first factor was conducted onthe group evaluation measures. The predictedGroup Targets’ Group Membership Attri-bution Style interaction was not significant( F < 1) and the only significant effect to emerge

was a main effect for group: the ingroup wasrated more favorably than the outgroup (M s =7.09 and 6.22) ( F (1, 58) = 21.26, p = .000).

Perceived group homogeneity The same analysison the perceived group homogeneity itemsrevealed a marginal main effect for attributionstyle: perceptions of group homogeneity werehigher when targets made group rather thanindividual attributions (M s = 5.41 and 4.66)( F (1, 58) = 3.98, p = .051). The predictedTargets’ Group Membership AttributionStyle interaction was marginally significant ( F (1, 58) = 3.09, p = .084). Perceptions of grouphomogeneity were higher after exposure toingroup targets who made group rather than

individual attributions ( F (1, 59) = 7.19, p = .01), whereas the attribution style of the outgrouptarget had no effect on perceptions of grouphomogeneity ( F < 1).

Achievement dimension importance An Achieve-ment Dimension Target’s Group Member-ship Attribution Style ANOVA with repeatedmeasures on the first factor was conducted onthe item checking the importance of sport andacademic achievement. The predicted Target’sGroup Membership Attribution Style inter-action was not significant ( F (1, 58) = 1.74, ns ).There was a main effect for achievement dimension ( F (1, 58) = 36.90, p < .001):academic achievement was considered moreimportant than sporting achievement (M s =7.19 and 5.87). This main effect was qualifiedby an Achievement Dimension Targets’Group Membership Attribution Style inter-action ( F (1, 58) = 10.70, p = .002). Inspection of


207

Table 2. Experiment 1: Means (SD) for group perception and self-evaluation measures as a function of target’s group membership and attribution style

Ingroup target Outgroup target

Individual Group Individual Groupattribution attribution attribution attribution

Group perceptions

Ingroup evaluation 7.17 6.93 7.27 6.98(1.25) (1.32) (1.03) (1.07)

Outgroup evaluation 5.98 6.18 6.29 6.44(1.69) (1.22) (1.56) (0.69)

Ingroup homogeneity 4.00 5.73 5.00 4.94(2.16) (1.34) (1.89) (1.34)

Outgroup homogeneity 4.50 5.60 5.13 5.38(2.13) (1.24) (2.17) (1.50)

Academic importance 6.88 7.07 7.13 7.69(1.86) (1.49) (0.99) (0.95)

Sporting importance 4.81 6.33 6.67 5.75(3.27) (1.29) (1.23) (1.73)

Self-evaluations

Own academic ability 6.23 6.60 6.80 6.63(1.09) (1.12) (1.01) (1.20)

Own sporting ability 4.88 6.33 5.87 5.31(2.03) (1.29) (1.41) (2.75)

Note : Higher values represent more favorable group evaluations, greater perceived group homogeneity,greater dimension importance, and more favorable self-evaluations.



the means shows that academic importance wasunaffected by the attribution style of theingroup or the outgroup target ( F s < 1.4). Theimportance of sporting achievement, however,

was heightened when participants were pre-sented with ingroup targets who made groupattributions relative to ingroup targets whomade individual attributions ( F (1, 59) = 4.24,

p = .044). The attribution style of the outgrouptarget did not significantly affect judgments of sporting importance ( F (1, 59) = 1.44, ns ).

Self-evaluations A Target’s Group Member-ship Attribution Style ANOVA was conductedon the two self-evaluation items. The predicted

two-way interaction did not emerge for self-evaluations of participants’ academic ability and there were no other significant effects onthis measure ( F s < 1). The predicted Target’sGroup Membership Attribution Style inter-action emerged on self-evaluations of partici-pants’ sporting ability ( F (1, 58) = 4.02, p =.050), and there were no other significant effects ( F s < 1). Inspection of the means showsthat participants’ evaluations of their ownsporting ability were higher after exposure toingroup targets who gave group rather thanindividual attributions for their success ( F (1,59) = 4.29, p = .043). The outgroup targets’attribution style did not affect participants’ per-ceptions of their sporting ability ( F < 1).

Mediational analyses Analyses were con-ducted to test whether the effect of the ingrouptarget’s attribution style on target evaluations

was mediated by group evaluations. The resultsso far reveal that attribution style had a clearand significant impact on the ingroup target’spersonal character and sporting achievement importance, thus, they were used as the depen-dent measure and the mediator respectively.Because the impact of attribution style on otherpossible mediators (e.g. perceived group homo-geneity, group evaluation) was nonsignificant or marginally significant, and theoretically, self-evaluations were thought to be too distal to act as a mediator, analyses using these variables asmediators are not presented.3 An overallmeasure of ingroup targets’ personal character

was computed by averaging the evaluations of the academic and sporting target’s personalcharacter. Consistent with Baron and Kenny (1986), full mediation is said to occur when the

proposed mediator is significantly related to thedependent measure and the inclusion of themediator renders the effect of the independent

variable nonsignificant. The Sobel test can beused to assess the significance of the reductionin effect. Sporting achievement importance wassignificantly related to the ingroup target’spersonal character ( = .51, p = .003). Analysesalso showed that the significant effect of attri-bution style on evaluations of the target’spersonal character ( = .38, p = .04) became

nonsignificant ( = .23, p = .16) once sportingachievement importance was included in theanalysis. However, the reduction in the effect of attribution style was not reliable according tothe Sobel test (z = 1.49, p = .13).

Discussion

Consistent with our predictions, ingroup positivedeviates who attributed their success to thegroup were evaluated more favorably than those

who took personal credit for their success. Thisfinding emerged for both the academic andsporting positive deviates on judgments of theirpersonal character and for the sporting positivedeviate on judgments of his or her attractivenessas a group member. In contrast, the attributionstyle of outgroup positive deviates did not influ-ence evaluations of these targets. This suggeststhat evaluations were motivated by social identity concerns rather than general expectations that specify, for example, the need for modesty onthe part of high achievers.

Further evidence that evaluations reflectedsocial identity concerns emerged on the groupand self-evaluation measures. Although overallgroup evaluations simply reflected the usualingroup bias effect (Brewer, 1979; Brewer &Kramer, 1985), on the indirect measures of group perceptions the influence of the ingroupmembers’ attributions was apparent. There wasa tendency for group members to heightenperceptions of group homogeneity wheningroup positive deviates made group attribu-tions. In situations where positive deviates’


208



success reflects on to the group, emphasizingsimilarity among ingroup and outgroupmembers may be a way for group members toenhance the positive distinctiveness of the

ingroup. This finding also accords withprevious research showing that group membersuse group variability judgments as a way of pro-tecting or enhancing ingroup identity (e.g.Doosje et al., 1995; Pickett & Brewer, 2001).

Judgments of the importance of academicand sporting achievement also suggest the useof strategic perceptions to enhance ingroupidentity. Not surprisingly given the salience of university affiliation, academic achievement wasconsidered more important than sporting

achievement. However, group membersupgraded the importance of a peripheraldimension (i.e. sporting achievement), wheningroup positive deviates made group attribu-tions for their success. By reflecting his/hersuccess onto the group the sporting positivedeviate may have provided group members withadditional means to enhance ingroup identity,a strategy reminiscent of other research that demonstrates that group members will differen-tially rate group defining dimensions as a way toprotect or enhance ingroup identity (e.g.Hinkle et al., 1998; Jackson et al., 1996).

Our argument that positive deviates are ratedmore favorably because they reflect their successonto the group gained tentative support in thisstudy. Mediational analyses showed that sportingachievement importance mediated the effect of attribution style on ratings of the personal char-acter of the ingroup targets, although the Sobeltest failed to attain significance. One possibleexplanation for the relatively weak mediatingeffects is the group evaluation measures used inthe study. It is possible that these measures didnot fully assess aspects of the group influencedby the targets’ attributions. A lack of power dueto low numbers may also account for the weakerresult.

The attributions that ingroup positivedeviates’ made for their success also impactedon group members’ self-evaluations. Partici-pants’ evaluations of their academic ability remained unaffected. However, on the dimen-sion of sporting ability group members had

higher evaluations of their own sporting ability after exposure to ingroup positive deviates whomade group attributions rather than individualattributions. Consistent with previous research

these results suggest that ingroup members’self-concept can benefit from an association

with high achieving others (Brewer & Weber,1994). The results indicate that when positivedeviates reflect their success onto the group,group members may be assimilating their ownperformance or ability to that of the positivedeviates. Although these results show that theingroup target’s attribution style impacted ongroup and self-evaluations, it must be acknowl-edged that without a control group we are only

able to establish the relative differences acrossconditions, not the absolute effects of attribu-tion style on group and self-evaluations.

It is interesting to note that self-evaluations were only influenced on a dimension that might be considered peripheral to ingroupidentity, that is, the sporting dimension.Indeed, it is evident that some of the predictedresults only emerged for the sporting target oron the dimension of sporting achievement.One explanation for this might relate to thetypicality of the sporting target. It was clear that the sporting target was judged as less typicalthan the academic target. Group members may have paid more attention to this group memberbecause he or she was unexpected (cf.Burgoon, 1993; Jones & McGillis, 1976; Jussim,Coleman & Lerch, 1987) and also becausesporting achievement is normative and thusimportant for the superordinate category of

Australian identity (Feather et al., 1991; Haslamet al., 1995). Group members might thereforebe keen to emphasize the success of this person,especially when the success can be generalizedto the group as a whole (i.e. when the positivedeviate makes group attributions). Thus, groupmembers judged ingroup sporting targets whoreflected their success onto the group as moreattractive group members and embraced theirachievement by heightening the importance of the sporting achievement dimension. Groupmembers’ evaluations of their own sportingability were also influenced. Another point

worth considering is that the centrality of the


209



academic dimension for the group identity may have stymied the ability of group members tobask in the reflected glory of the academictarget’s achievements (cf. self evaluation model;

Tesser, 1988). That is, academic positivedeviates have the potential to be threateningrather than augmenting to the self-concept.

The first experiment rules out the alternativeexplanation that responses to positive deviatessimply reflect general expectations about modesty or group allegiance. We argue that positive deviates are being evaluated more favor-ably when they make group attributions becausethese types of explanations help to reflect theirsuccess onto the group as a whole. However, it

might also be argued that any ingroup members,regardless of their level of success, who attributetheir performance to the group will beupgraded because group attributions affirmgroup identity and help to promote ingroupcohesion (cf. Taylor & Tyler, 1986). The secondexperiment was designed to rule out this expla-nation by varying the level of achievement of theingroup target. Moreover, we sought to general-ize our findings to a different group context.

Experiment 2

In Experiment 2, Australian participants werepresented with an Australian athlete whocompeted at the Olympic games and who either

won a gold medal or who was eliminated in thefirst round of competition. An Olympic athleteis a positive deviate by virtue of achieving at alevel far above the average group member.Simply having a fellow group member achieveat such a level should have benefit for thegroup and its members. However, it is likely that a positive deviate who attains the glory of a goldmedal at the Olympics has a greater capacity toenhance the group’s image than one who iseliminated from the competition. The positivedeviate described his/her performance eitherin terms of individual qualities or in terms of the group. Thus, the experiment involved a 2(Level of Achievement: success, failure) 2(Attribution Style: individual, group) between-subjects design.

Consistent with our general prediction, apositive deviate who has gained success and

who reflects his/her success onto the group viagroup attributions should attract more positive

evaluations than one who makes individualattributions. As we have argued above, this typeof ingroup member should especially contrib-ute to a positively distinct group identity and sogroup evaluations should also be more positivein this condition. If this logic is correct then theeffect of the successful target’s attribution styleon target evaluations should be mediated by his/her impact on group evaluations.

Predictions regarding the positive deviate who fails are less straightforward. We might

argue that group members who attribute theirfailure to the group will attract negativereactions because of the potential damage they do to the group’s image. However, in thepresent context the target is not truly a failurebut instead a positive deviate who has not achieved as highly as he or she might. For thisreason, group members may pay less attentionto the explanatory style of this person and attri-butions for performance may have little impact on evaluations of the failed positive deviate.

Method

Participants Participants were 80 introductory psychology students at the University of Queensland who participated in exchange forcourse credit. There were 65 females and 15males with a mean age of 20.28 (SD = 5.77).Participants were randomly assigned to one of the four conditions, and were tested in sessionsof between 1 and 10 people.

Procedure The experimental session was pre-sented to participants as two separate studies.Participants were told that they would initially fill out a questionnaire that was part of an inter-national study comparing the importance of sport across different countries, and that theirresponses would contribute to the Australiandata collection. This cover story and the first questionnaire were used as a means of making

Australian identity and the sporting achieve-ment dimension salient.


210



In the ‘second’ study, participants were toldthat they would read an article in which an

Australian scientist, actor, sporting figure, orpolitician was interviewed, and that they would

be asked to evaluate this person on the basis of the information in the article. In fact, all partici-pants received the article about the sportingfigure and the different types of targets wereonly mentioned to minimize the similarity between the first and second part of the study.

All participants were informed that somedetails, such as the name and gender of theathlete, and the date of the article had beenomitted so that they would not recognize theathlete and base their evaluations on precon-

ceived judgments of the athlete. Our aim inremoving this information was to reduce thelikelihood that evaluations of the target wouldbe influenced by shared interpersonal similarity or preconceived judgments of known athletes.The interview was also presented in a font andformat that imitated a genuine media article.

The interview article was presented directly after the introduction. The article began withan introductory paragraph that was followed by an interview in question and answer format. Inthe introductory paragraph the target wasdescribed as an Australian athlete who hadrecently returned from competing at theOlympics. The achievement level of the athlete

was manipulated at this point in the article. Anathlete who had failed was described assomeone who had been eliminated in the first heat of the competition at the Olympics. Anathlete who had succeeded was someone whohad won a gold medal at the Olympics.

Attribution style was manipulated within theinterview itself. This manipulation was presentedin the response to the question, ‘What do youthink contributes to your performance at thislevel of competition?’ Athletes in the individualattribution condition attributed their success totheir own individual qualities. They referred toskill level, hard work, and determination as con-tributing to their performance. Athletes in thegroup attribution condition attributed theirsuccess to the support they receive from theirgroup (i.e. Australia). Their response empha-sized that support from Australian fans and

access to Australian sporting programs had con-tributed to their performance. The remainderof the interview involved questions that probedother aspects of the athlete’s life, for example,

whether they took time off during the year and what activities they pursued during time off. After reading the article, participants com-pleted a questionnaire assessing checks on themanipulations, target evaluations, and groupevaluations. Participants were then debriefedand thanked for their participation.

Measures The first questionnaire includeditems assessing demographics, group identifi-cation, and the importance of sporting achieve-

ment. Participants were asked their sex, age,and nationality. Only participants who identi-fied as Australian citizens were included in thesample. As in Experiment 1 the group identifi-cation items were used to prime group identity and showed that identification as an Australian

was relatively high (M = 6.90). For the sake of the cover story there were five items asking par-ticipants about the importance of sport andsporting achievement for Australians. Anexample item is: ‘In your opinion how import-ant do you think sport is to Australians?’ (e.g. 1not very important , 9 very important ).

After reading the media article, a secondquestionnaire assessed target and group evalu-ations. All items were measured on 9-point scales (e.g. 1 not very much , 9 very much ). Fiveitems assessed the likability of the target, how

well the target represents Australia, how muchthe target should be supported by Australia,how much the target deserves to be successful,and how much the target deserves to be publicly recognized. The last three items were adaptedfrom Feather’s (1989) scale measuring reactionsto high achievers. The five items were averagedto form a measure of group member attractiveness ( = .88). Participants then rated how modest,sincere, honest, humble, giving and fair thetarget was. These items formed a measure of the target’s personal character ( = .80).

Two items assessed group perceptions. Oneitem simply asked participants to rate theiroverall impression of Australians. The other wasa comparative measure that asked participants


211



to rate Australia’s overall image compared toother countries.

Two items checked on the manipulation of achievement level. Participants were asked how

they would rate the athlete’s performance, and what level of achievement the athlete attained.These two items were highly correlated (r = .85)and were averaged to form a single scale. Tocheck on the manipulation of attribution style,participants were asked, ‘In your opinion, how much did the target think their performance wasdue to individual factors?’ and, ‘In your opinion,how much did the target think their perform-ance was due to support from their country?’

Results

Preliminary analyses including participant sexdid not reveal any main effects or interactionsinvolving this factor, therefore it was not included in the focal analyses.

Manipulation and background checks

Achievement level As expected, participantsrated the performance of the positive deviate

who had succeeded more highly than thepositive deviate who had failed (M s = 8.56 and6.06) ( F (1, 76) = 83.60, p = .000). There wereno other significant effects on this check.

Attribution style Participants rated the positivedeviate in the individual attribution condition

as attributing their performance more to individual factors than in the group attribution con-dition (M s = 7.68 and 5.03) ( F (1, 76) = 37.70,

p = .000). Participants also rated the positive

deviate in the group attribution condition asattributing his/her performance more to groupfactors than in the individual attribution con-dition (M s = 6.53 and 2.90) ( F (1, 76) = 100.80,

p = .000). There were no other significant effects on these checks ( F s < 1). Thus themanipulation of achievement level and attribu-tion style were clean and effective.

Target evaluation

Group member attractiveness Means and standard

deviations for target and group evaluations arepresented in Table 3. There were significant main effects for achievement level ( F (1, 76) =4.59, p = .035), and attribution style ( F (1, 76) =4.52, p = .037). Participants rated targets who

were successful as more attractive to the group(M = 7.62) than those that failed (M = 7.03),and targets that made group attributions asmore attractive to the group (M = 7.62) thanthose that made individual attributions fortheir performance (M = 7.03). The predicted

Achievement Level Attribution Style inter-action was also significant ( F (1, 76) = 4.67,

p = .034). As predicted, when the target failedattribution style did not affect ratings of thetarget’s attractiveness as a group member


212

Table 3. Experiment 2: Means and standard deviations (in parentheses) for target evaluations and groupevaluations as a function of achievement level and attribution style

Achievement level

Fail Succeed

Individual Group Individual Groupattribution attribution attribution attribution

Personal character 6.10 6.91 5.89 7.19(1.38) (0.91) (1.15) (0.77)

Group member attractiveness 7.03 7.02 7.03 8.22(1.18) (1.37) (1.62) (0.50)

Overall group impression 7.75 7.45 7.10 7.85(0.97) (1.19) (1.33) (0.99)

Comparative group impression 7.80 7.10 6.74 7.85

(0.77) (1.48) (1.52) (0.99)



( F < 1). However, successful targets were ratedas more attractive group members when they made group rather than individual attributionsfor their performance ( F (1, 76) = 9.19, p = .003).

Personal character The predicted Achievement Level Attribution Style interaction was not significant ( F (1, 76) = 1.08, ns ), and there wasonly a significant main effect for AttributionStyle ( F (1, 76) = 19.32, p < .000), indicating that targets who made group attributions wereevaluated as having a more positive personalcharacter than targets who made individualattributions (M s = 7.05 and 5.99 respectively).

Group evaluation On the item measuring theoverall impression of Australians only the pre-dicted Attribution Style Achievement Levelinteraction was significant ( F (1, 76) = 4.32,

p = .041). Group evaluations were not affectedby attribution style when the target failed( F < 1), but when the target succeeded andmade group attributions for success, partici-pants had a more positive impression of thegroup than when the target succeeded andmade individual attributions for success ( F (1,76) = 4.41, p = .039).

On the item comparing Australia’s image toother countries only the Attribution Style

Achievement Level interaction was significant ( F (1, 75) = 10.75, p = .002). There was atendency to rate Australia’s image as lessfavorable than other countries when the target failed and made group attributions relative to

when they failed and made individual attribu-tions ( F (1, 75) = 3.25, p = .076). In contrast,group members rated Australia’s image as morefavorable than other countries when the target succeeded and made group rather than individ-ual attributions ( F (1, 75) = 8.00, p = .006).

Mediational analyses Analyses were againconducted to test whether the significant effect of attribution style on target evaluations ismediated by group evaluations. In Experiment 1, sporting achievement importance was thegroup evaluation variable used as the mediator.In Experiment 2, the results show a clear andsignificant effect of attribution style on the

successful target’s attractiveness as a groupmember and the group evaluations measures of overall impression of Australia and comparativefavorability of Australia. Thus, group member

attractiveness of the successful target was used asthe dependent measure and overall impressionof Australia and comparative favorability of

Australia were used as the proposed mediators.Separate analyses were conducted testing for themediating effect of the two group evaluation

variables. Overall impression of Australia wassignificantly related to the targets’ attractivenessas a group member ( = .47, p = .002) and thesignificant effect of attribution style on groupmember attractiveness ( = .48, p = .002)

became less significant ( = .26, p = .02) onceoverall impression of Australia was included inthe analysis. The reduction in the effect of attri-bution style was marginally significant accordingto the Sobel test (z = 1.73, p = .07). Regressionanalyses testing whether ratings of the compara-tive favorability of Australia mediated therelationship between attribution style and thetarget’s attractiveness as a group memberrevealed that the proposed mediator was not significantly related to the dependent measure( = .25, p = .12) and, thus, mediation could not be established.

Discussion

Consistent with predictions and with thefindings of Experiment 1, evaluations of ingroup positive deviates were influenced by the type of attributions they made for their per-formance. We argued that positive deviates that succeeded and implicated the group in theirsuccess would be evaluated more positively thanthose who attributed their success to individualfactors. This pattern emerged when groupmembers were judging the positive deviate’sattractiveness as a group member. When thepositive deviate failed, however, attribution styledid not influence the positive deviate’s attrac-tiveness as a group member. It is probably not surprising that evaluations that relate to atarget’s attractiveness as a group membershould be sensitive to actions and behavior that impact on the group and its image. When

judging personal character, only attribution


213



style influenced evaluations. Positive deviates who made group attributions for their perform-ance were judged more favorably than those

who made individual attributions.

The logic for our predictions rests on a socialidentity explanation: successful group members

who reflect their success onto the group satisfy group enhancement motives. In support of thisreasoning, group evaluations were influencedby the successful positive deviates’ attributionstyle. Australians in general were evaluatedmore favorably and Australia’s image was con-sidered more favorable in comparison to othercountries when the successful positive deviatemade group rather than individual attributions.

Moreover, analyses showed that the effect of attribution style on the successful target’s groupmember attractiveness was partially mediated by overall impressions of Australia. Consistent withExperiment 1, this suggests that responses topositive deviates are determined, at least in part,by the impact they have on the group’s image.

The attributions of the failed positive deviatedid not influence overall evaluations of Aus-tralians but there was a tendency to downgradethe image of Australia relative to other countries

when the positive deviate failed and made groupattributions. This pattern gives some suggestionthat attributions that reflect a group member’sfailure on to the group can have a negativeimpact on the group’s image. However, as thispattern only emerged on one measure and isonly marginally significant it should be treated

with caution. As noted in Experiment 1, the lackof control condition in the study only allows usto assess the impact of the manipulations ongroup evaluations in relative not absolute terms.

Although there was some evidence from thisexperiment that ingroup members who makegroup attributions receive higher ratingsregardless of their level of achievement, thispattern emerged on one measure and thismeasure assessed the individual characteristicsof the positive deviates. On the more group-related evaluations, group attributions only ledto higher ratings when the positive deviate wassuccessful. Taken together, the results of thisexperiment lend further support to ourargument that responses to positive deviates are

primarily motivated by their ability to enhancethe ingroup’s image.

General discussionIn the introduction we posed a question: What do we know about reactions to outstanding orexceptional individuals? We went on to arguethat an important factor influencing ourreactions to positive deviates is social categoriz-ation and the identity-based motivations that result from categorizing as a group member.Central to the social identity perspective is theproposition that individuals are motivated tofeel good about themselves and that this need

can be satisfied in group contexts by member-ship in a positively evaluated group. To theextent that a positive deviate’s successes andachievements are reflected onto the group, andits members can bask in the reflected glory of the achievements (Cialdini et al., 1976; Cialdini& De Nicholas, 1989; Cialdini & Richardson,1980; Wann & Branscombe, 1990), positivedeviates should garner favorable reactions. Any actions on the part of positive deviates that interfere with the reflective process shouldreduce the positive deviate’s ability to benefit the group and its members; the likely outcomeis less positive reactions to the positive deviate.

One important way that positive deviates canreflect or de flect their successes and achieve-ments onto or away from the group is by theattributions they make for their performance.Our general prediction is that positive deviates

who make group-related attributions for theirsuccess will be evaluated more positively thanthose who make individual-related attributions.There was support for this prediction acrosstwo studies. Group members upgraded thepersonal character of ingroup positive deviates

who gave credit to the group for their perform-ance relative to those positive deviates who only took personal credit. Moreover, in Experiment 1, sporting positive deviates who made groupattributions were considered more attractivegroup members than those who made individ-ual attributions. In the second Experiment,ingroup positive deviates who were successfuland attributed that success to the group were


214



judged as more attractive group members thanthose who attributed their success to individualfactors.

Together, these findings provide clear

evidence that being a successful group memberand implicating your group in that success couldbe one way to garner approval and support fromfellow ingroup members. In addition, thereasoning that positive deviates who make groupattributions can help satisfy group-enhancement motives was supported by the results on thegroup and self-evaluation measures. In Experi-ment 1, group members rated their ownsporting ability higher after reading about highachieving ingroup members who attributed

their success to the group compared to ingroupmembers who took personal credit for theirsuccess. They also tended to increase percep-tions of group homogeneity and emphasizedthe importance of the sporting dimension, adimension that was largely peripheral to groupidentity. In Experiment 2, group identity was

judged as more favorable overall and as morefavorable than other countries when a successfulingroup member made group attributions forhis or her success. These results suggest that ingroup positive deviates who make group attri-butions contributed to a positively distinctiveingroup identity.

Results of the mediational analyses providetentative support for the logic underpinningour predictions. We argued that ingrouppositive deviates who make group attributionsare evaluated more favorably because their attri-butions reflect success on to the group, thereby enhancing ingroup image. In Experiment 1, theeffect of attribution style on judgments ofthe ingroup target’s personal character wasmediated by a group evaluation measure (i.e.sporting achievement dimension) and in Exper-iment 2 the effect of attribution style on ratingsof the successful target’s attractiveness was par-tially mediated by the overall impression of thegroup. Together these results indicate that target evaluations are, at least in part, beingdriven by group-related concerns. As noted pre-

viously, the failure to find stronger mediationeffects may perhaps reflect the group evaluationmeasures used in the two studies. The inclusion

of more comprehensive and sensitive groupevaluations measures in future studies may helpto tease out the nature of the relationshipbetween the variables. Low power due to small

numbers may also have played some part.Our findings are consistent with previous

research examining reactions to groupmembers’ attributions. Results from researchby Bond et al. (1984) and Taylor and Tyler(1986) indicated that group members preferother ingroup members who make attributionsthat enhance the image of the group. Whereasthe previous research has focused on indi-

viduals’ attributions for group outcomes (Bondet al., 1984; Forsyth et al., 1981; Forsyth &

Mitchell, 1979; Taylor & Tyler, 1986) thecurrent research instead looked at how percep-tions are affected by the attributions individualgroup members make for their own outcomes.Despite the difference in focus, there isconsistency in the finding of more favorableevaluations of ingroup members who aregroup-enhancing in their explanations.

A number of alternative explanations for ourresults can be ruled out. The first is that target evaluations might simply reflect general expec-tations that individuals should be modest orshow group allegiance. This explanation wouldsuggest that all individuals regardless of groupmembership would be upgraded for makinggroup attributions. This clearly was not the case:attribution style only influenced evaluations of ingroup positive deviates. It could also be arguedthat group members may have greater expec-tations that ingroup members will conform tonorms (cf. black sheep research; e.g. Marques &Paez, 1994). If this were the case then evalu-ations in the second experiment should only beinfluenced by attribution style and not by

whether the ingroup positive deviate succeededor failed. Yet on three of the four evaluationmeasures group attributions only significantly influenced evaluations when the positive deviatesucceeded. The weight of evidence seems toindicate that positive deviates who made groupattributions were not upgraded just becausethey adhered to normative expectations, but because ingroup members were able to benefit from the positive deviates’ achievements.


215



One limitation of our research is the lack of a baseline control condition in the designs.

Although this omission is not uncommon insocial comparison research (e.g. Brewer &

Weber, 1994; Major, Sciacchitano, Crocker,1993; see also Collins, 1996), without this typeof comparison condition, no firm conclusionscan be drawn about the direction of effects.Thus, although we propose that exposure topositive deviates who make group attributionsfor their success should result in more positivegroup and self-evaluations, it is possible that these evaluations are simply less negative.Future research including baseline control con-ditions is needed to address this issue.

Although the present research has focusedon positive forms of deviance, it is possible that the attributions that group members make forpoor performance or low achievement may alsobe important for how they are evaluated. Expla-nations for low achievement that relate to thegroup have the potential to threaten thepositive distinctiveness of ingroup identity anda negative deviate who makes these types of attributions may be harshly evaluated. On theother hand, negative deviates who personalizetheir failure may be evaluated more kindly.

Attributions are just one way that deviates couldrally support or ward off harsh treatment; otherstrategies could include statements of pride,allegiance, and identification with their group.

In sum, previous research suggests a numberof factors that may influence reactions topositive deviates. Shared group membership isone important factor that influences reactions:high achieving ingroup members receive morepositive reactions in salient intergroup contexts(Schmitt et al., 2000) and their achievementsbenefit other group members (Brewer & Weber,1994). Our research suggests that, in addition toshared group membership, positive deviates canfurther influence members’ reactions by theattributions they make for their achievement.Positive deviates who implicate the group intheir success stand to garner more supportand greater acceptance from other ingroupmembers. The take-home message is that highachievers seeking to gain support from othergroup members or those wishing to avoid beingthe target of jealousy or antagonism may be able

to do this by emphasizing their shared groupidentity and the ways in which the group hascontributed to their success.

Notes

1. Although it could be argued that targets whomake group attributions would also be judged asmore modest, the pattern of results showed thepredicted interaction and no main effect forattribution style on this item. Thus, this item waskept as a measure of personal character.

2. An unexpected Attribution Style Achievement Dimension interaction also emerged ( F (1,57) =5.07, p = .028). An inspection of the means showsthat the interaction arises because the sporting

target is perceived as giving less credit to thegroup than the academic target in the individualattribution condition (M s = 3.00 and 4.39;

F (1,59) = 15.91, p = .000) whereas there is nodifference between the sporting and academictargets in the group attribution condition(M s = 6.60 and 6.87; F < 1). It should be noted,however, that in support of the attribution stylemanipulation both the academic target ( F (1,59)= 24.82, p = .000) and the sporting target( F (1,59) = 79.20, p = .001) were perceived asgiving more credit to the group for their success

when they made group rather than individualattributions.

3. Analyses testing for the effect of perceived grouphomogeneity and own sporting ability as potentialmediators also revealed that neither variable wassignificantly related to the dependent measure.

Acknowledgment

The authors would like to thank Matthew Hornsey and Winnifred Louis for their helpful comments onan earlier draft of this article.

References

Abrams, D., Marques, J., Bown, N., & Dougill, M.(2002). Anti-norm and pro-norm deviance in thebank and on the campus: Two experiments onsubjective group dynamics. Group Processes &

Intergroup Relations, 5 , 163–182. Abrams, D., Marques, J. M., Bown, N., & Henson, M.

(2000). Pro-norm and anti-norm deviance withinand between groups. Journal of Personality and

Social Psychology, 78 , 906–912.Baron, R. M., & Kenny, D. A. (1986). Themoderator-mediator variable distinction in social


216



psychological research: Conceptual, strategic, andstatistical considerations. Journal of Personality and

Social Psychology, 51, 1173–1182.Bond, M. H., Chiu, C., & Wan, K. (1984). When

modesty fails: The social impact of group-effacingattributions following success or failure. European

Journal of Social Psychology, 14 , 335–338.Brewer, M. B. (1979). Ingroup bias in the minimal

intergroup situation: A cognitive-motivationalanalysis. Psychological Bulletin, 86 , 307–324.

Brewer, M. B., & Kramer, R. M. (1985). Thepsychology of intergroup attitudes and behavior.Annual Review of Psychology, 36 , 219–243.

Brewer, M. B., & Weber, J. G. (1994). Self-evaluationeffects of interpersonal versus intergroup socialcomparisons. Journal of Personality and Social

Psychology, 66 , 268–275.Burgoon, J. K. (1993). Interpersonal expectations,

expectancy violations, and emotionalcommunication. Journal of Language and Social

Psychology, 12 , 30–48.Cialdini, R., Borden, R., Thorne, A., Walker, M.,

Freeman, S., & Sloan, L. (1976). Basking inreflected glory: Three (football) field studies.

Journal of Personality and Social Psychology, 34 ,366–375.

Cialdini, R. B., & De Nicholas, M. E. (1989).Self-presentation by association. Journal of

Personality and Social Psychology, 57 , 626–631.Cialdini, R. B., & Richardson, K. D. (1980). Two

indirect tactics of image management: Baskingand blasting. Journal of Personality and Social

Psychology, 39 , 406–415.Collins, R. L. (1996). For better or worse:

The impact of upward social comparison onself-evaluations. Psychological Bulletin, 119 , 51–69.

Doosje, B., Spears, R., & Koomen, W. (1995). Whenbad isn’t all bad: Strategic use of sampleinformation in generalization and stereotyping.

Journal of Personality and Social Psychology, 69 ,

642–655.Feather, N. T. (1989). The fall of the high achiever:

Research into attitudes toward the tall poppy. In J. P. Forgas & J. M. Innes (Eds.), Recent advances in

social psychology: An international perspective

(pp. 7–20). Amsterdam: Elsevier.Feather, N. T. (1993). The rise and fall of political

leaders: Attributions, deservingness, personality,and affect. Australian Journal of Psychology, 45 ,61–68.

Feather, N. T. (1994). Attitudes towards highachievers and reactions to their fall: Theory and

research concerning tall poppies. Advances in Experimental Social Psychology, 26 , 1–73.

Feather, N. T., Volkmer, R. E., & McKee, I. R. (1991). Attitudes towards high achievers in public life: Attributions, deservingness, personality, and affect.Australian Journal of Psychology, 43 , 85–91.

Fielding, K. S., & Hogg, M. A. (1997). Social identity,self-categorization, and leadership: A field study of small interactive groups. Group Dynamics: Theory,

Research, and Practice, 1, 39–51.Forsyth, D. R., Berger, R. E., & Mitchell, T. (1981).

The effects of self-serving vs. other-serving claimsof responsibility on attraction and attribution ingroups. Social Psychology Quarterly, 44 , 59–64.

Forsyth, D. R., & Mitchell, T. (1979). Reactions toothers’ egocentric claims of responsibility. Journal

of Psychology, 103 , 281–285.Haslam, S. A., Oakes, P. J., Turner, J. C., & McGarty,

C. (1995). Social categorization and grouphomogeneity: Changes in the perceivedapplicability of stereotype content as a function of comparative context and trait favourableness.British Journal of Social Psychology, 34 , 139–160.

Hinkle, S., Taylor, L. A., Fox-Cardamone, L., &Ely, P. G. (1998). Social identity and aspects of social creativity: Shifting to new dimensions of intergroup comparison. In S. Worchel, J. F.Morales, D. Páez, & J.-C. Deschamps (Eds.), Social

identity: International perspectives (pp. 166–179).London: Sage.

Hogg, M. A. (1993). Group cohesiveness: A criticalreview and some new directions. European Review of

Social Psychology, 4 , 85–111.Hogg, M. A. (2001). A social identity theory of

leadership. Personality and Social Psychology Review,

5 , 184–200Hogg, M. A. (2003). Social identity. In M. R. Leary &

J. P. Tangney (Eds.), Handbook of self and identity

(pp. 462–479). New York: Guilford.Hogg, M. A., & Abrams, D. (1988). Social

identifications . London: Routledge.Hogg, M. A., Cooper-Shaw, L., & Holzworth, D. W.

(1993). Group prototypicality and depersonalizedattraction in small interactive groups. Personality

and Social Psychology Bulletin, 19 , 452–465.Hogg, M. A., Fielding, K. S., & Darley, J. (2005).

Fringe dwellers: Processes of deviance andmarginalization in groups. In D. Abrams,

J. Marques, & M. A. Hogg (Eds.), Social psychology

of inclusion and exclusion .(pp. 191–210). New York:Psychology Press.

Hogg, M. A., & Hains, S. C. (1996). Intergrouprelations and group solidarity: Effects of groupidentification and social beliefs on depersonalized

attraction. Journal of Personality and Social Psychology, 70 , 295–309.


217



Hogg, M. A., & Hardie, E. A. (1991). Socialattraction, personal attraction, andself-categorization: A field study. Personality and

Social Psychology Bulletin, 17 , 175–180.

Jackson, L. A., Sullivan, L. A., Harnish, R., & Hodge,C. N. (1996). Achieving positive social identity:Social mobility, social creativity, and permeability of group boundaries. Journal of Personality and

Social Psychology, 70 , 241–254. Jones, E. E., & McGillis, D. (1976). Correspondent

inferences and the attribution cube: A comparative reappraisal. In J. H. Harvey,

W. J. Ickes, & R. F. Kidd (Eds.), New directions in

attribution research (Vol. 1, pp. 389–420). Hillsdale,NJ: Erlbaum.

Jussim, L., Coleman, L. M., & Lerch, L. (1987). The

nature of stereotypes: A comparison andintegration of three theories. Journal of Personality

and Social Psychology, 52 , 536–546.Lockwood, P., & Kunda, Z. (1997). Superstars and

me: Predicting the impact of role models on theself. Journal of Personality and Social Psychology, 73 ,91–103.

Major, B., Sciacchitano, A. M., & Crocker, J. (1993).In-group versus out-group comparisons andself-esteem. Personality and Social Psychology Bulletin,

19 , 711–721.Marques, J. M., Abrams, D., Páez, D., &

Martinez-Taboada, C. (1998). The role of categorization and in-group norms in judgmentsof groups and their members. Journal of Personality

and Social Psychology, 75 , 976–988.Marques, J. M., Abrams, D., & Serôdio, R. G. (2001).

Being better by being right: Subjective groupdynamics and derogation of in-group deviants

when generic norms are undermined. Journal of

Personality and Social Psychology, 81, 436–447.Marques, J. M., & Páez, D. (1994). The ‘black sheep

effect’: Social categorization, rejection of ingroupdeviates, and perception of group variability.

European Review of Social Psychology, 5 , 37–68.Pickett, C. L., & Brewer, M. B. (2001). Assimilation

and differentiation needs as motivationaldeterminants of perceived in-group and out-grouphomogeneity. Journal of Experimental Social

Psychology, 37 , 341–348.Schmitt, M. T., Silvia, P. J., & Branscombe, N. R.

(2000). The intersection of self-evaluationmaintenance and social identity theories:Intragroup judgment in interpersonal andintergroup contexts. Personality and Social Psychology

Bulletin, 26 , 1598–1606.

Segal, M. W. (1979). Varieties of interpersonalattraction and their interrelationships in naturalgroups. Social Psychology Quarterly, 42 , 253–261.

Tajfel, H., & Turner, J. (1979). An integrative theory of intergroup conflict. In W. G. Austin &S. Worchel (Eds.), The social psychology of intergroup

relations (pp. 33–47). Monterey, CA: Brooks/Cole.

Taylor, D. M., Doria, J. R., & Tyler, J. K. (1983).Group performance and cohesiveness: Anattribution analysis. Journal of Social Psychology, 119 ,187–198.

Taylor, D. M., & Tyler, J. K. (1986). Group members’responses to group-serving attributions for successand failure. Journal of Social Psychology, 126 ,775–781.

Tesser, A. (1988). Toward a self-evaluationmaintenance model of social behavior. InL. Berkowitz (Ed.), Advances in experimental social

psychology ( Vol. 20, pp. 181–227). New York:

Academic Press.Turner, J. C. (1982). Towards a cognitive redefinition

of the social group. In H. Tajfel (Ed.), Social

identity and intergroup relations (pp. 15–40).Cambridge, UK: Cambridge University Press.

Turner, J. C., Hogg, M. A., Oakes, P. J., Reicher, S. D.,& Wetherell, M. S. (1987). Rediscovering the social

group: A self-categorization theory . Oxford, UK:Blackwell.

Wann, D. L., & Branscombe, N. R. (1990). Die-hardand fair-weather fans: Effects of identification andBIRGing and CORFing tendencies. Journal of Sport

and Social Issues, 14 , 103–117.

Paper received 21 May 2003; revised version accepted

15 May 2005.

Biographical notes

KE LL Y S. FIELDING is a lecturer in psychology andresearch methods at the University of Queensland,Brisbane, Australia. Her research interests are insocial identity and self-categorization processes,

leadership, intragroup differentiation, andenvironmental sustainability.

M IC HA EL A . HOGG is professor of social psychology,and an Australian Research Council ProfessorialFellow at the University of Queensland, Australia.His research interests broadly span groupprocesses, intergroup relations, social identity, andself-conception with a current focus on leadership,deviance, uncertainty reduction, and subgrouprelations.

N I CO L E A N NA N DA L E completed her Honours inPsychology at the University of Queensland. She iscurrently working as a counselor in Canada.


218

Group Processes Intergroup Relations-2006-Fielding-199-218.pdf

Documents