SHARED LEADERSHIP AND PERFORMANCE IN TEAMS:
THE EFFECT OF EMPOWERING LEADERSHIP,
DIRECTIVE LEADERSHIP, BOUNDARY SPANNING, AND
TASK COMPLEXITY
Master Thesis
Author: Robert van Dijk Student number: 10499318
Date of submission: June 24, 2015, final version
Executive Programme in Management Studies – Leadership and Management Track Amsterdam Business School (ABS), University of Amsterdam (UvA)
2
Statement of Originality
This document is written by Student Robert van Dijk who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
3 In loving memory of my parents
Leonardus van Dijk (* 29-04-1937 † 01-07-2014) Gerrie Vis (* 24-11-1945 † 24-07-2013)
4
TABLE OF CONTENT
LIST OF TABLES AND FIGURES ... 5
ABSTRACT ... 6
CHAPTER 1. INTRODUCTION ... 7
CHAPTER 2. THEORY AND HYPOTHESES ... 9
CHAPTER 3. RESEARCH DESIGN ... 17
CHAPTER 4. RESULTS ... 22
CHAPTER 5. DISCUSSION AND CONCLUSION ... 31
ACKNOWLEDGEMENTS ... 37
REFERENCE LIST ... 38
APPENDIX A: SURVEY INVITATION EMAIL ... 42
APPENDIX B: ITEMS FOR PRIMARY MEASURES ... 43
5 LIST OF TABLES AND FIGURES
Table 1: Summary of Exploratory Factor Analysis Results for Team Performance Table 2: Descriptive Statistics and Correlations
Table 3: Results of hierarchical regression analysis of planning and shared leadership Table 4: Results of hierarchical regression analysis of performance abilities and shared leadership
Table 5: Results of hierarchical regression analysis of shared leadership measure 3 and empowering leadership and directive leadership
Table 6: Results of hierarchical regression analysis of team viability and shared leadership Table 7: Results of hierarchical regression analysis of team commitment and shared leadership
Figure 1: The conceptual model including hypotheses. Figure 2: Standardized path coefficients.
Figure 3: Moderating effects of task complexity on shared leadership measure 3-planning, when directive leadership is low.
Figure 4: Moderating effects of directive leadership on shared leadership measure 2-planning, when task complexity is low.
Figure 5: Moderating effects of directive leadership on shared leadership measure 3-planning, when task complexity is low.
Figure 6: Moderating effects of task complexity on shared leadership measure 2-performance ability, when directive leadership is low.
Figure 7: Moderating effects of task complexity on shared leadership measure 3-performance ability, when directive leadership is low.
Figure 8: Moderating effects of directive leadership on shared leadership measure 2-performance ability, when task complexity is low.
Figure 9: Moderating effects of boundary spanning on shared leadership measure 3-planning. Figure 10: Moderating effects of boundary spanning on shared leadership measure
6
ABSTRACT
I conducted surveys in 125 teams of ASML to investigate under which circumstances vertical leadership behavior effectively facilitates and coexists with shared leadership in relation to team performance. Results showed that empowering leadership was a positive predictor of shared leadership. In turn, shared leadership was found to be positively related to team viability and team commitment. My results also point in the direction that high task complexity has a negative impact on the shared leadership-team performance relationship. Directions for further research on performance, a vertical leader’s influence on shared leadership and performance, and task complexity are discussed.
7 CHAPTER 1. INTRODUCTION
“You can accomplish anything in life, provided that you do not mind who gets the credit.” – Harry S. Truman (McCullough 1992, p. 564).
What is the best way to lead a team in today’s dynamic, competitive, and complex business environment? Should a leader coach and give direction (Hackman 2002) or is there more to making organizations, teams, and individuals perform? Organizations recognize the need to, among others, fully utilize human resources to compete in the business environment (Kotter 2012, Pearce and Barkus 2004). One of the ways organizations accomplish this utilization is by the synergetic effect of work performed in teams (Pearce and Barkus 2004). The number of organizations using teams is significant and rising (Burke et al. 2006a, Burke et al. 2006b, Carson, Tesluk and Marrone 2007, Morgeson, DeRue and Karam 2010, Nicolaides et al. 2014, Pearce and Barkus 2004). Large amounts of resources in organizations work in teams, many organizations focus on teamwork, and team effectiveness is important to organizations as it has a significant impact on organizational performance (Carson, Tesluk and Marrone 2007).
Factors influencing team performance are numerous and a significant portion of these factors have been extensively researched and described in both academic and management literature. Searching Google Scholar using “team performance” provides over 120,000 search results [Accessed January 28, 2015] and the same search on amazon.com provides 3,582 results [Accessed January 28, 2015]. Leadership is one of the factors that is known to significantly influence team performance (Carson, Tesluk and Marrone 2007). Within the various leadership styles, shared leadership is a relatively new style that distributes the leadership roles and responsibilities within the team and has been found to explain a significant amount of variance in team performance (Ensley, Hmieleski and Pearce 2006, Nicolaides et al. 2014). Shared leadership therefore provides valuable insight into one of the main mechanisms that enable organizations to influence team and organizational performance. This same research has, however, also identified a number of research gaps that need to be closed to better understand the shared leadership phenomena and to know how to and under what circumstances shared leadership is most effective. It is therefore needed to also research contextual factors linked to the shared leadership-team performance relationship, and the combination of the more classical vertical leadership and relatively new shared leadership (Nicolaides et al. 2014). My thesis not only contributes to closing the mentioned gaps, but is the first in its kind to study the combination of shared leadership, its impact on performance, and the circumstances under which shared leadership coexists with other forms of leadership. I thereby provide valuable insights in factors influencing team performance, and address the role of vertical leadership in combination with shared leadership by answering the research question: when and under what
circumstances does vertical leadership effectively facilitate and coexist with shared leadership in relation to team performance?
The person responsible for team leadership has a number of different roles and responsibilities in relation to team performance. One of these roles and responsibilities of the team leader is to facilitate team member behavior that leads to individual and team performance in line with team and organizational goals. In that context empowering activities from the team leader are important when looking at the coexistence with shared leadership as it can provide
8
us with valuable insights into what facilitates shared leadership behavior in a team (Morgeson, DeRue and Karam 2010). Another role and responsibility for the team leader is to provide clear guidance and direction to followers. This form of directive vertical leadership however appears to conflict with shared leadership as enabling shared leadership behavior within a team requires team member to exercise and accept influence from each other and not have team members focus on top-down centralized leadership (Pears and Sims 2002). It is therefore important to analyze the coexistence of shared leadership with both empowering vertical leadership and directive vertical leadership.
Next to what facilitates leadership behavior, contextual factors determine much of the circumstances that teams operate in. Task complexity is one of these contextual factors and is often mentioned as having a positive impact on the shared leadership-team performance relationship (D’Innocenzo, Mathieu and Kukenberger 2014, Nicolaides et al. 2014, Wang, Waldman and Zhang 2014). Another contextual factor is the level of information processing and external representation by the team leader, also known as boundary spanning. This factor should improve shared leadership behavior within a team and thereby team performance by improved collaboration with others outside the team and increasing required information for task completion within the team (Burke et al. 2006a).
I have therefore structured my thesis as follows. In the next chapter, I review the theory on teams and team performance, shared leadership, the coexistence of shared leadership and vertical leadership by addressing the facilitating roles of empowering vertical leadership and directive vertical leadership, and the moderating effects of task complexity and boundary spanning. I also present my hypotheses, conceptual model, and the exploratory research into multi-factor interaction between shared leadership, empowering leadership, directive leadership, task complexity, and boundary spanning in relation to team performance. I then discuss the research design in chapter 3 and present the results in chapter 4. Finally, I present the discussion and conclusion in chapter 5.
9 CHAPTER 2. THEORY AND HYPOTHESES
“The greatest enemy of knowledge is not ignorance, it is the illusion of knowledge.” (Boorstin 1983).
Teams and Team Performance: an Introduction
So far, only limited research has been done to explain why and how individuals within teams collectively influence team performance (DeChurch et al. 2010). My research contributes by expanding that body of knowledge with insights in the influence of leadership and contextual factors on team performance. Before I continue it is important to define the construct ‘team’ as it is a central element in my research and multiple definitions exist. I define a team as an entity that consists of three or more persons and one that adheres to Hackman’s (2002) four elements of a real team. Hackman’s elements are 1) the presence of a team task, 2) a certain degree of team membership stability, 3) clear boundaries, and 4) someone who has the role to manage work processes.
Hackman’s first element, the presence of a team task, provides team members with a common goal and distinguishes a team from a collection of people. Team member stability, the second element, is important as constant changes in team membership can have a negative impact on team performance (Hackman 2002). This team performance construct covers a wide variety of measures used as indicators of team success and sustainability. Boselie, Dietz and Boon (2005) for example drew a distinction between three different outcomes; human resource related, financial, and organizational, to capture this broad construct. These three performance areas provide valuable distinctions when further definition and analysis of performance and outcomes is needed. My research, however, focuses on team performance as a whole and I therefore define, based on Carson, Tesluk and Marrone (2007), team performance more generally and holistically as the measure of the effectiveness of a team in achieving its goals. Hackman’s third element, clear boundaries, is used in organizations to define who within the organization has the formal authority to apply certain organizational rules (Aldrich and Herker 1977). This boundary element together with Hackman’s fourth element, someone who has the role to manage work processes, are components of the vertical leadership role, a role performed by a single formal and hierarchical leader (Mathieu et al. 2008).
Leadership is a crucial element to enable a team to perform (Carson, Tesluk and Marrone 2007). A relatively new leadership style that is described in the academic literature is that of shared leadership, leadership provided by informal and non-hierarchical team members (Wang, Waldman and Zhang 2014). Shared leadership has been found to explain significant variance in team performance, even beyond the levels of vertical leadership (Ensley, Hmieleski and Pearce 2006, Nicolaides et al. 2014). I therefore next first address the influence of shared leadership on team performance, after which I describe the moderating effect of task complexity as a contextual factor of the shared leadership-team performance relationship. I continue this theory section with a section on the coexistence of vertical and shared leadership, addressing empowering leadership, directive leadership, and the influence of the contextual factor boundary spanning. Very limited research has been done in these areas, areas that are expected to provide a more detailed understanding of the mechanisms that lead to team performance (Carson, Tesluk and Marrone 2007, D’Innocenzo, Mathieu and Kukenberger 2014, Nicolaides et al. 2014, Morgeson, DeRue and Karam 2010, Wang, Waldman and Zhang 2014). I finalize this theory and hypotheses chapter with my conceptual model and an explanation of the
10
exploratory research that I have performed on multi-factor interaction between shared leadership, empowering leadership, directive leadership, task complexity, and boundary spanning in relation to team performance.
Shared Leadership
Carson, Tesluk and Marrone (2007) conceptualize leadership in relation to its strength of influence or to the source of influence while Morgeson, DeRue and Karam (2010) conceptualize leadership based on the sources of leadership; the locus of leadership and the formality of leadership. The strength of the leadership influence is a measure of the effectiveness or quality of the leadership function. The source of influence on the other hand is an indication of who exercises the leadership influence, being a single team member or multiple team members. Viewing leadership from a locus of leadership perspectives raises the question if the leader is a member of the team (internal vs. external), while the formality of leadership is an indicator of the level of leadership responsibility formalization (formal vs. informal). Formal leadership can be expressed in roles like a team leader or project manager and is a role that is also referred to as vertical leadership. Informal internal leadership from multiple sources is an expression of leadership behavior called shared leadership, a leadership style that distributes team leadership across multiple team members (Carson, Tesluk and Marrone 2007, D’Innocenzo, Mathieu and Kukenberger 2014, Nicolaides et al. 2014, Wang, Waldman and Zhang 2014).
Shared leadership can be viewed as a leadership construct that varies from other leadership constructs in two main aspects. One is that it involves leadership that is attributed from within the team and not as a top-down process via an external leader. The other is that the leadership function is distributed across multiple people within the team and not to the single person as manager or supervisor in the vertical leadership role. When comparing shared leadership to vertical leadership, these forms of leadership can be conceptualized as a continuum. The range of this continuum extends from vertical leadership where team members follow a single individual and formalized leader, to shared leadership where team members display a high degree of leadership influence in the team (Nicolaides et al. 2014).
D’Innocenzo, Mathieu and Kukenberger (2014), Nicolaides et al. (2014), and Wang, Waldman and Zhang (2014) found mixed results in their meta analyses on the shared leadership-team performance relationship. Most studies mentioned in these meta analyses reported a positive relationship, but some a neutral or even a negative relationship (e.g. Bergman et al. 2012, Boies, Lvina and Martens 2010, Neubert 1999, and Wellman 2013; Sample 1). The general consensus among the authors of the shared leadership meta analyses, however, is that shared leadership should have a positive influence on team performance. Zhou (2013) for example studied 144 entrepreneurial teams and his results suggest a moderately strong relation between shared leadership on team performance. The positive effect of shared leadership on team performance can be explained based on the following five arguments. 1) leadership, independent of the source of that leadership, is crucial to the performance of teams (Carson, Tesluk and Marrone 2007). 2) Shared leadership roles within a team require more coordination among team members and thereby positively influences the amount of information sharing needed to perform as a team (D’Innocenzo, Mathieu and Kukenberger 2014). 3) Shared leadership increases the amount of resources that are available within the team to perform tasks by providing the leadership by and to team members, thereby positively influencing team
11 performance. 4) The previous three items combined will lead to higher team member commitment towards successful task completion within the team, thereby also positively influencing team performance. 5) In order for shared leadership to be effective, trust and respect is required from the team members. D’Innocenzo, Mathieu and Kukenberger (2014) have proven that these trust and respect characteristics have a positive relationship with team performance. I therefore, in conclusion, propose:
Hypothesis 1. The degree of shared leadership within a team is positively related to that team’s performance.
The Moderating Effect of Task Complexity
Currently organizational structures are becoming flatter and we see more and more self-managed teams (Kotter 2012). These developments increase the need for leadership from within the team to ensure or enhance team performance when factors like time, task complexity, cognitive limitations, and other contextual factors limit the vertical leader’s ability to effectively lead the team. The leadership role in teams is fundamentally orientated around the satisfaction of critical team needs (Carson, Tesluk and Marrone 2007) and team members in knowledge related work tend to seek autonomy in how they apply their knowledge, skills and abilities (Morgeson, DeRue and Karam 2010). Sharing the leadership role therefore requires additional coordination within the team, that might influence the effectiveness and efficiency of this leadership style in varying contextual situations like task complexity.
Task complexity is a topic that was long considered to be fully defined based on Wood’s (1986) and Campbell’s (1988) articles. Researchers, however, were unable to use this body of knowledge to fully explain complex situations that involve the distribution of tasks between multiple actors (Haerem, Pentland and Miller 2014). These actors can be organizational (sub)units, machines, and people. Wood (1986) and Campbell (1988) viewed task complexity as a linear function of three factors; component complexity, coordinative complexity, and dynamic complexity, and defined these factors as follows. Component complexity is a factor of task complexity that deals with the amount of steps that are “nonredundant with other acts” (Wood 1986, p. 67) and tasks with more steps are more complex. Coordinative complexity deals with the “required interactions in sequencing and timing” (Wood 1986, p. 69) between actors and the greater the amount of interactions the greater the complexity. And finally, dynamic complexity deals with “changes in the cause-effect chain […] during performance of the task” (Wood 1986, p. 71) and tasks with more changes are more complex. Haerem, Pentland and Miller (2014), however, argue that task complexity is not a linear function as argued by Wood and Campbell but an exponential function of the three previously mentioned factors. Haerem, Pentland and Miller also argue that the level of coordinative complexity is the main factor that determines the complexity of a task. Coordinative complexity sets a critical point after which task complexity starts to increase exponentially. A coordinative complexity of 14 ties marks this critical point and the level of component and dynamic complexity act as multipliers (Haerem, Pentland and Miller 2014).
Team tasks vary in task complexity and the level of ambiguity that teams are faced with. The basic question if a single leader can be successful to perform all leadership functions is also relevant in this context. The influence of task complexity in combination with the need for internal team alignment, coordination, and leadership is one that, however, requires additional research (Carson, Tesluk and Marrone 2007, Wang, Waldman and Zhang 2014). Shared
12
leadership behavior enables the sharing of leadership responsibilities within the team to facilitate task completion (D’Innocenzo, Mathieu and Kukenberger 2014) and thereby positively influences the team’s performance when complexity increases. According to Wood (1986) increased task complexity increases the need for resources (information, finances, material, and people). If task complexity is low, individuals from within the team can provide their own resources or can use their job discretion to obtain the required resources to successfully complete the required task(s). When task complexity increases, the required number of interactions between people, the number of tasks, the coordination, or a combination of these factors increases exponentially (Haerem, Pentland and Miller 2014). The person performing the shared leadership role is able to provide the required resources and perform the required coordination to successfully complete the task(s) as long as task complexity does not go beyond a critical point. When this critical point is not passed shared leadership improves team performance compared to performance of these tasks without the shared leadership role (Carson, Tesluk and Marrone 2007, D’Innocenzo, Mathieu and Kukenberger 2014, Nicolaides et al. 2014, Wang, Waldman and Zhang 2014). I therefore propose:
Hypothesis 2. Task complexity moderates the shared leadership-team performance relationship, such that when task complexity increases, the relationship between shared leadership and team performance is enhanced.
I assume that beyond the critical point of task complexity as mentioned above, the person in the shared leadership role will be limited by his or her own cognitive abilities and the limitations of access to required resources to successfully complete the task. I therefore expect that when this occurs a form of vertical leadership is needed to support both the person in the shared leadership role and the team members by fulfilling the leadership roles of task coordination and access to resources. This however implies a complex and hard to predict multi-way interaction between vertical leadership constructs, task complexity, and the shared leadership-team performance relationship. I have therefore included this topic in the exploratory research question that I use as a basis for my exploratory research instead of defining additional hypotheses on the task complexity based interactions. I have included a brief outline on my exploratory research later on in this chapter that addresses the exploratory research question and the additional research next to testing the hypotheses.
The Coexistence of Shared Leadership and Vertical Leadership
Leadership roles can be rotated within a team to the person that best fits the required leadership role to accomplish a certain team task, for example based on that person’s specific abilities, skills, or knowledge (Pearce and Barkus 2004). That however raises the question of the role of the person in the vertical leadership position and addresses a fundamental question that keeps coming back in leadership literature: under what circumstances is which type of leadership best suited to obtain organizational objectives? This question can be defined more specifically in the context of my research: what vertical leadership styles facilitate shared leadership and what elements of the vertical leadership style moderate the shared leadership-team performance relationship so that organizations can leverage the positive effect of shared leadership?
Using Yukl’s (2005, p. 5) definition of leadership; “influence processes involving determination of group’s or organization’s objectives, motivating task behavior in pursuit of these objectives, and influencing group maintenance and culture”, a number of team leadership
13 conceptualizations can be made that help define various types of effective leadership styles. Burke et al. (2006a) identified two distinct behavioral categories in this context: team leader behavior that facilitates team development and/or interaction, and team leader behavior that has a focus on accomplishing tasks. These categories, like Yukl’s definition, however do not take into account that team leadership can originate from multiple and non-hierarchical sources at the same time, is dynamic and will therefore change over time (Morgeson, DeRue and Karam 2010). Combining Carson, Tesluk and Marrone (2007) and Morgeson, DeRue and Karam (2010) conceptualizations of team leadership provides insight into the dynamic nature of team leadership and the fact that it can originate from multiple sources at the same time. In this thesis I therefore apply Morgeson, DeRue and Karam’s (2010) view on the role of team leadership; the person who assumes responsibility for satisfying the team’s needs to be successful is the person who takes on the role of team leadership. I further address in the next section multiple elements of vertical leadership that have an influence on shared leadership or the shared leadership-team performance relationship.
Empowering Vertical Leadership
Individual followers could ‘take’ the autonomy and leadership needed for shared leadership behavior based on the available discretion within the team. This however could negatively impact trust and respect within the team and thereby potentially negatively impact both shared leadership behavior and team performance (Srivastava, Bartol and Locke 2006). Empowering leadership behavior by the vertical leader on the other hand ‘gives’ and ‘shares’ the autonomy, power, and leadership within a team and thereby increases followers’ intrinsic motivation. According to Zhang and Bartol (2010) empowering vertical leadership can be conceptualized as leadership behavior displaying and communicating confidence towards followers on high performance, focusing on the importance of the work that needs to be done, participating followers in decision making, and removing formal, administrative, and bureaucratic limitations. These leadership behaviors are important to followers in teams as they increase follower confidence and enable them to exercise influence on team behavior and thereby influence team performance (Mathieu et al. 2008).
It might appear that the often mentioned transformational leadership style and empowering leadership style have a significant overlap as leadership constructs. Upon further inspection, however, one sees that although both leadership styles are strongly related, they are different in their focal points. Transformation leadership on the one hand emphasizes leadership over management, transformation, inspiration, and vision, creating a willingness to go above and beyond the call of duty (Den Hartog and Koopman 2001). Empowering leadership on the other hand is all about motivating people by interacting in shaping team’s and individual’s performance (Mathieu et al. 2008). Next to that, transformational leadership typically, but not exclusively, manifests itself before empowering leadership, facilitating empowering leadership. I therefore expect empowering leadership behavior and not transformational leadership to facilitate the development of shared leadership behavior within a team as empowering leadership behavior will motivate team members to exercise influence (Mathieu et al. 2008). This collective influence is the essence of shared leadership behavior (Carson, Tesluk and Marrone 2007) and teams that are empowered have high levels of ability and autonomy in executing their tasks (Srivastava, Bartol and Locke 2006). Pearce and Manz (2005, p. 134)
14
state: “Shared leadership could be considered a case of fully developed empowerment in teams”. I therefore propose:
Hypothesis 3. Empowering vertical leadership positively facilitates shared leadership behavior within teams.
Directive Vertical Leadership
Pears and Sims (2002) define directive leadership based on leadership behavior that is expressed by the assignment of goals and giving commands and instructions, a form of vertical formal leadership that provides followers with clear focus, structure, and direction. Directive vertical leadership can be effective in situations where shared leadership is not well positioned to perform leadership tasks (Pearce and Barkus 2004).
Directive vertical leadership is different from shared leadership among others in the distribution of the leadership role and the approach to followers. Directive leadership centralizes the leadership at the formal hierarchical leader, while shared leadership distributes leadership within the team. Directive leadership minimizes conflict and ambiguity within a team (Burke et al. 2006a) and appears to be effective in crisis situations where quick decision making is needed (D’Innocenzo, Mathieu and Kukenberger 2014). With regards to approaching followers, directive leadership exercises influence over followers, while shared leadership motivates team members to exercise influence. Contrary to empowering vertical leadership, directive vertical leadership does not stimulate follower autonomy or trust (Trevelyan 2001), both important characteristics that need to be stimulated to exhibit shared leadership behavior in teams (D’Innocenzo, Mathieu and Kukenberger 2014). Empowering and directive leadership therefore appear to be opposites with regards to facilitating shared leadership behavior at followers. I therefore propose:
Hypothesis 4. Directive vertical leadership hinders shared leadership behavior within teams.
Boundary Spanning
Next to specific leadership roles, team members have a need for knowledge and information exchange to adequately complete their tasks and be able to adhere to corporate guidelines and policies (Aldrich and Herker 1977, Tortoriello, Reagans and McEvily 2012). Team members also want to be motivated by the acknowledgement that they are working on topics that are needed and in line with the company’s mission and vision, something that increases their team- and self-efficacy (Shea and Howell 1999). These team member needs require a flow of information to and from, and interaction with people outside the team that is referred to as the boundary spanning role.
Boundary spanning basically consists of two elements: information processing and external representation (Aldrich and Herker 1977). Proper information processing prevents an overload of information at team members. This is done by controlling the amount of information to followers and indicating what information requires followers’ attention and with what priority, a typical role for the vertical leadership role of the team leader. Next to that, Morgeson, DeRue and Karam’s (2010) state that the person in the shared leadership role is not well positioned to perform the following three leadership functions and that therefore other forms of leadership are needed at the same time.
15 1) The function of composing teams; informal internal leaders often do not have the status outside the shared leader’s team and the formal role required to change and set-up the required team composition. Formalized leadership roles are better positioned within the organization to perform this leadership function. 2) The function of providing resources; having access to, obtaining and providing information, finances, material, and people that are needed for the team in their task completion. Having access to adequate resources is a motivating factor (Morgeson, DeRue and Karam 2010) as it sends out a message of support towards team members. Team members in the shared leadership role however have limited and in most cases less access to these resources when compared to formal leaders. Organizations often structure the distribution of resources via formal hierarchies that shared leaders are not a member of. Formalized leadership roles are therefore better positioned within the organization to perform this leadership function. 3) The function of challenging the team; challenging performance, assumptions, methods, and processes might not be the best way to accomplish team tasks when the leader is a member of that same team and is also involved in the execution of the tasks. This phenomena will be further emphasized when task complexity is high as in those cases being able to challenge for example assumptions requires a certain level of emotional and content distance to the work in order to be able to come up with different ways of approaching the work, challenges, and activities (Pearce and Barkus 2004). A leader should be able to challenge followers’ status-quo, but shared leadership behavior requires acceptance by followers of the shared leader. That informal role can become compromised when a shared leader challenges followers on the status-quo. In conclusion, formal leadership characteristics such as boundary spanning are better than shared leadership positioned to perform these three leadership functions.
The information processing element of boundary spanning ensures that required information for task completion is available to team members. The external representation element of boundary spanning is a role in which the team is represented towards the environment outside of the team. It allows team members to focus on their tasks and to contribute to the team’s legitimacy perception within the organization (Aldrich and Herker 1977). This external representation role is a leadership task for which vertical leadership is better positioned than shared leadership (Pearce and Barkus 2004). Both elements of the team leader’s boundary spanning role should increase follower’s confidence and enable them to exercise influence on team behavior (Mathieu et al. 2008), thereby improving the shared leadership-team performance relationship. I therefore propose:
Hypothesis 5. Team leader boundary spanning activities as perceived by team members moderate the shared leadership-team performance relationship, such that when boundary spanning increases, the relationship between shared leadership and team performance is enhanced.
The effectivity of the boundary spanning role within a team in relation to the shared leadership-team performance relationship is among others dependent on the level of task complexity that team members are faced with (Aldrich and Herker 1977). As described in the task complexity section, this however implies a complex and hard to predict multi-way interaction. I have therefore also included this topic in the exploratory research question that I use as a basis for my exploratory research instead of defining additional hypotheses on the boundary spanning based interactions.
16
Conceptual Model
The conceptual model of this thesis is visualized in Figure 1 and also includes an overview of the five hypotheses.
Figure 1. The conceptual model including hypotheses.
Exploratory Research
The influence of and interaction between task complexity, boundary spanning and the described vertical leadership elements on the shared leadership-team performance relationship have, next to the mentioned hypothesis for the individual interactions above, an expected multi-factor interaction with each other. This interaction makes it hard to predict the outcome in various scenarios and define hypotheses for these scenarios. I will therefore, next to testing my hypotheses, perform exploratory research to investigate the various effects of the coexistence of shared leadership, empowering leadership, directive leadership, task complexity, and boundary spanning in relation to team performance. The next chapter addresses the research design for both the hypothesis testing and the exploratory research.
H5 H3 Shared Leadership Team Performance Contextual Factors: 1) Task Complexity (H2) 2) Boundary Spanning (H5) Facilitating Factors:
1) Empowering Vertical Leadership (H3)
2) Directive Vertical Leadership (H4)
H1 H4
17 CHAPTER 3. RESEARCH DESIGN
“Isn’t it astonishing that all these secrets have been preserved for so many years just so that we could discover them!” – Orville Wright, June 7, 1903 (Smithsonian National Air and Space Museum 2015).
The research that was conducted was based on a deductive approach to test the theory (Saunders, Lewis and Thornhill 2012) and my hypothesis mentioned in chapter 2. I applied a quantitative research design using the survey methodology to perform a cross-sectional examination of the relationships between the variables as described in the hypothesis and in the exploratory research. According to De Leeuw, Hox and Dillman (2008) “a survey can be seen as a research strategy in which quantitative information is systematically collected from a relatively large sample taken from a population”. The survey methodology therefore allowed me to collect data based on standardized survey questions to describe and infer the relationships between my variables. I used the remainder of this chapter to describe the sample, measures, and data analysis used in my research.
Sample
The studied sample included 125 teams (n = 290) consisting of 125 managers and 165 followers all from the multinational ASML, a manufacturer of complex lithography systems for the semiconductor industry. ASML employees often work in teams and are subjected to high and varying levels of task complexity in more than 70 locations in 16 countries and with 84 nationalities, strengthening the likelihood that results on shared leadership, team performance, task complexity, empowering leadership, directive leadership, boundary spanning, and the interaction between these measures would be generalizable to non-ASML populations.
I selected the sample within ASML based on convenience sampling as ASML Corporate Communications indicated that a number of departments should be excluded from the survey. ASML Corporate Communications limits the number of surveys being sent out within ASML and indicates to survey owners inside and outside ASML what teams and individuals should be excluded in conducting a certain survey. I approached all the members from the ASML teams that ASML Corporate Communications did not exclude and requested them via email to participate in the survey. Appendix A contains the text of the survey invitation email. ASML Human Resources provided me with generic survey participant data that included participant’s email, office location (country in which the participant’s office was located), the name of the participant’s manager, the name of the participant’s department, and the sector that the team is part of (development & engineering, operations, marketing and sales, IT, corporate, and ASML Center of Excellence). These items were, among others, used as control variables.
The survey, consisting of one questionnaire for managers and one questionnaire for followers, was designed and distributed via Qualtrics (www.qualtrics.com) as an internet web-based self-administered survey. Data from managers was collected using a dedicated questionnaire focusing on boundary spanning and team performance and was sent out on April 07, 2015. Data from team members was collected using a dedicated questionnaire focusing on task complexity, empowering leadership, directive leadership, and shared leadership and was sent out on April 07, 2015. The survey, including both questionnaires, was closed three weeks later on April 24, 2015. The survey response rate from team managers was 29% (202 of 686) and the rate from team members was 4% (165 of 4,107). I excluded responses from team
18
managers who’s followers did not participate in the survey. In total 125 teams were included in the analysis out of the 686 teams that I initially approached.
Teams ranged in team member size from five to 52 with a mean team member size of 17.48 (SD = 11.54). The average organizational tenure of managers was 11.77 years (SD = 7.02) and the average team tenure was 3.64 years (SD = 3.58). The average organizational tenure of followers was 9.14 years (SD = 7.47) and the average team tenure was 3.83 years (SD = 4.31). I received completed surveys from 227 people from the Netherlands, two from Ireland, 39 from Taiwan, 26 from Korea, seven from China, two from Japan, two from Singapore, and 60 from the United States of America. Ages ranged from 20 to 68 with a mean age of 42.28. 8.17% of the sample members were female.
Measures and data analysis
All questions that were addressed in both the manager and the follower questionnaires were derived from English studies, no translations were made. All measures, except for the control variables, were collected using a 5-point Likert scale (Likert 1932) to optimize the internal consistency (coefficient alpha reliability) of the scale (Hinkin 1998). The individual measures are described below. The questions and response options offered to both manager and follower respondents are included in appendix B. The Cronbach alpha procedure (Saunders, Lewis and Thornhill 2012) was used to assess the internal consistency of each of the used scales, and is mentioned below per measure. It is also included in the correlation matrix, table 2 in chapter 4.
Team performance: Team performance was measured using the scale of Mathieu et al. (2008). I used 16 subjective performance items (Cronbach’s alpha = .82) in the manager questionnaire to assess team performance, where higher values indicated greater levels of team performance. According to Wall et al. (2004) subjective performance measures can be interpreted as equivalent to objective measures. Limited objective team performance data was available and I therefore used the subjective measures. None of the 16 items were reverse coded. I conducted a principal axis factor analysis on the 16 team performance items with orthogonal rotation (varimax). The Kaiser-Meyer-Olkin (KMO) measure verified the sampling adequate for the analysis, KMO = .73, but three KMO values for individual items (question numbers 18, 28 and 30) were below the acceptable limit of .50 (Field 2013). I therefore removed these three questions from the team performance analysis and ran the analysis again with the remaining 13 items (Cronbach’s alpha = .80). The Kaiser-Meyer-Olkin measure verified the sampling adequate for the analysis, KMO = .78, and all KMO values for individual team performance items were greater than .60, which is above the acceptable limit. I also ran an initial analysis to obtain eigenvalues for each factor in the data. Four factors had eigenvalues of over Kaiser’s criterion of 1 and in combination explained 58.33% of the variance. The scree plot was ambiguous and showed inflexions that would justify retaining either two or four factors. I decided to retain two factors as they already explained 40.84% of the variance. Table 1 shows the factor loadings after rotation. The team performance items that cluster on the same factor suggest that factor 1 represents planning (question numbers 21, 23, and 24) and factor 2 performance abilities (question numbers 20, 22, and 25). I therefore averaged the scores of the two team performance items per factor, planning (Cronbach’s alpha = .67) and performance abilities (Cronbach’s alpha = .64), to yield two different team performance scores. An example questionnaire item was “please compare the performance and effectiveness of your team with
19 the performance and effectiveness of teams that perform similar tasks. Please do so regarding the following dimension: efficiency”.
Table 1
Summary of Exploratory Factor Analysis Results for Team Performance (n = 125)
Rotated Factor Loadings
Item Planning Performance Abilities Question # 16 .12 -.08 Question # 17 .29 .28 Question # 19 -.02 .34 Question # 20 .19 .60 Question # 21 .74 .09 Question # 22 -.01 .78 Question # 23 .72 -.04 Question # 24 .73 .31 Question # 25 .37 .68 Question # 26 .05 .09 Question # 27 .24 .22 Question # 29 .04 .13 Question # 31 .51 .10 Eigenvalues 3.95 1.36 % of variance 30.37 10.47
Note. Factor loadings over .40 appear in bold.
Team viability: Team viability was measured as an additional outcome variable using the scale of Marrone, Tesluk and Carson (2007). I used three items (Cronbach’s alpha = .83) in the follower questionnaire to assess the level of team viability, an indicator of team effectiveness via team member’s willingness and ability to work together as a team in the future (Marrone, Tesluk and Carson 2007). None of the items was reverse coded. I averaged the scores of the team viability items of all team members within one team to yield a single team viability score per team. An example questionnaire item was “please indicate the extent to which you agree with the following statement about your team: team members find being a member on this team a very satisfying experience”.
Team commitment: Team commitment was measured as a second additional outcome variable using the scale of Kirkman and Rosen (1999). I used three items in the follower questionnaire to assess the level of team commitment, but removed one (question number 63) because of the positive and relatively high impact on overall reliability (Cronbach’s alpha increased from .66 to .83) and no impact on the theoretical coverage. Team commitment is included as it also is an outcome indicator of team effectiveness and team attitude (Kirkman and Rosen 1999). None of the items was reverse coded. I averaged the scores of the team commitment items of all team members within one team to yield a single team commitment score per team. An example questionnaire item was “please indicate the extent to which you agree with the following statement about your team: in our team we are loyal to each other”.
20
Shared leadership: Shared leadership was measured in three different ways. The first measurement was done using a social network approach (Mayo, Meindl and Pastor 2003). I used two items in the follower questionnaire to have every team member rate his or her team peers: “how many colleagues are in your team (excluding your manager and yourself)? And for each of your colleagues, please answer the following question: does your team rely on this individual for leadership [yes or no]?”. The shared leadership score per team was calculated by summing all questionnaire values (yes = 1, no = 0) per team and dividing that outcome by the total number of possible ties among followers within that team, resulting in shared leadership density scores that ranged from 0.00 (no shared leadership within the team) to 1.00 (very high levels of shared leadership within the team). This shared leadership measurement was however excluded from the analysis. All team members needed to participate in the survey, but they did not, in order to calculate the density score that would represent the level of shared leadership within the team. The average number per team of people from the teams that participated was 1.32 while the average team size was 17.48.
Shared leadership was also measured using the scales of Hoch and Kozlowski (2014) and Hiller, Day and Vance (2006). I used 29 items, 14 based on Hoch and Kozlowski (2014) and 15 based on Hiller, Day and Vance (2006), in the follower questionnaire to assess shared leadership, where higher values indicated greater levels of shared leadership. One of the 29 items was reverse coded. In case of the reverse coded items a relatively high score indicates relatively low levels of shared leadership. I first transformed the reverse coded scores by recoding it into a different variable (transforming one to five, two to four, three to three, four to two, and five to one) and then averaged the scores of the Hoch and Kozlowski items (Cronbach’s alpha = .82) of all team members within one team to yield a shared leadership measure 2 score per team. I also averaged the scores of the Hiller, Day and Vance items (Cronbach’s alpha = .92) of all team members within one team to yield a shared leadership measure 3 score per team. An example questionnaire item was “please indicate the extent to which you agree with the following statements about your team: in regular sequences we take the time to figure out how to improve the team processes in our team”.
Task complexity: Task complexity was measured using the scales of Early (1985) and Wood (1986). I used four items (Cronbach’s alpha = .78) in the follower questionnaire to assess the level of task complexity within the team, where higher values indicated greater levels of task complexity. None of the items was reverse coded. I averaged the scores of the task complexity items of all team members within one team to yield a single task complexity score per team. An example questionnaire item was “Please indicate the extent to which you agree with the following statement: our tasks require us to coordinate many different things at the same time”.
Empowering leadership: Empowering leadership was measured using the scale of Kirkman and Rosen (1997). I used 18 items (Cronbach’s alpha = .92) in the follower questionnaire to assess the level of follower perceived empowering leadership from the manager within the team, where higher values indicated greater levels of empowering leadership. One of 18 the items was reverse coded. In case of the reverse coded item a relatively high score indicates relatively low levels of empowering leadership. I first transformed the reverse coded scores by recoding it into a different variable (transforming one to five, two to four, three to three, four to two, and five to one) and then averaged the empowering leadership scores of all team members to yield a single empowering leadership score per team. An example
21 questionnaire item was “please indicate to what extent you agree or disagree with the following statement: in general, my team manager gives my team many responsibilities”.
Directive leadership: Directive leadership was measured using the scale of Zhang and Bartol (2010). I used three items (Cronbach’s alpha = .62) in the follower questionnaire to assess the level of directive leadership of the team manager, where higher values indicated greater levels of directive leadership. None of the items was reverse coded. I averaged the directive leadership scores of all team members to yield a single directive leadership score per team. An example questionnaire item was “please indicate to what extent you agree or disagree with the following statement: in general my manager gives the team instructions on how to carry out tasks”.
Boundary spanning: Boundary spanning was measured using the scale of Ancona and Caldwell (1992). I used 15 items (Cronbach’s alpha = .72) in the manager questionnaire to assess the level of boundary spanning. None of the items was reverse coded. I averaged the scores of the boundary spanning items per manager to yield a single boundary spanning score per team. An example questionnaire item was “please indicate how often the following is applicable: how often do you obtain things the team needs from other groups or people in the company?”.
Control variables: Next to the earlier mentioned control variables, I also included team size, gender, age, organizational and team tenure (in years), employment type (full-time or part-time contract), transformational leadership, and task interdependence as control variables to capture some of the forces that are related to the team and the ASML environment. Task interdependence, an indicator of how dependent individual followers are upon others within the group to effectively perform their tasks (Saavedra, Earley and Van Dyne 1993), was taken into account as it is known to moderate the indirect influence of empowering leadership on team performance (Kirkman et al. 2004). It is also closely linked to task complexity (Saavedra, Earley and Van Dyne 1993). Task interdependence was measures using a combination of three scale items from Kirkman et al. (2004) and five scale items based on Kirkman et al. (2004). I excluded the average of the eight items from further analysis due to low internal consistency (Cronbach’s alpha = .31) and used the average score per team of the three Kirkman et al. (2004) scale items (Cronbach’s alpha = .64) for further analysis.
I combined the manager and follower responses by adding the responses from a manager from a specific team to all of the responses of the followers from that team, using Microsoft Excel. I used the Statistical software Package for Social Sciences (SPSS) to perform the statistical analysis at a team level. The results of this analysis are presented in the next chapter.
22
CHAPTER 4. RESULTS
“To consult a statistician after an experiment is finished is often merely to ask him to conduct a post mortem examination. He can perhaps say what the experiment died of.” (Fisher 1938).
Factor analysis and hypothesis testing
I have used table 2 to present the means, standard deviations, and correlations for the independent, dependent, and control variables of this study. I found 15 significant relationships with a large effect (r ≤ -.50 or r ≥ .50) between the mentioned variables. There was however no significant relationship between the main independent variable shared leadership and the main dependent variable team performance based on the correlation analyses of their measures; shared leadership measure 2 and 3, planning, and performance abilities. Shared leadership measure 2 was not significantly related to planning, r = -.13, p = .162, and also not to performance abilities, r = .06, p = .497. Shared leadership measure 3 was not significantly related to planning, r = -.09, p = .301, and also not to performance abilities, r = .11, p = .231. The 15 significant relationships with a large effect are listed below.
Team viability was significantly related to team commitment, r = .51, p < .000, shared leadership measure 2, r = .70, p < .000, shared leadership measure 3, r = .62, p < .000, empowering leadership, r = .55, p < .000, and transformational leadership, r = .57, p < .000. Team commitment was significantly related to shared leadership measure 2, r = .51, p < .000, and shared leadership measure 3, r = .50, p < .000. Shared leadership measure 2 was significantly related to shared leadership measure 3, r = .76, p < .000, empowering leadership,
r = .52, p < .000, and transformational leadership, r = .57, p < .000. Shared leadership measure
3 was significantly related to empowering leadership, r = .50, p < .000. Empowering leadership was significantly related to transformational leadership, r = .75, p < .000. Directive leadership was significantly related to transformational leadership, r = .50, p < .000. Age followers was significantly related to organizational tenure followers, r = .62, p < .000. Organizational tenure followers was significantly related to team tenure followers, r = .53, p < .000.
Hypothesis 1 states that shared leadership is positively related to team performance. Hierarchical multiple regression was performed to test this hypothesis. I investigated the ability of shared leadership measure 3 (highest internal consistency of the two remaining shared leadership measures) to predict levels of team performance in the form of the dependent variables planning and performance abilities, after controlling for team size, gender, age, organizational tenure, employment type, office location, and sector. In the first step of the regression the control variables were entered. The control variables model was not statistically significant (F (13, 111) = 1.68; p = .07). After entry of shared leadership measure 3 at step 2 the total variance explained by the shared leadership and planning model as a whole was not statistically significant (p = .16). In the final model one out of 14 predictor variables was statistically significant; team tenure managers (β = .21; p < .05). After entry of shared leadership measure 3 at step 2 the total variance explained by the shared leadership and performance ability model as a whole was also not statistically significant (p = .84). In the final model one out of 14 predictor variables was statistically significant; team tenure managers (β = .26; p < .01). Thus, hypothesis 1 was not supported. Tables 3 and 4 presents these results.
23 Ta bl e 2 De sc riptiv e S tatist ics a nd C orr elations Va ri ab le M ea n SD 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 1. P la nning 3. 35 0. 53 (. 67) 2. P er fo rm a n ce a b il it ies 3. 55 0. 53 .40 * * (. 64) 3. T ea m viab il it y 3. 66 0. 65 -. 08 -. 00 (. 83) 4. T ea m c o mm it m e n t 3. 77 0. 68 -. 12 .14 .51 * * (. 83) 5. S h ar ed le ade rs hi p m ea su re 2 3. 47 0. 42 -. 13 .06 .70 * * .51 * * (. 82) 6. S h ar ed le ade rs hi p m ea su re 3 3. 61 0 .53 -. 09 .11 .62 * * .50 * * .76 * * (. 92) 7. T as k co m p le xi ty 3. 72 0. 64 -. 02 -. 13 .08 -. 07 .08 -. 15 (. 78) 8. E m powe ring l ea de rs hip 3. 65 0. 52 .01 .06 .55 * * .28 * * .52 * * .50 * * .03 (. 92) 9. Dir ec tiv e le ade rs hi p 3 .51 0. 63 .06 .15 .28 * * .15 .39 * * .38 * * .00 .37 * * (. 62) 10. B o unda ry s pa nning 3. 51 0. 35 .10 .03 .02 -. 14 -. 02 .04 -. 18 * .03 .03 (. 72) 11. T ea m s ize 17 .48 11 .54 .01 .03 .13 .14 .06 .10 .13 .08 .05 -. 04 12. Ge n de r fo ll o we rs 1. 17 0. 37 .19 * .06 -. 03 -. 07 .01 -. 01 -. 16 -. 08 .13 .18 * -. 12 13. Ge n de r m a n a g er s 1. 04 0. 20 .04 -. 08 -. 02 -. 08 -. 09 -. 13 .03 .03 -. 02 .00 -. 09 .02 14. Ag e fo ll owe rs 40. 37 9. 16 -. 20 * -. 08 .02 .10 -. 01 -. 04 .07 -. 19 * -.0 8 .05 -. 07 -. 09 -. 02 15. Ag e m a n a g e rs 43. 53 6. 17 .06 .08 .16 .17 .16 .19 * .01 .02 .10 .01 .06 -. 09 -. 05 .05 16. Or g a ni za ti o n a l te nur e fo ll owe rs 9. 14 7. 47 -. 18 -. 12 -. 07 .00 -. 10 -. 17 .24 * * -.25 * * -. 09 .01 .11 -. 13 -. 01 .62 * * - .05 17. Or g a ni za ti o n a l te nur e m a n a g er s 11. 77 7. 02 .13 .12 .02 .07 -. 03 .01 -. 06 .01 .11 -. 14 .24 * * -. 01 -. 03 .01 .18 -. 08 18. T ea m te nu re f o ll owe rs 3. 83 4. 31 .02 .08 -. 01 .02 -. 04 .01 .03 -.29 * * .06 .11 .14 .07 .00 .40 * * - .08 .53 * * -. 07 19. T ea m te nu re m a n a g e rs 3. 64 3. 58 .19 * .28 * * .04 .02 .03 .08 .05 .09 .00 -. 04 .03 .04 .08 .11 .07 .10 .20 * .15 20. C on tr ac t ty pe fo ll ower s 1. 04 0. 17 .01 .01 -. 20 * .05 -. 07 .05 .00 -. 12 .09 -. 16 -. 05 -. 09 -. 05 -. 13 .06 -. 02 .06 -. 05 .06 21. C on tr ac t ty pe m a n a g er s 1. 02 0. 15 -. 04 -. 06 .09 .15 .06 .09 -. 01 .03 .02 -. 17 .01 -. 07 .24 * * .00 .15 .00 .20 * -. 04 -. 06 - .04 22. T ra n sf o rm at io n a l le ade rs hi p 3. 54 0. 59 .00 .08 .57 * * .29 * * .57 * * .47 * * .06 .75 * * .50 * * -. 03 -. 03 -. 01 -. 06 -. 13 .08 -. 21 * .01 -.22 * .00 - .02 .05 (. 90) 23. T as k int er de pe n de n ce 3. 46 0. 38 .09 .05 .18 * .03 .11 .09 .35 .10 -. 38 .01 .18 * -. 16 .11 .13 .08 .09 .06 .01 .04 .08 .03 .16 (. 64 ) 24. O ff ic e loc at io n 2. 93 2. 65 .08 .02 -. 01 -. 09 -. 08 -. 09 -. 04 .09 -. 05 .08 .10 .09 -. 15 .04 .06 .05 -.20 * .11 -.19 * - .14 - .06 .04 .03 25. S ec tor 2. 14 1. 31 -. 05 -. 19 * -. 07 -. 13 -. 12 -. 12 .08 -. 08 -. 11 .07 .00 -. 05 -. 15 .17 .04 .05 -.21 * -. 09 -. 13 - .05 - .10 -. 06 -. 04 .09 * p < . 05 . ** p < . 01
24 Table 3
Results of hierarchical regression analysis of planning and shared leadership
Variable R R2 R2 change B SE β t Step 1 .41 .17 Team size .00 .00 -.05 -0.53 Gender followers .18 .13 .13 1.39 Gender managers .18 .25 .07 0.72 Age followers -.01 .01 -.23 -1.88 Age managers .00 .01 .07 0.80
Organizational tenure followers -.01 .01 -.10 -0.78
Organizational tenure managers .01 .01 .15 1.48
Team tenure followers .02 .01 .13 1.15
Team tenure managers .03 .01 .20* 2.10
Employment type followers -.02 .28 -.01 -0.08
Employment type managers -.20 .32 -.06 -0.61
Office location .02 .02 .15 1.62 Sector .02 .04 .06 0.59 Step 2 .43 .18 .02 Team size .00 .00 -.03 -0.29 Gender followers .18 .13 .13 1.40 Gender managers .12 .25 .04 0.47 Age followers -.01 .01 -.21 -1.70 Age managers .01 .01 .09 1.01
Organizational tenure followers -.01 .01 -.14 -1.10
Organizational tenure managers .01 .01 .13 1.24
Team tenure followers .02 .01 .14 1.33
Team tenure managers .03 .01 .21* 2.22
Employment type followers .00 .28 .00 -0.11
Employment type managers -.14 .32 -.04 -0.42
Office location .03 .02 .14 1.45
Sector .02 .04 .04 0.39
Shared leadership measure 3 -.13 .09 -.13 -1.41
25 Table 4
Results of hierarchical regression analysis of performance abilities and shared leadership
Variable R R2 R2 change B SE β t Step 1 .40 .16 Team size .00 .00 -.01 -0.08 Gender followers .01 .13 .00 0.03 Gender managers -.28 .25 -.10 -1.12 Age followers .00 .01 -.03 -0.24 Age managers .01 .01 .08 0.82
Organizational tenure followers -.01 .01 -.19 -1.52
Organizational tenure managers .00 .01 .04 0.39
Team tenure followers .02 .01 .14 1.28
Team tenure managers .04 .01 .27** 2.85
Employment type followers -.06 .28 -.02 -0.20
Employment type managers -.17 .32 -.05 -0.52
Office location .01 .02 .07 0.72 Sector -.06 .04 -.15 -1.59 Step 2 .40 .16 .00 Team size .00 .01 -.01 -0.11 Gender followers .01 .13 .00 0.03 Gender managers -.27 .26 -.10 -1.05 Age followers .00 .01 -.03 -0.26 Age managers .01 .01 .07 0.78
Organizational tenure followers -.01 .01 -.18 -1.41
Organizational tenure managers .00 .01 .04 0.42
Team tenure followers .02 .01 .14 1.25
Team tenure managers .04 .01 .26** 2.81
Employment type followers -.06 .28 -.02 -0.21
Employment type managers -.18 .33 -.05 -0.54
Office location .01 .02 .07 0.73
Sector -.06 .04 -.15 -1.53
Shared leadership measure 3 .02 .10 .02 0.21
26
Hypothesis 2 states that task complexity moderates the shared leadership-team performance relationship. A moderated regression analyses was performed to test this hypothesis. I examined the incremental contribution of the relation between shared leadership and team performance for various levels of task complexity, after controlling for the main effects of shared leadership, task complexity, and the control variables team size, gender, age, organizational tenure, employment type, transformational leadership, office location, and sector. Shared leadership and task complexity were mean centered prior to the analysis. The model with shared leadership measure 3 as independent variable, planning as dependent variable, and task complexity as moderator (F (17, 107) = 1.50; p = .11) was not statistically significant. The interaction effect (p = .23) was also not statistically significant. The model with shared leadership measure 3 as independent variable, performance ability as dependent variable, and task complexity as moderator (F (17, 107) = 1.47; p = .11) was not statistically significant. The interaction effect (p = .12) was also not statistically significant. No further analysis was performed as none of the models and interaction effects were statistically significant. Thus, hypothesis 2 was not supported.
Hypothesis 3 states that empowering leadership positively facilitates shared leadership while hypothesis 4 states that directive leadership hinders shared leadership. Hierarchical multiple regression was performed to test these hypotheses in a combined model. I investigated the ability of empowering leadership and directive leadership to predict levels of shared leadership in the form of the variable shared leadership measure 3, after controlling for team size, gender, age, organizational tenure, employment type, transformational leadership, office location, and sector. In the first step of the regression the control variables were entered. This model was statistically significant (F (14, 110) = 4.45; p < .001), explaining 36.2% of the variance in shared leadership. After entry of empowering leadership and directive leadership at step 2 the total variance explained by the model as a whole was 42.7% (F (15, 109) = 5.20; p < .001). The introduction of empowering leadership and directive leadership explained an additional 6.5% in variance in shared leadership, after controlling for the mentioned control variables (R2 Change = .065; F (2, 108) = 10.34; p < .05). In the final model four out of 16 predictor variables were statistically significant. Gender managers (β = -.17; p < .05), organizational tenure followers (β = -.27; p < .05), and team tenure followers (β = .21; p < .05) were the control variables with a significant relationship with shared leadership. These indicate on the one hand that a female team manager and employees that spent a longer time in an organization are a negative predictor of shared leadership behavior in a team. On the other hand, results indicate that employees that spent a longer time in a team is a positive predictor of shared leadership behavior in a team. Next to that empowering leadership had a direct relationship with shared leadership (β = .39; p < .01), making it a positive predictor of shared leadership in a team, while no statistically significant direct relationship was found between directive leadership and shared leadership. Thus, hypothesis 3 was supported and hypothesis 4 was not supported. Table 5 presents these results.
Hypothesis 5 states that boundary spanning moderates the shared leadership-team performance relationship. A moderated regression analyses was performed to test this hypothesis. I examined the incremental contribution of the relation between shared leadership and team performance for various levels of boundary spanning, after controlling for the main effects of shared leadership and boundary spanning and previously mentioned control variables. Shared leadership and boundary spanning were mean centered prior to the analysis. The model with shared leadership measure 3 as independent variable, planning as dependent
27 Table 5
Results of hierarchical regression analysis of shared leadership measure 3 and empowering leadership and directive leadership
Variable R R2 R2 change B SE β t Step 1 .60 .36 Team size .01 .00 .17* 1.96 Gender followers .01 .12 .01 0.10 Gender managers -.40 .22 -.15 -1.85 Age followers .01 .01 .16 1.48 Age managers .01 .01 .13 1.65
Organizational tenure followers -.02 .01 -.29** -2.60
Organizational tenure managers -.01 .01 -.16 -1.78
Team tenure followers .02 .01 .19 1.92
Team tenure managers .01 .01 .06 0.70
Employment type followers .19 .25 .06 0.75
Employment type managers .37 .28 .11 1.30
Transformational leadership .41 .07 .45** 5.69 Office location -.03 .02 -.15 -1.77 Sector -.05 .03 -.11 -1.36 Step 2 .65 .43 .07 Team size .01 .00 .11 1.28 Gender followers .03 .11 .02 0.24 Gender managers -.46 .21 -.17* -2.17 Age followers .01 .01 .17 1.72 Age managers .01 .01 .14 1.82
Organizational tenure followers -.02 .01 -.27* -2.52
Organizational tenure managers -.01 .01 -.14 -1.63
Team tenure followers .03 .01 .21* 2.19
Team tenure managers .00 .01 .03 0.31
Employment type followers .31 .24 .10 1.28
Employment type managers .40 .27 .12 1.49
Transformational leadership .10 .11 .11 0.92 Office location -.02 .02 -.09 -1.14 Sector -.04 .03 -.10 -1.20 Empowering leadership .39 .13 .39** 3.14 Directive leadership .10 .08 .12 1.35 * p < .05. ** p < .01
