The dark side of Core self-evaluations

Master Thesis

The dark side of Core self-evaluations

Faculty of Economics and Business

MSc in Business Administration – Leadership & Management track

MSc BA

Roy Versteegh 10003989

Thesis supervisors:

Annebel De Hoogh Deanne Den Hartog

Final Version 28th _{of June 2015}

2

Statement of Originality

This document is written by Student Roy Versteegh, 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

Abstract

Research on the construct of Core self-evaluations (CSE) has shown that CSE is positive for leaders to score high on. Positive outcomes, like motivation, job satisfaction and job

performance, have been found to be higher when leaders score high on CSE. However, researchers have not taken into consideration that CSE might have a dark side to it. Today’s literature assumes that the higher a leader scores on the construct of CSE the better (more positive) the outcomes are likely to be. Very high level of CSE are also known as Hyper CSE. Consequently, this gives need to understand if CSE has a dark side to it. This study set off to look for a (positive) relationship between CSE and three dark sides of leadership. These three dark sides consist of two dark traits, Narcissism and Machiavellianism and one leadership style, namely autocratic leadership. In addition to looking at the whole construct of CSE, the four individual traits which encompass CSE (emotional stability, self-efficacy, self-esteem and locus of control) are also tested to see if they have a relationship with the above mentioned dark sides of leadership. Through a study of 162 leader/employee dyads from different companies in the Netherlands, this thesis explores if (Hyper) CSE has a dark side. Using correlation and regression analysis it is shown that CSE has no relationship with narcissism and autocratic leadership, while CSE is shown to be negative related to Machiavellianism. The results also show that emotional stability is negatively related to Machiavellianism and self-efficacy is positively related to Narcissism. These findings contribute to the

understanding on the dark side of leadership and CSE. This study has not found a dark side for the whole construct of CSE, while one of its trait (self-efficacy) does have a dark side to it.

Keywords:Core self-evaluations; narcissism; Machiavellianism; autocratic leadership; dark traits; leadership; self-efficacy; emotional stability

4

Table of content

2.1.4 Neuroticism / Emotional stability…………...……….……….………...….11

2.2 Possible Dark traits of CSE and autocratic leadership……….………11

2.2.1 Narcissism……….……….…………...……12

2.2.2 Machiavellianism………….………...….13

2.2.3 Autocratic leadership……….……….14

2.3 Research question and hypotheses ……….……….….15

3. Method…………...……….……….22

3.1 Procedure……….…..23

3.2 Sample………..………..24

3.3 Measures……….24

4. Data analysis and results………26

4.1 Data analysis………..………..…….26

4.1.1 Recoding……….26

4.1.2 Reliability ………..…27

4.1.3 Means, Standard deviations and correlations………....28

4.2 Regression………..……….…...29

5. Discussion……….………34

5.1 Discussion of the results………..34

5.2 Practical and theoretical implications……….……40

5.3 External findings………...43 5.4 Future research……….44 5.5 Limitations………..44 6. Conclusion………..………..45 Bibliography………47 Appendix………..51

5

1. Introduction

In today’s literature, job performance, job satisfaction and motivation are among one of the most researched topics. These three concepts are of utmost importance for successful companies. In search of what explains the role of these three concepts Judge, Locke and Durham (1997) came up with the term Core self-evaluations. This construct was originally proposed as a measurement of job satisfaction. However later research have also argued that the construct should be related to work motivation and, ultimately, to job performance (Judge, Erez, & Bono, 1998). Core self-evaluations (CSE) consists of four specific traits which play a central role for performance, satisfaction and motivation. These four traits are self-esteem, generalized self-efficacy, locus of control and emotional stability (also known as

neuroticism).

After the first paper by Judge, Locke and Durham (1997) about CSE more research was conducted on this construct, again showing the importance of CSE for job satisfaction, job performance and motivation (Judge & Bono, 2001; Judge & Erez, 2001; Bono & Judge, 2003). Thus the traits of CSE have proven to be important for employees to have and businesses should therefore search for employees who score high on these traits. However, CSE goes much further than just explaining the relationship between employees and performance. Research shows that if a leaders or managers score high on CSE it enhances follower’s perceptions of transformational leadership, leading to improved performance and motivation (Wang, Liden & Sun, 2012; Kim, Liden, Kim & Lee, 2014). Researchers have argued that the higher a manager scores on the CSE scale, the better it reflects positive behaviour of followers. Clearly CSE is a very powerful construct and it is important for good leader to score high on the four traits, and therefore on the construct of CSE.

However, while CSE is without a doubt very important for managers to score high on, research has thus far not checked if there is a possibility of scoring too high on CSE or if CSE

6

has a dark side to it. Very high levels of CSE are also known as Hyper CSE. Hyper CSE was first used in an empirical research by Hiller and Hambrick (2005). They define the concept of hyper CSE as very high levels of CSE and argue that it may have a negative effect on, for example, job performance. This may sound counter-intuitive for some, but when thinking beyond the bright layer of CSE and looking at extreme forms of, for example, self-esteem, the possible dark side becomes visible. Extremely positive self-views can have the same adverse effects associated with narcissism and hubris (Judge, Piccolo & Kosalka, 2009). This can happen because, for example, people may feel so good about themselves that they will act in a way that is close to being narcissistic. This hypothesized non-linear effect is shown in figure 1, here I question the assumption that higher levels of CSE are always positive for leaders to have. Figure 1 shows that when leaders get beyond the point of high CSE, were CSE has positive effects on job performance etc., and into the area of hyper CSE, negative traits and behaviours also come into the equation, making the benefit of high CSE go down. A possible linear effect could be that negative traits are positively related to CSE, this would mean that the higher levels of CSE a leader has, the higher they will score on negative leadership traits. In this paper I will look for these negative traits and behaviours by researching if leaders show higher signs of narcissism, Machiavellianism and autocratic leadership in the case of hyper CSE.

In summary, researchers have shown that high CSE is important for leaders to have, however little research has been done about a possible dark side of CSE and possible negative effects of Hyper CSE. In this thesis I will peer deeper into the potential dark side of CSE and try to answer if it is possible that hyper manager/leader Core self-evaluation will lead to showing more autocratic leadership and whether (hyper) CSE relates positively to other possible dark traits such as Machiavellianism and Narcissism.

7

Researchers generally expect CSE to be positive, if this research proves otherwise, the view of CSE has to be altered and CSE can no longer be seen as always good for leaders to have (at least not in the case of hyper CSE). If this research does not find a relationship between CSE and a dark side of leadership it will strengthen the general view of CSE being good no matter what.

First I present the theoretical background on CSE and the four traits of CSE. Then I will look at the negative traits (Narcissism and Machiavellianism) and a leadership style (autocratic leadership) of leaders. Finally hypotheses on how they are linked are developed. Next the research methodology will be described. Subsequently the results of the data analysis are reported and interpreted. After these results the discussion section will interpret the results obtained in the previous part, give theoretical and practical implications, as well as

limitations, external findings and possible suggestions to future research. Finally the conclusion will summarize this paper and its findings.

Figure 1 (The possible negative effect of Hyper CSE on the total benefit of CSE)

-2 -1 0 1 2 3 4 5 6 7 8

Low CSE High CSE Hyper CSE

8

2. Theoretical background

This theoretical background is first used to explain CSE and the four different traits which encompass CSE (self-esteem, generalized self-efficacy, locus of control and emotional stability). Secondly this part will look at two dark traits (Narcissism, Machiavellianism) of leaders and one leadership style (autocratic leadership), which may arise when leaders or managers have hyper CSE. Finally a research question and hypotheses will be developed.

2.1 Core self-evaluations

Many researchers have questioned which factors provide job satisfaction, job performance and motivation. Work characteristics fairness perceptions and value fulfillment were some of the factors that were investigated to find out how to provide job satisfaction in the workplace. These were the so-called situational/job characteristics approach studies. Dispositional factors were also investigated to be a part of job satisfaction because of its stability across time and specific job circumstance that affected the result in the dispositional approach. However these dispositional factors of job satisfaction got criticized for not taking the individual traits in consideration such as, individual subconscious, fundamental evolutions about themselves, own abilities and own control (Chang, Ferris, Johnson, Rosen & Tan, 2012). To prevent these limitations of individual’s traits, Judge, Locke and Durman (1997) extended the earlier ideas of characteristics and dispositional factors to integrate the individual traits. To do this they developed the construct of core evaluation (CSE). So the construct of core

self-evaluations is a relative new construct which measures the personality of an individual. The construct of CSE was originally developed for explaining job satisfaction. However later research has found that CSE is also related to job performance and motivation (Judge & Bono, 2001). The four personality dimensions involved in CSE (locus of control, emotional stability, generalized self-efficiency and self-esteem) can be perceived as a positive indicators for employees’ values and personal traits (Judge, Locke and Durman, 1997). These

9

four traits will be further discussed below. Individuals who score high on the construct of CSE believe in their own capabilities and see themselves as likely to succeed in, for example, getting a job done. CSE is an appraisal of one’s self-worth. However, CSE is broader than self-worth in that it also reflects beliefs in one’s capabilities (to control one’s life) and one’s competence (to perform, cope, persevere, and succeed) and a general sense that life will turn out well for oneself (Judge, 2009). CSE also appears to be important when it comes to leader effectiveness (Judge, Piccolo & Kosalka, 2009). Leaders who score high on the four traits, and thus score high on CSE, are better in leading an organization and will achieve higher organizational performance. Employee outcomes will therefore be higher when a leader shows high levels of CSE than when a leader scores low on overall CSE.

2.1.1 Self-esteem

Self-esteem is one of the four traits of CSE, self-esteem is a person’s overall emotional evaluation of own worth. Self-esteem can be seen as the global evaluations of oneself

(Baumeister, Smart & Boden, 1996). People with high self-esteem feel good about themselves and generally view things like their behaviour and looks as good and above average. People with high self-esteem are more confident and optimistic when completing tasks at work or at home. High self-esteem also leads to less depression and better motivation of employees. Leader with high self-esteem seem more confident towards their followers which enhances the overall evaluation of the leader. Traits such as confidence and optimism crop up when analysts think about very effective leaders (Baumeister, Campbell & Krueger, 2003), a leader with high self-esteem usually possesses these traits.

2.1.2 Generalized self-efficacy

Generalized self-efficacy is the second trait which is part of the construct CSE. Self-efficacy is the believe in one’s capabilities to organize and execute courses of action required in managing prospective situations. Efficacy beliefs influence how people act, think, feel, and

10

motivate themselves (Bandura, 1997). People with high self-efficacy are more motivated and believe more in their ability to complete tasks and reach their goals. Several studies have demonstrated the importance of self-efficacy for improving performance in the organizational context (Gist and Mitchell, 1992). Leaders who have a high levels of self-efficacy are more able to produce desired outcomes. Organizational and employees performances improves when the leader or managers shows high efficacy. If the leader expresses high

self-efficacy employees are also more likely to follow and thus show more desired outcomes (Gist and Mitchell, 1992).

2.1.3 Locus of control

Third trait which is part of CSE is locus of control. Locus of control simply means the extent to which individuals think they can influence and control events around them. Locus of control is divided in two segments, internal locus of control and external locus of control. Individuals with internal locus of control believes outcomes of their actions lies in their own control. People with internal locus of control believe they can control a broad array of factors

in their lives (Judge & Bono, 2001). For example when getting good results they will praise

themselves, similarly they also blame themselves when tests results are not good. Individuals with internal locus of control should be more job satisfied because they are less likely to stay in a dissatisfying job and are more likely to be successful in organizations (Judge & Bono, 2001). Individuals with external locus of control believes other events or external factors to be the cause of an outcome, so outcomes do not lie in their own control. For example they will blame the test as being too hard when getting a bad result or blame the teacher for not teaching well. Similarly they will say the test was easy when getting a good result. Good leaders are more likely to have internal locus of control rather than external locus of control and leaders with internal locus of control achieve superior performance (Anderson & Schneier, 1978).

11 2.1.4 Neuroticism / Emotional stability

Final and fourth trait of CSE is neuroticism. Neuroticism can be explained as reflecting the

tendency to be confident, secure, and steadiness (Judge & Bono, 2001). So neuroticism is an

emotional state related to guilt, envy, anger and anxiety. Neuroticism is related to lower well-being because individuals who score high on neuroticism are predisposed to experience negative effects, which in turn is related to negative job satisfaction (Judge & Bono, 2001). Good leaders generally score low on neuroticism. Individuals with high levels of neuroticism are less likely to be perceived as leaders generally and are more likely to be perceived as weak leaders (Peterson, Smith, Martorana & Owens, 2003). However, neuroticism, when it is used in the construct of CSE, is conceptualized as the opposite, namely emotional stability. This means that when leaders score low on neuroticism, they actually score high on emotional stability. From now on this thesis will mostly refer to neuroticism as its opposite, emotional

stability.

CSE is made up of the four above mentioned traits, hyper CSE encompass the same traits as CSE only than in extreme high form. For example, leaders with hyper CSE have extreme high emotional stability making them extreme confident, secure and have almost no anxiety. Leaders with hyper CSE will also have a very high internal locus of control, very high self-efficacy and have a very high self-esteem.

2.2 Possible Dark traits of CSE and autocratic leadership

In search of the potential ‘dark side’ of (hyper) CSE this paper will research the link of CSE with two traits associated with this dark side, namely narcissism and Machiavellianism. These two traits are chosen because they are part of the dark triad of personality (Paulhus &

Williams, 2002). Further, this paper searches if (hyper) CSE can be linked with autocratic leadership. Not many researchers have considered the possibility of CSE being bad but rather

12

think the higher CSE a leader has, the higher the positive effects are. Only a few researchers are considering the possibility of (hyper) CSE having a negative side but have yet to test it.

2.2.1 Narcissism

The first trait which is considered as a dark trait in relationship to CSE is narcissism.

Narcissists are individuals who have a high opinion of themselves, are self-centred, are given to grandiose fantasies, and are interpersonally manipulative (Judge & Hurst, 2007).

Furthermore, narcissists have a grandiose sense of self-importance, a tendency to exaggerate achievements, fantasies of power, excessive self-admiration, hostility toward criticism, and intolerance toward compromise (Resick, Whitman, Weingarden & Hiller, 2009). A person who is narcissistic does not only just expresses his or hers high sense of self-worth, research has also shown that narcissists tends to be aggressive (Rosenthal & Pittinsky, 2006). All of these traits and behaviours contribute to the fact that narcissism is generally not a positive trait for people or leaders to have, and that this inflated self-concept does not belong in a business environment.

As mentioned above, narcissists usually have a high need for power, therefore it is logical to assume that narcissistic people are frequently found in leadership positions. Maccoby (2003) argues that many narcissists are found in the top ranks of many

organizations, religious groups and governments. Leaders with high level of narcissism are generally motivated by their needs for power and admiration rather than empathetic concern for the constituents and institutions they lead, thus being bad for employees and

organizational performance (Rosenthal & Pittinsky, 2006). In their study the authors are defining narcissistic leadership as occurring “when leaders’ actions are principally motivated by their own egomaniacal needs and beliefs, superseding the needs and interests of the constituents and institutions they lead” (Rosenthal & Pittinsky, 2006, p. 629).

13

that narcissistic leaders are often perceived as arrogant and self-confident (De Vries & Miller, 1985). The link between self-esteem and narcissism is also found in research by Judge, Lepine & Rich (2006). Here the authors found a relationship between self- and other ratings of leadership and narcissism. In this paper I argue that people with extreme levels of CSE will act narcissistic because, for example, their high self-esteem and high self-efficacy will cause them to act as arrogant and view themselves as ‘untouchable’ (narcissistic). The possibility of CSE relating to narcissism is also suggested by Judge & Hurst (2007) who state that another means of looking at an overly positive self-concept is to consider narcissism. For example, high internal locus of control of leaders could possibly be perceived as the leaders being narcissistic. Imagine performance at a company being outstanding, a leader with extreme high internal locus of control will, instead of contributing the success partly to his followers, take credit for the success alone because he feels he and he alone is responsible for the success. High self-esteem can also be perceived as being narcissistic, individuals are so full of themselves that they become narcissistic.

2.2.2 Machiavellianism

The second dark trait investigated in this research is Machiavellianism. This trait is the employment of cunning and duplicity in statecraft or general conduct. This trait leads to mistrust in human nature, lack of conventional morality, opportunism, and lack of affect in interpersonal relationships (Drory & Gluskinos, 1980). People who scored high on

Machiavellianism are found to have a strong tendency to manipulate other people (Geis & Christie, 1965). So in summary, Machiavellianism can be defined as a strategy of social conduct that involves manipulating others for personal gain, often against the other's self-interest (Paulhus & Williams, 2002).

Machiavellianism is also part of the dark triad of personality (Paulhus & Williams, 2002), this personality, just like narcissism, is generally also bad for leaders to possess and is

14

bad for business. The intrinsic meaning of work is at risk to be lost under Machiavellian supervision. This is because these kinds of leaders are less willing to adhere to procedure or pursue lofty ethical and moral standards, instead concerned with maximizing opportunities to craft their own personal power (Judge, Piccolo & Kosalka, 2009). Leaders with hyper CSE may have such high self-worth that they will not listen to other people because they feel they know best. Machiavellianism is found to be negatively related to agreeableness (Ashton, Lee & Son, 2000). Individuals who score high on Machiavellianism tend to achieve their goals also at the expense of other, or at least without regard of others. Leaders with hyper CSE may have such high emotional stability (which as explained earlier is related to emotional feelings like guilt) that they might have no problem with reaching their goals at the expense of others. Also they could have such high self-efficacy that they may be not so willing to hand out tasks to others, or at least they will try to interfere with/manipulate the tasks in some way. Leader with hyper CSE may believe they are so good in reaching their goals they want to complete the tasks of other, making the leaders manipulate their employees into doing what they feel is best for completing the task. So in this paper I will research if leaders with hyper CSE also show higher signs of Machiavellianism.

2.2.3 Autocratic leadership

Finally, autocratic leadership will be researched in this thesis to check for a possible

association with hyper CSE. Autocratic leadership is associated with the way in which leader make decisions. Autocratic leaders makes decisions alone and have all of the control in decisions, this without discussion/consulting with their subjects. Autocratic leaders dictate methods and stages of goal attainment and are unconcerned about followers’ autonomy and personal development (Lippitt, 1940). Autocratic leaders are low on consideration and also show low supervisor support to their followers (Judge, Piccolo, & Ilies, 2004). Autocratic leaders will usually make decisions alone without consulting others, this leads to valuable

15

contributions from followers being lost because of this. Not having any input will also lower commitment of employees on decisions made by their leader (Kosgaard, Schweiger & Sapienza, 1995) and employee’s motivation and satisfaction will also decrease (De Cremer, 2006). These traits do not only affect individual commitment and contributions, also group work is negatively affected by autocratic leadership. Research has shown that autocratic leaders negatively influence group stability and effectiveness (Van Vugt, Jepson, Hart, & De Cremer, 2004).

In this paper the question if hyper CSE will lead to more autocratic leadership is researched. The leader may have such high self-esteem and such high confidence in their own-self that he does not want others to interfere or co-decide with him. He believes he is so good in making decisions that he will make all of these decisions alone. Also extremely high internal locus of control (another trait of hyper CSE) may lead the leader to the conclusion that he alone is responsible for outcomes produced, therefore he does not want (or is scared) to let others make decisions.

In summary, the above review of the literature has shown that successful and good leaders score high on the construct of CSE. Consequently they score high on self-esteem, high on self-efficacy, high on emotional stability (low on neuroticism) and have a high internal locus of control. However, the current literature has overlooked the possibly of CSE having negative effects on leaders functioning, this research gap will be addressed in this paper. In this paper I will investigated if scoring (extreme) high on these positive traits can lead to leaders having higher levels narcissistic, Machiavellian and showing more autocratic leadership.

2.3 Research question and hypotheses

Based on the above theoretical background and the research gap explained in the introduction the following research question is developed: Are managers/leaders with hyper Core

self-16 evaluations perceived as higher on autocratic leadership than managers/leaders with normal Core self-evaluations and does (hyper) manager/leader Core self-evaluations relate positively to the dark traits narcissism and Machiavellianism?

In order for this research question to be answered I will look at the four traits of CSE separately and at these traits combined in the construct of CSE. This is done in order to check which trait of CSE relates the most with the dark traits and autocratic leadership, if (hyper) CSE is related to these traits at all. A second benefit of this approach is, it makes sure that if hyper CSE is not related to the dark traits or autocratic leadership, it checks if possibly one of the traits of CSE is related, even though the construct as a whole is not. Therefore based on the explanations provided above, several hypotheses are developed, these hypotheses are also depicted in a conceptual model (see table 2).

As noted above, high leader CSE is usually linked to good leader attributes as leading an organization better and better organizational performance. CSE also encompasses believe in own capabilities and leader with high CSE will see themselves as likely to succeed. CSE is an appraisal of one’s self-worth, as explained above. In summary leaders with high CSE feel they are good at their job, feel confident at their work, and will be perceived as inspirational and motivational by followers. However very high levels of CSE might be a bad trait of a leader. When leaders have high CSE they might not accept input or help from others, because they think they are the best person for the job. Thus, leaders could shift from being perceived as inspirational and motivational by employees to being perceived as autocratic. Therefore the first hypothesis developed in this paper is:

H1a: Hyper CSE of leaders is related with higher levels of autocratic leadership than general CSE is with autocratic leadership

Next to this, it could be that when leaders show high levels of CSE they might also show signs of narcissism because they feel too good about themselves. They might have a very high

17

opinion of themselves and have grandiose fantasies because they have such high CSE. Therefore the next hypothesis is:

H1b: Hyper CSE of leaders is related with higher levels of narcissism than general CSE is with narcissism

Next to leaders being narcissistic and show autocratic leadership, leaders might also show signs of Machiavellianism. Because of their high CSE leaders might feel others are not good at their job, or at least not as good as they are themselves, leading to mistrust in employees. This could lead to the feeling that others should be manipulated in order for employees to do a better job. This leads to the next hypothesis:

H1c: Hyper CSE of leaders is related with higher levels of Machiavellianism than general CSE is with Machiavellianism

Besides checking for the whole concept of CSE, I will also check the relationship of the four traits which encompass CSE with the dark traits Machiavellianism and narcissism and with autocratic leadership. The four traits which encompass CSE (locus of control, self-esteem, self-efficacy and emotional stability) are different traits by themselves. So it could be possible that, for example, high self-efficacy is related with narcissism but high locus of (internal) control is not. For this reason the four traits will also be tested separately.

First trait for which hypotheses are developed is self-efficacy. As explained above self-efficacy is believe in one’s capabilities to organize and execute courses of action required in managing prospective situations. Managers with extremely high self-efficacy might believe their capabilities are so good, that they do not believe they need input from others. Leaders with high self-efficacy might feel they can complete the task at hand without help from others. These believes can possibly lead to leaders showing signs of autocratic leadership. Another reason for extreme high self-efficacy leading to autocratic leadership is that they are so motivated to complete a task they might simply overlook the possibility of working

18

together. Thus, managers with extremely high self-efficacy may show signs of higher autocratic leadership. Therefore the second main hypothesis formulated is:

H2a: Extremely high self-efficacy of leaders is related with higher levels of autocratic leadership than general CSE is with autocratic leadership

If leaders have so much confidence in their own ability to complete tasks and in their ability to reach their goals, it could cause them to develop a very high positive opinion of

themselves. As stated above narcissists are people who have very high opinions about

themselves. High self-efficacy might lead to a high opinion in themselves which in turn might cause leaders to be narcissistic. Leaders are so confident in their own capabilities and abilities that they start showing more narcissistic behaviour than leader who are less confident.

Therefore I will research in this paper if extreme high self-efficacy of leaders will lead to leaders being narcissistic. This bring us to the following hypothesis:

H2b: Extremely high self-efficacy of leaders is related with higher levels of narcissism leadership than general CSE is with narcissism

High self-efficacy could also lead managers to the idea that they are right and their method of working is the best way of getting things done. If leaders believe so highly in their own capabilities they might feel their way is the best way and people should act according to the leaders’ behaviours. This may lead to the idea that employees should be manipulated into the ways of doing things as the manager would. That is why I will also test the following

hypothesis in this paper:

H2c: Extremely high self-efficacy of leaders is related with higher levels of Machiavellianism than general CSE is with Machiavellianism

Next individual trait of CSE to be tested is locus of control. As explained in the

19

they believe outcomes of their actions lie in their own control. Leaders with extreme high internal locus of control might feel everything lie in their own control. If leaders believe that outcomes lie in their own control they might not want to give this control to other people, or they might not see a reason for discussing decisions with followers if they feel it is all in their own control. Similarly if leaders believe that bad results are their own fault they might feel they want close control over events/tasks and are not so keen on delegating these tasks. These feelings of high internal locus of control could lead to autocratic leadership. Therefore the third main hypothesis is:

H3a: Extremely high internal locus of control by leaders is related with higher levels of autocratic leadership than general locus of control is with autocratic leadership

The believe that they alone are responsible for outcomes may also lead to narcissism.

Especially in case of good performance or positive outcomes the leader might feel they alone are responsible for this success, because they feel they are responsible for the good

performance of the company. For example, a company has made a huge profit this year, leaders with high internal locus of control will praise themselves for this achievement and because their internal locus of control is so high they feel they and they alone, are responsible for these good results. This feeling of achievement could result in narcissistic behaviour. This high internal locus of control could also be explained as a feeling of “everything comes down to me” this can be explained as a typical narcissistic thought. Therefore next hypothesis to be researched is:

H3b: Extremely high internal locus of control by leaders is related with higher levels of narcissism than general locus of control is with narcissism

If leaders feel outcomes lie in their own control, they may not want to give tasks to other people in the organization. Leaders who have this high internal locus of control could be scared to hand out tasks because if the tasks goes wrong they will feel it is their fault. While it

20

is impossible most of the time to do all tasks themselves, leaders might want to manipulate their employees into doing what they feel is best for completing the tasks. This leads to the following hypothesis:

H3c: Extremely high internal locus of control by leaders is related with higher levels of Machiavellianism than general locus of control is with Machiavellianism

Third trait of CSE, emotional stability, will also be researched. This trait reflects the

tendency to be confident, secure and steady. If leaders have high emotional stability, meaning

that they are not easily thrown of their game when somethings go bad and are highly

confident, it could mean that they will act autocratic. This could happen because they have

such high levels of emotional stability and are so confident they feel they are up for any task

and do not see the need to delegate. This extremely high emotional stability could therefore

lead to autocratic leadership. So the third main hypothesis purposed in this paper is:

H4a: Extremely high emotional stability of leaders is correlated with higher levels of autocratic leadership than general emotional stability is with autocratic leadership

Next, because leaders with high emotional stability are confident and secure most of the time, they may feel they are better in some way than other people who show lower signs of

emotional stability. This might lead to leaders having a feeling of superiority over these people. As explained above people with extreme high emotional stability show very high confidence in themselves. As explained above narcissists are usually very self-confident. This feeling of extreme high confidence related to high emotional stability and feeling of

superiority could lead to leaders being narcissistic. Therefore the next hypothesis is:

H4b: Extremely high emotional stability of leaders is correlated with higher levels of narcissism than general emotional stability is with narcissism

Finally leaders who have extremely high emotional stability could also feel they can or need to control their employees who are not as emotionally stable as he/she is. If leaders have very

21

high emotional stability they are less likely to feel guilty, so they might have less problems with manipulating and controlling followers. These feelings could lead to leaders showing higher signs of Machiavellianism. This leads to the following hypothesis:

H4c: Extremely high emotional stability of leaders is correlated with higher levels of Machiavellianism than general emotional stability is with Machiavellianism

Last trait to be researched is self-esteem. Self-esteem is a person’s overall emotional evaluation of self-worth. Leaders with high self-esteem could possibly have the feeling they are up for anything and any job. They feel so good about themselves they do not feel they need help or input from their followers because they feel they already know what needs to happen. Extreme high self-esteem might lead to leaders thinking they can do the job without help from others and even the believe that they know what is best for everyone. They might feel so self-confident they want to do everything themselves. This high level of self-esteem might therefore lead to leaders acting autocratic. So the fifth main hypothesis purposed in this paper is:

H5a: Extremely high self-esteem of leaders is correlated with higher levels of autocratic leadership than general self-esteem is with autocratic leadership

as stated above in the theoretical background, narcissistic leadership has some overlap with esteem, researchers have found that narcissistic leaders are often perceived as self-confident. The line between self-esteem and narcissism is sometimes very thin. Extreme high self-esteem might just cross this line and cause leader to be narcissistic because of multiple reasons. Extreme high self-esteem of leaders could lead to grandiose fantasies, high opinion of themselves and self-centredness so that this trait leads to narcissism. So in this paper I also purpose the following hypothesis:

H5b: Extremely high self-esteem of leaders is correlated with higher levels of narcissism than general self-esteem is with narcissism

22

Finally esteem could also be related with Machiavellianism. Leaders with high self-esteem might want to manipulate others because they feel confident they can actually pull it off. They feel so confident and optimistic they want to make sure everything goes the way they want them to go because they feel their way is the best way of getting things done. In order to get the desired behaviour leaders could try to manipulate their employees. This could cause leaders to act in a way which can be viewed as Machiavellianism. Therefore the

following hypothesis will also be researched:

H5c: Extremely high self-esteem of leaders is correlated with higher levels of Machiavellianism than general self-esteem is with Machiavellianism

Figure 2 (Possible outcomes of hyper CSE)

3. Method

This method section will show how I am going to conduct the research needed to answer the research question and hypotheses in this study. First the data collection procedure is given,

Machiavellianism Narcissism Autocratic leadership Extremely high self-esteem Extremely high generalized self-efficacy Extremely high emotional stability of leaders Extremely high internal locus of control Hyper CSE

23

this will also include relevant information about the questionnaire. Then the sample’s characteristics are given and finally the measures of the (relevant) items included in the questionnaire are discussed.

3.1 Procedure

This thesis is part of the Master in Business Administration at the FEB department of the University of Amsterdam. The questionnaire of this thesis was developed by five students, with the supervision of two professors. The reason for five students developing the

questionnaire is that more manager/employee dyads could be collected. The goal was to collect at least 120 dyads in total. Two separate self-administered questionnaires were

developed, one for leaders/managers and one for followers/employees. The questionnaire was distributed online via the website Qualtrics.com to various functions, departments and

different hierarchical levels from a range of organisations in the Netherlands. The respondents consisted mostly of people in the close proximity of the researcher. So using convenience sampling, participants are asked to complete the survey using an online questionnaire.

Managers and employees were invited by email to fill in the questionnaire, this email served as an introduction, explanation and guarantee of anonymity of the survey. The email also including the purpose of the survey, this email (in Dutch) can be found in the Appendix A. In some instances the manager or employee were asked to send the link off the

questionnaire to their own employee/boss. The surveys were recorded anonymously but matching codes were used to link the leader and employee questionnaires. The data collection started in February 2015 and stopped in April 2015.

The questionnaire was only available in Dutch as only Dutch employees and leaders would fill in the questionnaire. However, the scales adopted for the questionnaire were from English studies, therefore the questions were first translated to Dutch. The questionnaire also collected data not relevant for this study in particular, but relevant for the other four students

24

mentioned above. While this increased the length of the questionnaire, it also increased the variety of the managers and employee and the amount of dyads collected. The leader survey would take around 10 minutes to complete and the employee survey around 15 minutes. Finally at the end of the questionnaire, respondents could fill in their email address to get the results of the master theses and it also offered the option to add comments or ask questions about the questionnaire.

3.2 Sample

In total 204 employees filled in the questionnaire, which gives us a response rate of 70.6%. After correcting and deleting unusable questionnaires we see that out of the 204 employees 180 employee questionnaires can be considered actually completed, giving a completion rate of 88.2%.

For the leader survey 215 leaders filled in the questionnaire which gives us a response rate of 74.9% for the leader questionnaire. After deleting non-usable responses 170

questionnaires could be considered completed, making the completion rate 79.1%.

After combining the questionnaires (matching the leaders with the employees) a total of 162 complete leader-follower dyads were collected. So the total of usable dyads which could be used for analysis were 162 dyads. These dyads are used for testing the relationship of CSE, locus of control, self-esteem, self-efficacy and emotional stability with the darks traits Machiavellianism and narcissism and with autocratic leadership.

3.3 Measures

As stated above the online questionnaire also gathered data not relevant for this paper. This part will only discuss the variables and data used in this paper, hence it will not discuss all data which was gathered. The survey consisted of several questions and items relevant for this paper. The questions used were adopted from extant literature. Questions are rated on a seven-point Likert scale, ranging from 1(Strongly disagree) to 7(Strongly agree) with 4 being

25

neutral. The overall scores are computed by averaging item means.

Core self-evaluations will be measured using 12 items based on based on Judge,

Locke, Durham & Kluger (1998). Within the questions about Core self-evaluations the four traits of CSE (Emotional stability, locus of control, self-esteem and self-efficacy) are embedded. Emotional stability and locus of control will be measured using three items, self-esteem will be measured using two items and self-efficacy will be measured using four items. Questions in this scale are for example: “Soms voel ik me depressief” (“Sometimes I feel depressed”) and “Ik bepaal wat er gebeurt in mijn leven” (“I decide what happens in my life”).

Leader Machiavellianism will be measured using 8 items based on Mach-IV scale by

Christie & Geis (1970). Examples of questions included in this scale are: “De beste manier om met mensen om te gaan is door ze te vertellen wat ze willen horen” (“The best way of interacting with people is by telling them what they want to hear”) and “Over het algemeen zullen mensen niet hard werken tenzij ze ertoe worden gedwongen” (“In general, people won’t work hard unless they are forced”).

Leader narcissism will be measured using 20 items based on Ames, Rose & Anderson

(2006). This scale includes questions as: “Ik zie mezelf als een goede leider” (“I see myself as a good leader”) and “Ik zal succesvol zijn” (“I will be successful”).

Autocratic leadership will be measured using 5 items based on De Hoogh, Den Hartog

& Koopman (2004). This leadership trait includes questions like: “Mijn leider is de baas en geeft bevelen” (“My leader is the boss and gives orders”) and “Mijn leider vindt dat er uiteindelijk één baas moet zijn” (“My leader thinks that in the end there should be only one boss”).

26

4. Data analysis and results

This section of the thesis will describe the steps taken which were needed to analyse the collected data and will also show the results of the analyses. First the basic data analysis will be explained, this includes the recoding and reliability analysis. After this I will use the whole range of data available to check for any correlation between the dark leadership traits,

autocratic leadership and CSE and the four traits. Finally (Curve estimation) regression analysis will be done in order to check for the hypotheses. To analyse the leader/employee dyads the computer program SPSS was used. This way we can establish if managers who have Hyper CSE and hyper self-efficacy, self-esteem, emotional stability and locus of control show higher signs of narcissism, Machiavellianism and autocratic leadership than leaders with normal levels of CSE.

4.1 Data analysis

This part will first explain how the data was recoded for analysis in the recoding section. After the recoding section a reliability analysis (Cronbach’s alpha test) will be conducted and explained. At the end a correlation is used to show relationships between the variables and a correlation table (table 4) is established.

4.1.1 Recoding

First items that are phrased so that an agreement with the item represents a low level of the construct being measured were recoded into different variables (counter-indicative items). In this research that means that some questions about Core-self evaluations (CSE) needed to be recoded. Of the twelve questions about CSE, six questions needed recoding. This were questions two, four, six, eight, ten and twelve.

27 4.1.2 Reliability

After recoding the items a reliability analysis was conducted to examine the consistence of the measurements. The reliability analysis was conducted the eight items used in this thesis namely, CSE, locus of control, self-esteem, self-efficacy, emotional stability,

Machiavellianism, Narcissism and Autocratic leadership. Cronbach’s alpha, which test if measures have the same construct (represent the estimator of internal consistency), was run on the eight items (see figure 3). Only items which scored Cronbach alpha above or equal to 0.6 will be used in this research. Normally only items with a Cronbach alpha above or equal to 0.7 are considered as reliable. But the four traits of CSE, which are measured in this thesis, only consist of a few questions as also explained above (for example, Locus of control is only measured by 3 questions). The threshold for Cronbach alpha used in this thesis is therefore lowered to 0.6. The Cronbach alpha depends partly on the number of questions used (Cortina, 1993), which in this case was too low at some items. Next to testing for the normal Cronbach alpha, reliability analysis also included the test if Cronbach’s alpha if would increase if an item was deleted. As can be seen below two of the traits (self-esteem and locus of control) did not have a high enough Cronbach alphas, these items are highlighted in the table. This means these traits will not be part of the analyses and also means that the hypotheses which include these items (H3 and H5) cannot be tested. The reason for this low internal consistency may be the low number of scales. Reliability analysis also showed that only with one variable deletion of an item was needed, this item was part of self-efficacy, here with the deletion of

item/question 8 the Cronbach alpha grew from 0.597 to 0.632. All other variables had high enough internal consistency so no more items needed to be deleted as this would not increase Cronbach’s alpha.

28

Figure 3 (Cronbach’s alpha’s)

4.1.3 Means, Standard deviations and correlations

With the use of SPSS a correlation matrix is established, here the combinations of variables are shown with the relevant correlations, see figure 4. This table also shows the means and standard deviation of each variable. In this paper the strength of the correlation will be determent with the correlation categorisation of Dancey and Reidy (2004), this means that a correlation of 1 is a perfect correlation, between 0.8-0.9 is a very strong correlation, a correlation between 0.5-0.8 means a strong correlation, between 0.3-0.5 is a moderate correlation, a correlation between 0.1-0.3 is a modest correlation, between 0-0.1 is a weak correlation and finally 0 is no correlation. The correlation table gives us six significant

correlations. First Leader CSE is negatively correlated with Leader Machiavellianism (-0.352) with p<0.01, this is a moderate negative correlation. Second Leader CSE is positively

correlated with leader emotional stability (0.761) with p<0.01, this positive correlation is strong. Third Leader CSE is positively related with self-efficacy (0.584) with p<0.05, this is a strong positive correlation. Fourth Leader Machiavellianism is modest positively correlated to leader Narcissism (0.190) with p<0.05. Fifth Leader Machiavellianism is moderate negatively correlated with emotional stability (-0.361) with p<0.01. Finally the sixth correlation between

Item Cronbach’s alpha

Core-self evaluations 0.743 Machiavellianism 0.807 Narcissism 0.888 Autocratic leadership 0.739 Locus of control 0.433 Self-esteem 0.309 Self-efficacy (Item 8 deleted) 0.632 Emotional stability 0.660

29

Leader Narcissism and self-efficacy is modest positively related (0.180) with p<0.05. However, for this research mainly the correlation of leader CSE, emotional stability and Self-efficacy to the dark traits (Machiavellianism and Narcissism) and the leadership style autocratic leadership is important. Which leaves three correlations, Leader CSE with Leader Machiavellianism, Emotional stability with Leader Machiavellianism and Self-efficacy with leader Narcissism. These three important correlations are highlighted in table 4.

Variables M SD 1 2 3 4 5 6 1.Leader CSE 5.54 0.63 (1) 2.Leader Machiavellianism 2.67 0.97 -0.352** (1) 3.Leader Narcissism 4.00 0.80 0.077 0.190* (1) 4.Autocratic lead 4.41 1.08 -0.069 0.077 0.044 (1) 5. Emotional stability 5.08 1.25 0.761** -0.361** -0.046 -0.079 (1) 6. Self-efficacy 5.93 0.63 0.584* -0.120 0.180* -0.047 0.147 (1)

** Correlation is significant at the 0.01 level (2-tailed) * Correlation is significant at the 0.05 level (2-tailed)

Table 4 (Means, Standard Deviations and Correlations)

4.2 Regression

In order to test the hypotheses stated above, a regression analysis will be conducted. This regression will test the relationship between the independent variables CSE, emotional stability and self-efficacy and the dependent variables Machiavellianism, narcissism and autocratic leadership. More specifically a (quadratic and linear) curve estimation regression will be conducted. This type of curvilinear regression will show for each of the dependent variables if they go up when the independent variables get higher. Meaning it can show if the ‘dark’ traits and autocratic leadership are higher when the independent variables are high. This type of regression is most suited to find an effect as predicted in the hypotheses. The

30

linear model will try to find a linear line and the quadratic model can see if the graph has a curve (for example, if narcissism gets higher when a point of hyper CSE is reached). First regression analysis will be conducted on Leader CSE and the dependent variables (Table 5 and 6), next regression analysis of Self-efficacy on the dependent variables (Table 7 and 8) will be done and finally regression analysis of emotional stability and the dependent variables (Table 9 and 10) will be conducted.

Independent: Leader CSE Dependent: Autocratic leadership

Equation R R Square Std. Error F t Sig. Constant B1 B2 Linear Quadratic .069 .069 .005 .005 1.038 1.042 .725 .364 -0.851 -0.156 .396 .695 5.038 5.510 -0.114 -0.289 .016

Independent: Leader CSE Dependent: Narcissism

Equation R R Square Std. Error F t Sig. Constant B1 B2 Linear Quadratic .077 .089 .006 .008 .814 .816 .908 .604 .953 -0.483 .342 .548 3.452 5.664 .101 -0.721 .075

Table 5 (Regression analysis with independent CSE)

As can be seen from this table only Leader CSE and Machiavellianism are statistically significant (p<0.05) with F=22.083, R=.124, R squared=.124, t=-4.699 and B=-0.546, with a Std.error of .922 for the linear model. The quadratic model is also significant (p<0.05) with F=11.316, R=.357, R squared=.127, t=-1.112, B1=-1.811 and B2=.115 with a Std.error of .923. Table 6 shows us the graph of the relationship between CSE and Machiavellianism for both the linear and quadratic model. Leader CSE and autocratic leadership are not significant related (p>0.05) and the same goes for Leader CSE and Narcissism (p>0.05) for both the

Independent: Leader CSE Dependent: Machiavellianism

Equation R R Square Std. Error F t Sig. Constant B1 B2 Linear Quadratic .352 .357 .124 .127 .922 .923 22.083 11.316 -4.699 -1.112 .000 .000 5.713 9.129 -0.546 -1.811 .115

31

linear and quadratic model, so no graph will be established in this thesis for these two non-significant relationships.

Table 6 (Curvilinear regression CSE and Machiavellianism)

Next regression analysis of the independent variable self-efficacy and dependent variables Machiavellianism, Autocratic leadership and Narcissism will be conducted (See table 7).

Independent: Self-efficacy Dependent: Autocratic leadership

Equation R R Square Std. Error F t Sig. Constant B1 B2 Linear Quadratic .047 .066 .002 .004 1.044 1.046 .350 .346 -0.592 -0.671 .555 .708 6.176 2.666 -0.047 -0.343 .300

Independent: Self-efficacy Dependent: Narcissism

Equation R R Square Std. Error F t Sig. Constant B1 B2 Linear Quadratic .180 .211 .032 .044 .795 .793 5.258 3.621 2.293 -1.031 .023 .029 4.516 2.838 .227 -0.531 .719

Table 7 (Regression analysis with independent Self-efficacy)

This table shows that only self-efficacy and narcissism are statistically significant related

Independent: Self-efficacy Dependent: Machiavellianism

Equation R R Square Std. Error F t Sig. Constant B1 B2 Linear Quadratic .120 .137 .014 .019 .968 .969 2.376 1.551 -1.541 0.606 .125 .215 5.235 0.975 -.120 0.306 -.432

32

(p<0.05) with F=5.258, R=.180, R squared=.032, t=2.293 and B=0.227, with a Std.error of .795 for the linear model. The quadric model is also statistically significant (p<0.05) with F=3.621, R=0211, R squared=.044, t=-1.031, B1=-0.531 and B2=.719, with a Std.error of 0.793. Table 8 shows the graph of the relationship between emotional stability and

Machiavellianism for the linear and quadric model. Because no significant relationship between self-efficacy and autocratic leadership and Self-efficacy and Machiavellianism has been found, no graphs will be established for these two relationships.

Table 8 (Curvilinear regression Self-efficacy and Narcissism)

Finally regression analysis of the independent variable emotional stability and dependent variables Machiavellianism, Autocratic leadership and Narcissism will be conducted (see table 9).

33 Independent: Emotional stability Dependent: Autocratic leadership

Equation R R Square Std. Error F t Sig. Constant B1 B2 Linear Quadratic .079 .126 .006 .016 1.043 1.041 1.010 1.278 -1.005 1.099 .316 .282 13.873 2.825 -.066 0.663 -.749

Independent: Emotional stability Dependent: Narcissism

Equation R R Square Std. Error F t Sig. Constant B1 B2 Linear Quadratic .046 .092 .002 .008 .806 .806 .331 .674 -.576 1.104 .566 .924 15.607 3.536 -.046 .558 -.609

Table 9 (Regression analysis with independent emotional stability)

This table shows that only emotional stability and Machiavellianism are statistically

significant related (p<0.05) with F=24.673, R=.361, R squared=.130, t=-4.967 and B=-.361 with a Std.error of .909 for the linear model. The quadratic model is also significant (p<0.05) with F=12.271, R=.361, R squared=.130, t=-0.773 and B1=-0.431 and B2=0.071 with a Std.error of .912. Table 10 shows the graph of the relationship between emotional stability and Machiavellianism for both the linear and quadratic model. Because no significant

relationship between emotional stability and autocratic leadership and emotional stability and Narcissism has been found, no graphs will be established for these two relationships.

Table 10 (Curvilinear regression emotional stability and Machiavellianism)

Independent: Emotional stability Dependent: Machiavellianism

Equation R R Square Std. Error F t Sig. Constant B1 B2 Linear Quadratic .361 .361 .130 .130 .909 .912 24.673 12.271 -4.967 -.773 .000 .000 13.897 4.122 -.361 -.431 .071

34

5. Discussion

The results above show that only one of the (sub) hypotheses (H2b) is supported. This part will first discuss the results and relate these results to the hypotheses. After this practical and theoretical implications for the findings will be given. Next the extra found significant correlations are discussed. This part will next give suggestions for future research after that. Finally this part will end with limitations of this research.

5.1 Discussion of the results

As can be seen from the results hypothesis 1 is not supported:

-H1a: Hyper CSE of leaders is related with higher levels of autocratic leadership than

general CSE is with autocratic leadership (Not supported)

-H1b: Hyper CSE of leaders is related with higher levels of narcissism than general CSE is with narcissism (Not supported)

-H1c: Hyper CSE of leaders is related with higher levels of Machiavellianism than general

CSE is with Machiavellianism (Not supported)

First concerning H1a, no significant relationship was found and this hypothesis is thus not supported. First the correlation table (table 4) did not find a significant correlation between CSE and autocratic leadership. This was the first evidence of no relationship between Hyper CSE and higher levels of autocratic leadership than normal CSE. Next to no significant correlation, the regression analysis (table 5) also did not shows any significant relationship between the two variables. This is indicating that Hyper CSE is not related with higher levels of autocratic leadership than normal CSE. This result shows that even when leaders think extremely positive of themselves and are extremely confident in their own abilities (as in the construct of hyper CSE) they will not show higher (or lower) signs of autocratic leadership than leaders with lower CSE, this in contradiction to what was hypothesized.

35

Second, also no significant relationship was found between CSE and narcissism

(H1b). Correlation analysis (table 4) did not show a significant correlation between these two

variables. The regression analysis (table 5) also did not show any significant relation between CSE and narcissism. These combined results lead to the conclusion that Hyper CSE is not related with higher levels of narcissism than normal CSE, dismissing hypothesis H1b. Meaning that when leaders have hyper CSE they will not be more (or less) narcissistic than leader with normal CSE.

Finally hypothesis H1c was also not supported. But while CSE had no (significant) relationship with Narcissism and autocratic leadership at all, a significant negative

relationship was found between CSE and Machiavellianism. The correlation table (table 4) gives us a negative correlation (-0.352, with p<0.01). The regression analysis also provides us with a significant negative effect, meaning that when CSE gets higher Machiavellianism get lower. This result is different from what was hypothesized, Machiavellianism got lower, even when CSE turned into Hyper CSE. However, while Machiavellianism got lower even to a point were Hyper CSE was reached, the effect negative effect of CSE on Machiavellianism got less strong. As can be seen in table 6, the quadric model becomes less linear when a point of Hyper CSE is reached. These results show that the higher leaders score on the construct of CSE, the less they will show signs of Machiavellianism. These findings were the opposite of what was hypothesized, Machiavellianism did not get higher when leaders showed Hyper CSE, instead levels of Machiavellianism kept going down (although not as fast as in normal CSE), thus hypothesis H1c is also not supported.

Next hypothesis 2 was partly supported:

- H2a: Extremely high self-efficacy of leaders is related with higher levels of autocratic leadership than general CSE is with autocratic leadership (Not supported)

- H2b: Extremely high self-efficacy of leaders is related with higher levels of narcissism leadership than general CSE is with narcissism (Supported)

36 - H2c: Extremely high self-efficacy of leaders is related with higher levels of narcissism than general locus of control is with narcissism (Not supported)

Hypothesis 2a is not supported, as no significant relationship between self-efficacy and autocratic leadership was found. Correlation analysis did not find a significant correlation between these two variables (p>0.05), indicating that autocratic leadership is not negatively or positively related with self-efficacy. The regression analysis (table 7) also did not shows any relationship between the two variables, indicating that extremely high self-efficacy of leader is not related with higher levels of autocratic leadership than normal self-efficacy. As explained in the theoretical background, people with high self-efficacy are more motivated and believe more in their ability to complete tasks and reach their goals. So even when leaders have extreme high self-efficacy they will not show higher (or lower) signs of autocratic

leadership than leaders with lower self-efficacy, this in contradiction to what was hypothesized.

Next hypothesis 2b is supported, a significant relationship was found between leader self-efficacy and leader narcissism. Looking at the correlation table, a significant positive relationship between self-efficacy and narcissism (0.180 with p<0.05) is found. This shows us that when self-efficacy of leaders get higher their level of narcissism also tend to rise. The regression analysis also provides us with a significant positive relationship between these two variables. The linear and quadratic model both show a significant relationship. As can be seen in the linear model (table 8), the higher leaders scored on the trait of self-efficacy the higher their level of narcissism will be. However, when looking at the Quadratic model a slightly different relation between self-efficacy and narcissism is found. It seems that when leaders have low/medium levels of self-efficacy leader narcissism drops (a negative effect), while there is an opposite effect when a leader shows high/Hyper levels of self-efficacy, leader narcissism rises, a positive effect (see table 8). This finding still supports the hypothesis but it places extra emphasis on the dark side of leader high levels of self-efficacy, moderate levels

37

of self-efficacy seems to be related to the lowest levels of leader narcissism. The results, thus, show support hypothesis 2b, extremely high self-efficacy of leaders is indeed related with higher levels of narcissism leadership than general CSE is with narcissism.

Finally hypothesis 2c is not supported, no significant relationship is found between leader self-efficacy and leader Machiavellianism. First, after conducting the correlation analysis no significant correlation was found (p>0.05), leading to the first signs of no

relationship between these two variables. Next to the correlation test, regression analysis also did not find a significant relationship, indicating that extremely high self-efficacy of a leader is not related with higher (or lower) levels of Machiavellianism than normal self-efficacy. This shows that even when leaders have extreme high self-efficacy they will not show higher (or lower) signs of Machiavellianism than leaders with lower self-efficacy, this in

contradiction to what was hypothesized.

As explained above hypothesis 3 could not be researched in this thesis due to the lack of internal consistency, therefore no discussion can be conducted on these hypotheses.

H3a: Extremely high internal locus of control by leaders is related with higher levels of autocratic leadership than general locus of control is with autocratic leadership (Not researched)

H3b: Extremely high internal locus of control by leaders is related with higher levels of narcissism than general locus of control is with narcissism (Not researched)

H3c: Extremely high internal locus of control by leaders is related with higher levels of Machiavellianism than general locus of control is with Machiavellianism (Not researched)

Next, the results shows us that Hypothesis 4 is not supported, no significant positive relationships were found between emotional stability and Machiavellianism:

H4a: Extremely high emotional stability of leaders is correlated with higher levels of autocratic leadership than general emotional stability is with autocratic leadership (Not supported)

38 H4b: Extremely high emotional stability of leaders is correlated with higher levels of

narcissism than general emotional stability is with narcissism (Not supported) H4c: Extremely high emotional stability of leaders is correlated with higher levels of

Machiavellianism than general emotional stability is with Machiavellianism (Not Supported)

First Hypothesis 4a is not supported, no significant relationship between emotional stability of leader and autocratic leadership is found. The correlation table shows no significant

correlation between these two variables (p>0.05), indicating that autocratic leadership is not correlated (negatively or positively) with emotional stability. The regression analysis (table 9) also did not shows any relationship between the two variables, showing that extremely high emotional stability of leaders is not related with higher levels of autocratic leadership than normal emotional stability. As explained in the theoretical background, emotional stability is reflecting the tendency to be confident, secure, and be steady. So even when leaders show

extreme high emotional stability they will not show higher (or lower) signs of autocratic leadership than leaders with lower emotional stability, this in contradiction to what was hypothesized.

Next hypothesis 4b is also not supported, no significant relationship is found between leader emotional stability and leader narcissism. Correlation analysis between the two

variables showed no significant correlation (p>0.05), this can be viewed as the first sign of no relationship (positive or negative) and thus no support for hypothesis 4b. Next to the

correlation analysis, regression analysis (table 9) also found no significant relationship

between the emotional stability of leaders and leader narcissism. These combined results from the correlation and regression analysis lead to the conclusion that extremely high emotional stability is not related with higher levels of narcissism than normal emotional stability, dismissing hypothesis H4b. The conclusion can be drawn that when leaders have extreme high emotional stability they will not be more (or less) narcissistic behaviour than leader with normal emotional stability.