carson et al 2007

7/23/2019 Carson Et Al 2007

1/19

SHARED LEADERSHIP IN TEAMS:AN INVESTIGATION OF ANTECEDENT CONDITIONS

AND PERFORMANCE

JAY B. CARSON

Southern Methodist University

PAUL E. TESLUKUniversity of Maryland

JENNIFER A. MARRONESeattle University

Shared leadership refers to a team property whereby leadership is distributed amongteam members rather than focused on a single designated leader. We examined ante-cedent conditions that lead to the development of shared leadership and the influenceof shared leadership on team performance in a sample of 59 consulting teams. Both theinternal team environment, consisting of shared purpose, social support, and voice,

and external coaching were important predictors of shared leadership emergence. Inturn, shared leadership was found to predict team performance as rated by clients. Weconclude by discussing the implications of these findings for team leadership andeffectiveness.

Leadership is considered crucial for enablingteam effectiveness (Cohen & Bailey, 1997; Hackman& Walton, 1986; Kozlowski, Gully, Salas, & Can-non-Bowers, 1996), and some researchers haveeven argued that it is the most critical ingredient(Sinclair, 1992; Zaccaro, Rittman, & Marks, 2001).Yet most existing research on team leadership has

focused narrowly on the influence of an individualteam leader (usually a manager external to a team),thus largely neglecting leadership provided byteam members (Kozlowski & Bell, 2003; Stewart &Manz, 1995). Several trends in team design, use,and structure, however, point to the importance ofinternal team leadership. First, the complexity andambiguity that teams often experience make it un-likely that a single external leader can successfullyperform all necessary leadership functions (Day,Gronn, & Salas, 2004). Second, current forms ofteamwork that emphasize knowledge-based work

rely on employees who have high levels of exper-tise and seek autonomy in how they apply theirknowledge and skills (DeNisi, Hitt, & Jackson,2003) and therefore desire greater opportunity toshape and participate in the leadership functionsfor their teams. Further, flatter organizational struc-

tures and the pervasive presence of self-managingteams, which are now well established and deeplyrooted in U.S. industry (Lawler, Mohrman, & Ben-son, 2001; Manz & Sims, 1987), emphasize the needfor leadership originating from within a team asopposed to that originating from a single individualelevated by hierarchy. Despite this transition inleadership responsibilities from formal managers toteam members, relatively little research has ad-dressed the implications of this evolutionary shiftto internally distributed forms of team leadership.

Early leadership scholars argued for the impor-tance of leadership being shared among team mem-

bers (Gibb, 1954; Katz & Kahn, 1978). Gibb, the firstto so argue, stated, Leadership is probably bestconceived as a group quality, as a set of functionswhich must be carried out by the group. This con-cept of distributed leadership is an importantone (1954: 884). Katz and Kahn (1978) also sug-gested that when team members voluntarily andspontaneously offer their influence to others insupport of shared goals, shared leadership can pro-vide organizations with competitive advantagethrough increases in commitment, in the personaland organizational resources brought to bear oncomplex tasks, in openness to reciprocal influencefrom others, and in the sharing of information. Con-sequently, they argued, Those organizations inwhich influential acts are widely shared are mosteffective (Katz & Kahn, 1978: 332). These early

We would like to thank Hank Sims, Lisa Dragoni, andHolly Slay, as well as Brad Kirkman and three anony-mous reviewers, for their helpful comments on earlierversions of this paper.

Academy of Management Journal

2007, Vol. 50, No. 5, 12171234.

1217Copyright 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 Carson Et Al 2007

2/19

perspectives challenged the convention that lead-ership is solely an individual phenomenon, butthere has been little empirical work on sharedforms of leadership until recently.

We define shared leadership as an emergent teamproperty that results from the distribution of lead-ership influence across multiple team members. It

represents a condition of mutual influence embed-ded in the interactions among team members thatcan significantly improve team and organizationalperformance (Day et al., 2004). Shared leadershipcontrasts with the conventional paradigm (referredto as vertical leadership by Pearce and Sims[2002]), which emphasizes the role of the managerwho is positioned hierarchically above and exter-nal to a team, has formal authority over the team,and is responsible for the teams processes andoutcomes (e.g., Druskat & Wheeler, 2003; Hackman& Walton, 1986; Kozlowski et al., 1996). Recentempirical work has demonstrated links betweenshared leadership and team performance (e.g.,Avolio, Jung, & Sivasubramaniam, 1996; Ensley,Hmieleski, & Pearce, 2006; Pearce & Sims, 2002;Sivasubramaniam, Murry, Avolio, & Jung, 2002).Although these studies have helped advance theconcept of shared leadership, several remaining re-search gaps motivated this study.

The first purpose of our study was to begin totheoretically identify and test conditions that sup-port shared leadership in work teams. To ourknowledge, no studies have directly explored theconditions that give rise to shared leadership. Re-

flecting a perspective on leadership in teams as adynamic process involving interactions betweenteam members and external team leaders (cf. Koz-lowski et al., 1996; Zaccaro & Klimoski, 2002), weconsider conditions both internal and external to ateam. Two proximal factors that are likely to influ-ence the development of shared leadership are in-ternal team environment, including a shared pur-pose, social support, and voice, and level ofexternal coaching support. Considering these ante-cedent conditions for shared leadership also addsto the scant literature on interactions between ex-

ternal and internal team leadership, which is im-portant since both forms of leadership operate con-currently and in conjunction with one another (e.g.,Manz & Sims, 1987).

A second purpose was to provide an improvedconceptualization and operationalization of theshared leadership construct that reflects its theoret-ical complexity. Scholars have called for more at-tention to theoretical models of team leadershipthat are developed at the team level rather than asmere extrapolations of existing dyadic leadershipapproaches (Kozlowski & Bell, 2003). Prior work on

shared leadership has relied primarily on aggregat-ing team members assessments of the degree towhich leadership responsibilities are shared or cer-tain behaviors are exhibited within a team, a pro-cedure that may fail to capture the relational natureof the patterns of shared influence in teams. Table1, below, provides details on this previous work.

Utilizing social network theory (e.g., Brass, 1995;Mehra, Smith, Dixon, & Robertson, 2006), we ad-vanced a more complete conceptualization of theemergent and relational nature of shared leadershipand took a social network measurement approachin order to better capture these overall patterns ofinfluence.

A third purpose of this study was to predictperformance outcomes. Existing research on sharedleadership has almost exclusively relied on teammembers and/or external leaders ratings of perfor-mance, a practice that raises concerns about bothcommon method variance and the ability to obtainan independent assessment of a teams perfor-mance. (Again, Table 1, below, provides details.)We studied the performance outcomes of sharedleadership using a source independent of teamsand their immediate leadership structures: specifi-cally, the end users of the teams work products.

LITERATURE REVIEW AND HYPOTHESES

Shared Leadership Defined

Gibb (1954) first suggested the idea of two forms

of team leadership: distributed and focused. Fo-cused leadership occurs when leadership resideswithin a single individual, whereas distributedleadership occurs when two or more individualsshare the roles, responsibilities, and functions ofleadership. Gronn (2002) argued that these twoconcepts of focused and distributed leadership beconsidered endpoints on a continuum rather thanrigid either-or categories.

To further develop the concept of how leadershipis shared among team members, we utilize Yuklsdefinition of leadership as influence processes in-

volving determination of the groups or organiza-tions objectives, motivating task behavior in pur-suit of these objectives, and influencing groupmaintenance and culture (1989: 5). Building onthe concept of leadership as influence and drawingon multilevel theory (Kozlowski & Klein, 2000;Morgeson & Hofmann, 1999), we define sharedleadership as an emergent team property that re-sults from the distribution of leadership influenceacross multiple team members. In keeping with thenotion of collective constructs (Morgeson & Hof-mann, 1999), we argue that shared leadership orig-

1218 OctoberAcademy of Management Journal

7/23/2019 Carson Et Al 2007

3/19

inates with individual members of a team engagingin activities that influence the team and other teammembers in areas related to direction, motivation,and support (Yukl, 1989) and through the series ofinteractions that team members have with eachother involving the negotiation and sharing of lead-ership responsibilities. The resulting collective

structure can be considered to be a leadership net-work that influences and shapes both team andindividual activities and outcomes.

Leadership can be conceptualized in relation toeither the strength of influence (i.e., its quality oreffectiveness) or the source of influence (i.e., singleversus multiple team members). Our definition is

TABLE 1Previous Definitions and Measures of Shared Leadership

Study Definition Measure Dependent Variable

Avolio, Jung, Murry, andSivasubramanium(1996)

No explicit definition given, but sharedleadership is essentially viewed astransformational leadershipmanifested at the group level inhighly developed teams.

Team Multifactor LeadershipQuestionnaire (TMLQForm 5X) aggregated to theteam level

Self-reported ratings(undergraduate project teameffectiveness)

Pearce and Sims (2002) Distributed influence from within theteam (p. 172).

Lateral influence among peers (p. 176).

Ratings (aggregated to teamlevel) on behavioral scalesfor five leadershipstrategies: aversive,directive, transactional,transformational, andempowering

Self-reported and managerratings of seveneffectiveness dimensions(automobile changemanagement teams)

Sivasubramanium,Murry, Avolio, andJung (2002)

Collective influence of members in ateam on each other (p. 68).

How members of a group evaluatethe influence of the group as opposedto one individual within or externalto the group (p. 68).

Team Multifactor LeadershipQuestionnaire (TMLQForm 5X) aggregated to theteam level

Team potency (self-ratings attimes 1 and 2) and teamgrades assigned byinstructor (undergraduateproject team effectiveness).

Pearce and Conger(2003)

A dynamic, interactive influenceprocess among individuals in groupsfor which the objective is to lead oneanother to the achievement of groupor organizational goals or both. . . .[L]eadership is broadly distributedamong a set of individuals instead ofcentralized in [the] hands of a singleindividual who acts in the role of asuperior (p. 1).

Not applicable Not applicable

Pearce, Yoo, and Alavi(2004)

Simultaneous, ongoing, mutualinfluence process within a team that

is characterized by serialemergence of official as well asunofficial leaders (p. 48).

Ratings (aggregated to teamlevel) on behavioral scales

for four leadershipstrategies: directive,transactional,transformational, andempowering

Self-ratings of problem-solvingquality and effectiveness

(virtual teams of studentsocial workers)

Ensley, Hmieleski, andPearce (2006)

Team process where leadership iscarried out by the team as a whole,rather than solely by a singledesignated individual (p. 220).

Ratings (aggregated to teamlevel) on behavioral scalesfor four leadershipstrategies: directive,transactional,transformational, andempowering

Growth index for newventures, consisting of theaverage of firm revenuegrowth and employeegrowth rates (new ventureTMTs)

Mehra, Smith, Dixon,and Robertson (2006)

Shared, distributed phenomenon inwhich there can be several (formally

appointed and/or emergent) leaders(p. 233).

Qualitative coding based onvisual analysis of

leadership networkdiagrams

Team sales divided by teamsize (financial services sales

teams)

2007 1219Carson, Tesluk, and Marrone

7/23/2019 Carson Et Al 2007

4/19

focused on multiplesourcesof influence and refersto widespread influence within teams rather thanto specific leadership behaviors, formal positions,specific types of influence, or the effectiveness ofthe leadership exhibited by these sources. Buildingon these ideas of distributed influence and drawingupon Gibbss (1954) original conceptualization, we

believe shared leadership can be conceptualizedalong a continuum based on the number of leader-ship sources (i.e., team members) having a highdegree of influence in a team. Anchoring the lowend of the continuum are cases in which teammembers follow the leadership of a single individ-ual. Although the leadership exhibited by this sin-gle individual might be quite strong, leadershiphere originates from only a single source. In con-trast, at the high end of the shared leadership con-tinuum are teams in which most, if not all, teammembers provide leadership influence to one an-other. Here, the source of leadership influence isdistributed among team members rather than con-centrated or focused in a single individual. In theseteams, team members both lead and follow oneanother in such a way that at a given time, membersare both providing leadership for certain aspects ofteam functioning and also responding to the lead-ership provided by other team members in differentareas. Teams with high levels of shared leadershipmay also shift and/or rotate leadership over time, insuch a way that different members provide leader-ship at different points in the teams life cycle anddevelopment.

Shared leadership is a relational1

phenomenoninvolving mutual influence between team mem-

bers as they work toward team objectives, andsocial network theory provides a natural theoret-ical and analytical approach to studying the re-lational influence structure in teams (Mehra etal., 2006). The exercise of leadership influence(Yukl, 1989) occurs in the context of team mem-

ber relationships and presumes the existence offollowers or influencees (Bedeian & Hunt,2006). Shared leadership creates patterns of re-ciprocal influence that further develop and rein-

force existing relationships among team mem-bers. Thus, social network theory is appropriateas it examines patterns of relationships among

individuals such as advice, information, andfriendship networks and emphasizes the relation-ship as the unit of analysis (Brass, 1995; Spar-rowe, Liden, Wayne, & Kraimer, 2001). Further,social network analysis allows for the study ofmultiple sources of leadership influence and theability to model patterns of influence within a

team and preserve rich data about the actual dis-tribution of influence (Mehra et al., 2006).In keeping with social network theory, we argue

that the pattern of emerging mutual influence inteams can be conceptualized as an increase in thedensity of the teams internal leadership networks.A leadership network is the pattern of individualswho rely on others for leadership within a team,and density increases as this reliance on one an-other for leadership grows. Density, as used con-ceptually in social network research, is a structuralproperty representing the pattern of relationshipswithin a team, and it describes the overall level ofdifferent types of exchanges among members of agiven social network (Sparrowe et al., 2001). Spar-rowe and his colleagues described this team-levelconstruct as follows: Density is analogous to themean number of ties per group member. The moreties each group member enjoys with the othergroup members, the greater the density of the net-work (2001: 317). Here, ties between team mem-

bers (also referred to as relationships) exist whenone team member perceives another as exertingleadership influence in the team. Thus, the densityof a leadership network is the mean number of

relationships (per team member) involving leader-ship influence. When more team members provideleadership to their peers, the density of this type ofnetwork increases. Operationally, network densityis a measure of the proportion of total possiblerelationships (actual ties over potential ties) thatexists in a given network (Wasserman & Faust,1994) and thus captures variance in the overallpatterns of relationships rather than variance inshared perceptions of a construct (as is the casewith aggregated behavioral scales). Accordingly,utilizing network density as a measure of shared

leadership appropriately reflects the extent towhich leadership influence is distributed among arelatively high or relatively low proportion of teammembers.

Relationship with Similar Constructs

Having defined and described the nature ofshared leadership, it is also helpful to describe

briefly its relationship to other similar constructs,such as team autonomy or self-management, teamempowerment, cooperation, team cognition (e.g.,

1 When we use the terms relational and relation-ships here, we are simply referring to the interactions

between two or more people. These terms are not meantto imply that shared leadership presumes the existenceof or is the result of close personal friendships amongteam members. We thank an anonymous reviewer forhighlighting the need to make this distinction.


7/23/2019 Carson Et Al 2007

5/19

transactive memory systems and team mental mod-els), and emergent leadership. Self-managing andautonomous team designs are those in which teammembers have greater responsibility for settingtheir own goals, monitoring their own progress,and making their own decisions than do teammembers in manager-led teams (Hackman, 1987).

Although self-managing team designs may promotethe development of shared leadership through in-creased self-management (Manz & Sims, 1987), orthrough heightened trust or autonomy (Langfred,2004), such designs in themselves do not necessar-ily result in leadership influence being widely dis-tributed in a team, as other factors, such as theinternal team environment and external coaching,may also influence shared leadership (Wageman,2001).

Team empowerment is a motivational constructand has been defined as the collective experienceof heightened levels of task motivation as a result ofteam members assessments of their teams tasks asproviding them with high levels of meaningfulness,autonomy, impact, and potency (Kirkman & Rosen,1997). From a temporally dynamic perspective,team empowerment can be viewed as an emergentstate that either precedes or follows team processes,depending on the stage of a teams developmentand performance cycle (cf. Marks, Mathieu, & Zac-caro, 2001). From this perspective, team empower-ment may facilitate the development of sharedleadership by motivating team members to exerciseinfluence. Conversely, shared leadership may also

lead to greater team empowerment by heighteningmembers sense of meaningfulness, autonomy, im-pact, or potency, depending on the stage of a teamsdevelopment. However, a team may experience ahigh level of empowerment yet still have a strongexternal leader providing most of the leadershipinfluence, with very little shared leadership exhib-ited by team members.

Shared leadership is related to but distinct fromother team processes such as cooperation or help-ing, which refer to team members working withand/or assisting other team members with their

tasks (Kozlowski & Bell, 2003). Although thesetypes of behavior relate to being an effective teammember and promote efficiency, they do not in-volve the active influence that is essential to lead-ership. In keeping with this conceptualization, arecent study showed only a moderate correlation

between shared leadership and cooperation orhelping (Ziegert, 2005).

Shared leadership is also distinct from team cog-nition constructs, such as transactive memory sys-tems (TMSs), or structures through which memberscan collectively encode, store, and retrieve infor-

mation and expertise (Wegner, 1986), and teammental models (TMMs), or shared understandingsabout the attributes of a team or its task at hand(Cannon-Bowers, Salas, & Converse, 1993). Con-ceptually, the primary distinction between sharedleadership and these team knowledge structures isthat the former concerns collective influence,

whereas the latter concerns collective cognition.This conceptual difference may perhaps best beseen in the distinction between measurement ap-proaches. Shared leadership assesses the distribu-tion of leadership among team members. TMS mea-sures capture team-level systems for utilizing andintegrating individually and collectively held ex-pertise (Lewis, 2003). TMM measures assess thesimilarity and accuracy of individual mental mod-els within a team (Marks, Zaccaro, & Mathieu,2000). Although these are three distinct concepts,TMMs and TMSs are likely both facilitated by

shared leadership, through the continual influence-based interactions and social exchanges (Klimoski& Mohammed, 1994) that occur as team membersshare leadership responsibilities. Reciprocally,through effective coordination of expertise and thedevelopment of mutual understandings, TMSs andTMMs likely enable the emergence of sharedleadership.

Finally, emergent leadership refers to groupmembers exerting significant influence over othermembers of their group although no formal author-ity has been vested in them (Schneier & Goktepe,

1983). Shared leadership is consistent with some ofthe early group research by Bales (1953), whofound that two informal leaders often emerge inleaderless groups: one focused on the group task,and one concentrating on relational issues. Thisresearch is similar to that on shared leadership inthat it typically concerns whether leadership isprovided informally by a group member (known asan emergent leader) in addition to or instead of

being provided by a formally appointed leader (e.g.,Wheelan & Johnston, 1996). However, emergentleadership research differs by focusing on the char-

acteristics of individuals and groups that predictinformal leadership emergence, as well as narrowlyconsidering only one or two persons as emergentleaders and ignoring the leadership influence ofothers. In sum, shared leadership is distinct fromemergent leadership in that the former can takeplace in a team with or without a designated leader,can be either formal or informal, and addresses thedistribution and sharing of leadership among allteam members, in contrast to only one or twoleaders.


7/23/2019 Carson Et Al 2007

6/19

Antecedent Conditions: Internal and External

Researchers studying shared leadership have ar-gued that for shared leadership to emerge, two setsof activities must occur (Katz & Kahn, 1978). First,the members of a team must offer leadership andseek to influence the direction, motivation, andsupport of the group. Second, the team as a whole

must be willing to rely on leadership by multipleteam members. For these individual and collective

behaviors to occur, team members must believethat offering influence to and accepting it fromfellow team members are welcome and construc-tive actions. We considered key factorsboth in-ternal and externalthat are likely to impact thedevelopment of shared leadership in teams throughthese mechanisms. The first condition is an inter-nal team environment that supports the develop-ment of shared leadership over time, and the sec-ond is the level of supportive coaching provided by

an external leader.We propose first that shared leadership is facili-

tated by an overall team environment that consistsof three dimensions: shared purpose, social sup-port, and voice. These dimensions, drawn from areview of research on shared leadership (e.g.,Avolio et al., 1996; Barry, 1991; Pearce & Conger,2003; Seers, 1996; Yukl, 1989), represent distinctconcepts that are also highly interrelated and mu-tually reinforcing, thereby representing a higher-order construct (cf. Edwards, 2001; Law, Wong, &Mobley, 1998). We refer to them here, collectively,

as an internal team environment enabling sharedleadership because they work together to producethe kind of team context that encourages teammembers willingness to both offer leadership in-fluence and rely on the leadership of other teammembers (Katz & Kahn, 1978).

Shared purpose, the first dimension of an inter-nal team environment enabling shared leadership,exists when team members have similar under-standings of their teams primary objectives andtake steps to ensure a focus on collective goals.Prior work has theorized and demonstrated that

team members who have a common sense of pur-pose and agreed-upon goals are more likely to feelmotivated, empowered, and committed to theirteam and work (Kirkman & Rosen, 1999; Liden,Wayne, & Sparrowe, 2000; OLeary-Kelly, Martoc-chio, & Frink, 1994). These heightened levels ofmotivation, empowerment, and commitment thatindividuals experience when their team possessesa shared purpose increase the willingness of teammembers to share the teams leadership responsi-

bilities (Avolio et al., 1996). In addition, with acommonly understood set of objectives and direc-

tion, team members are more likely to establishgoals and take actions that support the activities ofother team members, thereby facilitating both goal-oriented and work-directive leadership behaviors

by team members (Seers, 1996), as well as a collec-tive direction to team activities (Yukl, 1989).

The second dimension of an internal team envi-

ronment that supports shared leadership is socialsupport, which is defined as team members effortsto provide emotional and psychological strength toone another. Team members support one anotherthrough encouraging and recognizing individualand team contributions and accomplishments(Marks et al., 2001). This helps to create an envi-ronment where team members feel that their inputis valued and appreciated. By actively participatingin a team and feeling supported, team members aremore likely to work cooperatively and develop asense of shared responsibility for team outcomes(Kirkman & Rosen, 1999). Social support is associ-ated with group maintenance and culture (Yukl,1989), leader support/supportive behaviors (Seers,1996), relational leadership (Barnard, 1938), anddeveloping and maintaining a team by providinginterpersonal glue that helps build a strong inter-nal social network (Barry, 1991).

The third dimension of this internal team envi-ronment is voice. No standard definition of voiceexists, as it has been used in a variety of researchareas to describe constructive change-orientedcommunication, participation in decision making,involvement, certain extra-role work behaviors,

due process, and employee grievance procedures(Van Dyne & LePine, 1998); however, at its core,voice connotes participation and input. We defineit here as the degree to which a teams membershave input into how the team carries out its pur-pose. Voice is associated with interaction facili-tation/participative behaviors in teams (Seers,1996), and these types of behaviors can result inhigher levels of social influence among team mem-

bers through increased engagement and involve-ment. In addition, voice has been associated withparticipation in decision making and constructive

discussion and debate around alternative ap-proaches to team goals, tasks, and procedures (DeDreu & West, 2001; Simons, Pelled, & Smith, 1999),which can improve the amount of collective influ-ence, involvement, and commitment relative to im-portant team decisions. Thus, the presence of highlevels of voice in a team should create an environ-ment where people engage in mutual leadership by

being committed to and becoming proactively in-volved in helping the team achieve its goals andobjectives and constructively challenging eachother in pursuit of group goals.


7/23/2019 Carson Et Al 2007

7/19

These three dimensions are mutually reinforcingand complementary. When team members are ableto speak up and get involved (voice), the likelihoodthat many of them will exercise leadership in-creases greatly. The opportunity for voice also fa-cilitates shared leadership by strengthening both acommon sense of direction and the potential for

positive interpersonal support in a team. Whenteams are focused on collective goals (shared pur-pose), there is a greater sense of meaning and in-creased motivation for team members to both speakup and invest themselves in providing leadershipto the team and to respond to the leadership ofothers. The motivation to participate and provideinput toward achieving common goals and a com-mon purpose can also be reinforced by an encour-aging and supportive climate. When team membersfeel recognized and supported within their team(social support) they are more willing to share re-sponsibility, cooperate, and commit to the teamscollective goals. Thus, these three dimensions worktogether to create an internal team environmentthat is characterized by a shared understandingabout purpose and goals, a sense of recognition andimportance, and high levels of involvement, chal-lenge, and cooperation. Therefore, we predict:

Hypothesis 1. An internal team environmentconsisting of shared purpose, social support,and voice is positively related to the level ofshared leadership in a team.

External Team Coaching

Scholars studying shared leadership and leader-ship in self-managing teams have noted the criticalrole of external team leaders in the development ofteam members motivation and capabilities to leadthemselves and become self-directed (Kozlowski etal., 1996; Manz & Sims, 1987). When discussingthis role, researchers have frequently stressed theimportance ofcoachingbehaviors, which Hackmanand Wageman defined as external team leadersdirect interaction with a team intended to help

team members make coordinated and task-appro-priate use of their collective resources in accom-plishing the teams task (2005: 269). Researchershave identified different types of team coaching,distinguishing between forms that are more sup-portive and reinforcing of a teams self-leadershipand those that focus on identifying team problemsand engaging in active task interventions that inter-fere with the teams autonomy and self-manage-ment (Morgeson, 2005; Wageman, 2001). Here, wespecifically refer to the former, which has beencalled supportive coaching (Morgeson, 2005) be-

cause it is more closely connected with the devel-opment of team self-management, initiative, andautonomy, whereas active coaching is more likelyto undermine these team characteristics and possi-

bly inhibit the development of shared leadership.Supportive coaching can also be distinguishedfrom other external team leadership functions,

such as designing a team and its task (Wageman,2001) and facilitating boundary management(Druskat & Wheeler, 2003).

Through supportive coaching, external teammanagers can contribute to the development ofshared leadership in a variety of ways. First, byengaging in behaviors such as encouraging, rein-forcing, and rewarding instances in which teammembers demonstrate leadership, supportivecoaching fosters a sense of self-competence andteam independence among team members (Manz &Sims, 1987). When team members believe that theyhave significant autonomy and are confident intheir skills for managing the work of their team,they should be more likely to demonstrate leader-ship. Supporting this assertion, Wageman (2001)found a positive association between supportivecoaching by a team manager and self-management

by team members. Second, by providing a teamwith encouragement and support, external coach-ing can help build a shared commitment to theteam and its objectives that can reduce free ridingand increase the likelihood that team members willdemonstrate personal initiative (Hackman & Wage-man, 2005). Third, by giving teams suggestions

about appropriate task strategies that ensure thattheir activities are well aligned with work require-ments and demands (Hackman & Wageman, 2005),supportive coaching provides team membersgreater clarity on how to best manage their workand processes. Thus, team members are more likelyto influence each other because they share thisunderstanding (Kozlowski et al., 1996). Therefore,we predict:

Hypothesis 2. External team coaching is posi-tively related to the level of shared leadershipin a team.

The second, more indirect, way in which exter-nal coaching may influence shared leadership is

based on a functional approach to team leadership.According to this approach, the role of an externalteam leader is to do whatever is not being ade-quately managed by the team itself (Hackman &Walton, 1986). When teams have a supportive in-ternal environment, team coaching by an externalteam leader is likely to be largely redundant withthis internal environment and therefore less criticalto the emergence of shared leadership among team


7/23/2019 Carson Et Al 2007

8/19

members. However, for teams that lack a strongshared purpose and do not promote full engage-ment and participation, and in which team mem-

bers are unable to provide each other with socialsupport, a functional leadership perspective sug-gests that external leaders coaching may be partic-ularly important. Specifically, effective team

coaching by an external leaderfocused on build-ing collective commitment to a team and its work,assisting the team with aligning activities with taskrequirements, and fostering independencecanhelp provide the motivational and consultativefunctions (Hackman & Wageman, 2005) that enableshared leadership but have not been adequatelydeveloped by the team internally. External teamleaders can also help team members understand thedifferent skills and capabilities of team membersand how they can be integrated to address the de-mands of the team task. This understanding canmotivate individual team members to initiate andengage in internal leadership activities and do so ina coordinated fashion resulting in an emergent pat-tern of shared leadership. In this fashion, an exter-nal team leaders supportive coaching can enableshared leadership to emerge when a team has yet todevelop a high level of social support, shared pur-pose, and voice. These relationships between aninternal team environment and an external leadersteam coaching suggest:

Hypothesis 3. Team coaching by an externalleader interacts with the internal team environ-ment in predicting shared leadership: coach-ing is more strongly related to shared leader-ship when the internal team environment isunsupportive.

Shared Leadership and Team Performance

Shared leadership is an important intangible re-source available to teams, and therefore it shouldenhance team performance on complex tasks (Dayet al., 2004). When team members offer their lead-ership to others and to the mission or purpose oftheir team, they should experience higher commit-

ment, bring greater personal and organizational re-sources to bear on complex tasks, and share moreinformation (Katz & Kahn, 1978). When they arealso open to influence from fellow team members,the team can function with respect and trust anddevelop shared leadership that in turn becomes anadditional resource for improving team process andperformance (Day et al., 2004; Marks et al., 2001).This intangible resource, which is derived from thenetwork relationships within the team, results ingreater effort, coordination, and efficiency (Na-hapiet & Ghoshal, 1998).

Only a handful of empirical studies have beenconducted with shared leadership as an explicitsource of leadership, but the results are promising(see Table 1). Avolio and colleagues (1996) ex-plored shared leadership among teams of under-graduate students and found a positive correlationwith self-reported effectiveness. Pearce and Sims

(2002) studied the relationship between sharedleadership and change management team effective-ness at a large automotive manufacturing firm andfound shared leadership to be a more useful pre-dictor than the vertical leadership of appointedteam leaders. Sivasubramaniam and colleagues(2002) found that team leadership, defined in amanner similar to previous definitions of sharedleadership as the collective influence of team mem-

bers on each other, was positively related to bothteam performance and potency over time in a sam-ple of undergraduate business students. Pearce,Yoo, and Alavi (2004) studied shared leadership invirtual teams engaged in social work projects andagain found that shared leadership was a strongerpredictor of team performance than vertical leader-ship. Ensley, Hmielski, and Pearce (2006) alsofound shared leadership to be a stronger predictorthan vertical leadership of new venture perfor-mance in a sample of top management teams.

Finally, there is also indirect support for sharedleadership predicting team performance. Taggar,Hackett, and Saha (1999) examined emergent lead-ership within teams and found that team perfor-mance was greatest when other team members, in

addition to the emergent leader, demonstrated highlevels of leadership influence. Failure of even asingle member to exhibit leadership behavior wasfound to be detrimental to team performance. Al-though shared leadership was not formally definedor measured, these findings seem to support thenotion that shared leadership may result in greatereffectiveness than the emergence of a single inter-nal team leader. Taken as a whole, these studiessuggest that shared leadership is an important pre-dictor of team performance and provides a resourcefor teams that goes beyond the leadership of any

single individual. Therefore, we predict:Hypothesis 4. The degree of shared leadershipin a team is positively related to teamperformance.

METHODS

Sample

The study sample included 59 consulting teamscomprised of MBA students (n 348) from a largeeastern university. Teams ranged in size fromfour


7/23/2019 Carson Et Al 2007

9/19

to seven members with a mean size of 5.93 indi-viduals. Sixty-seven percent of the sample mem-

bers were male, and ages ranged from 24 to 42 years(mean age 29). Participants were 56 percentwhite, 33 percent Asian, 5 percent black, and 5percent Hispanic. This sample was well suited totesting our hypotheses because the teams tasks

were highly similar and their life cycles wereidentical, thus ruling out mitigating factors oftenpresent in empirical team research. Teams wereengaged in real consulting projects and workedclosely with their clients over a five-month periodthat concluded with a significant deliverable (aclient presentation and an accompanying re-port). Thus, the likelihood that findings would

be generalizable to nonstudent populations wasstrengthened.

The universitys MBA Consulting Program officeassigned students to their teams, which were mul-tifunctional in terms of members areas of concen-tration and expertise. Each team worked on a cur-rent problem or business need for an existing firmand had a faculty advisor who served as an externalleader and also assigned grades. Each external teamleader acted much like a partner in a consultingfirm who supervises multiple projects; he or shewas available to provide general guidance and sup-port for the team in working with the clientthroughout the course of the project. Teams did nothave formally appointed internal leaders.

Data from teams were collected through surveysadministered approximately two-thirds of the way

through the projects, and data from clients andfaculty advisors were collected through surveys ad-ministered after the projects and final deliverableswere completed and presented to the clients. Wesurveyed clients to obtain independent sources ofteam performance and used the faculty surveysprimarily to gather information on the level ofproject demands faced by the teams. Surveys werereceived from 348 team members (a 100 percentresponse rate), from faculty advisors for 51 ofthe teams (an 86 percent response rate), and fromclient contacts for 56 of the teams (a 95 percent

response rate).

Measures

Team performance. Client contacts who hadworked closely with a team and were the end usersof the project results were asked to rate the effec-tiveness of each team in terms of project deliver-ables, presentation, and helpfulness of recommen-dations. Seven items began with the stem Howeffective was this team in . . . and included meet-ing your expectations in terms of the quality of the

final deliverables, providing a quality presenta-tion of the final deliverables, and overall, meet-ing your needs and goals for this project. Respon-dents rated these items on a scale ranging from 1,extremely ineffective, to 7, extremely effective.Principal components analysis yielded a single fac-tor ( .93).

Shared leadership.We measured shared leader-ship following a social network approach (Mayo,Meindl, & Pastor, 2003) by using density, which isa measure of the total amount of leadership dis-played by team members as perceived by others ona team (x 3.16, median 3.15, and range 2.403.90). Every team member rated each of his/her peers (1, not at all, to 5, to a very greatextent) on the following question: To what de-gree does your team rely on this individual forleadership? To calculate density, we followed themeasurement approach for valued relations setforth by Sparrowe and colleagues (2001) by sum-ming all values (here, the team members ratings ofeach others leadership) and then dividing that sum

by the total number of possible ties, or relation-ships, among team members. Thus, following ourdefinition of shared leadership as a team propertyreflecting the distribution of leadership amongmultiple team members, teams in which manyteam members are rating many of their peers asleaders will appropriately yield higher densityscores than will teams in which only one memberor a few members are perceived as exerting leader-ship. Agreement across the respondents ratings of

their team members was assessed and demon-strated adequate interrater reliability (medianrwg.65, ICC[1] .34, and ICC[2] .78).

To illustrate the density measure visually, wecreated leadership sociograms for each team (Mayoet al., 2003). Leadership network ratings were firstdichotomized: values of 4 (to a great extent) or 5 (toa very great extent) were assigned a value of 1, andvalues of 3 or less were assigned a 0.2 Figure 1presents the sociograms for the lowest-, middle-,and highest-scoring teams on the shared leadershipmeasure. The circles are nodes representing team

members. Arrows represent leadership relations:An arrow pointing from one member (A) to another(B) means that member B is perceived as a source ofleadership by member A. Two-headed arrows mean

2 It is important to note that the dichotomous (0, 1)data set was used onlyfor the purpose of creating thesegraphic depictions of the leadership relations within theteam. All substantive analyses in the paper used the fullyvalued ratings (15) for calculation of the density scorewhich was our measure of shared leadership.


7/23/2019 Carson Et Al 2007

10/19

that two individuals perceive one another as asource of leadership.

Internal team environment. Members ratedtheir teams internal environment using ten items(1, strongly disagree, to 5 strongly agree) con-sisting of three separate, theoretically derived sub-scales: shared purpose, social support, and voice.Voice was measured using four items based onprevious work by VanDyne and LePine (1998) andDeDreu and West (2001), and the shared purposeand social support scales were developed to specif-ically fit the context of our sample. The Appendixlists all items. To test for discriminant validity, weperformed a confirmatory factor analysis using EQS(Bentler & Wu, 2005), specifying a higher-order fac-tor with three dimensions (indicated by the tenitems) that yielded a good fit to the data (232 67.87; AIC 3.87; CFI .98; GFI .96; AGFI .94; SRMR .04; RMSEA .06), thus demonstrat-ing support for the hypothesized structure.3, 4 Toinvestigate the convergent validity of the structure

for internal team environment, we examined thecorrelations among the three subscales. The zero-order correlations were high, ranging from .72 to.80 (p .001), which provided evidence that the

three subscales represented highly interrelated di-mensions. Given the overall support for the hy-pothesized model, we first aggregated these threesubscales to the team level and then averaged thescores to produce a single variable representinginternal team environment ( .94). We tested forwhether aggregation was appropriate using the

within-team agreement statistic (rwg; James, Dema-ree, & Wolf, 1993) and used intraclass correlationcoefficients (ICC[1] and ICC[2]) to assess the extentto which team responses differed among teams andthe reliability of the team-level means (Bliese,2000). The meanrwgof .96 indicated a high level ofagreement among members of teams on rating theirinternal team environment; the ICC(1) of .14 dem-onstrated that team membership accounted for sig-nificant variance; and the ICC(2) of .71 suggestedthat the team-level means were reliable.

Coaching.Team members were asked to rate thelevel of supportive coaching (Morgeson, 2005) pro-vided by their external leader (faculty advisor) us-ing a three-item scale. Items included expresseshis/her confidence in the capabilities of our team,effectively motivates and guides our team towardaccomplishing challenging goals for this project,and is sensitive to the needs of our team and triesto help us however he/she can. These items cap-tured the motivational and consultative functionsof external leaders that have been suggested asparticularly important for fostering both commit-ment to a team and independence (Hackman &Wageman, 2005; Morgeson, 2005). Items were rated

on a scale ranging from 1 (strongly disagree) to5 (strongly agree) ( .92), and responses wereaggregated to the team level and demonstratedstrong levels of within-team agreement, between-team differences, and reliable team-level means(meanrwg .83; ICC[1] .51; ICC[2] .80).

Control variables. We included controls for theeffects of team size, project demands, gender diver-sity, and race diversity in order to address thesepossible alternate explanations for shared leader-ship and team performance. Differences in teamsize may influence resources and workload require-

ments that may influence team performance andtherefore team size was included as a control vari-able (Kirkman & Rosen, 1999). Environmental fac-tors can impact team outcomes (e.g., Tesluk &Mathieu, 1999), and more demanding projects maythus have a detrimental impact on shared leader-ship and team performance. Previous research hasalso shown significant effects for demographic het-erogeneity on team outcomes (Williams & OReilly,1998), and we therefore included it as a control.Team size was measured as the actual number ofteam members on each consulting team. Project

3 We also examined the fit statistics for a one-factormodel, which was not a very good fit to the data. Achi-square difference test also indicated that our theoret-ical model was a significantly better fit.

4 AIC is the Akaike Information Criterion and iscalculated as 2 2df(Bentler & Wu, 2005).

FIGURE 1Leadership Sociograms


7/23/2019 Carson Et Al 2007

11/19

demands was measured by a seven-item Likertscale that captured faculty advisor ratings of thedegree to which the team had to manage difficult

project challenges. Sample items include chang-ing client demands during the course of theproject, difficulties in accessing data or informa-tion necessary for completing the work, andproblems or changes in the project timeline thatwere outside the teams direct control. Items wererated on a scale ranging from 1 (not at all present)to 7 (very much present) ( .75).Gender diver-sityandrace diversitywere measured using Teach-mans index,5 which captures how team membersare distributed among the possible categories of avariable (Teachman, 1980).

RESULTS

Table 2 gives means, standard deviations, andzero-order correlations. To test Hypotheses 1through 3, we used moderated regression analysis.In step 1, we entered all of the control variables. Instep 2, we entered the two main effect variables,internal team environment and coaching. In step 3,we tested for interactions by entering the product ofinternal team environment and coaching. Table 3presents these results. Team size was the only con-trol variable with a significant relationship withshared leadership ( 0.37, p .01). In step 2,internal team environment had a direct relation-ship with shared leadership ( 0.25, p .05,

one-tailed), as did external coaching ( 0.26,p .05, one-tailed), supporting Hypotheses 1 and 2. Instep 3, we found the interaction effect between

coaching and internal team environment to be sig-nificant ( 4.06, p .05) and explaining anadditional 5 percent of the variance. FollowingAiken and Wests (1991) methods for plotting in-teractions, we graphed these relationships; Figure 2presents this graph. It shows that internal teamenvironment was significantly and positively re-lated to shared leadership for teams that had low

5 Although the current tendency is to use Blaus indexfor diversity, the only difference between Blau andTeachman is standardization, and there is no conceptualor empirical reason to favor one over the other. We aregrateful to David Harrison for noting this point in aposting to the Academy of Management Research Meth-ods Division listserv group (RMNet).

TABLE 2Descriptive Statistics and Correlationsa

Variable Mean s.d. 1 2 3 4 5 6 7

1. Team performanceb 5.81 1.02 (.93)2. Shared leadership 3.16 0.35 .46*3. Internal team environment 4.08 0.41 .19 .33* (.94)4. Coaching 3.76 0.64 .03 .37* .24 (.92)5. Team Size 5.93 0.72 .10 .28* .42* .156. Project demandsb 4.38 1.19 .26 .20 .22 .22 .04 (.75)7. Gender diversity 0.55 0.18 .20 .22 .14 .15 .00 .058. Race diversity 0.74 0.27 .03 .03 .02 .24 .02 .14 .05

a n 59 for most variables (n 51 for variable 6 and 56 for variable 1, because of missing data). Scale reliabilities are in parenthesesalong the diagonal.

b These variables were measured using seven-point Likert scales. All other scales were measured using five-point Likert scales.*p .05

TABLE 3Results of Regression Analyses

VariableShared

LeadershipaTeam

Performanceb

Step 1Team size 0.37* 0.11Project demands 0.22 0.25Gender diversity 0.24 0.21Race diversity 0.04 0.02

R2 0.23* 0.12

Step 2Internal team environment 0.25* 0.25Coaching 0.26* 0.14

R2 0.39* 0.16

R2

0.16* 0.04Step 3

Internal team environment coaching

4.06*

Shared leadership 0.65*

R2 .44* .42*R2 .05* .26*

Adjusted R2 .32

a Estimates are standardized regression coefficients; n 51.b Estimates are standardized regression coefficients; n 49.

*p .05


7/23/2019 Carson Et Al 2007

12/19

coaching support and was not related to sharedleadership for teams that had high coaching sup-port. Teams with an unsupportive internal teamenvironment were still able to develop high levelsof shared leadership, so long as they received ahigh level of coaching. Thus, Hypothesis 3 wasalso supported.

To test Hypothesis 4, we used hierarchical re-gression, regressing team performance on shared

leadership after controlling for team size, projectdemands, gender diversity, and race diversity, aswell as the main effects of internal team environ-ment and coaching, to determine whether therewas significant additional explained variance. Theresults, presented in Table 3, indicate that sharedleadership is a strongly positive predictor of ateams performance as rated by the end users of theteams work ( 0.65, p .001) and accounts forsignificant variance in team performance exceedingthat accounted for by the control variables, internalteam environment, and coaching (R2 .26, p

.001). Thus, Hypothesis 4 received strong support.

DISCUSSION

Our study makes three key contributions to theliterature on team leadership. First, we examinedantecedent conditions for shared leadership andfound that a teams internal environment andcoaching by an external leader are important pre-cursors for shared leadership. Second, our findingsshow that coaching provided by an external teamleader is particularly important for the develop-

ment of shared leadership when teams lack a stronginternal team environment. Third, the findings ex-tend previous research suggesting positive effectsof shared leadership on team performance using anetwork-based measure of shared leadership that

better captures the patterns of mutual influenceinherent in the construct and a measure of perfor-mance that is less subject to common source vari-ance and rating biases.

Theoretical Implications

Despite early suggestions by scholars that sharedinternal leadership is important (e.g., Gibb, 1954;Katz & Kahn, 1978), team leadership theory hascontinued to focus primarily on the role of externalleaders and to use existing models of dyadic lead-ership extrapolated to the team level (Kozlowski &Bell, 2003). Our findings suggest that a promisingfuture direction is to move the lens inward to in-vestigate how team members themselves share

leadership responsibilities. Indeed, we found thatteams relying on multiple members for leadershipperformed better than those in which internal lead-ership was relatively scarce. Importantly, this find-ing is based on performance ratings provided byclients who focused on the quality of a teams finalproduct rather than on its process and functioning.This argument suggests that shared leadership has

benefits for work teams beyond just improved teamprocesses. Shared leadership can occur in a teamwith a designated formal leader or in a team with-out one. The results of this study do not mean that

FIGURE 2The Moderating Effect of Coaching on the Relationship between

Internal Team Environment and Shared Leadership


7/23/2019 Carson Et Al 2007

13/19

vertical leadership needs to be abandoned in favorof shared leadership; rather, these two importantsources of team leadership should be studied incombination (Kozlowski & Bell, 2003).

Building on previous work that has advocatedthe use of social network methodology to under-stand relationships in teams (Mayo et al., 2003;

Sparrowe et al., 2001), we advocate a network ap-proach to conceptualizing and measuring sharedleadership as an important step forward. Ratherthan capturing a teams overall central tendency bytaking the mean of a Likert scale, this approachincorporates the pattern of leadership presentthroughout the team. By examining all possiblerelationships in the team, this density measure cap-tures the degree to which the team as a whole reliesheavily on most of its members for leadership. Itthus allows for a closer approximation to the theo-retically rich concept of shared leadership.

This study also presents an initial understandingof the antecedent conditions, both internal and ex-ternal to teams, that enable shared leadership todevelop. We found that when a team has an inter-nal environment characterized by a clear and uni-fying direction that is well understood within theteam, a strong sense of interpersonal supportwhereby team members feel recognized and en-couraged, and a high level of voice and involve-ment within the team, it is able to develop a lead-ership network characterized by high levels ofmutual influence and sharing of leadership respon-sibilities. Our findings also demonstrate the impor-

tance of coaching by an external leader for support-ing the emergence of shared leadership, as well aswhen this coaching support is most necessary(Hackman & Wageman, 2005). When an internalteam environment is supportive, coaching by theexternal leader is less critical for the emergence ofshared leadership; however, when an internal teamenvironment is unsupportive, coaching interven-tions are important for filling a role that is not beingfilled by the team (cf. Hackman & Walton, 1986).

Managerial Implications

This study has important implications for teamleaders and managers. First, the findings suggestthat organizations should help develop strong in-ternal leadership patterns within their teams to

bolster effectiveness. Organizations can promoteinternal leadership by setting expectations and en-couraging members when teams initially form toview themselves and their fellow team members asleaders and to engage in shared, mutual leadership.Organizations can also provide training that fostersa shared leadership perspective and disseminates

best practices. Second, our results point to spe-cific dimensions of internal team environmentsshared purpose, social support, and voicethatsupport the development of shared leadership inteams. Managers should therefore ensure that eachteam has a clear and shared sense of direction andpurpose, promote and establish norms of participa-

tion and input into the teams activities and strat-egies, and seek to foster a positive environmentwhere team members encourage one another andactively recognize each others contributions. Or-ganizations may further support these conditions

by institutionalizing a team charter processwhereby teams, upon their formation, collectivelyidentify and agree upon a common goal and set ofpriorities, team roles and responsibilities, and teamnorms. Third, our findings suggest that externalleaders should engage in supportive coaching ofteams to facilitate the development of shared lead-ership. This coaching can be in the form of encour-aging, reinforcing, and rewarding instances inwhich team members demonstrate leadership, as-sisting teams when internal conflicts arise (e.g.,over sharing leadership responsibilities), providinggeneral encouragement to a team as a whole, and

being available for suggestions or input into theteams task strategies as needed (Hackman & Wage-man, 2005). Team leaders should pay particularattention to teams that may have weaker internalenvironments in order to provide additional moti-vation, guidance, and support. However, for teamswith supportive internal environments, stronger

coaching may not provide much additional assis-tance in developing shared leadership.

Limitations and Future Research

Our study has several limitations that need to beaddressed in future research. First, the partiallycross-sectional design did not allow for testing cau-sality. Although we did measure shared leadershipafter it had been given time to develop and col-lected outcome data from third parties after teamprojects were complete, shared leadership is an

emergent phenomenon, and longitudinal designsare needed to understand how shared leadershipdevelops over time by looking at changes in a lead-ership network over stages of team development.Second, although the teams we studied performedreal consulting engagements and were responsibleto their clients for delivering completed projects,team members were MBA students, not full-timeemployees. For team members who are full-timeemployees working in different organizational set-tings, shared leadership may operate differently.Third, common source variance may have influ-


7/23/2019 Carson Et Al 2007

14/19

enced the relationship between internal team envi-ronment and shared leadership, since both mea-sures were taken from team members. However, itis important to note that internal team environmentwas a perceptual measure of an entire teams be-havior and actions, whereas shared leadership wasa network measure compiled from ratings of each

individual team member. This distinction helpsmitigate the likelihood that common source biasinfluenced the relationship.

Additionally, both the strengths and weaknessesof our measurement approach for shared leadershipshould be highlighted. Our measurement of sharedleadership as network density represents an im-provement in this research field, notably by captur-ing the overall patterns of shared influence withinteams and overcoming a primary limitation of be-havioral scales that restrict influence to a set ofprescribed behaviors. Operationally, by broadlyasking respondents the extent to which team mem-

bers exerted leadership rather than detailing spe-cific leadership behaviors, our measure of sharedleadership captured the respondents personal andimplicit theories of leadership, and was consistentwith the approach others have used in similar con-texts (e.g., Mehra et al., 2006). However, the notablelimitations to such an approach should also behighlighted. Because it neither specified the mean-ing of leadership nor primed specific behaviors forrespondents, it is possible that our measure tappedsomething other than leadership influence, such asparticipation and engagement, helping and cooper-

ation, or respect and listening among team mem-bers. Thus, future research along these lines shouldinclude efforts to provide leadership definitionsand/or behavioral examples to minimize the influ-ence of differences in respondents attributions.Further, a richer conceptualization and operation-alization might be developed that, in addition toidentifying leadership sources, captures the qualityand nature of leadership offered by each teammember.

Future work should focus on a more detailedunderstanding of the nature of shared leadership,

its development, and boundary conditions on itseffectiveness. It seems clear that relying on manyteam members for leadership can be an effectiveapproach to team leadership, yet there are manyleadership styles that team members might employ,such as directive, transactional, transformational,and empowering leadership, as well as varyingleadership roles that team members may adopt.Future research is needed to address both howdifferent leadership styles and roles interrelate andcomplement one another when they are shared inteams and the relationship between shared leader-

ship and external leadership beyond the effects ofcoaching.

Additional predictors of shared leadership devel-opment should also be examined, such as teamempowerment (e.g., Kirkman & Rosen, 1997), teamcomposition (e.g., Cohen & Bailey, 1997), and con-textual factors (e.g., Gladstein, 1984). Teams that

are highly empowered should be more likely todevelop shared leadership as a result of the auton-omy and meaningfulness of the work they are doingas well as strong collective beliefs about their po-tency and impact (Kirkman & Rosen, 1997). Thecomposition of a team in terms of its size andcharacteristics such as experience, expertise, andpersonality, as well as demographic compositionalpatterns such as the existence of faultlines (Lau &Murnighan, 1998), may also play a role in the de-velopment of shared leadership. Finally, in addi-tion to external coaching, other contextual factorsmay also impact shared leadership, such as rewardand recognition systems, training in importantteamwork and leadership skills, and the nature of ateams task itself (e.g., Hackman, 1987).

Potential mediating mechanisms linking sharedleadership to performance and other effectivenesscriteria should be explored as well. For instance,the reciprocal interactions and influential ex-changes between team members may facilitate de-velopment of team knowledge structures such astransactive memory systems or shared mental mod-els. In addition to serving as a potential antecedent,team empowerment may be another mediating

mechanism through which leadership becomesshared among team members (Kirkman & Rosen,1997).

A number of important boundary conditions forthe effectiveness of shared leadership should beexamined, such as the distribution of task compe-tence, task interdependence, task complexity, teamlife cycle, and cultural values (Pearce & Conger,2003). Shared leadership is likely to be more effec-tive when team members have a high level of taskcompetence, when a task is relatively complex,when task interdependence is high, and when the

team life cycle allows for the development ofshared leadership. It is also likely to be affected bycultural values, particularly power distance andcollectivism (Gibson & Zellmer-Bruhn, 2001; Kirk-man & Shapiro, 1997), and it may be more likely todevelop and thrive in cultures that are low inpower distance and in cultures that are high incollectivism (Carson, 2005).

Finally, team size was found to have a stronglypositive relationship with shared leadership. Ourinterpretation of this finding is that teams withmore members have greater potential leadership


7/23/2019 Carson Et Al 2007

15/19

resources available for sharing. However, there maybe a nonlinear relationship between team size andshared leadership in teams larger than those in ourstudy. Beyond a certain team size, either a detri-mental or ceiling effect on shared leadership mayoccur, perhaps as the result of social loafing or freeriding. Future research should explore these non-

linear possibilities further by sampling teams withgreater range on team size.

Conclusion

As organizations continue to devote vast re-sources to the use of teams and teamwork, the needfor a better understanding of effective team leader-ship continues to grow. This study provides animportant contribution by highlighting the impor-tance of leadership input from multiple team mem-

bers and suggests that shared leadership is a critical

factor that can improve team performance from theviewpoint of customers or end users of a teamswork. Although not a final statement on the topic,this study adds to the growing body of evidencethat a team does well when it relies on leadershipprovided by the team as a whole rather than look-ing to a single individual to lead it.

REFERENCES

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

Park, CA: Sage.Avolio, B. J., Jung, D. I., & Sivasubramaniam, N. 1996.

Building highly developed teams: Focusing onshared leadership processes, efficacy, trust, and per-formance. In M. M. Beyerlein & D. A. Johnson (Eds.),

Advances in interdisciplinary study of workteams: Team leadership, vol. 3: 173209. Green-wich, CT: JAI Press.

Bales, R. F. 1953. The equilibrium problem in smallgroups. In T. Parsons, R. F. Bales, & E. A. Shils (Eds.),Working papers in the theory of action: 111161.Glencoe, IL: Free Press.

Barnard, C. I. 1938. The functions of the executive.

Cambridge, MA: Harvard University Press.

Barry, D. 1991. Managing the bossless team: Lessons indistributed leadership. Organizational Dynamics,20(1): 3147.

Bedeian, A. G. & Hunt, J. G. 2006. Academic amnesia andvestigial assumptions of our forefathers. LeadershipQuarterly,17: 190205.

Bentler, P. M., & Wu, E. J. C. 2005.EQS 6.1 for Windows.Encino, CA: Multivariate Software.

Bliese, P. D. 2000. Within-group agreement, non-inde-pendence, and reliability: Implications for data ag-

gregration and analysis. In K. J. Klein & S. W. J.Kozlowski (Eds.), Multilevel theory, research, and

methods in organizations:349381. San Francisco:Jossey-Bass.

Brass, D. J. 1995. A social network perspective on humanresources management.Research in Personnel and

Human Resources Management, 13: 3979.

Cannon-Bowers, J. A., Salas, E., & Converse, S. A. 1993.Shared mental models in expert team decision mak-ing. In N. J. Castellan (Ed.), Current issues for fieldsettings: 221246. Hillsdale, NJ: Erlbaum.

Carson, J. 2005. Shared leadership and culture: Poten-tial emergence and global application. In N. S. Huber& M. C. Walker (Eds.), Emergent models of glo-bal leadership: A volume in building leadershipbridges: 116. College Park, MD: The James Mac-Gregor Burns Academy of Leadership.

Cohen, S. G., & Bailey, D. E. 1997. What makes teamswork: Group effectiveness research from the shopfloor to the executive suite. Journal of Management,

23: 239290.

Day, D. V., Gronn, P., & Salas, E. 2004. Leadership capac-ity in teams. Leadership Quarterly, 15: 857880.

De Dreu, C. K. W., & West, M. A. 2001. Minority dissentand team innovation: The importance of participa-tion in decision making. Journal of Applied Psy-chology,86: 11911201.

DeNisi, A. S., Hitt, M. A., & Jackson, S. E. 2003. Theknowledge-based approach to sustainable competi-tive advantage. In S. E. Jackson, M. A. Hitt, & A. S.DeNisi (Eds.), Managing knowledge for sustainedcompetitive advantage: 333. San Francisco: Jos-

sey-Bass.Druskat, V. U., & Wheeler, J. V. 2003. Managing from the

boundary: The effective leadership of self-managingwork teams.Academy of Management Journal, 46:435457.

Edwards, J. R. 2001. Multidimensional constructs in or-ganizational behavior research: An integrative ana-lytical framework.Organizational Research Meth-ods,4(2): 144192.

Ensley, M. D., Hmieleski, K. M., & Pearce, C. L. 2006. Theimportance of vertical and shared leadership withinnew venture top management teams: Implicationsfor the performance of startups. Leadership Quar-terly,17: 217231.

Gibb, C. A. 1954. Leadership. In G. Lindzey (Ed.), Hand-book of social psychology, vol. 2: 877917. Reading,MA: Addison-Wesley.

Gibson, C. B., & Zellmer-Bruhn, M. E. 2001. Metaphorsand meaning: An intercultural analysis of the con-cept of teamwork. Administrative Science Quar-terly,46: 274303.

Gladstein, D. L. 1984. Groups in context: A model of taskgroup effectiveness. Administrative Science Quar-terly,29: 499517.


7/23/2019 Carson Et Al 2007

16/19

Gronn, P. 2002. Distributed leadership as a unit of anal-ysis. Leadership Quarterly, 13: 423451.

Hackman, J. R. 1987. The design of work teams. In J. W.Lorsch (Ed.), Handbook of organizational behav-

ior: 315342. Englewood Cliffs, NJ: Prentice Hall.

Hackman, J. R., & Wageman, R. 2005. A theory of teamcoaching. Academy of Management Review, 30:

269287.Hackman, J. R., & Walton, R. E. 1986. Leading groups in

organizations. In P. S. Goodman & associates (Eds.),Designing effective work groups: 72119. SanFrancisco: Jossey-Bass.

James, L. R., Demaree, R. G., & Wolf, G. 1993. rwg: Anassessment of within-group interrater agreement.

Journal of Applied Psychology,78: 306309.

Katz, D., & Kahn, R. L. 1978. The social psychology oforganizations(2nd ed.). New York: Wiley.

Kirkman, B. L., & Rosen, B. 1997. A model of work teamempowerment. In W. A. Pasmore & R. W. Woodman

(Eds.),Research in organizational change and de-velopment: An annual series featuring advances

in theory, methodology, and research,vol. 10: 131167. Greenwich, CT: JAI Press.

Kirkman, B. L., & Rosen, B. 1999. Beyond self-manage-ment: Antecedents and consequences of team em-powerment.Academy of Management Journal, 42:5874.

Kirkman, B. L., & Shapiro, D. L. 1997. The impact ofcultural values on employee resistance to teams: To-ward a model of globalized self-managing work teameffectiveness. Academy of Management Review,22: 730757.

Klimoski, R., & Mohammed, S. 1994. Team mental mod-el: Construct or metaphor?Journal of Management,20: 403437.

Kozlowski, S. W. J., & Bell, B. S. 2003. Work groups andteams in organizations. In W. C. Borman & D. R. Ilgen(Eds.), Comprehensive handbook of psychology:

Industrial and organizational psychology,vol. 12:333375. New York: Wiley.

Kozlowski, S. W. J., Gully, S. M., Salas, E., & Cannon-Bowers, J. A. 1996. Team leadership and develop-ment: Theory, principles, and guidelines for trainingleaders and teams. In M. M. Beyerlein, D. A. John-

son, et al. (Eds.), Advances in interdisciplinarystudy of work teams: Team leadership,vol. 3: 253292. Greenwich, CT: JAI Press.

Kozlowski, S. W. J., & Klein, K. J. 2000. A multilevelapproach to theory and research in organizations. InK. J. Klein & S. W. J. Kozlowski (Eds.), Multileveltheory, research, and methods in organizations:390. San Francisco: Jossey-Bass.

Langfred, C. W. 2004. Too much of a good thing? Nega-tive effects of high trust and individual autonomy inself-managing teams. Academy of Management

Journal,47: 385399.

Lau, D. C., & Murnighan, J. K. 1998. Demographic diver-sity and faultlines: The compositional dynamics oforganizational groups. Academy of Management

Review,23: 325341.

Law, K. S., Wong, C. S., & Mobley, W. H. 1998. Toward ataxonomy of multidimensional constructs. Acad-emy of Management Review, 23: 741755.

Lawler, E. E., Mohrman, S. A., & Benson, G. 2001.Orga-nizing for high performance: Employee involve-ment, TQM, reengineering, and knowledge man-agement in the Fortune 1000. San Francisco: Jossey-Bass.

Lewis, K. 2003. Measuring transactive memory systemsin the field: Scale development and validation. Jour-

nal of Applied Psychology,88: 587604.

Liden, R. C., Wayne, S. J., & Sparrowe, R. T. 2000. Anexamination of the mediating role of psychologicalempowerment on the relations between the job, in-terpersonal relationships, and work outcomes. Jour-

nal of Applied Psychology,85: 407416.

Manz, C. C. & Sims, H. P., Jr. 1987. Leading workers tolead themselves: The external leadership of self-managing work teams. Administrative ScienceQuarterly,32: 106130.

Marks, M. A., Mathieu, J. E., & Zaccaro, S. J. 2001. Atemporally based framework and taxonomy of teamprocesses. Academy of Management Review, 26:356376.

Marks, M. A., Zaccaro, S. J., & Mathieu, J. E. 2000. Per-formance implications of leader briefings and team-interaction training for team adaptation to novel en-vironments. Journal of Applied Psychology, 85:971986.

Mayo, M., Meindl, J. R., & Pastor, J. C. 2003. Sharedleadership in work teams: A social network ap-proach. In C. L. Pearce & J. A. Conger (Eds.),Shared

leadership: Reframing the hows and whys of lead-ership:193214. Thousand Oaks, CA: Sage.

Mehra, A., Smith, B., Dixon, A., & Robertson, B. 2006.Distributed leadership in teams: The network ofleadership perceptions and team performance.

Leadership Quarterly,17: 232245.

Morgeson, F. P. 2005. The external leadership of self-managing teams: Intervening in the context of novel

and disruptive events. Journal of Applied Psychol-ogy,90: 497508.

Morgeson, F. P., & Hofmann, D. A. 1999. The structureand function of collective constructs: Implicationsfor multilevel research and theory development.

Academy of Management Review,24: 249265.

Nahapiet, J., & Ghoshal, S. 1998. Social capital, intellec-tual capital, and the organizational advantage.Acad-emy of Management Review, 23: 242266.

OLeary-Kelly, A. M., Martocchio, J. J., & Frink, D. D.1994. A review of the influence of group goals on


7/23/2019 Carson Et Al 2007

17/19

group performance.Academy of Management Jour-nal,37: 12851301.

Pearce, C. L., & Conger, J. A. 2003. Shared leadership:Reframing the hows and whys of leadership.Thousand Oaks, CA: Sage.

Pearce, C. L., & Sims, H. P. 2002. The relative influence ofvertical vs. shared leadership on the longitudinal

effectiveness of change management teams. GroupDynamics: Theory, Research, and Practice, 6(2):172197.

Pearce, C. L., Yoo, Y., & Alavi, M. 2004. Leadership,social work and virtual teams: The relative influenceof vertical vs. shared leadership in the nonprofitsector. In R. E. Riggio & S. Smith-Orr (Eds.), Im-

proving leadership in nonprofit organizations:180203. San Francisco: Jossey-Bass.

Schneier, C. E., & Goktepe, J. R. 1983. Issues in emergentleadership: The contingency model of leadership,leader sex, leader behavior. In H. H. Blumberg, A. P.Hare, V. Kent, & M. F. Davies (Eds.), Small groups

and social interactions, vol. 1: 413421. Chichester,U.K.: Wiley.

Seers, A. 1996. Better leadership through chemistry: To-ward a model of emergent shared team leadership. InM. M. Beyerlein & D. A. Johnson (Eds.),Advances inthe interdisciplinary study of work teams: Team

leadership, vol. 3: 145172. Greenwich, CT: JAIPress.

Simons, T. L., Pelled, L. H., & Smith, K. A. 1999. Makinguse of difference: Diversity, debate, and decisioncomprehensiveness in top management teams.

Academy of Management Journal,42: 662673.

Sinclair, A. L. 1992. The tyranny of a team ideology.Organization Studies, 13: 611626.

Sivasubramaniam, N., Murry, W. D., Avolio, B. J., & Jung,D. I. 2002. A longitudinal model of the effects ofteam leadership and group potency on group perfor-mance. Group & Organization Management, 27:6696.

Sparrowe, R. T., Liden, R. C., Wayne, S. J., & Kraimer,M. L. 2001. Social networks and the performance ofindividuals and groups. Academy of Management

Journal,44: 316325.

Stewart, G. L., & Manz, C. C. 1995. Leadership for self-managing work teams: A typology and integrativemodel.Human Relations,48: 747770.

Taggar, S., Hackett, R., & Saha, S. 1999. Leadership emer-gence in autonomous work teams: Antecedents andoutcomes.Personnel Psychology, 52: 899926.

Teachman, J. D. 1980. Analysis of population diversity.Sociological Methods and Research, 8: 341362.

Tesluk, P. E., & Mathieu, J. E. 1999. Overcoming road-blocks to effectiveness: Incorporating managementof performance barriers into models of work groupeffectiveness. Journal of Applied Psychology, 84:200217.

Van Dyne, L., & LePine, J. A. 1998. Helping and voiceextra-role behaviors: Evidence of construct and pre-dictive validity.Academy of Management Journal,41: 108119.

Wageman, R. 2001. How leaders foster self-managingteam effectiveness: Design choices versus hands-oncoaching.Organization Science, 12: 559577.

Wasserman, S., & Faust, K. 1994. Social network anal-ysis: Methods and applications. Cambridge, U.K.:Cambridge University Press.

Wegner, D. M. 1986. Transactive memory: A contempo-rary analysis of the group mind. In B. Mullen & G. R.Goethals (Eds.), Theories of group behavior: 185208. New York: Springer-Verlag.

Wheelan, S. A., & Johnston, F. 1996. The role of informalmember leaders in a system containing formal lead-ers. Small Group Research, 27: 3355.

Williams, K. Y., & OReilly, C. A. 1998. Demography anddiversity in organizations: A review of 40 years of

research. In B. M. Staw & R. Sutton (Eds.), Researchin organizational behavior,vol. 20: 77141. Green-wich, CT: JAI Press.

Yukl, G. A. 1989.Leadership in organizations(2nd ed.).Englewood Cliffs, NJ: Prentice Hall.

Zaccaro, S. J., & Klimoski, R. 2002. Special issue intro-duction: The interface of leadership and team pro-cesses. Group and Organization Management, 27:413.

Zaccaro, S. J., Rittman, A. L., & Marks, M. A. 2001. Teamleadership.Leadership Quarterly, 12: 451483.

Ziegert, J. C. 2005. Does more than one cook spoil the

broth? An examination of shared team leadership.Unpublished doctoral dissertation, University ofMaryland, College Park.

APPENDIX

Items Assessing Internal Team Environment forShared Leadership

Shared PurposeThe members of my team . . .1. Spent time discussing our teams purpose, goals,

and expectations for the project.2. Discuss our teams main tasks and objectives to

ensure that we have a fair understanding.3. Devise action plans and time schedules that allow

for meeting our teams goals.

Social SupportThe members of my team . . .4. Talk enthusiastically about our teams progress.5. Recognize each others accomplishments and

hard work.6. Give encouragement to team members who seem

frustrated.


7/23/2019 Carson Et Al 2007

18/19

Voice7. People in this team are encouraged to speak up to

test assumptions about issues under discussion.8. As a member of this team, I have a real say in how

this team carries out its work.

9. Everyone on this team has a chance to participateand provide input.

10. My team supports everyone actively participatingin decision making.

Jay B. Carson ([email protected]) is an assistantprofessor at the Edwin L. Cox School of Business,Southern Methodist University. He received his Ph.D.from the Robert H. Smith School of Business, Univer-sity of Maryland. His research interests are in leader-ship, teams, and cross-cultural issues, with a currentfocus on shared leadership and internal leadership inteams.

Paul E. Tesluk ([email protected]) is an as-sociate professor at the Robert H. Smith School ofBusiness at the University of Maryland, College Park.His research fo-cuses on team effectiveness, leadership

development, and innovation processes in organiza-tions. He received his Ph.D. from Pennsylvania StateUniversity.

Jennifer A. Marrone ([email protected]) is anassistant professor at the Albers School of Businessand Economics, Seattle University. She received herPh.D. from the Robert H. Smith School of Business,University of Maryland. Her research interests in-clude team processes and performance, leadership,and strategic human resource management, with anemphasis on applying multilevel perspectives andtechniques.


7/23/2019 Carson Et Al 2007

19/19

carson et al 2007

Documents