Multiple Identity Configurations: The Benefits of Focused ... Files/Multiple Identity...

This article was downloaded by: [128.103.149.52] On: 19 September 2017, At: 11:14Publisher: Institute for Operations Research and the Management Sciences (INFORMS)INFORMS is located in Maryland, USA

Organization Science

Publication details, including instructions for authors and subscription information:http://pubsonline.informs.org

Multiple Identity Configurations: The Benefits of FocusedEnhancement for Prosocial Behaviorhttp://orcid.org/0000-0002-9979-5354Lakshmi Ramarajan, Ida E. Berger, Itay Greenspan

To cite this article:http://orcid.org/0000-0002-9979-5354Lakshmi Ramarajan, Ida E. Berger, Itay Greenspan (2017) Multiple IdentityConfigurations: The Benefits of Focused Enhancement for Prosocial Behavior. Organization Science 28(3):495-513. https://doi.org/10.1287/orsc.2017.1129

Full terms and conditions of use: http://pubsonline.informs.org/page/terms-and-conditions

This article may be used only for the purposes of research, teaching, and/or private study. Commercial useor systematic downloading (by robots or other automatic processes) is prohibited without explicit Publisherapproval, unless otherwise noted. For more information, contact [email protected].

The Publisher does not warrant or guarantee the article’s accuracy, completeness, merchantability, fitnessfor a particular purpose, or non-infringement. Descriptions of, or references to, products or publications, orinclusion of an advertisement in this article, neither constitutes nor implies a guarantee, endorsement, orsupport of claims made of that product, publication, or service.

Copyright © 2017, INFORMS

Please scroll down for article—it is on subsequent pages

INFORMS is the largest professional society in the world for professionals in the fields of operations research, managementscience, and analytics.For more information on INFORMS, its publications, membership, or meetings visit http://www.informs.org

http://pubsonline.informs.org

https://doi.org/10.1287/orsc.2017.1129


http://pubsonline.informs.org/page/terms-and-conditions

http://www.informs.org

ORGANIZATION SCIENCEVol. 28, No. 3, May–June 2017, pp. 495–513

http://pubsonline.informs.org/journal/orsc/ ISSN 1047-7039 (print), ISSN 1526-5455 (online)

Multiple Identity Configurations: The Benefits of FocusedEnhancement for Prosocial BehaviorLakshmi Ramarajan,a Ida E. Berger,b Itay Greenspanc

aHarvard Business School, Boston, Massachusetts 02163; bTed Rogers School of Business Management, Ryerson University, Toronto, OntarioM5B 2K3, Canada; cPaul Baerwald School of Social Work and Social Welfare, Hebrew University of Jerusalem, Mt. Scopus, Jerusalem 91905,IsraelContact: [email protected], http://orcid.org/0000-0002-9979-5354 (LR); [email protected] (IEB); [email protected] (IG)

Received: May 16, 2014Revised: July 31, 2015; February 28, 2016;September 2, 2016; November 2, 2016;December 19, 2016Accepted: January 13, 2017Published Online in Articles in Advance:May 23, 2017


Copyright: © 2017 INFORMS

Abstract. This paper introduces a configurational approach to the study of multiple iden-tities. Specifically, it examines how prosocial identity combines with collective and indi-vidualistic identities in conflicting and enhancing ways to affect prosocial behavior inorganizational settings. We examine an unexplored intuition in the multiple identitiesliterature that when all identities are enhancing (a mutual enhancement configuration),it will be best for prosocial outcomes. Our results show, however—across two field stud-ies and two experiments—that enhancement between prosocial and collective identities(a focused enhancement configuration) results in the highest levels of prosocial behav-ior. Furthermore, we trace this result to the greater self-serving orientation activated in amutual enhancement configuration, where one’s individualistic identity enhances one’sother identities. Our work demonstrates the value of a configurational approach to thestudy of multiple identities, and it challenges the assumption that a mutual enhancementconfiguration is always desirable.

Supplemental Material: The online appendix is available at https://doi.org/10.1287/orsc.2017.1129.

Keywords: identity • identity configurations • multiple identities • prosocial behavior

IntroductionProsocial behavior—behavior that protects, promotesor contributes to the welfare of others (Grant 2007)—iscrucial for organizations and societies to function well.Organizations have sought to promote prosocial behav-ior by activating their members’ prosocial identities—identities such as “mentor” or “volunteer”—that orienta person toward helping others (Piliavin and Callero1991,Grant et al. 2008). Yet, a person’s prosocial identitydoes not exist in a vacuum. Scholars have long recog-nized that people have multiple identities (James 1890).Organizations are a fertile context for the activation oftheir members’ multiple identities, not just prosocialidentities, but also individualistic and collective identi-ties (Bartel 2001, Blader 2007). However, little researchhas examined how multiple identities affect prosocialbehavior, particularly in organizational contexts.Individualistic and collective identities are funda-

mental aspects of the self-concept and therefore espe-cially important to examine in combination withprosocial identities (Brewer and Gardner 1996). Indi-vidualistic identities are egoistic, “me” orientedaspects of the self-concept and include idiosyncraticpersonal characteristics such as “trailblazer” or “star.”Collective identities are “we” oriented aspects of theself-concept, and include one’s group memberships,such as ethnic or professional affiliations (Brewer andGardner 1996). Prosocial identities are “you” oriented

aspects of the self-concept and include those partsof the self that are in a giving or helping relation-ship with others (Grant and Dutton 2012, Piliavin andCallero 1991). In this paper, we examine how individu-als’ prosocial, individualistic, and collective identitiescombine to affect their prosocial behavior.

The perspective we take hinges on understandinghow individuals experience their multiple identities.Building on past research showing that two iden-tities can be related in conflicting (Benet-Martinezand Haritatos 2005, Greenhaus and Beutell 1985) orenhancing ways (Dutton et al. 2010, Rothbard andRamarajan 2009), we take a novel, configurationalapproach to multiple identity research, examiningpatterns of conflict and enhancement among threeidentities. Moreover, we investigate an unexploredintuition in the multiple identities literature that multi-ple identities will be most effective for outcomes whenthey are all mutually enhancing, because the full setof meanings, values, and behaviors associated withone’s identities can be applied to a task (Ashforth 2007,Dutton et al. 2010, Pratt and Foreman 2000). By explic-itly considering all three specific types of identities(prosocial, collective, and individualistic), we questionthis assumption, both theoretically and empirically.Instead, we propose that a person with an identityconfiguration characterized by focused enhancement (inwhich prosocial and collective identities are mutually

495

Dow

nloa

ded

from

info

rms.

org

by [

128.

103.

149.

52]

on 1

9 Se

ptem

ber

2017

, at 1

1:14

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.

http://pubsonline.informs.org/journal/orsc/

mailto:[email protected]

http://orcid.org/0000-0002-9979-5354




Ramarajan, Berger, and Greenspan: Multiple Identity Configurations496 Organization Science, 2017, vol. 28, no. 3, pp. 495–513, ©2017 INFORMS

enhancing but both are unrelated to an individual-istic identity) will exhibit higher levels of prosocialbehavior than a person with an identity configurationcharacterized by mutual enhancement (mutual enhance-ment among all three identities—prosocial, collective,and individualistic). Across two field studies and twoexperiments, we consistently find that identity config-urations of focused enhancement between prosocial andcollective identities result in higher prosocial outcomesthan configurations of mutual enhancement among allthree identities. Our work contributes to research onboth multiple identities and prosocial behavior bydemonstrating the value of a configurational approachthat highlights the importance of the specific types ofidentities that enhance one another, rather than assum-ing that mutual enhancement is always desirable.

Prosocial Behavior and Multiple IdentitiesProsocial Behavior in OrganizationsProsocial behavior in organizational contexts captureshow organizational members engage with one anotheras givers and receivers, as well as how they engagewith beneficiaries in the larger society. Prosocial orga-nizational behavior is defined as behavior which is“(a) performed by a member of an organization, (b)directed toward an individual, group, or organizationwith whom he or she interacts while carrying out hisor her organizational role and (c) performed with theintention of promoting the welfare of the individual,group, or organization toward which it is directed”(Brief and Motowidlo 1986, p. 711). The organizationalimportance of prosociality is evident from studies ofa wide range of behaviors, such as donating (O’Reillyand Chatman 1986), volunteering (Grant and Sumanth2009), and helping in groups (Blader and Tyler 2009)as well as studies in a range of contexts, including vol-untary organizations (Schaubroeck and Ganster 1991),university alumni (O’Reilly and Chatman 1986), andamong employees who volunteer inside or outsidetheir organizations (Bartel 2001, Grant et al. 2008, Grantand Sumanth 2009).

Identities: Types and RelationshipsIdentity Types. Identities, “the set of meanings thatdefine who one is” (Burke and Stets 2009, p. 3), area fundamental driver of individual behavior in orga-nizations (Bartel et al. 2007). Individualistic identitiesdrive behaviors that promote one’s personal welfareand goals; collective identities drive behaviors thatpromote or stem from feelings of group belonging(Brickson 2013, Brewer and Gardner 1996) and proso-cial identities drive behaviors that promote the wel-fare of others (Grant 2007). Prosocial identity differsfrom collective identity because it is defined in termsof helping others regardless of whether the recipientshares the giver’s group memberships. For instance,

Grant and colleagues (2008) show that employee par-ticipation in a corporate giving program affected bothprosocial identity and organizational identity (a collec-tive identity) independently.

Two of the three types of identities we consider havebeen found to influence prosocial behavior in organi-zational contexts. A prosocial identity drives proso-cial behavior (Grant 2007, Penner et al. 2005). Forexample, identifying oneself with an altruistic role,such as donor, increases how much, how often, andhow consistently one donates to others (Grube andPiliavin 2000, Piliavin and Callero 1991). A collectiveidentity also leads to prosocial behavior. For exam-ple, alumni identification with their university influ-ences their donation behavior (Mael and Ashforth1992, O’Reilly and Chatman 1986). Whether individu-alistic identity also drives prosocial behavior remainsan open question.Relationships Among Identities. Scholars have begunto recognize that two or more identities may be simul-taneously salient or coactivated for individuals, rais-ing questions about how one’s identities are related(Ashforth and Johnson 2001, Blader 2007, Rothbardand Ramarajan 2009). Research suggests that simulta-neously salient identities may be related in conflicting(Benet-Martinez and Haritatos 2005, Greenhaus andBeutell 1985) or enhancing ways (Dutton et al. 2010), orthey may be unrelated and operate independently ofone another (Lipponen et al. 2005).

When people experience identity conflict, they feelcaught between opposingmeanings that they associatewith their different identities. Their cognition narrows,their energy is depleted, and negative affect abounds(Hugenberg and Bodenhausen 2004, Rothbard 2001,Hirsh and Kang 2016), often leading to diminishedwell-being and poor task performance (Brook et al.2008, Cheng et al. 2008). When people experienceidentity enhancement, they feel that the values andmeanings associated with their various identitiesare complementary. Energy and positive affect areabundant, creating engagement and inspiring growth(Creary et al. 2015, Dutton et al. 2010, Rothbard 2001).Consistent with research on the independence of pos-itive and negative affect and motivation (Carver andWhite 1994, Watson et al. 1988), low identity conflict isnot equivalent to high identity enhancement. Further-more, identity conflict and enhancement can coexist.For example, some aspects of one’s work identity canbe in conflict with some aspects of one’s family iden-tity, while other aspects of both identities can enhanceone another (Tiedje et al. 1990).

Multiple Identity ConfigurationsThe above research has largely examined relationshipsbetween two identities (i.e., whether identities A andB are related in conflicting or enhancing ways). Yet,

Dow

nloa

ded

from

info

rms.

org

by [

128.

103.

149.

52]

on 1

9 Se

ptem

ber

2017

, at 1

1:14

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.

Ramarajan, Berger, and Greenspan: Multiple Identity ConfigurationsOrganization Science, 2017, vol. 28, no. 3, pp. 495–513, ©2017 INFORMS 497

examining a single relationship of conflict or enhance-ment between a pair of identities (such as work andfamily identities or organizational and professionalidentities) overlooks the reality that there may be morethan two identities operating and therefore manymorerelationships among those identities to consider. Bymoving from two to three identities, as we do here,more complex patterns of relationships between iden-tities emerge. For instance, people may experience allthree identities as unrelated, two identities as enhanc-ing and the third as unrelated, two identities as enhanc-ing and the third conflicting with both, and so on.Scholars have called for greater understanding of the

complexity of people’s experience and management ofmore than two identities, suggesting a variety of pat-terns that better characterize how multiple identitiesmay operate as a “whole,” rather than simply as indi-vidual parts (Ashforth 2007, Pratt and Foreman 2000,Roccas and Brewer 2002, Ramarajan 2014). We use theterm “multiple identity configurations” for patterns ofconflict and enhancement among more than two iden-tities, as “configuration” implies “a single set of rela-tions among many components” such that they “canbe considered a single entity” (Schachter 2004, p. 170).

Although many different multiple identity configu-rations may exist, the dominant intuition in the liter-ature is that a configuration characterized by mutualenhancement among all simultaneously salient identi-ties will be the most effective one for fostering manyorganizationally relevant behaviors (Ashforth 2007,Dutton et al. 2010, Pratt and Foreman 2000). The rea-soning is that in this configuration, each identity facil-itates the meanings, values, and behaviors of the oth-ers. Yet, this intuition overlooks the specific types ofidentities that may be enhancing one another. Below,we question this intuition by considering both thetypes of identities and the relationships among themto hypothesize how prosocial, collective, and individ-ualistic identities jointly operate to influence prosocialbehavior.

Identity Configurations andProsocial BehaviorTo develop hypotheses about how configurationsof prosocial, collective, and individualistic identitiesaffect prosocial behavior, we define three theoreticallyrelevant patterns of relationships.

Mutual ConflictThis identity configuration is characterized by a patternof conflicting relationships between an individual’sprosocial, collective, and individualistic identities. Wepropose that high levels of conflict between all threeidentities lead to low levels of prosocial behavior.Whenpeople perceive their identities as conflicting, they feelcaught between opposing values, expectations, and

meanings (Hirsh and Kang 2016, Brook et al. 2008). Theexperience of feeling torn between different aspects of“who one is” creates a sense of psychological stressand insecurity (Marcussen 2006), which may detractfrom one’s ability to advance the other party’s wel-fare. Identity conflict is also likely to turn one’s focusinward toward oneself rather than outward toward oth-ers, because one is deciding between various aspects ofwho one is. Regulating and choosing among aspects ofwhoone is candeplete one’s cognitive andmotivationalresources, leaving less energy and attention to give toothers (Baumeister et al. 1998, Marks 1977, Rothbard2001). Thus, we have the following:

Hypothesis 1 (H1). Identity configurations characterizedby high mutual conflict between an individual’s prosocial,collective, and individualistic identities will be associatedwith low levels of prosocial behavior.

Mutual EnhancementThis identity configuration is characterized by a pat-tern of mutual enhancement between an individual’sprosocial, collective, and individualistic identities.Consistent with the dominant intuition, we proposethat high levels of enhancement between all threeidentities should increase prosocial behavior. Identityenhancement can expand the cognitive and motiva-tional resources people bring to their work, fosteringmore task-focused behavior (Creary et al. 2015, Rama-rajan 2014). This is because when a person feels thatall aspects of herself are facilitating one another, shecan focus on the task at hand rather than on who sheis (Baumeister et al. 1998, Marcussen 2006). Further,the more people engage with the groups and relation-ships associated with their various identities, the morelikely they are to gain meaning and energy, which theycan then devote to the task (Marks 1977, Rothbard2001). Identity enhancement can also provide relevantresources to apply to the task because it enables accessto the variety of identity-based knowledge, skills, andbehaviors one possesses (Caza andWilson 2009, Chenget al. 2008).

Identity enhancement may also create a sense of psy-chological security, because different aspects of whoone is are being expressed and validated simultane-ously (Stets andHarrod 2004, Swann 1983). The greaterone’s sense of psychological security, themore onemaybe able to transcend focusing on oneself (i.e., be lessself-oriented) and the more one is likely to take othersinto account (i.e., be more other-oriented) (Pratt et al.2012). This should make one more likely to engage inprosocial behavior (Mikulincer et al. 2005). Thus wehave the following:

Hypothesis 2 (H2). Identity configurations characterizedby high mutual enhancement between an individual’sprosocial, collective, and individualistic identities will be

Dow

nloa

ded

from

info

rms.

org

by [

128.

103.

149.

52]

on 1

9 Se

ptem

ber

2017

, at 1

1:14

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


associated with higher levels of prosocial behavior thanconfigurations in which all three identities are mutuallyconflicting.

Focused EnhancementWhile the above hypothesis is both intuitively appeal-ing and supported by research, a closer examinationof the types of identities under consideration suggeststhat a focused enhancement identity configurationmay lead to even higher levels of prosocial behavior.Enhancement in this configuration is relatively greaterbetween the prosocial and collective identities, withthe individualistic identity unrelated to the other two.We explain why this configuration should exhibit thehighest levels of prosocial behavior by unpacking howenhancement between different identity types influ-ences prosocial behavior.

Prosocial–Collective Identity Enhancement. Prosocialand collective identities compel a person, in one wayor another, to focus on the welfare of others (Brickson2013, Penner et al. 2005). Hence, configurations inwhich prosocial and collective identities enhance oneanother are likely to foster prosocial behavior becausethey motivate and mutually reinforce other-orientedactions. For example, imagine a lawyer (collectiveidentity) who is also a volunteer (prosocial identity).Enhancement between her lawyer and volunteer iden-tities might well encourage her to seek out and takeon pro bono work: her lawyer identity motivates her tocontribute on behalf of her profession, and her volun-teer identity motivates her to contribute her services toclients who cannot afford to pay for legal services.

Individualistic–Prosocial and Individualistic–Collec-tive Identity Enhancement. Adding to the mix en-hancement between one’s individualistic identity, onthe one hand, and one’s prosocial and collective iden-tities, on the other—i.e., full mutual enhancement—islikely to curtail prosocial behavior relative to a con-figuration in which prosocial and collective identitiesare unrelated to the individualistic one—i.e., focusedenhancement. Individualistic identities focus a personon his or her personal goals and achievements (Brewerand Gardner 1996, Brickson 2013). As a result, whenan individualistic identity enhances an other-orientedidentity, the other-oriented identitymay be co-opted bythe individualistic identity’s self-oriented pursuits. Forexample, imagine a lawyer (collective identity) who isa volunteer (prosocial identity) but who also definesherself as a high-achieving legal expert (individualidentity). While enhancement between her lawyer andvolunteer identities would foster her interest in volun-teering, she might only take on pro bono work thatwould advance her own skills. Likewise, enhancementbetween her lawyer and expert identities might propelher to take on only those pro bono cases that would

increase her standing in the profession. Thus, con-figurations with prosocial–collective identity enhance-ment in which these other-oriented identities are alsoenhanced by individualistic identities may result insome prosocial behavior, but the individualistic iden-tity, because of its self-serving orientation toward help-ing, may function as a countervailing or limiting forceon prosocial behavior.

Research on self-serving motives for prosocialbehavior supports the argument that individualisticidentity enhancement may ultimately suppress proso-cial behavior by activating a more self-serving orien-tation toward helping, though this idea has not beendirectly tested. This work identifies a number of ego-istic, self-serving reasons for engaging in prosocialbehavior, including improving one’s own psychologi-cal or material state, gaining career or knowledge bene-fits, and making a good impression (Batson et al. 1983,Clary et al. 1998, De Dreu and Nauta 2009).

Importantly, this research suggests that while self-serving motives to engage in prosocial behavior canencourage a certain amount of prosocial behavior,they may also limit it, especially in comparison toother-oriented motives. For instance, when peopleview others in need, those motivated by a desire toalleviate their own distress engage in more limitedhelping behavior than those motivated by empathy(Batson et al. 1983). Engaging in prosocial behavior tofeel intrinsic pleasure can also divert attention awayfrom the goal of prosocial activities, thus limitinghow much help one actually gives to those in need(Andreoni 1990, Bales 1996, Grant 2008). In work set-tings, a common self-serving motive for helping othersis impression management (Rioux and Penner 2001).When people are motivated to help others to make agood impression on their boss or colleagues and gainrewards such as praise or promotions, their proso-cial behavior is less extensive than when they aremotivated by other-oriented reasons (Grant and Mayer2009). In sum, configurations in which prosocial andcollective identities are enhanced by an individualisticidentity (i.e., the mutual enhancement configuration)may invoke a self-serving orientation toward helping,ultimately resulting in lower prosocial behaviors rel-ative to configurations in which there is prosocial-collective identity enhancement but individualisticidentity is unrelated (i.e., the focused enhancementconfiguration). Thus, we have the following:

Hypothesis 3 (H3). Identity configurations characterizedby high focused enhancement—prosocial and collectiveidentities are mutually enhancing but unrelated to anindividualistic identity—will be associated with higher lev-els of prosocial behavior than configurations in which allthree identities are mutually enhancing.

Dow

nloa

ded

from

info

rms.

org

by [

128.

103.

149.

52]

on 1

9 Se

ptem

ber

2017

, at 1

1:14

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


Overview of StudiesWe use a mixed-method research design to testour hypotheses because research on multiple iden-tity configurations is still emerging (Edmondson andMcManus 2007). In Study 1, we take an exploratoryapproach to understanding identity configurations,using qualitative and quantitative data from a lon-gitudinal field study. In Study 2, we extend thegeneralizability of our results by examining identityconfigurations and prosocial behavior in a team set-ting. In Study 3, we test causal relationships using anexperimental design and also examine logically gen-erated configurations that did not emerge in our fieldstudies. In Study 4, we experimentally examine self-serving motives as an underlying mechanism for thehypothesized effect of focused versus mutual enhance-ment on prosocial behavior.

Study 1Research Setting, Sample, and ProcedureOur setting involves two nonprofit organizations inIsrael that raise funds for charities through bike rides.The first, Alyn Children’s Hospital (Alyn), involvesapproximately 300 Jewish riders who collectively raiseover $2 million per year. The second, the AravaInstitute for Environmental Studies (Arava), involvesapproximately 200 Jewish riders who raise close to$700,000 per year. Riders in both events cycle roughly60 miles per day, pay their own transportation costs(approximately $1,000 for North Americans) and regis-tration fees ($400), and are required to raise aminimumamount, which is $2,500 for Alyn and $3,600 for Arava.Event participation and fundraising represent majorinvestments of time and money and suggest a strongcommitment to charity, cycling, and Jewish ethnicity.To verify the appropriateness of our setting for

testing hypotheses concerning the rich dynamics ofmultiple identities, we conducted 33 semistructuredinterviews with 25 informants (8 were interviewedpre- and post-ride), prior to fielding our survey. Infor-mants’ ages ranged from 40 to 65 years old; 44% werefemale. Interviews lasted from 30 to 60 minutes andwere taped and transcribed by the authors. The inter-view protocol covered the participants’ motivationsfor joining the ride, their identities, self-perceptions ineach domain, and their training and fundraising expe-riences. Our interview data illustrate the salience ofmultiple identities to our informants and reveal rela-tionships of conflict and enhancement between thoseidentities (see online appendix for illustrative exam-ples). Quantitative data were collected using pre- andpost-ride, 20-minute, web-based surveys. In 2007, thesurvey was distributed to 281 Alyn riders, and in 2014,it was distributed to 321 Alyn riders and 160 Arava rid-ers. A total of 162 participants completed the surveys

(response rates: Alyn 1 � 37%, Alyn 2 � 40%, Arava �23%). The response rates are within the commonlyreported range for longitudinal organizational fieldresearch (Roth and BeVier 1998). We combine the datafrom all three rides and control for the year the surveywas taken and the beneficiary organization (Alyn orArava).

MeasuresIndependent Variables.Identities. In 2007, participants’ identities were mea-sured using Bergami and Bagozzi’s (2000) overlap-ping circles scale. Participants were asked to choosethe extent of overlap between a pair of circles—one representing themselves and the other represent-ing the target identities: prosocial (altruist), individ-ualistic (cyclist), and collective (Jewish). Eight pairsof visual circles were rated on a scale from 1 (farapart) to 8 (completely overlapping), and the meanswere as follows: prosocial (altruist) identification (M �

5.7, SD � 1.68); individualistic (cyclist) identification(M � 4.4, SD � 1.84); and collective (Jewish) identifica-tion (M � 6.5, SD � 1.7).In 2014, participants’ identities weremeasured using

three items of the Luhtanen and Crocker (1992) iden-tity subscale: “this identity is important to how I seemyself”; “this identity is an important reflection ofwhoI am”; and “this identity has very little to dowithwho Iam” (reverse coded). These items were rated on a scalefrom 1 (strongly disagree) to 5 (strongly agree), foreach identity, and the means were as follows: prosocial(altruist) identification (M � 3.66, SD � 1.09); individu-alistic (cyclist) identification (M � 3.31, SD � 1.28); andcollective (Jewish) identification (M � 4.58, SD � 0.87).We standardized the variables before combining themacross samples.1

Identity configurations. Our focal independent vari-able was ameasure of multiple identity configurations.We used cluster analysis to identify the configurationsbecause it is a commonly used technique to reducecomplex data in a manageable way and to buildtheory inductively by identifying common characteris-tics among units (Punj and Stewart 1983, Sonensheinet al. 2014).

To identify the clusters, we first measured iden-tity conflict and enhancement between all threepairs of identities: prosocial–collective, individualistic–collective, and individualistic–prosocial. Measureswere adapted from existing multiple identity scales(Benet-Martinez and Haritatos 2005, Brook et al. 2008).The item stem was as follows: “Of the times whenyou think of yourself as a [X] and [Y], how often doyou think . . . ” which was then followed by three con-flict and three enhancement statements for each dyad(counterbalanced). The three conflict items were as fol-lows: “I struggle to maintain an [X] and [Y] way of

Dow

nloa

ded

from

info

rms.

org

by [

128.

103.

149.

52]

on 1

9 Se

ptem

ber

2017

, at 1

1:14

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


Figure 1. Study 1 Configurations of Multiple Identities—Empirical Results

1. Mutual enhancement configuration

5.3

7.0 6.4

Cyclist

Jewish Altruist

4.8 5.1

6.2

2.6 1.7

2.2

2. Independent configuration

3.8

5.9 5.0

Cyclist

Jewish Altruist

1.31.4

1.8 1.1

1.4

3.3

3. Mutual conflict configuration

4.5

6.4 5.5

Cyclist

Jewish Altruist

3.33.2

5.1

2.7 2.2

2.6

4. Focused enhancement configuration

4.3

6.9 6.5

Cyclist

Jewish Altruist

1.9 2.3

5.6

1.3 1.1

1.5

Strength of relationship = Width of line; Strength of identity = Size of circleKey: Enhancement = Conflict =

Note. Means presented are from Alyn Ride 2007 sample for illustrative purposes.

doing things”; “Being a ‘good’ [X] interferes with beinga ‘good’ [Y]”; and ”I feel a [X] way of doing thingsand a [Y] way of doing things are opposed.” The threeenhancement items were “I am a better [X] becauseof my [Y] identity”; “I rely on both [X] and [Y] wayof doing things”; and “I appreciate being an [X] morebecause I am a [Y].” Responses were on a scale from1 (never) to 7 (always). The reliabilities for the iden-tity conflict and identity enhancement scales rangedfrom 0.7 to 0.8 for both identity conflict and identityenhancement across all pairs. We averaged the threeitems for each pair of identities.

For robust cluster analysis, Punj and Stewart (1983)recommend K-means clustering, and theoretically pre-determining the number of clusters. Relying on priorresearch suggesting that conflict and enhancementare independent dimensions (Tiedje et al. 1990),2we specified four clusters that are based on fourpossible patterns of multiple identity relationships:

low enhancement–low conflict, low enhancement–highconflict, high enhancement–low conflict, and highenhancement–high conflict.3 Figure 1 displays the fourconfigurations resulting from the cluster analysis.

While cluster analysis is “quantitative,” it remainsan inductive technique calling for the careful interpre-tation of clusters (Punj and Stewart 1983). Here, theinterpretation needed to take into account the meaningof the identities along with their pattern of relation-ships. The enhancement-conflict patterns of the fourclusters showed that one cluster most resembled aconfiguration characterized by enhancement across allthree pairs, with relatively low conflict; we labeled thismutual enhancement. Another was characterized by lowlevels of both enhancement and conflict; we labeled itindependent. A third was characterized by high levels ofconflict, with some enhancement; we labeled it mutualconflict, and the fourth cluster displayed a pattern offocused relationships with strong enhancement of the

Dow

nloa

ded

from

info

rms.

org

by [

128.

103.

149.

52]

on 1

9 Se

ptem

ber

2017

, at 1

1:14

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


Table 1. Study 1 Mean Scores for Conflict and Enhancement for Identity Dyads by Configuration

Individualistic– Prosocial– Individualistic– Average ofcollective identities collective identities prosocial identities 3 identity dyads

Enhancement Conflict Enhancement Conflict Enhancement Conflict Enhancement ConflictConfiguration (1) (2) (3) (4) (5) (6) (7) (8)

Mutual enhancement 0.82 −0.09 0.40 −0.14 1.07 −0.09 0.76 −0.11(1.06) (0.60) (0.79) (0.56) (0.88) (0.59) (0.57) (0.43)

Independent −0.54 −0.16 −0.92 −0.37 −0.78 −0.48 −0.75 −0.34(0.90) (0.91) (1.07) (0.52) (0.66) (0.28) (0.52) (0.44)

Mutual conflict 0.60 1.13 0.14 0.82 0.11 1.40 0.28 1.11(0.68) (1.59) (0.82) (1.19) (0.57) (1.49) (0.44) (0.93)

Focused enhancement −0.55 −0.34 0.40 0.01 −0.55 −0.40 −0.23 −0.24(0.46) (0.45) (0.73) (1.22) (0.54) (0.19) (0.35) (0.48)

Notes. Standard deviations are in parentheses. Number of observations are as follows: mutual enhancement � 53; independent � 39; mutualconflict � 23; focused enhancement � 47; total � 162. Because of the combined samples, the presented means and standard deviations arestandardized values.

collective–prosocial dyad, but relatively low enhance-ment of the individualistic identity; we labeled thispattern focused enhancement. We describe each config-uration in greater detail below (see Table 1 for meanscores for conflict and enhancement by configuration).Illustrative examples are from our qualitative data.1. Mutual enhancement configuration. The distin-

guishing characteristic of this configuration is that allthree identity dyads exhibit high levels of enhance-ment. In comparison to the other configurations, wecan observe this in the average enhancement score,which was the highest of the four configurations(Menh � 0.76) (Table 1, Column 7). The prosocial–collective enhancement score (Mpro–coll enh � 0.40) in thisconfiguration and in the focused enhancement con-figuration were identical (Table 1, Column 3). How-ever, the prosocial–collective enhancement score wasthe lowest of the three enhancement scores in this con-figuration (Table 1, Columns 1 and 5). There was also amoderate degree of identity conflict in this configura-tion (Mconf �−0.11) (Table 1, Column 8).

In the qualitative data, we saw this configurationexpressed by participants who felt that all three facetsof the ride were complementary. They found it difficultto rank order the various aspects of the ride. Instead,they spoke of a “perfect fit” among the components:

I’m participating because it is something that in everyaspect is close to my heart. It’s Eretz Israel [Land ofIsrael], Tzedakah [charity] and Ofnayim [bike], so mythree passions. (Interview #13)I think that there’s very little gap in the importance ofeach of them. I think it’s a confluence rather than aranking. It’s a convergence of events that make it forme, what makes it happen. . . . If any one of those rea-sons weren’t there, it would diminish the focus for me.(Interview #18)

These quotes illustrate the participants’ sense thateach identity domain was unique (e.g., my three pas-sions, any one of those reasons) but each positively

reinforced the other (e.g., convergence) and that theircoming together formed a whole that was greater thanthe sum of its parts.

2. Independent configuration. The distinguishing char-acteristic of this cluster was the low level of both con-flict and enhancement between identities. In essence,the three identities did not interact with one anotherpositively or negatively, but rather they operated rel-atively separate from one another. In comparison tothe other configurations, this configuration rankedthe lowest of the four in terms of both enhancement(Menh � −0.75) and conflict (Mconf � −0.34) (Table 1,Columns 7 and 8).

In our qualitative data, we saw this configurationexpressed by participants who described their iden-tities as fairly distinct, with the event providing anopportunity to enact each identity separately. In con-trast to the mutual enhancement group, who had dif-ficulty responding to questions about how they wouldrank their identities, participants within this configu-ration could rank their identities. This configurationwas, therefore, consistent with dominance of identi-ties as noted in the literature (Ashforth 2007, Pratt andForeman 2000, Roccas and Brewer 2002):

I think Israel is number one and I guess number tworeally is doing the charity work in a way that’s a per-sonal and a physical challenge for me. I think it is Israel,charity and a physical bike ride. (Interview #21)

It [is] the biking, then Israel, and then the fundraising.(Interview #14)

As the quotes above suggest, the most importantidentity domain differed from person to person; forsome it was the collective identity, for others it was theindividualistic one. Thus, the common experience forpeople with this configuration was fairly little conflictor enhancement between their identities, suggestingthat all three identities were independent.

Dow

nloa

ded

from

info

rms.

org

by [

128.

103.

149.

52]

on 1

9 Se

ptem

ber

2017

, at 1

1:14

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


3. Mutual conflict configuration. The distinguishingcharacteristic of this configuration is the high levelof conflict. These participants reported high conflictbetween all three identity dyads. Comparatively, theaverage conflict for this configuration was the highestof the four configurations (Mconf � 1.11) (Table 1, Col-umn 8). It is notable that participants in this cluster alsoexperienced some enhancement (Menh � 0.28) (Table 1,Column 7).In our interviews, people who most resembled this

configuration were individuals feeling torn betweenthe demands of various identities. One female ridernoted, for example, the conflict between her individ-ualistic (cyclist) and collective (Jewish) identities andhow she was debating which to prioritize:

When I go into Jerusalem, am I going to be wearingshorts or am I going to be wearing a skirt on top, Idon’t know. It’s kind of an issue, like here [in NorthAmerica] I ride in shorts but I don’t know about EretzIsrael . . . .When you’re in Jerusalem, the holy city, we’regoing to be riding around, they’ll be all kinds of Yeshiva[religious education] students there, I might put on askirt for the Jerusalem ride . . . I don’t think it will slowme down . . . that’s the thing, I don’t know. I might weara shorter skirt, you know just a skirt that will cover myknees, or maybe riding pants. (Interview #20)

She was also struggling to reconcile her desire to goon the ride, which was driven by her collective identity(“There’s something in my nature that draws me to theLand of Israel”), with her desire to support the causeas much as possible, which was driven by her prosocialidentity (“I donate to all kinds of organizations. I giveway more than 10 percent of my income to Tzedakah[charity] every year”). For example, although she wasgoing on the ride, she was still questioning the moneyspent on the trip, and whether or not it was a “waste”compared to giving more to the organization:

I spent more money going there than I raised. If youthink about it, the plane ticket is 1,500 dollars and thenyou get to Israel and you’re going to spend at least 500dollars hanging around there and then you stay an extratwo weeks and that’s at least an extra 1,000 dollars . . . .I wasted or spent more money than I raised. (Inter-view #20, continued)

While people with this configuration exhibitedgreater conflict relative to other clusters, they alsoexperienced some enhancement between identities.For example, the above woman also described how ful-filling one aspect of herself enhanced her fulfillmentwith another:

Bicycle riding—I really enjoy it . . . . Today, I cycleddowntown . . . . See my little hole in my arm. I donateblood platelets for cancer patients, so that’s why I havea hole in my arm. I rode downtown and I gave blood,takes about three hours, and then I rode for about two

and a half hours after I gave blood. I thought that wasvery easy. I did the three hours today, nothing. (Inter-view #20, continued)

4. Focused enhancement configuration. In this clus-ter, participants experienced their prosocial and col-lective identities as enhancing one another and theirindividualistic identity as more separate, neither con-flicting nor enhancing. Compared to other configu-rations, focused enhancement ranks just above theindependent configuration in both average enhance-ment (Menh � −0.23) and conflict (Mconf � −0.24)(Table 1, Columns 7 and 8). However, this configurationis distinguished by high prosocial–collective enhance-ment (Mpro–coll enh � 0.40) relative to the individualis-tic identity dyads (Mind–coll enh �−0.55 and Mind–pro enh �

−0.55, Table 1, Columns 1 and 5). This is unlike themutual enhancement configuration, in which the samelevel of prosocial–collective enhancement (Mpro–coll enh�

0.40) was relatively lower than the two individualisticidentity dyads (Mind–coll enh � 0.82 and Mind–pro enh � 1.07,Table 1, Columns 1 and 5).

In our qualitative data, we saw focused enhancementamong participants who spoke about the enhancementproduced through being Jewish and altruist, whilefeeling low conflict and enhancement with their cyclistidentity. For example, the descriptions of one womanabout her contributions illustrate theways inwhich herprosocial identity and her collective identity mutuallyenhance one another:

I don’t feel like I’m taking from Israel but I’m giving. Notjustmy tourist dollars butmyself . . . and there’s the issueof . . .giving to an organization that’s a worthy organiza-tion. I think there’s a really good feelingwith that. If youreally believe in something, an organization or a charityor a cause, and you really work hard and contribute toit, it’s a very satisfying feeling you know. . . . I feel verystrongly that it’s a very good organization so I feel verysatisfied in working hard for it. (Interview #12)

Another participant noted the enhancement ofprosocial and collective identities and pointed to theimportance of a prosocial cause that served more thanthe collective he was part of:

I feel I’m supporting a hospital plus . . . it’s a Jewish hos-pital but I know they treat everyone . . . I know that theytreat not only Jewish children but Arab children as well.To me, it is great. They should treat everyone who hasany type of problem. But the fact that it’s in Jerusalem,it’s in Israel, has everything to do with the work I’mdoing for Alyn . . . [and] when I see these children whohave these terrible problems, it gives me extra energy totry and work for them. (Interview #15)

Despite his focus on Jewish and altruist identities, hewas not ignoring his cyclist identity; indeed, in describ-ing himself he noted, “[As a cyclist], I would say I amsort of untamed, individualistic, and a little reckless

Dow

nloa

ded

from

info

rms.

org

by [

128.

103.

149.

52]

on 1

9 Se

ptem

ber

2017

, at 1

1:14

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


even . . . I take a lot of risks that maybe people may nottake when it comes to dodging traffic and doing offroad riding and some dangerous stuff.” Yet his cyclingexperience was not necessarily enhanced by his proso-cial or collective identities. He said, “it’s great to ridethere, I love it, but as for a physical challenge . . . it willnot [be] a huge one.”In sum, we found four configurations of multi-

ple identities in this setting: mutual enhancement, inwhich all identities enhanced one another; indepen-dent, in which all identities were largely separate fromone another with both low conflict and low enhance-ment; mutual conflict, distinguished by strong conflictamong identities; and focused enhancement, in whichthe collective–prosocial identities were enhancing andthe individualistic identity was less enhanced.We usedthese four identity configurations as independent vari-ables to test our hypotheses about the effects of config-uration on prosocial behavior.

Dependent Variable.

Prosocial behavior. We operationalized prosocial be-havior as the self-reported amount of money (in USD)raised by each participant for their respective charityorganizations. Prior research on prosocial behavior hasused the amount raised as a meaningful criterion infundraising contexts (Grant and Sumanth 2009). Ourinterviewswith ride organizers suggested that this wasalso an important outcome for the organizations. Themean amount raised was 6,509 USD.

Control Variables.Our control variables include the following: gender(binary) and age (continuous) of riders, survey year(2007 or 2014), and the ride (Alyn or Arava). We alsocontrolled for prior experience with similar activities:biking experience—respondents’ self-reported cyclingexpertise (1 � novice to 5 � experienced), which wedichotomized as 1 (expert cyclists) and 0 (novice andintermediate cyclists); as our qualitative data sug-gested, novice and intermediate riders were more sim-ilar to one another (they tend to take on-road routes)compared to experts (who took off-road routes). Altru-ist experience was measured by asking participantsapproximately how much of their annual income theydonated to charity on a scale from 1 (<1% of income)to 5 (>10% of their income). Charity ride experiencewas measured by the number of times participantshad been on their respective bike rides (never, once,two to three times, more than three times), which wedichotomized as 1 (three times or more on the ride)and 0 (less than three times). Our interviews withride organizers suggested that people who went sev-eral times were more likely to return than those whowent just a few times. (See Table 2 for descriptives andcorrelations.) Table2.

Stud

y1Su

mmaryStatistic

s

Varia

bles

12

34

56

78

910

1112

1314

15

1Gender(

1�Fe

male)

—2

Age

−0.1

3—

3Su

rvey

Year

(1�

2014)

−0.0

80.

19∗

—4

Ride

(1�Alyn)

−0.0

3−0.2

3∗∗−0.4

2∗∗∗

—5

Charity

ridee

xperien

ce(1≥

3)−0.2

4∗∗

0.13

0.16∗

0.22∗∗

—6

Biking

experie

nce(

1�Ex

pert)−0.2

0∗0.

15∗−0.0

70.

100.

18∗

—7

Altr

uistexperie

nce

0.02

0.10

−0.0

60.

16∗

0.06

0.07

—8

Individu

alisticidentity(cyclist)−0.0

70.

120.

06−0.0

30.

14∗

0.35∗∗∗

0.06

—9

Prosocialidentity

(altr

uist)

−0.0

60.

01−0.0

20.

040.

040.

070.

35∗∗∗

0.28∗∗∗

—10

Collectiveidentity

(Jewish)

0.02

−0.0

4−0.0

1−0.0

8−0.0

2−0.1

5∗0.

140.

33∗∗∗

0.57∗∗∗

—11

Mutua

lenh

ancement

0.08

−0.0

80.

18∗

−0.2

5∗∗

0.01

−0.0

50.

000.

28∗∗∗

0.14

0.28∗∗∗

—12

Independ

ent

0.08

0.08

−0.2

3∗∗

0.10

−0.0

00.

00−0.1

7∗−0.3

5∗∗∗−0.4

7∗∗∗−0.5

3∗∗∗−0.3

9∗∗∗

—13

Mutua

lcon

flict

−0.0

1−0.1

3∗−0.2

0∗∗

0.10

−0.1

20.

03−0.0

00.

02−0.0

20.

04−0.2

8∗∗∗−0.2

3∗∗

—14

Focusedenhancem

ent

−0.1

50.

110.

18∗

0.09

0.09

0.03

0.17∗

0.03

0.31∗∗∗

0.19∗

−0.4

5∗∗∗−0.3

6∗∗∗−0.2

6∗∗

—15

Amou

ntraise

d−0.0

80.

08−0.2

2∗∗

0.20∗∗

0.18∗

0.02

0.18∗

0.05

0.15

0.03

−0.1

7∗−0.1

00.

050.

23∗∗

—Descriptiv

estatistic

sMean

0.27

52.3

60.

630.

770.

260.

363.

36−0.0

60.

080.

090.

330.

240.

140.

296,

509.

18SD

0.45

11.5

30.

480.

420.

440.

481.

171.

011.

010.

970.

470.

430.

350.

466,

828.

68

Note.

Num

bero

fobserva

tions

�16

2.∗ p<

0.05

;∗∗ p<

0.01

;∗∗∗

p<

0.00

1.

Dow

nloa

ded

from

info

rms.

org

by [

128.

103.

149.

52]

on 1

9 Se

ptem

ber

2017

, at 1

1:14

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


ResultsAn analysis of variance (ANOVA) of funds raised byconfiguration type was significant (F3,158 � 3.78, p �

0.01). In support of Hypothesis 3, participants with thefocused enhancement configuration raised the mostmoney (M � 8,993.15, SD � 10,988.27). This amountwas significantly greater than the mutual enhance-ment configuration (M � 4,846.11, SD � 2,971.03), (t �3.11, p < 0.01) and the independent configuration (M �

5,330.96, SD � 3,285.91), (t � 2.54, p < 0.05). However,the focused enhancement configuration was not signif-icantly greater than the mutual conflict configuration(M � 7,263.35, SD � 5,136.04), (t � 1.02, p � ns), and themutual enhancement, mutual conflict, and indepen-dent configurations were also not significantly differ-ent from one another, indicating H1 and H2 were notsupported.Because there is over-dispersion in the dependent

variable (indicated by the variance being greater thanthe mean), we further tested our hypotheses using anegative binomial regression. The regression results(see Table 3) indicate that the configurations have asignificant effect on prosocial behavior. The key con-figuration variables are significant after accountingfor the control variables as well as the individualeffects of each identity (Models 1 and 2). Further-more, in Model 3, using the mutual enhancementconfiguration as the reference category we see thatonly the focused enhancement configuration has agreater effect on the amount of money raised (β �

0.44, p < 0.01). The amount of money raised by theindependent (β � −0.26, ns) and mutual conflict (β �

0.20, p � ns) configurations are not significantly differ-ent from the mutual enhancement configuration.4 Insum, the focused enhancement configuration exhibitedgreater prosocial behavior than the mutual enhance-ment configuration.

Discussion. Study 1 provides several important in-sights regarding multiple identity configurations andprosocial behavior. First, we observed that participants’experiences of conflicting and enhancing relationshipsbetween the three identities could be meaningfullycaptured both quantitatively and qualitatively. Weidentified four configurations: independent, mutualconflict, mutual enhancement, and focused enhance-ment. Second, as hypothesized, we found that thefocused enhancement configuration (in which theprosocial–collective identities were mutually enhanc-ing but both were unrelated to the individualistic iden-tity) resulted in higher levels of prosocial behavior thanthe mutual enhancement configuration. This suggeststhat the type of identity being enhanced may matter.

Self-serving motives for prosocial behavior. To fur-ther understand why the mutual enhancement con-figuration had lower levels of prosocial behavior

Table 3. Study 1 Negative Binomial Regression forFunds Raised

Model 1 Model 2 Model 3Variables B (SE) B (SE) B (SE)

ControlsGender (female) 0.08 0.06 −0.03

(0.13) (0.13) (0.12)Age 0.01∗∗ 0.01∗∗ 0.01∗∗

(0.00) (0.00) (0.00)Survey year (2014) 0.58∗∗∗ 0.60∗∗∗ 0.70∗∗∗

(0.14) (0.14) (0.13)Ride (Alyn) −0.19 −0.18 −0.02

(0.13) (0.13) (0.13)Charity ride experience (>3) −0.45∗∗ −0.44∗∗ −0.46∗∗

(0.17) (0.16) (0.15)Cyclist experience (expert) 0.10 0.22 0.27

(0.14) (0.15) (0.14)Altruist experience 0.13∗∗ 0.11∗ 0.09

(0.05) (0.05) (0.04)Identities

Individualistic (cyclist) 0.11 0.11(0.08) (0.07)

Prosocial (altruist) 0.09 0.04(0.08) (0.06)

Collective (Jewish) −0.10 −0.20∗(0.08) (0.08)

Identity configurationsIndependent −0.26

(0.17)Mutual conflict 0.20

(0.15)Focused enhancement 0.44∗∗

(0.17)Constant 7.73∗∗∗ 7.71∗∗∗ 7.67∗∗∗

(0.37) (0.36) (0.32)Log likelihood −1,543.75 −1,541.80 −1,538.46

Notes. N � 162. Reference category is the mutual enhancement con-figuration. When focused enhancement configuration is the refer-ence category, the coefficients are as follows: mutual enhancement(β � −0.44, SE � 0.17, p < 0.01); independent (β � −0.70, SE � 0.21,p < 0.01); mutual conflict (β �−0.25, SE� 0.20, p � ns).∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; two-tailed tests.

relative to the focused, we returned to our qualita-tive data. Quotes from our interviews suggest that themutual enhancement configuration may contain a self-serving orientation. For instance, mutual enhancementis described by a participant as follows:

It’s this unique chemistry . . .where the light bulb wasturned on and I said yeah, that sounds good. So it’s var-ious components and I’m not sure what ranks highest.(Interview #22)

This interviewee further admitted his self-serving ori-entation. He continued,

There is something self-serving about thewhole processfor me. If it wasn’t a bike ride and just money I maybewouldn’t be giving Alyn 2,000 dollars. But on top of that

Dow

nloa

ded

from

info

rms.

org

by [

128.

103.

149.

52]

on 1

9 Se

ptem

ber

2017

, at 1

1:14

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


there’s the expense of the hotel and the adjustments forthe bike and the flight and so on. So it’s way more thanjust the charitable contribution I’ve given to Alyn. Sothere is something self-serving about it and while Israelmay call me, so does the bicycle. So it’s a fun thing.(Interview #22, continued)

Similarly, the mutual enhancement rider that wequoted previously as feeling a “confluence” and“convergence” (interview #18) was highly personallyinvested in going. He stated, “something happened lastweek and I was thinking I might not be able to go andI was very disappointed personally. If I could predictthe future, I would like to do this for as long as I can foras often as I can.” He further described his reluctanceto raise as much money as he possibly could because itmight compromise his ability to go on the ride again:

I got an e-mail from the chairman of Alyn . . . and it sayshe got so overwhelmed by Alyn and what they do andhowgreat it is that he decided to surpass his fundraisingminimum of 2,000 and I thought I was going to read4,000, 5,000, and he said he raised 25,000 dollars last yearand he encouraged everyone to do that. I’m not going todo that. I probably could if I tried but then I would neverbe able to raise money again because I would basicallytap every source I had. I’m trying to do it in a way sothat if I want to do it again next year, and I will want to,then I can go back and raise a similar amount. I decidedto raise 5,000 dollars and I’m there. (Interview #18)

For this rider, fulfilling his personal desire to ride thefollowing year limited the amount of money he waswilling to raise in the short term for the organization.These examples suggest that the smaller amounts ofmoney raised by those with the mutual enhancementconfigurationmight be the result of a more self-servingmotive for participating, namely promoting one’s ownpleasure and fulfillment.

In contrast to H3, we did not find support for H1and H2; the mutual conflict configuration was not dif-ferent from either the focused or mutual enhancementconfigurations. This may be because of the voluntarynature of the ride. Those experiencing pure conflictamong their identities could have self-selected out ofthe event, possibly limiting the level of conflict. Infact, we chose this setting because it allowed us togain deeper insight into the understudied dynamics ofenhancement (Dutton et al. 2010). However, it wouldbe useful to examine settings in which identity conflictis likely to be more salient.

In addition to the lack of conflict, Study 1 is alsoin a unique context. The Jewish identity is a collectiveidentity containing religious and cultural components(Berger and Gainer 2002). In addition, as with otherreligions (Shariff and Norenzayan 2007), the religiousaspects of Jewish identity may overlap with the altru-ist identity. Further, the individualistic identity, cyclist,

was related to a leisure activity. Therefore, examin-ing other instantiations of the three types of identitiesand prosocial behavior in a nonvoluntary, work-relatedcontext would be useful. Accordingly, in Study 2 weexamined prosocial behavior in a team setting (Tylerand Blader 2003).

Study 2In Study 2, our goal was to replicate and extend thegeneralizability of our findings in a team context witha different operationalization of individualistic, collec-tive, and prosocial identities.

Sample and ProcedureParticipants were 70 executiveMBA students at a Euro-pean business school enrolled in a course on corporatesocial responsibility taught in English. Respondentswere 77% male with an average age of 35 and wereworking full time in their respective organizations.Data were gathered via a questionnaire that studentswere required to complete as part of their coursework.Students had been assigned to study teams of four tosix members at the start of the semester with ongo-ing teamwork as a required element of the curricu-lum. Consequently, teams were meaningful collectivesin which work was accomplished and collective iden-tities were built. We focused on identities likely tobe salient for participants in this setting: individualis-tic (high-achieving student), collective (team member),and prosocial (helper of their teammates unrelated togroup tasks). Survey items asked about conflict andenhancement between each pair of identities, whichwe used to construct our independent variable, iden-tity configurations. We also asked participants to ratethe prosocial behavior of other study group mem-bers, which served as our dependent measure. Wematched individual survey responses to their team-mates’ ratings.

MeasuresIndependent Variables.Identity configurations. Identity configurations againserved as the independent variable and were con-structed in the following three steps:

(1) Identities. First, as in Study 1, participants ratedthe extent to which they identified with each of thethree identities using the “circles overlap” measure(Bergami and Bagozzi (2000); Mind � 5.53 (SD � 1.43);Mcoll � 5.73 (SD � 1.44); Mpro � 5.47 (SD � 1.88)).(2) Identity conflict and enhancement. Second, similar

to the procedures of Study 1, participants were askedto indicate the extent to which they experienced con-flict and enhancement between identities. They rateda single statement about conflict and a single state-ment about enhancement for each pair of identities asfollows: “Think of your [individualistic identity as ahigh achieving student] and your [collective identity

Dow

nloa

ded

from

info

rms.

org

by [

128.

103.

149.

52]

on 1

9 Se

ptem

ber

2017

, at 1

1:14

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


as a member of your study group]: How frequently dothese identities conflict with one another? And howfrequently do these identities enhance one another?”Items were rated on a scale from 1�never to 7� always(Mconf � 2.61 (SD � 0.91) and Menh � 4.80 (SD � 0.89)).(3) Configurations. Reflecting the focus of our

hypotheses and building on the findings of Study 1,we used the identity conflict and enhancement scoresto derive the configurations. Using the same K-meansclustering approach (Punj and Stewart 1983), we againspecified a four-cluster solution. To understand andinterpret the resulting configurations, we looked atboth the inter-cluster and the intra-cluster patterns ofconflict and enhancement. As in Study 1, we labeledthe clusters based on their closeness to our theorizedconfigurations:

(a) Focused enhancement (FE): This configurationwas characterized by having the highest level ofprosocial–collective identity enhancement and lowidentity conflict (Mconf � 2.61 (0.99); Mpro–coll enh � 6.32(0.48)).

(b) Mutual enhancement (ME): This configurationwas characterized by the lowest level of conflict andequivalent levels of identity enhancement among theidentities (Mconf � 2.19 (0.65); Mpro–coll enh � 5.16 (0.37)).

(c) Mutual conflict (MC): This configuration wascharacterized by the highest levels of conflict andmoderate identity enhancement (Mconf � 3.67 (0.67);Mpro–coll enh � 4.63 (0.52)).

(d) Independent (IND): This configuration wascharacterized by the lowest level of identity enhance-ment coupled with a moderate level of conflict (Mconf �

2.81 (0.77); Mpro–coll enh � 3.67 (0.49)).

Dependent Variable.

Prosocial behavior. We operationalized prosocialbehavior in this team context as voluntary actionsdirected toward helping other teammates’ personalgoals or needs in ways not related to the team’s tasks.Such discretionary helping behavior has been recog-nized as a critical form of prosocial behavior in orga-nizations (Brief and Motowidlo 1986). We specificallychose behaviors that were not related to the core workof the team to distinguish our measure of prosocialbehavior from helping behavior that may be motivatedby personal benefits (such as helping the team per-formwell so one could be a high performer). Thus, thismeasure is consistent with altruistically (versus egois-tically) oriented helping (Batson et al. 1983) and somemeasures of extra-role behavior in teams that empha-size helping behavior when actors do not expect creditfor such behavior (Blader and Tyler 2009).To measure prosocial behavior, participants were

asked to think about their experienceworkingwith [X],who is a teammate, and answer the following question:“Please think about how often [X] does the following

in his/her altruistic role as a person who helps [study]group members on issues NOT related to the group’stask performance.” The items, rated on a scale from1�never to 7� always, were as follows: (1) helps team-mates advance their personal (nongroup-related) goalsand (2) is generally helpful toward others, even whenit is not personally beneficial to him/her. For each par-ticipant, we then used the average of team members’ratings as the dependent variable. Cronbach’s α � 0.85;M � 4.96 (SD � 0.69).

ResultsAn ANOVA of prosocial behavior means by config-uration type was significant (F(3, 66) � 3.03, p < 0.05).Specifically, team members rated those with a focusedenhancement identity configuration as more helpfulto teammates compared to those with other identityconfigurations, thereby exhibiting the greatest proso-cial behavior (MFE � 5.26 (SD � 0.72) > MME � 4.86(SD � 0.51) > MIND � 4.75 (SD � 0.78) > MMC � 4.62(SD � 0.71)). Supporting Hypothesis 3, pairwise con-trasts show that those with the focused enhancementconfiguration exhibited significantly greater prosocialbehavior than those with mutual enhancement (t �

2.13, p < 0.05). Those with focused enhancement alsoexhibited greater prosocial behavior than those withthe mutual conflict (t � 2.40, p < 0.05) and the indepen-dent (t � 2.21, p < 0.05) configurations. Further, consis-tent with Study 1, the mutual enhancement, mutualconflict, and independent configurations were not sig-nificantly different from one another, once again sug-gesting that H1 and H2 were not supported.5

Discussion. The results of Study 2 demonstrate thatthose with a focused enhancement configurationengaged in greater prosocial behavior compared tothose with a mutual enhancement configuration. Fur-thermore, Study 2 extends the generalizability of ourfindings by showing that the identity configurations ofworking adults engaged in task groups function in asimilar fashion to those of the volunteers in Study 1.We also found that mutual conflict exhibited signifi-cantly lower prosocial behavior than focused enhance-ment in this study, suggesting a greater role for conflictin this nonvoluntary setting. However, one limitationof Studies 1 and 2 is that we could only examineidentity configurations that emerged naturally in ourfield settings. As a result, we do not know whetherthe focused enhancement configuration was beneficialspecifically because the prosocial and collective iden-tities were mutually enhancing and the individualis-tic identity was unrelated, or whether any configura-tion with two enhancing identities and the third setapart would have similar effects. In addition, it is pos-sible that the mutual conflict that emerged in thesefield settings was more similar to configurations withboth conflict and enhancement rather than pure con-flict. It would, therefore, be useful to experimentally

Dow

nloa

ded

from

info

rms.

org

by [

128.

103.

149.

52]

on 1

9 Se

ptem

ber

2017

, at 1

1:14

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


manipulate alternative identity configurations to morethoroughly test ourhypotheses. Therefore,wedesignedStudy 3 to constructively replicate and extend the find-ings from Studies 1 and 2 by experimentally inducingwork-related identity configurations.

Study 3In Study 3, we conducted a scenario experiment thatallowed us to test causal relationships and extend gen-eralizability by using a different context and differ-ent operationalizations of identities. We also aimedto compare the focused enhancement configuration tostructurally similar identity configurations that did notemerge in Studies 1 and 2 and to examine a configura-tion of pure mutual conflict.

Sample and ProcedureParticipants were adults recruited through AmazonMechanical Turk (n � 1,399). Participants’ average agewas 35 years (SD � 11.34), 47.2% were female, and allwere located in the United States. Each participant waspaid $1 for the study.The study design was a seven-condition scenario

experiment. The scenario presented an architecturefirm trying to balance multiple goals—profitability,environmental sustainability, and good design—andasked participants to imagine they were employeesin the firm. Participants were told they thought ofthemselves in terms of the following: (1) their indi-vidualistic identities as star “hotshot” employees inthe firm, which oriented them toward their personalsuccess; (2) their collective identities as professionalarchitects, which oriented them toward their profes-sion; and (3) their prosocial identities as environmentalactivists, which oriented them toward the welfare ofwider society.

Participants were also told they were members of avoluntary “fellows” program in the firm and as partof that program they could mentor junior employ-ees; this was meant to be an indicator of prosocialbehavior within the organization. They were told theycould raise money as part of the sustainability effortsof the firm and, as in Study 1, this served as an indi-cator of philanthropic prosocial behavior. The scenariothen presented one of seven identity configurationsdescribed below (see the online appendix) and askedrespondents to imagine the configuration representedtheir own feelings regarding their identities and torespond to the outcomes described below.

MeasuresIndependent Variables.

Identity configurations. Participants were randomlyassigned to one of seven conditions. Three described

the identity configurations reflected in our hypotheses:(1) puremutual conflict, inwhich all three identities areonly conflicting; (2) mutual enhancement, in which allthree identities enhance one another; and (3) focusedenhancement, in which prosocial and collective iden-tities are enhancing and the individualistic identity isseparate.

We included four additional configurations for com-parison purposes: (4) independent, in which all threeidentities are separate (as observed in Study 1);(5) divided, in which all three are conflicting andenhancing. This configuration was to help us comparepure mutual conflict as theorized with the mutual con-flict configuration we empirically observed in Stud-ies 1 and 2, in which there was mutual conflict andsome enhancement. We expected that the mutual con-flict configuration would have lower prosocial behav-ior than this divided configuration. The last twoadditional configurations, (6) individualistic–prosocialenhancement with the collective separate; and (7) in-dividualistic–collective enhancement with the proso-cial separate, were to help us examine whether thebeneficial effect of the focused enhancement configura-tion in Studies 1 and 2 was simply structural (three ver-sus two mutually enhancing identities with the thirdset apart) or specifically due to the lack of enhance-ment of the individualistic identity as we proposed. Ifthe latter, then the focused enhancement configurationshould exhibit greater prosocial behavior than thesetwo additional “partial” configurations as well.Dependent Variable.Prosocial intentions. Two items measured intentionsto behave prosocially:

(1) Intention to mentor young professionals. This itemasked, “How much time do you want to put into men-toring young professionals who could be future fel-lows?” on a scale from 1� none to 7� a lot.

(2) Effort to raise funds. This item asked, “How mucheffort would you devote to raising money for this pro-gram?” on a scale from 1 � none to 7 � a lot. Thesemeasures are consistent with research suggesting thattime and effort given in a prosocial task are indicatorsof prosocial behavior (Grant and Sumanth 2009).

ResultsAn ANOVA examining the means of prosocialintentions by experimental condition shows thatconfiguration type had a significant impact on bothmentoring (F6, 1,392 � 23.76, p < 0.001) and fundraising(F6, 1,392 � 10.46, p < 0.001) (see Table 4 for means).First, consistent with Studies 1 and 2 and in supportof H3, planned comparisons show that participants inthe focused enhancement condition were significantlymore likely to spend time mentoring young employ-ees (t � 2.21, p < 0.05) and expend greater effort inraising funds (t � 2.71, p < 0.01) than those in the

Dow

nloa

ded

from

info

rms.

org

by [

128.

103.

149.

52]

on 1

9 Se

ptem

ber

2017

, at 1

1:14

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


Table 4. Study 3 Analysis of Variance of Prosocial Behaviorby Configurations

Intention to Effort toConfiguration mentor raise funds

Focused enhancement 5.07c 5.23b(prosocial–collective enhancement; (1.26) (1.37)individualistic identity separate)

Mutual enhancement 4.77a 4.80a

(1.29) (1.46)Independent 4.29b 4.61a

(1.32) (1.47)Divided (conflict and enhancement) 4.26b 4.55a

(1.29) (1.55)Individualistic–prosocial enhancement 4.06b 4.72a

(and collective identity separate) (1.55) (1.75)Individualistic–collective enhancement 3.91b,d 4.21c

(and prosocial identity separate) (1.47) (1.67)Mutual conflict (pure) 3.71d 4.19c

(1.53) (1.68)Mean (across configurations) 4.30 4.62

(1.46) (1.60)

Notes. Standard deviations are in parentheses. Number of observa-tions are as follows: mutual enhancement � 197; independent � 208;divided� 202; focused enhancement� 204; individualistic–collectiveenhancement � 203; individualistic–prosocial enhancement � 193;mutual conflict (pure) � 192; and total � 1,399. Means that do notshare a superscript differ at p < 0.05.

mutual enhancement condition. The focused enhance-ment configuration was also significantly greater thanthe two additional partial enhancement configurations(the individualistic–collective enhancement and theindividualistic–prosocial enhancement), both for men-toring and fundraising (all p’s < 0.01). This findingoffers some support for the role of identity type inthe focused enhancement configuration. That is, allthree configurations containing individualistic iden-tity enhancement led to lower prosocial behaviorthan the focused enhancement configuration, sug-gesting that focused enhancement may be beneficialfor prosocial behavior because of the types of iden-tities being enhanced (prosocial–collective) and notenhanced (individualistic). Last, consistentwithH1andH2, the pure mutual conflict configuration resulted insignificantly lower prosocial intentions for both men-toring and fundraising than the mutual enhancement,focused enhancement, and divided configurations (allp’s < 0.05).Discussion. This study constructively replicated thefocused enhancement effect with a set of work-relatedindividualistic and collective identities (e.g., hotshotand professional) and a distinct prosocial identity (e.g.,environmental activist) in a context similar to employeevolunteering initiatives (Bartel 2001). Furthermore, incontrast to the earlier studies, in this study we foundthat the experimental manipulation of pure mutualconflict resulted in lower levels of prosocial behavior

than focused and mutual enhancement. This may bebecause the naturally emergent mutual conflict con-figurations in Studies 1 and 2 may not have been aspurely or as strongly conflicting as the experimentalmanipulation. However, this remains an open questionfor future work. Last, by comparing focused enhance-ment to mutual enhancement as well as the two otherpartial configurations, this study provides some sug-gestive evidence for the potential negative impactof individualistic identity enhancement on prosocialbehavior. However, Study 3 still does not directly testthe proposed mechanism through which the focusedversus mutual enhancement effect operates. Specifi-cally, Hypothesis 3 posited that the weaker proso-cial behavior of the mutual enhancement configurationwas rooted in the enhancement of the individualisticidentity. Such enhancement may amplify self-servingmotives and thereby dampen prosocial behavior. Totest this explanation for Hypothesis 3more thoroughly,we undertook one final experimental study.

Study 4In Study 4, we conducted a scenario experiment toexamine self-serving motives for prosocial behavioras a possible mediator for the focused versus mutualenhancement effect.

Sample and ProcedureParticipants were 397 adults recruited through Ama-zon Mechanical Turk. Their average age was 36.6 years(SD � 12.74), 57.2% were female, and all were locatedin the United States. Participants were paid $2.50 forthe study.

The study design was a two-condition, between-subjects experiment. We created an organizationallyrelevant scenario involving a corporate volunteeringprogram similar to that used in Study 3 and mod-eled after an existing corporate volunteering programat IBM. The setting was a consumer goods firm calledAMI. Participants were told that they had three identi-ties: (1) their individualistic identity as expert employ-ees in the firm, oriented toward their personal careersuccess; (2) their collective identity as AMI employ-ees, oriented toward the welfare of the organization;and (3) their prosocial identity as volunteers, orientedtoward helping beneficiaries. Participants were pre-sented with a configuration and asked to imagine thatit represented their own experience of their identi-ties within AMI (see the online appendix). Participantsresponded to the survey measures described below.

MeasuresIndependent Variables.

Identity configurations. Participants were randomlyassigned to one of two scenario conditions that

Dow

nloa

ded

from

info

rms.

org

by [

128.

103.

149.

52]

on 1

9 Se

ptem

ber

2017

, at 1

1:14

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


operationalized identity configurations: (1) mutualenhancement, in which they were told all three iden-tities enhance one another; (2) focused enhancement,in which they were told their collective and prosocialidentities were enhancing and the individualistic iden-tity was not.Dependent Variables.Prosocial behavioral intentions. We measured proso-cial intentions with two scales: involvement in corpo-rate volunteering and helping behavior as a corporatevolunteer.(1) Involvement in corporate volunteering. This was

measured as a composite of four items: participation,time, energy, and effort devoted to the AMI program.Using a 5-point scale (1 � not likely, 5 � very likely),these items asked, “How likely are you to participate inthe voluntary service activities?” “How much time doyou expect to devote to helping beneficiaries on theirproblems?” “How much energy will you give to help-ing beneficiaries on their problems?” and “How mucheffort will you give to helping beneficiaries with theirproblems?” Cronbach’s α � 0.90, M � 4.13 (SD � 0.75).

(2) Helping behavior. This was measured using sixitems from Blader and Tyler’s (2009) extra-role behav-ior scale adapted to the corporate volunteering con-text: (1) Do things that are not expected of volunteersto help the beneficiaries; (2) Volunteer to help orientnew volunteers; (3) Offer to help other volunteerswhenthey have heavy volunteer workloads; (4) Put an extraeffort into doing your volunteer work well, beyondwhat is normally expected; (5) Share your knowledgewith beneficiaries or other volunteers even when youwill not receive credit; and (6) Work extra volunteerhours even when you will not receive credit for doingso. These items were measured on a 7-point scale(1 � never, 7 � always). Cronbach’s α � 0.93, M � 5.36(SD � 1.14).Mediator.Self-serving motives for prosocial behavior. Scholarshave examined various self-serving motives for engag-ing in prosocial behavior, ranging from pleasure to thedesire to gain career rewards (Batson et al. 1983, Claryet al. 1998). In Study 1, the interviews revealed a self-servingmotive that was relevant in a leisure context (toseek pleasure and personal fulfillment). In this study,we use a self-serving motive that has been examinedin a work context: being helpful to impress one’s col-leagues and managers at work so one can gain work-related material and social rewards (Clary et al. 1998,Grant and Mayer 2009, Rioux and Penner 2001). Thismotive has been measured with a 12-item scale labeledimpression management motives to engage in proso-cial behavior and has been validated and used in workcontexts (Rioux and Penner 2001). The stem was as fol-lows: “How important is each of the following reasons

in your decision to volunteer?” Sample items includedthe following: “to avoid looking bad in front of oth-ers; to look better than my co-workers; because I wanta raise.” Items were measured on a 6-point scale (6 �extremely important, 1 � not at all important). Cron-bach’s α � 0.95, M � 2.76 (SD � 1.37).

ResultsAn overall ANOVA examining the means of prosocialintention by configuration type showed that those inthe focused enhancement condition reported greaterinvolvement (MFE � 4.23 (SD � 0.69) > MME � 4.03(SD � 0.79), F1,395 � 7.12, p < 0.01) and helping behav-ior (MFE � 5.50 (SD � 1.13) > MME � 5.20 (SD � 1.14),F1,395 � 6.83, p < 0.01) than the mutual enhancementcondition. An ANOVA examining the means of self-serving orientation by configuration type was also sig-nificant (MFE � 2.48 (SD � 1.29) < MME � 3.05 (SD �

1.40), F1,395 � 17.70, p < 0.001). A mediation analysisusing Hayes’ (2013) PROCESS macro with 1,000 boot-strap samples and a 95% CI shows that consistentwith our expectations, self-serving motives mediatedthis effect. Specifically, the indirect effect estimate forinvolvement was B � −0.07 (SE � 0.03), the 95% CIdid not include zero (LL � −0.13 to UL � −0.03), andthe indirect effect estimate for helping was B � −0.14(SE � 0.04), and the 95% CI did not include zero (LL �

−0.24 to UL � −0.07). These results support the argu-ment that self-serving orientation mediates the nega-tive effect of the mutual enhancement configuration onprosocial behavior.6

Discussion. Consistent with the earlier studies andsupporting Hypothesis 3, Study 4 shows that thefocused enhancement configuration exhibited greaterprosocial intention than the mutual enhancement con-figuration. Study 4 also replicates and extends the find-ings of the earlier studies by demonstrating that, astheorized, self-serving motives exhibited by those withthe mutual enhancement configuration are an under-lying mechanism driving their lower level of prosocialbehavior relative to the focused enhancement configu-ration. Whether this mechanism also differentiates thefocused enhancement configuration from other config-urations remains an avenue for future exploration.

General DiscussionProsocial behavior, which protects, promotes, and con-tributes to the welfare of others, is crucial for thehealthy functioning of organizations and societies.This paper set out to understand how configurationsof prosocial, collective, and individualistic identitiescombine to influence prosocial behavior. Intuitively, aconfiguration of mutual enhancement—in which allthree identities are enhancing—should lead to thehighest levels of prosocial behavior; however, across

Dow

nloa

ded

from

info

rms.

org

by [

128.

103.

149.

52]

on 1

9 Se

ptem

ber

2017

, at 1

1:14

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


four studies, utilizing multiple methods and multiplecontexts, we found that a focused enhancement con-figuration, characterized by high prosocial–collectiveidentity enhancement and limited individualistic iden-tity enhancement, resulted in the highest levels ofprosocial behavior.Our paper contributes to the literatures on both

prosocial behavior and multiple identities in organi-zations. First, by taking a configurational approachto identity relationships, we have introduced a newway to study multiple identities that extends beyondexisting single- and dual-identity approaches. In addi-tion, the results of Study 1 suggest that identityconfigurations better explain prosocial behavior thansingle, salient identities. Interestingly, other studieshave also found that single identifications were notsignificantly related to outcomes while identity rela-tionships were (Cheng et al. 2008). Reading our resultsalongside the existing literature suggests that whenseveral identities are simultaneously salient, a configu-rational approachmay bemore informative than exam-ining identities alone.

Second, by attending to both identity types andrelationships, we questioned the dominant intuitionregarding the mutual enhancement configuration andinstead demonstrated that a focused enhancementconfiguration could lead to higher levels of proso-cial behavior. Our research also articulates and pro-vides support for a mechanism through which focusedenhancement maximizes prosocial behavior. Specifi-cally, we suggest that individualistic identity enhance-ment may be counterproductive, because it ampli-fies a self-serving orientation when helping others(Batson et al. 1983, Grant and Mayer 2009, Rioux andPenner 2001).

Third, by focusing on prosocial behavior—an other-oriented outcome (Creary et al. 2015)—this papermoves research on multiple identities beyond per-sonally oriented outcomes such as well-being, workengagement, and creativity (Brook et al. 2008, Chenget al. 2008, Rothbard 2001, Thoits 1983). This opens upnew opportunities for studying the links between mul-tiple identities and prosocial behavior.

The results are bolstered by the multimethodapproach we used to unpack the complexity of mul-tiple identity configurations. Across the four studies,we used qualitative, quantitative, and experimentaldesigns, and we consistently found that the focusedenhancement configuration leads to higher levels ofprosocial behavior than the mutual enhancement con-figuration. Furthermore, the strengths of some stud-ies offset the limitations of others (McGrath 1981).For example, although Study 1 examines a voluntarysetting, the subsequent studies examine work-relatedcontexts. The concerns with utilizing “minimal” iden-tity configurations, hypothetical scenarios, and hypo-thetical behavioral intentions in Studies 3 and 4 are

somewhat mitigated by the field settings and situ-ated identities of participants in Studies 1 and 2. Like-wise, the experimental approaches of Studies 3 and 4help address concerns with causality and configura-tions that did not emerge in the field. Moreover, weused multiple measures of prosocial behavior (ourdependent variable), ranging from actual donations, toprosocial behavior rated by others, to prosocial inten-tions measured by validated scales. Taken together, thefour studies illustrate that our key findings hold acrossvarious operationalizations of identity configurationsand prosocial behavior.

Future research on both multiple identities andprosocial behavior can build on this paper’s findingsin several ways. First, a key boundary condition ofour studies is that we examine the effects of identityconfigurations on prosocial behavior. Extending thesefindings to other outcomes is an important directionfor future work. Specifically, we do not imply that thefocused enhancement configuration will be more effec-tive than the mutual enhancement configuration for alltasks and outcomes. Our pattern of findings may notextend, for example, to outcomes such as satisfaction orcareer success. Even within the literature on prosocialbehavior there is a common distinction between altru-istic and egoistic prosociality (Batson et al. 1983, Briefand Motowidlo 1986). The effects of focused enhance-ment may be most applicable to prosocial behaviorthat is largely altruistic, while themutual enhancementconfiguration may be no different, or better, for proso-cial behavior that jointly benefits oneself and others.Relatedly, perhaps the mutual enhancement configura-tion optimizes performance across multiple outcomes,while the focused enhancement configuration is mosteffective for prosocial behavior alone.

Second, future work should examine the relation-ship between collective and prosocial identities ingreater detail. In our studies, we did not manipu-late whether givers and receivers shared a collectiveidentity. Our findings suggest that the benefits offocused enhancement are present for helping mem-bers of one’s ingroup (Study 2) as well as help-ing nongroup members (Studies 1, 3, 4). However,ingroup/outgroup dynamics may moderate theseeffects. Focused enhancement may make it easier tohelp outgroup members, while mutual enhancementmay be more effective for helping ingroup mem-bers because individualistic identity enhancementmaybackfire lesswhen one can gain statuswithin the groupby helping groupmembers. A related boundary condi-tion may be that the benefits of focused enhancementare most likely to occur in contexts where the collectiveidentity and the prosocial identity somewhat overlap;for example, among organizations with social respon-sibility programs.

Dow

nloa

ded

from

info

rms.

org

by [

128.

103.

149.

52]

on 1

9 Se

ptem

ber

2017

, at 1

1:14

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


Third, future work should also examine the general-izability of these results beyond the identities studiedin this paper. For example, future work could explorethe valence of these identities (e.g., a has-been insteadof a hot-shot). It is likely that people will still focus onfulfilling their personal needs even when the individ-ualistic identity is negatively valenced because peoplemaintain negative self-views (Swann 1983). However,future work needs to explore whether the valence ofthe identities alters the impact of individualistic iden-tity enhancement. Future work could also examine dif-ferent types of identities (e.g., relational identities inthe work domain, such as a boss or subordinate).

Contemporary trends suggest that organizationswill increasingly try to integrate prosocial behaviorwith other goals (Battilana and Lee 2014, Grant 2012).Organizational members are also increasingly attempt-ing to blend prosociality with personal and profes-sional aspects of their lives (Bartel 2001, Grant et al.2008). This paper proposes a configurational approachto investigate how people experience and manage thecomplexity of their multiple identities in the face ofsuch trends. We suggest that managing our prosocialidentities along with other personal- and work-relatedidentities has important consequences for prosocialbehavior in organizational settings: mutual enhance-ment may backfire, while focused enhancement mayprovide important benefits.

AcknowledgmentsThe authors are grateful to the organizers and participantsof the Alyn and Arava Rides, who graciously supported theresearch. The authors are grateful to the Division for FacultyResearch and Development at Harvard Business School, theTed Rogers School of Management, and the Paul BaerwaldSchool of Social Work and Social Welfare, for enabling thisresearch. The authors thank their colleagues for feedbackon earlier versions of this paper: Michel Anteby, Blake Ash-forth, Sigal Barsade, Julie Battilana, Steven Blader, JenniferBerdahl, Stephanie Creary, Tracy Dumas, Robin Ely, MaryAnn Glynn, Francesca Gino, Adam Grant, Spencer Harrison,Jochem Menges, Celia Moore, Tsedal Neeley, Otilia Obo-daru, Jennifer Petriglieri, Jeff Polzer, Michael Pratt, Erin Reid,Nancy Rothbard, and SteffanieWilk. The authors are gratefulto seminar participants at Boston College, INSEAD, LondonSchool of Economics, London Business School, University ofBritish Columbia, University of Toronto, and the WhartonJunior Faculty OB Conference for their feedback. The authorsthank Breeshna Javed, Christy Ley, Ayelet Oreg, and JeanSohn for their assistance with data collection. The authorsare grateful to Emily LeRoux-Rutledge and Sharon Benheimfor their editing assistance.

Endnotes1Given the load on participants, we did not use bothmeasures in tan-dem in the field study. However, in a separate survey study (n � 500)onMturk the correlations between the visual scale and the three-itemmeasure for the same identity ranged between 0.7 to 0.8, providing

some evidence that the two measures were valid measures of thesame construct. Therefore, we standardized the measures.2The correlation between conflict and enhancement for a given pairof identities was between 0.1 and 0.25, and the correlation betweenidentity conflict and enhancement averaged across all the pairs was0.17, p < 0.05. This also supported the idea that conflict and enhance-ment are not opposites of one another.3We also examined the four-cluster solution against two-cluster andthree-cluster solutions based on theoretical grounds. The four-clustersolution was the most efficient clustering solution, providing max-imum variation between clusters while still exhibiting coherencewithin cluster.4We conducted several robustness checks. First, we examined theresults for a subsample of participants that raised over $2,500 (theminimum for Alyn) (n � 133) because we did not have a measure ofincome as a control variable. That is, participants who could affordto do so may have simply paid the minimum. The results are con-sistent with those presented here: the focused enhancement config-uration raises significantly more than the mutual enhancement andindependent; in addition, it is also significantly higher than mutualconflict. Second, we also had two outliers in the data, and we testedour results without those two observations to ensure they were notbeing driven by the outliers. The results are consistent with thosereported here. Third, we also examined the results separately for thetwo Alyn samples (2007, n � 60 and 2014, n � 65; note the Arava ridesample was too small to test the hypothesis (n � 37)). The resultswere also consistent with those reported here. Last, OLS results arealso consistent with those presented here. All additional results areavailable from the first author.5Because group members rated one another, we calculated the intr-aclass correlation coefficient, ICC(1), which measures how muchvariance in prosocial behavior is accounted for by nesting measureswithin a given unit (in this case, the group). ICC(1)s that are over 0.25are considered a “large” effect (LeBreton and Senter 2008), mean-ing that one-quarter or more of the variability in the outcome isdue to systematic within-group differences. The ICC(1) was 0.08. Inaddition, the smaller the number of observations within a group,the lower the impact of a small ICC(1) (Kreft and de Leeuw 1998).Here we had four to six observations per group. Thus, we expect theimpact of the within-group variability to be low. However, becausethere was still some variation explained by the group, we also exam-ined amultilevel model accounting for the group. Becausemultilevelmodeling typically requires much larger sample sizes at both levels(here we had 13 groups and 70 people; Maas and Hox 2005) andbecause there were no differences between the mutual enhancement,independent, and mutual conflict configurations, we examined onlythe focused enhancement configuration compared to the others. Wefound that our results are robust (Bfocused � 0.40, p � 0.014).6We conducted a replication of Study 4 to examine an additionalmeasure of self-serving orientation toward helping. We used thesame design, manipulation, and population from which to sample(n � 418 Mturk workers). We included both the measure of impres-sion management motives (Rioux and Penner 2001) used in Study 4and a four-item self-oriented helping scale (Roth 2008). The correla-tion between the two scales was 0.57. A mediation analysis as aboveshows that both measures of self-serving orientation toward help-ing were significant. Specifically, the indirect effect estimate of theimpression management motives scale (Rioux and Penner 2001) wasB�−0.06 (SE�0.02), and the 95%CI did not include zero (LL�−0.11to UL � −0.03), and the indirect effect estimate for the self-orientedhelping scale (Roth 2008) was B � −0.06 (0.02), and the 95% CI didnot include zero (LL � −0.11 to UL � −0.02). This suggests that ourresults are robust to these two different operationalizations of self-serving orientation toward helping in this context.

Dow

nloa

ded

from

info

rms.

org

by [

128.

103.

149.

52]

on 1

9 Se

ptem

ber

2017

, at 1

1:14

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


ReferencesAndreoni J (1990) Impure altruism and donations to public goods:

A theory of warm-glow giving. Econom. J. 100(401):464–477.Ashforth BE (2007) Identity: The elastic concept. Bartel CA, Blader

S, Wrzesniewski A, eds. Identity and the Modern Organization(Lawrence-Erlbaum, Mahwah, NJ), 85–96.

Ashforth BE, Johnson SA (2001) Which hat to wear? The relativesalience of multiple identities in organizational contexts. HoggMA, Terry DJ, eds. Social Identity Processes in Organizational Con-texts (Psychology Press, Philadelphia), 31–48.

Bales K (1996) Measuring the propensity to volunteer. Soc. PolicyAdmin. 30(3):206–226.

Bartel C (2001) Social comparisons in boundary-spanning work:Effects of community outreach on members’ organizationalidentity and identification. Admin. Sci. Quart. 46(3):379–313.

Bartel C, Blader S, Wrzesniewski A, eds. (2007) Identity and the Mod-ern Organization (Lawrence ErlbaumAssociates, Inc., Publishers,Mahwah, NJ).

Batson CD, O’Quin K, Fultz J, Vanderplas M, Isen AM (1983) Influ-ence of self-reported distress and empathy on egoistic ver-sus altruistic motivation to help. J. Personality Soc. Psych. 45(3):706–718.

Battilana J, LeeM (2014) Advancing research on hybrid organizing—Insights from the study of social enterprises. Acad. ManagementAnn. 8(1):397–441.

Baumeister RF, Bratslavsky E,MuravenM, TiceDM (1998) Ego deple-tion: Is the active self a limited resource? J. Personality Soc. Psych.74(5):1252–1265.

Benet-Martinez V, Haritatos J (2005) Bicultural identity integration(BII): Components and psychosocial antecedents. J. Personality73(4):1015–1050.

Bergami M, Bagozzi RP (2000) Self-categorization, affective commit-ment and group self-esteem as distinct aspects of social identityin the organization. British J. Soc. Psych. 39(4):555–577.

Berger IE, Gainer B (2002) Jewish identity, social capital and giving.Adv. Consumer Res. 29(1):408–413.

Blader SL (2007) Let’s not forget the“ me” in “team”: Investigatingthe interface of individual and collective identity. Bartel CA,Blader S, Wrzesniewski A, eds. Identity and the Modern Organiza-tion (Lawrence-Erlbaum, Mahwah, NJ), 61–84.

Blader SL, Tyler TR (2009) Testing and extending the group engage-ment model: Linkages between social identity, procedural jus-tice, economic outcomes and extrarole behavior. J. Appl. Psych.94(2):445–464.

Brewer MB, Gardner W (1996) Who is this “we”? Levels of collec-tive identity and self representations. J. Personality Soc. Psych.71(1):83–93.

Brickson SL (2013) Athletes, best friends, and social activists: Anintegrative model accounting for the role of identity in organi-zational identification. Organ. Sci. 24(1):226–245.

Brief AP, Motowidlo SJ (1986) Prosocial organizational behaviors.Acad. Management Rev. 11(4):710–725.

Brook AT, Garcia J, Fleming MA (2008) The effects of multiple iden-tities on psychological well-being. Personality Soc. Psych. Bull.34(12):1588–1600.

Burke PJ, Stets J (2009) Identity Theory (Oxford University Press, NewYork).

Carver CS, White TL (1994) Behavioral inhibition, behavioral activi-ation, and affective responses to impending reward and punish-ment: The BIS/BAS scales. J. Personality Soc. Psych. 67(2):319–333.

Caza B, WilsonM (2009)Me, Myself, and I: The Benefits of Work IdentityComplexity (Psychology Press, London).

Cheng C, Sanchez-Burks J, Lee F (2008) Connecting the dotswithin: Creative performance and identity integration. Psych.Sci. 19(11):1178–1184.

Clary EG, Snyder M, Ridge RD, Copeland J, Stukas AA, Haugen J,Miene P (1998) Understanding and assessing the motivationsof volunteers: A functional approach. J. Personality Soc. Psych.74(6):1516–1530.

Creary SJ, Caza BB, Roberts LM (2015) Out of the box? How man-aging a subordinate’s multiple identities affects the qualityof a manager-subordinate relationship. Acad. Management Rev.40(4):538–562.

De Dreu CK, Nauta A (2009) Self-interest and other-orientationin organizational behavior: Implications for job performance,prosocial behavior, and personal initiative. J. Appl. Psych.94(4):913–926.

Dutton JE, Roberts LM, Bednar JS (2010) Pathways for positive iden-tity construction at work: Four types of positive identity andthe building of social resources. Acad. Management Rev. 35(2):265–293.

Edmondson A, McManus SE (2007) Methodological fit in manage-ment field research. Acad. Management Rev. 32(4):1155–1179.

Grant AM (2007) Relational job design and the motivation to make aprosocial difference. Acad. Management Rev. 32(2):393–417.

Grant AM (2008) Does intrinsic motivation fuel the prosocial fire?Motivational synergy in predicting persistence, performance,and productivity. J. Appl. Psych. 93(1):48–58.

Grant AM (2012) Giving time, time after time: Work design and sus-tained employee participation in corporate volunteering. Acad.Management Rev. 37(4):589–615.

Grant AM, Dutton JE (2012) Beneficiary or benefactor: Are peoplemore prosocial when they reflect on receiving or giving? Psych.Sci. 23(9):1033–1039.

Grant AM, Mayer DM (2009) Good soldiers and good actors: Proso-cial and impression management motives as interactive pre-dictors of affiliative citizenship behaviors. J. Appl. Psych. 94(4):900–912.

Grant AM, Sumanth JJ (2009) Mission possible? The performanceof prosocially motivated employees depends on manager trust-worthiness. J. Appl. Psych. 94(4):927–944.

Grant AM, Dutton JE, Rosso BD (2008) Giving commitment:Employee support programs and the prosocial sensemakingprocess. Acad. Management J. 51(5):898–918.

Greenhaus JH, Beutell NJ (1985) Sources of conflict between workand family roles. Acad. Management Rev. 10(1):76–88.

Grube JA, Piliavin JA (2000) Role identity, organizational experi-ences, and volunteer performance. Personality Soc. Psych. Bull.26(9):1108–1119.

Hayes AF (2013) Introduction to Mediation, Moderation, and ConditionalProcess Analysis (Guilford Press, New York).

Hirsh JB, Kang SK (2016) Mechanisms of identity conflict: Uncer-tainty, anxiety, and the behavioral inhibition system. Pers. Soc.Psych. Rev. 20(3):223–244.

Hugenberg K, Bodenhausen GV (2004) Category membership mod-erates the inhibition of social identities. J. Experiment. Soc. Psych.40(2):233–238.

James W (1890) The Principles of Psychology (Harvard UniversityPress, Cambridge, MA). [Reprint 1983.]

Kreft I, de Leeuw J (1998) Introducing Multilevel Modeling (Sage Pub-lications, Thousand Oaks, CA).

LeBreton JM, Senter JL (2008) Answers to 20 questions about inter-rater reliability and interrater agreement. Organ. Res. Methods11(4):815–852.

Lipponen J, Helkama K, Olkkonen ME, Juslin M (2005) Predict-ing the different profiles of organizational identification: A caseof shipyard subcontractors. J. Occupational Organ. Psych. 78(1):97–112.

Luhtanen R, Crocker J (1992) A collective self-esteem scale: Self-evaluation of one’s social identity. Personality Soc. Psych. Bull.18(3):302–318.

Maas CJM, Hox JJ (2005) Sufficient sample sizes for multilevel mod-eling.Methodology 1(3):86–92.

Mael F, Ashforth BE (1992) Alumni and their alma mater: A partialtest of the reformulated model of organizational identification.J. Organ. Behav. 13(2):103–123.

Marcussen K (2006) Identities, self-esteem, and psychological dis-tress: An application of identity-discrepancy theory. Soc. Per-spect. 49(1):1–24.

Dow

nloa

ded

from

info

rms.

org

by [

128.

103.

149.

52]

on 1

9 Se

ptem

ber

2017

, at 1

1:14

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.


Marks SR (1977)Multiple roles and role strain: Some notes on humanenergy, time and commitment. Amer. Sociol. Rev. 42(6):921–936.

McGrath JE (1981) Dilemmatics: The study of research choices anddilemmas. Amer. Behav. Scientist 25(2):179–210.

Mikulincer M, Shaver PR, Gillath O, Nitzberg RA (2005) Attach-ment, caregiving, and altruism: Boosting attachment securityincreases compassion and helping. J. Personality Soc. Psych.89(5):817–839.

O’Reilly CA, Chatman J (1986) Organizational commitment andpsychological attachment: The effects of compliance, identifica-tion, and internalization on prosocial behavior. J. Appl. Psych.71(3):492–499.

Penner LA, Dovidio JF, Piliavin JA, Schroeder DA (2005) Proso-cial behavior: Multilevel perspectives. Annual Rev. Psych.56(February):365–392.

Piliavin JA, Callero PL (1991) Giving Blood: The Development of anAltruistic Identity (Johns Hopkins University Press, Baltimore).

Pratt MG, Foreman PO (2000) Classifying managerial responsesto multiple organizational identities. Acad. Management Rev.25(1):18–42.

Pratt MG, Fiol CM, O’Connor EJ, Panico P (2012) Promoting positivechange in physician–administrator relationships: The impor-tance of identity security in managing intractable identity con-flicts. Golden-Biddle K, Dutton JE, eds. Using a Positive Lens toExplore Social Change and Organizations: Building a Theoretical andResearch Foundation (Routledge, New York), 267–288.

Punj G, Stewart DW (1983) Cluster analysis in marketing research:Review and suggestions for application. J. Marketing Res.20(2):134–148.

Ramarajan L (2014) Past, present and future research on multipleidentities: Toward an intrapersonal network approach. Acad.Management Ann. 8(1):589–659.

Rioux SM, Penner LA (2001) The causes of organizational citizen-ship behavior: A motivational analysis. J. Appl. Psych. 86(6):1306–1314.

Roccas S, Brewer MB (2002) Social identity complexity. PersonalitySoc. Psych. Rev. 6(2):88–106.

Roth G (2008) Perceived parental conditional regard and auton-omy support as predictors of young adults’ self- versus other-oriented prosocial tendencies. J. Personality 76(3):513–534.

Roth PL, BeVier CA (1998) Response rates in HRM/OB sur-vey research: Norms and correlates, 1990–1994. J. Management24(1):97–117.

Rothbard NP (2001) Enriching or depleting? The dynamics of engag-ment in work and family roles. Admin. Sci. Quart. 46(4):655–684.

Rothbard NP, Ramarajan L (2009) Checking your identities at thedoor? Positive relationships between nonwork and work identi-ties. Roberts LM, Dutton JE, eds. Exploring Positive Identities andOrganizations (Psychology Press, New York), 125–144.

Schachter EP (2004) Identity configurations: A new perspectiveon identity formation in contemporary society. J. Personality72(1):167–200.

Schaubroeck J, Ganster DC (1991) Beyond the call of duty: A fieldstudy of extra-role behavior in voluntary organizations. HumanRelations 44(6):569–582.

Shariff AF, Norenzayan A (2007) God is watching you priming Godconcepts increases prosocial behavior in an anonymous eco-nomic game. Psych. Sci. 18(9):803–809.

Sonenshein S, DeCelles K, Dutton J (2014) It’s not easy being green:Self-evaluations and their role in explaining support of environ-mental issues. Acad. Management J. 57(1):7–37.

Stets JE, Harrod MM (2004) Verification across multiple identities:The role of status. Soc. Psych. Quart. 67(2):155–171.

Swann WB (1983) Self-verification: Bringing social reality into har-mony with the self. Suls J, Greenwald AG, eds. Social Psycholog-ical Perspectives on the Self, Vol. 2 (Psychology Press, New York),33–66.

Thoits PA (1983) Multiple identities and psychological well-being:A reformulation and test of the social isolation hypothesis.Amer.Sociol. Rev. 48(2):174–187.

Tiedje LB, Wortman CB, Downey G, Emmons C, Biernat M, Lang E(1990) Women with multiple roles: Role-enhancement percep-tions, satisfaction, and mental-health. J. Marriage Family 52(1):63–72.

Tyler TR, Blader SL (2003) The group engagement model: Proceduraljustice, social identity, and cooperative behavior. Personality Soc.Psych. Rev. 7(4):349–361.

Watson D, Clark LA, Tellegen A (1988) Development and validationof brief measures of positive and negative affect: The PANASscales. J. Personality Soc. Psych. 54(6):1063–1070.

Lakshmi Ramarajan is assistant professor of organi-zational behavior at Harvard Business School, HarvardUniversity. She received her Ph.D. from the Wharton School,University of Pennsylvania. Her research examines the man-agement and consequences of individuals’ multiple identi-ties in organizations.

Ida E. Berger is professor of marketing and a memberof the graduate faculty at the Ted Rogers School of Busi-ness Management at Ryerson University. She received herPh.D. from the University of Toronto. Her research interestsare social identity, ethnic diversity, social integration, socialalliances, socially responsible consumer behavior, and phi-lanthropy.

Itay Greenspan is lecturer of nonprofit management andleadership at the Paul Baerwald School of Social Work andSocial Welfare, Hebrew University of Jerusalem. He receivedhis Ph.D. from the University of Pennsylvania. His researchinterests include nonprofit organizations and civil society,prosocial and proenvironmental behavior, identity forma-tion, and environmental civic participation.

Dow

nloa

ded

from

info

rms.

org

by [

128.

103.

149.

52]

on 1

9 Se

ptem

ber

2017

, at 1

1:14

. Fo

r pe

rson

al u

se o

nly,

all

righ

ts r

eser

ved.

Multiple Identity Configurations: The Benefits of Focused ... Files/Multiple Identity...

Documents

Transcript of Multiple Identity Configurations: The Benefits of Focused ... Files/Multiple Identity...