The relationship between personality traits and leadership preferences: An analysis of the leadership preferences of Defence cyber employees

The relationship between personality traits and leadership preferences

An analysis of the leadership preferences of Defence cyber employees

Supervisor:

Prof. dr. S.M. Groeneveld Second reader:

Dr. J. van der Voet

Marije Timmer S2361663

m.timmer@umail.leidenuniv.nl

14 May 2020 MSc Public Administration Track: Public Management and Leadership

List of contents

2.3 Personality characteristics: the Big Five ... 17

2.4 The association between personality and leadership preferences ... 18

2.4.1 Theoretical mechanisms ... 18

2.4.2 Hypotheses ... 20

3. Methodology ... 25

3.1 Organisational context ... 25

3.2 Data collection ... 26

3.3 Operationalisation and initial data processing ... 27

3.3.1 Measuring the independent variables ... 27

3.3.2 Measuring the dependent variables ... 29

3.3.3 Measuring the control variables ... 34

3.4 Data analysis strategy... 35

3.5 Descriptive statistics and correlation ... 36

4. Results ... 41

4.1 Assumptions of linear regression ... 41

4.2 Relationship between personality and leadership preferences ... 43

4.2.1 Preference for directive-task leadership ... 43

4.2.2 Preference for supportive-relations leadership ... 47

4.2.3 Preference for empowering-change leadership ... 50

4.3 Robustness checks ... 53

5. Discussion and conclusion ... 55

5.1 Discussion of results ... 55

5.1.1 Discussion of preference for directive-task leadership ... 56

5.1.2 Discussion of preference for supportive-relations leadership ... 58

5.1.3 Discussion of preference for empowering-change leadership ... 59

5.1.4 Discussion of differences between OLS and logistic regression ... 60

5.3 Practical considerations and recommendations ... 63

References... 66

Appendices ... 74

Appendix A: Survey ... 74

4

List of tables

Table 1. Taxonomy of Leader Behaviour ... 16

Table 2. Descriptive statistics ... 39

Table 3. Correlations ... 40

Table 4. OLS regression coefficients for Degree of preference for directive-task leadership ... 45

Table 5. Logistic regression coefficients for Ranking of directive-task leadership (Dummy) ... 46

Table 6. OLS regression coefficients for Degree of preference for supportive-relations leadership ... 48

Table 7. Logistic regression coefficients for Ranking of supportive-relations leadership (Dummy) ... 49

Table 8. OLS regression coefficients for Degree of preference for empowering-change leadership ... 51

Table 9. Logistic regression coefficients for Ranking of empowering-change leadership (Dummy) ... 52

5

Acknowledgements

I would like to thank several people for their support and encouragement during the research process that resulted in this thesis.

First and foremost I would like to express my gratitude to my thesis supervisor Prof. dr. Sandra Groeneveld, for introducing me to this fascinating project and guiding me throughout the research and writing process.

I would also like to thank Senior Advisor Anouk Schipper for supporting my research efforts at various stages and advising me on the organisational intricacies of the Defence Cyber Command.

Furthermore, I would like to show my appreciation of Commander of the Defence Cyber Command Air Commodore Elanor Boekholt-O’Sullivan and Head of the Defence Cyber Security Centre Colonel Kees Verdonk for giving me the opportunity to conduct my research within their respective organisations.

Lastly, I want to thank all the employees of the Defence Cyber Command and Defence Cyber Security Centre for their honesty and for taking the time out of their day to complete my survey.

6

1. Introduction

Today’s rapidly changing world poses many challenges to public organisations. Increased volatility, uncertainty, complexity, and ambiguity are said to plague the public sector. Technological advancement, societal changes, and global power shifts are only a few of the characterising trends of the 21st _{century that fundamentally change the operating environment of}

public organisations (Van der Wal, 2017). While these circumstances are challenging for all parts of the public sector, it also creates new issues for government departments responsible for defence policy. To respond to these challenges, the Dutch Ministry of Defence1 _{put forward a}

White Paper in 2018 outlining a new vision. The concept of the ‘adaptive armed forces’ was announced as key in the future of Defence. This perspective emphasises the need to be able to respond appropriately to changing circumstances, cooperate with partners, and continue to innovate and improve the inner workings of the organisation. Such a vision requires, among other things, a more flexible personnel system with a larger emphasis on the recruitment and retention of employees (Ministry of Defence, 2018).

The introduction of a more flexible, innovative, and responsive organisation arguably makes adaptive leadership and employee empowerment more important. The idea on which adaptive leadership is based is that behaviour needs to adapt as the situation changes in an effort to most appropriately respond to certain circumstances (Yukl & Mahsud, 2010). This leadership philosophy aims to promote adaptability in organisations and stimulate capacity for change (Heifetz & Linsky, 2002). Such adaptability is what is being strived towards by Defence with the introduction of the ‘adaptive armed forces’ concept, which consequently seeks to promote adaptive leadership in the organisation. What this entails in practice is a shift in focus to less traditional leadership behaviours, as adaptive leadership calls for more emphasis on empowering employees, stimulating critical thinking, and promoting the taking of initiatives by employees (Lichtenstein, Uhl-Bien, Marion, Seers & Orton, 2006; Yukl, Gordon & Taber, 2002). This contrasts with more ‘conventional’ task-oriented behaviours such as planning activities and monitoring performance, and relationship-oriented behaviours such as providing support and encouragement for employees (Yukl, Gordon & Taber, 2002). A shift towards adaptive

leadership could thus bring about significant changes in the leadership behaviour that employees are confronted with.

7

However, as this change implies the introduction of a general model of leadership for the entire organisation, some difficulties might occur in relation to the different personalities of employees. The renewed efforts of Defence to invest in the recruitment of new employees, using sometimes innovative methods such as partnerships with other organisations, aims in part ‘‘to make the organisation more diverse in terms of its human capital’’ (Ministry of Defence, 2018). It is therefore likely to also increase the variety of personalities of employees in the organisation, especially in fields that are at the centre of efforts to attract personnel, such as cybersecurity and IT. In addition, the workforce in the cyber domain of Defence consists not only of a mixture of military and civilian employees but also includes substantial amounts of cyber reservists and people from different professional backgrounds. A consequence of having a more diversified employee group is that these employees might also have more diverse preferences as to their supervision. Hence, this research is conducted within the Defence Cyber Command (DCC) and the Defence Cyber Security Centre (DCSC), the main players within Defence with regards to cybersecurity.

Previous research has indeed shown that personal characteristics can be important factors in influencing leadership preferences. For example, Ehrhart & Klein (2001) argue that values and personality help predict the leadership preferences of followers. Besides, Boatwright & Forrest (2000) show that age, educational level, and gender can be useful predictors in this respect. Similarly, Stinson & Robertson (1973) and Vecchio & Boatwright (2002) point to a relationship between factors such as employee maturity and gender and preferences for leadership styles. Moss & Ngu (2006) take a somewhat different approach, by demonstrating the relationship between personality traits and leadership preferences.

A problem arises if existing leadership styles and behaviours, or those that are being developed, do not fit the preferences of the employees. Yukl (1971) argues that it is the

discrepancy between the preferences and the experience of an employee that determines his or her satisfaction, with ‘‘the less the discrepancy between preferences and experience, the greater the satisfaction’’ (p. 419). Considering that job satisfaction and intent to leave tend to be negatively related (Hellman, 1997), maintaining the satisfaction of employees seems to be essential to retain them. Crucially in this respect, the White Paper 2018 does not only emphasise the need to recruit additional employees, but also the need to retain personnel. This makes achieving a fit between leadership preferences and actual leadership behaviour very important. In addition, taking leadership preferences of employees seriously is essential because of the common assumption that leadership preferences mediate the effectiveness of leadership behaviour itself (Singer & Singer, 1990). In fact, it has been repeatedly argued that followers are crucial in

8

realizing the potential beneficial effects of leadership, including effects on performance, productivity, and worker satisfaction (Boatwright & Forrest, 2000; Moss & Ngu, 2006). The problems that a mismatch between leadership behaviour and preferences could potentially cause, including lower job satisfaction and employee retainment, justify an examination of what those preferences are.

Employees are also essential in adaptive leadership since this leadership style aims to ‘empower’ employees. Empowerment can hardly be said to be a one-directional influence from leaders to followers. Indeed, Chiles & Zorn (1995) point out that the follower’s perspective is essential in creating and maintaining empowerment. While these authors examine individuals’ interpretations of leadership behaviour and not their preferences as such, it certainly warrants more attention to the individual’s point of view when it comes to the issue of empowerment. Furthermore, identifying what employees’ leadership preferences are is an important step in preparing the organisation for upcoming changes. Leadership is important in creating the capacity to change and making organisational changes successful (Van der Voet, 2016; Van der Voet, Kuipers & Groeneveld, 2015), and adaptive leadership more specifically is very focused on the creating capacity for change (Heifetz & Linsky, 2002). Yet, a mismatch between leadership behaviour and leadership preferences could apparently endanger the success of such changes. Individual preferences for leadership behaviours are thus crucial to take into account when implementing adaptive leadership.

Nevertheless, as Boatwright & Forrest (2000) point out, ‘‘the role of the worker as a fundamental factor in co-constructing or ‘legitimizing’ effective leadership is often neglected’’ (p. 19). Indeed, despite the seeming importance of leadership preferences, there have been relatively few studies on the preferences of employees (Singer & Singer, 1990; Vecchio & Boatwright, 2002; Ehrhart, 2012). In general, there has been an astounding number of studies on leadership styles and behaviour during the last few decades. However, while many aspects of ‘leadership’ have been examined, the focus has often been on employees’ perceptions of leaders as well as on the effects of leadership on employees (for example see Jacobsen & Andersen, 2015; García-Morales, Jiménez-Barrionuevo & Gutiérrez-Gutiérrez, 2012). As a result, there has generally been a lack of studies on the preferences of employees in regards to leadership. Whenever preferences are examined, gender or age differences tend to be the central factors examined (Boatwright & Forrest, 2000; Zander & Romani, 2004). Personality has been widely under-researched in regards to leadership preferences, with exceptions being, for example, Moss & Ngu (2006) and Ehrhart & Klein (2001). Yet, as their research pointed to possible relationships between personality factors and leadership preferences, it provides an interesting avenue for further research. Indeed, Ehrhart

9

& Klein (2001) themselves speculate that perhaps the ‘Big Five’ personality traits are factors in determining follower preferences to specific leadership styles, and suggest further research be done to shed light on this issue. Moss & Ngu (2006) use these personality traits to examine associations with leadership preferences but focus on more general leadership styles instead of more specific leadership behaviours.

Therefore, this thesis examines whether employees with different personality traits do indeed have different preferences regarding leadership. Considering that most previous research on this topic has examined characteristics such as gender and age, a focus on personality brings a more refreshing look at the topic of leadership preferences. Subordinates’ personalities can be argued to be essential to the study of leadership. Early studies of leadership often focused mostly on the personalities of leaders, but more recent efforts view leadership as an interaction between leader and follower, or even as processes and practices of interaction between people in general (Crevani, Lindgren & Packendorff, 2010). Since personality is commonly accepted to be a very important factor in determining someone’s perception and behaviour in social interactions (Furnham & Heaven, 1999), it is only logical to assume that this might also be the case in leadership relations. For example, an extrovert person is likely to have different preferences in social interactions than an introvert person (for example see Hutchinson & Gul, 1997). While this seems to be common knowledge, it is odd that there has only been very limited research into the relationship with leadership preferences. Therefore, research is warranted to examine if there is a connection between personality traits and leadership preferences.

Specifically, the goal is to examine which leadership behaviours are being preferred by employees of the DCC and DCSC. This knowledge is important in an organisation that aims to invest more in its people, recruit new employees and retain current personnel (Ministry of Defence, 2018). Awareness of the preferences of employees would facilitate the introduction of organisational changes, including a change in leadership styles. In addition, it sheds light on an area of study that tends to be underexposed; namely, the actual leadership preferences of employees. Thus, the examination of the possible relationship between personality factors and leadership preferences has both academic and societal relevance.

The central research question is as follows: ‘‘How do different personality characteristics of employees translate into different preferences as to their supervision?’’. This question allows a careful examination of the leadership preferences of DCC and DCSC employees, and in addition makes it possible to explore the link between their preferences and their personality traits. The research is thus conducted at the level of the leader-employee relationship. To answer this question, a quantitative research project is carried out. More specifically, a vignette survey is

10

employed as the main method to test hypotheses on employees’ leadership preferences, and their possible relationship with personality traits.

In this thesis, the following chapters are included. The introductory chapter is followed by a chapter that covers the theoretical framework applied in this study. It introduces the concept of adaptive leadership, covers the different typologies of leader behaviour, specifies the

differences between needs and preferences, and introduces the Big Five personality model as well as the hypotheses that follow from it. The third chapter elaborates on the methodology,

describing the research design and the data collection strategy, as well as the initial processing and measurement of data and descriptive statistics. Afterwards, the results of the vignette analysis and robustness checks are discussed in chapter four, which are subsequently discussed in chapter five. This last chapter not only offers an interpretation of the results and attempts to answers the research question defined above, but also provides practical recommendations and suggestions for further research, and outlines the limitations of the study.

11

2. Theoretical framework

This chapter reviews the relevant academic literature on adaptive leadership and leadership preferences, and the personality factors that might impact these preferences. In doing so, it also lays down the theoretical foundation for the subsequent empirical study. The hypotheses that are formulated form the basis for that study and guide data collection and analysis.

First, the concept of adaptive leadership is explained and contrasted with similar, but fundamentally different concepts. The introduction of adaptive leadership within Defence provides the backdrop against which this research is conducted. Secondly, different types and categories of leadership behaviour are discussed and a taxonomy of leadership behaviour is presented. Thirdly, existing literature on personality traits and their possible relationship with leadership preferences is examined, and hypotheses concerning these relationships are formulated.

2.1 Adaptive leadership

Given recent societal changes, relational and collective approaches to leadership have gained more attention in the public sector (Ospina, 2016). This trend is also visible in Defence, for which typical 21st-century challenges such as technological advancements and increased unpredictability of the international sphere can be especially concerning. More varied state and non-state actors, new technologies, and vulnerability to social disruption are all part of the developments that put new and additional pressures on defence organisations around the world (Ministry of Defence, 2018; Cojocar, 2011). As a response to such circumstances, the Australian, Canadian, and British defence organisations, among others, adopted more adaptive visions of leadership. Following these developments, in 2014 the Dutch Ministry of Defence revised its perspective on leadership by adopting a vision promoting a more adaptive leadership style (Dalenberg, Folkerts & Bijlsma, 2014). This developed further with the introduction of the ‘adaptive armed forces’ in which adaptive leadership plays a key role (Ministry of Defence, 2018).

Despite this popularity of adaptive leadership as an organisational vision, it is not always completely clear what the concept entails. Adaptive leadership is most commonly considered to be a collective leadership perspective. Such a perspective does not focus on the leader-follower level of interaction but instead looks at the collective, system level. In general, ‘leadership’ in such a perspective is assumed to be ‘‘an emergent, interactive process intended to cultivate group members’ capacity and adaptability to navigate complexity’’ (Ospina, 2016, p. 281). This is exactly

12

the idea behind adaptive leadership: to enable an organisation to become more adaptive by creating the right conditions to allow knowledge, preferences, and behaviours of group members to change when necessary, and in such a way tackling problems and seizing opportunities when they present themselves (Lichtenstein et al., 2006). Leadership in this definition is thus mostly focused on the collective dimension and interactive dynamics in groups and organisations.

Nevertheless, a more leader-focused version of the adaptive leadership concept has been developed that gives more attention to the leader and his or her behaviour. Heifetz and Linsky (2002) contend that for each problematic situation, leaders can create the right conditions that allow a solution to be found through ‘adaptive work’ (Leigh, 2003). This work entails

experimenting, adjusting, and taking risks, as well as changing the attitudes and behaviours of employees to adjust to new circumstances. Through phrases such as ‘get on the balcony’ and ‘orchestrate the conflict’, they advise leaders on how to act in response to ‘adaptive challenges’ (Heifetz & Linsky, 2002). According to these authors, an adaptive challenge occurs when there is no simple solution or response that can solve the issue. Instead, an adaptive challenge requires a change in behaviour and attitudes of both leader and follower to respond appropriately to the circumstances (Heifetz & Linsky, 2002). Yet, while the work of Heifetz & Linksy thus gives more attention to the leader, it is still very much focused on interactive dynamics in groups. Moreover, the ultimate aim of leadership in their perspective is to increase the adaptive capabilities of the organisation and to enable the organisation as a whole to ‘‘meet the ongoing stream of adaptive challenges posed by a world ever ready to offer new realities, opportunities, and pressures’’ (Heifetz, Grashow & Linsky, 2009, p. 17).

While adaptive leadership thus involves changing behaviour to adapt to the circumstances (Yukl & Mahsud, 2010) and it, therefore, bears some similarity to situational leadership, there are fundamental differences between these two concepts. Situational leadership theory was initially developed by Hersey and Blanchard (1996) and contends that the behaviour of leaders should change in accordance with what the situation requires. Depending on the maturity of the followers, the level of either supportive or directive behaviour should be altered. In general, situational or contingency models are thus very much concerned with creating a fit between a situation and the leadership style used, although they differ in their focus on the type of

‘situation’. Many contingency theories do not look at the maturity of followers but consider other factors. For example, Shamir & Howell (1999) argue that the effectiveness and emergence of charismatic leadership depend on the organisational context, and Jacobsen (2015) examines the effect of organisational publicness on the occurrence of transformational leadership. The difference with adaptive leadership is that adaptive leadership is less concerned with the leader

13

changing his or her leadership style to fit the situation and more with creating the capacity in a group or organisation to change and adapt to that situation. In addition, adaptive leadership tends not to look at the micro-level of leadership interactions, whereas situational leadership usually does. The aim of adaptive leadership is thus very much to enhance the adaptability of a group or organisation. Adaptation is key in this respect since adaptive leadership fundamentally concerns dealing with change and processes of change (Lichtenstein et al., 2006).

Additionally, what follows from this characterisation of adaptive leadership is that it is mostly concerned with the adaptiveness of an organisation itself, and not as much with the individual effects that a leader has on an employee, except for some of Heifetz & Linsky’s work. In general, the concept of adaptive leadership entails that a leader also does not have to be in a position of authority for leadership to occur, because any individual can act as a leader ‘‘when they mobilize people to seize new opportunities and tackle tough problems’’ (Lichtenstein et al., 2006, p. 4). Leadership in this sense is much more of a dynamic than an effect. Most popular leadership theories, such as charismatic or transformational leadership (Antonakis, d’Adda, Weber & Zehnder, 2014; Bass, 1990), focus on individuals, formal supervisors, and the effects their leadership style can have on their subordinates, whether that may be in terms of

productivity or motivation or otherwise. In contrast, adaptive leadership is much more

concentrated on leading-following processes in groups, which can be more fluid, and which in turn enables a group to adapt more easily to different situations (DeRue, 2011).

This is where the concept of employee empowerment comes in. Adaptive leadership is not just concerned with ‘‘getting followers to follow the leader’s wishes’’, but with getting employees to use their skills and experience when the situation asks for it and thus effectively acting as a leader when necessary to accommodate to changes (Lichtenstein et al., p. 4). Therefore, adaptive leadership aims to empower employees and increase their self-efficacy to make a group or organisation as a whole more adaptive. In addition, adaptive leadership envisions a leader as being very aware of the environment that the group operates in, noticing relevant changes and challenges and proposing plans to deal with such occurrences. When necessary, such an ‘adaptive leader’ will take risks or promote the taking of risks to adapt to the newly changed environmental circumstances (Heifetz, Grashow & Linsky, 2009, p. 172).

All in all, the introduction of adaptive leadership as an organisational vision seems to entail the adoption of different forms of leadership behaviour. In this context, actual types of leadership behaviour and employees’ preferences for these styles are examined.

14 2.2 Leadership behaviour

Considering that this thesis aims to measure employees’ preferences for actual leader behaviour, and not their preferences for the overall leadership paradigm of the organisation, the concept of adaptive leadership is on its own not enough to be analysed. As stated before, this is mostly an organisational or collective perspective and is, in general, less concerned with actual leader behaviour at the micro-level. However, it was concluded that implementing adaptive leadership also implies a change in leadership behaviour. Therefore, it is important to delve into the existing taxonomies of leader behaviour that are identified in the literature. This makes it possible to see what leadership styles exist, which style fits the adaptive framework at the level of the leader-employee relationship and compare this to the other leadership styles.

While it is possible to identify as many as 39 different leadership styles in the literature, such a great number of types of leadership complicates research (Hassan, Asad, Hoshino, 2016). Luckily, the many different leadership styles can be organized into several broader, more

overarching forms. Most scholars use a twofold distinction of leader behaviour that was developed in the 1950s and remains popular today. These categories are sometimes known by different names, which is largely due to the proliferation of leadership behaviour taxonomies. As a result of this, essentially very similar concepts are often used under different terms (Yukl, Gordon & Taber, 2002). On the one hand, there is task-oriented or job-centred leadership behaviour, which is primarily concerned with goal achievement and enhancing performance and productivity. This includes scheduling, planning, supervising and coordinating activities of employees, as well as clearly defining the roles of employees and the communication structures between them (Boatwright & Forrest, 2000; Ehrhart, 2012; Tabernero, Chambel, Curral & Arana, 2009). On the other, there is relationship-oriented, or worker-centred leadership behaviour. This includes behaviour that expresses respect for employees, establishes mutual trust, shows

kindness, and expresses appreciation. It allows for more participation and two-way communication between employee and supervisor then a task-oriented style usually does (Boatwright & Forrest, 2000; Ehrhart, 2012; Tabernero et al., 2009). Others prefer to use the terms ‘consideration’ and ‘structuring’ to denote these types of leadership behaviour (Vecchio & Boatwright, 2002). Initiating structure and consideration are roughly similar to task-oriented and relationship-oriented leadership (Tabernero et al., 2009).

Other categories of leadership styles that are often used are directive and supportive behaviour. Cruz, Henningsen & Smith (1999) define directive leadership as focusing on

15

where directive leadership can be linked to task-oriented leadership, which is very much focused on the completion of the task at hand. Combining the two perspectives as directive-task behaviour makes for a leadership style that is focused on goal achievement, close supervision, and the control and direction of follower actions.

Supportive leadership behaviour is ‘‘directed toward the satisfaction of subordinates needs and preferences’’ (House, 1996, p. 326), and includes being kind, considerate and helpful towards followers, being approachable, and looking out for the personal wellbeing of individuals (House, 1971). Therefore this behaviour is characterised by providing ‘‘emotional, informational, instrumental, and appraisal support to followers’’ (Rafferty & Griffin, 2006, p. 39). In practice, this translates into showing appreciation for the work of employees, giving constructive feedback, and encouraging employees to participate in decision-making (Yilmaz & Altinkurt, 2012). This is similar to how relations-oriented leadership behaviour has been defined above. Therefore, supportive-relations behaviour makes for a leadership style that is focused on creating a warm environment where employees are supported and recognised in their work, can participate in decision-making, and are respected for who they are. Many other types of leadership styles have been defined over the years, most of which tend to overlap with the definitions above in one way or another. For example, some scholars consider participative behaviour to be distinct from supportive behaviour (House, 1996). However, for this research, preferences for respectively directive-task and supportive-relations behaviour are considered as two of the dependent variables.

The third dependent variable and type of leadership behaviour that is examined is one that occasionally emerges next to the more established two categories. In an attempt to integrate the different works on leadership behaviour, Yukl, Gordon & Taber (2002) created a taxonomy. They argued that the traditional twofold distinction between task and relations behaviour does not reflect all leadership behaviours properly, and therefore point to the overlooked category of change-oriented leadership. Change-oriented leadership behaviour is what Ehrhart (2012) calls charismatic behaviour, and also carries some similarities to transformational leadership, as pointed out by Yukl, Gordon & Taber (2002) themselves. Many of the change-oriented leadership behaviours are what would be expected under an adaptive leadership framework. Identifying threats and opportunities in the external environment, providing new ideas, taking risks when necessary to promote changes, and stimulating and encouraging the creative thinking of employees are all change-oriented behaviours (Yukl, Gordon & Taber, 2002), and are also important factors in adaptive leadership. The difference between transformational leadership and change-oriented leadership is that change-oriented leadership does not necessarily include the

16

collective component of transformational leadership. In addition, a very central aspect to adaptive leadership is ‘empowering’, which Yukl, Gordon & Taber (2002) consider as belonging to the category of relations behaviour, instead of change behaviour. Increasing the self-efficacy of employees enables an organisation to be more adaptive and therefore allows for organisational change (Malone, 2001). This is why for this research, empowering-change behaviour is considered as the third leadership style and the style that can be expected to be visible at relations-level when an adaptive leadership vision is implemented.

What follows from this discussion is an adapted version of the taxonomy of Yukl, Gordon & Taber (2002), with the main differences being that empowering behaviour is

combined with change behaviour, and task-oriented behaviour is more explicitly assumed to be in its directive variant.

Table 1. Taxonomy of Leader Behaviour

Directive-Task Behaviour

• Planning, scheduling, and coordinating activities of employees • Clarifying tasks and roles of employees

• Close supervision and controlling/directing employee actions and performance • Focused on goal achievement and enhancing the productivity of employees

Supportive-Relations Behaviour

• Providing support and attention to the needs and preferences of employees • Expressing respect and appreciation for employees and their contributions • Giving constructive and positive feedback

• Consulting employees and emphasising communication

Empowering-Change Behaviour

• Identifying threats and opportunities in the external environment • Providing new and innovative ideas, strategy, or vision

• Taking risks when necessary to promote changes

• Stimulating creative and innovative thinking of employees

• Increasing the self-efficacy of employees to take the initiative in solving problems

17 2.3 Personality characteristics: the Big Five

Personality has not escaped the interest of scholars in the field of leadership. Bono & Judge (2004) point out that from the articles on leadership published between 1990 and 2004, around 12% included the keywords ‘personality’ and ‘leadership’. However, most of this interest has been directed to the personality of the leaders, in an attempt to discover which personality traits are linked to the exercise of which leadership styles (Bono & Judge, 2004). Personality in relation to the leadership preferences of employees has not received nearly the same amount of attention.

Nevertheless, the examination of personality traits is far from a new endeavour. Digman (1990) shows how over the course of the 20th _{century, different factor models have been}

proposed to bring more structure to the study of personality. The five-factor, or ‘Big Five’, model has emerged as the most commonly accepted version as a result of extensive analyses of

personality traits over many decades (Judge & Ilies, 2002). There have been many investigations into the reliability and validity of the Big Five model, and as a result, it has been widely debated (see for example Becker, 1999). While the model is not without criticism, it is still the ‘‘most pervasive and accepted characterization of personality traits’’ (Moss & Ngu, 2006, p. 72). Remarkably, most alternatives to the Big Five model are largely the same and often merely extensions of the original version (Moss & Ngu, 2006, Barrick & Mount, 1991). As Goldberg (1993) points out, the popularity of the Big Five is largely due to the fact that there have been various attempts to replace it by its critics, and yet all these attempts failed. Likewise, Barrick & Mount (1991) show that compelling evidence has been acquired that supports the five-factor model. They argue that the ‘‘model provides a meaningful framework for formulating and testing hypotheses relating to individual differences’’ (p. 23).

The ‘Big Five’ model assumes the existence of five different personality dimensions, that together accurately capture a person’s personality. These dimensions or factors are Extraversion, Openness, Agreeableness, Conscientiousness and Neuroticism. Each dimension is an overarching concept which includes many different more specific traits (Judge & Ilies, 2002; Goldberg, 1993). John, Naumann & Soto (2008) provide some basic definitions of these personality dimensions. Extraversion is defined as an ‘‘energetic approach toward the social and material world and includes traits such as sociability, activity, assertiveness, and positive emotionality’’ (p. 120). Individuals who score high on the Extraversion domain thus have a tendency to be very social, outgoing, dominant, talkative, and confident, in contrast to individuals that tend to be more reserved, passive, silent, serious and compliant (Goldberg, 1993; Watson & Clark, 1997). Openness is defined as describing ‘‘the breadth, depth, originality, and complexity of an individual’s mental and experiential life’’ (John, Naumann & Soto, 2008, p. 120). Openness has occasionally been

18

termed ‘openness to experience’, ‘open-mindedness’, ‘intellect’, or ‘culture’, and consists of tendencies to be more creative, curious, flexible, imaginative, and unconventional (Judge & Ilies, 2002; Goldberg, 1993). In contrast, a low score indicates a tendency for being down to earth, conservative, and pragmatic, and perhaps even shallow or ignorant. High scorers are also likely to have a broader range of interests (Goldberg, 1993; McCrae, 1987). Agreeableness can be

conceptualised as ‘‘a prosocial and communal orientation toward others … and includes traits such as altruism, tender-mindedness, trust, and modesty’’ (John, Naumann & Soto, 2008, p. 120). High scorers tend to be trustworthy, amiable, warm, good-natured, compassionate and

cooperative, whereas low scorers are argued to be more sceptical, hard-headed, and competitive, perhaps even hostile, distrustful, and selfish (Goldberg, 1993; Hogrefe Ltd, 2017).

Conscientiousness is explained as describing ‘‘socially prescribed impulse control that

facilitates task- and goal-directed behaviour, such as thinking before acting, delaying gratification, following norms and rules, and planning, organizing, and prioritizing tasks’’ (John, Naumann & Soto, 2008, p. 120). High scoring individuals are supposed to have a tendency for organising and planning, are achievement-focused, reliable, diligent, and thorough. Low scorers, on the other hand, are said to be more careless, unorganised, unreliable, chaotic, and easy-going (Goldberg, 1993; Judge & Ilies, 2002; Hogrefe Ltd, 2017). Neuroticism is concerned with ‘‘negative

emotionality, such as feeling anxious, nervous, sad, and tense’’ (John, Naumann & Soto, 2008, p. 120). This implies that individuals with a high score have a tendency to be worried, insecure, nervous, moody, embarrassed or even depressed. In contrast, low scorers are generally more relaxed and less affected by stressful situations, are more emotionally stable, confident and calm (Goldberg, 1993; Hogrefe Ltd, 2017).

2.4 The association between personality and leadership preferences The formulation of the hypotheses that guide this research is based on two theoretical mechanisms: similarity attraction and need satisfaction. These are explained before the hypotheses are formulated.

2.4.1 Theoretical mechanisms

To investigate whether personality characteristics are in some way related to the leadership preferences of individuals, hypotheses must be developed that posit such linkages. A useful starting point to formulate these hypotheses is provided by the assumptions that Ehrhart & Klein

19

(2001) make in their research into follower preferences of leadership styles. One assumption that they make that is relevant for this research is that ‘‘individuals’ preferences for … particular types of leaders are based, to a considerable extent, on (a) similarity attraction, and/or (b) need

satisfaction’’ (p. 155). Thus, according to this argument, there are two types of mechanisms that determine someone’s preferences for leadership behaviour.

Firstly, regarding the similarity attraction mechanism, the argument is as follows: people tend to prefer leadership behaviour that matches or is similar to their personal characteristics, attributes, and values (Ehrhart & Klein, 2001). The similarity attraction mechanism is far from new and has been a common explanatory mechanism since the 1970s. It has been declared one of the strongest possible relationships in behavioural sciences and has been backed up by

numerous studies, although it has also been criticized for ignoring other possibly relevant factors (Montoya, Horton & Kirchner, 2008). Montoya, Horton & Kirchner (2008) explain that similarity attraction theory is based on the idea by Byrne (1971) that if individuals share the same attributes or characteristics, they are more likely to view the other in a positive light. Similar individuals are likely to reinforce one’s worldview and logic and are therefore looked upon positively. However, these authors also point to perceived similarity as being potentially just as powerful as actual similarity in creating attraction. This assumption informs the hypotheses that are formulated, in the sense that it is assumed that individuals will tend to prefer leadership behaviour that is similar to their own personality characteristics. For example, someone who is very quiet and cautious is not likely to prefer a leader who is very outgoing and brash, because this is too dissimilar to the personality of the follower.

Secondly, the need satisfaction mechanism posits that needs can also inform the

leadership preferences of individuals. The assumption is that followers will prefer and be drawn to that type of leadership behaviour that fulfils their needs (Ehrhart & Klein, 2001). The need satisfaction argument is frequently used, praised, but also criticized in explanatory accounts of job satisfaction, performance, wellbeing, and other work-related attitudes (see for example Salancik & Pfeffer, 1977; Baard, Deci & Ryan, 2004). The core argument is that individual preferences develop in relation to the extent to which their immediate environment satisfies their needs. Following this perspective, individuals prefer situations that allow their needs to be met (Salancik & Pfeffer, 1977). Therefore, individuals are argued to prefer leaders whose behaviour meets their leadership needs. Together the two theoretical mechanisms of similarity attraction and need satisfaction form the basis for the hypotheses that guide the analysis.

A clarification might be required when it comes to the terms ‘needs’ and ‘preferences’. While the hypotheses are based on the assumptions of similarity attraction and need satisfaction,

20

this does not imply that this research focuses on both preferences and needs, as it solely aims to examine the preferences of employees. Yet, there can be a lot of confusion surrounding these terms. Baard, Deci & Ryan (2004) point out that there is a difference between needs defined as desires, or needs defined as nutriments. In this thesis, the former is defined as ‘preferences’ and the latter as ‘needs’. Desires, or preferences, ‘‘refer to a person’s conscious wants, desires, or motives’’ (Baard, Deci & Ryan, 2004, p. 2046). Preferences in the context of this thesis, therefore, concern the leadership behaviours that an employee ‘wants’ or ‘desires’ the most. A person can have a positive attitude towards more than one style of leadership behaviour. Therefore, what they ‘prefer’ is the leadership style towards which they have the most positive attitude. While individuals tend to be consciously aware of what they prefer, this is not necessarily the case for what they need. What an employee needs relates to that which would be best for them, which is not always identical to what they prefer or desire. Needs in this view are defined ‘‘in term of the nutriments (whether physiological or psychological) that are essential for survival, growth, and integrity of the individual’’ (Baard, Deci & Ryan, 2004, p. 2046). In addition, the satisfaction of needs can be a factor in determining an individual’s preference for a type of leadership, as explained by Ehrhart & Klein (2001). Piedmont, McCrae & Costa (1991) show that the Big Five personality factors correlate with 15 different kinds of psychological needs. Therefore, in this study, the needs are considered satisfied when it meets the psychological demands that follow from a specific personality trait. Thus, while needs can inform preferences, they are themselves not synonymous to those preferences. Therefore, this thesis only considers needs when they are argued to inform preferences through the mechanism of need satisfaction.

2.4.2 Hypotheses

First of all, following both the similarity attraction and need satisfaction argumentation,

Extraverts are hypothesised to prefer empowering-change leadership behaviour. The personality dimension of Extraversion is argued to be significantly related to a factor called ‘positive

emotionality’ (Watson & Clark, 1997). Extraverts are more likely to experience higher levels of positive emotional mood states, such as enthusiasm, energy, joy and engagement. They are therefore more active, experience less uncertainty, are more ambitious, and have more risk-taking tendencies than introverts. Nettle (2005) indeed also argues that individuals with high

Extraversion levels are more likely to take risks and seek new experiences. Therefore, considering that empowering-change leadership behaviour is centred around risk-taking, change, and allowing

21

employees to take the initiative, extraverted individuals may prefer this style because it is similar to their personality.

Moreover, Piedmont, McCrae & Costa (1991) point out that Extraversion is correlated with psychological needs that can point to a preference for empowering-change leadership behaviour. On the one hand, Extraverts’ psychological needs of Aggression and Achievement do not make it seem unlikely for them to appreciate directive-task leadership behaviour, which is inherently more focused on achievements, conflictual and authority-based than the other leadership styles. On the other hand, their strong need for Dominance and the negative

correlation with the need for Deference (to seek subordinate roles) might point to a preference for being in the role of a leader instead of passively ‘following’, as would be the case in a

directive-task leadership situation. The needs of Autonomy and Exhibition similarly point in the direction of preference for empowering leadership, as this would allow more freedom for the employees to behave autonomously and be in the spotlight themselves. A need for Change furthermore indicates their preference for exploring new ways of doing things. Therefore, it is hypothesised that Extraverts prefer empowering-change leadership behaviour as this is both more similar to their personalities, and satisfies most of their psychological needs. Furthermore, negative correlations between Extraversion and Succorance (to get emotional support) make it unlikely for Extraverts to have a strong preference for supportive-relations leader behaviour (Piedmont, McCrae & Costa, 1991).

Hypothesis 1: High levels of Extraversion will be associated with a preference for

empowering-change leadership behaviour

Secondly, Openness is fundamentally concerned with being flexible towards change and is therefore hypothesised to be associated with a preference for empowering-change leadership behaviour. People with a high Openness score are more likely to question authority (Hogrefe Ltd, 2017). Therefore, it is unlikely that these individuals would prefer directive-task behaviour, which is heavily authority-based. Rather, because of their innate tendency to test limits and embrace new and innovative ways of doing things (Judge & Ilies, 2002; Goldberg, 1993), empowering-change behaviour might be closest to their taste. Both because of the similarity of this leadership style to their fondness for experimentation and because it satisfies their need for Change

(Piedmont, McCrae & Costa, 1991), empowering-change leadership behaviour may be preferred by these individuals.

22

Hypothesis 2: High levels of Openness will be associated with a preference for

empowering-change leadership behaviour

Thirdly, Agreeableness is hypothesised to be associated with a preference for supportive-relations leadership behaviour. The characteristics of agreeable individuals imply that these individuals are concerned with establishing and maintaining positive relations in a group, and therefore avoid confrontation in favour of cooperation. This makes it reasonable to argue that they are likely not to prefer directive-task behaviour, which is less focused on cooperating and more on achieving goals. Similarly, Judge & Ilies (2002) find a negative correlation between Agreeableness and goal-setting motivation, which they justify by arguing that it is likely that ‘‘agreeable individuals set less ambitious performance goals because they are motivated more by communion’’ (p. 803).

Therefore, of the three identified forms of leadership behaviour, the type of leader that seems to be most similar to agreeable individuals is the leader that exercises supportive-relations

behaviour. This is also the style that most satisfies their needs for Intraception (to understand behaviour and emotions), Nurturance (to provide benefits for others), and Affiliation, and fits with the negative correlation between Agreeableness and the need for Aggression. The negative correlation with a need for Autonomy furthermore makes a preference for empowering-change leadership behaviour less likely, as they seem to have no strong need to be empowered

(Piedmont, McCrae & Costa, 1991).

Hypothesis 3: High levels of Agreeableness will be associated with a preference for

supportive-relations leadership behaviour

Fourthly, arguing from both the similarity attraction and need satisfaction assumptions, individuals with a high Conscientiousness score may be mostly drawn to leaders with directive-task leadership behaviour. Their tendency to be directive-task and goal-oriented is similar to what is to be expected from directive-task leadership behaviour, as is their fondness for organising and

planning. Not only is this leadership style very similar to the Conscientiousness personality factor, but in theory, it also meets the needs of Conscientious individuals. These needs are Achievement, Endurance (to carry through a task), and Order. They thus need to be well-organised and focused on completing the task at hand. Negative correlation with a need for Succorance also dismisses the possibility of supportive-relations leadership behaviour as preferred by Conscientious individuals, as they do not have a great need for sympathy and emotional support from their supervisors. They rather need more task-oriented guidance, which directive-task leadership

23

behaviour can provide (Piedmont, McCrae & Costa, 1991). For example, Colbert & Witt (2009) show how goal-oriented leadership contributes to the productivity of Conscientious employees.

Hypothesis 4: High levels of Conscientiousness will be associated with a preference for

directive-task leadership behaviour

Fifth, hypothesising the preference of people with high Neuroticism using the similarity attraction and needs satisfaction assumptions is a bit more complicated than for the previously discussed personality domains. Intuitively, arguing from the assumption of similarity attraction, it does not seem reasonable that an anxious follower would prefer a similarly anxious leader. Arguing from the assumption of need satisfaction, it is not immediately apparent if neurotic individuals need firm guidance, or to be let alone. However, both similarity attraction and need satisfaction in this instance might be a bit more nuanced. Felfe & Schyns (2006), in their research into the perceptions of leaders, emphasise the importance of similarity between leader and follower, also when it comes to Neuroticism. They conclude that ‘‘there is no evidence … that weak followers tend to perceive or prefer more transformational or charismatic leadership’’ (p. 727), in which ‘weak followers’ are those with high Neuroticism scores. They, therefore, reject the suggestion by Yukl (1999) that perhaps anxious and insecure followers are more inclined to favour transformational leadership. The rejection of the idea that neurotic individuals prefer more empowering-change oriented leaders is thus in line with the similarity attraction argument because neurotic people are less optimistic (Sharpe, Martin & Roth, 2011), and thus seemingly the opposite of leaders who optimistically advocate change and experimentation.

While the similarity attraction assumption thus dismisses the possibility of high scorers preferring empowering-change leadership behaviour, it does the same for the other previously defined styles of leadership. Judge & Ilies (2002) point out that in general Neuroticism is related to a lesser likeliness to be goal-oriented because anxiety leads to a lesser ability to control the emotions required to focus on a task. In addition, while they might enjoy structure, Moss & Ngu (2006) argue that the potential of conflict that a transactional leadership style embodies is very threatening to people with high Neuroticism levels. This line of argumentation rejects the

possibility that neurotic individuals prefer directive-task leadership behaviour. Besides, it has been shown that Neuroticism is related to a higher level of social anxiety, and in general individuals with a high Neuroticism score seem to have poorer social skills (Argyle & Lu, 1990). Therefore, from the similarity attraction argument, it similarly seems unlikely that high scorers prefer

24

supportive-relations behaviour, who emphasise the social dynamics that neurotic individuals shy away from.

These lines of argument all point to what neurotic individuals do not prefer, which raises the question of what they do prefer. The research by Moss & Ngu (2006) might shed light on this question. Their findings indicate that individuals with high Neuroticism levels favour a laissez-faire type of leadership, as it is argued that any type of involved leader might be considered a threat by neurotic individuals. Therefore, what is most similar to these individuals is behaviour that is more passive and avoidant. Thus, following the similarity attraction argument, Neurotic individuals are not likely to prefer leaders that are very ‘present’, whether that may be by advocating change, directing behaviour, or emphasising social relations. Piedmont, McCrae & Costa (1991) furthermore indicate that Neuroticism is not correlated with any of the

psychological needs that they define, except for a correlation with Succorance. From the need satisfaction point of view, it seems reasonable to assume that Neurotic individuals have a psychological need for emotional support from others, as would be provided by supportive-relations behaviour. Yet, the similarity attraction mechanism contradicts a preference for supportive-relations behaviour. Therefore, no conclusive hypothesis can be put forward regarding the possible relationship between Neuroticism and a preference for supportive-relations leadership behaviour. All in all, the assumptions of similarity attraction and need satisfaction, therefore, indicate that Neuroticism might be negatively associated with directive-task and empowering-change leadership behaviour.

Hypothesis 5: High levels of Neuroticism will be negatively associated with a preference for

25

3. Methodology

This chapter explains how the analysis is conducted. The data is collected using a vignette survey among Defence employees, and more specifically, those employees working in the DCC and DCSC. First, an overview is given of the organisational context in which the survey is conducted. Second, the strategy for data collection is explained. Third, the operationalisation of the variables is discussed, before fourth, the data analysis strategy is presented and fifth, the descriptive statistics and correlation analysis are explored.

3.1 Organisational context

The research is situated in the Defence Cyber Command (DCC), which was established on 25 September 2014 by then Minister of Defence Jeanine Hennis-Plasschaert (Folmer, 2014). Officially, the activities of the DCC focus on three core areas: defensive, offensive, and intelligence capabilities, all to ensure the cybersecurity of Defence (Ministry of Defence, n.d.). These are considered necessary efforts due to the worsening of international (cyber)security and rising geopolitical tensions (Ministerie van Defensie, 2018). The digital domain is now considered one of the domains of military activities, next to land, air, sea, and space (Claver, 2018; Folmer, 2014). The DCC is largely located in The Hague, and consists of the following departments: Staff, Cyber Operations, Cyber Reservists, and the Cyber Warfare and Training Center (CWTC). The CWTC is not located in The Hague, but in Soesterberg. As the centre of expertise, it enhances the knowledge of the cyber domain and strengthens the capacities of all parties

responsible for executing military cyber operations. The department Cyber Operations plans and executes military cyber operations at an operational and strategic level. The Cyber Reservists provide additional support for Defence in the cyber domain in varying capacities. Besides these departments, the Staff of the DCC supports its general functioning. Employed within the DCC is not only military and civilian personnel, but also a considerable number of reservists. Because of a shortage of cyber experts, and because of the added benefits of tapping into multiple sources of expertise, employing reservists, or part-time military personnel, has become increasingly

important to the DCC.2 _{In general, employing reservists has become one way to make the armed}

forces more ‘adaptive’, by employing people from all background in society with varieties of useful knowledge and expertise (Visser, 2019).

26

Next to the DCC, employees of the Defence Cyber Security Centre (DCSC) are also included in the study, as they are another major player in the cyber domain within Defence. The DCSC is part of the Joint IT Command (JITC), the IT supplier of Defence. The DCSC

contributes to enhancing the cyber resilience of Defence. It does so by preventing and

monitoring the security threats that present themselves and mitigating those where possible. In addition, the DCSC provides information on threats and vulnerabilities where necessary, including on operational, technical, and strategic level.3 _{This provides the context where the}

survey is conducted, which includes military, reservist, and civilian employees from all departments and staff of both the DCC and the DCSC.

3.2 Data collection

To test the possible relationships between personality characteristics and preferences for certain leadership styles, a survey was conducted among personnel of the DCC and DCSC. This was done by distributing paper-and-pen questionnaires during the last week of January and the month of February in 2020, in the DCC locations in The Hague and Soesterberg, at a general meeting of the cyber reservists, and at the DCSC location in Huis ter Heide. In this way, all departments and staff were covered in the survey, and plenty of time was taken to ensure a wide audience was reached within the DCC and DCSC. In addition, this ensures the inclusion of all three types of employees: military, reservist, and civilian. The choice for conducting the surveys in a paper-and-pen fashion was made because of the importance of protecting the privacy of individual

employees of the DCC, and the inevitable risk in this respect that using internet-based surveys would entail. To complete the survey, approximately 10 minutes were necessary. After more than a month of conducting the surveys at the locations of the DCC and DCSC, 184 employees had participated and completed the survey. Subsequently, the collected data was entered manually into a data analysis programme and the required analyses were conducted.

Of the 220 surveys that were distributed, 184 were completed, amounting to a response rate of roughly 84%. Although this is not identical to the actual response rate, it indicates the relatively high response to the survey. Being physically present in all locations of the DCC and DCSC, and distributing the surveys in a paper format, arguably enabled me to ensure a higher response rate than using e-mail or web-based surveys (see for example Shih & Fan, 2009). Due to the sensitive nature of the work of both the DCC and the DCSC, I am unable to report details of

27

the population of these organisations. This includes details concerning the size of the population and its exact composition. Nevertheless, there is no reason to assume that the sample gives a distorted image of the variables that are investigated. Similarly, while I cannot report any details on representativeness, there is no reason to assume that the sample in any way misrepresents the population. However, should reliability and validity checks warrant an examination of the response rate and representativeness, the justification can be requested from the author.

3.3 Operationalisation and initial data processing

The surveys were conducted to examine possible relationships between personality factors of employees (the independent variables), and their preferences for types of leadership behaviour (the dependent variables). These two types of variables, as well as the control variables, had to be operationalised to be included in the survey and examined in the analysis. The measurement of the variables is explained, as well as the initial processing of the data including the making of dummy variables, reliability tests, and a factor analysis.

3.3.1 Measuring the independent variables

The independent variables were measured by using the Big Five Inventory-2 Short Form (BFI-2-S), developed by Soto & John (2017). This questionnaire was created to provide a shorter form of the widely-used BFI-2 questionnaire, which consists of a total of 60 items that measure the Big Five personality domains Extraversion, Openness, Agreeableness, Conscientiousness and Neuroticism. Yet, for some research, responding to 60 items can take too much time and effort for participants, especially when this is only one part of a larger survey. Soto & John (2017) indicate that carelessness, fatigue, and annoyance of participants can in some instances warrant a shorter version of the questionnaire to be used. In addition, some research only calls for a quick measurement of personality traits before moving on to assess other relevant factors. For these reasons, the BFI-2-S was created, which includes 30 items that assess the same five domains, although the used terms are slightly different. Neuroticism is termed Negative Emotionality and Openness becomes Open-Mindedness. Nevertheless, the content of the domains remains the same.

The short form was used in this study for the same reasons as to why it was created. This paper does not aim for an in-depth investigation of personality but instead focuses on the more

28

general links between the larger personality domains and leadership preferences. The BFI-2-S is suitable because this research does not look at the lower-level facets in the Big Five domains, but only at the domains itself. This is important because, on facet-level, the short form of the BFI is only recommended for relatively large samples, due to otherwise limited reliability and validity (Soto & John, 2017). Although there is a slight loss of predictive power for the BFI-2-S

compared to the BFI-2, ‘‘the BFI-2 short forms capture the full measure’s domain-level pattern of substantive associations with external criteria quite well’’ (Soto & John, 2017, p. 75).

Therefore, since this research is only interested in the domain level, the increased efficiency that the short form offers is worth the slight decrease in predictive power that this entails.

In order to conduct the survey, the BFI-2-S has been translated into Dutch. An official translated version of the BFI-2 already existed, which I used as the baseline for my translation in as far as the same items were used in both questionnaires (see Denissen, Geenen & Van Aken, n.d.). Moreover, I had the questionnaire translated by two fellow students, after which I contrasted their translation with mine to make sure there would not be any discrepancies in interpretation.4

The Dutch translation of the BFI-2-S exists of a total of 30 questions, of which five sets of six questions combine to measure the Big Five personality traits. The traits are scored from 6 to 30. Three of the six questions that measure one personality factor are originally phrased negatively. The responses to these questions are reverse coded (i.e., recoded so that a 5 becomes a 1, etc.) and subsequently the results of the six questions are added up. The lowest score consists of six questions scoring 1, the highest score consists of six questions scoring 5; thus creating the 6 to 30 scale. Neuroticism is expected to be a mirror image of the other four personality traits, as this is considered an inherently unfavourable personality trait (Jackson & Wilson, 1994). For this study, the score itself is not as meaningful as is its comparison to scores on other traits. This is important to keep in mind when interpreting the results.

As each Big Five personality domain is created by combining the results of six questions, it is necessary to check for dimensionality and reliability. Cronbach’s alpha tests reported a reliability coefficient of over 0.7 for two of the personality traits and close to 0.7 for the other three. A factor analysis was run to check whether the dimensions observable in the data match the theoretically defined personality dimensions. While it was expected to find five dimensions, the factor analysis initially showed nine components. This can be explained by the facets that make up the dimensions. The five personality factors are not wholly unidimensional, but consist of a few facets that, when combined, indicate a personality trait. However, because these facets

29

measure intrinsically different aspects of a personality trait, it is not unlikely for individuals to score slightly different on one facet compared to another. Except for Extraversion, the other four personality traits are split into two components, thus explaining the nine components that the factor analysis reported, and the few Cronbach’s alpha scores that fall just below 0.7.

When forcing a factor analysis to result in five factors, the resulting components

correspond with the personality dimensions of the Big Five. Only two of the 30 aspects could be placed in a different factor than the theory postulates. The first of these concerns ‘has difficulty getting started on tasks’, which is a reversely coded part of the productiveness aspect of

Conscientiousness. Yet, the factor analysis shows it can also be placed in the Neuroticism component. This makes theoretical sense, as multiple studies show the relationship between procrastination and negative emotionality (see for example Flett, Blankstein & Martin, 1995; Saddler & Sacks, 1993). The second aspect is ‘feels secure, comfortable with self’, which is a reversely coded part of the depression component of Neuroticism. The factor analysis shows that it can also (in similarly reversed coding) be considered a part of the Extraversion dimension. Existing research confirms this association between Extraversion and positive emotionality (see for example Watson & Clark, 1997).

3.3.2 Measuring the dependent variables

The dependent variable consists of the employees’ preferences for leadership behaviour. The three types of leadership behaviour that have previously been discussed were directive-task behaviour, supportive-relations behaviour, and empowering-change behaviour. The preferences for these types of leadership behaviour were measured by a vignette survey.

Vignette studies have been frequently lauded as one of the most promising

methodologies in the social sciences, and are therefore increasingly used to study a wide variety of topics (Herzog, 2017; Liebe, Moumouni, Bigler, Ingabire, & Bieri, 2017). Vignettes can be defined as ‘‘short, systematically varied descriptions of situations or persons … to elicit the beliefs attitudes, or behaviours of respondents with respect to the presented scenarios’’ (Atzmüller & Steiner, 2016, p. 52). There are several advantages to this approach. A major advantage is pointed out by Herzog (2017), who argues that more conventional survey research tends to demand very simplistic answers to sometimes very complex questions. Vignette studies address this issue by making for potentially more realistic and accurate responses due to the more life-like decisions respondents have to make when answering the survey questions (Herzog, 2017). A related advantage of vignette studies is that since the questions are not asked directly, as

30

they would in more conventional survey research, it minimises the risk of social desirability bias. Since respondents are confronted with hypothetical situations or descriptions instead of direct questions, they are more likely to give honest answers when it comes to sensitive topics (Jilke & Van Ryzin, 2017). Specifically for this research, a third major advantage is the opportunity to assess the attitudes and preferences of participants. Regarding this advantage, it is important to make the distinction between explicit and implicit vignette studies, as indicated by Aguinis & Bradley (2014). They point out that implicit vignette studies, such as conjoint analysis, can be used to measure implicit processes of which the participants are usually not aware. Implicit vignette studies include an experimental component which makes it harder for the participants to be fully aware of what part of their judgement is being examined. In contrast, explicit vignette studies, or ‘paper people studies’, can be used for examining ‘‘explicit decisions, judgements, and choices or … behavioural preferences’’ (p. 354). These vignettes are descriptions of fictional situations or people that are evaluated by the participants of a study. Considering that this research focuses on preferences, this type of paper people study is employed.

These two types of vignette studies correspond to two different approaches to this kind of research methodology: a simple scenario approach and an experimental version. In the former, all participants evaluate the same scenarios, whereas, in the latter, the scenarios that participants get presented are created by randomly selecting variables from the different available variables. The simpler version is more common in the paper people type of study and is also used in this research. This does pose a disadvantage; namely that it allows for less experimental control over the research (Herzog, 2017). However, the advantage of the ‘simpler’ version is that it can be conducted in pen-and-paper based surveys, which is a requirement for conducting this research in the DCC and DCSC. A disadvantage of vignette research, in general, is the trade-off described by Jilke & Van Ryzin (2017) between on the one hand the complexity of vignettes, which

determines how realistic the vignettes are, and on the other the limit on the number of vignettes that can realistically be put to participants. While it might be an admirable goal to strive for a vignette that is as realistic and complex as possible, to mirror the real world closely, it is not feasible to present a large number of vignettes to participants.

The paper people type of vignette was used for examining employees’ preferences for different leadership styles. For each previously defined leadership style one vignette was created, which describes leaders acting in accordance with either directive-task, supportive-relations, or empowering-change behaviour. The three vignettes that resulted from this were presented to the participants. The participants were subsequently asked questions about their preferences for these fictional leaders. This is in some ways similar to Ehrhart & Klein’s (2001) research, which also