GrantGinoHofmann Reversing

Embed Size (px)

Citation preview

  • 7/23/2019 GrantGinoHofmann Reversing

    1/23

    REVERSING THE EXTRAVERTED LEADERSHIP ADVANTAGE:THE ROLE OF EMPLOYEE PROACTIVITY

    ADAM M. GRANTUniversity of Pennsylvania

    FRANCESCA GINOHarvard University

    DAVID A. HOFMANNUniversity of North Carolina at Chapel Hill

    Extraversion predicts leadership emergence and effectiveness, but do groups performmore effectively under extraverted leadership? Drawing on dominance complementa-rity theory, we propose that although extraverted leadership enhances group perfor-mance when employees are passive, this effect reverses when employees are proactive,

    because extraverted leaders are less receptive to proactivity. In Study 1, pizza storeswith leaders rated high (low) in extraversion achieved higher profits when employees

    were passive (proactive). Study 2 constructively replicates these findings in the labo-ratory: passive (proactive) groups achieved higher performance when leaders actedhigh (low) in extraversion. We discuss theoretical and practical implications forleadership and proactivity.

    Scholars have spent more than a century seekingto understand the characteristics of effective lead-ers (Zaccaro, 2007). Research now suggests thatleading in an extraverted manner is a key to suc-cess. Extraversion is best understood as a tendencyto engage in behaviors that place oneself at thecenter of attention, such as seeking status and act-

    ing dominant, assertive, outgoing, and talkative(Ashton, Lee, & Paunonen, 2002). In a meta-analy-sis of the relationship between personality andleadership emergence and effectiveness, Judge,Bono, Ilies, and Gerhardt (2002: 765) found thatextraversion is the most consistent correlate ofleadership across study settings and leadership cri-teria. Their results indicated that extraverted em-

    ployees are significantly more likely to (1) emergeas leaders in selection and promotion decisionsand (2) be perceived as effective by both supervi-sors and subordinates. In another meta-analysis,Bono and Judge (2004) found that extraversion wasthe best personality predictor of transformationalleadership: leaders high in extraversion were

    more likely to express charisma, provide intellec-tual stimulation, and offer individualized consid-eration to their employees. Primary studies havefurther shown that extraversion is the only person-ality trait that predicts both typical and maximumratings of transformational leadership performance(Ployhart, Lim, & Chan, 2001), that the link be-tween extraversion and transformational leader-ship can be traced to genetically heritable sources(Johnson, Vernon, Harris, & Jang, 2004), and thatU.S. presidents are perceived as more effectivewhen they appear to be extraverted rather than

    introverted (Rubenzer & Faschingbauer, 2004;Young & French, 1996).This research suggests that in leadership roles,

    extraverts have a clear advantage. However, schol-ars have begun to question whether this conclusionoverstates the benefits of extraversion in leadershiproles and overlooks the costs (Judge, Piccolo, &Kosalka, 2009; McCormack & Mellor, 2002). In par-ticular, existing studies have focused on observersperceptions of leadership effectiveness (Judge etal., 2002), overlooking the objective performance ofthe groups and organizations that leaders guidea

    For helpful feedback on drafts, we are grateful to As-sociate Editor Elizabeth Morrison, three anonymous re-

    viewers, Jim Detert, Alison Fragale, Adam Galinsky, andDave Mayer, as well as participants in the New Direc-tions in Leadership Conference and the seminar series atthe University of Toronto, University of British Colum-

    bia, Rice University, and the Tuck School of Business.For insightful discussions, we thank Brian Little andSusan Cain. For assistance with data collection and cod-ing, we thank Andy Duvall, Jennifer Fink, Lisa JonesChristensen, Leslie Talbott, and Brett Yates. This articlewas accepted before Adam Grant became an associateeditor for this journal.

    Editors Note:The manuscript for this article was ac-cepted during Duane Irelands term as editor.

    Academy of Management Journal

    2011, Vol. 54, No. 3, 528550.

    528Copyright of the Academy of Management, all rights reserved. Contents may not be copied, emailed, posted to a listserv, or otherwise transmitted without the copyright holders express

    written permission. Users may print, download or email articles for individual use only.

  • 7/23/2019 GrantGinoHofmann Reversing

    2/23

    paramount indicator of leaders actual effectiveness(Kaiser, Hogan, & Craig, 2008). Accordingly, it ispossible that although extraversion is a consistentpredictor of supervisor and subordinate percep-tions of leadership effectiveness, extraverted lead-ership may not always contribute positively togroup performance. Research has shown that indi-

    viduals tend to hold implicit theories of leaders asextraverts, suggesting that supervisor and subordi-nate evaluations of extraverted leaders may be en-hanced by a halo effect. This may occur becauseextraverted leaders match the prototypes of charis-matic leaders that dominate both Western cultures(Lord, De Vader, & Alliger, 1986) and Eastern cul-tures (Leung & Bozionelos, 2004) and are especiallyprevalent in business (Agle, Nagarajan, Sonnen-feld, & Srinivasan, 2006). Illustrating this point, inan online survey of over 1,500 senior leaders earn-ing at least six-figure salaries, 65 percent viewedintroversion as a barrier to leadership, and only 6percent believed that introverts were more effectiveleaders than extraverts (Jones, 2006).

    To advance theory, research, and practice onleader characteristics, it is critical to understandhow they affect group performance (Kaiser et al.,2008). Our goal in this article is to examine theconditions under which extraverted leadershipcontributes to versus detracts from group perfor-mance. We propose that when employees are notproactive, extraverted leadership contributes tohigher group performance, but when employees areproactive, this relationship reverses to a negative

    one. There is good reason to believe that in a chang-ing business world, less extraverted leaders bringimportant strengths to the table. As organizationallife becomes more dynamic, uncertain, and unpre-dictable, it has become increasingly difficult forleaders to succeed by merely developing and pre-senting their visions top-down to employees (Grif-fin, Neal, & Parker, 2007). More than ever before,leaders depend on employees to proactively ad-vance bottom-up change by voicing constructiveideas (Van Dyne & LePine, 1998), taking charge toimprove work methods (Morrison & Phelps, 1999),

    and engaging in upward influence (Dutton, Ash-ford, ONeill, & Lawrence, 2001). However, re-search suggests that many leaders see these proac-tive behaviors as threats or distractions, and theythus fail to benefit from employees contributions(Detert & Burris, 2007; Grant, Parker, & Collins,2009; Parker, Williams, & Turner, 2006; Seibert,Kraimer, & Crant, 2001).

    We expect that leaders who are low rather thanhigh in extraversion are more receptive to bot-tom-up proactive behaviors from employees. As aresult, we hypothesize that when employees are

    proactive, extraverted leadership is negativelyrather than positively associated with group perfor-mance. We base these predictions on dominancecomplementarity theory (Carson, 1969; Kiesler,1983), according to which high-quality interactionsare facilitated when dominance and assertivenessfrom one party are balanced by compliance, obedi-

    ence, and submissiveness from the other party.From this perspective, employees proactive behav-iors have the potential to complement the quiet,more reserved behavior of leaders low in extraver-sion, but challenge the authority of leaders high inextraversion.

    In field and laboratory studies, we examinewhether employee proactivity reverses the effect ofextraverted leadership on group performance. Inour first study, we use field data to examinewhether a positive relationship between extra-verted leadership and the profits of pizza deliveryfranchises is reversed when employees are proac-tive. In our second study, we use a laboratory ex-periment to test whether groups led in a highlyextraverted manner perform better when employ-ees are not proactive, while groups led in a lessextraverted manner perform better when employ-ees are proactive. The combination of field andlaboratory data supports both the external and in-ternal validity of our hypotheses, and we findparallel results across naturally occurring and ex-perimentally manipulated variations in leaders ex-traversion and employees proactivity for predict-ing two objective performance outcomes: profits

    and output.Our research makes important theoretical contri-

    butions to the literatures on leadership and proac-tivity, offering a more balanced perspective thatrecognizes both the strengths and weaknesses ofextraverted leadership. First, we highlight howleadership research can gain traction by attendingto how follower behaviorsand leader reactions tothese behaviors create a context that alters theimpact of leader characteristics. Our studies docu-ment that leader extraversion can be either an assetor a liability for group performance, depending on

    the degree to which employees are proactive. Indoing so, we show how leadership research canbenefit from careful consideration of an expandedset of outcomes focusing on group performance, notonly on leader emergence and perceived effective-ness. Second, we provide a theoretical and an em-pirical account of how employees proactive behav-iors are not only caused by leader characteristics;they can also moderate the effects of leader charac-teristics. This perspective accentuates the value ofexamining the group performance consequences ofproactivity. Third, our findings reveal a potential

    2011 529Grant, Gino, and Hofmann

  • 7/23/2019 GrantGinoHofmann Reversing

    3/23

    irony of modern organizational life: although organ-izations often promote both extraverted leadershipand employee proactivity, these two characteristicsin combination may yield suboptimal group perfor-mance. Our research suggests that group perfor-mance is maximized when highly extraverted lead-ership is paired with less proactive employee

    behavior, or when less extraverted leadership ismatched with more proactive employee behavior.

    EXTRAVERTED LEADERSHIP ANDGROUP PERFORMANCE

    The purpose of this article is to examine contin-gencies that moderate the relationship between ex-traverted leadership and group performance. Bygroup performance, we refer to the effectiveness ofa unit in achieving collective goals (Campbell,1990). To specify the characteristics that defineextraverted leadership, it is important to gain adeeper understanding of personality. Psychologistshave developed two different approaches to con-ceptualizing personality (Hogan, 1991). The traitview of personality emphasizes the underlying psy-chological processes, or cognitive and affective ten-dencies, that cause stability in individual charac-teristics over time or across situations (House,Shane, & Herold, 1996; McCrae & Costa, 1990). The

    behavioral view of personality emphasizes theovert, observable actions in which individuals en-gage over time or across situations (Buss & Craik,1983). Recently, psychologists have theoretically

    and empirically integrated these approaches bydemonstrating that traits can be conceptualized asdensity distributions of behavioral acts (Fleeson,2001). Behavioral acts are the building blocks oftraits, and the stronger an individuals propensitytoward a trait, the more frequently and intenselythe individual enacts a corresponding set of behav-iors (Fleeson, 2001). This integration of traits and

    behavioral acts is consistent with longstanding ev-idence showing that the same content dimensionsand descriptors underlie both traits and behavioralacts and that the boundary between the two is fuzzy

    rather than categorically discrete (for a review, seeChaplin, John, and Goldberg [1988]).Given the value of both trait and behavioral ap-

    proaches to personality, our focus is on how leadercharacteristicswhich capture both leaders traitsand their behaviorsaffect group performance. Inparticular, we are interested in the leader charac-teristic of extraversion. Psychologists have longrecognized that extraversion is one of the funda-mental dimensions along which personality varies(e.g., Eysenck, 1973; Fleeson, 2001). Althoughscholars have debated about the defining features

    of extraversion, recent research suggests that thereal core of extraversion is the tendency to be-have in ways that attract social attention (Ashtonet al., 2002: 245). To do so, extraverted individualsoften seek out status and act assertive, interperson-ally dominant, talkative, and outgoing (Caspi, Rob-erts, & Shiner, 2005; DeYoung, Quilty, & Peterson,

    2007; Trapnell & Wiggins, 1990). A number of or-ganizational scholars have identified these quali-ties as critical to leadership (e.g., Bass, 1990; Kirk-patrick & Locke, 1991; Mann, 1959; Yukl, 1998),and some neurobiologists have even described theenjoyment of leadership roles as a prototypicalfeature of extraversion (Depue & Collins, 1999:492). As noted previously, extensive evidence sug-gests that extraverted individuals are more likely toemerge as leaders, receive high ratings of effective-ness from supervisors and subordinates, and dis-play transformational behaviors (for reviews, see

    Bono and Judge [2004], Judge et al. [2002], andOnes and Dilchert [2009]). An extraverted leader-ship style often involves engaging with followers,

    building networks, and influencing others with ef-fective inspiration and ingratiation strategies (e.g.,Cable & Judge, 2003; House & Howell, 1992).

    Our contribution lies in introducing employeeproactivity as a boundary condition for the extra-verted leadership advantage outlined above. Wepredict that when employees are proactive, theymay be more effective when their leaders are lessextraverted. Proactive behaviors are anticipatory

    actions that employees take to create change (Crant,2000; Grant & Ashford, 2008). Because leaders can-not always predict, envision, and control key inter-nal and external events, they rely on employees totake initiative to create constructive change (Griffinet al., 2007). Among the most widely studied pro-active behaviors are voice, taking charge, and up-ward influence (Grant et al., 2009; Parker & Collins,2010). Voice refers to speaking up with useful sug-gestions (Van Dyne & LePine, 1998). Taking chargerefers to exercising initiative to improve workstructures, processes, methods, and practices (Mor-rison & Phelps, 1999). Upward influence refers tomaking active attempts to persuade leaders to en-dorse and implement changes (Dutton et al., 2001;Kipnis & Schmidt, 1988). When employees engagein these behaviors, they exercise proactivity; whenthey do not engage in these behaviors, in keepingwith prior research, we describe their actions aspassive (Crant, 2000; Parker et al., 2006). We usethe term passive not to denote a lack of energy andeffort, but rather to capture reacting to others vi-sions and ideas instead of initiating ones own(Crant, 2000; Grant & Ashford, 2008).

    530 JuneAcademy of Management Journal

  • 7/23/2019 GrantGinoHofmann Reversing

    4/23

    To explain group performance, we conceptualizeand study employee proactivity at the group level.Although proactivity is often measured in terms ofindividual-level behaviors, empirical findings have

    begun to converge around the view that proactivitycan exist as a group-level phenomenon (Grant &Ashford, 2008). Researchers have demonstrated the

    existence of team-level proactivity (Kirkman &Rosen, 1999), firm-level proactivity (Aragon-Cor-rea, 1998; Henriques & Sadorsky, 1999), proactiveclimates (Fay, Luhrmann, & Kohl, 2004), andclimates for initiative (Baer & Frese, 2003). The-oretically, work groups are likely to develop normsfor high levels of proactivity under contextual con-ditions of environmental uncertainty, interdepen-dence, and autonomy. When environmental uncer-tainty is high, there is widespread recognition thatemployees need to be more proactive in order toanticipate and act upon threats and opportunities(Griffin et al., 2007). Furthermore, when interde-pendence is high, employees are more likely tocoordinate their levels of proactivity to facilitatetask completion (Griffin et al., 2007). In addition,research has shown that when employees are givenautonomy, they often work together to coordinateefforts to take charge, undertaking collaborative ac-tivities to improve work processes and methods(Leana, Appelbaum, & Shevchuk, 2010). Thus, de-pending on uncertainty, interdependence, and au-tonomy, different work groups can develop normsfor different levels of proactivity.

    A Dominance Complementarity Perspective

    Drawing on dominance complementarity theory,we propose that when employees engage in proac-tive behaviors, extraverted leadership is negativelyrather than positively associated with group perfor-mance. According to dominance complementaritytheory (Carson, 1969; Kiesler, 1983), effective inter-actions are achieved when dominant, assertive be-havior from one party is matched by submissive,passive behavior from another. A core tenet of dom-inance complementarity is that people seek balance

    in interpersonal interactions (Leary, 1957): whenone acts dominant, the other is expected to actsubmissive, and this pairing allows them to coor-dinate their actions and interactions effectively.Kiesler (1983: 198) proposed that complementarityexists when the individual contextually recog-nized as being superior is in the one-up or primaryposition, whereas the person recognized as beinginferior is in the one-down or secondary posi-tion. Dominance complementarity represents theexistence of a status hierarchy (Tiedens & Fragale,2003) in which both members are able to confirm

    . . . self-definitions (Kiesler, 1983: 198). The dom-inant parties gain validation of their power andstatus, and the submissive parties gain support andsecurity. For both parties, uncertainty about au-thority roles is reduced, preventing conflict andcompetition and facilitating the pursuit of commongoals (Bendersky & Hays, in press; Tiedens, Chow,

    & Unzueta, 2007; Wiltermuth, 2009). When com-plementarity is lacking, individuals often focus onjockeying for position, which can create anxietyand distract attention and energy away from taskcompletion (Smelser, 1961).

    As a result, matching dominance and submis-siveness can have important consequences forgroup performance. As Kristof-Brown, Barrick, andStevens (2005: 939) explained, There is extensiveevidence that complementarity in dominance (acomponent of extraversion) and submissiveness isrelated to higher quality interpersonal interac-tions. Studies have shown that individuals, dyads,and groups achieve greater productivity whendominance complementarity exists, wherein moreextraverted and dominant members work withmore introverted and submissive members (e.g.,Ghiselli & Lodahl, 1958; Hoffman & Maier, 1961;Kristof-Brown et al., 2005; Smelser, 1961). How-ever, these studies have focused primarily on peerrelationships, overlooking how dominance com-plementarity applies to hierarchical relationships

    between leaders and followers.We extend dominance complementarity theory

    to inform the interplay of extraverted leadership

    and employee proactivity. As noted previously,highly extraverted leadership typically involvesengaging in dominant, assertive behaviors andavoiding quieter, more reserved behaviors (see alsoMcCrae & Costa, 1989). Employees proactive be-haviors can be viewed as a form of dominance.Researchers have defined proactivity as an exerciseof control (Frese, Garst, & Fay, 2007; Parker et al.,2006), an expression of agency (Grant & Ashford,2008; Grant & Parker, 2009), and an effort to changeand challenge the status quo (Crant & Bateman,2000; Van Dyne, Cummings, & McLean Parks,

    1995). Voicing ideas, taking charge, and exertingupward influence are all displays of assertiveness.As a result, dominance complementarity theorysuggests that the more extraverted a leader is, theless employees will perceive him or her as recep-tive to their proactive ideas and suggestions. Per-ceptions of leader receptivity refer to the degree towhich employees view leaders as open to and will-ing to implement changes (Ashford, Rothbard, Pi-derit, & Dutton, 1998; Detert & Burris, 2007).

    Employees are likely to perceive more extra-verted leaders as less receptive to proactivity. Ac-

    2011 531Grant, Gino, and Hofmann

  • 7/23/2019 GrantGinoHofmann Reversing

    5/23

    cording to dominance complementarity theory,when one party acts dominant, he or she expectsthe other party to accept a lower-status positionand communicate in a more submissive manner(Kiesler, 1983). Building on this logic, extravertedleaders may respond to employee proactivity as athreat, seeking out dominance and communicating

    in assertive, forceful ways. In contrast, less extra-verted leaders may be receptive to employee pro-activity as a valuable source of input, communicat-ing in ways that signal openness and interest.Below, we discuss how extraverted leadership mayhave negative implications for receptivity to em-ployee proactivity using two conceptual lenses: sta-tus conflict and communication styles.

    Status conflict. Researchers have frequently ob-served that employees proactive behaviors can bethreatening to leaders, as they have the potential tointroduce unwelcome changes (Grant et al., 2009;Miceli & Near, 1995), make leaders feel embar-rassed, incompetent, or vulnerable by exposingtheir flaws and weaknesses (Morrison & Milliken,2000), and usurp leaders authority by venturingbeyond what management wants their employeesto do (Frese & Fay, 2001: 171). We predict thatextraverted leaders will be especially prone to ex-periencing employees proactive behaviors asthreatening. Extraverted leaders tend to seek outstatus (Anderson, John, Keltner, & Kring, 2001; Bar-rick, Stewart, & Piotrowski, 2002), which allowsthem to maintain a hierarchy in which their asser-tiveness and dominance are complemented by obe-

    dience and submissiveness on the part of employ-ees. When their status is threatened, extravertedleaders may be willing to engage in conflict and touse defensive and/or unconstructive tactics withothers (Ames & Flynn, 2007: 309). In general, ex-traverted leaders tend to exercise influence not byseeking ideas from others, but rather by gainingdominance over their employees (Peterson, Smith,Martorana, & Owens, 2003) and using direct per-suasion techniques to build others commitment totheir own ideas (Anderson, Spataro, & Flynn, 2008;Cable & Judge, 2003; Caldwell & Burger, 1997).

    Thus, extraverted leaders may reject employeesproactive ideas in favor of their own existingpractices.

    In keeping with these arguments, Judge et al.(2009: 868) noted that extraverted leaders tend tobehave in bold, aggressive, and grandiose ways.They like to be the center of attention. . . . As such,extraverted leaders may be less likely to solicitinput from subordinates and colleagues, poten-tially alienating organizational members who pre-fer that attention and credit be shared. We thuspredict that the combination of extraverted leader-

    ship and employee proactivity has the potential tocreate a power struggle, as both leaders and em-ployees seek to gain control and exercise influence.To do so, extraverted leaders may dismiss employ-ees proactive behaviors or work harder to increasetheir own control and influence, focusing on claim-ing status and asserting their authority instead of

    facilitating effective group performance. Indeed, re-search suggests that seeking dominance and poweris likely to discourage leaders from consideringemployees perspectives (Galinsky, Magee, Inesi, &Gruenfeld, 2006). Employees, in turn, are likely tofeel rejected and slighted by leaders lack of con-sideration for their ideas (McFarlin & Sweeney,1996), leading them to experience helplessness andpowerlessness (Magee & Galinsky, 2008).

    In contrast, less extraverted leaders tend not to beas concerned with status and power (Anderson etal., 2001; Barrick et al., 2002), and they may bemore receptive to employees proactive behaviors.When employees take charge to introduce newwork methods, less extraverted leaders may em-

    brace their ideas, using them to develop more effi-cient and effective systems (Moon, Kamdar, Mayer,& Takeuchi, 2008), correct errors in faulty proce-dures (Staw & Boettger, 1990), and identify newtechniques for preventing errors and problems inthe future (Wall, Jackson, & Davids, 1992). Further-more, less extraverted leaders may take particularnotice of employees efforts to voice importantproblems, which can enhance group performance(Detert & Burris, 2007; Van Dyne & LePine, 1998),

    and respond to upward influence by mobilizingattention and effort toward solving problems,which can enable them to make meaningfulchanges with respect to strategically significant is-sues (Dutton et al., 2001).

    Communication styles. In addition, less extra-verted leaders may listen more carefully to employ-ees proactive ideas and suggestions. Whereashighly extraverted individuals tend to be assertiveand confident in building visions and expressingideas, individuals lower in extraversion tend to bequieter and more reserved (e.g., Fleeson, 2001; Gos-

    ling, John, Craik, & Robins, 1998). Research indi-cates that on average, less extraverted individualsspeak more quietly than their highly extravertedcounterparts (Scherer, 1978), which suggests thatless extraverted leaders communication stylesleave more room for employees to be proactive.Furthermore, less extraverted individuals tend tospend more time listening and less time talking(Ramsay, 1966), and feel more apprehensive aboutinitiating communication in groups (Opt & Loffredo,2000) and speaking in public (Feingold, 1983). As aresult, less extraverted leaders may send more ver-

    532 JuneAcademy of Management Journal

  • 7/23/2019 GrantGinoHofmann Reversing

    6/23

    bal and nonverbal signals that they are open toproactivity and may take more time to hear andconsider ideas and suggestions.

    Of course, leaders are likely to derive the great-est benefits from these proactive behaviors whenthe ideas underlying them are relevant to organ-izational goals (Frese & Fay, 2001). However,

    even when employees proactive behaviors arenot relevant to organizational goals, there are tworeasons to believe that less extraverted leaderswill respond to them in ways that enhance groupperformance. First, because less extraverted lead-ers listen more carefully, employees will be morelikely to feel that their ideas are considered andappreciated, which will enhance and sustaintheir motivations to contribute (e.g., Dutton et al.,2001; McFarlin & Sweeney, 1996). Second, evenif employees proactive suggestions are mis-guided, they can still facilitate constructivechanges. Employees proactive behaviors can in-troduce a novel perspective, which may stimu-late divergent attention and thought. As a result,even when they are wrong they contribute to thedetection of novel solutions and decisions that,on balance, are qualitatively better (Nemeth,1986: 23). Accordingly, since less extravertedleaders may be more willing to listen to divergentopinions and perspectives, they may be morecapable of using misguided ideas constructively,reinforcing for employees that their ideas arevalued.

    Perceived leader receptivity and group perfor-

    mance.We thus propose that when employees areproactive, more extraverted leaders will respondless receptively to ideas and suggestions. In turn,perceiving a lack of leader receptiveness will dis-courage employees from working hard on behalf oftheir leaders. From the standpoint of organizationalsupport theory (Rhoades & Eisenberger, 2002),when employees feel that leaders value their con-tributions, they reciprocate by working harder.Feeling valued strengthens employees perceivedobligation to contribute, increases their identifica-tion with leaders, and enhances their confidence

    that their contributions will be worthwhile andrewarded (Rhoades & Eisenberger, 2002). Indeed,research has shown that when employees feel thattheir contributions are valued, they respond by in-creasing their effort (Grant, 2008; Grant & Gino,2010). When employees perceive that leaders arenot receptive to their contributions, they feel lessvalued and thus less motivated to reciprocate.

    Consequently, we predict that employees per-ceptions of leaders receptivity mediate the moder-ating effect of employee proactivity on the relation-ship between leaders extraversion and group

    performance. When employees are proactive,highly extraverted leaders respond less receptively,which decreases employees motivation to performeffectively. This sequence constitutes a first-stagemoderation model (Edwards & Lambert, 2007). Em-ployee proactivity moderates the effect of leaderextraversion on employees perceptions of leader

    receptivity in such a way that employees only per-ceive highly extraverted leaders as less receptiveunder conditions of high proactivity. Employeeproactivity threatens leaders ability to be the cen-ter of attention, reducing their receptiveness toideas and suggestions. Our predictions are summa-rized in the following hypotheses:

    Hypothesis 1. Employee proactivity moderatesthe association between leader extraversionand group performance. When employees arepassive, leader extraversion is positively re-

    lated to group performance, but when employ-ees are proactive, leader extraversion is nega-tively related to group performance.

    Hypothesis 2. Employee perceptions of recep-tivity mediate the moderating effect of em-ployee proactivity on the relationship betweenleader extraversion and group performance.

    Overview of the Present Research

    To test these hypotheses, we conducted twostudies. Guided by the trait and behavioral ap-proaches to personality, we operationalized extra-version in terms of leaders traits (Study 1) and

    behavioral acts (Study 2). In Study 1, we testedHypothesis 1 with multisource field data frompizza delivery stores. In Study 2, we conducted alaboratory experiment to constructively replicateour test of Hypothesis 1, support causal inferences,and test Hypothesis 2. Since our hypotheses rely onmotivational mechanismswhen extraverted lead-ers are not receptive to proactive suggestions, em-ployees are less motivated to put forth effortitwas important to use tasks in which motivationdrives differences in performance. In both studies,to observe motivational effects on group perfor-mance, we focused on relatively structured, simpletasks in which motivation is a central determinantof performance. By contrast, more complex tasksopen the door for ability to play a more central role(Locke & Latham, 2002; Schmidt & Hunter, 2004),potentially suppressing performance differencescaused by variations in leader extraversion and em-ployee proactivity.

    2011 533Grant, Gino, and Hofmann

  • 7/23/2019 GrantGinoHofmann Reversing

    7/23

    STUDY 1: METHODS

    Sample and Procedures

    We obtained data from a U.S. national pizza de-livery company that runs franchises. The leaders inthe company are the overall store managers (1 perstore), who independently oversee all store opera-

    tions, including inventory control, personnel hir-ing and scheduling, and management of supervi-sors. We sent questionnaires for leaders andemployees to 130 stores and obtained completedata from 57, achieving a response rate of 43.1percent. The respondents were 57 store leaders and374 employees (mean 6.56 employees per store,minimum 2, maximum 15). Because of confi-dentiality concerns, we collected demographic andemployment data using five-point scales. With re-spect to employment with the organization, 1.8 per-cent of respondents had been with the company forless than one year; 5.4 percent had been employedfor one to three years; 37.5 percent had been em-ployed for three to five years; 44.6 percent had beenemployed for five to ten years; and 10.7 percent had

    been employed for at least ten years. With respectto employment in their current store, 19.3 percenthad been employed less than one year; 33.3 percenthad been employed for one to three years; 28.1percent had been employed for three to five years;17.5 percent had been employed for five to tenyears; and 1.8 percent had been employed for atleast ten years.

    Measures

    Unless otherwise indicated, all items used a Lik-ert-type scale anchored at 1, to a very small ex-tent, and 5, to a great extent.

    Group performance.We measured group perfor-mance in terms of each stores overall profitabilityduring the seven weeks following survey comple-tion, after controlling for the average price of pizzaorders and worker hours. This lagged seven-weekperiod allowed us to calculate the reliability ofweekly profits, which demonstrated excellent in-

    ternal consistency (

    .98). Store profits are theultimate metric on which the company evaluatesleaders performance, and it is the basis on whichleaders are rewarded. It was important to controlfor the average price of orders and worker hours

    because these variables heavily influence profit-ability but are largely determined by a stores loca-tion (e.g., college campus vs. urban street vs. ruraltown), which are factors beyond the control of lead-ers and employees. Controlling for these location-

    based input factors allowed us to examine howfeatures of internal operations that leaders and em-

    ployees could influence (i.e., delivery costs, foodcosts, labor costs, planning, scheduling) were af-fected by leaders personality traits and employeesproactive behaviors.

    Leaders personality traits. Store leaders ratedtheir own personality traits using ten items for eachof the Big Five traits from Goldbergs (1992) ad-

    jective scale. The extraversion scale included ad-jectives such as assertive, talkative, bold,introverted (reverse-scored), reserved (reverse-scored), and energetic ( .73). Although ourhypotheses focused on extraversion, since this traitis correlated with other Big Five traits (e.g., Olson,2005), to reduce the likelihood of spurious associ-ations, we controlled for the other four traits: emo-tional stability, agreeableness, openness, and con-scientiousness. In particular, extraversion typicallycorrelates positively with openness and emotionalstability and also frequently correlates positivelywith agreeableness and conscientiousness (e.g.,DeYoung et al., 2007; John & Srivastava, 1999). Toisolate the unique variance attributable to leaderextraversion, it is important to control for thesetraits and their interactions with employee proac-tivity. The emotional stability scale included adjec-tives such as relaxed, touchy (reverse-scored),moody (reverse-scored), emotional (reverse-scored), high-strung (reverse-scored), and calm( .66). The agreeableness scale included adjec-tives such as agreeable, cooperative, cold (re-verse-scored), helpful, rude (reverse-scored), andconsiderate ( .75). The openness scale included

    adjectives such as imaginative, creative, intel-lectual, innovative, shallow (reverse-scored),and complex ( .72). The conscientiousnessscale included adjectives such as organized, de-pendable, sloppy (reverse-scored), conscien-tious, careless (reverse-scored), and systematic( .76).

    Group proactivity. Employees rated the averagelevels of proactive behaviors occurring in theirstore. Across the stores, 374 employees providedratings of store proactivity. Specifically, each em-ployee was asked to rate the typical behavior of

    employees in your store (excluding the store man-ager). With respect to level of measurement, theseitems represent a referent shift compositionalmodel (Chan, 1998). We provide additional evi-dence justifying the aggregation of these scales tothe store level below.

    We measured taking charge using the five high-est-loading items from the taking charge scale de-veloped by Morrison and Phelps (1999), includingTry to bring about improved procedures for thework unit, Try to correct a faulty procedure orpractice, and Try to implement solutions to

    534 JuneAcademy of Management Journal

  • 7/23/2019 GrantGinoHofmann Reversing

    8/23

    pressing organizational problems (individual-level .86, store-level .87). We measuredvoiceusing the four highest-loading items from thevoice scale developed by Van Dyne and LePine(1998), which includes items such as Speak upwith ideas for new projects or changes in proce-dures, Communicate opinions about work issuesto others even if their opinions differ or othersdisagree, and Develop and make recommenda-tions concerning issues that affect this store (indi-vidual-level .88, store-level .89). Wemeasuredupward influenceusing the four highest-loading items adapted from the upward communi-cation scale developed by Hofmann and Morgeson(1999), including Discuss production issues withthe store leader and Discuss work issues with thestore leader (individual-level .83, store-level .80).

    To examine whether it was appropriate to aggre-gate the three proactive behaviors into a higher-order proactivity construct, we conducted confir-matory factor analyses using EQS software version6.1 with maximum-likelihood estimation proce-dures (e.g., Kline, 1998). We found that a three-factor model with a latent higher-order proactivityfactor achieved acceptable fit with the data andsignificantly better fit than all alternative nestedmodels. Table 1 presents these results. All factorloadings were statistically significant and rangedfrom .70 to .80 for taking charge, .58 to .82 for voice,

    and .71 to .91 for upward influence. The factorloadings for the three proactivity constructs on thehigher-order proactivity factor were .83 for takingcharge, .88 for voice, and .69 for upward influence.These results suggest that the three proactivity con-structs could be analyzed as a single higher-orderproactivity construct (individual-level .91,store-level .97).

    In light of the referent shift compositional model,we also investigated the appropriateness of aggre-gating the measure of proactivity to the store level.The medianrwgvalue was .93. This indicates strong

    within-group agreement, justifying the view thatstore proactivity existed and could be aggregated tothe group level (LeBreton & Senter, 2008). A one-way analysis of variance (ANOVA) indicated thatproactivity varied significantly across stores (F[56,317] 1.43,p .05). Interrater reliability betweenemployees was moderate (ICC1 .06, ICC2 .30,

    bothp .05). In conjunction, these results indicatehigh agreement within stores, but relatively lowvariance between storesperhaps because thestores operate in the same industry and do notdiffer dramatically in the contextual levels of un-certainty, interdependence, and autonomy that cre-ate variations in group proactivity. The limited de-gree of between-store variance served to attenuateour results at the store level and resulted in a con-servative investigation of our hypotheses.

    STUDY 1: RESULTS AND DISCUSSION

    Means, standard deviations, and correlations forour key variables appear in Table 2. We tested ourhypotheses using the moderated regression proce-dures recommended by Aiken and West (1991). Westandardized the leader personality and employeeproactivity variables and then multiplied them tocreate interaction terms. In our regression analyses,we controlled for the average price of orders andworker hours, as well as for the other four Big Fivepersonality traits and their interactions with each

    proactive behavior. Table 3 displays the results ofour regression analyses.1 As specified in Hypothe-sis 1, there was a significant interaction between

    1 We also conducted the analyses separately for takingcharge, voice, and upward communication and found thesame patterns, with one exception: a significant interac-tion between leader emotional stability and employeeproactivity in predicting taking charge. Simple slopesindicated that leader emotional stability predicted highergroup performance when employees did not take charge,

    but lower group performance when they did. Since this

    TABLE 1Study 1: Results of Confirmatory Factor Analysis of Proactivity Itemsa

    Model 2 df CFI SRMR 2

    1. Three-factor model with latent higher-order proactivity factor 232.49 62 .94 .042. Two-factor model a: Voice and taking charge on same factor 417.54 64 .87 .06 2(2) 185.05***3. Two-factor model b: Taking charge and upward influence on same factor 693.80 64 .77 .09 2(2) 461.31***4. Two-factor model c: Voice and upward influence on same factor 606.19 64 .80 .09 2(2) 373.70***5. Single-factor model 857.13 65 .71 .10 2(3) 624.64***

    a Values in bold indicate the best-fitting model.*p .05

    **p .01***p .001

    2011 535Grant, Gino, and Hofmann

  • 7/23/2019 GrantGinoHofmann Reversing

    9/23

    leaders extraversion and employee proactivity inpredicting store profits. To interpret the form of theinteraction, we plotted the simple slopes for therelationship between leaders extraversion andstore profits at one standard deviation above and

    below the mean of employee proactivity (see Figure

    1). When employees engaged in low proactivity,extraverted leadership was associated with higherstore profits ( .25, p .02). When employeesengaged in high proactivity, the simple slopes in-dicated that extraverted leadership was associatedwith lower store profits ( .23,p .05). Thus,

    extraverted leadership predicted higher store per-formance when employees were passive, but lowerstore performance when employees were proactive.

    These results show that employee proactivity mod-erates the relationship between leaders extraversionand group performance, and the form of this moder-

    ating effect is consistent with the dominance comple-mentarity perspective. Although these results arepromising, they suffer from several limitations, in-cluding a small sample size and modest responserate, moderate between-store variance in proactivity,and an unusual coefficient for conscientiousness. Assuch, it was important to examine whether the resultscould be constructively replicated (Lykken, 1968) in adifferent sample with different operationalizations ofleadership, employee proactivity, and group perfor-mance. In particular, we measured leaders extraver-sion as a trait, but it was important to test whether ourfindings held when extraversion was operationalizedas a set of behaviors. This investigation appearedespecially worthwhile in light of evidence that lead-ers traits have their impact through their expressionin overt behaviors that mediate the relationship be-tween traits and performance (e.g., Anderson & Sch-neier, 1978). In addition, our observational data didnot support causal inferences, and we were not ableto measure the proposed mediating mechanism ofperceived leader receptivity. Furthermore, employeesmight communicate proactivity differently to moreversus less extraverted leaders.

    was not a hypothesized interaction and it did not emergefor the other two proactive behaviors, it is important to becautious in interpreting it, as it may be a methodologicalartifact. However, if it can be replicated, one explanationis that leaders low in emotional stability experience lev-

    els of anxiety that can be productive or counterproduc-tive, depending how they manage their anxiety (e.g.,Norem & Cantor, 1986; Tamir, 2005). When employeestake charge, leaders low in emotional stability may beable to channel their anxiety in the productive directionof supporting, honing, and scaling up the developmentand implementation of new work processes. Voice andupward influence, because they involve verbal commu-nications but not necessarily action taken to develop andimplement a concrete change in work processes, maymaintain uncertainty and thus fail to give leaders a clearfocus for channeling their anxiety into constructiveimprovements.

    TABLE 2Study 1: Descriptive Statistics and Correlationsa

    Variable Mean s.d. 1 2 3 4 5 6 7 8 9 10 11 12

    1. Groupperformance(profits)

    $6,206.78 1,766.65 (.98)

    2. Average price of

    orders

    $17.03 1.70 .10

    3. Worker hours 465.58 86.82 .83*** .224. Leader

    extraversion3.76 0.57 .05 .29* .08 (.73)

    5. Leader emotionalstability

    3.18 0.59 .01 .17 .01 .29* ( .66)

    6. Leaderagreeableness

    4.04 0.52 .15 .02 .05 .10 .34** (.75)

    7. Leader openness 3.76 0.54 .17 .18 .15 .36** .08 .16 (.72)8. Leader

    conscientiousness4.05 0.52 .26 .04 .03 .35** .08 .33** .43** ( .76)

    9. Store takingcharge

    3.28 0.41 .06 .12 .06 .10 .10 .05 .05 .19 (.86/.87)

    10. Store voice 3.31 0.42 .01 .16 .04 .02 .15 .15 .16 .19 .60** (.88/.89)11. Store upward

    influence3.75 0.52 .03 .06 .07 .02 .03 .17 .08 .16 .48*** .55*** (.83/.80)

    12. Store proactivity 3.45 0.37 .01 .07 .07 .03 .03 .11 .00 .22 .81*** .85*** .84*** (.91/.97)

    a n 57 (57 stores for variables 13, 57 store leaders for variables 48, and 374 employees for variables 912). Variables 311 aremeasured on a five-point Likert-type scale. Cronbachs alphas appear in parentheses on the diagonal. For the group proactivitymeasures (912), the first coefficient is the individual-level internal consistency, and the second is the store-level internalconsistency.

    536 JuneAcademy of Management Journal

  • 7/23/2019 GrantGinoHofmann Reversing

    10/23

    To address these issues, our second study used anexperimental method to demonstrate that the interac-tions of extraverted leadership behaviors with em-ployee proactivity caused variations in group perfor-mance. This experimental design also allowed us totest whether more extraverted leadership behaviorscaused higher group performance when employeeswere passive, while less extraverted leadership be-haviors caused higher group performance when em-ployees were proactive. It also enabled us to holdconstant the nature and form of employee proactivityto rule out the possibility that employees expresstheir proactive behaviors as a function of leader char-acteristics. Furthermore, we tested Hypothesis 2 byexamining the role of perceptions of leader receptiv-ity in explaining the observed effects.

    STUDY 2: METHODS

    Sample and Procedures

    One hundred sixty-three college students froma university in the southeastern United States

    (mean age 20.90, s.d. 1.24; 101 male, 62 fe-male) participated in the study in exchange forcredit in an introductory organizational behaviorclass. After explaining that we were interested inunderstanding the factors that influence group per-formance, we told participants that they would beleading a group of four members to fold as manyT-shirts as possible in ten minutes. There were 56groups, each of which contained three focal partic-ipants: a leader and two followers. In all groups,two additional undergraduate research assistants(both male) were present as confederates, posing asadditional followers. These two confederates wereinstructed to fold approximately the same numberof T-shirts in every session regardless of experi-mental condition, leaving variance in group perfor-mance attributable to the leader and the two otherfollowers.

    In each session, one participant selected a cardfrom a hat identifying him/her as the leader. In fact,all cards read leader, so that the first person whochose a card became the leader. The two other

    TABLE 3Study 1: Moderated Regression Analyses Predicting Group Performancea

    Variable

    Step 1 Step 2 Step 3

    b s.e. t b s.e. t b s.e. t

    Average price of orders 88.72 79.64 .09 1.11 51.51 86.04 .05 0.60 47.21 8.82 .05 0.58

    Worker hours 17.31 1.56 .85 11.10*** 16.94 1.61 .83 1.52*** 17.46 1.53 .86 11.45***Store employeeproactivity

    3.53 136.37 .00 0.03 135.83 137.63 .08 0.99

    Leader extraversion 89.92 168.27 .05 0.53 13.18 16.71 .01 0.08Leader emotional

    stability78.13 159.96 .04 0.49 42.81 157.13 .02 0.27

    Leader agreeableness 133.17 159.27 .08 0.84 3.90 157.48 .00 0.03Leader openness 11.15 161.94 .06 0.68 99.51 152.14 .06 0.65Leader

    conscientiousness421.07 169.96 .24 2.48* 401.93 159.78 .23 2.52*

    Emotional stability proactivity

    127.97 161.19 .07 0.79 231.03 156.39 .13 1.48

    Agreeableness proactivity

    15.93 185.40 .01 0.09 31.54 174.22 .02 0.18

    Openness

    proactivity

    117.89 169.25 .06 0.70 33.40 162.18 .02 0.21

    Conscientiousness proactivity

    142.50 19.81 .07 0.75 277.64 186.45 .13 1.49

    Extraversion proactivity

    423.89 161.52 .24 2.62*

    R2: F(2, 54): R2: F(12, 44): R2: F(10, 44): R2: F(13, 43): R2: F(1, 43):.69 62.59*** .70 12.11*** .01 1.31 .74 13.21*** .04 6.89*

    a n 57. Future research is necessary to explain why leaders conscientiousness predicted lower store profits. Since conscientiousnessscores were positively skewed, the pattern may be a partial function of restricted range, whereby the majority of leaders are above thethreshold necessary for effective leadership. Moderately high leader conscientiousness may be optimal: extremely high scores may signala level of detail orientation that distracts attention away from bigger-picture issues, a tendency to micromanage employees, or excessivecautiousness, risk aversion, and resistance to change (Judge et al., 2009; Le, Oh, Robbins, Ilies, Holland, & Westrick, 2011).

    *p .05**p .01

    ***p .001

    2011 537Grant, Gino, and Hofmann

  • 7/23/2019 GrantGinoHofmann Reversing

    11/23

    participants and the two confederates assumed therole of followers. To encourage participants to careabout the outcome, we established instrumentalityand valence by informing participants that themembers of the groups in the top 10 percent ofproductivity would win iPod Nanos. We manipu-lated two factors between subjects: leaders extra-version (high vs. low) and followers proactive be-havior (high vs. low).

    Manipulation 1: Leaders extraversion. Weadapted procedures developed by Fleeson, Mala-

    nos, and Achille (2002) to temporarily manipulateparticipants enactment of extraverted leadership

    behaviors. We asked participants to read a sum-mary about why it is important for leaders to act ina manner indicating either high or low extraversionand then to write a short description of times whenthey had done so. In the high-extraversion condi-tion, participants read about evidence for the im-portance of extraversion in leadership: Scientificresearch now shows that behaving in an extra-verted manner is the key to success as a leader. Like

    John F. Kennedy, Martin Luther King, Jr., and Jack

    Welch, great leaders are extraverted: their behavioris bold, talkative, and assertive. This enables themto communicate a strong, dominant vision that in-spires followers to deliver results. The descriptionthen summarized the results of a meta-analysisshowing that extraversion is the most importanttrait of leaders and effective leadership (Judge etal., 2002: 773) and provided highlights from severalprimary studies illustrating this pattern. Then, tomake extraverted behavior more palatable, in-formed by research on self-persuasion (Heslin,Latham, & VandeWalle, 2005), we asked partici-

    pants to reflect on a time when they had engaged inextraverted leadership: Now, to get ready for yourrole, think of a time when you led a group effec-tively by acting bold, talkative, outgoing, and asser-tive. Write a paragraph about what you said anddid, and why these behaviors helped to ensure thesuccess of your group.

    In the low-extraversion condition, participantsread about evidence for the importance of a lessextraverted leadership style: Scientific researchnow shows that behaving in an introverted manneris the key to success as a leader. Like MahatmaGandhi, Abraham Lincoln, and Socrates, great lead-ers are introverted: their behavior is quiet, shy, andreserved. This enables them to empower their peo-ple to deliver results. The description then sum-marized studies showing the potential costs ofhighly extraverted leadership behavior and the po-tential benefits of acting less extraverted (e.g., Judgeet al., 2009; McCormack & Mellor, 2002). Partici-pants then reflected on a time when they had en-gaged in less extraverted leadership: Now, to getready for your role, think of a time when you led agroup effectively by acting quiet, shy, and reserved.Write a paragraph about what you said and did, andwhy these behaviors helped to ensure the successof your group.

    After they had completed the extraversion ma-nipulation, the experimenter (a research assistant)asked leaders to explain the task to the two follow-ers. The groups then started the task of foldingT-shirts, timed by the experimenter. During thefolding task, we introduced the manipulation offollowers proactive versus passive behavior.

    FIGURE 1Study 1: Simple Slopes for Store Profits

    $5,000

    $5,500

    $6,000

    $6,500

    $7,000

    Low Leader Extraversion High Leader Extraversion

    Low employeeproactivity

    High employeeproactivity

    Leadership

    538 JuneAcademy of Management Journal

  • 7/23/2019 GrantGinoHofmann Reversing

    12/23

    Manipulation 2: Followers proactive behavior.During the folding task, we varied the degree towhich the two confederate followers behaved pro-actively versus passively. In the passive condition,the two confederate followers simply acted accord-ing to the leaders instructions throughout the fold-ing task. In the proactive condition, after 1 minute

    and 30 seconds, one of the confederates remarked,I wonder if theres a more efficient way to do this.The other confederate then said to the leader, Ac-tually, I have a friend from Japan who has a fasterway. It might take a minute or two to teach it, butdo we want to try it? We chose this proactive

    behavior because it involves elements of voice(speaking up with an idea), taking charge (introduc-ing a new work method and process), and upwardinfluence (attempting to change the leaders strat-egy). We also selected it as a prototypical exampleof a proactive behavior that had the potential to

    create improvements but required an investment oflearning time up front.

    If the leader said yes, the second confederatetaught the new method to the group. If the leadersaid no, then the confederates continued with thetraditional, slower method. The new method con-sisted of folding T-shirts as the task is often per-formed in Japan. The method involves laying theT-shirt flat, with the left sleeve pointing at ones

    body, and following these steps: (1) The right handpinches the top next to the collar, (2) the left handpinches in the middle of the shirt parallel to the

    first pinch, (3) the right hand folds the top to thebottom and (4) pinches the two together, (5)the right hand flips the T-shirt over and uncrossesthe arms, and (6) the T-shirt is set down with theright sleeve folded under (see www.youtube.com/watch?vAn0mFZ3enhM&featurerelated). Priorto the study, the confederate making the suggestionwas trained to fold T-shirts according to thismethod, so that he could teach it to the rest of thegroup.

    At the end of the session, the experimentercounted the number of T-shirts that the leader andtwo followers folded in the ten-minute period. Theexperimenter also counted the number of shirts theconfederates folded so that their performance could

    be eliminated from the analyses. Confederate per-formance did not vary significantly by condition.Upon completion of the task, leaders completed asurvey that assessed the extent to which they be-haved in an extraverted manner during the task andthe degree to which their followers were proactive.The followers also completed a survey asking themto assess leader extraversion, follower proactivity,and leader receptivity.

    Measures

    Dependent variable: Group performance. Wemeasured group performance by counting the num-

    ber of T-shirts each group had folded in the allo-cated period. Group performance is the sum of theleader and follower output; it does not include thenumber of T-shirts folded by the confederates.

    Mediator: Leader receptivity. To measure per-ceived leader receptivity, we adapted items fromexisting measures of leader openness (Ashford etal., 1998; Detert & Burris, 2007). We asked the fol-lowers to evaluate the leaders on five items (1 disagree strongly, 7 agree strongly): open tonew ideas, receptive to suggestions, interestedin our ideas, rejected new ideas (reverse-scored), and dismissed suggestions (reverse-scored) ( .94 for follower 1, and .89 forfollower 2). Since the two followers ratings dem-

    onstrated good interrater reliability (ICC1 .70,ICC2 .83, p .001), we averaged them to com-pute an overall score for perceived leader receptiv-ity. We used this aggregate measure in the analysespresented below.

    Manipulation check 1: Leaders extraverted

    behavior. Leaders indicated the extent to whichthey displayed behaviors characteristic of extra-verts during the task (1 extremely inaccurate,9 extremely accurate). We used Goldbergs(1992) 20-item adjective scale, which consisted of10 positively worded items, including assertive,

    talkative, and extraverted, and 10 reverse-scored items, such as introverted, quiet, andshy ( .97). Followers also completed the same20-item scale to rate their leaders extraversion (.97 for both followers). Because the two followersachieved good interrater reliability (ICC1 .61,ICC2 .76, p .001), we averaged their ratings.

    Manipulation check 2: Followers proactive be-haviors. Leaders rated their followers proactivebehaviors by indicating the extent to which follow-ers as a group displayed such behaviors (1 dis-agree strongly, 7 agree strongly). We used aten-item scale including items such as Came upwith ideas to improve the way in which the taskwas done and Put forward ideas to improve per-formance ( .98). This scale was adapted fromitems developed by Griffin et al. (2007), Morrisonand Phelps (1999), and Van Dyne and LePine(1998). The two followers also rated their teamsproactivity using the same scale ( .96 for bothfollowers). Since interrater reliability was good(ICC1 .81, ICC2 .89,p .001), we averaged thetwo followers ratings.

    2011 539Grant, Gino, and Hofmann

  • 7/23/2019 GrantGinoHofmann Reversing

    13/23

    STUDY 2: RESULTS AND DISCUSSION

    Means and standard deviations by condition forour focal variables appear in Table 4.

    Manipulation Checks

    We started by examining whether our leadershipmanipulation was effective using two (leader extra-version: high vs. low) times two (followers behav-ior: proactive vs. passive) between-subjects ANOVAs.Followers rated leaders as more extraverted in thehigh-extraversion condition (mean 5.42, s.d. 1.49) than in the low-extraversion condition(mean 4.19, s.d. 1.15;F[1, 51] 10.65,p .01,2 .16). Leaders ratings of their own extraversionduring the task were consistent with the followersratings; leaders in the high-extraversion conditionreported acting in a more extraverted manner(mean 5.36, s.d. 1.80) than did those in thelow-extraversion condition (mean 4.21, s.d. 1.41;F[1, 52] 6.98,p .02,2 .12). Neither ourproactivity manipulation nor the interaction be-tween the two manipulations significantly affectedthese ratings.

    We also used both leaders and followers ratingsof followers proactive behaviors to test the validityof our manipulation of followers proactive behav-ior. As expected, leaders rated the behavior of fol-lowers in the proactive condition as more proactive(mean 5.41, s.d. 1.03) than the behavior offollowers in the passive condition (mean 3.04,s.d. 1.38; F[1, 52] 52.40, p .001, 2 .50).Furthermore, followers in the proactive conditionrated group proactivity as significantly higher(mean 4.85, s.d. 0.43) than did the followers inthe passive condition (mean 2.78, s.d. 0.70;F[1, 51] 174.10, 2 .76). Neither the extravertedleadership manipulation nor the interaction be-tween the two manipulations had significant ef-fects. Taken together, these results indicate that ourmanipulations were effective.

    Performance Effects

    We tested our hypotheses by conducting a 2(leader extraversion) 2 (followers proactive be-havior) between-subjects ANOVA using the num-

    ber of T-shirts a group folded as the dependentvariable. We controlled for group size because fivegroups had four rather than five members (only one

    follower instead of two). The two main effects werenot statistically significant. In keeping with Hy-pothesis 1, we found a significant interaction be-tween leaders extraversion and followers proac-tive behavior (F[1, 51] 7.17, p .01, 2 .12).Figure 2 is a graph of this interaction. Within eachlevel of the proactivity manipulation, we used sim-ple effects to examine whether groups led in ahighly (less) extraverted manner performed betterwhen the confederates were passive (proactive).When the confederates were passive, groups in thehigh-leader-extraversion condition (mean

    167.93, s.d. 24.31) outperformed those in thelow-leader-extraversion condition (mean 137.64,s.d. 48.30; F[1, 52] 5.70,p .03). The reverseoccurred when the confederates were proactive:groups with leaders in the low-extraversion condi-tion (mean 175.43, s.d. 30.22) outperformedthose whose leaders were in the high-extraversioncondition (mean 142.00, s.d. 25.83; F[1, 52] 6.95,p .02).

    To ascertain whether these effects were driven byleader or follower performance, we conducted ad-ditional 22 between-subjects ANOVAs on the

    number of T-shirts folded by each leader and twofollowers. With follower performance as the depen-dent variable, the analyses showed a significantinteraction between the leader extraversion and fol-lower proactivity manipulations (F[1, 51] 6.38,p.02, 2 .11), and no other effects were signifi-cant. When the confederates were proactive, fol-lowers with leaders in the low-extraversion condi-tion (mean 59.64, s.d. 10.80) outperformedthose with leaders in the high-extraversion condi-tion (mean 46.61, s.d. 10.62; F[1, 52] 6.63,p .01). When the confederates were passive, the

    TABLE 4Study 2: Means and Standard Deviations by Condition for Key Variablesa

    ConditionGroup

    PerformanceIndividual Follower

    PerformanceIndividual Leader

    PerformanceLeader

    Receptivity

    Low extraversion, passive followers 137.64 (48.30) 50.32 (19.86) 52.07 (15.27) 4.50 (1.13)High extraversion, passive followers 167.93 (24.31) 55.21 (9.70) 57.50 (13.07) 5.34 (1.01)Low extraversion, proactive followers 175.43 (30.22) 59.64 (10.80) 56.14 (11.13) 5.35 (0.87)High extraversion, proactive followers 142.00 (25.83) 46.61 (10.62) 51.57 (10.46) 4.32 (0.98)

    a Standard deviations are in parentheses.n 14.

    540 JuneAcademy of Management Journal

  • 7/23/2019 GrantGinoHofmann Reversing

    14/23

    performance of followers with leaders in the high-extraversion condition was higher (mean 55.21,s.d. 9.70) than the performance of followers withleaders in the low-extraversion condition (mean 50.32, s.d. 19.86), but unlike the results for thefull groups performance, the difference did notachieve statistical significance (F[1, 52] 0.93,n.s). In contrast, the leader extraversion and fol-lower proactivity manipulations had no significantmain or interactive effects on leader performance.These analyses demonstrate that the differences ingroup performance caused by the interaction of

    leader extraversion and follower proactivity weredue to followers performance, not leaders perfor-mance. Thus, supporting our hypotheses, when fol-lowers were proactive, they achieved higher perfor-mance when leaders acted in a less extravertedmanner.

    Mediation Analyses

    To assess whether followers perceptions of re-ceptivity explained the interactive effects on groupperformance, we conducted moderated mediation

    analyses using the procedures recommended byPreacher, Rucker, and Hayes (2007) and Edwardsand Lambert (2007). A 22 ANOVA showed thatthe interaction of leader extraversion and followerproactivity had a significant effect on followersperceptions of leader receptivity (F[1, 51] 9.47,p .01,2 .15), and the two main effects were notsignificant. Simple effects showed that as pre-dicted, when the confederates were proactive, fol-lowers rated leaders in the low-extraversion condi-tion as more receptive (mean 5.35, s.d. 0.87)than leaders in the high-extraversion condition

    (mean 4.32, s.d. 0.98;F[1, 52] 7.42,p .01).When the confederates were passive, this patternreversed: followers actually viewed leaders in thehigh-extraversion condition as more receptive(mean 5.34, s.d. 1.01) than those in the low-extraversion condition (mean 4.50, s.d. 1.13;F[1, 52} 4.98, p .05). One explanation for thispattern is that when followers are passive, theshier, more reserved style of less extraverted lead-ers is interpreted as a lack of interest. When follow-ers are proactive, on the other hand, this very styleis interpreted as openness. These results show that

    as predicted, employee proactivity moderated theeffect of leader extraversion on followers percep-tions of leader receptivity.

    In a hierarchical regression analysis predictingfollower performance, when we entered leader re-ceptivity, the interactive effect of the leader extra-version and follower proactivity manipulations de-creased to nonsignificance, and leader receptivitywas a significant predictor of follower perfor-mance. Table 5 presents the results of this analysis.We completed the test of mediation by testing thesize of the indirect effects of our manipulations on

    follower performance through perceived leader re-ceptivity. We used bootstrap procedures to con-struct bias-corrected confidence intervals based on1,000 random samples with replacement from thefull sample (Shrout & Bolger, 2002). The 95% con-fidence interval for the indirect interaction effectthrough perceived leader receptivity excluded zero( 0.89, 16.43), indicating statistical significanceand supporting mediation of the moderating effect

    by perceived leader receptivity. Additional analy-ses showed that when the confederates were pro-active, the 95% confidence interval for the indirect

    FIGURE 2Study 2: Results for Group Performance

    120

    130

    140

    150

    160

    170

    180

    190

    200

    Low Extraversion High Extraversion

    Passivefollowers

    Proactivefollowers

    T-ShirtsFolded

    Leadership Condition

    2011 541Grant, Gino, and Hofmann

  • 7/23/2019 GrantGinoHofmann Reversing

    15/23

    effect of leader extraversion on follower perfor-mance through perceived leader receptivity ex-cluded zero (0.68, 17.68). In contrast, when theconfederates were passive, the 95% confidence in-terval for the indirect effect of leader extraversionon follower performance through perceived recep-tivity included zero (3.66, 1.53). The first-stagemoderation effect was significantly stronger in thehigh proactivity condition than the low proactivitycondition (95% confidence interval for the differ-ences: 0.21, 1.60), as was the overall indirect effect(95% confidence interval for the differences: 0.27,

    15.79). These results support Hypothesis 2, show-ing that perceptions of leader receptivity mediatedthe moderating effect of employee proactivity onthe relationship between leader extraversion andgroup performance.

    Alternative Explanation

    A rival account of our findings is that leaders inthe high- (low-)extraversion condition were less(more) likely to adopt the proactive suggestion, andusing an inferior (superior) method caused their

    groups to perform less (more) effectively. Accord-ing to this perspective, followers perceptions ofleader receptivity may be a by-product of leadersactual decisions about whether to accept or rejectthe proactive suggestion made by the confederates.To test this possibility, we examined the datawithin the proactive conditions. The proportions ofleaders who accepted and adopted the new methodwere 8/14 (57.1%) in the low-extraversion condi-tion and 9/14 (69.2%) in the high-extraversion con-dition, and these proportions did not differ statis-tically (2[1] 0.42, n.s.). Thus, the extraversion

    manipulation did not influence the likelihoodthat leaders accepted and adopted the proactivesuggestion.

    Furthermore, the performance of the groups thatadopted and did not adopt the new method did notsignificantly differ. Followers whose leader ac-cepted the idea (mean 161.82, s.d. 34.87) didnot fold significantly more T-shirts than thosewhose leader rejected the idea (mean 156.60,s.d. 29.26; t[25] 0.40, n.s.). Our observationssuggest that, although the new method had thepotential to be more efficient, the ten-minute inter-

    val did not provide most teams with the opportu-nity to achieve this potential, as the time lost inteaching and learning the new method offset thepotential gains.

    These results rule out the alternative explanationthat adopting a more effective or efficient methoddrove the differences in the performance of proac-tive groups in the high- versus low-leader-extraver-sion conditions. The moderated mediation analy-ses support our hypothesis that differencesin perceived leader receptivity are what motivatethe observed differences in performance. When

    the confederates were proactive, participants per-ceived the more extraverted leaders as less recep-tive to ideas, and they invested less effort in thetask.2 In addition to supporting our proposed mo-

    2 This raises a critical question about how the leadersin the high-extraversion condition came to be perceived

    by followers as less receptive than their counterparts inthe low-extraversion condition. Although further re-search is necessary to shed light on the relevant micro-mediators, the high-extraversion manipulation may have

    TABLE 5Study 2: Results of Moderated Regression Analyses Predicting Follower Performancea

    Variable

    Step 1 Step 2

    b s.e. t b s.e. t

    Group size 3.73 6.93 .08 0.54 5.04 6.75 .10 0.75

    Leader extraversion 5.96 5.47 .22 1.09 3.17 5.48 .12 0.58Follower proactivity 10.39 5.47 .38 1.90 7.57 5.48 .27 1.38Leader extraversion follower proactivity 19.26 7.62 .60 2.53* 12.71 8.05 .40 1.58Perceived leader receptivity 3.75 1.82 0.29 2.06*

    R2: F(4, 51): R2: F(5, 50): R2: F(1, 50):.13 1.94 .20 2.50* .07 4.23*

    a n 56. When we entered the interaction term in a separate step between the first and second, variance explained increased by 11percent, from R2 .02 to R2 .13 (F[1, 51] 6.38, p .02). We replicated these patterns of results for group performance (the sum offollower and leader T-shirts) as the dependent variable.

    *p .05**p .01

    ***p .001

    542 JuneAcademy of Management Journal

  • 7/23/2019 GrantGinoHofmann Reversing

    16/23

    tivational mechanism of perceived leader receptiv-ity, this study provides a conservative test of ouroverall hypotheses, as it shows that lower leaderextraversion can improve the performance of pro-active groups even when their ideas are not actu-ally superior or more efficient.

    GENERAL DISCUSSION

    When does extraverted leadership contribute tohigher group performance? In both a field and alaboratory study, we found that when employeeswere not proactive, extraverted leadership was as-sociated with higher group performance. However,when employees were proactive, this patternreversed, so that extraverted leadership was as-sociated with lower group performance. We dem-onstrated this crossover interaction using twodifferent measures of group performance, bothnaturally occurring and controlled proactive be-haviors, and operationalizations of leader extra-version as a personality trait and a behavioralstyle. Our findings offer meaningful theoreticalcontributions to the literatures on leadership andproactivity in organizations.

    Theoretical Contributions

    Our primary contribution lies in identifying animportant boundary condition for when groups ledin an extraverted manner perform more effectively.Our research suggests that complementarity be-

    tween leadership style and employee proactivitycontributes to group performance. Specifically, thehighest level of group performance was achievedeither when a lack of proactivity from employeeswas paired with a more extraverted leadership styleor when employee proactivity was paired with aless extraverted leadership style. Group perfor-mance was hindered when both employees andleaders acted in a more dominant, agentic manner(i.e., proactive employees, extraverted leader) andwhen neither employees nor leaders acted in this

    way (i.e., passive employees, less extravertedleader).

    Our research represents a step toward theoreti-cally integrating trait and contingency perspectiveson leadership. The recent resurgence of research onleader characteristics has focused primarily ontheir direct associations with effectiveness out-

    comes (Judge et al., 2009); less research has exam-ined contingencies that moderate these associa-tions. A handful of studies have examined howleader extraversion moderates the relationship ofleader-member exchange with performance andturnover (e.g., Bauer, Erdogan, Liden, & Wayne,2006) and strengthens the relationship betweenemotion recognition and transformational leader-ship (e.g., Rubin, Munz, & Bommer, 2005). How-ever, little research has identified contingenciesthat moderate the effects of extraverted leadershipon group performance. In the spirit of classic con-tingency theories of leadership (e.g., Fiedler, 1971;Vroom & Yetton, 1973), we introduced employeeproactivity as an important contingency for thegroup performance effects of extraverted leader-ship. Our results provide an explanation for thereversal of the extraverted leadership advantage byshowing that when employees are proactive, thosewho lead in a less extraverted style are viewed asmore receptive to employee proactivity. By show-ing that followers proactivity strengthens the per-formance of groups led by less extraverted individ-uals, our research lends conceptual and empiricalrigor to assertions in the popular press that extra-

    verted leadership is not necessarily a requirementfor group and organizational effectiveness (Ba-daracco, 2002; Collins, 2001; Jones, 2006).

    Our research also advances knowledge aboutproactivity in organizations. Researchers fre-quently assume that employees proactive behav-iors contribute to group performance but haverarely tested this assumption, focusing instead onthe antecedents of proactive behaviors (Grant et al.,2009). The few studies that have examined theconsequences of proactive behaviors have done soat the individual level, examining implications

    for employees performance evaluations (e.g., VanDyne & LePine, 1998) and career success (e.g., Seib-ert et al., 2001). Although scholars have presentedconceptual arguments that proactive behaviors fa-cilitate higher group performance, little empiricalresearch has tested these specific linkages (Grant &Ashford, 2008; Griffin et al., 2007). Our research isamong the first efforts to examine the group perfor-mance consequences of employees proactive

    behaviors. We found that whether employees pro-active behaviors increase or decrease group perfor-mance depends on the degree to which their lead-

    encouraged leaders to interrupt more frequently, attemptto reassert their visions and authority, show less enthu-siastic facial expressions, and actively discourage furtherideas and suggestions, and the low-extraversion manip-ulation may have encouraged leaders to listen carefullyand show their appreciation for followers contributions.Because of these differences in receptive responses to theconfederates proactivity, followers with leaders in thehigh-extraversion condition may have felt less valuedand thus less motivated to contribute, and followers withleaders in the low-extraversion condition may have feltmore valued and thus more motivated to contribute.

    2011 543Grant, Gino, and Hofmann

  • 7/23/2019 GrantGinoHofmann Reversing

    17/23

    ers act extraverted. Our research reveals howemployees proactive behaviors contribute to groupperformance when leaders are quiet and reserved

    but can actually undermine group performancewhen leaders are active and assertive. These resultsaddress calls to understand the conditions underwhich employees proactive behaviors have a neg-

    ative rather than positive impact on group perfor-mance (Grant & Ashford, 2008).

    Our findings thereby move toward synthesizing theliteratures on leadership and proactivity, presenting anew perspective on employee proactivity as a contin-gency for leadership effects. Existing research hastreated leadership as an antecedent of employeesproactive behaviors (e.g., Detert & Burris, 2007; Grif-fin, Parker, & Mason, 2010; Parker et al., 2006), over-looking the possibility that leadership interacts withemployees proactive behaviors to influence perfor-mance outcomes. Our studies provide new evidence

    that leadership is not only an influence on employeeproactivity; its effects can also be shaped and altered

    by employee proactivity.In addition, our research raises questions about

    whether job enrichment and empowerment interven-tions may have unintended consequences. A numberof studies have shown that enriching jobs to provideautonomy and empowerment is associated withhigher levels of employee proactivity (Grant & Parker,2009). However, this research has yet to examine howleaders traits and styles affect the consequences ofthese proactive behaviors. Our studies point to the

    provocative possibility that when extraverted leadersenrich jobs to provide autonomy and empowerment,they may respond to employees ensuing proactive

    behaviors in ways that undermine their potential tocontribute to improved performance.

    Finally, our research extends knowledge aboutthe objective, not only the subjective, consequencesof dominance complementarity theory. Previousstudies have focused primarily on the psychologi-cal consequences of complementarity, including at-traction, liking, cohesion, and observer perfor-mance ratings (e.g., Kristof-Brown et al., 2005;Fragale, Tiedens, & Lee, 2003). In contrast, littleresearch has linked dominance complementarity toobjective performance outcomes, especially at thegroup level, which is also a critical oversight inleadership research (Kaiser et al., 2008). Our re-search begins to fill this gap by documenting theobjective performance benefits of complementarity,operationalized in terms of the pairing of extra-verted, assertive leadership with minimal em-ployee proactivity or less extraverted, more re-served leadership with high employee proactivity.

    Limitations and Future Directions

    Our studies are subject to a number of limitationsthat suggest directions for future research. In bothof our studies, leaders were in charge of groupsresponsible for relatively structured, repetitive,effort-based tasks (delivering pizzas and foldingT-shirts). It remains to be seen whether the patternsgeneralize to more difficult, complex, or creativetasks. It is possible that differences in leader extra-version may have a stronger effect in more complextasksnot only through the motivational mecha-nism of perceived leader receptivity, but alsothrough the knowledge mechanism of enablingleaders to utilize better task strategies. Uncertaintyabout leader power and status (Kramer, 1998) may

    be another important boundary condition. Whenleaders have high referent power and highachieved or ascribed status, they may be open toproactivity regardless of their levels of extraver-

    sion, as they feel that their standing and authorityare not being threatened. Similar predictions may

    be made for goal interdependence and value con-gruence, which may help to align leaders and em-ployees around particular forms of proactivity thatfacilitate goal pursuit and value expression. In ad-dition, our results may be circumscribed to rela-tively constructive forms of proactive behaviors. InStudy 1, we did not measure whether employeesproactive behaviors were targeted in productive di-rections, and in Study 2, our confederates sug-gested an idea that had the potential to improve

    performance. Will less extraverted leaders still bereceptive to more self-serving or destructive ex-pressions of proactivity? Will more extravertedleaders be even less receptive to these forms ofproactivity?

    On a related note, extraversion is a multifacetedtrait, and we did not unpack which particular fac-ets were responsible for the effects observed. Is itdominance-assertiveness, sociability, or a combina-tion of these facets that drives the different re-sponses of leaders to employees proactive behav-iors? In addition, we did not test the psychological,

    behavioral, and social processes through whichleaders with different levels of extraversion reacteddifferently to employees proactive behaviors. Itwill be critical for researchers to test several mech-anisms. Do less extraverted leaders listen morecarefully? Are they more focused and less dis-tracted? Do they experience proactivity as less ego-threatening and actively encourage more proactiv-ity? When employees are proactive, are lessextraverted leaders less overconfident (see Ng, Ang,& Chan, 2008) or less overwhelmed by the socialand intellectual demands of leadership roles (see

    544 JuneAcademy of Management Journal

  • 7/23/2019 GrantGinoHofmann Reversing

    18/23

    Little & Joseph, 2006; McGregor, McAdams, & Lit-tle, 2006). We also did not examine the mecha-nisms through which leaders reactions to employ-ees proactive behaviors influenced groupperformance. Further research will enable a deeperunderstanding of the explanatory processes under-lying our findings.

    Interestingly, our findings appear to contrastwith research on leader prototypicality, which sug-gests that when leaders are viewed as similar to andrepresentative of their groups, these groups per-form more effectively (Lipponen, Koivisto, & Olk-konen, 2005; van Knippenberg & Hogg, 2003; vanKnippenberg & van Knippenberg, 2005). From thisperspective, proactive employees may be particu-larly motivated when working with extravertedleaders who share their assertive tendencies. Al-though additional research is necessary to addressthis issue in further depth, one interpretation of thediscrepancy is that employees responses to proto-typical leaders vary as a function of the contentdimension under consideration. Employees mayrespond favorably to leaders who share their val-ues, attitudes, beliefs, and characteristics, but dom-inance complementarity emerges with respect to

    behaviors that are zero-sum: it is difficult for highlyextraverted leaders to be the center of attentionwhen employees are proactive, and it is equallychallenging for proactive employees to advance

    bottom-up change when highly extraverted leadersimpose their ideas. Highly extraverted leaders andproactive employees have different goals and ex-

    pectations about the degree of leader control versusemployee input that is appropriate, which may bewhy prototypicality is not ideal with respect tothese content dimensions. When proactive employ-ees work with less extraverted leaders, on the otherhand, the ensuing complementarity establishes areciprocal relationship in which both sides havetheir needs met. . . . This type of mutually fulfillinginteraction is desirable to both sides (Kristof-Brown et al., 2005: 939940). If this interpretationis accurate, it suggests that leader prototypicalityeffects are bounded to dimensions of similarity that

    are not zero-sum and can be shared by both sides.

    Practical Implications and Conclusion

    Our research offers valuable practical insights forboth leaders and employees. For leaders, our studieshave three key implications. First, our findings pro-vide less extraverted leaders with a new set of toolsfor directing their groups toward effective perfor-mance. The popular press is replete with suggestionsfor individuals low in extraversion to build on theirquiet strength by practicing their public speaking

    skills (Kahnweiler, 2009), achieve the introvert ad-vantage by smiling more frequently (Laney, 2002),leverage introvert power by taking breaks andscheduling time to think (Helgoe, 2008), and taketheir companies from good to great by being quietand reserved but still strong-willed (Collins, 2001). Incontrast to these speculations, our research highlights

    a theoretically sound, empirically supported strategywhereby less extraverted individuals can facilitategroup performance: actively encourage proactive be-haviors on the part of employees. By being receptiveto employees efforts to voice ideas, take charge toimprove work methods, and exercise upward influ-ence, less extraverted leaders can develop more effi-cient and effective practices that enhance groupeffectiveness.

    Second, our findings may provide highly extra-verted leaders with action steps for improvinggroup performance. In settings and situations inwhich proactive suggestions are important, leaderswho naturally tend to be assertive may wish toadopt a more reserved, quiet style. Since our labo-ratory experiment indicated that individuals cantemporarily change their patterns of behavior, thisprescription may be tenable. Third, in the spirit ofdeveloping future leaders, existing leaders mayfind it useful to train more extraverted managers totake notice of, utilize, recognize, and reward em-ployees proactive behaviors. Finally, for employ-ees, our findings suggest that proactive behaviorsmay be more effective with quieter leaders who aremore receptive. It may be wise, then, for employees

    to make particular efforts to voice suggestions, takecharge, and exert upward influence when workingwith less extraverted leaders. Employees may alsoseek out such leaders as audiences for their proac-tive ideas. In conclusion, our findings reveal howleader extraversion can be a liabilitynot only anassetfor group performance.

    REFERENCES

    Agle, B. R., Nagarajan, N. J., Sonnenfeld, J. A., & Sriniva-san, D. 2006. Does CEO charisma matter? An empir-

    ical analysis of the relationships among organization-al performance, environmental uncertainty, and topmanagement team perceptions of CEO charisma.

    Academy of Management Journal,49: 161174.

    Aiken, L. S., & West, S. G. 1991. Multiple regression:Testing and interpreting interactions. NewburyPark, CA: Sage.

    Ames, D. R., & Flynn, F. J. 2007. What breaks a leader:The curvilinear association between assertivenessand leadership. Journal of Personality and Social

    Psychology,92: 307324.

    Anderson, C., John, O. P., Keltner, D., & Kring, A. M. 2001.

    2011 545Grant, Gino, and Hofmann

  • 7/23/2019 GrantGinoHofmann Reversing

    19/23

    Who attains social status? Effects of personality andphysical attractiveness in social groups. Journal of

    Personality and Social Psychology,81: 116132.

    Anderson, C., Spataro, S. E., & Flynn, F. J. 2008. Person-ality and organizational culture as determinants ofinfluence.Journal of Applied Psychology, 93: 702710.

    Anderson, C. R., & Schneier, C. E. 1978. Locus of control,leader behavior and leader performance among man-agement students. Academy of Management Jour-

    nal,21: 690698.

    Aragon-Correa, J. A. 1998. Strategic proactivity and firmapproach to the natural environment. Academy of

    Management Journal,41: 5565