To Delegate or Not to Delegate: Gender Differences in ...

r Academy of Management Journal2018, Vol. 61, No. 4, 1467–1491.https://doi.org/10.5465/amj.2016.0662

TO DELEGATE OR NOT TO DELEGATE: GENDERDIFFERENCES IN AFFECTIVE ASSOCIATIONS AND

BEHAVIORAL RESPONSES TO DELEGATION

MODUPE AKINOLAColumbia University

ASHLEY E. MARTINStanford University

KATHERINE W. PHILLIPSColumbia University

Effectively delegating work to others is considered critical to managerial success, as itfrees up managers’ time and develops subordinates’ skills. We propose that femaleleaders are less likely than male leaders to capitalize on these benefits of delegating.Although delegation has communal (e.g., relational) and agentic (e.g., assertive) prop-erties, we argue that female leaders, as compared tomale leaders, find it more difficult todelegate tasks due to gender-role incongruence. In five studies, we draw upon social roleand backlash theories to show that women imbue delegation with more agentic traits,have more negative associations with delegating, and feel greater guilt about delegatingthan men. These associations result in women delegating less than men and, when theydo delegate, having lower-quality interactions with subordinates. We further show thatreframing delegation as communal attenuates women’s negative associations with del-egation. These findings reveal that even when a given behavior has both agentic andcommunal elements, perceptions of agency can undermine women’s engagement inthem. However, emphasizing the communal nature of seemingly agentic acts may en-courage women’s engagement in such critical leadership behaviors. These findings havetheoretical and practical implications for research on gender differences and leadershipbehavior in the workplace.

Delegation, defined as the assignment of re-sponsibilities to subordinates and conferral of au-thority to carry out assigned tasks, is considered animportant and effective leadership behavior. Forleaders, delegation can reduce work overload andimprove the speed and quality of decisions, whilesimultaneously enabling subordinates to viewleaders as participative (De Pater, Van Vianen, &Bechtoldt, 2010). For subordinates, delegation canincrease intrinsic motivation and provide opportu-nities for professional development, as it empowers

subordinates, enables efficacy, and strengthensrelationship building between leaders and sub-ordinates (Chen & Aryee, 2007; Leana, 1986, 1987;Yukl & Fu, 1999). For organizations, delegation en-hances task coordination, productivity, and perfor-mance by drawing upon the specializations andskills of different employees (Becker & Murphy,1992; Miles, Snow, Meyer, & Coleman, 1978).

Definitions of delegation in much of the manage-ment literature have painted this leadership behav-ior as one that is primarily communal in nature, asit is relational and can help leaders develop em-ployees. Since communality is a prescriptive genderstereotype that is rewarded when exhibited bywomen (Rudman & Phelan, 2008), one might arguethat women leaders should benefit from engagingin delegation (Amanatullah & Morris, 2010; Carli,LaFleur, & Loeber, 1995; Heilman & Okimoto,2007). Yet the act of delegating can also be consid-ered agentic, as it requires assertiveness and

The first two authors contributed equally to this work.We thankSarahAnolik,HayleyBlunden,ZachBucknoff,

Ashley Culver, Zach Heinemann, Yael Warach, LauraFitzelle, and Lucie Yang for their research assistance. Wealso thank Daniel Ames, Adam Galinsky, Astrid Homan,Jessica Kennedy, Michael Slepian, and the Social In-teractions in Groups and Organizations Lab at ColumbiaBusiness School for their thoughtful comments and sug-gestions regarding this research.

1467

Copyright of the Academy of Management, all rights reserved. Contents may not be copied, emailed, posted to a listserv, or otherwise transmitted without the copyright holder’s expresswritten permission. Users may print, download, or email articles for individual use only.

https://doi.org/10.5465/amj.2016.0662

intentionality. Evidence has suggested that pre-scriptive gender stereotypes that reward agentic be-havior for men can sanction this same behavior forwomen (Eagly, 2007; Heilman, 2001; Rudman &Phelan, 2008). Thus, delegation can be considereda gender-role-paradoxical behavior for women: insome respects, delegation is alignedwith female rolestereotypes, thus presumably helping women; inother respects, delegation is misaligned with femalerole stereotypes, thus potentially hurting womenwhen they engage in this behavior. In this paper, wepropose that women will associate beneficial lead-ership behaviors that are simultaneously construedas both agentic and communal with more negativeattributes, and will engage in these behaviors lessfrequently than men.

Specifically, we examine whether women viewdelegation with more negative affect than men do,and if so, why. Research on social role and backlashtheories has demonstrated that women leaders facea double bind, as the same qualities that make them“good leaders” (e.g., agency) also make them “badwomen” (e.g., not communal) (Tannen, 1990: 244).In contrast, male leaders are often free to exercisetheir agency without the constraint of simulta-neouslyneeding to engender communality (Heilman& Okimoto, 2007; Rudman, 1998). Drawing on thesetheories, we argue that when leadership behaviorscan be viewed as both agentic and communal, theagentic nature of the behaviorwill bemore salient forwomen than the communal nature. Thus, engagingin such behaviors will be more difficult for womenthan for men.

When applied to the context of delegation, wetheorize that women are more likely to associatedelegation with agency than men are, as women arepenalized more for showing agency than they arerewarded for showing communality (Heilman &Chen, 2005). Since women often see themselves as,and are expected to be, less agentic (i.e., assertive,aggressive) thanmen (Bem, 1981; Rudman& Phelan,2008), we propose that they will be more sensitive tothe act of delegating, ascribing it with more agenticcharacteristics compared to men. In turn, women’saffective associations (i.e., emotions and feelingsassociated with a behavior) with delegation will bemore negative, affecting their propensity to delegate.Accordingly, we argue that the gender-role-paradoxical nature of delegation will lead women,compared to men, to associate delegation with morenegative emotions, being more sensitive to punish-ment that can ensue from engaging in gender-roleincongruent behavior (Brescoll, 2012; Moss-Racusin

& Rudman, 2010). Since affective associations witha given behavior are powerful predictors of engage-ment in suchbehavior,we expect thatwomen’smorenegative associations with delegation will decreasetheir likelihood of delegating and will negativelyaffect the quality of their interactions with sub-ordinates. We also identify ways to mitigate genderdifferences in delegation. We argue that for women,emphasizing the communal aspects of delegationmay allow its communal properties to outweigh itsagentic ones. Thus, we propose that framing dele-gation as communal, by highlighting its benefits forsubordinates, has the potential to change women’snegative associations with delegation.

The current research contributes to managementtheory in four key ways. First, to our knowledge, weoffer the first empirical evidence demonstrating thatwomen and men differ in their delegation behavior.Gender researchers have largely examined how dif-fering gender-role stereotypes shape how womenexperience and perceive leadership behaviors thatare construed as either agentic or communal, such asself-promotion and negotiating on behalf of others(Amanatullah & Morris, 2010; Heilman, 2001;Rudman & Phelan, 2008). Instead, we offer insightinto how women experience and perceive gender-role-paradoxical leadership behaviors, such as del-egation, that could be simultaneously construed asboth agentic and communal.

Second, we provide the first empirical evidence ofthe processes throughwhich gender differencesmayinfluence delegation behavior. We propose thataffective associations are an important predictor ofdelegation behavior for women. Men’s and women’sdiffering affective associations with delegation maymake women less likely to delegate and can affectperformance. Based on social role and backlashtheories (Eagly, 1987; Eagly & Karau, 2002; Rudman,1998; Rudman & Phelan, 2008), we advance thetheoretical argument that for tasks that are bothagentic and communal, agentic qualities loom largerthan communal ones for women, dominating theirperceptions and generating greater negative affectas compared to men.

Third, effective delegation entails having a pro-ductive relationship with subordinates, yet little isknown about the quality of managers’ interactionswith subordinates in the context of delegation.We propose that women’s negative associationswith delegation can influence the quality of theirinteractions with subordinates. Additionally, weexamine delegation in a controlled laboratory ex-periment. In doing so, we experimentally test

1468 AugustAcademy of Management Journal

whether there are gender differences in delegationwhen position, task, and employees are heldconstant.

Fourth, by clarifying why women differentiallydelegate,weoffer insight intopotential interventionsthat may help address gender disparities in leader-ship behaviors. Specifically, our research suggeststhat when leadership behaviors that have bothagentic and communal properties are reframed ascommunal in nature, women may view them morefavorably. Thus, this research not only offers insightinto the challengeswomenmay face as they continueto ascend to top leadership positions in organiza-tions, particularly in engaging in behaviors that areboth agentic and communal, but also provides po-tential remedies to these challenges.

In what follows, we draw upon literature on so-cial role (Eagly, 1987, 2007) and backlash (Rudman,1998) theory to describe the theoretical rationale forwhy women may have more negative associationswith delegation relative to men, and for how thesenegative associations may influence women’s del-egation behavior. We build a theoretical case forhow reframing delegation to be more gender-rolecongruent may change women’s negative associa-tions with delegation.We then describe five studiesthat test our hypotheses that women andmen differin their perceptions of, and behavior during, dele-gation, and that women’s affective associationscan be changed by reframing how they perceivedelegation.

SOCIAL ROLE THEORY: GENDER STEREOTYPESAND ROLE INCONGRUENCE

Gender stereotypes pervade society and influ-ence how men and women are perceived and ex-perience their lives (Eagly, 2007; Heilman, 2001).According to social role theory, the roles occupiedby, and division of labor between, men and womencreate expectancies about what women are like andhow they are expected to behave. As such, womenare expected to be communal, enacting behaviorsthat are cooperative and other-oriented, whereasmen are expected to demonstrate agency, enactingbehaviors that are independent and action-oriented.

Both men and women internalize gender roles ata very young age through a variety of means. Media,parents, and peers teach children to behave ingender-normative ways, rewarding girls for beingcommunal and boys for being agentic (Bem, 1983;Bryant & Check, 2000). For example, parents en-courageandrewardgender-congruentbehaviors inplay

and household chores (Weisner & Wilson-Mitchell,1990), and teachers encourage agency in boys bycalling on them more often and giving them moretalking time than girls (Kindlon & Thompson, 2000).By adulthood, men and women tend to self-identifywith these gender stereotypes, with women oftenrating themselves as more communal and behavingmore communally than men, and men rating them-selves as more agentic and behaving more agenti-cally (Bem, 1981; Valian, 1998).

Those who violate these gender norms are oftensocially and economically penalized (Rudman,1998; Rudman & Glick, 2001). For example,agenticwomen are evaluated less favorably, receivelower salaries, and are less likely to be hired for thesame position compared to agentic men (Heilman,Wallen, Fuchs, & Tamkins, 2004; Rudman, 1998;Rudman&Fairchild, 2004). As a result, bothwomenand men are sensitive to situations that elicit so-cial sanctions for counter-stereotypical behavior(Brescoll, 2012; Moss-Racusin, Phelan, & Rudman,2010). This social sensitivity to gender-role in-congruence may have implications for how womenperceive delegation, as the social penalties forstereotype-incongruent behavior may loom largerin their minds than the rewards for congruency(Heilman & Chen, 2005; Rudman & Phelan, 2008).Thus, although delegation has qualities of bothagency (stereotype incongruence for women) andcommunality (stereotype congruence for women),we argue women will be especially sensitive tothe agentic nature of delegation, as compared tomen, due to social costs outweighing gains (i.e.,communality does not help women as much asagency hurts them [Cuddy, Fiske, & Glick, 2004]).With regard to the communal nature of delegation,communality is often perceived more positivelythan agency (Abele & Wojciszke, 2014), and peopledescribe themselves with more communal thanagentic traits (Bruckmuller & Abele, 2013); thus, wedo not expect to see gender differences in percep-tions of delegation as communal. We thereforehypothesize:

Hypothesis 1. Women will associate the act of dele-gating (in general) with more agentic traits (e.g., aggres-siveness, power) compared to men.

GENDER-ROLE STEREOTYPES AND AFFECTIVEASSOCIATIONS WITH DELEGATION

Both men and women are sensitive to situationsthat evoke penalties for gender-role incongruent

2018 1469Akinola, Martin, and Phillips

behavior (Moss-Racusin & Rudman, 2010; Vandello,Bosson, Cohen, Burnaford, & Weaver, 2008).Awareness of these repercussions creates negativeassociations with engaging in gender-role in-congruent behaviors, where women are more likelyto associate agency with negative emotions com-pared to men. These negative associations can takemany forms, including feeling heightened anxietyin anticipation of engaging in agentic behavior(Babcock & Laschever, 2003; Spencer, Steele, &Quinn, 1999), feeling guilt for exhibiting agenticbehavior (Benetti-McQuoid & Bursik, 2005), andfearing backlash from engaging in agentic behavior(Brescoll, 2012).

One specific emotion related to gender-role viola-tion is guilt. Guilt, an unpleasant emotion associatedwith the recognition that one has violated a per-sonal or social standard (Kugler & Jones, 1992),can arise from a sense of responsibility to up-hold cultural norms for social behavior (Izard,1977). Women who engage in gender-role-paradoxical behaviors that they perceive asagentic (e.g., delegating) are likely to recognize thattheir actions run counter to the female gender normof behaving communally. Consequently, they mayexperience feelings of guilt, as engaging in situa-tions that are incongruent with one’s gender rolecan influence the emergence of guilt (Benetti-McQuoid & Bursik, 2005). Thus, if women associ-ate delegation with agentic traits more than men do(as proposed in Hypothesis 1), they may havegreater negative associations with delegation com-pared to men.

A second negative association that has beenfound to co-occur with agentic behavior for womenis fear of backlash. Fear of backlash, or trepidationabout potentially incurring social and economicpenalties from others (Rudman 1998; Rudman &Phelan, 2008), has been shown to be prevalent forwomen in a variety of contexts. For example,women experience a fear of backlash for self-promoting, as they view this behavior as more as-sertive and gender-role incongruent than men do(Moss-Racusin & Rudman, 2010). Likewise, womenavoid behaving aggressively in negotiations(Amanatullah &Morris, 2010) and minimize powerdisplays in political and organizational settings(Brescoll, 2012) for fear of backlash. This researchsuggests that because women may have negativeassociations with engaging in agentic behaviors,they may feel guilty for violating communalityexpectations and may also fear backlash. We pro-pose that these negative associations extend to

the domain of delegation through the followinghypotheses:

Hypothesis 2a. Women, as compared to men, willhave more negative affective associations when an-ticipating or engaging in the act of delegation.

Hypothesis 2b. Women, as compared to men, willhave the specific negative affective association ofguilt when anticipating or engaging in the act ofdelegation.

Hypothesis 2c. Women, as compared to men, willhave the specific negative affective association offear (of backlash) when anticipating or engaging inthe act of delegation.

GENDER-ROLE STEREOTYPES ANDDELEGATION BEHAVIOR

Weargue that perceiving delegation as agentic andhaving negative associations with delegation canimpact whether women choose to delegate whengiven the opportunity, and influence women’s be-haviorwhendelegating. Previousmodels of decisionmaking have assumed that individuals engage ina rational process of weighing the costs and benefitsof behavioral choices and selecting the course ofaction with the most favorable ratio of costs to ben-efits (Weinstein, 1993).However, significant supporthas emerged for the idea that the emotions, or affect,associated with a behavior play an important rolein decision making (Crites, Fabrigar, & Petty, 1994;Kiviniemi, Voss-Humke, & Seifert, 2007; Simons &Carey, 1998). Specifically, behavioral choices areinfluenced by the anticipation of feeling particularemotions (such as regret, guilt, pride, or happiness)as a result of engaging in a specific behavior (Richard,van der Pligt, & de Vries, 1996).

Emotions associated with a particular behavioralpractice are powerful predictors of one’s likelihoodof participating in specific behaviors, with morepositive affective associations leading to a greaterlikelihood of engaging in a behavior (Kiviniemi &Bevins, 2008). For example, more positive and lessnegative associations with exercise are related toincreased motivation to engage in physical activity(Laverie, 1998). Having a less negative associationwith medical testing is related to having a greaterlikelihood of partaking in preventive medical tests(Kiviniemi, Jandorf, & Erwin, 2014). And holdingmore negative associations with marijuana, nico-tine, and alcohol is related to reduced smokingand drinking (Simons & Carey, 1998; Trafimow &


Sheeran, 1998). Together, these results suggest that ifwomen have negative associations with delegation,including guilt and fear of backlash, then these neg-ative affective associations should deter womenfrom delegating. In contrast, since for men agency isprescribed and rewarded, they should not experi-ence these same negative affective associations andwill be free to engage in delegating. We thereforehypothesize:

Hypothesis 3a. Women will be less likely to delegateas compared to men.

Hypothesis 3b. Men’s and women’s differential af-fective associationswith delegationwillmediate theirlikelihood of delegating.

Outside the realm of health psychology, social-psychological research has shown that affectiveassociations influence behavior toward others. Forexample, individuals who have more negative asso-ciations with a particular social group are less likelyto interact with individuals in that category (Rooth,2010) and to have more negative interactions whentheydo interactwith those individuals (McConnell &Leibold, 2001). Negative associations can increaseanxiety and avoidant behavior (Plant, Devine, &Brazy, 2003; Stephan & Stephan, 1985), depletingcognitive capacity andheightening self-regulation insocial interactions (Apfelbaum, Sommers, & Norton,2008). These forces can divert attention away fromthe social interaction itself, resulting in poorer-quality interactions (Eysenck, 1992) and impairedperformance (Richeson & Shelton, 2003). Thus, ifwomen have negative associations with delegationthen when they do delegate, the quality of their in-teractions with subordinates should be lower thanthose of male delegators, who are less likely to havenegative associations with delegation due to itsgender-role congruent nature for men. Thus, wehypothesize:

Hypothesis 3c. Women’s interactions with sub-ordinates when delegating will be perceived by bothsubordinates and independent observers as beingpoorer in quality compared to men’s behavior whendelegating.

Further, given that delegation has been found toprovide critical performance benefits for managers(Leana, 1987; Yukl & Fu, 1999), we more broadlyhypothesize:

Hypothesis 3d. Delegators will outperform non-delegators in the tasks they have been given tocomplete.

ATTENUATING NEGATIVE ASSOCIATIONSWITH DELEGATION

It is possible that women’s affective associationswith regard to gender-role-paradoxical behaviors maybe influenced by how these behaviors are framed.A large body of research in judgment and decisionmaking (e.g., Kahneman & Tversky, 1984) has high-lighted that using language that changes how in-formation is framed can influence individuals’behavior in numerous domains. For instance, framingrisks in terms of losses versus gains changes in-dividuals’ decision-making processes. Specifically,because losses feel more painful than gains feel plea-surable, framing risks in terms of losses fosters riskacceptance, while framing risks as gains fosters riskavoidance (Kahneman & Tversky, 1984; Rothman &Salovey, 1997). These framing effects have been at-tributed to the feelings and construals evoked by dif-feringdescriptionsof the same information (Liberman,Samuels, & Ross, 2004; McFarland & Miller, 1994;).

Gender researchers have capitalized on framingeffects in an effort to reduce gender differences inbehavior, particularly in the domain of negotiation(Bowles, Babcock, & Lai, 2007; Bowles, Babcock, &McGinn, 2005; Small, Gelfand, Babcock, & Gettman,2007). Since negotiation is seen as an agentic be-havior, women are often more reluctant to negotiatewhen “given the opportunity to negotiate” comparedto men (Small et al., 2007); however, when the samesituations are described more communally (e.g., “anopportunity to ask”) or communal properties arehighlighted (e.g., negotiating on behalf of others)gender differences are reduced (Amanatullah &Morris, 2010; Small et al., 2007). Highlighting thecommunal properties of agentic acts aligns femalebehavior with their communal prescriptions andreduces the (fear of) social repercussions involved inviolating gender roles (Heilman & Okimoto, 2008).Taken together, this research suggests that framinggender-role-paradoxical behaviors as communal byemphasizing how these behaviors benefit others hasthe potential to change women’s negative associa-tions with behaviors that are both communal andagentic. Thus, we hypothesize:

Hypothesis 4. Framing delegation in a manner thatemphasizes its communal properties will attenuatewomen’s negative associations with delegation.

OVERVIEW OF THE PRESENT RESEARCH

We conducted five studies to test our hypotheses.In Study 1, using a survey of Master of Business


Administration (MBA) students, we tested whetherwomen perceive delegation to be more agentic thanmen do (Hypothesis 1). In Study 2, a scenario study,we instructed men and women to either delegate ornot delegate a set of tasks, then examined whethermen and women have differential associations withdelegation using text analysis (Hypothesis 2a). InStudy 3, a correlational study, we again used textanalysis to examine whether men and women differin their associations with delegation (Hypotheses 2aand 2b) among a sample of participants with mana-gerial experience. In Study 4, we examined the de-gree to which men’s and women’s associations withdelegation differentially predict their actual delega-tion behavior in an experimental setting involvinga face-to-face interaction with an ostensible sub-ordinate (Hypotheses 3a–3d). Finally, in Study 5, wetested whether framing delegation as communal innature will attenuate women’s negative associationswith delegation (Hypothesis 4).

STUDY 1

Through a survey administered to MBA studentsat a large, private university, we tested the hypoth-esis that women perceive delegation to be moreagentic than men and that these asymmetrical per-ceptions do not extend to perceptions of delegationas communal.

Method

Participants. We recruited 104 participants fromthe MBA program of an East Coast University, astheir recent work experience likely entailed dele-gating. Seven participants with no delegation expe-rience were excluded, leaving a sample of 97participants (63men; 34women)with an average ageof 28.05 years (SD 5 2.81). Fifty students were inmanagerial positions prior to pursuing their MBAs,and participants had 13.78 (SD 5 39.82) sub-ordinates on average. The racial composition of thesample was 46% White, 33% Asian, 7% Hispanic,and 13% other.

Procedure. Participants were asked questions re-garding delegation embedded in a decision-makingsurvey administered during a leadership course.Specifically, they completeddemographic questionsand were asked, “How much do you associate thefollowing adjectives with delegating?” on scalesranging from 1 (not at all) to 7 (very much). Based onKoenig and Eagly (2014) and Diekman and Eagly(2000), we selected four adjectives reflecting

agency—powerful, confident, aggressive, and con-trolling (a 5 .51)1—and four adjectives reflectingcommunality—other-oriented, supportive, develop-mental, and considerate (a 5 .79).

Results

Supporting Hypothesis 1, we found that womenassociated delegation with agency (M 5 4.57, SD 5.84) more than men did (M 5 4.16, SD 5 1.03), F(1,95) 5 3.86, p 5 .05, hp

2 5 .039. With regard to asso-ciations of communality with delegation, there wasno difference between women (M5 5.24, SD5 1.18)and men (M 5 5.17, SD 5 1.18), F(1, 95) 5 .07, p 5.80, hp

2 , .00.

Discussion

In support of Hypothesis 1, female participantsviewed delegation as more agentic than male par-ticipants did. Since both men and women per-ceive communality more positively than agency(Bruckmuller & Abele, 2013), we did not expect anddid not find gender differences in communality.Importantly, both men and women perceived dele-gation as more communal than agentic overall;however, both communality and agency ratings fordelegation were above the midpoint, offering evi-dence that delegation is indeed considered bothagentic and communal in nature. Further, thesefindings support our theoretical argument thatwomen may be more sensitive to the gender-roleincongruence of delegation, perceiving it as moreagentic thanmen do.We argue that this sensitivity tothe agentic nature of delegation may lead womento view delegation more negatively than men do,a hypothesis we tested in Study 2.

STUDY 2

Study 2 was designed to understand women’s andmen’s differential affective responses when told ei-ther to delegate or not delegate.Using a 2 (participantgender: male vs. female)3 2 (condition: delegate vs.do not delegate) 3 2 (subordinate gender: male vs.female) design, we examined men’s and women’s

1 Low alphas have been found in samples with signifi-cant international diversity (Zou, Tam,Morris, Lee, Lau, &Chiu, 2009). Given the low reliability, we examined the avalues separately for U.S. and international students andfound higher values for U.S. students (n 5 31; a 5 .69)relative to international students (n 5 66; a 5 .35).


free-response affective associations with delegation.We manipulated the gender of the subordinate, butdid not expect effects of subordinate gender on af-fective associations with delegation. We tested Hy-pothesis 2a, that women would have more negativeassociations relative tomen in thedelegate conditionas compared to the do-not-delegate condition. Fur-ther, we explored the tenor of women’s negative af-fective associations by examining whether men andwomen differentially associate delegation with guilt(Hypothesis 2b). To explore the possibility thatwomen’s negative associations with delegation aredue to concerns that others will view them as lesscompetent (i.e., unable to handle their workload) ortheir subordinates as less competent, we includedcompetence measures as control variables.

Method

Participants. Two hundred participants (93women; Mage: 36.42 years) were recruited from Am-azon’s Mechanical Turk (MTurk). The sample was77%Caucasian, 10%Black, 6%Hispanic, 6%Asian,and 3% other, with an average of 15.06 (SD 5 8.09)years of work experience. Participants resided inthe United States and 90% had post-secondaryeducation.

Procedure. The study was described as an exam-ination of workplace experiences. Participants weretold to imagine that they were an employee ata marketing company specializing in telecommuni-cations, and that their position involved managing,coordinating, and executing the company’s strategicvision. Participants were told that they personallyhad a busyweek ahead as theywere launching a newmarketing campaign, and that a junior employeewhoworkedunder them (Michael orMichelle) couldhelp them complete their tasks.

Measures. Participants first wrote about their af-fective associationswith delegating or not delegatingand responded to measures capturing guilt andcompetence on scales ranging from 1 (strongly dis-agree) to 7 (strongly agree).

After reading the scenario, participantswere askedto write about how they felt about delegating ornot delegating to their employee, and to list anythoughts, feelings, or adjectives associatedwith theirbehavior (delegating or not delegating).

To capture guilt that may be associated withoverburdening an employee through delegation (orthe relief of not delegating and not overburdening anemployee), participants in the delegate conditionwere asked towhat extent they felt. . . “baddelegating

because your subordinate has other things to do,”“some guilt delegating because you think your sub-ordinate might be overburdened,” and “concernedthat your subordinate will have to stay at work latebecause they are working on your tasks.” Partici-pants in the do-not-delegate condition were askedabout the extent to which they felt. . . “good not del-egating because your subordinate has other things todo,” “some relief not delegating because you thinkyour subordinate might be overburdened,” and“happy that your subordinate will not have to stay atwork late because they are working on your tasks”(a 5 .93).

To capture whether women’s negative associa-tions with delegation were influenced by the con-cern that others would perceive them as lesscompetent for delegating or not delegating, partici-pants were asked, when deciding whether to dele-gate, the extent to which they would. . . “think thatothers might see you as less competent?” “feel likeyou cannot manage your time effectively at work?”“feel like it seems like you can’t handle the job onyour own?” “worry that your employees might viewyou less favorably?” “think that youmight not get thecredit you deserve from others at work?” and “thinkthat you won’t get respect you want from others atwork?” (a 5 .95). This scale was used as a controlvariable in all analyses.

We also asked three questions to capture whetherwomen’s negative associations with delegationmight be influenced by concerns about the compe-tence of their subordinate: in deciding whether todelegate, to what extent would the participant. . .“think you could accomplish the tasks much fasteron your own?” “think you could accomplish thetasks much better yourself?” and “worry about thetasks gettingdone quickly and effectively?” (a5 .71).This scale was used as a control variable in allanalyses.

Participants were asked to identify the name oftheir junior employee, their department, and theirrole from a number of options.

Data-analysis strategy. Written responses wereanalyzed using Linguistic Inquiry Word Count(LIWC) software (Pennebaker, Francis, & Booth,2001). LIWC computes percentages of words ina text within specified categories. Given our interestin affective associations with delegation, we ana-lyzed negative and positive emotions. We analyzedeach participant’s response separately to yield thepercentage of words pertaining to negative emo-tions and positive emotions. Due to the skewness ofthe distribution of the LIWC output, data were


transformed with a square-root function to attain anormal distribution (Tabachnick & Fidell, 2007).

Results

Participant attrition. Overall, participants accu-rately reported the name of their junior employee,their department, and their role; 86% passed the at-tention checks. The final sample therefore included171 participants (83 women; Mage 5 37.0 years).

Affective associations with delegation. On av-erage, participants used 19.5 (SD 5 14.18) words intheir free-response text, 95% of which were recog-nized by the LIWC program.2 We first examined thethree-way interaction of participant gender, delega-tion condition, and subordinate gender on negativeand positive affect; as expected, we found no signif-icant effects.3 We therefore collapsed across sub-ordinate gender for all subsequent analyses.Next,weexamined the effects of gender, delegation condition,and their interaction on negative affect. There werenomaineffectsofgender,F(1,170)5 .49,p5 .49,hp

25.003, or delegation condition, F(1, 170) 5 .19, p 5.66, hp

2 5 .001, on negative affect. However, con-sistentwith our prediction,weobserved a significantgender3delegationcondition interaction,F(1, 170)54.67, p 5 .03, hp

2 5 .03. As shown in Figure 1a,women who were told to delegate associated dele-gation with greater negative affect (M 5 1.36, SD 51.91) relative to men who were told to delegate (M5.66, SD 5 1.38), F(1, 170) 5 4.47, p 5 .04, hp

2 5 .03.There was no difference in negative affect betweenmen (M 5 1.41, SD 5 1.70) and women (M 5 .84,SD5 1.54) told not to delegate, F(1, 170)5 0.86, p5.35, hp

2 5 .005. Nor were there significant differ-ences between the delegate and do-not-delegate con-ditions for women (p5 .23) or for men (p5 .07).

We had intended to examine the negative affectexperienced by women with greater granularity byanalyzing the LIWC negative emotion subscalesof anger, sadness, and anxiety. However, becauseonly 5%, 8%, and 20% of the sample listed wordsassociated with anger, sadness, and anxiety, re-spectively, we did not have sufficient statisticalpower to analyze these subscales.

Wenext examined the effects of gender, delegationcondition, and their interaction on positive affect.

We observed no significant effects of gender, F(1,170)5 0.24, p5 .62, hp

2, .01, delegation condition,F(1, 170) 5 3.25, p 5 .07, hp

2 5 .02, or gender 3delegation condition interaction on positive affect,F(1, 170) 5 0.30, p 5 .58, hp

2 , .001. (Figure 1b).Guilt of overburdening subordinates.4 Our ex-

amination of guilt associated with delegating (orrelief associatedwithnot delegating) revealed amaineffect of delegation condition, F(1, 170)5 34.86, p,.0001, hp

2 5 .16, such that those in the do-not-delegate condition felt more relief from not having todelegate (M55.10,SD51.54) relative to the guilt feltby those in the delegate condition (M 5 3.65, SD 51.59).We observed no effects of gender on feelings ofguilt or relief, F(1, 170) 5 0.15, p 5 .70, hp

2 5 .001,nor was there a significant gender 3 delegationcondition interaction on guilt or relief, F(1, 170) 50.05, p 5 .82, hp

2 5 .0002.

Discussion

Supporting Hypothesis 2a, Study 2 showed thatwomen have different affective associations withdelegation compared to men, such that women asso-ciate delegation with more negative affect comparedto men. This finding is consistent with researchdemonstrating that engaging in counter-stereotypical

2 We did not include word count as a covariate, but re-sults do not change when including word count in theanalysis.

3 Correlations across all study variables are availableupon request from the first author.

4 We thank two anonymous reviewers for raising thepoint that gender differences in delegation may be attrib-uted to concerns about being perceived as less competentfor delegating. To rule out this alternative explanation, inStudy 2 we tested whether women were more concernedabout their own competence and their subordinate’scompetence compared to men. Counter to prevailing hy-potheses, men were more concerned about how theircompetence would be perceived (M 5 2.68; SD 5 1.47)thanwerewomen (M5 1.82; SD5 1.09), F(1, 170)5 18.71,p, .001,hp

25 .10, collapsing across conditions.Menwerealso more concerned about their subordinate’s compe-tence (M 5 4.03; SD 5 1.30) than were women (M 5 3.57;SD5 1.56), F(1, 170)5 4.45, p, .04, hp

2 5 .03, collapsingacross conditions. Further, we included competence con-cerns (both self and subordinate) as control variables in ouranalysis and saw no changes in our results. Third, ina separate study we asked 46 female managers (Mage 544.76; SD5 8.07) to rate the extent towhich their concernsabout their competence, their subordinate’s competence,and guilt of overburdening subordinates prevented themfrom delegating.We found self-competence concerns to bethe lowest-rated reason for not delegating (M5 2.01; SD51.45), relative to guilt (M 5 3.43; SD 5 1.49) and sub-ordinate competence (M54.32;SD51.40). These findingssuggest that competence is not an alternative explanationfor our effects.


(e.g., agentic) behavior can result in women feelingconcerned about appearing too agentic (Moss-Racusin & Rudman, 2010). Importantly, controllingfor competence perceptions allowed us to rule outthe alternative explanation that women’s negativeassociations with delegation are influenced by con-cerns about being perceived as less competent fordelegating or by concerns regarding their sub-ordinates’ competence.

STUDY 3

In Study 3, we sought to test whether our findingsin Study 2 would replicate in a sample of managersbased on their daily delegation experiences. Wepredicted that women delegators would have more

negative associations with delegation compared tomale delegators (Hypothesis 2a) and that womenwould feel more guilt for overburdening their sub-ordinates compared to men (Hypothesis 2b).

Method

Participants. We recruited 164 participants (79women; Mage 5 32.90) from MTurk. Since holdinga managerial role was a requirement for participa-tion, participants were asked to indicate whetherthey were in a managerial role, list their job de-scription, and report the kinds of tasks they tended todelegate. The sample consisted of 77% Caucasian,8%Asian, 7% Black, and 5%Hispanic participants,with an average of 14.33 years of work experience.

FIGURE 1Study 2: Percentage of (A) Negative Affect and (B) Positive Affect Words Used by Men and

Women by Delegation Condition

0.0

0.0

0.5

1.0

1.5

2.0

2.5

3.0

A

B 3.5

Did Not Delegate

Did Not Delegate

Delegation Condition


Delegated

Delegated

% P

osit

ive

Aff

ect

Wor

ds

(Squ

are

Roo

t)%

Neg

ativ

e A

ffec

t W

ord

s(S

quar

e R

oot)

Male

Male

Female

Female

0.5

1.0

1.5

2.0

2.5

3.0

3.5

Note: Error bars show standard errors.


Participants resided in the United States and 51%had at least some post-secondary education. Partic-ipants on average had worked at their organizationfor 5.76 (SD 5 4.80) years, had 14.23 (SD 5 39.17)people working under them, and worked 42.15(SD 5 9.93) hours per week.

Procedure.The studywas described as examiningday-to-day experiences and behaviors at work. Par-ticipants were first asked about their job descriptionand tasks that they had authority to delegate. Thosewith the authority to delegate (our key exclusioncriterion) were allowed to proceed and were askeda number of questions about their self-reported del-egation behavior, as well as their associations withdelegation.

Measures. Participants were asked to respondto self-report measures on scales ranging from 1(strongly disagree) to 7 (strongly agree). To test par-ticipants’unprimedassociationswithdelegation,weasked them to list any thoughts, feelings, and emo-tions they associated with delegation in an open-ended format. We asked two questions designed tocapture guilt that may be associated with delegationand overburdening an employee: “I sometimes feelbad delegating because my employee has otherthings to do” and “I feel some guilt delegating be-cause my employee might be overburdened” (r 5.81). We also asked two questions on how muchparticipants delegate (r 5 .59): “Compared to otherpeople in your position, how much do you delegateto your employees?” and “How effective do youthink you are at delegating?” As an attention checkduring the study, participants received a question,embedded within our scales, asking them to click“strongly disagree” to show that they were payingattention.

Data-analysis strategy. As in Study 2, we ana-lyzed written responses using LIWC. We tabulatedthe overall number of positive and negative emotionwords for each response and examined three

subscales of negative emotions: anger, sadness, andanxiety. Due to skewness, data were transformedwith a square-root function. Additionally, we con-ducted analyses of variance on each of the measureslisted above to examine gender differences on eachvariable.

Results

Participant attrition.Nineteenparticipants (12%)were removed for lacking authority to delegate. Ofthe remaining participants, 14 (10%) did not reporttheir work experience or failed the attention check.The final sample was 130 participants (66 women;Mage 5 34.02 years).

Affective associations with delegation. To testwhether men and women had different associationswith delegation, we examined participants’ free-response texts. On average, participants used 9.78(SD 5 13.97) words, and LIWC captured 90% of thewords used. An examination of the captured nega-tive and positive affect words associated with dele-gation revealed that, consistent with Hypothesis 2a,there was a significant main effect of gender onnegative affect, F(1, 128)5 6.99, p5 .009, hp

2 5 .05,such that women listed significantly more negative-affect-based words (M 5 3.31, SD 5 3.57) than didmen (M 5 1.87, SD 5 2.55). Means by gender for allkey variables are presented in Table 1.

An examination of the three subscales of negativeaffect revealed that only one participant (, 1% ofsample) used “sad” or “angry” affect-based words;as such, we did not analyze these subscales. A largerproportion of participants (40%) used anxiety-related words, which allowed us to analyze theLIWC anxiety subscale. We observed a significantmain effect of gender on anxiety, F(1, 128) 5 10.26,p5 .002,hp

25 .07), withwomen listing significantlymore anxiety-based words (M 5 2.93, SD 5 3.58)compared to men (M 5 1.26, SD 5 2.16).

TABLE 1Study 3: Means by Gender and Analysis of Variance Results for Key Dependent Variables

Men Women

M SD M SD F p hp2

Self-Reported Delegation 4.64 1.18 4.78 1.08 0.50 0.480 0.00Negative Affect (Sqrt) 1.87 2.55 3.31 3.57 6.99 0.009 0.05Anxiety (Sqrt) 1.26 2.16 2.93 3.58 10.26 0.002 0.07Positive Affect (Sqrt) 3.82 3.27 2.77 3.02 3.63 0.059 0.03Guilt 4.02 1.53 4.55 1.42 4.07 0.046 0.03

Note: n 5 130.


Finally, therewas amarginally significant effect ofgender on positive affect, F(1, 128) 5 3.63, p , .06,hp

2 5 .03, such that women listed marginally fewerpositive-affect-based words (M 5 2.77, SD 5 3.02)compared to men (M 5 3.82, SD 5 3.27).

Guilt of overburdening subordinates.Therewas asignificant main effect of gender on guilt, F(1,129) 54.07, p , .05, hp

2 5 .03, such that women expressedgreater guilt about their employee being over-burdened (M 5 4.55, SD 5 1.42) compared to men(M5 4.02, SD5 1.53), supporting Hypothesis 2b.

Self-reported delegation. We examined whetherself-reported delegation behavior differed betweenmen and women. Interestingly, we observed nogender differences in participants’ reporting of howmuch they delegated, F(1, 128)5 .50, p5 .48, hp

2 ,.00,withmen (M5 4.64,SD5 1.18) reporting similarlevels of delegation as women (M5 4.78, SD5 1.08).We examined the two measures (i.e., how muchparticipants delegated and how effective they wereat delegating) independently and observed no sig-nificant effects (pvalues . .19).

Discussion

Consistent with Study 2, the results of Study 3offer support for Hypothesis 2a and demonstratethat women associate delegation with more nega-tive affect, namely anxiety, as well as with lesspositive affect, than do men. Further, we extendedthese findings to show that women associate del-egation with greater guilt compared to men withregard to overburdening subordinates, supportingHypothesis 2b. Although Study 3 offers pre-liminary evidence of gender differences in dele-gation, we found that women’s self-reports ofdelegation indicated that they did not delegateany less than men. Despite the lack of gender

differences in our self-reported measure, we didfind that negative affect was negatively correlatedwith self-reported delegation (see Table 2); thosewho had stronger negative associations with dele-gation self-reported less delegation behavior.Since women have more negative associationsrelative to men, this finding suggests an indirectrelationship between gender and self-reporteddelegation through negative associations. Giventhe limitations of self-report behavioral measures(Baumeister, Vohs, & Funder, 2007) and the sub-jective nature of our measures, which allowedparticipants to qualify the quality and quantity oftheir delegation, it is possible that the items weselected to capture delegation were not sufficientto capture differential delegation behavior be-tween women and men. Further, the job charac-teristics of the sample (i.e., industry, tenure,occupation, position) varied significantly; thus,a variety of important employment circumstancesmay have affected men’s and women’s self-reported delegation. Therefore, in Study 4, we ex-amined gender differences in actual delegationbehavior and controlled for potential confoundspresent in Study 3.

STUDY 4

In Study 4,we turned our attention to the impact ofnegative associations with delegation on women’spropensity to delegate. In addition, we examinedhow these negative associations affected men’s andwomen’s interactions with their subordinates andinfluenced their performanceona set of tasks. To thisend,we studied delegation in a controlled laboratoryexperiment employing a confederate as a sub-ordinate. We sought to replicate and extend Hy-pothesis 2a by testing whether women feel more

TABLE 2Study 3: Means, Standard Deviations, and Correlations

M SD 1 2 3 4 5 6

1. Gender 1.51 0.50 –

2. Negative Affect (Sqrt) 2.60 3.18 .23** –

3. Anxiety (Sqrt) 2.11 3.07 .27** .91** –

4. Positive Affect (Sqrt) 3.28 3.18 –.17† –.36** –.29** –

5. Guilt 4.29 1.49 .18* .17† .18* .07 (0.81)6. Self-Report Delegation 4.71 1.13 .06 –.21* –.12 .11 –.17* (0.59)

Notes: n5 130. Gender is coded such that 15Male and 25 Female. Reliabilities appear on the diagonal in parentheses (where applicable).† p , .10*p , .05

**p , .01


negative affect while delegating, and to providerichness to our measures by collecting emotionsbefore and after the opportunity to delegate (ratherthan through affective associations devoid of con-text). Specifically, wemeasured anxiety before andguilt after delegation, and included a fear-of-backlash scale to test our argument that the anxi-ety associated with delegation is due to a fear ofbeing perceived as too agentic. Finally, we exam-ined the downstream consequences of havingnegative associations with delegation, testing ourpredictions that women will be less likely to del-egate than men (Hypothesis 3a), that women willfear backlash for delegating more than men will(Hypothesis 2c), and that women’s interactionswith their subordinates when delegating will beperceived as being of lower quality than men’s(Hypothesis 3c). We further explored whetherwomen’s and men’s differential negative associa-tions with delegation will mediate their likelihoodof delegating (Hypothesis 3b), and the implicationsof delegation on actual performance in this context(Hypothesis 3d).

Method

Participants. We recruited 148 participants (81women;Mage5 22.77) from a behavioral research labat a large, private university. Participants werescheduled for a 90-minute session and paid $20 plusa $4 performance bonus. The majority (62%) ofparticipants were undergraduate students, while theremaining 38% were graduate students or commu-nity members. The sample was 77% Caucasian and11%Asian, while 3% identified with more than onerace.

Procedure. The study, described as being aboutidentity and performance, had three phases. InPhase 1, participants first answered personalityquestionnaires. In Phase 2, they engaged in an ac-tivitywith apartner (a trained confederate).5 InPhase3, participants completed a questionnaire about theirinteractions and experience with their partner dur-ing Phase 2.

Phase 1: Participants entered the lab with a personwho was ostensibly another participant (confeder-ate); the twowere escorted into separate rooms. Afterproviding consent, participants were asked to com-plete an “Activity Role Assessment” and told that

this assessment would be used to determine theirroles in an upcoming leadership activity.6

Participants were then told that, based on theirassessments, theywouldbeput into the senior role of“Sponsorship Coordinator” for a “University GoesGreen” campus campaign and that their partner(confederate) would be put into the junior role of“Committee Member.” This paradigm has beenfound to create status and power differences inexperimental settings (Akinola & Mendes, 2013).Participants were told that both they and the “Com-mittee Member” had six tasks to complete, but thattheir tasks and compensation would differ based ontheir leadership roles, and as the leader, they coulddelegate to their partner if they would like to. Theconfederate then went to a separate workspace andparticipants completed questionnaires assessingtheir associations with delegation, and their inten-tions to delegate in the upcoming activity.

Phase 2: Participants were given 35 minutes tocomplete an activity consisting of six tasks in their roleof sponsorship coordinator: (1) draft an email to thestudentbodyabout thecampaign, (2) enterdata intoanExcel spreadsheet, (3) redact or interpret the data en-tered into the Excel spreadsheet, (4) proofread a Pow-erPoint presentation, (5) format a newsletter, and (6)design a logo for the campaign (see Appendix A fordetails on each task). Eachparticipantwas videotapedduring the session in an effort to capture his or herdelegation behavior. After 35 minutes had elapsed,participants were informed that the activity was over.

Phase 3: Participants then filled out a final ques-tionnaire and were debriefed, thanked, and paid.

Dependent measures prior to the delegation activ-ity (Phase 1). Prior to engaging in the delegationactivity, participants completed the role assessmentquestionnaire as well as questionnaires assessingtheir affective associations with, and likelihoodof, delegating.

This questionnaire served as a cover story forassigning participants to the senior role of “Sponsor-ship Coordinator.” The questionnaire included theTen Item Personality Inventory (Gosling, Rentfrow, &Swann, 2003) and three Rorschach Inkblot Tests

5 Wemanipulated and controlled for confederate genderin all analyses.

6 While participants waited for their assessments to bescored, they answered questions about themselves foreight minutes either alone or with the confederate. Wewere interested in whether forming a relationship with anemployee would affect women’s propensity to delegate.We observed no effect of this manipulation on our keydependent variables. In all analyses, we controlled for self-reflection condition and confederate gender.


(Karson, 2007), in which participants viewed ambig-uous pictures and described what they saw.

To assess the changes in emotions that may ac-company delegation, we assessed participants’ self-reported emotions using the Positive and NegativeAffect Schedule (PANAS) (Watson & Clark, 1994) attwo points: (1) upon arrival at the lab in the activityrole assessment questionnaire (Time 1) and (2) afterlearning about the delegation activity and their lead-ership role (Time 2). Participants rated themselves on15 negative and 19 positive items. However, we wereparticularly interested in the anxiety subscale of thePANAS given our findings that women had morenegative associations, namely anxiety, with delega-tion in Study 3. Participants rated how nervous,afraid, and scared they were on scales ranging from 1(not at all) to 5 (a great deal), resulting in an anxietyindex (alphas ranged from .71 to .85). We also exam-ined three positive emotions (confident, proud, andstrong), as these words were most relevant based ontheopen-endedcontentofStudies2and3, resulting ina positive affect subscale indexing agentic emotions(alphas ranged from 78 to .82).We calculated changesin emotions by subtracting participants’ emotion rat-ings just prior to the delegation task (Time 2) fromthose collected upon arrival at the lab (Time 1).

Dependent measures during delegation activity(Phase 2). We assessed whether participants dele-gated, examined their behavior toward their sub-ordinatewhen they did delegate, andmeasured theirperformance. We assessed delegation by havingconfederates indicate whether the participant dele-gated any tasks to them and how many tasks theydelegated. Additionally, independent coders exam-ined each videotape to confirm whether the partici-pant delegated. For each task that was delegated, wemeasured the number of seconds that elapsed be-tween the time the participant entered the confed-erate’s room to delegate a task and the time theparticipant exited the room.

We assessed the quality of participants’ interac-tions with the confederate7 in two ways, using items

based on the “showing concern” dimension of theempowering leadership questionnaire (Arnold,Arad, Rhoades, & Drasgow, 2000). First, immedi-ately following each interaction, we asked eachconfederate to rate the participant on six questions:(1) “the participant seemed to consider how theirdelegation would affect me,” (2) “the participantkept me motivated throughout the task,” (3) “I feltlike the participant had confidence in me to com-plete the tasks,” (4) “I felt like the participant trustedme to complete the tasks,” (5) “the participant fol-lowed up with me to make sure I was completing mytasks,” and (6) “the participant checked in on me tomake sure I was completing my tasks.” Ratings wereon scales from 1 (strongly disagree) to 7 (stronglyagree) (a 5 .68).

Second, five research assistants watched eachparticipant’s video and rated participants on thefollowing four questions: To what extent did theparticipant in the video. . . (1) “ask about the con-federate’s to-do list to see if they had the capacity totakeon the task,” (2) “compliment the confederate onhis or her ability to accomplish the task,” (3) “expressgratitude for taking on or finishing the task,” and (4)“offer to help, if there was extra time.” Coders ratedthese items, which capture how considerate theparticipant appeared, on scales ranging from1 (not atall) to 7 (very much). Reliability between the coderswas high (a5 .88), as was the reliability between theitems (a 5 .90). Further, confederate ratings of theirinteractions and coder ratings were strongly corre-lated, r 5 .57, p , .001, indicating consistency intheir perceptions of interaction quality.

Each of the six tasks was scored by two researchassistants on several dimensions (see Appendix A).We created an average performance score by stan-dardizing thedifferentmetrics of performance acrossour tasks (using a z-score). Participants who dele-gated a task to the confederate received full pointsfor that task.

Dependent measures following the delegationactivity (Phase 3).8 To understand the theoreticaldrivers of negative associations found in previousstudies, we asked participants who had delegated tocomplete a questionnaire assessing guilt and fear ofbacklash.

Wemeasured guilt associatedwith overburdeningwith the same two items adapted fromStudy 3 (“I feltbad delegating because the Committee Member had

7 To ensure gender-role stereotyping did not explain theratings of interaction quality, we transcribed participants’interactions and had two research assistants, blind toparticipant gender, rate each transcript on the same di-mensions on which participants were rated by confeder-ates and video coders. Blind transcript coders’ ratingswere significantly correlated with those of confederates(r5 .39, p, .001) and with those of video coders (r5 .46,p , .001), suggesting that ratings of consideration werenot unduly influenced by gender-role stereotypes.

8 Since we were interested in men’s and women’s af-fective responses to delegation, these items were onlyasked of participants who delegated.


other things to do” and “I felt some guilt delegatingbecause I thought the Committee Member might beoverburdened”) using scales ranging from1 (stronglydisagree) to 7 (strongly agree). We also added a thirditem, “I was concerned that the Committee Membermaynot receive their full payment because theywereworking on my tasks” (a 5 .82).

To assess participants’ concerns about backlash,they were told to imagine that a group of people waswatching a tape of their interaction with the con-federate and asked to what extent they would beconcerned that. . . “[they] might be disliked?” “peo-ple would think [they] dominated the interaction?”“others would perceive [them] as too controlling?”and “[they] might be judged for being too assertive?”(a5 .83). These items were anchored at 1 (not at all)to 7 (very much) (see Moss-Racusin & Rudman,2010).

Results

Participant attrition. Eleven participants wereremoved from the analysis due to their suspicionsabout their interaction with the confederate, as dis-covered in participant debriefing and in confederatenotes. The final sample included 137 participants(75 women; Mage 5 21.97). Additionally, threevideos were not coded due to technical malfunc-tions, and one participant did not complete thePANAScorrectly. Varying degrees of freedom reflectthis data loss.

Our primary prediction was that women wouldfeel greater increases in anxiety relative tomenwhenanticipating delegation. In line with our hypothesis,there was no main effect of gender on changes inanxiety, nor was there a main effect of delegation onchanges in anxiety (ps . .54). However, we did ob-serve a significant gender 3 delegation interaction,F(1, 130) 5 4.04, p , .05, hp

2 5 .03. See Figure 2a.Planned contrasts revealed that women who

delegated had greater increases in anxiety prior tobeing given the opportunity to delegate (M 5 .64,SD 5 .88) than women who did not (M 5 .31, SD 5.68), F(1, 130) 5 3.57, p 5 .06, hp

25 .03. For men,there were no differences in anticipatory anxietybased on whether they chose (M5 .28, SD5 .71) orchose not (M5 .51,SD5 .90) to delegate,F(1, 130)51.07, p 5 .30, hp

2 5 .01.We conducted a similar analysis for our positive

emotions subscale. We found no main effect of gen-der, F (1, 130) 5 .83, p 5 .36, hp

25 .01. However,there was a main effect of delegation on positive af-fect, F(1, 130) 5 4.67, p 5 .03, hp

25 .04, such that

those who delegated felt more positive emotionsprior to being given the opportunity to delegate(M 5 .11, SD 5 .67) than did those who did notdelegate (M52.10, SD5 .67). This main effect wasqualified by a gender 3 delegation interaction, F(1,130)53.90,p5 .05,hp

25 .03.As shown inFigure 2b,planned contrasts reveal that for women there wasno difference in positive affect prior to being giventhe opportunity to delegate, regardless of whetherthey chose (M5 .04, SD5 .64) or chose not (M5 .02,SD5 .65) to delegate, F(1, 130)5 .02, p5 .88, hp

2 ,.01. In contrast, men who did not delegate experi-enced significantly greater decreases in positive af-fect prior to being given the opportunity to delegate(M 5 2.32, SD 5 .67), and men who did delegateexperienced increases in positive affect prior to be-ing given the opportunity to delegate (M 5 .17, SD 5.70), F(1, 130)5 7.28, p 5 .01, hp

2 , .05.We used a binary logistic regression to test

whether men and women differentially delegatedand observed a significant main effect of gender,such that women (51%) delegated less than did men(68%), B 5 2.72, SE 5 .36, p , .05, Wald z 5 4.01.Interestingly, for those who delegated, there was nodifferencebetweenmenandwomen in the amount ofitems delegated, F(1, 76) 5 .42, p 5 .52, hp

2 , .01.There was a significant main effect of gender on

the time spent interactingwith the confederatewhiledelegating, F(1, 70) 5 4.85, p 5 .03, hp

25 .07,9 withwomen spending fewer seconds with the confeder-ate when they delegated (M 5 33.28, SD 5 25.98)compared to men (M 5 54.20, SD 5 59.68).

There was also a significant main effect of genderon confederates’ ratings of their interactions withparticipants, F(1, 75) 5 11.07, p 5 .001, hp

25 .13).Confederates rated female participants as being lessconsiderate, trustworthy, motivating, and support-ive (M 5 3.25, SD 5 1.01) compared to male partic-ipants (M 5 3.96, SD 5 1.15). Additionally,independent coder ratings yielded a marginally sig-nificant effect of gender with regard to how consid-erate the participant appeared, F(1, 73) 5 3.48, p ,.07, hp

2 5 .05. Namely, the coders rated female par-ticipants as less considerate (M 5 2.94, SD 5 1.22),compared tomale participants (M5 3.44, SD5 1.16)in their interactions.

While we observed no effect of gender on totalperformance score,F(1, 129)5 .02,p5 .89,hp

2, .00,there was a significant main effect of delegation on

9 We controlled for number of items delegated to theconfederate, which could have affected the amount of timespent interacting with the confederate.


performance, F(1, 129) 5 10.44, p 5 .002, hp2 5 .08.

Participants who chose to delegate performed sig-nificantly better (M 5 .11, SD 5 .45) than those whodid not delegate (M 5 2.15, SD 5 .44).

In sum, we found that women delegated less thanmen when given the opportunity to do so. Further,women spent less time and showed less consider-ation toward their subordinate when they did

FIGURE 2Study 4: Change in (A) Anxiety and (B) Positive Affect for Men and Women by Delegation

Ch

ange

in

Pos

itiv

e A

ffec

t (T

2–T

1)C

han

ge i

n A

nxi

ety

(T2–

T1)

Did Not Delegate Delegated

Did Not Delegate Delegated

Male

Female

MaleFemale

.90

.80

.70

.60

.50

.40

.30

.20

.10

.00

.60

.40

.20

.00

-.20

-.40

-.60

-.80

A

B



delegate. Finally, we found that delegation influ-enced performance, as those who delegated com-pleted more tasks.

Dependent measures following the delegationactivity were only given to those who delegated, asthe questions were specific to their interaction andhow they felt about delegating. There was a signifi-cant main effect of gender for those participantswho delegated in the amount of guilt they felt aboutoverburdening their subordinate, F(1, 76) 5 5.24,p , .03, hp

2 5 .06, such that women who delegatedfelt more guilt (M 5 5.36, SD 5 1.41) compared tomen who delegated (M 5 4.65, SD 5 1.49). Therewas also a significant main effect of gender for par-ticipants who delegated in terms of how much theyfeared others viewing them negatively, F(1, 75) 54.26, p 5 .04, hp

2 5 .05, such that women whodelegated felt more fear of backlash (M 5 2.91,SD 5 1.38) than did men who delegated (M 5 2.34,SD 5 1.10).

Given that we found significant differences inmen’s and women’s positive and negative affectiveresponses to delegation, as well as their likelihoodto delegate, and delegation behavior, it is possiblethat men’s and women’s differential affective re-sponses mediated their likelihood of delegating(Hypothesis 3b) and associated behaviors. Wetested for the presence of an indirect effect of genderon delegation and interaction quality by calculating95% confidence intervals (CIs) using the PROCESSmacro for SPSS (Hayes, 2013) with 5,000 bootstrapresamples. We specifically examined negative andpositive affect as mediators predicting delegationand interaction quality, as both of these measureswere assessed prior to participants decidingwhether to delegate. We were unable to examineguilt and fear of backlash as mediators, as they wereassessed after delegation and only for those whodelegated.

Results from mediation analyses examining neg-ative and positive affective responses to delegation,using binary logistic regression, indicated that neg-ative affective responses did not mediate the effectsof gender on delegation (CI95 5 2.02, .23); nor didpositive affective responses (CI95 5 2.12, .18), asboth confidence intervals contained zero. Negativeaffective responses also did notmediate the effects ofgender on the quality of women’s interactions withsubordinates when delegating (CI95 5 2.007, .28),nor did positive affective responses (CI95 5 2.21,.03). Despite our nonsignificant mediations, wefound a marginal relationship between genderand interaction quality through anxiety (indirect

effect 5 .07, SE 5 .07, CI90 5 .002, .24). Though in-conclusive, this finding suggests that greater anxietymay influence women’s delegation behavior withsubordinates, even if it does not influence their de-cision to do so.

Discussion

The goal of Study 4 was to replicate our previousfindings that women havemore negative associationswith delegation, and to expand on these findings toshow that these negative associations have implica-tions for actual (and not just self-reported) delegationbehavior. In a controlled experiment we found thatwomen delegated less than did men, consistent withHypothesis 3a, and that when women did delegate,they spent less time with their subordinates and hadlower-quality interactions, a finding aligned withHypothesis 3c. Further, we offer evidence of the dif-ferent types of negative associations women havewith gender-role-paradoxical behaviors by showingthat women experience relational guilt and fear ofbacklash about delegating, supporting Hypotheses 2band 2c. Notably, keeping performance of the confed-erate constant, we found that participants who diddelegate performed better on the task than those whodid not, as they had more time to accomplish theirtasks. Specifically, women delegators outperformedwomen who chose not to delegate. These findingsreplicate prior research showing the adaptive natureof delegation (Leana, 1987) and offer evidence thatwomen’s negative associations with delegation mayhave negative repercussions for workplace perfor-mance when they choose not to delegate.

The fact that we did not observe a gender effector a significant gender 3 delegation interaction onperformance suggests that women did not performworse than men on the task, even though they dele-gated less; instead, they essentially outperformedmen through their own contributions. However, it isunclear whether women could sustain this out-performance across multiple tasks or days. Becausethe benefits of delegating may be cumulative, easingthe superior’s workload over time, the potentialnegative performance implications for women of notdelegating might be difficult to capture based ona single point in time.

In addition, Study 4 provides insight into the roleaffect plays in delegation decisions. We found thatwomen who delegated self-reported more anxietyprior to delegating than did men. This finding sug-gests that prospective anxiety does not necessarilyhinder women from delegating. On one hand, it is


possible that the negative associations women havewith delegation prevent them from delegating, suchthat women who associate delegation with anxietydo not delegate. On the other hand, it is possible thatsome women press ahead and delegate despite feel-ing anxious about doing so. Our results in Study 4suggest that prospective anxiety may not only in-fluence the decision to delegate, but also has impli-cations for the quality and quantity of women’sinteractions with their subordinates when theydo delegate. Nonetheless, we did not findconclusive—though we did find suggestive—evidence that prospective anxiety (i.e., negative af-fect) mediated the effects of gender on delegationbehavior.

Given our findings in Study 1 demonstrating thatwomen perceive delegation as more agentic thanmen, our findings in Study 4 suggest that womeninternalize these social role perceptions, as evidentin their emotions before, their behavior during, andtheir reflections after delegating. Women’s concernsabout violating their social role may degrade theirinteractions with subordinates, leading them to rushthrough delegating (evidenced by shorter interactiontimes) and to become more self-focused (evidencedby being rated as less considerate) as compared tomen. Interestingly, men who did not delegate expe-rienced less positive affect relative to those who diddelegate. These male nondelegators violated genderroles by not acting agentically, which influencedtheir emotions. This finding supports the conceptthat gender roles play an important part in responsesto delegation behavior for both men and women.

STUDY 5

In Study 5, we sought to explore one possiblefactor that might attenuate gendered responses todelegation behavior, particularly for women. Be-cause we theorized that women have negativeassociations with delegation due to its gender-role-paradoxical nature of being communal yet agentic,we tested whether framing delegation as morecommunal would attenuate women’s negative as-sociations with delegation. We examined whetherhighlighting how delegation can empower and helpsubordinates would reduce women’s negative as-sociations with delegation.

Method

Participants. One hundred and thirty-five MBAstudents (47 women; Mage 5 28.11), enrolled in an

introductory organizational behavior course at alarge private university on the East Coast, partici-pated in the study. Fifty-nine students were ina managerial position prior to pursuing their MBA,and participants on average had 14.35 (SD 5 43.84)subordinates. The sample was 46% White, 27%Asian, 11% Hispanic, 3% Black, and 13% other.

Procedure.We told participants that they wouldengage in an in-class feedback-giving and -receiv-ing role play between a manager and subordinate,and randomly assigned them to either be themanager(feedback giver) or the subordinate (feedback re-ceiver). Participants first read their role materials,which provided context for the role play. Sub-ordinates all read the exact same role materials, inwhich they were informed that they would meetwith their manager to receive performance feed-back and that some tasks might be delegated tothem. Managers were told that they would givetheir subordinate performance feedback and thatthere were three tasks that could be delegated tosubordinates. After reading the role materials, butprior to engaging in the role play, both managersand subordinates answered questions about theiremotions.

We manipulated whether managers were told thatdelegating tasks canhelpsubordinatesdevelop, learn,and grow (delegation-as-communal condition) orwere told nothing about the communal benefits thatcould accrue to the subordinate from performingdelegated tasks (control condition). We chosea manipulation that emphasized the communalproperties of delegation, andnot one that emphasizedthe benefits of engaging in agentic behaviors, for tworeasons. First, much research has shown that high-lighting communal properties of agentic behaviorscan attenuate fear and increase the likelihood of en-gaging in such behaviors (see Amanatullah & Morris,2010;Kray, Galinsky, &Thompson, 2002; Small et al.,2007). Thus, we expected highlighting communalityto be beneficial here in the domain of delegation, justas it has been shown to be beneficial for negotiations.Second, agency is generally perceived to be adaptivein leadership roles. Thus, it is not that women do notknow the benefits of engaging in agentic behavior; it ismore likely that gender-role prescriptions and back-lash prevent women from engaging in agentic be-haviors (Rudman&Phelan, 2008).Wewerenot able toinclude other control conditions but see this initialtest of intervention as a first step toward potentiallymitigating gendered concerns about delegation.

Materials and measures. Participants wereasked to read the role play (with the embedded


manipulation for managers) and respond to mea-sures capturing their affective responses, guiltfrom overburdening their subordinate (for man-agers), and evaluations of their manager (forsubordinates).

Participants in the role of manager were ran-domly assigned to either a control condition or adelegation-as-communal condition prior to engag-ing in the feedback role play. These conditionsdiffered only in the instructions participants weregiven at the end of the exercise. All managers weretold the following:

You are planning to delegate a few tasks on your plateto your subordinate. Among these are: (1) preparingmaterials for a client meeting, (2) reconciling ex-penses, and (3) running a training session. You rec-ognize that [your subordinate] has beenworking hardand is very busy, but you need help.

In the delegation-as-communal condition, this state-ment was followed by:

Further, you know that by assigning these tasks, youarebeing a goodmentor, helping [your subordinate] todevelop, and teaching your subordinate critical skillsthat are needed for the senior management position.

As in Study 4, participants rated how nervous,afraid, and scared theywere on scales ranging from 1(not at all) to 5 (a great deal), resulting in an anxietyindex (a 5 .76). We also examined three positiveemotions (confident, proud, and strong), resulting ina positive affect scale (a 5 .67).

To capture guilt associatedwith overburdening anemployee, we asked those in the manager role thethree questions used in Study 4 (a 5 .77) on scalesranging from 1 (strongly disagree) to 7 (stronglyagree).

FIGURE 3Study 5: (A) Anxiety and (B) Positive Affect for Men and Women by Delegation Condition

0.0

2.7

2.8

2.9

3.0

3.1

3.2

3.3

3.4

3.5

3.6

3.7

0.5

1.0

1.5

2.0

2.5

3.0

Delegation as Communal

Delegation as Communal

Control

Control



Male

Male

Pos

itiv

e A

ffec

tA

nxi

ety

A

B

Female

Female



To ensure that the manager and the subordinatewere both acclimated to their respective roles, weasked those in the subordinate role about theirdesire to take on their managers’ tasks using twoquestions, “I am motivated to take on my man-ager’s tasks” and “I am happy to take on my man-ager’s work” (r 5 .71), on scales ranging from 1(strongly disagree) to 7 (strongly agree). Since therewas no manipulation in the subordinate role, wedid not expect any gender differences in thesemeasures.

Results

Affective responses to delegation. Our primaryprediction was that women managers in the controlcondition would feel greater anxiety than those inthe delegation-as-communal condition when antic-ipating delegating to the subordinate. We did notexpect any differences in anxiety by delegationcondition for male managers. In line with our pre-diction, there was no main effect of gender on anxi-ety, F(1, 68)5 0.01, p5 .90, hp

2 , .00, nor was thereamain effect of delegation condition on anxiety, F(1,68) 5 2.68, p 5 .11, hp

2 , .04. However, as can beseen in Figure 3, we did observe a marginally sig-nificant gender 3 delegation condition interaction,F(1, 68) 5 3.55, p 5 .06, hp

2 5 .05.Planned contrasts revealed that women in the

control condition had greater anxiety prior to dele-gating (M 5 2.31, SD 5 .76) than women in thedelegation-as-communal condition (M5 1.67, SD5.55),F(1, 23)54.75,p5 .03,hp

25 .07. Formen, therewere no differences in anxiety by delegation condi-tion, F(1, 44) 5 0.04, p 5 .84, hp

2 5 .00.We conducted a similar analysis for our positive

emotions subscale and found no main effects orsignificant interactions (ps . .29). We also con-ducted this analysis for subordinates’ ratings ofanxiety and positive emotions. As expected, be-cause there was nomanipulation in the subordinaterole, we found no differences between men andwomen (ps . .57).

Guilt of overburdening subordinate (managersonly). There was a significant main effect of delega-tion condition in the amount of guilt managers feltabout overburdening their subordinate in anticipa-tion of delegating, F(1, 68)5 7.04, p, .01, hp

25 .10,such thatmanagers in the control condition feltmoreguilt (M5 4.03, SD5 1.13) than did managers in thedelegation-as-communal condition (M5 3.37, SD51.27). However, there was no main effect of gender,F(1, 68) 5 1.46, p 5 .23, hp

2 5 .02, nor was there

a significant gender3 delegation condition effect forguilt of overburdening one’s subordinate, F(1, 68) 51.99, p 5 .16, hp

2 5 .03.Desire to take on manager’s tasks (subordinates

only). Since there was no manipulation in the sub-ordinate role, we did not expect there to be differ-ences betweenmenandwomen in their desire to takeon their manager’s tasks. Consistent with our ex-pectations, we found no gender differences in thisvariable (p 5 .48).

Discussion

The goal of Study 5 was to demonstrate that asimple manipulation emphasizing the communalaspects of delegation can attenuate women’s neg-ative affective responses to delegation. We foundthat women who were told delegation could helptheir subordinate develop, learn, and grow felt lessanxiety and less guilt about overburdening theirsubordinate prior to delegating, compared towomen who were not given these instructions andcompared to male managers. Taken with Study 4,in which we found that women’s negative associ-ations had implications for their actual delegationbehavior and performance, our findings in Study 5suggest that reducing women’s negative associa-tionswith delegationmay increase their likelihoodof delegating, improving interactions with sub-ordinates and potentially positively affecting theirperformance.

GENERAL DISCUSSION

The goal of the current research was to examinewhether men and women have different associa-tions with delegation, and to explore how these as-sociations may differentially affect their propensityto delegate, behavior during delegation, and ulti-mately performance. To this end, in Study 1we findthat women perceive delegation as more agenticthan men do. In Studies 2 and 3, we find that wo-men associate delegation with more negative affectcompared to men. Although women’s self-reportsof delegation indicated that they did not delegateany less than men, our behavioral examination ofdelegation in Study 4 demonstrated that womendid indeed delegate less than men. Further, whenwomen did delegate, they spent less time with theirsubordinates and had lower-quality interactionscompared to men. Confederate ratings highlightedthe nature of these interactions by demonstratingthat womenwho delegatedwere less likely to check


in and follow up than were men, leading to lower-quality interactions in which the subordinate feltunsupported, less trusted, and less motivated.Moreover, women felt greater guilt after delegatingthan men did and were more likely to fear backlashfor being perceived as too agentic in their delegationbehavior. Importantly, women who did delegateoutperformed those who did not. Finally, in Study5, we show that reframing delegation as communalattenuates women’s negative affective associationswith delegation.

Theoretical Contributions

Collectively, our findings contribute to the litera-ture on gender and leadership in several key ways.First, our findings suggest that even when a leader-ship behavior, such as delegation, could potentiallybe seen as both communal and agentic in nature,women seem to focus more on the agentic charac-teristics than on the communal characteristics. Al-thoughdelegation canbe considered communal, as itcan give subordinates opportunities to engage inhigher-level activities (De Pater et al., 2010), theagentic nature of delegation, in which one exhibitsdominance and power, looms larger for women thanfor men. This perspective may help to explain whywomen experience greater fear of backlash and guiltwhen delegating. Building on the body of researchthat touts the advantages of women’s more commu-nal, transformational style of leadership (Eagly &Karau, 2002), our findings demonstrate that, incertain contexts, even when a behavior could beconstrued and enacted in a communal manner, tra-ditional notions of agency associated with certainbehaviors have a powerful effect on women’s abilityto enact a communal style of leadership, ironicallyresulting in them leading with a less trans-formational and communal style. As women con-tinue to ascend to higher levels in organizations, itwill be important to further develop our under-standing of the role of gender-role incongruence inwomen’s propensity to engage in behaviors that areboth agentic and communal in nature. Beyond del-egation, research on managerial behaviors such asnegotiation, and even the act of leadership itself, of-fer convergent support that women experience be-haviors that are considered both agentic andcommunal differently than men do, in part due togender-role incongruence. Further research isneeded to enhance our understanding of other cru-cial leadership behaviors that fall under this categoryof being gender-role paradoxical.

Second, our findings offer insight into how engag-ing in gender-role-paradoxical behaviors may influ-ence women’s relationships with their subordinatesand subsequent performance. Because effective del-egation requires a productive relationship with sub-ordinates, our finding thatmen andwomen treat theirsubordinates differently when delegating has impli-cations for the development of effective and pro-ductive long-term work relationships. Ironically,rather than exhibiting gender-role-congruent behav-iors, such as showing warmth and sensitivity whendelegating, women appeared less considerate andspent less timewith their subordinatesascompared tomen. Given that employees are affected by the moodand behaviors of their leaders (Sy, Cote, & Saavedra,2005), the behavioral spillover effects of havingmorenegativeand fewerpositive emotionsassociatedwith delegating have important implications forwork-related outcomes, including performance, co-ordination, and relationship development amongemployees. Moreover, our finding that delegatorsperformed better on the task than did nondelegatorsfurther emphasizes the importance of overcomingthese negative emotions associated with delegation.

Interestingly, although our focus was on the im-plications of gender-role-paradoxical behaviors forwomen’s experiences in organizations, we foundthatmenwho did not delegate, presumably violatingtheir expected social role (e.g., being agentic and self-assertive), experienced less positive affect. Giventhat research on gender and leadership has mostlyexamined the repercussions of gender-role conflictfor women (Rudman & Phelan, 2008), our findingsamong males highlight a need to further examinehow gender-role expectations negatively influenceboth women and men, potentially preventing menfrom enacting more communal and less hierarchicalstyles of leadership.

Finally, and critically, our research uncovers animportant remedy that can help encourage womento engage in gender-role-paradoxical behaviors. Con-sistent with research demonstrating the powerful ef-fects of framing in shaping behavior (Kray et al.,2002; Small et al., 2007), we show that women’saffective associations with gender-role-paradoxicalbehaviors can be changed by emphasizing the com-munal properties of these behaviors.

Limitations and Future Directions

While the present studies offer implications forour understanding of gender and delegation, it doesso with several limitations, many of which offer


promising avenues for future research. First, the factthat we did not observe differences in self-reporteddelegation behavior between male and female man-agers inStudy3, but did observedifferences in actualbehavior in Study 4 with a student sample, makes itimportant for future research to test delegation be-haviorally across men and women in organizationswith variedwork experience and across hierarchicallevels within organizations. Given that we did ob-serve differential associationswith delegation acrossmen andwomen in Studies 1 and 3,whichwere bothstudies of managers, it is possible that these associ-ations only affect behavior among younger and lessexperienced managers.

Additionally, delegation is typically predicated onhaving anongoing relationshipwith one’s subordinate.Given that our participants in Study 4 had no historywith one another, it is possible that any delegation be-havior we observed might have been suppressed orexaggerated relative to the behavioral responses thatwould occur in actual work situations (Heilman &Okimoto, 2008). We tried to address this concern byincluding a “fast friends”manipulation (shown to cre-ate close, intimate, and continued friendships insideand outside the laboratory [Page-Gould, Mendoza-Denton, & Tropp, 2008]). Although this manipulationdid affect how close participants felt to their sub-ordinates, it did not affect how much or whether par-ticipants delegated, suggesting that these results mayextend to close relationships outside of the laboratory.Future research could benefit from exploring this phe-nomenon longitudinally, inducing familiarity andobserving delegation behavior over time. Moreimportantly, delegation research in organizationsshould examine thedegree towhich familiarity and thelength of the relationshipmay influence the likelihoodof and comfort with delegation among women.

We did not find conclusive evidence that anxiety(i.e., negative affect) mediated the effects of genderon delegation or delegation behavior. One reasonwhy we did not find the expected mediation may bedriven by the way in which we measured the emo-tions associated with delegation in Study 4. InStudies 2 and 3, participants were asked to list anythoughts, feelings, and emotions they associatedwith delegation in an open-ended format. Thismeasure captures associated anxiety with delega-tion more broadly, whereas in Study 4 participantscompleted a PANAS, which instead captures felt orprospective anxiety. That is, anxiety in the case ofStudy 4 more accurately reflects the anxiety aroundprospective delegation in the moment by those whodo indeed delegate (as shown by the greater anxiety

of women who chose to delegate versus those whochose not to delegate), and does not necessarilyprevent women from delegating. Instead, we wouldexpect that general anxiety associated with delega-tion might prevent women from delegating. Fromthese results, it is clear that distinguishing betweenthese different measures of affect and examining theeffects of associated, prospective, and even retro-spective anxiety on the decision to delegate and ondelegation behavior is an important direction forfuture research.Additionally, future research shouldtest a broader range of negative affective associa-tions, including anticipatory guilt, to gain greaterinsight into the mediating mechanisms underlyinggender differences in delegation.

Finally, we operationalized delegation, particularlyin Studies 2 and4, in amore agenticmanner byplacingemphasis on time scarcity, as both the employee andsubordinatewere told theyhadaverybusyweek aheadand many tasks to complete. Although we chose thismanipulation because managers who are busy andhave many tasks to complete have been shown to del-egate (Yukl&Fu,1999), this setupmayhaveminimizedthe more communal aspects of delegation in partici-pants’minds. Our results in Study 5 highlight the ben-efits of framing delegation as communal for women.Future studies should not only operationalize delega-tion in amanner that emphasizes its communal nature,but also explore different manipulations intended tochange how women perceive delegation. It is possiblethat a noncommunal, and less prosocial, framing ofdelegation could also attenuate its negative effects.Women might respond well to a manipulation thatemphasizes the importance of delegation for gettingtheir ownworkdoneand for signalingwho is in charge.This framing could also legitimize women behavingagentically, giving them confidence to do so, and thusrescripting gender role expectations. Drawing fromnegotiation research demonstrating that when situa-tional ambiguity is reduced, women are more likely tofollow nongender-specific scripts and achieve betternegotiation outcomes (Bowles et al., 2005), we wouldexpect that emphasizing the communal aspects of anygender-role-paradoxical leadership behavior, or sanc-tioning enactment of the agentic aspects, would reduceambiguity about whether it is appropriate to engage inthe behavior, constraining the potential for gender toinfluence its enactment.

CONCLUSIONS

Delegation is a critical managerial skill that canenhance performance and allow leaders to effectively


manage their time.Our findings suggest that ifwomenhavemorenegative associationswith the gender-role-paradoxical behavior of delegating, theymaydelegateless thanmen, a choice that can hinder their ability tocomplete tasks effectively and engage in adaptive andsymbiotic relationships with their subordinates. Be-cause gender roles are evolving (Diekman & Eagly,2000), it is possible that, over time, the gender-roleincongruence evoked through delegation may be-come less pronounced. However, our research im-plies that a more promising avenue for increasingwomen’s propensity to delegate may be to reframetheir associations with delegation, such that theyperceive it as amore communal and selfless behaviorrather than as an agentic and self-serving one.

REFERENCES

Abele, A. E., &Wojciszke, B. 2014. Communal and agenticcontent in social cognition: A dual perspectivemodel.Advances in Experimental Social Psychology, 50:195–255.

Akinola, M., & Mendes, W. B. 2013. It’s good to be the king:Neurobiological benefits of higher social standing. So-cial Psychological & Personality Science, 5: 43–51.

Amanatullah, E. T., & Morris, M. W. 2010. Negotiatinggender roles: Gender differences in assertive negoti-ating are mediated by women’s fear of backlash andattenuated when negotiating on behalf of others.Journal of Personality and Social Psychology, 98:256–267.

Apfelbaum, E. P., Sommers, S. R., & Norton, M. I. 2008.Seeing race and seeming racist? Evaluating strategiccolorblindness in social interaction. Journal of Per-sonality and Social Psychology, 95: 918–932.

Arnold, J. A., Arad, S., Rhoades, J. A., & Drasgow, F. 2000.The empowering leadership questionnaire: The con-struction and validation of a new scale for measuringleader behaviors. Journal of Organizational Behav-ior, 21: 249–269.

Babcock, L., & Laschever, S. 2003. Women don’t ask:Negotiation and the gender divide. New York, NY:Princeton University Press.

Baumeister, R. F., Vohs, K. D., & Funder, D. C. 2007. Psy-chology as the science of self-reports and fingermovements: Whatever happened to actual behavior?Perspectives on Psychological Science, 2: 396–403.

Becker, G. S., &Murphy, K. M. 1992. The division of labor,coordination costs, and knowledge. Quarterly Jour-nal of Economics, 107: 1137–1160.

Bem, S. L. 1981. Gender schema theory: A cognitive ac-count of sex typing. Psychological Review, 88: 354–364.

Bem, S. L. 1983. Gender schema theory and its implicationsfor child development: Raising gender-aschematic chil-dren in a gender-schematic society. Signs, 8: 598–616.

Benetti-McQuoid, J., & Bursik, K. 2005. Individual differ-ences in experiences of and responses to guilt andshame: Examining the lenses of gender and genderrole. Sex Roles, 53: 133–142.

Bowles, H. R., Babcock, L., & Lai, L. 2007. Social incentivesfor gender differences in the propensity to initiatenegotiations: Sometimes it does hurt to ask. Organi-zational Behavior and Human Decision Processes,103: 84–103.

Bowles, H. R., Babcock, L., & McGinn, K. L. 2005. Con-straints and triggers: Situational mechanics of genderin negotiation. Journal of Personality and SocialPsychology, 89: 951–956.

Brescoll, V. L. 2012. Who takes the floor and why: Gender,power, and volubility in organizations. Administra-tive Science Quarterly, 56: 622–641.

Bruckmuller, S., &Abele, E. A. 2013. The density of the bigtwo: How are agency and communion structurallyrepresented? Social Psychology, 44: 63–74.

Bryant, A., & Check, E. 2000. How parents raise boys andgirls. Newsweek: Your child, 136: 64–65.

Carli, L. L., LaFleur, S. J., & Loeber, C. C. 1995. Inter-personal relations and group processes: Nonverbalbehavior, gender, and influence. Journal of Person-ality and Social Psychology, 68: 1030–1041.

Chen, Z. X., & Aryee, S. 2007. Delegation and employeework outcomes: An examination of the cultural con-text of mediating processes in China. Academy ofManagement Journal, 50: 226–238.

Crites, S. L., Fabrigar, L. R., & Petty, R. E. 1994. Measuringthe affective and cognitive properties of attitudes:Conceptual and methodological issues. Personalityand Social Psychology Bulletin, 20: 619–634.

Cuddy, A. J. C., Fiske, S. T., & Glick, P. 2004. When pro-fessionals become mothers, warmth doesn’t cut theice. The Journal of Social Issues, 60: 701–718.

De Pater, I. E., Van Vianen, A. E., & Bechtoldt, M. N. 2010.Gender differences in job challenge: A matter of task al-location.Gender,WorkandOrganization, 17:433–453.

Diekman, A. B., & Eagly, A. H. 2000. Stereotypes as dy-namic constructs: Women and men of the past, pres-ent, and future. Personality and Social PsychologyBulletin, 26: 1171–1188.

Eagly, A. H. 1987. Sex differences in social behavior: Asocial-role interpretation. Hillsdale, NJ: LawrenceErlbaum Associates Inc.

Eagly, A. H. 2007. Female leadership advantage and dis-advantage: Resolving the contradictions. Psychologyof Women Quarterly, 31: 1–12.


Eagly, A. H., & Karau, S. J. 2002. Role congruity theory ofprejudice toward female leaders. Psychological Re-view, 109: 573–598.

Eysenck,M.W. 1992.Anxiety: The cognitive perspective.Hillsdale, NJ: Erlbaum.

Goff,K.,&Torrance,E.P.2002.AbbreviatedTorrancetest foradults. Bensenville, IL: Scholastic Testing Service, Inc.

Gosling, S. D., Rentfrow, P. J., & Swann,W. B. 2003. A verybrief measure of the big-five personality domains.Journal of Research in Personality, 37: 504–528.

Hayes,A. F. 2013. Introduction tomediation,moderation,and conditional process analysis: A regression-based approach. New York, NY: Guilford Press.

Heilman, M. E. 2001. Description and prescription: How gen-der stereotypes prevent women’s ascent up the organiza-tional ladder. The Journal of Social Issues, 57: 657–674.

Heilman,M. E., & Chen, J. J. 2005. Same behavior, differentconsequences: Reactions to men’s and women’s al-truistic citizenship behavior. The Journal of AppliedPsychology, 90: 431–441.

Heilman, M. E., & Okimoto, T. G. 2007. Why are womenpenalized for success at male tasks? The impliedcommunality deficit. The Journal of Applied Psy-chology, 92: 81–92.

Heilman, M. E., & Okimoto, T. G. 2008. Motherhood: Apotential source of bias in employment decisions.TheJournal of Applied Psychology, 93: 189–198.

Heilman,M. E.,Wallen, A. S., Fuchs, D., & Tamkins,M.M.2004. Penalties for success: Reactions to women whosucceed at male gender-typed tasks. The Journal ofApplied Psychology, 89: 416–427.

Izard, C. E. 1977. Human emotions. New York, NY:Plenum.

Kahneman, D., & Tversky, A. 1984. Choices, values, andframes. The American Psychologist, 39: 341–350.

Karson, M. 2007. Rorschach inkblot test. In N. Salkind(Ed.), Encyclopedia of measurement and statistics:849–850. Thousand Oaks, CA: SAGE Publications.

Kindlon, D., & Thompson, M. 2000. Raising Cain: Pro-tecting the emotional life of boys. New York, NY:Random House.

Kiviniemi, M. T., & Bevins, R. A. 2008. Role of affectiveassociations in the planning and habit systems ofdecision-making related to addiction.Behavioral andBrain Sciences, 31: 450–451.

Kiviniemi, M. T., Jandorf, L., & Erwin, D. O. 2014. Dis-gusted, embarrassed, annoyed: Affective associationsrelate to uptake of colonoscopy screening. Annals ofBehavioral Medicine, 48: 112–119.

Kiviniemi,M. T., Voss-Humke, A.M., & Seifert, A. L. 2007.How do I feel about the behavior? The interplay of

affective associations with behaviors and cognitivebeliefs as influences on physical activity behavior.Health Psychology, 26: 152–158.

Koenig, A. M., & Eagly, A. H. 2014. Evidence for the socialrole theory of stereotype content: Observations ofgroups’ roles shape stereotypes. Journal of Person-ality and Social Psychology, 107: 371–392.

Kray, L. J., Galinsky,A.D., &Thompson, L. 2002.Reversingthe gender gap in negotiations: An exploration of ste-reotype regeneration. Organizational Behavior andHuman Decision Processes, 87: 386–409.

Kugler, K., & Jones, W. H. 1992. On conceptualizing andassessing guilt. Journal of Social and Clinical Psy-chology, 62: 318–327.

Laverie, D. A. 1998. Motivations for ongoing participationin a fitness activity. Leisure Sciences, 20: 277–302.

Leana, C. R. 1986. Predictors and consequences of delega-tion.Academy of Management Journal, 29: 754–774.

Leana, C. R. 1987. Power relinquishment versus powersharing: Theoretical clarification and empirical com-parison of delegation. The Journal of Applied Psy-chology, 72: 228–233.

Liberman, V., Samuels, S.M., & Ross, L. 2004. The name ofthe game: Predictive power of reputations versus sit-uational labels in determining prisoner’s dilemmagame moves. Personality and Social PsychologyBulletin, 30: 1175–1185.

McConnell, A. R., & Leibold, J. M. 2001. Relations amongthe Implicit Association Test, discriminatory behav-ior, and explicit measures of racial attitudes. Journalof Experimental Social Psychology, 37: 435–442.

McFarland, C., &Miller, D. T. 1994. The framing of relativeperformance feedback: Seeing the glass as half emptyor half full. Journal of Personality and Social Psy-chology, 66: 1061–1073.

Miles, R., Snow, C., Meyer, A., & Coleman, H. 1978. Orga-nizational strategy, structure, and process. Academyof Management Review, 3: 546–562.

Moss-Racusin, C. A., Phelan, J. E., & Rudman, L. A. 2010.When men break the gender rules: Status incongruityand backlash againstmodestmen.Psychology ofMen& Masculinity, 11: 140–151.

Moss-Racusin, C., & Rudman, L. A. 2010. Disruptions inwomen’s self-promotion: The backlash avoidancemodel.PsychologyofWomenQuarterly, 34: 186–202.

Page-Gould,E.,Mendoza-Denton,R.,&Tropp,L.R.2008.Witha little help from my cross-group friend: Reducing in-tergroup anxiety through cross-group friendship. Journalof Personality and Social Psychology, 95: 1080–1094.

Pennebaker, J. W., Francis, M. E., & Booth, R. J. 2001. Lin-guistic inquiry and word count: LIWC 2001. Mah-wah, NJ: Erlbaum Publishers.


Plant, E. A., Devine, P. G., & Brazy, P. C. 2003. The boguspipelineandmotivations to respondwithoutprejudice:Revisiting the fading and faking of racial prejudice.Group Processes & Intergroup Relations, 6: 187–200.

Richard, R., van der Pligt, J., & de Vries, N. 1996. Antici-pated affect andbehavioral choice.BasicandAppliedSocial Psychology, 18: 111–129.

Richeson, J. A., & Shelton, J. N. 2003.When prejudice doesnot pay: Effects of interracial contact on executivefunction. Psychological Science, 14: 287–290.

Rooth,D. 2010.Automatic associations anddiscriminationin hiring: Real world evidence. Labour Economics,17: 523–534.

Rothman, A. J., & Salovey, P. 1997. Shaping perceptions tomotivate healthy behavior: The role of message fram-ing. Psychological Bulletin, 121: 3–19.

Rudman, L. A. 1998. Self-promotion as a risk factor forwomen: The costs andbenefits of counter-stereotypicalimpression management. Journal of Personality andSocial Psychology, 74: 629–645.

Rudman, L. A., & Glick, P. 2001. Prescriptive gender ste-reotypes and backlash toward agentic women. TheJournal of Social Issues, 57: 743–762.

Rudman, L. A., & Fairchild, K. 2004. Reactions to coun-terstereotypic behavior: The role of backlash in cul-tural stereotype maintenance. Journal of Personalityand Social Psychology, 87: 157–176.

Rudman, L. A., & Phelan, J. E. 2008. Backlash effects fordisconfirming gender stereotypes in organiza-tions. Research in Organizational Behavior, 28:61–79.

Simons, J., & Carey, K. B. 1998. A structural analysis ofattitudes toward alcohol and marijuana use. Person-ality and Social Psychology Bulletin, 24: 727–735.

Small, D. A., Gelfand, M., Babcock, L., & Gettman, H. 2007.Who goes to the bargaining table? The influence of gen-der and framing on the initiation of negotiation. Journalof Personality and Social Psychology, 93: 600–613.

Spencer, S. J., Steele, C. M., & Quinn, D. M. 1999. Stereo-type threat and women’s math performance. Journalof Experimental Social Psychology, 35: 4–28.

Stephan,W. G., & Stephan, C.W. 1985. Intergroup anxiety.The Journal of Social Issues, 41: 157–175.

Sy, T., Cote, S., & Saavedra, R. 2005. The contagiousleader: Impact of the leader’s mood on the mood ofgroup members, group affective tone, and groupprocesses. The Journal of Applied Psychology, 90:295–305.

Tabachnick, B. G., & Fidell, L. S. 2007.Usingmultivariatestatistics. Boston, MA: Pearson/Allyn & Bacon.

Tannen, D. 1990. You just don’t understand: Women andmen in conversation. New York, NY: Morrow.

Trafimow, D., & Sheeran, P. 1998. Some tests of the dis-tinction between cognitive and affective beliefs. Jour-nal of Experimental Social Psychology, 34: 378–397.

Valian, V. 1998. Why so slow? The advancement ofwomen. Cambridge, U.K.: MIT Press.

Vandello, J. A., Bosson, J. K., Cohen,D., Burnaford, R.M., &Weaver, J. R. 2008. Precarious manhood. Journal ofPersonality and Social Psychology, 95: 1325–1339.

Watson, D., & Clark, L. A. 1994. The PANAS-X: Manualfor the positive and negative affect schedule—Expanded Form. Iowa City, IA: University of Iowa.

Weinstein, D. D. 1993. Testing four competing theories ofhealth-protective behaviors. Health Psychology, 12:324–333.

Weisner, T. S., & Wilson-Mitchell, J. E. 1990. Non-conventional family life-styles and sex typing in six-year-olds. Child Development, 61: 1915–1933.

Yukl, G., & Fu, P. P. 1999. Determinants of delegation andconsultation by managers. Journal of OrganizationalBehavior, 20: 219–232.

Zou, X., Tam,K. P.,Morris,M.W., Lee, S. L., Lau, I. Y.M., &Chiu, C. Y. 2009. Culture as common sense: Perceivedconsensus versus personal beliefs as mechanisms ofcultural influence. Journal of Personality and SocialPsychology, 97: 579–597.

ModupeAkinola ([email protected]) is the SanfordC. Bernstein & Co. Associate Professor of Leadership andEthics at Columbia Business School. She received her PhDin organizational behavior from Harvard University. Herresearch explores how stress affects workplace perfor-mance. She also examines workforce diversity.

Ashley Martin ([email protected]) is an assis-tant professor at the Stanford Graduate School of Business.She received her PhD in management from ColumbiaBusiness School. Her research focuses on the challengesand benefits associated with gender and racial diversity inorganizations.

Katherine W. Phillips ([email protected]) is theReuben Mark Professor of Organizational Character at Co-lumbia Business School. She received her PhD in organiza-tional behavior from Stanford University. Her researchfocuses on the impact of diversity on performance, and re-lationshipbuildingacross racial boundaries inorganizations.


mailto:[email protected]



APPENDIX A

TABLE A1Study 4 Tasks

Task Name Description Scoring

1. Influencing Email Write an email to the school to influence students todonate to the organization. Six relevant points hadto be included in the body of the text.

This task was scored by two different raters (a 5 .75)on four dimensions: length, completion, writingquality, following instructions.

2. Organizing Donations(Pt. 1)

Copy a list of student names,majors, school years, anddonation amounts into a spreadsheet. Participantswere given a computer and Excel file in which toenter this information.

This task was scored based on completion (% of listcompleted) and accuracy (number of spellingerrors).

3. Organizing Donations(Pt. 2)

Participants were asked three questions about theinformation accumulated from the donation list: (1)what was the average amount donated? (2) whichmajor donated the most money? and (3) which yearwas most represented in new memberships?

This task required the participant to have completedthe first part of the task accurately and havecalculated these numbers correctly. This metricwas scored from 023 based on how many of thesequestions they answered correctly.

4. ProofreadingPowerPoint

Participants were told to proofread an eight-pagePowerPoint presentation. The presentation wasprinted on 8” 3 11” paper, with each slide ona separate page, and contained 38 errors in total.

This task was scored based on how many of the 38errors the participant found, with scores rangingfrom 0238.

5. Formatting Newsletter Participantswere given a printed copy of a newsletter,which contained a variety of fonts, colors, andsizing. Their goal was to format a document inMicrosoft Word, which included the unformattedrawmaterials, to look like the newsletter they weregiven.

Performance on this task was scored based on the 34changes that needed to be made to the document.Scores ranged from 0232.

6. ATTA (Creativity Task) This taskwas based on theAbbreviated Torrance Testfor Adults (ATTA; Goff & Torrance, 2002).Participants were given a sheet with nine trianglesand were told to create creative and novel logos forthe environmental campaign.

This task was scored by two research assistants, blindto the hypothesis, on (1) overall creativity, (2)fluency, (3) flexibility, and (4) flexibility. Anydiscrepancies were discussed and resolved by thetwo research assistants.


Copyright of Academy of Management Journal is the property of Academy of Managementand its content may not be copied or emailed to multiple sites or posted to a listserv withoutthe copyright holder's express written permission. However, users may print, download, oremail articles for individual use.

To Delegate or Not to Delegate: Gender Differences in ...

Documents

Transcript of To Delegate or Not to Delegate: Gender Differences in ...