Narcissistic leaders’ performance evaluations of agreeable and extraverted employees

30-6-2014

Student Maaike Postema

Student number 10632441 Supervisor A.H.B de Hoogh

Master Business Studies, Leadership and Management Faculty Economics and Business

University University of Amsterdam

Narcissistic leaders’ performance

evaluations of agreeable and

Table of contents

Abstract ... 2

Introduction ... 3

Literature review ... 4

Narcissism ... 4

Narcissism, preference and evaluation ... 5

Personality traits ... 9

Narcissistic leaders and agreeableness ... 9

Narcissistic leaders and extraversion ... 10

Narcissistic leaders and perceived respect ... 12

Method ... 13

Sample and procedure ... 13

Measurements ... 15 Dependent variable ... 15 Independent variable ... 15 Mediating variable ... 16 Moderating variables ... 16 Control variables ... 16 Results ... 18

Missing values and recoding ... 18

Reliability ... 20

Hypotheses ... 20

Altruism and assertiveness ... 26

Discussion ... 27

Future research ... 32

Limitations ... 34

Conclusion and practical implications ... 35

References ... 37

Appendix ... 44

Survey supervisor ... 44

Abstract

The theory of threatened egotism indicates that narcissists could react negatively to negative feedback, because they want to protect their self-esteem. It could be possible that narcissistic leaders feel more threatened by dominant and assertive (characteristics of the personality trait extraversion) employees and feel more supported by agreeable and modest employees. Thus, it is expected that the personality traits extraversion and agreeableness moderate the relationship between narcissism and performance evaluation. Furthermore, it is proposed that the extent in which the narcissistic leader feels respected by his/her employee mediates this relationship. Supervisors and employees were asked to fill in a questionnaire and the researchers used their network to collect data. In total, 161 supervisor-employee dyads completed the survey. The results show that none of the hypotheses is supported by the data. Nevertheless, narcissism, perceived respect and performance evaluation are significant related to each other. However, after testing these variables in a mediation model, there is no significant result. In addition, there are some signals that extraversion and agreeableness moderate the proposed indirect effect. Future research is necessary to clarify when narcissistic supervisors feel more or less respected by their employees. For example, the extent to which an employee admires his/her supervisor could be an interesting moderator.

Introduction

Most people know someone in their environment who is a narcissist, a self-loving, overconfident and arrogant person (Campbell, Goodie & Foster, 2004; Maccoby, 2004). Although there is a paradox in the literature whether narcissists will create more positive or negative effects in organizations, most studies have shown that eventually narcissists will take decisions which are mainly beneficial for themselves (e.g. Morf & Rhodewalt, 2001). In addition, they are more likely to become a leader than non-narcissists (Brunell et al., 2008; Nevicka, De Hoogh, Van Vianen, Beersma & McIlwain 2011a), especially in an uncertain context (Nevicka, De Hoogh, Van Vianen & Ten Velden, 2013). Consequently, Maccoby (2004) states that nowadays many leaders may have narcissistic characteristics.

Although many researchers have investigated the possible positive or negative effects of narcissistic leaders in organizations (e.g. Morf & Rhodewalt, 2001), little is known so far about their preference to work with different employees. Narcissists prefer certain characteristics in people who are close to them (Campbell, 1999; Tanchotsrinon, Maneesri & Campbell, 2007). Besides this, narcissists have inflated self-views and a grandiose self-image (Judge, LePine, & Bruce, 2006). Their constant struggle to maintain their grandiose self-image may affect how they see and evaluate their subordinates.

Drawing on the theory of threatened egotism, narcissists will react negatively to situations that threaten their self-esteem. This reaction is due to their defence mechanism, they want to protect their self-image at all cost (Baumeister, Bushman & Campbell, 2000). Therefore, narcissistic leaders are likely to prefer employees who enhance their self-esteem and disfavour employees who threaten their ego. Consequently, this could result in higher respectively lower performance evaluations of their employees.

For this study two personality dimensions (extraversion and agreeableness) of ‘The Big Five’ (Judge, Bono, Ilies, & Gerhardt, 2002) were investigated. It is expected that narcissistic leaders feel more threatened by extraverted employees, because people who are extravert are dominant and assertive (John & Srivastava, 1999; Judge et al., 2002). In contrast, the expectation is that narcissistic leaders prefer agreeable employees, because people who are agreeable are trustful and compliant (John &

Srivastava, 1999; Judge et al., 2002). In order to explain this relationship, it is also interesting to investigate the extent to which narcissistic leaders feel respected by their employees. Twenge and Campbell (2003) argue that narcissists feel disrespected and threatened when they receive negative feedback. This is because narcissists feel socially rejected. Therefore, it is expected that narcissistic leaders feel disrespected by extraverted employees and, in contrast, feel respected by agreeable employees. Consequently, this study examines the moderating effects of extraversion and agreeableness on the relationship between narcissism and performance evaluations. In addition, it is expected that leaders’ perceived respect mediates this moderating effect. In the following paragraphs, there will be a discussion of the present literature. In addition, hypotheses will be formulated and a conceptual model will be provided.

Literature review

Narcissism

Much research has been done on the topic narcissism. A narcissist can be described as someone who is overconfident (Campbell et al., 2004), self-loving, arrogant (Maccoby, 2004), skeptical, cautious (Hogan & Hogan, 2001) and as someone who has an overly high self-esteem (Morf & Rhodewalt, 2001). Furthermore, narcissists are searching for confirmation for their grandiose self-image (Maccoby, 2004) and they take decisions and actions that are mostly focused on self-enhancement (Campbell & Campbell, 2009). Therefore, they need power (Kets de Vries & Miller, 1985), achievements (Chatterjee & Hambrick, 2006) and people who admire them (Emmons, 1987). Although narcissists have a negative image, they also have positive qualities, namely they are able to attract followers with their great visions (Maccoby, 2004) and they are charming (Campbell, Hoffman & Campbell, 2011).

Besides the interest in different traits of a narcissist, recent research has also investigated how effective narcissists are as leaders in organizations. There is a paradox in the literature whether narcissistic leaders will create more positive or negative effects in organizations (Morf & Rhodewalt, 2001). Some researchers found that they have negative effects on performance and effectiveness in an organization (e.g. Blair, Hoffman & Helland, 2008; Nevicka, Ten Velden, De Hoogh & Van Vianen,

2011b) and some researchers found no effects or links (e.g. Chatterjee & Hambrick, 2006). Furthermore, the effect of narcissistic leaders on employees is investigated in different contexts. For example, when employees are proactive and eager to solve problems, they are less likely to be negatively affected by narcissistic leaders (Hochwarter & Thompson, 2012). Campbell et al. (2011) found that narcissistic leaders can be destructive in multiple ways, especially in stable organizations and in the long-term.Nevertheless, narcissistic leaders are charming in the recruitment process and they are more popular at first sight than non-narcissistic leaders. An explanation for this could be their manipulation skills, neat clothing, humor, charm and self-assurance(Back, Egloff & Schmukle, 2010). It seems that over time this popularity changes and that others perceive narcissists as hostile and arrogant (Paulhus, 1998). Apparently, they have difficulty to remain charming and display commitment (e.g. Campbell & Foster, 2002). Grijalva and Harms (2014) argue that this is because narcissists are not able to maintain healthy relationships.

Although narcissists have difficulty to maintain long-term relationships, they are more likely to emerge as leaders than non-narcissists. Narcissism appears to be a reliable predictor of emergent leadership that has been investigated by different researchers (Brunell et al., 2008; Nevicka et al., 2011a). Furthermore, research indicates that the preference for narcissistic leaders enhances in uncertain contexts (Nevicka et al., 2013). Hence, nowadays many leaders and managers are associated with narcissistic characteristics (Maccoby, 2004). In conclusion, although it is not clear whether narcissistic leaders will create more positive or negative outcomes in organizations, they are likely to emerge in leadership positions.

Narcissism, preference and evaluation

Although many researchers have studied the effects of narcissistic leaders, it is also interesting to investigate their preference to work with different employees. Little is known so far about which personality characteristics of employees are preferred by the narcissistic leaders. Clearly, narcissists prefer certain qualities in people who are close to them. Campbell (1999) shows that narcissists prefer people who admire them. Furthermore, Tanchotsrinon et al. (2007) also argue that they are attracted

to people who have a high status. This has been investigated by measuring the level of narcissism and preferences trough self-questionnaires. There are two reasons to explain these preferences. Firstly, narcissists are searching for people who admire them in order to enhance their self-esteem (Maccoby, 2004; Morf & Rhodewalt, 2001). Secondly, narcissists use relationships to maximize their social status. When someone else has a high status, they could use this for networking and increasing their own status. However, is it important that narcissists can use the status of their partner for their own interests and are not threatened by this. Otherwise, they could react negatively (e.g. Baumeister et al., 2000), since threat from a high status person is perceived as really intense for a narcissist. They will do everything to protect themselves, because their self-esteem is all what matters. Horton and Sedikides (2009) show that narcissists will not hesitate to negatively rate their evaluator, when they receive negative feedback. Whereas non-narcissists are more likely to show some mercy for low status evaluators, narcissists will evaluate both high and low status evaluators more negatively after a threat out of self-protection. Consequently, it could be possible that, in a supervisor-subordinate relationship, narcissists will also not hesitate to negatively rate the performance of an employee who threatens their self-esteem to restore their own self-image. Overall, it is important for the narcissist that there is a potential for self-enhancement. Campbell (1999) proposes that more research should focus on the preference of narcissists with different methods and in different situations, for example in an organization.

Furthermore, researchers have shown that narcissistic leaders are more likely to evaluate themselves with higher ratings compared to non-narcissists. However, employees rate their narcissistic leaders less positive or even negative. Consequently, narcissistic leaders see themselves (undeserved) as favourable leaders and this reflects their grandiose self-image (Judge et al., 2006). Although they seem to flatter their own ratings, it is not clear whether they are also more likely to provide inflated evaluations of their employees. Turban, Jones and Rozelle (1990) show that in general, supervisors’ liking of employees positively influences how they treat their employees and how they evaluate the performance of their employees. This effect is direct and indirect. Supervisor liking is

positively related to the work-relationship with the employee, which in turn could affect the performance of the employee. Besides this, supervisors’ liking of employees results directly in higher performance evaluations, since the supervisors are more likely to bias the evaluations of their employees.

Drawing on the theory of threatened egotism, people with a high self-esteem will react negatively when they feel threatened by negative feedback. This reaction is a defence mechanism, since they want to protect their self-image at all cost (Baumeister et al., 2000). Therefore, narcissistic leaders could feel threatened by certain types of employees, who are more likely to provide negative feedback. This threat could negatively influence the narcissist’s liking of these employees. Consequently, this could affect (positively or negatively) the evaluations of their employees, which depends on the narcissist’s liking of the employee and on their need for self-protection.

In addition, Grijalva and Harms (2014) are interested in the Narcissistic Leaders and Dominance Complementarity Model. They suggest to use this framework to measure to what extent the employees dominance level has an influence on the effectiveness of a narcissistic leader. The framework indicates that friendly behavior creates friendly reactions and that hostile behavior results in the opposite. Furthermore, dominant behavior is more likely to be related to submissive behavior. Grijalva and Harms (2014) argue that, since narcissists are more likely to have dominant characteristics (Emmons, 1987), submissive subordinates will create a harmonious relationship with their narcissistic leaders, whereas dominant subordinates will threaten the narcissistic leaders. The study of Shechtman and Horowitz (2006) shows that persons react more negatively to each other, when they both have dominant characteristics. Overall, based on the theory of threatened egotism and the dominant framework, it is expected that narcissists will have more harmonious relationships with submissive subordinates and less harmonious relationships with dominant subordinates. Combining this with the supervisors’ liking theory of Turban et al. (1990), this could result in respectively higher or lower performance evaluations. Grijalva and Harms (2014) suggest that more research is needed to further explain the relationship between the subordinates dominance level and the narcissistic leader.

Besides this, Judge, LePine, and Bruce (2006) acknowledge the fact that more research is needed on how narcissists react to threat, especially in an organizational context. In addition, Stucke and Sporer (2002) argue that the reactions of narcissists to their employees should be investigated in different contexts outside the laboratory. Although they investigated, inter alia, the effects of negative feedback on narcissists, this has been done in a laboratory setting.

In conclusion, there is a gap in the literature about narcissistic leaders’ preference to work with different types of employees. In addition, there is no information available about how narcissistic leaders react to different types of employees. Clearly, narcissistic leaders prefer specific characteristics in their partners (Campbell, 1999) and are likely to negatively influence evaluations when they feel threatened, because this positively influences their self-esteem (Baumeister et al., 2000; Horton & Sedikides, 2009). Furthermore, leaders are likely to bias performance evaluations when they have harmonious relationships with their employees (Turban et al., 1990). Combining this, narcissistic leaders could prefer certain personality characteristics in their employees. In addition, by protecting their self-esteem, this could result in higher or lower performance evaluations for employees. This depends on the extent the narcissistic leader feels threatened by the employee. Consequently, the following question is interesting to investigate: Which personality traits of employees will be preferred by narcissistic leaders and result in a higher performance evaluation? In the next paragraphs personality traits will be discussed in more detail and hypotheses will be formulated (see Figure 1 for the conceptual model of the question).

Figure 1: Research question

Personality trait employee

Narcissism Performance

Personality traits

Employees have different personality traits and researchers assume that these traits are linked to certain behavior. These reactions are seen as habits and people will show this behavior frequently (Digman, 1990). Often, to define personality traits of people, ‘The Big Five’ is used. ‘The Big Five’ has been investigated during the past century by many researchers to define relationships and to test the reliability. Although not every researcher concluded that ‘The Big Five’ is a reliable measurement-instrument, the majority accepts this measurement-instrument for assessing personality (e.g. Goldberg & Saucier, 1995; Mount & Barrick, 1998). The five different dimensions can be divided into neuroticism, extraversion, openness, agreeableness and conscientiousness (Judge et al., 2002). This study will focus on the two dimensions extraversion and agreeableness. Based on the theory of threatened egotism, it is expected that subordinates’ agreeableness and extraversion affect the degree to which narcissistic leaders feel respected. They search for employees who confirm their own inflated views (Morf & Rhodewalt, 2001). Therefore, they prefer employees who enhance their self-esteem and feel rejected by those who do not (Twenge & Campbell, 2003). As a result of this, the narcissistic leaders may be more inclined to evaluate the performance of an agreeable employee positive and the performance of an extraverted employee negative.

In this paragraph, the two dimensions will be further explained and the relationship with narcissistic leaders will be discussed. Subsequently, hypotheses will be formulated and a detailed conceptual model will be provided.

Narcissistic leaders and agreeableness

An agreeable person is associated with trustful, warm, compliant, gentle and sympathetic characteristics. They are less likely to be stubborn, demanding and to show-off (John & Srivastava, 1999; Judge et al., 2002). Furthermore, researchers investigated the link between agreeableness, aggression and vengefulness and they found negative relationships (e.g. Barlett & Anderson, 2012; McCullough et al., 2001). In addition, agreeableness and counterproductive work behavior are negatively related (Bolton, 2010; Mount, Ilies & Johnson, 2006). As a result of this, the narcissistic leader will probably prefer an employee with this personality trait. The agreeable person could be

associated with submissive behavior. It is more likely that the submissive employee and the narcissistic leader have a harmonious relationship with each other, based on the Narcissistic Leaders and Dominance Complementarity Model of Grijalva and Harms (2014).

Furthermore, narcissists need to exhibit their superiority (Morf & Rhodewalt, 2001) and they want to force their ideas in order to achieve the highest performance. Therefore, they need people who admire them to enhance their self-image and people who accept their ideas without any contradiction (Rosenthal & Pittinsky, 2006). Consequently, when narcissistic leaders argue that they want more teamwork, they are actually looking for a group of people who agree with their visions (Maccoby, 2004).

In conclusion, the personality trait agreeableness matches with the preferences of the narcissistic leader, therefore it is expected that they will evaluate these employees more positively. This could be conscious (they want them to stay nearby) or unconscious (they like them more than other employees), which may result in higher performance evaluations. The following hypothesis can be formulated:

H1: Agreeableness of the employee moderates the relationship between leader narcissism and the

performance evaluation of the employee, such that when agreeableness is high narcissistic leaders will evaluate employees more positively.

Narcissistic leaders and extraversion

People who are extravert are characterized to be social, assertive, dominant, active, warm, energetic, adventurous, zeal and enthusiastic (John & Srivastava, 1999; Judge et al., 2002). Although there is no link between extraversion and aggression (Barlett & Anderson, 2012) and there is a negative relationship between extraversion and counterproductive work behavior (Bolton, 2010; Mount et al., 2006), it is expected that a narcissistic leader does not prefer an employee with this personality trait. Since they are looking for people who accept and support their ideas, they do not prefer assertive and dominant employees who provide their own opinions. When narcissistic leaders feel being threatened

by someone, they react with negative emotions and aggression (Barry, Chaplin & Grafeman, 2006; Bushman & Baumeister, 1998; Stucke & Sporer, 2002; Twenge & Campbell, 2003). They try to boost themselves by devaluing this particular person and report more negative views (Morf & Rhodewalt, 2001). In addition, they are sensitive for critique and they react defensively when someone expresses their (negative) feelings (Maccoby, 2004). This defensive behavior of narcissistic leaders can be explained by the theory of threatened egotism. Since they have a high favourable view of themselves, they defend this view at all cost against someone who undermines this (Baumeister, et al., 2000). It is possible that an extraverted person is more likely, in their enthusiasm, to react (negative) to the ideas of their narcissistic leader. According to Martinez, Zeichner, Reidy and Miller (2008), it is not even necessary that narcissists are threatened directly with negative feedback. When there is a chance in the future that narcissists will receive ego-threat from someone, they behave more negatively than non-narcissists. Thereby, they have no sense for empathy and therefore they do not care what the effect of their behavior is on their employees (Maccoby, 2004).

In addition, drawing on the Narcissistic Leaders and Dominance Complementarity Model, narcissists are less likely to work harmoniously with other dominant persons (Grijalva & Harms, 2014). Persons who have dominant characteristics and have to work together, are more likely to show anger (Shechtman & Horowitz, 2006). Again, this could be explained by the theory of threatened egotism, since the dominant persons could feel threatened by each other.

In conclusion, the personality trait extraversion does not match with the preferences of the narcissistic leader, therefore it is expected that they will evaluate these employees more negatively. As mentioned before, this could be conscious or unconscious. The following hypothesis can be formulated:

H2: Extraversion of the employee moderates the relationship between leader narcissism and

performance evaluation of the employee, such that when extraversion is high narcissistic leaders will evaluate employees more negatively.

Narcissistic leaders and perceived respect

Although an agreeable person is associated with less career success than an extraverted person (Seiber & Kraimer, 2001), it is expected that the narcissistic leader evaluates the agreeable employee more positively and the extraverted employee more negatively. As mentioned before, this is due to the expectation that narcissistic leaders feel threatened by extraverted employees and prefer agreeable employees. When the narcissists feel threatened, they feel insecure (Maccoby, 2004) and social rejected (Twenge & Campbell, 2003). Twenge and Campbell (2003, pp. 262.) quoted a statement of a (probably) narcissistic person: “I insist upon getting the respect that is due me”. Clearly, the narcissistic leaders do not feel respected by their employees when they are socially rejected and when their arguments are not supported. Obviously, they need confirmation about their own inflated self-views (Morf & Rhodewalt, 2001) and their visions (Maccoby, 2004). Therefore, it is expected that perceived respect has a mediating role in hypotheses 1 and 2. The more narcissistic leaders feel respected by agreeable employees, the more preference they have for these employees and this results in higher performance evaluations. In addition, it is expected that this works vice versa with the personality trait extraversion. The following two hypotheses could be formulated (see figure 2 for conceptual model):

H3: Leaders’ perceived respect mediates the moderating effect of employee agreeableness on the

relationship between leader narcissism and performance evaluation

H4: Leaders’ perceived respect mediates the moderating effect of employee extraversion on the

relationship between leader narcissism and performance evaluation

Figure 1: Conceptual model

Narcissism Perceived respect Performance

evaluation employee Personality trait

Method

Sample and procedure

To assess the relationships between ‘leader narcissism’, ‘employee personality traits’, ‘performance evaluation of employee’ and ‘perceived respect’, this study uses quantitative research techniques. A survey was chosen to collect the data and this survey was provided at one single point in time to the respondents. Two surveys were designed by six researchers. The first survey was developed for the supervisor and the second survey for the employee.

The population consist of dyads of supervisors and employees working in organizations. There are a few specific requirements for selecting the respondents. Obviously, the supervisor and the employee should have regular and direct contact with each other. In addition, the work needed to be the primary occupation of the respondent. For example, students who work in a supermarket to earn extra money were excluded from the study. In order to succeed, the convenience sample technique was used (Saunders & Lewis, 2012). The group of six researchers used their network and asked family, friends, and/or neighbors to participate in this study. Since it is important to know which dyads belong together, the surveys were coded and linked. The surveys were provided both via the internet and on paper. The researchers decided per respondent what the most preferable approach was. However, this could result in bias and therefore we coded whether the survey was distributed on paper or whether it was used on the internet.

This has resulted in two different procedures. For the internet versions, the employees and supervisors were asked to complete the survey by email. The employees and supervisors received different links that directed them to the questionnaire. For the paper versions, the surveys were provided by person. The questionnaires were handed out in envelopes and respondents returned the completed questionnaires in sealed envelopes. In 23 cases, the employee provided the questionnaire in an envelope to the supervisor and in 10 cases it was vice versa. Employees and supervisors were not exposed to each other’s questionnaire and were not allowed to discuss the answers. They were told that this research is part of the project ‘Leadership and Behavior in Organizations’ which is conducted for the University of Amsterdam. Supervisors were asked to answer questions about themselves and

the behavior of their employees. Employees were asked to answer questions about themselves, their working conditions and the behavior or their supervisor. During the data collection, the ethical requirements of the University of Amsterdam were maintained. Participation was on voluntary basis and anonymity was guaranteed. Respondents were ensured that the gathered data only would be used for research purposes and will not be spread for others to use.

The aim was to collect data on 150 supervisor-employee dyads. In total, 464 persons were asked to participate in this study (232 supervisor-employee dyads). After four weeks we received 161 complete dyads, which results in a response rate of 69%. This is relatively high due to the fact that people in the researchers’ network were approached. They were more likely to participate in this study. Overall, 61 male and 100 female participants within the age of 20 to 61 (M= 30.86 and SD= 10.19) completed the employee survey. In total, 103 male and 58 female participants within the age of 22 to 62 (M= 40.78 and SD= 10.39) finished the supervisor survey. Consequently, men and women are not equally represented. In addition, the age category 20 to 29 of the employees is overrepresented, whereas the age categories of the supervisors are more equally spread (see Table 1). The education level of the employees and supervisors varies between High school, College, Post-Bachelor, Post-Bachelor, Master and PhD. In addition, the organizational context is different across the participants. Furthermore, 71.4% of the surveys is distributed using the internet and 28.6% of the surveys was on paper. Lastly, 83.2% of the respondents is Dutch and 16.8% is non-Dutch.

Table 1: Age categories

Age Employee

Age Supervisor

Frequency

Percentage

Frequency

Percentage

20-29

108

67,1

29

18

30-39

23

14,3

53

32,9

40-49

17

10,6

38

23,6

50-62

13

8,1

41

25,5

Measurements

All of the variables were measured using Likert-scales. Respondents were asked to indicate the degree to which they did agree with a certain statement. The Likert-scales are based on 7-points (1= very low agreement and 7= very high agreement), since this provides the respondent a broad range of options to choose from. Indeed, a 10-point scale has also a broad range, however the difference between option 9 and 10 is negligible (McDaniel & Gates, 2012). In addition, the measurement scales are translated from English into Dutch. To check the accuracy of the translation, the Dutch scales were back translated to English.

Questionnaire supervisor

Dependent variable

Leader Narcissism; The variable ‘leader narcissism’ is based on 40 items (Cronbach’s α: 0.934)

(Raskin & Terry, 1988). Originally, this scale is measured based on true-false items. However, there seems to be a new trend in the literature. Different researchers argue that by using a Likert-scale instead of true-false questions, the results will be more useful for statistical analyses and interpretations (Corry, Merritt, Mrugs & Pamp, 2008; Kubarych, Deary & Austin, 2004). In addition, this format will not negatively affect the outcomes (Barelds & Dijkstra, 2010). Lastly, the total questionnaire is based on Likert-scale questions, therefore this format matches with the other questions. An example statement of the Raskin and Terry (1988) scale is: ‘I have a natural talent for influencing people’.

Independent variable

Performance evaluation; The variable ‘performance evaluation’ is based on 7 items (Cronbach’s α:

0.852) (Williams & Anderson, 1991). This scale measures the perceived in-role behavior of the employees by their supervisor. This is the only variable used in this study with counter indicative items (items 6 and 7). An example item of the Williams and Anderson (1991) scale is: ‘My employee adequately completes assigned duties’.

Mediating variable

Perceived respect; The variable ‘perceived respect’ is based on 3 items (Cronbach’s α: 0.777) which

are part of the leader-member exchange scale (Liden, Wayne & Stilwell, 1993). Instead of measuring the extent to which employees respect their supervisor, the scale is adapted so that superiors rated the extent to which they felt respected by their employees. An example of an original item of the Liden, Wayne and Stilwell (1993) is: ‘I respect my supervisor's knowledge of and competence on the job’ and this is changed into: ‘I have the feeling that the employee respects my knowledge of and competence on the job’.

Questionnaire employee

Moderating variables

Extraversion and agreeableness; The variables ‘extraversion’ and ‘agreeableness’ are part of ‘The

Big Five’ personality traits and were measured by using the mini IPIP scale developed by Donnellan, Oswald, Baird and Lucas (2006). Ten items were used (5 items for measuring extraversion (Cronbach’s α: 0.729) and 5 items for measuring agreeableness (Cronbach’s α: 0.763). An example of an extraversion item of the Donnellan et al. (2006) scale is: ‘I talk to a lot of different people at parties’ and an example of an agreeableness item is ‘I am interested in other people’s problems’.

Control variables

The following variables are controlled for, since they could unintendedly influence the relationships between the variables in this study.

Gender; Men and women differ by their personality traits and behavior. In general, men seem to

score higher on narcissism than women and men are more inclined to behave negatively after receiving negative feedback (e.g. Ames, Rose & Anderson, 2006; Bushman & Baumeister, 1998). Since men are more sensitive for negative feedback, they could be more likely to influence the performance evaluations of their employees. Gender is measured as a dichotomous variable (1 = male, 2 = female).

Age; Young people are more likely to score high on narcissism than older people. Therefore,

Campbell & Twenge, 2003; Twenge, Konrath, Foster, Campbell & Bushman, 2008). Age is measured by an open-ended question and is used as a continuous variable.

Education; Although researchers have not investigated the relationship between the level of

education and narcissism yet, there seems to be an increase of narcissistic students at college level (Twenge et al., 2008). To ensure that the variable education does not affect the variables under study, the effect of this variable is controlled for. Education is measured by a multiple choice question (1 = High school, 2 = College, 3 = Bachelor, 4 = Master, 5 = Other). The item ‘Other’ is recoded into Post-Bachelor and PhD, to ensure that the results are ranked in ascending order (1 = High school, 2 = College, 3 = Post-Bachelor, 4 = Bachelor, 5 = Master 6 = PhD).

Tenure; As mentioned before, narcissistic leaders are popular at first sight (Back et al., 2010) and

motivate their employees by inspiring visions (Maccoby, 2004). Nevertheless, on the long-term they could negatively affect the performance of their team (Campbell et al., 2011; Nevicka, et al. 2011b). Therefore, the variable tenure has to be controlled for. Tenure is measured by an open-ended question and is used as a continuous variable. Months were recoded into years when it was necessary.

Type of organization; The personality traits extraversion and agreeableness could be called ‘niche

traits’. They only predict performance when critical criteria are present, such as a cooperative and helpful environment for the agreeable person. On the other hand, extraversion is stronger correlated with performance when training or management skills are important (Barrick & Mount, 2005; Barrick, Mount & Judge, 2001). Therefore, extraverted employees could receive higher evaluations than agreeable employees in a particular organizational context and vice versa. In order to control for this effect, the variable ‘type of organization’ is measured by an open-ended question and later divided into 11 dummy variables (1 = Healthcare, 2 = Business services, 3 = Technology, 4 = Education, 5 = Government, 6 = Logistics, 7 = Culture, 8 = Retail, 9 = Recruitment, 10 = Consultancy and 11 = Others).

Nationality: Culture has an important influence on how people act and behave. Consequently,

characteristic than others (Foster et al., 2003). This variable is measured by an open-ended question and later divided into two groups, namely 1 = Dutch and 2 = Non-Dutch.

Procedure (paper/internet questionnaires); Some respondents had to fill in the questionnaire on

paper and some online using the internet. There is a possibility that the results are biased, since different procedures are used. The perceptions of confidentially and anonymity could differ by the procedures. It seems that people perceive less privacy at internet surveys (Bates & Cox, 2008; Dennistion et al., 2010), although other researchers found no effect (Campos, Zucolotoa, Bonaféa, Jordani & Maroco, 2011). To ensure that the use of different procedures does not unintendedly influence variables under study, the (possible) effect of paper and internet questionnaires are controlled for. This is measured as a dichotomous variable (0 = paper and 1 = internet).

Results

Missing values and recoding

In the survey, the respondents were not forced to answer questions. Consequently, they had a free choice whether they wanted to respond to different statements. However, this has resulted in missing values for some variables. The HOTDECK function of SPSS is used to deal with the missing values. HOTDECK is not recommended when the missing values for one item are more than 10%. In addition, another variable that is correlated to the variable under study, but not part of the research study, is needed to use this function. The variables extraversion (1.2%), narcissism (0.6%) and perceived in-role behavior (0.6%) include missing values. These percentages do not exceed the maximum norm. The next step is to replace the missing values, based on another item/variable which correlates with the variable under study. By using the Pearson Correlation test, different matches are found. Extraversion correlates with assertiveness, narcissism with leader entitlement and perceived in-role behavior with organizational commitment (see Table 2 for these correlations). By using the correlations, the missing values are replaced.

Perceived in-role behavior is the only variable that includes counter indicative items (items 5 and 6). These items are reverse coded to ensure that the results are interpreted correctly.

Table 2: Mean, Standard Deviations and correlations for HOTDECK

Variables

Mean

SD

1

2

3

4

5

1. Narcissism

4.26

.73

2. Leader entitlement

3.36

1.19

.579**

3. Extraversion

4.69

.93

.049

-.034

4. Assertiveness

4.83

.78

-.044

-.054

.611**

5. Perceived IR behav. 5.79

.75

-.023

-.179*

-.035

.088

6. Org. commitment

4.99

1.09

.082

-.023

.033

-.007

.292**

* p<0.05, **p<0.01, one tailed test

Table 3: Mean, Standard Deviations and correlations for hypotheses

Variables

Mean

SD

1

2

3

4

5

1. Narcissism

4.26

.73

2. Perceived IR behavior 5.79

.75

-.023

3. Extraversion

4.69

.93

.049

-.035

4. Agreeableness

5.38

.75

-.015

.094

.408**

5. Perceived respect

5.09

.87

.414**

.159*

.033

-.094

Reliability

The reliability of the variables is computed using Cronbach’s alpha. It is important that the Cronbach’s alpha is at least 0.6, this indicates that the reliability of the variable is acceptable. The higher the score of Cronbach’s alpha, the higher the reliability of the variable (above 0.7 = good and above 0.9 = excellent) (Zinbarg, Revelle, Yovel and Li, 2005). The Cronbach’s alpha of narcissism (40 items) is 0.934. Therefore, this variable reliability is excellent. Furthermore, the reliability could not be improved by deleting certain items. The variables extraversion and agreeableness (both 5 items) score respectively 0.729 and 0.763 (good). Again, these scores could not be increased by removing items. Furthermore, the Cronbach’s alpha of perceived in-role behavior (7 items) is 0.852. Although the reliability could be improved by deleting item 6 to 0.881, this item is not excluded. The difference is too small to create a notable change. The Cronbach’s alpha of perceived respect (3 items) is 0.777. This score is good and no further changes could be made to improve the reliability.

Hypotheses

The variables are all based on Likert-scales. In order to test the hypotheses, the means of the items have to be computed. Before analysing the hypotheses, the data is checked on outliers. Although some items could be defined as outliers, none of the items is deleted. When removing these outliers from the analyses, the outcomes mainly remain the same.

The mean, standard deviation and correlations of these variables are presented in Table 3. Leader narcissism is significantly positively related to perceived respect (r = .414, p = .000). This indicates that narcissistic leaders are more likely to feel respected by their employees. In addition, perceived respect is significantly positively related to perceived in-role behavior (r = .159, p = .022). This means that leaders who feel more respected by their employees, evaluate the performance of these employees

higher. Lastly, extraversion and agreeableness are significantly positively related to each other

(r = .408, p = .000). Thus, employees who score high on the personality trait extraversion, are also more

Subsequently, a multiple regression analysis is conducted to test Hypotheses 1 and 2. Initially, I controlled for gender, age, education, tenure, type of organization, nationality and procedure. However, I found no significant correlations between these controls and the study variables and including them in the analyses did not affect the results. Therefore, the control variables were left out from further analyses. In addition, following Jaccard, Wan and Turrisi (1990) (see also Jaccard & Turrisi, 2003), the mean of the variables is subtracted from the total score to avoid multicollinearity. The check for multicollinearity indeed showed that the VIF score and Tolerance score were improved extremely and that there is only a small chance for multicollinearity.

Hypothesis 1 states that when employees score high on agreeableness, high narcissistic leaders will evaluate these employees more positively. This hypothesis predicts a moderating effect. Shown by the results (see Table 4), in step 1 narcissism (β = -.016, p = .836) and agreeableness (β = .129, p = .139) are not significantly related to perceived in-role behavior. In step 2 the interaction effect is added. The results still show that narcissism (β = -.012, p = .877) and agreeableness (β = .104, p = .240) are not significantly related to perceived in-role behavior. The interaction effect is indeed positive (β = .074), but not significant (p = .376). For a graphical presentation of the non-significant interaction effect, see Figure 3. Thus, although it seems that leaders high on narcissism are more likely to evaluate agreeable employees as more positive, results are not significant and Hypothesis 1 is not supported.

Hypothesis 2 states that when an employee scores high on extraversion, high narcissistic leaders will evaluate these employees more negatively. The results show (see also Table 4), that in step 1 extraversion (β = -.089, p = .320) is not significantly related to perceived in-role behavior. In step 2 the interaction effect is added. Extraversion (β = -.050, p = .577) is again not significantly related to perceived in-role behavior. In this model the interaction effect is indeed negative (β = -.126), but again not significant (p = .141). See Figure 4 for a graphical presentation of the non-significant interaction effect. As proposed, there is some indication that leaders high on narcissism are more likely to evaluate extraverted employees lower on performance. Nevertheless, these relations are also not significant and therefore Hypothesis 2 is not supported.

Table 4: Results of moderated multiple regression analyses for narcissism, agreeableness

and extraversion explaining perceived in-role performance

Perceived in-role behavior

Step 1

Step 2

Step 1 Narcissism

-.016

-.012

Agreeableness

.129

.104

Extraversion

-.087

-.050

Step 2 Narcissism X Agreeableness

.074

Narcissism X Extraversion

-.126

F

.823

.975

R² / Adj. R²

.015/-.003

.030/-.001

Δ

_F

_1.200

Δ

_R²

_.015

* N=161, standardized regression coefficients are shown * p<0.05, **p<0.01

Table 5: Results of the moderated mediation model

Predictor

B

SE

F

R²

Perceived respect

Constant

.002

.063

Narcissism

.482**

.088

Agreeableness

-.143

.094

Extraversion

.071

.077

Narcissism X Agreeableness .182

.124

Narcissism X Extraversion

-.027

.095

7.426**

.193

Perceived in-role performance

Constant

5.791

.058

Narcissism

-.102

.088

Perceived respect

.186*

.074

Agreeableness

.131

.087

Extraversion

-.053

.071

Narcissism X Agreeableness .070

.116

Narcissism X Extraversion

-.127

.088

1.885

.068

Figure 1: Interaction effect of narcissism and agreeableness on perceived in-role behavior

Figure 2: Interaction effect of narcissism and extraversion on perceived in-role behavior

Low agreeableness

High extraversion Low extraversion

Hypothesis 3 states that the leaders’ perceived respect mediates the moderating effect of employee’s agreeableness on the relationship between leader narcissism and performance evaluation. This means that narcissistic leaders feel more respected by an agreeable person and this results in a higher performance evaluation. As shown before, agreeableness of the employees does not significantly moderate the relationship between narcissism and performance evaluation. However, agreeableness could moderate the relationship between narcissism and perceived respect. In order to test Hypothesis 3, model 10 from PROCESS is used (Hayes, 2012). The first step is to check whether agreeableness of the employee moderates the effect of narcissism on perceived respect. In this model (see Table 5), narcissism is significantly positively related to perceived respect (β = .482, p = .000) and agreeableness is not significantly related to perceived respect (β = -.142, p = .131). The interaction effect is indeed positive but not significant (β = .182, p = .145). This means that the model does not meet the first condition of the moderated mediation model and further analyses are not required. In conclusion, Hypothesis 3 is not supported.

Hypothesis 4 states that the leaders’ perceived respect mediates the moderating effect of employee’s extraversion on the relationship between leader narcissism and performance evaluation. This means that narcissistic leaders feel less respected by an extraverted person and this results in a lower performance evaluation. As shown before, employee extraversion does not significantly moderate the relationship between narcissism and performance evaluation. Nevertheless, extraversion could moderate the relationship between narcissism and perceived respect. Shown by the results of PROCESS model 10 (see also Table 5), extraversion is not significantly related to perceived respect (β = .071, p = .360). The interaction effect is indeed negative but not significant (β = -.027,

p = .776). The first step of the moderated mediation model does not meet the first condition, which is

necessary for a significant moderated mediated effect. Thus, Hypothesis 4 is not supported.

Although it is not necessary to further discuss model 10, an extra Table is added in this paper (see Table 6). These results show some interesting findings. The indirect effect is in certain situations significantly moderated by agreeableness and extraversion. When both moderators are low (- 1 SD),

the indirect effect is significantly positively moderated (L = .004, U = .191). This is partly consistent with Hypotheses 3 and 4, since a lower score on extraversion should lead to higher perceived respect and performance evaluations, but a lower score on agreeableness not. In addition, when agreeableness is

average and extraversion low, the indirect effect is significantly positively moderated (L = .013,

U = .226). This is consistent with Hypothesis 4, because the lower the employee scores on extraversion,

the more the supervisor feels respected and the more likely the employee receives a high performance evaluation. Furthermore, when agreeableness is high and extraversion average, the moderation is again significant (L = .006, U = .255). As proposed in Hypothesis 3, the higher the employee scores on agreeableness, the more the supervisor feels respected and the more likely the employee receives a high evaluation. Lastly, the interaction effect is significant when agreeableness is high and extraversion is low (L = .018, U = 281). This is consistent with Hypothesis 3 and 4.

It could be possible that there is a 3-way interaction between agreeableness, extraversion and narcissism explaining perceived respect and perceived in-role behavior. To check this, model 11 and 12 are used from PROCESS. Nevertheless, after conducting these tests, the results show no significant 3-way interactions.

Table 6: Bootstrapping results for test of conditional indirect effects at specific values of

the two moderators

Mediator

Conditional

indirect effect

Agreeableness

Conditional

indirect effect

Extraversion

SE

Lower

Upper

Perceived respect -.750*

-.935*

.047

.004

.191

-.750

.000

.049

-.001

.188

-.750

.935

.061

-.011

.216

.000*

-.935*

.053

.013

.226

.000

.000

.052

-.000

.205

.000

.935

.062

-.002

.233

.750*

-.935*

.065

.018

.281

.750*

.000*

.062

.006

.255

.750

.935

.069

-.000

.269

Altruism and assertiveness

Two other multiple regression analyses are conducted to test Hypotheses 1 and 2 with similar variables. Although this is not hypothesized, the personality traits altruism and assertiveness of the employees could significantly influence the relationship between narcissism and performance evaluation. Altruism is the most relevant factor in agreeableness and assertiveness is the most relevant factor in extraversion for testing the hypotheses (Goldberg, 1999). Results based on these personality traits could be stronger, because less important aspects of agreeableness and extraversion are left out of analysis and cannot cloud relationships under investigation. Altruistic employees could please a narcissistic leader, because they are self-effacing. In this way the narcissistic leader could receive more attention from others. In contrast, assertive employees could threaten a narcissistic leader, since they are more likely to act in a dominant way (Goldberg, 1999). Assertiveness (Cronbach’s α: 0.818) and altruism (Cronbach’s α: 0.790) are both measured by 10 items, based on Goldberg (1999) and Goldberg et al. (2006).

Firstly, agreeableness in Hypothesis 1 is replaced by altruism. Therefore, it is expected that when employees score high on altruism, high narcissistic leaders will evaluate these employees more positively. After conducting the regression analysis, the results show (see Table 7) that in this model narcissism is not significantly (β = .007, p = .927) related to perceived in-role behavior. Altruism is significantly positively (β = .205, p = .011) related to perceived in-role behavior. As proposed, the interaction effect is positive (β = .085), but not significant (p = .288). In conclusion, Hypothesis 1 is also not supported, when agreeableness is replaced by altruism.

Secondly, extraversion in Hypothesis 2 is replaced by assertiveness. Consequently, the expectation is that when employees score high on assertiveness, high narcissistic leaders will evaluate these employees more negatively. Shown by the results (see also Table 7), in this model the interaction effect is positive (β = .019) and again not significant (p = .814). In conclusion, Hypothesis 2 is also not supported, when extraversion is replaced by assertiveness.

Table 7: Results of moderated multiple regression analyses for narcissism, altruism and

assertiveness explaining perceived in-role performance

Perceived in-role behavior

Step 1

Step 2

Step 1 Narcissism

-.001

.007

Altruism

.217**

.205*

Assertiveness

.071

.077

Step 2 Narcissism X Altruism

.085

Narcissism X Assertiveness

.019

F

3.032*

2.041

R² / Adj. R²

.055/.037

.062/.031

Δ

_F

_.578

Δ

_R²

_.007

* N=161, standardized regression coefficients are shown * p<0.05, **p<0.01

Discussion

This is the first time that a study examined the influence of personality traits of employees on the relationship between narcissism and performance evaluation. In order to explain the preference of narcissistic leaders, the theory of threatened egotism is used (Baumeister, et al., 2000). The purpose was to investigate whether narcissistic leaders feel less threatened by an agreeable employee and more threatened by an extraverted employee, which would result in respectively higher or lower performance evaluations. In other words, it was expected that the two personality traits moderate the relationship between narcissism and performance evaluation. In order to explain this relationship, perceived respect was investigated as mediator. The expectation was that narcissistic leaders feel more respected by an agreeable employee and less respected by an extraverted employee, which would influence the performance evaluations.

The first expectation was that an agreeable employee positively moderates the relationship between narcissism and performance evaluation. This is in line with the expectations of Baumesiter et al. (2000) and Grijalva and Harms (2014), since the narcissistic leaders feel less threatened by their employees. Although the interaction effect is indeed positive, the hypothesis is not significantly

supported. Different reasons could be given for this. Firstly, narcissists prefer persons who confirm their own inflated self-views (Morf & Rhodewalt, 2001) and who admire them (e.g. Campbell, 1999; Emmons, 1987; Maccoby, 2004) in order to enhance their self-esteem (Rosenthal & Pittinsky, 2006). This is extremely important for them. It could be possible that an agreeable person is not good enough for the narcissists. They are not only looking for people who agree with them and who accept their ideas (Rosenthal & Pittinsky, 2006), but they also need employees who really admire them. In this way they could create their own group of employees who admire them. Furthermore, narcissists prefer partners with a high social status, in order to increase their own status (Tanchotsrinon et al., 2007). Consequently, the agreeable employee is not interesting enough for the narcissistic leader in two ways. The employee does not enhance the narcissists’ self-esteem by admiration and does not increase the narcissists’ social status. Lastly, Horton and Sedikides (2009) show that narcissists do not hesitate to influence ratings when they are threatened. However, maybe they will not do this the other way around. This could indicate that the narcissistic leader does not evaluate employees more positively, when he/she is not threatened. The performance evaluation could remain stable.

The second expectation was that an extraverted employee negatively moderates the relationship between narcissism and performance evaluation. This is based on different researchers who argue that narcissists will react more negatively and aggressively, when they are threatened (Baumesiter et al., 2000; Barry et al., 2006; Bushman & Baumeister, 1998; Stucke & Sporer, 2002; Twenge & Campbell, 2003). Although the interaction effect is indeed negative, the hypothesis is not significantly supported. It could be possible that extraverted employees do not threaten narcissistic leaders. The personality trait extraversion is not related to aggression (Barlett & Anderson, 2012) or to counterproductive work behavior (Bolton, 2010; Mount et al., 2006). Nevertheless, extraverted persons are dominant (Judge et al., 2002) and it is likely that this threatens narcissists (Baumesiter et al., 2000). It could be possible that only extreme dominant employees threaten narcissists (Grijalva & Harms, 2014) and not extroverted employees. Furthermore, proactive employees could also be a threat for narcissistic leaders. Grant, Gino and Hoffmann (2011) have found in their study that dominant leaders could

increase the performance of a group. This was only the case when the subordinates were passive. When the subordinates showed dominant and proactive behavior, the effect reversed and the group performance decreased. Grant et al. (2011) suggest that dominant and proactive subordinates threatened the leader and therefore the leader ignored ideas from their team and made decisions on their own. This affected the motivation of the subordinates as they felt that their contributions were not recognized.

Another explanation could be, again, the level of admiration of employees. There is no information gathered whether extraverted employees admire their leader or not. It could be possible that this influences the moderating effect of employee’s extraversion on the relationship between leader narcissism and performance evaluation. As mentioned before, it is important for narcissists that people admire them to enhance their self-esteem (e.g. Campbell, 1999; Emmons, 1987; Maccoby, 2004). Consequently, when extraverted employees admire their narcissistic leader, this leader could evaluate these employees according to standard rules, or maybe even more positively, since they receive confirmation about their own grandiosity.

An additional expectation was that an altruistic employee positively moderates the relationship between narcissism and performance evaluation and that an assertive employee negatively moderates this relationship. Altruism and assertiveness are the most relevant factors in agreeableness and extraversion for testing the hypotheses (Goldberg, 1999). Therefore, the outcomes could become stronger, since less important aspects of agreeableness and extraversion are left out of analysis. Even when agreeableness is replaced by altruism and extraversion by assertiveness, the results remain insignificant. Consequently, altruism and assertiveness do not change the outcomes in this study significantly.

The third and fourth expectations were that perceived respect mediates the moderating effect of agreeableness and extraversion on the relationship between narcissism and performance evaluation. The expectation was that agreeableness positively moderates this relationship, because narcissistic leaders are looking for people who agree with their visions (Maccoby, 2004). When the agreeable

person supports the ideas of the narcissistic leaders, they feel accepted and more respected (Twenge & Campbell, 2003). In contrast, it was expected that extraversion negatively moderates this relationship, since when narcissists feel threatened, they feel lack of respect for their ideas (Twenge & Campbell, 2003). The first step was to check whether agreeableness and extraversion moderates the effect of narcissism on perceived respect. Indeed, the interaction effect of agreeableness is positive and the interaction effect of extraversion is negative, but both are not significant. A possible explanation for these two insignificant results could be that more variables influence this relationship. As mentioned before, the level of admiration of the employee could be very important for the narcissistic leader.

The most interesting finding is that narcissism is significantly positively related to perceived respect. In addition, perceived respect is significantly positively related to perceived in-role behavior. Nevertheless, when narcissism, perceived respect and perceived in-role behavior are tested in a mediation model, the results are not significant. This indicates that more variables need to be taken into account, to find significant relationships. For example, Stucke and Sporer (2002) made a distinction in their study between individuals who score respectively high and low on narcissism and individuals who score high and low on concept clarity. Although narcissists mostly have an inflated self-esteem, some of them could be insecure and unstable and this is defined as low self-concept clarity. Stucke and Sporer (2002) show that narcissists who have a high but unstable self-esteem, are more likely to react with anger and aggression on negative feedback than narcissists who have a high and stable self-esteem.

However, there are some outcomes which suggest that agreeableness and extraversion could moderate the relationship between narcissism, perceived respect and perceived in-role behavior. Firstly, it is important to mention that the sample is not representative (see paragraph Limitations). Supervisors and employees who have a poor-quality relationship, were less likely to participate in this study. This could result in biased outcomes, since harmonious relationships are overrepresented.

Secondly, it could be questioned whether ‘type of organization’ really measured the environment of an organization. Apparently, an agreeable person performs better in an cooperative and helpful environment, whereas the performance of an extraverted person increases when management skills are important (Barrick & Mount, 2005; Barrick, Mount & Judge, 2001). However, the control variable ‘type of organization’ does not completely cover the differences in organizational context. It could be possible that this unintendedly influences the relationships between the variables in this study.

Thirdly, shown by Figure 1 and 2, these outcomes seem to support the expectations. The higher an employee scores on agreeableness, the higher the performance evaluation. In addition, the higher an employee scores on extraversion, the lower the performance evaluation. Combining this with the underrepresentation of poor-quality relationships between supervisor and employee, the outcomes may become significant when the sample includes more variation. This means that poor-quality relationships also need to be included.

Fourthly, there is another outcome which seems to support the defined expectations (see Table 6). As proposed, the indirect effect is significantly moderated when agreeableness is high and/or extraversion is low. Only one moderation effect is partly consistent with the expectations, when agreeableness and extraversion are both low. Overall, this indicates that the personality traits agreeableness and extraversion could significantly moderate the relationship between narcissism, perceived respect and perceived in-role behavior.

In conclusion, three explanations could be given for the non-significant results. Firstly, there is a possibility that other variables significantly influence the relationship. Secondly, extraversion and agreeableness moderate the relationship, but the data is not representative enough to create support for the hypotheses. Lastly, it could be possible that extraversion and agreeableness only moderate the relationship, when another variable is high. For example, the level of admiration of the employee for the leader.

Future research

More research is needed in order to understand the relationship between narcissism, perceived respect and perceived in-role behavior. This study has shown some interesting relations between these variables. The variables narcissism-perceived respect and perceived respect-perceived in-role behavior are significantly related to each other. However, after testing these variables in a mediation model, there is no significant result. This indicates that other variables could influence the relationship. For future research, it is interesting to research in which situation narcissistic leaders’ perceived respect influences the performance evaluation of the employee. Different moderators should be investigated to gain more in-depth knowledge about this relationship.

According to the literature, narcissists are highly sensitive to threat (e.g. Baumeister et al., 2000; Judge et al., 2006; Stucke & Sporer, 2002) and narcissists are more likely to evaluate a person negatively after a threat (Horton & Sedikides, 2009). Future studies could examine in which situation a narcissist feels threatened by an employee. For example, Grant et al. (2011) suggest that dominant and proactive employees could threaten a dominant leader.

On the other hand, narcissists are highly sensitive to admiration (e.g. Campbell, 1999; Emmons, 1987; Maccoby, 2004). Whereas narcissists react negatively to employees who threaten them, they could react positively to employees who admire them. It is interesting to investigate whether narcissistic leaders are more likely to evaluate admiring employees positively and whether they are more likely to evaluate dominant and proactive employees negatively. In addition, more research is needed to examine if narcissistic leaders feel more respected by admiring employees and less respected by dominant and proactive employees.