The dark side of leadership inclusiveness : moderated mediation effects on employee burnout in teams

(1)

The dark side of leadership inclusiveness:

Moderated mediation effects on employee burnout in teams

Joëlle Coenen 11420464 June 12, 2017 – Master Thesis MSc Business Administration – Leadership and Management track

Amsterdam Business School – University of Amsterdam Mw. R. E. van Geffen MSc

(2)

2

Statement of originality

This document is written by Student Joëlle Coenen 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)

3

Abstract

Due to the fact that an increasingly number of employees is suffering from burnout, the role of leadership is more important than ever before. Therefore the aim of this study is to investigate the dark side of inclusive leadership by exploring both contextual factors and personal factors of employees, which could negatively impact the relation between leadership and employee burnout. 215 surveys of employees in different organizations located in the Netherlands were gathered by means of one questionnaire. PROCESS analysis was used to test the proposed model. As expected, inclusive leadership was negatively related to burnout. Additionally, it was found that the leader-member exchange (LMX) fully mediates the relationship between inclusive leadership and burnout. This study also expected a moderating effect of task independence on between inclusive leadership and burnout, however the results did not support this conception. In addition, no support was found for the moderating mediation of emotional stability on the relationship between inclusive leadership, LMX and burnout. The findings imply that leaders who are inclusive, will have higher quality LMX relationships with their followers, which subsequently negatively impacts burnout. Thus, a dark side of leader inclusiveness due to contextual and personal factors was not found. Theoretical and practical implications are discussed, and limitations and future research directions are given.

Keywords: inclusive leadership; burnout; leader-member exchange; task independence;

(4)

4

Index of figures and tables

Figure 2.1 Research model of hypothesized relationships Table 4.1 Descriptive Statistics and Correlations

Table 4.2 Hierarchical Regression Model of Burnout (Disengagement) Table 4.3 Hierarchical Regression Model of Burnout (Emotional exhaustion) Table 4.4 Moderation of task independence (Disengagement)

Table 4.5 Moderation of task independence (Emotional exhaustion) Table 4.6 Moderated mediation with independent variable Disengagement Table 4.7 Moderated mediation with independent variable Emotional Exhaustion

Index of appendices

Appendix A Survey Instructions Appendix B Scale Items

(6)

6

1. Introduction

Burnout is a tedious malady of the world we live in today. It is even “the biggest occupational hazard of the twenty-first century”, which costs 60.000 euros on average per burnout sufferer (Leiter & Maslach, 2005, p.3; ‘’Arboned slaat alarm,’’ 2016). One in seven people in the Netherlands have burnout symptoms and this number, as well as the duration of burnouts, is rapidly increasing (‘’Arboned slaat alarm,’’ 2016; CBS, 2015). It makes clear that prevention of burnout needs more attention. Burnout replies to stressors at work, which can be a variety of things such as workload, work related conflicts and the relationship with the leader (Maslach, Schaufeli & Leiter, 2001). Although leaders do have influence on employee burnout, the precise relationship between leadership and employee burnout remains ambiguous (Halbesleben & Bowler, 2007; Hetland, Sandal & Johnsen, 2007).

Studies argue that the role of the leader and the strength of the relationship

between leaders and followers could impact employee well-being positively by appreciating followers’ work, creating a supportive work environment and being a role model (Skakon, Nielsen, Borg & Guzman, 2010; Shore et al., 2011). Nevertheless, leaders can also be a source of employee burnout, which could lead to friction between employees and their leader that has been acknowledged to result in employee burnout (Hetland et al., 2007; Maslach, Leiter & Schaufeli, 2008; Zopiatis & Constanti, 2010). Therefore, leaders should be able to provide their employees with sufficient resources, such as appreciating followers, listening to followers and giving support (Väänänen et al., 2003; Hollander, 2012; Bakker, Demerouti & Euwema, 2005). These actions could be seen as buffering mechanisms for employees to cope with, for instance, work load and

(7)

7 work pressure, which are related to burnout (Spreitzer, Lam & Fritz, 2010 in Saks &

Gruman, 2014; Väänänen et al., 2003).

The aforementioned characteristics a leader should exhibit are in line with these of a relatively new leadership style, inclusive leadership (Hollander, 2012). Inclusive leaders enable inclusive decision-making and value different kinds of team members and their viewpoints by creating an environment where individuals feel appreciated, inspired and supported (Nembhard & Edmondson, 2006). This study argues that inclusive leadership contributes to affective employee well-being and lowers employee stress by, for example, improving employees’ stress coping and building strong relationships with their followers (Skakon et al., 2010; Schaufeli & Enzmann, 1998, in Skakon et al., 2010). Thus, at first glance, inclusive leaders can be categorized as leaders, who positively impact employees as well as the organizational performance.

However, employee perceptions of a leader may vary per person due to, for

instance, different personalities or different contextual factors (Graham, 2009; Maslach et al., 2001; Bakker, Demerouti & Sanz-Vergel, 2014). Therefore the concept underpinning this study is that people do not work in isolation (Maslach et al., 2001). This subsequently means that benefits of inclusive leadership for employees with different personalities in different situations are disputable (Hollander, 2012).

From a different point of view, a leader could pose a so-called ‘’dark side’’,

which has become an important topic in the leadership discourse (Judge, Piccolo & Kosalka, 2009). Even charismatic and transformational leaders pose a dark side, since they are eager to reach their goals with their clear vision and drive, but may lose sight of personality and contextually related problems (Einarsen, Aasland & Skogstad, 2007; De

(8)

8 Hoogh & Den Hartog, 2009; Stone, Russell & Patterson, 2003).

The aim of this research is therefore two-folded: first, this study examines the

direct relationship between leader inclusiveness and employee burnout moderated by task independence and second, the research investigates the relationship between leader inclusiveness and employee burnout, moderated by emotional stability and mediated by the construct of Leader-member exchange (LMX). Based on the profound antecedents of employee burnout, the research question is:

To what extent is the direct relationship between leader inclusiveness and employee burnout mediated by LMX and moderated by emotional stability and task independence?

Focusing on the potential dark side of leadership inclusiveness results in both

theoretical and practical implications. As current research has exclusively focused on work characteristics in predicting burnout, this paper aims to incorporate both personality factors of employees and contextual factors in order to give some critical insights about leader inclusiveness (Langelaan, Bakker, Doornen & Schaufeli, 2006; Halbesleben & Buckley, 2004; Nembhard & Edmondson, 2006). To gain a deeper understanding of how contextual and personal factors interact with inclusive leadership, this study will specifically elaborate on task independence and emotional stability, as suggested by Wageman (1995) and Alarcon, Eschleman and Bowling (2009), respectively. As little is known about the role of followers in the leader-follower relationship, this research replies to the call of Kramer (2004 in Graham, 2009) that strong leader-member relationships may have a negative impact on employees’ well-being due to personal antecedents.

(9)

9

Leaders play an important role in preventing or causing employee burnout by

determining which tasks employees should execute and rating their performance for instance, which subsequently makes or breaks a leader’s success in terms of well-being (Macey & Schneider, 2008 in Saks & Gruman, 2014). Therefore, this study contributes to the understanding of how inclusive leaders influence individual-level outcomes such as employee burnout under certain circumstances, which is useful for organizations in selecting the right leader or in training current leaders in order to prevent employee burnout (Leiter & Maslach, 2005). Moreover, results of this study will contribute to reducing absenteeism and turnover costs by preventing employee burnout (Maslach et al., 2001).

This study is based on the following theories and key concepts: burnout as a

psychological phenomenon (Maslach et al., 2001), the job demands-resource model (JD-R) model (Bakker et al., 2005), Leader inclusiveness (e.g. Hollander, 2012; Nembhard & Edmonson, 2006; Mitchell et al., 2015), Leader-member exchange (LMX) theory (Gerstner & Day, 1997), the construct emotional stability of the Big-Five personality factors (Langelaan et al., 2006), and task independence (Wageman, 1995). These theories and concepts will be explained in the theoretical background in the next section.

The following study starts with an overview of the existing literature on burnout, LMX, task independence and emotional stability. Subsequently, the research methodology is described. Next, the results based on the data analyses are presented and interpreted. The discussion companies both theoretical and practical implications of this study, limitations and possible future research directions. Last, overall conclusions are drawn concerning all the different effects.

(10)

10

2. Theoretical background

The following section will explain the job demands-resources model, to elaborate on the concept of employee burnout. Next, leader inclusiveness is explained and subsequently, the relationship between burnout and inclusive leadership through LMX is described. Furthermore, the moderating effects of emotional stability and task independence will be specified.

2.1 Job Demands-Resources Model

In order to explain the concept of burnout, the complementary model of job demands-resources (JD-R), cannot be disregarded since research has highlighted that an overload of job demands and a lack of job resources is related to burnout (e.g. Bakker et al., 2005; Saks & Gruman, 2014). Therefore, the model is used to explain the development of employee burnout through two processes which both will be related to inclusive leadership. First, the model demonstrates that high job demands could lead to emotional exhaustion and second, that a lack of resources result in disengagement (Saks & Gruman, 2014). Many burnout scholars have studied this interplay of demands-resources in relation to positive and negative aspects of work, such as getting compliments and having stress respectively (Demerouti, Bakker, Nachreiner & Schaufeli, 2001). The JD-R model will be applied in this study since it is expected an inclusive leader to provide resources, such as, optimism, self-esteem and self-efficacy in order to cope with job demands (Xanthopoulou, Bakker, Demerouti & Schaufeli, 2009a in Saks & Gruman, 2014; Xanthopoulou, Bakker, Demerouti, & Schaufeli, 2009b in Saks & Gruman, 2014).

Job demands can be defined as “those physical, social, or organizational aspects of the job that require sustained physical or mental effort and are therefore associated

(11)

11 with certain physiological or psychological costs” (Demerouti et al., 2001, p. 501).

Examples of job demands are a lack of autonomy, recipient contact, work overload, work environment, role ambiguity, role conflict, and time pressure (e.g. Witt, Andrews & Carlson, 2004; Wright & Hobfoll, 2004; Demerouti et al., 2001). These job demands are strongly related to a loss of passion for the job, sleepiness and fatigue, which can be seen as a response to overwhelming job demands (Demerouti, et al., 2001; Maslach et al., 2001; Maslach et al., 2008). It is proposed that job demands are even the most crucial antecedent of burnout since an overload of these demands could result in for example chronic stress (Schaufeli & Bakker, 2004; Maslach et al., 2001; Demerouti et al., 2001). In order to cope with the job demands, job resources are needed, which is called ‘’the buffering effect’’ (Graham, 2009). An inclusive leader could be able to make such a buffering effect possible.

Researchers have investigated that a lack of job resources may cause employee

burnout (e.g. Graham, 2009; Maslach et al., 2001; Bakker et al., 2014). Job resources are defined as ‘’the tools we use to address job demands’’ (Halbesleben & Buckley, 2004, p. 863). The underlying mechanism is that job resources enable employees to cope with job demands and can lead to participation, satisfaction, motivation, inspiration and positive attitudes (Mulki, Jaramillo & Locander, 2006). Examples of job resources are rewards, job control, participating in decision-making, task variety and social support of the leader. These different facets could lead to engaged and healthy employees (Antonovski, 1987 in Demerouti et al., 2001). This approach is relevant for this study, since it will help to understand whether an inclusive leader is able to provide relevant resources to cope with the previously mentioned job demands.

(12)

12

Nowadays, employees are expected to go the extra mile in terms of, skills, time, efforts and flexibility, which can ask a lot of employees (“Burn-out groeiend probleem’’, 2015; Demerouti et al., 2001). However, employees receive less in return in terms of lifetime employments and career opportunities. The underlying mechanism responsible for affecting employee well-being is that job demands are not compensated with job resources, which could harm employee well-being (Maslach et al., 2001). Such an imbalance between job demands and job resources or even chronic job demands is disastrous since it could lead to job strain or even burnout. This subsequently results in absenteeism, higher turnover rates and related costs for organizations, which is of great practical importance for organizations (Bakker et al., 2014).

2.2 Burnout

The previous section is based on a relationship between the interplay of job demands-resources and employee burnout. The JD-R model constitutes the basis for understanding the evolvement of burnout, since.it demonstrates how job resources can buffer the impact of job demands on employee burnout (Bakker et al., 2005). It is interesting to note that while burnout is a well-known concept nowadays, the concept was initially vague, slippery and was lacking agreement about what it is and what it does for employees (Demerouti et al., 2001). Maslach was the first scholar who bundled all the burnout research and defined burnout as “a syndrome of emotional exhaustion, depersonalization, and reduced personal accomplishment that can occur among individuals” (1982, p.3). In addition, burnout is a work-related stress reaction of employees in a variety of occupations (Bakker, Demerouti & Schaufeli, 2002). Specifically, it is a response to a

(13)

13 combination of chronic interpersonal and emotional stressors at work, which can be

characterized by three related dimensions as already indicated in the definition: emotional exhaustion, cynicism and reduced personal accomplishment (Maslach et al., 2001).

Chronic and overwhelming job demands result in emotional exhaustion (Maslach

et al., 2008). This is the first dimension of employee burnout and refers to the feelings of overextension and depletion of someone’s emotional as well as physical resources (Maslach, 1982; Halbesleben & Demerouti, 2005). Emotional exhaustion is related to the level of stress someone copes with and is the most evident manifestation of burnout. However, the critical relationships people have with their work are neglected in this dimension (Maslach et al., 2001). When employees are exhausted, they create an emotional and cognitive distance towards colleagues. Furthermore, they are getting detached from their job, which subsequently leads to the second symptom of burnout:

cynicism (also labelled depersonalization) (Maslach et al., 2008). Cynicism arises

because employees are more able to manage their demands when these are considered as impersonal matters of work. The underlying mechanism is that employees use cognitive distancing by developing a cynical stand when they feel disheartened and exhausted (Maslach et al., 2001).

The third dimension is reduced personal accomplishment, which occurs when employees experience that their job efforts are no longer effective in accomplishing job responsibilities (Bakker, Emmerik & Euwema, 2006). Working conditions with overwhelming and long-lasting job demands that result in emotional exhaustion or cynicism are likely to decrease someone’s sense of effectiveness (Maslach et al., 2001). This is a sequential process that evolves slowly (Langballe, Falkum, Innstrand &

(14)

14 Aasland, 2006). Reduced personal accomplishment seems to arise from a disturbed

balance between job demands and job resources, whereas emotional exhaustion and cynicism emerge from social conflict or work overload (Maslach et al., 2001). To conclude, individuals who suffer from burnout are most likely tired out, feeling empty, are less concentrated, behave indifferently toward colleagues and clients, and they think they have failed in their performance (Maslach et al., 2008).

Nonetheless, this study will focus on disengagement and exhaustion, thereby excluding personnel accomplishment, based on the Oldenburg Burnout Inventory, (Langballe et al., 2006; Demerouti et al., 2003 in Langballe et al., 2006). Emotional exhaustion and cynicism are covered by exhaustion and professional accomplishment is replaced for disengagement in this two-factor model (Langballe et al., 2006). The underlying reason for this is that the dimensions emotional exhaustion and cynicism were found to collapse into one factor, namely emotional exhaustion (Leiter & Schaufeli, 1996 in Langballe et al., 2006). This means that physical, cognitive as well as affective aspects are covered in the dimensions exhaustion and disengagement (Leiter and Schaufeli, 1996, in Langballe et al., 2006). Additionally, professional efficacy was found as the weakest burnout dimension, where exhaustion and disengagement were found as the core elements of burnout (Langballe et al., 2006). Disengagement in this research is regarded as employees that distance themselves from their work (Demerouti et al., 2003).

In conclusion, an imbalance between job demands and job recourses could result

in burnout, which is a complicated mechanism instead of a simple individual stress reaction. The leader could impact most of previously mentioned job demands and job resources that can result in exhaustion and disengagement. Therefore, leadership might

(15)

15 play a major role in diminishing job demands and serve as a job resource by for example

supporting followers, giving performance feedback and being optimistic about employees’ work and goals (Halbesleben & Buckley, 2004; Maslach et al., 2001; Graham, 2009). In the next paragraph, the role of an inclusive leadership style in relation to the interplay of job demands-resources and burnout will be discussed.

2.3 Inclusive leadership

The role of the leader is significant in the sense-making of employees and will influence their behaviours directly, which means that leadership is a process of social influencing at its most fundamental level (O’Driscoll & Beehr, 1994; Day & Dragoni, 2015). This implies that leader attention to follower’s needs and interests is essential for employee well-being and for an organization to be effective (Hollander, 2012). A suitable leader is able to build and assist in encouraging a strong bond such as trust and dedication within a team and between the leader and followers (Hollander, 2012).

An ‘’inclusive process’’ where everyone is involved in and where followers are

treated with respect since leaders listen to them, is a leadership style that an ever-increasing diverse world needs (Eagly & Chin, 2010). A relatively new leadership style that evolved as a significant theme in the literature responds to these needs, called inclusive leadership (Hollander, 2009). This leadership style focuses on the active and participative role of the follower to include them in leadership tasks and does this by looking at what is needful in the environment (Hollander, 2012). While research on inclusive leadership is still in its early days, it has been discussed as the determining factor of performance in (diverse) work groups through open discussions and knowledge sharing (Shore et al., 2011).

(16)

16 Leader inclusiveness directly refers to diminishing power and status differences,

as it entails equality of team members on one hand and fosters inclusion in decision-making on the other hand (Nembhard & Edmondson, 2006). Specifically, this leadership style could be defined as behaviors through which leaders are open, accessible and available to their unit members (Edmondson, Kramer & Cook, 2004; Nembhard & Edmondson, 2006). Also, inclusive leaders respect followers’ diverse opinions and involve followers by listening (Hollander, 2012).

The aim of inclusive leaders is creating equality and horizontal relationships by facilitating advocating, learning, and dialogues and execute inclusive decision- and policy-making strategies in order to include everyone and to master the inhibiting effects of diversity (Nembhard & Edmondson, 2006). Inclusive leadership therefore leads to more knowledge sharing and open conversations because of the consultation with followers (Mitchel et al., 2015).

Prior research has found that inclusive leadership behaviour is related to positive outcomes, such as a psychological climate and team performance (Hirak, Peng, Carmeli & Schaubroeck, 2012; Mitchell et al. 2015). Additionally, inclusive leadership could affect the maintenance and employee well-being and it is even suggested that leaders have a moral responsibility to ensure this employee well-being (Graham, 2009). As noted before, job demands-resources can help to relate inclusive leadership to the burnout concept. This will be explained further below.

Maslach (1982) stated that leaders with their supervisory actions have a key role in exacerbating or reducing burnout amongst employees. For instance, leaders who make themselves available and accessible for followers, who exhibit a high level of openness

(17)

17 and who share power, will improve the relationship between a leader and followers since

followers are part of the decision-making processes within teams rather than receiving outcomes of decisions (Nishii & Mayer, 2009; Hollander, 2012). This is based on the mechanism where leaders influence the follower’s self-concept, which results in more feelings of competence and self-efficacy to execute tasks (Maslach et al., 2001; Hollander, 2012).

Furthermore, an inclusive leader gives employees opportunities to learn by providing social support, listening to followers and giving them decision rights (Hollander, 2012; Hirak, et al., 2012). These characteristics of an inclusive leader could improve employees’ well-being through the social support from the leader, employees’ perceptions of a fair process and having a voice (Hollander, 2012). Specifically, all these ways of behaving of an inclusive leader can be seen as job resources or facilitators in order to buffer for job demands of employees, which implies that an inclusive leader protects followers against feelings of overextension and depletion of physical resources (Halbesleben & Demerouti, 2005). Through the mechanisms of giving support, appreciating subordinates and their work and including different opinions, employees may not create a cynical stand, since they feel less exhausted as well as disheartened (Maslach et al., 2001). These resources could prevent emotional exhaustion among employees (Maslach et al., 2001; Langballe et al., 2006).

Employees could even express themselves freely because inclusive leadership gives employees a feeling of psychological safety where employees can share their opinions and concerns (Carmeli, Reiter-Palmon & Ziv, 2010). This means that employees can share with team members and their leader when work is too demanding or when they

(18)

18 need certain resources from their inclusive leader (Hollander, 2009). This way of

empowering consequently extends employees’ self-efficacy and rejects outspokenness (Hollander, 2012). Employees’ feelings of reduced disengagement could be the result because of the provided resources by an inclusive leader (Maslach et al., 2001).

To conclude, the described behavior of inclusive leadership is vital for employees to feel less overwhelmed, exhausted and disengaged. Therefore, leaders can be seen as variables that may predict burnout through providing relevant resources to their employees. An inclusive leader meets the requirements in order to buffer for job demands, which will reduce the risk of employee burnout by providing resources such as support, employees’ voice and appreciating employees and their work (Mitchell et al., 2015; Hollander, 2009; Hollander, 2012). Bases on the arguments given before, when a follower perceives a lot of inclusion from the leader, a negative relationship is expected for leader inclusiveness with disengagement and emotional exhaustion. This leads to the first hypotheses:

H1a: Inclusive leadership is negatively related to disengagement. H1b: Inclusive leadership is negatively related to emotional exhaustion.

As mentioned, major strengths of inclusive leadership are putting followers in the foreground and focuses on the relationships between leaders and followers (Hollander, 2006). Nowadays, the role of followers is more important than ever before and seems to be essential in achieving organizational success in terms of employee well-being and other organizational outcomes, such as productivity and efficiency (Hollander, 2012).

(19)

19 This means that the relationship between leader and follower(s) is indispensable. For

these reasons, this study focuses on the dyadic relationship, which is explained by the leader-member exchange (LMX) (Graham, 2009). A more detailed description of LMX will be discussed in the next section, where LMX will be positioned as a mediator between inclusive leadership and employee burnout.

2.4 The mediating role of leader-member exchange

LMX is the main theory that describes the relationship between the leader and the follower, which means that leaders have different relationships with each of their followers (Sparrowe & Liden, 1997). This development of LMX is rooted in both role and exchange theories that focus on how different actors feel about a certain exchange relationship based on their perceptions about fairness (e.g. Graen & Uhl-Bien, 1995; Graham, 2009, Lee, 2008; Howell & Hall-Merenda, 1999; Gerstner & Day, 1997). This, however, declares actors’ input in the LMX relationship. Also, through the interaction between a leader and follower an effective relationship will arise and followers are able to gain access to benefits of this relationship (Graen & Uhl-Bien, 1991). This LMX relationship can vary over time and influences follower perceptions of the organizational environment and attitudes (Davis & Gardner, 2004 in Graham, 2009). Therefore, the better the LMX relationship the more effective this relationship will be in terms of individual outcomes (e.g. employee well-being) and organizational outcomes (e.g. turnover rate) (Graen & Uhl-Bien, 1995).

The level of analysis can vary from low-quality relationships to high-quality

(20)

20 obligation and mutual trust (e.g. Graham, 2009; Wang et al., 2005; Howell &

Hall-Merenda, 1999; Wayne, Shore & Liden, 1997). This means that the mechanism is based on the way both leaders and followers perceive and react to actions and qualities of each other (Hollander, 2012). Subordinates in a low quality relationship, for example, do not enjoy the same levels of emotional support or trust and do not get the same exclusive rights (Harris & Kacmar, 2006). A low quality relationship is based on the formal job description only and leads to a distance between leader and follower (Graham, 2009). In contrast, followers in a high-quality relationship receive various benefits in terms of resources and information (Erdogan & Enders, 2007). Examples of these advantages are personal favours, increased communication with their leader and rewards (Wayne et al., 1997 in Harris & Kacmar, 2006). The higher the quality of the relationship, the more mutual trust, influence, and obligation will emerge (Howell & Hall-Merenda, 1999). In addition, a high-quality relationship seems to be associated with employees who engage in organizational behaviour and go for the extra mile, which contributes to performance ratings in return (Wayne, Shore, Bommer & Tetrick, 2002). As the burnout mechanism is related to job demands and job resources, a high LMX relationship may serve as a resource for employees and reduces feelings of emotional exhaustion and disengagement (Thomas & Lankau, 2009). These differences between a low and high LMX relationship underline the assumption that leaders can be seen as providers of resources (Erdogan & Enders, 2007).

As this study focuses on inclusive leadership, an inclusive leader values personal qualities and actions of all the followers, which will we positively received by all of them (Hollander, 2012). An inclusive leader meets the requirements of being supportive, being

(21)

21 an effective communicator and being trustworthiness to all the employees (Hollander,

2012). Also, inclusive leaders are characterized as effective communicators, which imply that all the followers stay updated by getting the information they need (Hollander, 2012). Last, all the followers will equally benefit from an inclusive leader (Hollander, 2012). Therefore, based on these characteristics, it is expected that an inclusive leader and its subordinates will have a high-quality LMX relationship.

Such a high-quality LMX relationship leads to a positive effect on diminishing

burnout, since a strong leader-member relationship can be seen as a job resource for employees (Bakker et al., 2014). When employees experience a high-quality relationship, a strong leader-member relationship could buffer for overwhelming demands such as work overload and social conflict, since employees do have a voice where they can indicate their boundaries in terms of these job demands (Maslach et al., 2001). As work overload and social conflict are related to exhaustion, it can be expected that a high-quality relationship may lead to less exhaustion (Maslach et al., 2001). Moreover, employees may perceive more demands in terms of support, appreciation and rewards that compensates for the hard working (Maslach et al., 2001). This subsequently could diminish feelings of disengagement among followers (Demerouti et al., 2003). Therefore, LMX is expected to mediate between inclusive leadership and burnout.

Both positive and negative relationships have been found between LMX and

burnout (e.g. Thomas & Lankau, 2009; Graham, 2009). However, it is expected that inclusive leadership improves the LMX relationship, which consequently leads to less burnout sufferers, due to the positive effects of inclusive leadership. Hence, the second group of hypotheses proposed for testing is the following:

(22)

22

H2a: LMX mediates the negative relationship between leader inclusiveness and disengagement.

H2b: LMX mediates the negative relationship between leader inclusiveness and emotional exhaustion.

2.5 The possible dark side of leader inclusiveness

Aforementioned, an increasing number of researchers are interested in the dark side of leadership (Griffin & O’Leary-Kelly, 2004). After the entire positive arguments, leader inclusiveness may be too much of a good thing. A leadership style may look very positive at first sight but could also be destructive in a certain way. For instance, a leader who stimulates the social good, empowers followers and elevates them by a transforming and inspirational leadership style could also be very instrumental in the way the leader utilizes followers as tools for personal gain (Clements & Washbush, 1999). This implies that beneficial outcomes of an inclusive leader, such as extra role behaviour or psychological safety, may result in something negative as workload, high expectations from the leader and work pressure (Nembhard & Edmondson, 2006; Maslach et al., 2001). As previously mentioned, these outcomes could result in exhaustion, which is part of the burnout concept (Graham, 2009).

However, not everything is within the control of the leader. An example is the

educational level of followers or employees’ roles that could impact the relation between leadership and employee well-being (Arnold et al., 2010). This implies that such factors can also influence the relationship between leader inclusiveness, LMX and burnout,

(23)

23 which means that inclusive leadership is maybe not the best matching leadership style

due to the interplay with employee personality and contextual factors. This is where two moderators come into play, which are related to contextual and personal precursors of burnout, task independence and emotional stability, respectively. Maslach et al. (2008) emphasize the importance of these factors in relation to employee burnout. In the next two sections, a more detailed description of the unwanted outcomes of inclusive leadership will be discussed. First, task independence as moderator will be discussed. Second, the role of emotional stability as personality trait will be initiated as moderator.

2.6 The moderating role of task independence

Task independence can be defined as the extent to which group members can interact and accomplish tasks without each other (Campion, Medsker & Higgs, 1993). The more task independence within a team, the less interaction and coordination among team members is required (Katz-Navon & Erez, 2005), whereas this coordination is a key premise of inclusive leadership. In addition, Mitchell et al. (2015) suggest that inclusive leadership might have greater effects when team tasks are interdependent over independent. However, not all the teams within organisations are highly task interdependent, which makes task independence an interesting concept combined with inclusive leadership (Mitchell et al., 2015). Several researchers opted for investigating the moderating effect of task independence as contextual factor on the relationship between leadership and well-being constructs (e.g. Wageman 1995; Kramer, 2004).

This study took steps towards answering the call of Wageman (1995) and Kramer

(24)

24 burnout. Aubé, Rousseau, Mama and Morin (2009) found support for a moderating effect

of the level of task independence, while having psychological well-being as dependent variable. This implies that task independence could impact well-being concepts, such as burnout. Four arguments are outlined below.

First, task independence could be contradicting with the inclusive leader’s way of working, since this leadership style requires a lot of meetings associated with discussions where all the team members can have their say (Mitchell et al., 2015). This is unnatural when being task independent and can be perceived as time consuming and inefficient. This consequently may lead to unpleasant outcomes such as strain that is related to exhaustion (Maslach et al., 2001). In addition, Kramer (2004) questions the increasing positive outcomes of a high quality leader-member relationship and focused on detrimental effects, such as stress, due to the amount of communication between leader and follower. Although communication leads to intellectual stimulation for all the followers, it might lead to more workload and subsequently to work stress (Kramer, 2004). When having an inclusive leader, less time is left to spend on individual tasks and group meetings are time-consuming, which can increase stress levels when these meetings are not needed for followers own task performance. In return, higher stress levels may negatively impact the expected negative relationship between inclusive leadership and burnout.

Second, in line with the previous reasoning, the prescribed behaviours through

which inclusive leaders value and appreciate different ideas and perspectives of team members imply a lot of knowledge sharing and open discussions (Mitchell et al., 2015). Sharing information, achieving mutual goals, participating and contributing in

(25)

decision-25 making could be seen as job demands rather than job resources for task independent

employees (Hollander, 2012), especially when task independence is implying that the core tasks do not require this interaction. An extremely high or low level of task independence within a team may lead to tensions between an individual and their leader since employees who are task independent tend to lose their autonomy when having an inclusive leader (Langfred, 2005). An inclusive leader may create feelings of ‘forced’ cooperation and smooth communication between group members, which can have a detrimental effect on these individualistic employees (Salancik & Pfeffer, 1978; Saavedra, Earley & Van Dyne, 1993). Given the tensions that can arise between leader and follower, task independence could negatively impact the expected relationship between inclusive leadership and burnout.

Third, the positive intentions of an inclusive leader to include everyone in most of the decision making processes, may be perceived as higher expectations from their leader, resulting in experienced pressure (e.g. Langfred, 2000; Arnold et al., 2010). The possible benefits of task independence – resources such as job autonomy and efficient decision making – may be diminished by an inclusive leader trying to involve all parties in each decision (Arnold et al., 2010). This could be seen as a lack of the needed resources for task independent employees and may result in a disturbed balance between their job demands and job resources (Maslach et al., 2001). Also, social conflicts between the leader and follower may arise (Langballe et al., 2006). Both outcomes can have detrimental effects on the relationship between inclusive leadership and burnout.

To conclude, it can be argued that the combination of inclusive leadership and

(26)

26 where support of an inclusive leader and the voice employees can not buffer for it. These

resources can be seen as demanding as well. Based on the previous given arguments, it can be expected that task independence will moderate the expected negative relationship between leader inclusiveness and employee burnout. Therefore, the following hypotheses are advanced:

H3a: Follower task independence is negatively related to the relationship between leader inclusiveness and disengagement, such that this negative relationship is weaker when follower task independence is higher.

H3b: Follower task independence is negatively related to the relationship between leader inclusiveness and emotional exhaustion, such that this negative relationship is weaker when follower task independence is higher.

Despite research on burnout has nearly solely focused on the work context, it is an intriguing point of discussion why differences exist in reporting burnout among employees working in the same environment (Langelaan et al., 2006). This means that followers’ personality can also influence the LMX relationship, which cannot be neglected (Clements & Washbush, 1999; Graham, 2009; Maslach et al., 2001). Therefore, the role of the follower influenced by personality will be discussed in the next section.

2.7 The moderating role of emotional stability

The level of emotional stability determines how easily employees’ emotions and associated responses are triggered (De Hoogh & Den Hartog, 2009). A low level of

(27)

27 emotional stability means that individuals are easily hurt by stress, experience negative

moods and have a negative view of the world (Bono & Judge, 2004; Stewart & Barrack, 2004; De Hoogh & Den Hartog, 2009). Previous research has found that low levels of emotional stability are related to negative emotional responses such as stress and feelings of personal insufficiency (De Hoogh & Den Hartog, 2009; Zellars, Perrewe, & Hochwarter, 2000 in De Hoogh & Den Hartog, 2009). In this study, the focus is on low levels of emotional stability that may help explain why the beneficial effects of inclusive leaders in reducing burnout may not hold for all the followers.

As previously discussed, inclusive leaders try to reach their goals by including all the followers in decision-making, giving employees a voice and appreciating their input (Nembhard & Edmondson, 2006; Mitchell et al., 2015). Despite these positive inclinations, it could also be seen as expectations from the leader since he or she is also giving something to the followers (Hollander, 2012). This implies that followers low in emotional stability may experience these expectations as harder to deal with than the ones high in emotional stability (De Hoogh & Den Hartog, 2009). Next to that, the frequent interactions within the team may increase the stress level and overload of subordinates low in emotional stability (Hetland et al., 2007). Both arguments can result in friction between leader and follower (Hollander, 2012; Langballe et al., 2006).

In line with this arguing, Bono and Vey (2007) found that employees low in

emotional stability are likely to perceive higher job demands, as compared to employees high in emotional stability to buffer for their negative emotions. In addition, emotional stability impacts someone’s job in terms of work style, performance and attitudes and in the other way around (Graham, 2009). Based on the theory, the less emotional stable

(28)

28 someone is, the less he or she can cope with stress (Graham, 2009). Stress as outcome,

can have a negative effect on the relationship between LMX and burnout.

This automatically means that the level of emotional stability impacts the

relationship with the leader. Previous discussed arguments lead to the expectation that low emotional stable individuals may experience the leader’s behaviour different than highly emotional stable employees (De Hoogh & Den Hartog, 2009). It is questionable whether the expected positive effects of inclusive leadership on LMX also hold for the effect on disengagement and emotional exhaustion for employees who score low on emotional stability. Employees low in emotional stability may thus benefit less from an inclusive leader, as they are less likely to manage job demands such as job strain. Therefore, this study will examine to what degree the personality factor of emotional stability moderates the relationship between LMX and the dimensions of employee burnout.

As discussed previously, it is expected that inclusive leadership is positively

related with a high-quality LMX relationship. For such a good leader-member relationship, input from both parties in terms of time and effort is required (Hollander, 2012). Investments for followers low in emotional stability in such a relationship can be emotionally demanding. Followers low in emotional stability may thus benefit less from an inclusiveness and subsequently from a high-quality LMX relationship.

In sum, individuals who are high in emotional stability might benefit more from

the supporting and inclusive behaviour of an inclusive leader that might result in a high-quality LMX relationship. That is why it is expected that emotional stability strengthens the negative relationship between leader inclusiveness and employee burnout. This leads

(29)

29 to the last two hypotheses:

H4a: Follower emotional stability is positively related to the negative relationship between leader inclusiveness and disengagement, so this relationship is weaker when follower emotional stability is higher.

H4b: Follower emotional stability is positively related to the negative relationship between leader inclusiveness and emotional exhaustion, so this relationship is weaker when follower emotional stability is higher.

(30)

30

3. Method

The full research execution is presented in the next section. The procedure, the respondent sample and the instruments that are used, are explained in detail. Subsequently, the preliminary analyses and data analyses are provided, together with the predictions.

3.1 Procedure

Data was collected in collaboration with six other Master students Business Administration and our thesis supervisor of the University of Amsterdam. The study was conducted among several Dutch occupations. The online survey, which has been made with Qualtrics was recruited via contacting personal contacts and companies. The requirements for teams to participate were that these consisted of at least three team members who work together more than two days a week. Data gathering started on March 14th 2017 and was closed in four weeks on April 7th 2017. In order to generalize the data, a large sample size in a limited time frame was needed. Therefore a convenience approach was used. As seven students used a non-probability sampling technique, a divers sample can be expected (Field, 2009).

Two questionnaires were designed, one for the leader, and the other for the

subordinates, referred to employees in this study. For this study, only the employee questionnaire is used. Both English and Dutch questionnaires were available. All items used in the survey were originated from English studies. Native speakers translated the English questionnaire in Dutch to ensure the comparability in meaning of the two versions. Both questionnaires contained an introduction, in which the general research goal, the length, and procedural facets were described. If they wanted, respondents were

(31)

31 offered a report with the general findings of this study. The surveys were filled in

anonymously, but all the participants could contact the researchers for any questions.

3.2 Sample

289 employee questionnaires were returned, corresponding to a response rate of 69%. As the norm for this study was having at least three team members and one leader per team, incomplete teams were deleted from the total sample consisting of 289 employees, which led to a sample of 215 (74%). Besides this, no more respondents were eliminated from the study since Statistical Package for Social Sciences (SPSS) of IBM is capable of detecting unanswered questions. After eliminating missing values, a total of 215 employees from 49 different teams working under the same supervisor filled in the questionnaire electronically. The exact number of respondents used per analyses depends on the missing items per scale. As this research only requires the employee sample, therefore, exclusively the employee sample is outlined below.

Taking all respondents, the most frequent age group of the included sample was

20-30 years (Mage = 3.14, SDage = 1.42), with a gender distribution of 43% female and 57% male. Employees’ tenure varies from two months up to 43 years (M = 8.58; SD = 11.09). The sample insured a range of educational backgrounds. Only 22,8% completed just a secondary education programme, 24,2% completed an educational programme at University of Professional Education and most of the respondents completed an educational programme at the University. Most of the respondents (46%) completed an education programme at the University.

(32)

32

3.3 Measures

The variables inclusive leadership (the independent variable), burnout (the dependent variable), LMX (the mediating variable) and emotional stability and task independence (the moderating variables) were all measured from the employee perspective. The questionnaires were not only used for this research paper, so therefore the remaining variables were not included in this research paper. Most of the items were rated on a 7-point likert scale, ranging from 1 = strongly disagree to 7 = strongly agree. Only the construct of LMX has an irregular scale, but is interpreted as a 7-point likert scale for the analysis.

Burnout measurement is based on 16-item Oldenburg Burnout Inventory (OLBI)

scale. The Oldenburg Burnout Inventory (OLBI) consists of two dimensions: disengagement and emotional exhaustion and is validated by Halbesleben and Demerouti (2005). Eight items were reversed coded. Cronbach’s alpha for disengagement and emotional exhaustion are .85 and .80, respectively. An example item is ‘’During my work, I often feel emotionally drained’’ (Halbesleben & Demerouti, 2005).

Inclusive leadership is measured using the 9-item scale of Carmeli et al. (2010) in

order to test the three different dimensions; openness, accessibility, and availability. It reflects the inclusive leadership behaviors perceived by employees. Cronbach’s alfa for inclusive leadership is .94. An example of an item is “The manager is ready to listen to my requests.’’

Emotional stability is assessed using the 4-item scale from the International

(33)

33 originally measures neuroticism, so the scale was reversed coded to align it with the

direction of the hypotheses. That is why it is also renamed into the opposite meaning, namely neuroticism into emotional stability. Cronbach’s alfa of emotional stability is .69. An example item is “I Am relaxed most of the time”.

Task independence is measured using the scale of Campion et al., (1993), which

consists of four items. It measures to what extent team members depend on each other to complete their tasks. Again, as this scales originally measures task interdependence, the scale was reversed coded to align it with the direction of the hypotheses. The Cronbach’s alfa is .75. An example of an item is “I cannot accomplish my tasks without information or materials from other members of my team”.

Control variables. Gender, tenure and age are controlled to diminish the potential

bias of the regression results. According to Bauer, Erdogan, Liden & Wayne (2006) differences in gender have been identified as facets that influence the relation between leaders and their followers. In particular, women are more sensitive for burnout than men (Purvanova & Muros, 2010). This implies that differences between men and women could result in different leader-member relationships that subsequently impact employee outcomes, such as burnout. Also, gender was recoded as 0 for female and 1 for male in SPSS. Age is controlled since studies have found a negative relationship between age and burnout (Randall, 2007). Specifically, the older people are, the more experienced they are in coping with stress, which means that it is expected that younger employees are more prone to getting a burnout. In addition, as burnout and LMX relationships develop over time, tenure can impact the results of this study (De Hoogh & Den Hartog, 2009).

(34)

34

3.4 Data analyses

As previously mentioned, Statistical Software Package for Social Sciences was used to perform statistical analyses. Table 4.1 provides the descriptive statistics consisting of means, standard deviations, correlations of the five variables, control variables and Cronbach’s alphas among the study variables. As the variables are measured with a scale from one to seven, it can be argued that the employees in this sample experience a high level of inclusive leadership (M = 5.65, SD = .93) as well as emotional stability (M = 5.08, SD = 1.09). In addition, employees experience moderate levels of disengagement (M = 3.29, SD = 1.04), emotional exhaustion (M = 3.08, SD .91), leader-member exchange (M = 3.71, SD = .59), and task independence (M = 3.03, SD = 1.23).

Regression analyses were executed in order to test the moderated and mediated

hypotheses. To make it more specific, direct relationships were tested by the use of hierarchical regression. Subsequently, a linear regression analysis defined the expected linear relationship between inclusive leadership and the two dimensions of burnout. Next, the moderating effect of task independence on the direct relationship will be tested via

model 1 of PROCESS analyses. Last, in order to measure the moderated mediating effect, model 14 of PROCESS analysis will be used.

By doing reliability analysis the internal validity was checked. Every scale is

suitable for use with a Cronbach’s Alfa of .70 or above except the emotional stability scale (α = .69). Therefore, it can be concluded that almost all the scales are sufficiently reliable (Field, 2009). Task independence (α = .75) is above the conventional acceptable norm, the scale is not highly reliable though. However, none of the items in the scales is deleted, since this would not substantially affect the reliability (Field, 2009).

(35)

35 Subsequently, all the scales were computed by averaging the items, into variables for

leader inclusiveness, burnout (disengagement and emotional exhaustion), task independence and emotional stability in order to run the analyses. The explore function was used to test for homoscedasticity, normality and linearity. It can be expected that the sample is normally distributed, but some outliers were found for burnout and LMX, which is not problematic since it is assumed that the large sample (N = 215) is representative for the population (Field, 2009).

3.5 Predictions

As mentioned before, it can be expected to find a negative direct relationship between inclusive leadership and the two dimensions of burnout, emotional exhaustion and disengagement, respectively. From now on, these dimensions will be discussed separately. In addition, it is expected to find an indirect effect of leader-member exchange on the direct relationship. Also, it is expected to find support for the moderating effect of emotional stability on the indirect relationship, suggesting that this relationship is weaker for low levels of emotional stability. Last, this study expects to find evidence for a conditional effect of task independence on the indirect relationship between inclusive leadership and the dimensions of burnout, so that the indirect relationship is weaker for higher levels of task independence.

4. Results

The next section represents the different steps taken in order to analyse the data. The analytical analyses as well as the results of those are described in detail in order to replicate the study in the future.

(36)

36

Descriptive statistics

As PROCESS analysis automatically excludes data listwise (Hayes, 2012), the analyses were drawn on a sample of 215 employees who answered the investigated questions. Subsequently, the Pearson’s correlation coefficient in Table 4.1 shows that the negative relationship inclusive leadership with disengagement (H1a) and emotional exhaustion (H1b) are supported, respectively r = -.40, p < .01 and r = -.17, p < .05. This indicates a stronger relationship between inclusive leadership and disengagement than between inclusive leadership and emotional exhaustion, which will be further elaborated on in the discussion.

As expected, the correlation between inclusive leadership and LMX is

characterized by a large effect, r = .68, p = < .01. So it does not come as a surprise that the results provide initial support for the indirect relationship of inclusive leadership with disengagement (H2a) and emotional exhaustion (H2b) through LMX, respectively

r = -.55, p < .01 and r = -.39, p < .01. A small effect size was found for the interaction

effect of inclusive leadership and task independence r = .14, p < .05. In addition, results indicate a medium effect for the moderating effect of emotional stability on the indirect relationship, with disengagement as outcome variable (H4a), r = -.28, p = < .01. The results even show a strong effect with emotional exhaustion as dependent variable (H4b),

(37)

Table 4.1. Descriptive Statistics and Correlations

** Correlation is significant at the 0.01 level (2-tailed). * Correlation is significant at the 0.05 level (2-tailed). Listwise N = 215. Cronbach's alpha on diagonal.

M SD 1 2 3 4 5 6 7 8 9 1. Disengagement (Burnout) 3.29 1.04 (.85) 2.Emotional exhaustion (Burnout) 3.08 .91 .52** (.80) 3.Leader inclusiveness 5.65 .93 -.40** -.17* (.94) 4.Leader-member Exchange 3.71 .59 -.55** -.39** .68** (.84) 5.Emotional stability 5.08 1.09 -.28** -.56** .05 .18** (.69) 6.Task independence 4.97 1.23 -.13 -.00 .14* -18** -.08 (.75) 7. Age 3.14 1.24 -.23** -.20** -.01 .02 .20** .05 (--) 8. Gender .58 .51 -.18** -.11 .06 .14* .05 .04 .21** (--) 9. Tenure 8.58 11.09 -.15* -.16* -.03 .01 .17* . 10 .72** .16* (--)

(38)

4.2 Linear regression

In order to test the expected negative relationship between inclusive leadership, disengagement and emotional exhaustion, regression analyses were conducted. These regression analyses can be generalized since the scatterplot as well as the histogram indicated homoscedasticity and a normal distribution in the data (Field, 2009). The results of these linear regression analyses are presented in Table 4.2 and Table 4.3. Table 4.2 provides the results of the linear regression results for disengagement as outcome variable, whereas Table 4.3 shows the results of the regression results with emotional exhaustion as dependent variable.

The first condition states that the independent and the dependent variable must be significantly related. In the first step (Model 1) of the regression, control variables were entered: age, gender and tenure. This model was statistically significant F (3, 208) = 5.30;

p < 0.01 and explained 7,1% of the variance in disengagement. After entering inclusive

leadership in Model 2, the total variance explained by the total model was 23% F (4, 207) = 15,48; p < .001. The introduction of inclusive leadership in Model 2 explained an additional 16% variance in job satisfaction, after controlling for age, gender and tenure (F (1, 207) = 42.80; p < .001. It is notably that in the total model three out of four predictor variables were statistically significant with the highest Beta value for inclusive leadership (β = -.40, p < 0.001). This means that if inclusive leadership increases for one, their disengagement will decrease for .40. Interesting is that on average women score 0.13 lower in disengagement than men, which indicates that inclusive leadership has stronger effects on disengagement for man than for woman. All in all, these results support H1a, stating that leader inclusiveness is negatively related to disengagement.

(39)

39 The second condition states that the independent and the dependent variable,

emotional exhaustion must be significantly related to each other. Therefore, H1a was also tested with the regression to investigate inclusive leadership to predict levels of emotional exhaustion. Again, age, gender and tenure were used as control variables and were entered as three predictors in the first step (Model 1). This model was statistically significant F (3, 208) = 3.36; p < 0.05 and explained 4,6% of the variance in emotional exhaustion. After entering inclusive leadership at Step 2 the total variance explained by the total model (Model 2) was 7,7% F (4, 207) = 4,31 p < 0.001. The introduction of inclusive leadership explained additional 3,1% variance in emotional exhaustion, after controlling for age, gender and tenure (F (1, 207) = 6.90; p < 0.01). Notably, in the final model only one out of four, namely inclusive leadership was statistically significant with recording a Beta value of β = -.18; p < .001. In other words, if inclusive leadership increases by one, their emotional exhaustion will decrease with 0.18. To conclude, the expected linear relations between inclusive leadership and both disengagement (H1a) and emotional exhaustion (H1b) that would have the expected negative effect are supported.

4.3 Conditional effect of task independence

This condition requires that the direct relationship between inclusive leadership and burnout would be weaker when the moderator, task independence is included. In order to test this interaction effect, inclusive leadership and task independence are standardized. This simplifies the interpretation afterwards and avoids multicolliearity (Hayes, Glynn & Huge, 2012). Although hypothesizes H3a and H3b proposed a moderation effect of task independence on the relationship of inclusive leadership with disengagement and emotional exhaustion, no support was found for these hypotheses (B = .03, t = .43, n.s.

(40)

40 and B = -.02, t = -.41, n.s. respectively). This implies that the nature of the relationship

did not change by the moderating variable task independence, so hypothesizes H3a and H3b are rejected. The results related to this specific regression are presented in Table 4.4 and Table 4.5.

4.4 Moderated mediation

The PROCESS approach of Hayes (2012) allows for testing the mediation of LMX between inclusive leadership, and between disengagement as well as emotional exhaustion. 5.000 bootstrap samples were extracted in order to find support for the moderated mediating effect. Aforementioned, the independent variables were disengagement and emotional exhaustion, the mediating variable was LMX, the moderator was task independence and the dependent variable was inclusive leadership. The analyses were performed for both disengagement and emotional exhaustion.

The significance levels of indirect effects in this study were determined using

bootstrapping. This is a nonparametric resampling procedure which is an addition advocated for testing mediation without imposing the assumption of normal sample distribution (Preacher & Hayes, 2008). Additionally, bootstrapping is a computational and intensive way that involves repeatedly sampling that estimated the indirect effect in each sample (Preacher & Hayes, 2008). By repeating this process 5.000 times, an empirical approximation of the sampling distribution can be given, which in turn is used to determine the confidence intervals for the indirect effect (Preacher & Hayes, 2008). In the following, the results for the moderated mediation for both disengagement and emotional exhaustion are described. Also, the results of these analyses are presented in Table 4.6 and Table 4.7, respectively.

(41)

41

4.4.1 Disengagement

The PROCESS results show mixed outcomes in terms of significance, as provided in Table 4.6. A large direct effect of inclusive leadership with LMX, ai = .67 means that two employees that differ by one unit on inclusive leadership are estimated to differ by .67 units on LMX. This effect is statistically significant p < .001 with a 95% confidence interval from .436 to .908. Another significant result holds for the relationship between LMX and disengagement b1i = -.47; p < .001, with a 95% confidence interval from -.643 to -.293). This effect indicates that two subordinates who differ by one unit in experiencing inclusive leadership determined to differ by .67 units in the reported LMX. Consequently, when having a high quality relationship with the leader, less disengagement will be the result. This theoretically means that when two employees differ by one unit in experiences of LMX estimated to differ by -.47 in the reported disengagement. Nevertheless, the direct effect of an inclusive leadership is not statistically significant anymore. The mediation effect of LMX (c’ = -.08, n.s.) is the reason for this. This is called a full mediation in statistically terms (Field, 2009). This implies that having a high quality relationship with the leader is of great importance in diminishing the level of disengagement (H2a).

The next condition requires that the indirect relationship between inclusive

leadership and disengagement through LMX, would be stronger when an interaction between LMX and emotional stability takes place. Nevertheless, it has turned out to be statistically insignificant (b3i = -.08, n.s.), thus the expected moderated mediation does not take place. Furthermore, Model 14 allows for testing the direct effect of the moderator on disengagement, b2 = -.17, which is statistically different from zero p < 0.01

(42)

42 with a 95% confidence interval from -304 to -.042. This implies a negative relationship

between emotional stability and disengagement. To make it clearer, if employee’s emotional stability increases for one, their disengagement will decrease with 0.17. In conclusion, the mediation (H2a) is supported, but de moderated mediation (H3a) is rejected. In the next paragraph, emotional exhaustion as outcome variable will be discussed.

4.4.2. Emotional exhaustion

Again, like the previous outcome variable, the results (Table 4.7) show a statically significant effect of .67 for inclusive leadership on LMX (p < .001, CE: .436 to .908). Also, an effect of LMX, b1i = -.32 on emotional exhaustion has been found which is statistically significant p < .01 with a 95% confidence interval of -.520 to -.125. Again, like the moderated mediation for disengagement, the direct relationship between inclusive leadership c’ = -.10 and emotional exhaustion is insignificant when adding LMX. This implies a full mediation between inclusive leadership and emotional exhaustion through LMX.

However, the regression coefficient for the interaction effect of LMX and

emotional stability b3i = -.07 on emotional exhaustion is not statically different from zero in the model of emotional exhaustion (b3i = -.07, n.s.). Thus, the conditional effect is not supported between LMX and emotional exhaustion. However, a negative effect of emotional stability, b2 = -.44 on emotional exhaustion has been found. This effect is statistically different from zero, p < .001 with a 95% confidential interval from -55 to -32. Furthermore, results show no statistically significant direct effect between inclusive leadership and emotional exhaustion has been found (c’ = -.10, n.s.).

(43)

43 The results, as also presented in Table 4.7, confirm that, LMX fully mediates the

relationship between inclusive leadership and both disengagement and emotional exhaustion. To conclude, the mediation (H2b) is supported, but the moderated mediation (H3b) is rejected.

(44)

Table 4.2 Hierarchical Regression Model of Burnout (Disengagement)

Disengagement

Step 1 Step 2 Adj. R2 ΔR2 ΔF

Step 1: Control variables .06 .07 5.30** Age -.206* -.207* Gender Tenure -.148* .018 -.126* .003 Step 2: .22 .16 42.80*** Inclusive leadership -.400***

(45)

45

Table 4.3 Hierarchical Regression Model of Burnout (Emotional exhaustion)

Emotional exhaustion

Step 1 Step 2 Adj. R2 ΔR2 ΔF

Step 1: Control variables .04 .05 3.36* Age -.15 -.15 Gender -.09 -.08 Tenure -.04 -.05 Step 2: .06 .03 6.70** Inclusive leadership -.17**

(46)

46

Table 4.4 moderation of task independence (Disengagement)

Coefficient SE t p

Inclusive leadership (X) b1 -.40 .07 -6.15 <.001

Task independence (M) b2 -.06 .07 -.87 .39

Task independence*inclusive leadership (XM) b3 .03 .06 .43 .66

Intercept i1 3.98 .20 19.44 <.001

R2= .234 p<0.001 F(6,205)= 10.430

Table 4.5 moderation of task independence (Emotional exhaustion)

Coefficient SE t p

Inclusive leadership (X) b1 -.17 .06 -2.72 <.0.05

Task independence (M) b2 -.04 .06 .71 .48

Task independence*inclusive leadership (XM) b3 -.02 .06 -.42 .68

Intercept i1 3.53 .19 18.13 <.001

R2=.080 p<0.05 F(6,205)= 2.973

The dark side of leadership inclusiveness : moderated mediation effects on employee burnout in teams