OF SHARED LEADERSHIP AND TEAM EFFECTIVENESS

DARE TO SHARE: THE MEDIATING EFFECTS OF ROLE CLARITY AND TEAM EFFICACY ON THE RELATIONSHIP BETWEEN THE DISTRIBUTION

OF SHARED LEADERSHIP AND TEAM EFFECTIVENESS

Master thesis, MSc Human Resource Management University of Groningen, Faculty of Economics and Business

June 14th, 2015

SABINE SILBEEK Studentnumber: 2013835

Parkweg 33A 9725 EB Groningen Tel: + 31 (6) 30894089 e-mail: S.L.Silbeek@student.rug.nl

Supervisor David de Geest

Co-assessor J. Van Polen

Acknowledgement: I would like to express my gratitude to my supervisor David de Geest, for

his guidance and valuable feedback during this period of writing my master thesis.

DARE TO SHARE: THE MEDIATING ROLE OF ROLE CLARITY AND TEAM EFFICACY ON THE RELATIONSHIP BETWEEN THE DISTRIBUTION

OF SHARED LEADERSHIP AND TEAM EFFECTIVENESS

ABSTRACT

A growing number of studies have investigated shared leadership in teams and its relationship with team performance or effectiveness. The current research addresses how shared leadership is most effective, i.e. how different distributions of leadership functions impact team effectiveness. The effects of different distributions of leadership functions and its effect on role clarity, team efficacy and team effectiveness are discussed. Seven hypotheses were tested in an experiment with 216 undergraduate students in the Netherlands. The results neither provide support for the positive effect of high distribution of shared leadership functions on team effectiveness, nor for the mediating effect of role clarity and team efficacy.

However, evidence was found for the positive effect of role clarity on team efficacy and the positive effects of both role clarity and team efficacy on team effectiveness. Possible explanations for the non-significant results, as well as directions for future research and theoretical and practical implications are discussed.

Keywords: Shared leadership, distribution, leadership functions, role clarity, team efficacy,

team effectiveness, student teams.

INTRODUCTION

In recent years, many companies introduced self-managing teams to increase productivity, job and customer satisfaction and organizational commitment. As Druskat and Wheeler (2004) have shown, 79% of Fortune 1000 companies and 81% of Fortune 1000 manufacturing companies are using self-managing teams. This increased use of self-managing and empowered teams (Druskat & Wheeler, 2004; Manz & Sims, 1993; Park, 2012) implies the need to study shared leadership compared to the more traditional, vertical leadership structures (Pearce and Conger, 2003; Wang, Waldman & Zhang, 2014). Shared leadership occurs when two or more members engage in the leadership of the team in an effort to influence and direct fellow members to maximize team effectiveness (Bergman, Rentsch, Small, Davenport & Bergman, 2012). These simultaneous, ongoing and mutual influential processes (Pearce, 2004) involve peer, lateral, upward or downward influences of team members (Conger & Pearce, 2003).

Three recent meta-analyses that were published on shared leadership have shown a positive relationship between shared leadership and team effectiveness (D’Innocenzo, Mathieu & Kukenberger, 2014; Nicolaides, LaPort, Chen, Tomassetti, & Weis, 2014; Wang et al., 2014). Similarly, an abundance of research has shown that shared leadership is positively related to team performance (e.g. Carson, Tesluk & Marrone, 2007; Hoch, 2014;

Pearce & Conger, 2003; Pearce & Sims, 2002), by increasing team coordination, group potency, team confidence, information sharing, intragroup trust and cohesion (e.g. Bergman et al., 2012; Hoch, 2014; Mathieu, Heffner, Goodwin, Salas, & Cannon-Bowers, 2000;

Nicolaides et al., 2014). However, evidence for shared leadership’s positive effect on team

effectiveness and performance is mixed, since some studies have not supported this

relationship and found a negative relationship between shared leadership and team

performance (Boies, Lvina, & Martens, 2010; Fausing, Jeppesen, Jønsson, Lewandowski, &

Bligh, 2013; Friedrich, Vessey, Schuelke, Ruark & Mumford, 2009). For instance, Boies et al.

(2010) found that shared leadership identified using transformational leadership had negative effects on team performance. Moreover, Fausing et al. (2013) demonstrated that work function significantly moderates the relation between shared leadership and team performance such that shared leadership exhibited a negative relationship with manufacturing team performance.

Despite the increased attention and research on shared leadership, there are a number

of questions that remain unanswered. First of all, previous research did not examine how

shared leadership is most effective (Yukl, 2013). For instance, should team members share all

the leadership functions without an explicit distribution of these functions? Or should all the

functions be explicitly divided such that one team member emphasizes team self-

management, while another focuses on providing resources, while yet another takes primary

responsibility for solving problems (Morgeson, DeRue & Karam, 2010)? Moreover, a recent

study of Harris, Li, Boswell, Zhang & Xie (2014) have shown positive links between

empowering leadership and role clarity. Previous research has not investigated how the

different distributions of shared leadership impact role clarity, which can be described as the

degree to which required information is provided about how someone is expected to perform

his/her job (Teas, Wacker & Hughes, 1979). Finally, Pearce & Conger (2003) argue that

members of shared leadership teams should have confidence in the ability of each team

member, assuming the leadership role at any given time. Without this belief in fellow team

member’s abilities, members may become reluctant to follow guidance of direction given by

that team member. This emphasizes the importance of collective efficacy, which is the group

member’s shared perspective about how capable their group is regarding a specific task

(Bandura, 1997). Previous research has not investigated the mediating role of team efficacy

on the relationship between the distribution of shared leadership and team effectiveness.

The main purpose of this research is to determine to what extent shared leadership functions have to be explicitly distributed among team members. The direct effect of different distributions of these leadership functions on team effectiveness and the indirect effect via role clarity and team efficacy will be explored. Figure 1 shows the conceptual model of this study:

Insert Figure 1 about here

In addressing the aforementioned gaps, this study makes the following contributions to the literature: First, in contrast to prior research of shared leadership, which is based on sharing all leadership functions, this study differentiates multiple distributions of leadership functions between team members. The different distributions of these leadership functions, from explicitly dividing all leadership function among team members to none explicit division, may impact the degree of role clarity. Second, this study answers the need for mediating mechanisms between shared leadership and team effectiveness (Boies et al., 2010;

Drescher, Welpe, Korsgaard, Picot & Wigand, 2014; Hoch & Kozlowski, 2014; Wang et al., 2014). Multiple different things can be happening in a team when shared leadership occurs, and I want to disentangle those different things by testing role clarity and team efficacy as mediators, because of the importance of role clarity (Rizzo, House & Lirtzman, 1970) and team efficacy (Pearce & Conger, 2003) in relation to team effectiveness.

This paper is structured as follows. The paper begins with a theoretical background of

the concept of shared leadership, including the role of role clarity and team efficacy. This will

be followed by the research methodology. Next, the results will be presented and discussed in

terms of practical and managerial implications. The final part consists of research limitations

and suggestions for future research.

THEORY AND HYPOTHESES

Shared leadership

In recent years, there has been an increasing amount of literature on shared leadership, which is frequently been compared to others forms of shared leadership, such as team leadership, collective leadership and distributed leadership (Wang et al., 2014). Shared leadership is best contrasted with traditional ‘vertical’ or ‘hierarchical’ leadership, which can be described as a downward influence on subordinates by an appointed or elected leader (Conger & Pearce, 2003). Shared leadership has to be distinguished from other concepts applied to work teams, such as team empowerment, cooperation, and autonomy or self- managing teams. In the next paragraphs these concepts will be described and contrasted with shared leadership.

Team empowerment is a motivational construct and can be defined as ‘‘increased task motivation that is due to team members’ collective, positive assessments of their organizational tasks’’ (Kirkman & Rosen, 2000). Teams experience empowerment on four dimensions: meaningfulness, autonomy, impact and potency (Kirkman & Rosen, 1999). Team empowerment can be seen as an emergent state that either precedes or follows team processes.

Therefore team empowerment may facilitate the development of shared leadership by motivating team members to exercise influence (Carson et al., 2007). Conversely, shared leadership may also lead to greater team empowerment by increase members’ sense of meaningfulness, autonomy, impact or potency, but this depends on the stage of a team’s development (Carson et al., 2007).

Shared leadership is related to but distinct from other team processes such as

cooperation or helping. These team processes refer to team members working with and/or

assisting other team members with their tasks (Kozlowski & Bell, 2003). Cooperation or

helping differ from shared leadership in that they do not involve active influence that is

essential to leadership (Carson et al., 2007).

Other shared leadership related concepts are team autonomy and self-managing teams.

Members in self-managing and autonomous teams have greater responsibility for setting their own goals, monitoring their own progress, and making their own decisions compared to members of manager-led teams (Hackman, 1987). Self-managing teams may promote the development of shared leadership through increased self-management (Manz & Sims, 1987) or increased trust or autonomy (Langfred, 2004), among other things. Besides the fact that self-managing teams can promote the development of shared leadership, Manz, Skaggs, Pearce and Wassenaar (2015) argue that the positive effects of shared and self-leadership will be strongest when combined.

Shared leadership and team effectiveness

A large and growing body of literature has investigated the relationship between shared leadership and team performance or team effectiveness. Shared leadership is positively associated with team and organizational outcomes in a range of different organizational settings and for a variety of types of teams (Avolio, Jung, Murry, & Sivasubramaniam, 1996;

Carson et al., 2007; Ensley, Hmieleski, & Pearce, 2006; Pearce, Yoo, & Alavi, 2004). A recent meta-analysis of 42 independent samples of shared leadership of Wang et al. (2014) demonstrates a positive effect of shared leadership on team effectiveness, but with the important notion that what is actually shared appears to matter with regard to team effectiveness. They distinguished three forms of shared leadership: (a) traditional leadership;

(b) new-genre leadership; and (c) cumulative, overall shared leadership. These various forms

of shared leadership differ in terms of content regarding leadership styles or behaviors and in

its effect on team effectiveness. Wang and colleagues (2014) argue that shared new-genre

leadership, which emphasizes change and development (Contractor, DeChurch, Carson,

Carter, & Keegan, 2012), or overall shared leadership, is more beneficial compared to the more traditional shared leadership, which relies on exchanges or the maintenance of the status quo. Moreover, compared to vertical leadership, Wang et al. (2014) found that shared leadership shows unique effects in relation to team performance.

The increased attention to shared leadership is rooted in the proposition that it contributes to team performance and effectiveness due to greater team coordination, information sharing, intragroup trust, cohesion and team confidence (Bergman et al., 2012;

Gupta, Huang & Niranjan, 2010; Hoch, 2014; Mathieu et al., 2000; Nicolaides et al., 2014;

Sivasubramaniam, Murry, Avolio & Jung, 2002). For example, Gupta et al. (2010) argue that cohesion mediates the impact of shared leadership on performance. Moreover, a longitudinal study with 142 groups engaged in a strategic simulation game over a 4-month period demonstrates the mediating effect of trust on shared leadership and team performance (Drescher et al., 2014). In addition, Day, Gronn & Salas (2004) argue that shared leadership can enhance team effectiveness by increasing team social capital, which includes knowledge, abilities and skills, through team information-processing and learning. Finally, a recent meta- analysis of Nicolaides et al. (2014) of 3882 teams shows that team confidence, which includes both collective efficacy and group potency, partially mediates the effects of shared leadership on team performance. Together, these studies indicate that shared leadership is positively related to team effectiveness and this relationship is mediated by e.g. team coordination, information sharing and intragroup trust.

Hypothesis 1: Shared leadership has a positive relationship with team effectiveness.

Distributions of shared leadership, role clarity and team effectiveness

Much of the available research on shared leadership deals with the question whether

shared leadership positively relates to team effectiveness, and if so, through which mechanisms. Besides these mechanisms, there are other things we do not know yet, like how the distribution of different leadership functions matter. Morgeson et al. (2010) have identified eight different leadership functions which can be shared among team members during the ‘action phase’, which is the portion of the team’s performance cycle where the team is focused on activities that directly contribute to accomplishing its goals (Marks, Mathieu & Zaccaro, 2001). These functions include: ‘’monitoring the team and its performance environment, managing the boundaries between the team and the broader organizational environment, challenging the team to continually improve, becoming involved in performing the team’s work, solving problems that the team encounters, acquiring resources for the team, encouraging the team to act autonomously, and cultivating a positive social climate within the team’’ (Morgeson et al., 2010). I exclude the ‘managing team boundaries’ leadership function in this study, because this function entails managing the relationship between the team and the larger organizational context (Ancona, 1990; Ancona &

Caldwell, 1992), which is not applicable to the lab setting in this study.

The seven remaining leadership functions are divided on a continuum from a low

distribution of leadership functions to a high distribution of leadership functions. A low

distribution of shared leadership functions refers to a situation where all leadership functions

are shared among the team members but where no strict and explicit distribution of these

leadership functions among the team members exists. In teams with a low distribution of

leadership functions the team members are together responsible for all the leadership

functions as provided by Morgeson et al. (2010). In contrast, a high distribution of shared

leadership functions refers to a situation where all leadership functions are shared and

explicitly divided among the team members. In such a discrete distribution of leadership

functions in a team, team member A is responsible for some of the leadership functions (i.e.

monitoring and challenging the team), while team member B concentrates on other leadership functions (i.e. helping the team performing the task, solving problems and providing resources) and while yet team member C is responsible for the remaining leadership functions (i.e. encouraging team self-management and supporting the social climate).

Roles are essential in the structure and performance of small groups (Bray & Brawley, 2002) and can be considered to be ‘‘a set of prescriptions that define the behaviors required of an individual member who occupies a certain position’’ (Bray & Brawley, 2002; Katz &

Kahn, 1978). Role clarity and role ambiguity are essential elements within role theory. High role clarity (i.e. low role ambiguity) represents clear job procedures, goals, criteria, and knowledge of consequences. Individuals who experience high levels of role clarity clearly understand the duties, tasks, objectives and expectations of their work roles (Hinkin &

Schriesheim, 2008; Katz & Kahn, 1978). In contrast, low role clarity (i.e. high role ambiguity) represents ambiguous procedures, goals, criteria and knowledge of consequences (Rizzo et al., 1970). Therefore, individuals who experience low levels of role clarity have a lack of clear information associated with their particular role (Kahn, Wolfe, Quinn, Snoek & Rosenthal, 1964).

Nowadays, an important shift can be seen in organizations, whereby tasks are

increasingly abstract and complex and employees have to work in several teams and report to

several managers. Due to these recent developments role clarity can be easily compromised in

organizations (Wong, DeSanctis & Staudenmayer, 2007). This phenomenon can be

detrimental to organizations, because if there is a lack of clear roles, employees are unlikely to

identify with their organization’s strategy (He, Lai & Lu, 2011) or align their behavior with

the organization’s strategy (Riel, Berens & Dijkstra, 2009). Moreover, the absence of role

clarity is negatively associated with job performance, job satisfaction, organizational

citizenship behaviors and organizational commitment (Churchill, Ford, Hartley & Walker,

1985; DeRuyter, Wetzels & Feinberg, 2001; Eatough, Chang, Miloslavic & Johnson, 2011;

Gilboa, Shirom, Fried & Cooper, 2008; Ngo, Foley & Loi, 2005; Slattery, Selvarajan &

Anderson, 2008; Tubre & Collins, 2000). These results stress the importance of ensuring that all organizational members understand their roles (Kauppila, 2014).

Previous research has established some antecedents of role clarity. While Mukherjee

& Malhotra (2006) focus on interpersonal antecedents such as feedback, autonomy, participation, supervisory consideration and team support, Harris and colleagues (2014) are more concerned with antecedents in terms of leadership. They demonstrated positive links between empowering leadership and role clarity.

Following this finding of Harris et al. (2014), I expect different effects of the various distributions of shared leadership on role clarity. On one hand, high distributions of leadership functions provide an explicit function allocation. This, in turn, will result in higher levels of role clarity, namely clear understanding of the duties, task, objectives and expectations of the roles. On the other hand, in teams with low distribution of leadership functions, the different functions are not explicitly assigned to members. This may cause role ambiguity, since individuals have a lack of clear information and it is unclear what is expected from them.

Hypothesis 2: The distribution of shared leadership has a positive effect on role clarity.

Thus far, several studies have reported consequences of role clarity. For instance,

DeRuyter et al. (2001) demonstrated that role clarity is linked with employees’ job

satisfaction, organizational commitment and improved performance. Moreover, Cohen (1980)

found that high role clarity enables individuals to preserve their mental energy and use it

effectively to accomplish the tasks. In another study about the effect of role clarity on

effectiveness, Jackson & Schuler (1985) have reported that high role clarity increases individual’s motivation and ability to perform effectively. This corresponds to a more recent longitudinal field study of Fried et al. (2003), which shows that higher levels of role clarity has a positive influence on performance. Therefore,

Hypothesis 3: Role clarity has a positive effect on team effectiveness.

Since the distribution of shared leadership will positively influence role clarity, and in turn team effectiveness, I expect that, aligned with the findings that shared leadership has a positive effect on team effectiveness (e.g. D’Innocenz et al., 2014; Nicolaides et al., 2014;

Wang et al., 2014), the distribution of shared leadership has a positive effect on team effectiveness.

Hypothesis 4: The distribution of shared leadership has a positive effect on team effectiveness.

Role clarity and team efficacy

Team efficacy, or collective efficacy, refers to “a shared belief in a group's collective capability to organize and execute courses of action required to produce given levels of goal attainment” (Bandura, 1997). Collective efficacy and team efficacy differ from each other at the unit of focus. Whereas collective efficacy can refer to teams, departments, organizations and even nations, team efficacy refers specifically to teams (Gully, Incalcaterra, Joshi &

Beaubien, 2002). Since this study focus mainly on groups and teams, I will use the term team

efficacy to indicate the specific level of interest. It is important to note that this is not just a

simple aggregate of self-efficacy across group members, but it refers to the shared perceptions

as a collective entity (Bandura, 2000; Chan, 1998; Gully et al., 2002). Team efficacy is highly related to group potency (Gully et al., 2002; Stajkovic, Lee & Nyberg, 2009), which is a

‘‘collective belief that the group can be effective’’ (Shea & Guzzo, 1987). Although the high inter-relatedness of these two constructs, they distinct in that team efficacy is task specific and group potency is a ‘‘shared group-level belief about its general effectiveness across multiple tasks and contexts’’ (Kozlowski & Ilgen, 2006). Since the current study includes only one team task in one context, team efficacy seems to be more appropriate than group potency.

In recent years, there has been an increasing amount of literature on the antecedents of team efficacy. For example, Chen & Bliese (2002) argue that leadership climate is an immediate predictor of collective efficacy. Leadership climate at higher organizational levels is more strongly associated with collective efficacy than leadership climate at lower organizational levels. In addition, a study of Cheng & Yang (2014) demonstrates that team knowledge, achievement motivation and knowledge integration capability positively influence collective efficacy. Another antecedent of team efficacy might be the level of role clarity.

Yukl (2013) has identified some actions that might be useful to raise team efficacy, including providing clear explanations when assigning tasks. If one does not know what demands must be fulfilled, one cannot accurately judge whether one has the requisite abilities to perform the task. This view is supported by Beauchamp & Bray (2001) who reports that role ambiguity is negatively associated with role-related efficacy beliefs. Even though this research was focused on the individual level, the results can be translated to the group level. Members who have a clear understanding of the task that they have to perform in a team context are likely to be more efficacious. Therefore,

Hypothesis 5: Role clarity has a positive effect on team efficacy.

Team efficacy and team effectiveness

According to Bandura (1982; 1986; 1997) the degree of team efficacy influences what a group choose to do as a team (i.e., goal setting), how much effort they will put in and its persistence in the face of failure. Recent research seems to support this assertion. For instance, Campion, Medsker, & Higgs (1993) found that team efficacy is the strongest predictor, out of nineteen group characteristics, of team effectiveness. Similarly, a meta-analysis of Gully et al.

(2002) of 67 empirical studies has shown a positive relationship between collective efficacy and team performance. In addition, a number of other studies have shown that team efficacy is positively related to team effectiveness (Jung & Sosik, 2002; Pearce, Gallagher & Ensley, 2002; Sosik, Avolio, & Kahai, 1997) or team performance (Gibson, 1999; Prussia & Kinicki, 1996).

As stated before, Bandura (1982; 1986; 1997) argues that what people choose to do in a group and how much effort they put into tasks depends on the degree of team efficacy. So, if individuals believe in their group members’ capability to perform the task successfully as a group, they are more likely to sustain their efforts until the goals are achieved. Therefore,

Hypothesis 6: Team efficacy has a positive effect on team effectiveness.

Distribution of shared leadership, role clarity, team efficacy and team effectiveness

Previous studies have reported that team efficacy mediates the effects of leadership

behaviors displayed by formal leaders on team performance (Bass, Avolio, Jung, & Berson,

2003; Lester, Meglino, & Korsgaard, 2002). Similarly, sharing the leadership functions

among team members should increase the collective belief that the team can organize and

execute courses of actions required to produce given levels of goal attainment. As a result of

sharing the leadership functions, members gain more voice in team direction and in the

management of team processes, which in turn should foster a shared commitment to team action and therefore lead to a more positive sense of the team’s ability to succeed (Nicolaides et al., 2014). In addition, Srivastava, Bartol & Locke (2006) performed a study within 102 management teams and found a positive relationship between empowering leadership and team efficacy, which in turn was positively related to performance. Furthermore, an exploratory study of Kumpfer, Turner, Hopkins & Librett (1993) reported a positive relationship between a group’s level of empowering leadership and levels of team efficacy and between team efficacy and group effectiveness.

Due to a higher distribution of leadership functions, members know better what is expected from them (i.e. higher role clarity), causing a more positive belief in a group’s collective capability to attain a goal, which in turn leads to higher team effectiveness.

Therefore,

Hypothesis 7: Role clarity and team efficacy mediate the positive relationship between distribution of shared leadership and team effectiveness.

METHODOLOGY Sample

To test my hypotheses, I relied on an experiment procedure using undergraduate

business students from the faculty Business and Economics at the University of Groningen. A

total of 216 undergraduate students participated in this 30-minutes experiment, which was

combined with another study about group decision making of 30 minutes due to laboratory

availability. Sixty-eight percent of the participants were female and the mean age was 21.39

years (SD = 2.62). There was a high diversity among the team members in terms of

nationality (36.1% Dutch, 16.7% German, 8.8% Chinese and 38.4% other, such as

Indonesian, Bulgarian, and Romanian). The students could participate in both studies for money (€8) or course credits.

Procedures

Upon arrival in the laboratory participants read and signed a consent form. The participants were scheduled in teams of three persons and each session was randomly assigned to a particular experimental condition before the experiment. The participants were seated together around a small table and randomly received a team member letter (A, B or C).

Next, participants completed an initial paper-and-pencil questionnaire that included control variables (e.g. age, gender, nationality) and questions concerning the Big-Five dimensions of John & Srivastava (1999). Hereafter, the students completed the group decision-making task and a questionnaire of the other study. The experimental manipulations were then enacted.

In the low distribution of shared leadership condition, participants were told that they would be working in a team without one leader, but that they would share all the leadership functions. These leadership functions include monitoring the team, challenging the team, helping to perform the team task, solving problems, providing resources, encouraging self- management of the team and supporting the social climate within the team.

Some possible actions that the team members might want to use in order to perform the leadership functions during the team exercise were provided (see Appendix A).

In the high distribution of shared leadership condition, participants were told that they

would be working in a team without one leader, but that they share all the leadership

functions. Moreover, they were told that each team member is responsible for a couple of

these leadership functions. Team member A was responsible for monitoring and challenging

the team, team member B was responsible for helping the team performing the task, solving

problems and providing resources and team member C was responsible for encouraging team

self-management and supporting the social climate. Some possible actions that the team members might want to use in order to perform his/her leadership functions during the team exercise were provided (see Appendix A).

The control group includes vertical leadership, since this form of leadership is best contrasted with shared leadership (Pearce & Conger, 2003; Wang et al., 2014). In the vertical leadership condition, the participants were told that they would be working in a team with one leader and two followers. Member A was the leader and was responsible for monitoring the team, challenging the team, helping to perform the team task, solving problems, providing resources, encouraging self-management of the team and supporting the social climate within the team. Some possible actions that the leader might want to use in order to perform his/her leadership functions during the team exercise were provided (see Appendix A). Members B and C were the followers and had to follow the leader’s directions.

After the experimental manipulations, the participants filled out a questionnaire

regarding role clarity and read the team exercise instructions. In these instructions the

participants could read that they have to build a rainforest animal with a LEGO construction

set within 15 minutes, that they have to perform the leadership functions that were assigned to

them and that the researcher would keep track of the time. When they were finished reading, I

asked the participants whether there were any questions. If not, I started the 15 minutes and

give them a 5-minutes warning and stopped the building process after 15 minutes. After the

exercise, the participants completed a final questionnaire that included a manipulation check

and questions concerning team efficacy, team viability, team satisfaction and transactive

memory system. At the end of the experiment, the participants read the debriefing and got

their money or course credits.

Measures

Role clarity. Role clarity was measured with the role clarity scale from Rizzo et al.

(1970). These items reflect certainty about duties, authority, allocation of time, and relationships with others; the clarity or existence of guides, directives, policies; and the ability to predict sanctions as outcomes of behavior. Since these scales were developed for use within organizations and not recently student-teams, the language used and contextual inferences made several items inappropriate. For the remaining 6 items I used the seven-point Likert scale. Each participant indicated to which extent he/she agreed or disagreed with the statements. Examples: ‘I feel certain about how much authority I have’ (1 = strongly disagree;

7 = strongly agree) and ‘I know what my responsibilities are’ (1 = strongly disagree; 7 = strongly agree). All the questionnaire items of role clarity are provided in Appendix B (α = .765).

Team efficacy. Team efficacy was assessed with the 5-item measure developed by Jung & Sosik (2002). Each participant indicated to which extent he/she agreed or disagreed with the statements. Sample items were ‘My team members work hard to fulfill the group’s overall responsibilities’ and ‘My team can find solutions to problems with its performance’.

Participants answered these questions using a 7-point scale ranging from 1 (strongly disagree) to 7 (strongly agree). All the questionnaire items of team efficacy are provided in Appendix B (α = .805). The variable team efficacy is conceptualized at the group level, requiring aggregation of the data collected from individuals.

Team effectiveness. As noted earlier, the teams had to build a rainforest animal with a LEGO construction set. This team exercise meets the task requirements for shared leadership:

the task is highly interdependent, requires a great deal of creativity and is complex (Pearce,

2004). The total steps completed (out of the 57 steps in total) are a measure of team

performance and constitutes a primary indicator of team effectiveness. In fact, based on these

team performances, the teams are ranked from 1 to 72.

Control variables. Based on a study with 96 dispersed teams of Muethel, Gehrlein &

Hoegl (2012), I controlled for three potential covariates: age, gender and nationality. Muethel et al. (2012) demonstrated that a team’s high mean age negatively relates to shared leadership behaviors, team’s high female-to-male ratio positively relates to shared leadership behaviors and national diversity positively related to shared leadership behaviors.

Data aggregation

Measurement errors can seriously affect statistical analysis and interpretation, so it is very important to assess the amount of such error by calculating a reliability index (Shrout &

Fleiss, 1979). There is a consensus among social scientists that data on group level may provide more reliable estimates of occurrence than do raw individual-level variables (Glick, 1985; James, 1982; Paunonen & Gardner, 1991).

In this study, the dependent variable was measured at team level, whereas the independent variables were measured at individual level. Therefore it was necessary to demonstrate that the individual level data are in agreement with one another (LeBreton &

Senter, 2008) before aggregating the individual scores to team scores. In order to justify the use of aggregated data throughout the analyses, the ICC(1) and ICC(2) statistics of role clarity and team efficacy were measured. The ICC(1) could be interpreted as the percentage of variance due the disagreement among the raters. The ICC(2) assesses the reliability of average ratings rather than the reliability of a single rating (Bartko, 1976).

The ICC(1) value for role clarity was .04; and the ICC(2) value was .10. ICC(1) value

for team efficacy was .20; and the ICC(2) value was .43. ICC(1) values should be between .05

and .20 (Bliese, 2000) and ICC(2) values should be between .70 and .85 to justify aggregation

(LeBrenton & Senter, 2008). These ICC(1) and ICC(2) scores makes it pretty clear that the

teams did not develop a shared perception of role clarity, though they seemed to for collective efficacy. Although the ICC values of role clarity do not justify aggregation, I decided however to continue the analysis with the aggregated data, because some similarity in answers could be predicted. The first explanation can be found in the small group size of 3 persons, which could result in high similarity among answers. Another reason for high similarity among the answers is because measurements of role clarity tend to show a slightly high mean score with small spread (Bray & Brawley, 2002; Kauppila, 2014; Mukherjee & Malhotra, 2006).

Analysis

To examine the effect of different leadership manipulations, several analyses were

performed. To start with, the data was aggregated from individual level to team level, which

resulted in one score per team for each measure. Next, the main descriptive and bivariate

correlations were analyzed to give a first glance of the data and to check the assumption of

normal distribution. In order to test whether shared leadership is more effective than vertical

leadership and to identify which distribution of shared leadership is more effective, Mann-

Whitney U-tests are performed. Moreover, to analyze if role clarity differs in the three groups,

I conducted a one-way ANOVA from leadership manipulation on role clarity. Furthermore,

the effect of role clarity and team efficacy on team effectiveness was analyzed by a multiple

regression. In addition, an OLS regression was used to analyze the relationship between role

clarity and team efficacy. Finally, to test the mediation in hypothesis 7, I used Tofighi and

MacKinnon’s (2011) R Mediation program to calculate asymmetric 95% confidence intervals

with the distribution of products of coefficients method (PRODCLIN; MacKinnon,

Lockwood, Hoffman, West, & Sheets, 2002; MacKinnon, Lockwood, & Williams, 2004).

Normality

Role clarity is normal distributed with a skewness of -.137 (SE = .283) and a kurtosis of -.765 (SE = .559). However, the analysis of the distribution of the variables team efficacy and team effectiveness reported non-normal distributions. The distribution of team efficacy is moderately skewed to the left (Skewness = -.638, SE = .283) with a more distinct peak (Kurtosis = 1,574, SE = .559). The distribution of the dependent variable team effectiveness is hardly skewed (Skewness = .065, SE = .283) with a more distinct peak to the left (Kurtosis = 1,243, SE = .559). I conducted a log-transform on team effectiveness as suggested by Bland &

Altman (1996). However, the transformation did not lead to a normal distribution for team effectiveness. The analyses were conducted with both the original team effectiveness variable and the transformed variable and no significant differences between analyses were observed.

Therefore, reported results are based on the analyses with the original variable.

RESULTS

Manipulation check

The leadership manipulation was assessed using a single item (‘’To which extent took each team member different and specific leadership functions, such as monitoring the team, challenging the team, helping to perform the team task, solving problems, providing resources, encouraging self-management of the team and supporting the social climate within the team?’’, where 1 = To a great extent, 2 = Somewhat, 3 = Very little, 4 = Not at all).

A Kruskal-Wallis test demonstrated that the manipulation moderately worked, χ2 (2) = 5.390, p = .068.

Descriptive statistics and correlations

Table 1 presents the descriptive statistics and Spearman correlations of all variables

used in this study. No correlations or associations were found between leadership manipulation and team effectiveness (r = -.092, p = .443). Neither were associations observed between role clarity and team effectiveness (r = -.229, p = .053) or leadership manipulation and team efficacy (r = .044, p = .715). However, team efficacy had a positive correlation with role clarity (r = .331, p = .005).

Team efficacy shows a high mean (M = 5.57, SD = .53), indicating that the participants had a positive belief in the team’s collective capability, which is consistent with means found in studies on team efficacy (Chen & Bliese, 2002; Jung & Sosik, 2002;

Moolenaar, Sleegers & Daly, 2012). Moreover, role clarity shows a high mean as well (M = 5.04, SD = .48), suggesting that the instructions and goals were pretty clear. This is consistent with means found in other studies about role clarity (e.g. Bray & Brawley, 2002; Kauppila, 2014). The standard deviation of team effectiveness was pretty high (M = 34.89, SD = 20.77) because of the ranking of the teams from rank 1 until 72.

The three groups do not significantly differ in terms of mean age (χ2(2) = 5,046, p = 0.080), gender (χ2(2) = .128, p = .938) and nationality (χ2(6) = 12.088, p = .060), but it is important to note that the significance levels of age and nationality are close to .05 and therefore close to a significant difference between the teams. Taken together, the covariates age, gender and nationality have no significant impact on the study variables and their relationships.

Insert Table 1 about here

Hypotheses testing

Hypothesis 1 proposed that shared leadership is positively related to team

effectiveness. However, results demonstrate no significant difference in team effectiveness

between the shared leadership teams and the vertical leadership teams (U = 538,500, p = .653,

r = .053). So there is no difference in team effectiveness between shared leadership teams

(Mdn = 35.00, n = 48) and vertical leadership teams (Mdn = 32.00, n = 24). Therefore, hypothesis 1 is rejected.

Hypothesis 2 proposed that the distribution of shared leadership has a positive effect on role clarity. Nevertheless, no significant differences in role clarity were found between the three groups, F(2, 69) = 1,073, p = .348. These results indicate that the level of role clarity does not differ between the teams with low distribution of shared leadership, high distribution of shared leadership and vertical leadership teams. Therefore, hypothesis 2 is also rejected.

Hypothesis 3 predicted that role clarity has a positive effect on team effectiveness. For testing this hypothesis I treated the ordinal dependent variable as an interval variable. A multiple regression was run to predict team effectiveness from role clarity and team efficacy.

These variables statistically predicted team effectiveness, F(2,69) = 5.229, p = .008, R

= .132.

The analysis shows that role clarity does significantly predict team effectiveness (β = -.319, t(71)= -2.691, p = .009), but in the opposite way as was predicted. Therefore, hypothesis 3 is rejected.

Hypothesis 4 proposed that the distribution of shared leadership has a positive effect on team effectiveness. However, a Mann-Whitney U test indicated no significant difference in team effectiveness between the teams with low distribution of shared leadership (Mdn = 38.50, n = 24) and the teams with high distribution of shared leadership (Mdn = 35.00, n = 24), U = 249.00, p = .420, r = .116.

Hypotheses 5 predicted that role clarity has a positive effect on team efficacy. A regression analysis revealed a significant positive effect of role clarity on team efficacy, R

= .105, F(1,70) = 8.236, p = .005. It was found that role clarity significantly predicted team efficacy, β = .358, p = .005. Therefore, hypothesis 5 is confirmed.

Hypothesis 6 proposed that team efficacy has a positive effect on team effectiveness.

The analysis of the multiple regression as was conducted for hypothesis 3, shows that team

efficacy significant predicted team effectiveness (β = .305, t(71) = 2.570, p = .012).

Therefore, hypothesis 6 is confirmed.

Finally, hypothesis 7 predicted the mediating effect of role clarity and team efficacy.

Using the PRODCLIN mediation technique, no mediation effect was found of role clarity in the relationship between shared leadership and team effectiveness (indirect effect = .909, SE = 1.596, 95% CI = [-1.472, 4.976]). Furthermore, there was no mediation effect found of team efficacy in the relationship between shared leadership and team effectiveness (indirect effect = 1.541, SE = 2.085, 95% CI = [-1.663,6.639]). Consequently, the analyses revealed no support for hypothesis 7 and I can generally conclude that there is no statistically significant indirect effect for the relationship of shared leadership with team effectiveness via changes in role clarity and/or team efficacy.

Descriptive statistics for role clarity, team efficacy and team effectiveness from the three manipulation groups are presented in table 2.

Insert Table 2 about here

DISCUSSION

Summary of results

The overall objective of the study was to examine to what extent shared leadership

functions have to be explicitly distributed among team members. Consistent with latest

arguments about shared leadership and recent findings by e.g. Carson et al. (2007), Drescher

et al. (2014), Pearce & Conger (2003) and Wang et al. (2014), it was hypothesized that shared

leadership would be related to higher levels of team effectiveness compared to vertical

leadership. On the basis of latest reasoning about role clarity (e.g. Mukherjee & Malhotra,

2006; Yukl, 2013) and team efficacy (e.g. Gibson, 1999; Gully et al., 2002; Jung & Sosik,

1999; Kozlowski & Ilgen, 2006), it was also hypothesized that a high distribution of shared

leadership would result in higher team effectivity compared to a low distribution of shared leadership. This experiment did not detect any evidence for the positive effects of the distribution of shared leadership on team effectiveness. Moreover, contrary to expectations, this study did not find a significant effect of the distribution of shared leadership on role clarity. However, this research did find a significant negative influence of role clarity on team effectiveness, suggesting that teams who have a clear understanding of their tasks and objectives perform worse. Furthermore, it was found that team efficacy was positively related to team effectiveness, indicating that a shared belief in a group’s collective capability to attain goals will result in better performances. Moreover, this study reveals a positive relationship between role clarity and team efficacy. This finding suggest that team members who know what is expected from them to perform their job and who have for instance clear job procedures and goals, have a more positive attitude towards the group’s collective capability to attain their collective goal.

Reflection on results

The absence of supporting evidence may be due to several reasons. In this section different explanations will be given. To start with, although the non-significant positive effect of shared leadership on team effectiveness differ from some published studies (e.g. Carson et al., 2003; Pearce & Conger, 2003; Sivasubramaniam et al., 2002), the result is consistent with those of Boies et al. (2010) and Friedrich et al. (2009) who did not supported this relationship.

It is possible, therefore, that the effectiveness of shared leadership differs per setting (i.e. team

or organization). So it may be the case that the student teams that are used in this study are not

more effective when they work in shared leadership teams relative to vertical leadership

teams. Moreover, the results of the manipulation checks indicate that the leadership

manipulation was only partially successful. It is therefore likely that the three different

manipulation groups did not perform their leadership functions that were assigned to them. As a result, the effectiveness of the three different groups must be interpreted with caution.

Another possible explanation for the non-significant results is that teams sampled from laboratory settings yields lower average effect sizes as compared to teams sampled from field setting (D’Innocenzo et al., 2014). In addition, the ICC(1) and ICC(2) results might explain some of the non-significant results. These scores make it pretty clear that the teams did not develop a shared perception of role clarity, though they seemed to for collective efficacy.

Therefore, it was hard to find an effect of distribution of shared leadership on role clarity.

Finally, some authors (e.g. Carson et al., 2007; Hoch & Kozlowski, 2014; Pearce & Conger, 2003) have seen shared leadership as an emergent team property that results from the distribution of leadership influence across team members. Yukl (2013) has argued that processes such as developing function distributions and building effective teams often require months or years of study. Others have highlighted the importance of formally appointing leadership functions across team members (Mehra, Smith, Dixon & Robertson, 2006).

Therefore, it can be suggested that teams where leadership functions are appointed to members may not be as effective as teams where leadership functions and its distribution develop over time. So, since in this study leadership functions are appointed to team members, this might be a reason for the non-significant results.

Theoretical and practical implications

The results of this research offer some important contributions to the literature of

shared leadership, role clarity, team efficacy and team effectiveness. An important finding

was that role clarity is positively related to team efficacy. This finding enriches the literature

on the concepts of role clarity and team efficacy, since previous research of Yukl (2013) and

Beauchamp & Bray (2001) focused only on the individual level of self-efficacy. So this

finding creates new insights that have not been found in any research concerning individual role clarity and its effect on team efficacy. Moreover, several authors (e.g. DeRuyter et al., 2001; Fried et al., 2003; Jackson & Schuler, 1985) have shown that role clarity is positively related to team effectiveness. This differs from the findings presented here, which shows a significant negative effect of role clarity on team effectiveness. This finding implies that team effectiveness might decrease when team members clearly understand what is expected from them, which is surprisingly. Finally, an additional finding is the direct relationship between team efficacy and team effectiveness, which was significantly positive. This finding confirms research of, among others, Campion and colleagues (1993) and Gully et al. (2012) concerning the positive relationship between team efficacy and team performance. This implies that an increase in the team’s shared belief in its group collective capability would increase team performance.

Besides theoretical implications, this research also provides valuable insights for practitioners. Despite the fact that this research did not reveal any significant positive effect of shared leadership on team effectiveness, three recent meta-analyses by D’Innocenzo et al.

(2014), Nicolaides et al. (2014) and Wang et al. (2014) did demonstrate this relationship. It

can be therefore assumed that shared leadership results in higher team performances. So,

managers should form teams in which leadership will be shared in order to increase, for

instance, cohesion (Mathieu, Kukenberg, D’Innocenzo & Reilly, 2015) and team performance

(D’Innocenzo et al, 2014). Moreover, since role clarity will result in higher team efficacy,

organizations should focus on the antecedents of role clarity, such as participation in decision-

making, team support and feedback (Mukherjee & Malhotra, 2006). In addition, since team

efficacy is positively related to team effectiveness, managers should focus on increasing team

efficacy. For instance, managers could improve training and professional development

practices, set reasonable goals and expectations, provide coaching and/or assign mentors and

expressing confidence in the team.

Limitations and future research

The focus of this study was on contributing to theory on shared leadership and its distribution in teams. Since the results failed to support the hypotheses and the underlying theories, it is important to address the limitations for the study that may also present fruitful avenues for future research. An important limitation of this study is the modest sample size of teams included in the analysis. With 216 participants in total and only 24 teams in each of the 3 conditions, the statistical power of the analyses were low and some effects might have failed to reach conventional levels of significance because of that. Due to this small sample size the generalizability remains in question. Therefore, replication of the current findings with a larger sample is warranted.

Another concern is that I treated role clarity as a group variable, because I wanted to explore the differences in experiencing role clarity between the three manipulation groups.

However, traditionally, it has been argued that role clarity is a construct on individual level (Beauchamp & Bray, 2001; Bray & Brawley, 2002; Harris et al., 2014; Mukherjee &

Malhotra, 2006) and therefore should not be aggregated to the team level. It is recommended that future research should focus on the emergent and collective function of this construct. In addition, I aggregated the data about role clarity from individual level to group level despite insufficient ICC values. Since the variation between the scores given to each item of role clarity by the raters is too high, there is no consensus among the team members about the level of role clarity. So, the aggregated data do not present a shared perception of role clarity.

There are also several limitations with regard to the experiment. The results on the

manipulation checks indicate only a moderate effect of the manipulation of the leadership

functions. Thus, I am forced to concede that the manipulation of leadership functions did not

have the predicted effect. One possible explanation for this might be the sequence of the study. The participants completed another decision-making task before the manipulations were enacted. During this decision making team task the team members might have already developed some role distribution and some natural leader might have been identified.

Therefore, it was more difficult to manipulate leader functions among the team members.

Another reason for this might be the use of the student teams in this manipulation. Students work a lot in temporal teams and are thus used to work within a team with new team members. In these teams there is often a lack of leadership, and instead role distributions occur naturally (Forman & Katsky, 1986; Hansen, 2006). Therefore, it was hard to manipulate these naturally emerging roles. The last possible explanation for the absence of a strong manipulation effect concerns the way of measuring this effect. The manipulation check was conducted on the basis of self-assessment rather than on an objective assessment. Bergman et al. (2012) suggest that shared leadership should be considered by either the number of team members of each team that engage in leadership behaviors and the total amount of leadership behavior displayed by the team. The former can be assessed by behaviorally coding videotapes of the teams’ process. The latter can be assessed by calculating the leadership ratings for each team member and for each type of behavior. Future research could for instance videotape the team process in order to objectively assess the leadership functions performed by each of the team members in combination with self-assessments. Another possible area of future research would be another study design that could make the manipulation stronger or more salient. For instance, larger teams could make it easier to distribute the seven leadership functions among the team members. Moreover, it is recommended to use a more complex task, since the effects of shared leadership are stronger when the work of team members is more complex (Pearce, 2004; Wang et al., 2014).

Another limitation concerns the research methodology. This study was, in contrast to

the majority of survey studies about leadership (Brown & Lord, 1999; Dipboye, 1990;

Wofford, 1999), performed in a laboratory context, because of the many beneficial features of laboratory settings including determining causality, manipulating leader behavior and assess their independent and joint effects, controlling for extraneous variables and measuring mediating processes (Pitariu & Ployhart, 2010; Yukl, 2013). But it should be noted that experimental studies would only benefit from these advantages if the study includes a strong manipulation (Yukl, 2013). However, lab samples are often criticized in organizational research because they are not representative of organizational settings (Highhouse &

Gillespie, 2008; Peterson, 2001). Alternatively, some studies demonstrated that laboratory research with student samples may generalize work teams when the focus is upon the underlying principles or the construct relationships in question (Driskell & Salas, 1992;

Mathieu et al., 2000). A recent meta-analysis of different forms of shared leadership and team performance relations of D’Innocenzo et al. (2014) has shown that teams sampled from laboratory settings yielded lower average effect sizes as compared to teams sampled from field settings. Future research should investigate whether the relationships argued for in this study demonstrates any significant effect in a field study, because a sample of experienced managers from the same industry is often more useful for a survey than a convenience sample of undergraduate students with little practical experience in organizations (Yukl, 2013).

The last limitation of this research focusses on the distribution of leadership functions.

The seven leadership functions from Morgeson et al. (2010) that were used were distributed

among only three team members. In addition, the distribution was done randomly and not

based on for instance, task related functions or interpersonal functions. So, there is abundant

room for further progress in determining if these leadership functions reflect the most

important functions, and if so, how these functions should be distributed among team

members.

CONCLUSION

As organizations continue to devote vast resources to the use of self-managing and

empowered teams, the need for a better understanding of effective shared leadership continues

to grow. The current research addresses how shared leadership is most effective, i.e. how

different distributions of leadership functions impact team effectiveness. Overall, the results

of this study do not show any support for the proposed hypotheses concerning the distribution

of shared leadership. No relationship between distribution of shared leadership and team

effectiveness was found, as well as there was no evidence for the mediating effect of both role

clarity and team efficacy on this relationship. However, this study reveals a positive effect of

role clarity on team efficacy. Moreover, evidence was found for the positive effects of role

clarity and team efficacy on team effectiveness.

REFERENCES

Ancona, D. G. 1990. Outward bound: Strategies for team survival in an organization.

Academy of Management Journal, 33: 334-365.

Ancona, D. G., & Caldwell, D. F. 1992. Bridging the boundary: External activity and

performance in organizational teams. Administrative Science Quarterly, 37: 634-665.

Avolio, B. J., Jung, D. I., Murry, W., & Sivasubramaniam, N. 1996. Building highly developed teams: Focusing on shared leadership processes, efficacy, trust, and performance. Advances in Interdisciplinary Studies of Work Teams, 3: 173–

209.

Bandura, A. 1982. Self-efficacy mechanism in human agency. American Psychologist. 37:

122-147.

Bandura, A. 1986. Social foundations of thought and action. Englewood Cliffs, NJ:

Prentice-Hall

Bandura, A. 1997. Self-efficacy: The exercise of control. New York: Freeman.

Bandura, A. 2000. Exercise of human agency through collective efficacy. Current directions in psychological science, 9(3): 75-78.

Bartko, J. J. 1976. On various intraclass correlation reliability coefficient. Psychological

Bulletin, 83(5): 762-765.

Bass, B. M., Avolio, B. J., Jung, D. I., & Berson, Y. 2003. Predicting unit performance by assessing transformational and transactional leadership. Journal of Applied

Psychology, 88(2): 207–218.

Beauchamp, M. R., & Bray, S. R. 2001. Role ambiguity and role conflict within interdependent teams. Small Group Research¸ 32(2): 133-157.

Bergman, J. Z., Rentsch, J. R., Small, E. E., Davenport, S. W., & Bergman, S. M. 2012. The shared leadership process in decision-making teams. The Journal of Social

Psychology, 152(1): 17-42.

Bland, J. M., & Altman, D. G. 1996. Transforming data. BMJ, 312: 770.

Bliese, P. D. 2000. Within-group agreement, non-independence, and reliability: implications for data aggregation and analysis. In: K. J. Klein, & S. W. Kozlowski (Eds.),

Multilevel theory, research, and methods in organizations: 349–381. San Francisco, CA: Jossey-Bass.

Boies, K., Lvina, E., & Martens, M. L. 2010. Shared leadership and team performance in a business strategy simulation. Journal of Personnel Psychology, 9(4): 195–202.

Bray, S. R., & Brawley, L. R. 2002. Role efficacy, role clarity and role performance

effectiveness. Small Group Research, 33(2): 233-253.

Brown, D. J., & Lord, R. G. 1999. The utility of experimental research in the study of transformational and charismatic leadership. Leadership Quarterly, 10: 531-539.

Campion, A. M., Medsker, J. G., & Higgs, A. C. 1993. Relations between work group characteristics and effectiveness: Implications for designing effective work groups.

Personnel Psychology, 46: 823-850.

Carson, J. B., Tesluk, P. E., & Marrone, J. A. 2007. Shared leadership in teams: an

investigation of antecedent conditions and performance, Academy of Management Journal, 50(5): 1217-1234.

Chan, D. 1998. Functional relations among constructs in the same content domain at different levels of analysis: A typology of composition models. Journal of Applied Psychology, 83: 234-246.

Chen, G., Bliese, P. D. 2002. The role of different levels of leadership in predicting self- and collective efficacy: Evidence for discontinuity. Journal of Applied Psychology, 87(3):

549-556.

Cheng, H., Yang, H. L. 2014. The antecedent of collective creative efficacy for information system development teams. Journal of Engineering & Technology Management, 33: 1-17.

Churchill, G. A., Ford, N. M., Hartley, S. W. and Walker, O. C. 1985. The determinants of

salesperson performance: a meta-analysis. Journal of Marketing Research, 22(2): 103-118.

Cohen, C. 1980. After effects of stress on human performance and social behavior: A review of research and theory. Psychological Bulletin, 88: 82–108.

Conger, J. A., & Pearce, C. L. 2003. A landscape of opportunities: Future research on shared leadership. In C. L. Pearce & J. A. Conger (Eds.), Shared leadership: Reframing the hows and whys of leadership: 285–304. Thousand Oaks, CA: Sage.

Contractor, N. S., DeChurch, L. A., Carson, J., Carter, D. R., & Keegan, B. 2012. The topology of collective leadership. The Leadership Quarterly, 23: 994–1011.

Day, D. V., Gronn, P., & Salas, E. 2004. Leadership capacity in teams. Leadership Quarterly, 15(6): 857-880.

DeRuyter, K., Wetzels, M., & Feinberg, R. 2001. Role stress in call centres: its effects on employee performance and satisfaction. Journal of Interactive Marketing, 15(2): 23- 35.

D’Innocenzo, L., Mathieu, J. E., & Kukenberger, M. R. 2014. A meta-analysis of different forms of shared leadership–team performance relations. Journal of Management.

Dipboye, R. L. 1990. Laboratory vs. field research in industrial and organizational

psychology. International Review of Industrial and Organizational Psychology,5: