1 Amsterdam Business School
The Effect of Narcissism on Reward Preferences
Student name: Bora Yilmaz Student number: 11080736 Thesis supervisor: M. Schabus MSc Date: 20 June 2017
Word count: 18.114
MSc Accountancy & Control, specialization Control Amsterdam Business School
2
Statement of Originality
This document is written by student Bora Yilmaz who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document are 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
Rewards are an important tool for companies to motivate their employees and eventually to achieve the company’s strategy. In order to have an effective reward strategy there should be a link between the reward provided by the company and the reward preferred by the employee. In the past there have been various studies which examined factors that
influence the reward preferences of employees such as gender or nationality. However, the relationship between narcissism, which has been increasing in recent years according to various studies, and rewards is unclear. In order to find this out the following research question has been formulated: does narcissism influences the reward preferences of employees? This research question has been answered in this study based on five
hypotheses. The findings indicate that narcissism does influence the reward preferences of employees: high narcissistic people prefer different rewards (such as recognition rewards) than low narcissistic people. The current management control tools do not take this into consideration and could be ineffective in guiding these employees. Organizations should therefore, in order to motivate narcissistic employees, choose specific rewards which matches their preferences.
4
Content
1 Introduction ... 6
1.1 Problem Definition and Motivation ... 7
1.2 Research Question ... 8
2 Theory and Hypothesis Development... 9
2.1 Theoretical Background ... 9
2.2 The Dependent Variable: Rewards ... 10
2.3 The Independent Variable: Narcissism ... 13
2.4 Hypothesis Development ... 15
3 Research Methodology ... 19
3.1 Method and Sample Selection ... 19
3.2 Rewards... 20
3.3 Narcissism ... 21
3.4 Control Variables ... 22
4 Results ... 24
4.1 Descriptive Statistics ... 24
4.2 Preliminary Analyses and Correlation Matrix ... 26
4.3 Regression Analyses ... 29
4.4 Additional Analyses ... 32
5 Conclusion, Limitations and Further Studies ... 35
5.1 Conclusion ... 35
5.2 Limitations ... 36
5.3 Further Studies ... 37
Literature ... 38
Appendix A: Narcissistic Personality Inventory-40 (NPI-40) ... 43
Appendix B: Narcissistic Personality Inventory-16 (NPI-16) ... 44
Appendix C: Surveys ... 45
Appendix D: Rotated Component Matrix ... 50
Appendix E: Variable Measurement ... 51
5 Figures
Table 1: The hypotheses of the thesis
Table 2: Descriptive statistics of the dependent and independent variables
Table 3: Gender and narcissism
Table 4: Faculty and narcissism
Table 5: Preliminary ANOVA tests on rewards
Table 6: Correlation matrix
Table 7: The standardized betas and significance
Table 8: The R square and Adjusted R square of NPI and rewards
Table 9: The standardized betas and significance of the individual power rewards
Table 10: The standardized betas and significance of the individual intrinsic rewards
Abbreviations
ANOVA – Analysis of Variance
CEO – Chief Executive Officer
GPA – Grade Point Average
KMO – Kaiser-Meyer-Olkin
MC – Management Control
MCS – Management Control Systems
NPD – Narcissistic Personality Disorder
NPI – Narcissistic Personality Inventory
6
1 Introduction
Companies accept that rewards help significantly to achieve their strategic goals. However, reward policies such as pay for performance should not just be copied from similar or good performing organizations. There should be a link with the reward practices and the human capital (Gross & Friedman, 2004). Gross and Friedman (p. 7, 2004) stated the importance of this link by the following: “But as a leading provider of business services discovered, unless your reward program elements are strategically integrated and effectively delivered, the outcome can be quite unexpected”. Therefore it is important that organizations take their total reward strategy in a considered manner (Gross & Friedman, 2004).
With the election of American president Donald Trump and trends such as the increasing usage of social media and self-focus it could be said that narcissism and
narcissistic individuals are unavoidable in today’s society (Buser, Cruz, Bolen, & Swift, 2016; Foster, Campbell, & Twenge, 2003; Somerville, 2015). In the past there have been various studies which examined the relationship between employee personalities / characteristics and the reward policy of companies in order to motivate employees as best as possible. Olejnik, Tompkins and Heinbuck (1982) researched the reward preferences of gender and found that male prefer equitable and females prefer equal rewards (Olejnik, Tompkins, & Heinbuck, 1982). Gunkel and Marjaana (2006) have compared and analyzed the preferences of employees from Germany and The United States of America. The researchers concluded that the preferences of non-financial rewards between the countries differ. Despite these researches there have been a few researches which have examined specifically the impact of narcissism and rewards (Dworkis, Nair, & Soderstrom, 2015; Young, Du, Dworkis, & Olsen, 2016).
In this thesis I will empirically research the preferences of narcissistic people on rewards. In order to link the reward practices with the human capital it is important to know the preferences of certain types of employees. This thesis contributes to the literature in various ways because in the existing literature there is a gap in the impact of narcissism on management control tools (or more specifically rewards), the reward preferences of various employee characteristics (or more specifically narcissistic employees) and specific rewards such as power rewards. This thesis will be in pursuance of closing the gap in the existing literature and provide useful insight for a total reward strategy.
7
1.1 Problem Definition and Motivation
In the past various studies have showed that narcissistic people are associated with counterproductive work behavior and/or fraud (Grijalva & Newman, 2015; Rijsenbilt & Commandeur, 2013). This is not surprising since narcissistic people have specific character traits such as being exploitative (taking advantage of others) and being egocentric
(American Psychiatric Association, 2013). A recent study by Young, Du, Dworkis and Olsen (2016) has shown the implications of narcissistic people and management control tools. This motivated me to examine the relationship between narcissistic people and rewards since narcissistic people have different attitudes and motives than non-narcissistic people. Furthermore, a misalignment in interest could lead as indicated above to significant negative effects for organizations.
There are various studies which have shown that narcissism has increased in recent years. Twenge et al (2008) have showed that since 1982 narcissism has increased in
comparison with the mean of 1979 – 1985. This is a complement for other studies which has found increasing individualistic traits such as assertiveness and extraversion (Twenge J. M., Konrath, Foster, & Bushman, 2008). In response to this research Twenge and Foster (2008) found between 2002 and 2007 that the increase in narcissism was doubled in comparison with the period which Twenge et al (2008) had researched. The findings implicate that students have more narcissistic characteristics than the previous generations (Twenge & Foster, 2008). These studies were conducted in the United States of America. However, narcissism is also been evident in other countries in the world (Cai, Kwan, & Sedikides, 2012).
Management control systems were developed in the 1950s, while the workforce between then and now are different. The increase in narcissistic employees in the workforce has led to different work attitudes and personalities. The current management control tools do not take these changes into consideration and could be ineffective in guiding these employees (Young, Du, Dworkis, & Olsen, 2016). News media like the Wall Street Journal are already noticing new forms of (reward) policies from companies such as Google which take the preferences into consideration of the newer generations (Kwoh, 2012). There should be a fit between the preferences of the employee and the management control design of an employer to avoid agency problems (more about this in the theory section). Previous studies have often focused on the impact of narcissism on Chief Executive Officer (CEO) level
8 (Rijsenbilt & Commandeur, 2013). This will be one the few studies that will empirically test if individual narcissistic people prefer management control tools different than
non-narcissistic people (Chatterjee & Hambrick, 2007; Dworkis, Nair, & Soderstrom, 2015).
1.2 Research Question
As stated in the introduction various factors do affect the reward preferences of employees and reward design of employers. The relationship between narcissism and rewards is not completely outlined. Dworkis, Nair, & Soderstrom (2015) have researched the preferences of narcissistic and non-narcissistic managers on how they design the reward structure of their company. However, it is questionable whether the findings hold for individual
employees. Employers and employees have a different perspective on control. For example, employers could prefer tight controls in order to avoid risk and obtain their goals with a higher probability. However, employees might not prefer tight controls since they could for example value creativity/freedom to act and want to obtain their goals without following strict rules. These implications show that the preferences of managers do not have to reflect the preferences as an employee. Next to this, Dworkis, Nair, & Soderstrom (2015) do not examine the relationship between narcissism and specific types of rewards. Their research focused more on specific characteristics of rewards such as public praise. Therefore I would like to further research the relationship between narcissism and reward preferences of employees. The following research question will be examined: Does narcissism influences the reward preferences of employees? The research question will be answered with five hypotheses which examine the preference differences of narcissistic and non-narcissistic people for recognition, individual/group, goal/non-goal, power and intrinsic rewards. In the second paragraph the relevant theories will be discussed such as the theoretical background, the dependent variable rewards and the independent variable narcissism. The second paragraph contains also the hypotheses of the thesis which are based on the theory. In the third paragraph the research method will be explained. The descriptive statistics and the findings of the research will be explained in paragraph four. The last paragraph will contain a final conclusion, the limitations of the study and the possibilities for further research.
9
2 Theory and Hypothesis Development
In this section I will explain the importance and the consequences of the agency theory in companies. Next to this I will explain how companies control the agency theory with help of management control tools. Furthermore, I will give a description of one of most used management control tools which are rewards. Lastly, an explanation of narcissism is given and five hypotheses will be developed based on the theory.
2.1 Theoretical Background
An agency relationship is a contract between an agent and the principal, where the principal give rights to the agent to make decisions on behave of the principal. This kind of a
relationship happens when there is a separation between ownership and control. The agency theory addresses an issue in this relationship, which is that the agent has a different kind of interest than the principal. The agent is able to take actions to increase his utility, while the principal bears the costs of these actions. In order to reduce the misalignment of interest, the principal can take certain actions. The principal can monitor the agent and the principal can control if the actions of the agent deviate from the desired actions (monitoring costs). Another way to reduce the misalignment between is to expend resources in order to compensate the agent for taking the desired actions on behave of the principal (bonding costs). It is not efficient to align all the actions of the agent which is the residual part of misalignment (residual loss). Monitoring and bonding costs are a part of management control and will be discussed in the next sections (Jensen & Meckling, 1976).
One of the first definitions for Management Control (MC) or Management Control Systems (MCS) was described by Simons (1994, p. 5), who formulated it as: “The formal, information-based routines and procedures used by managers to maintain or alter patterns in organizational activities”. MC has changed over time and various broader definitions are used nowadays. According to Merchant and Van Der Stede (2012, p. 723) a more exhaustive definition is as follow: “Management control includes all the devices or systems managers use to ensure that the behaviors and decisions of their employees are consistent with the organization’s objectives and strategies. The systems themselves are commonly referred to as the management control systems”. MCS help organizations to align the interests of the agent and the principal in order to achieve the objectives of the organization. Therefore it is a solution for agency problems.
10 The agency theory is the core of this thesis. The misalignment of interest led to the development of management control tools such as rewards which is the dependent variable of this thesis. The effectiveness or efficiency of management control tools depends however on situational factors (contingency theory) such as the independent variable narcissism. The tools which were created to control employees did not take the different characteristics of people into account (Merchant & Stede, 2012). In the past there have been various studies which examined the relationship between employee characteristics and the reward policy. Graham, Harvey and Puri (2013) have researched the psychological traits and attitudes of CEOs. They have found that the CEOs optimism and their risk attitude influence the corporate financial policies and their compensation (Grahama, Harveya, & Puri, 2013). As stated in the introduction gender and country also influence the preferences. Thus, the reward strategy to use as a company differs per individual. The one-size fits all approach could lead to a misalignment between the effort and goals of employees and employers and will eventually lead to agency problems (Snelgar, Renard, & Venter, 2013). In this thesis the needs and preferences of narcissistic individuals will be compared to non-narcissistic individuals. A reward strategy will be defined for these individuals based on the differences or not.
2.2 The Dependent Variable: Rewards
Pitts (1995) describe the definition of rewards as follow: “reward is the benefit that arise from performing a task, rendering a service or discharging a responsibility” (Pitts, 1995, p. 11). Rewards are used to motivate people in order to maximize their efforts regarding a performance (Armstrong & Stephens, 2007). In order to understand this concept it is important to understand what motivation is. A motive is the reason why someone is choosing to act in a certain way. Employees are motivated if they expect that acting in a certain way will eventually lead to the realization of an objective (theory of Vroom) and will receive a reward which satisfies their needs (Maslow’s theory) (Armstrong & Stephens, 2007).
Maslow’s hierarchy of needs theory describe five various types of needs, which are the physiological, safety, love, esteem and self-actualization needs (Griffin, Ledbetter, & Sparks, 2014). The physiological needs are very basic in nature such as food and sleep. If these needs are not met the individual will suffer from hunger or fatigue. The individual will
11 seek to meet his body's requirements in order to find a balance. After the individual have found physical comfort, they will seek to find stability and consistency (safety needs). This helps the individual to find certainty and predictability in their future. The needs of love are giving and receiving love. Giving love is based on the comprehension and accepting others for who they are. Receiving love is based on being with and getting accepted by others. The esteem needs have also two distinct types which are competence (being able to fulfill tasks or being skillful) and admiration (being in authority, social recognition of others). The self-actualization needs are based on maximizing the individual’s capacity. This can be fulfilled in various ways such as seeking more knowledge or improve in beauty (Griffin, Ledbetter, & Sparks, 2014).
According to Maslow the needs should be met in sequence. For instance, before satisfying the safety needs someone is expected to want his physiological needs be met. However, there has been a significant amount of studies done which did not found the hierarchical order or the theory. For instance, a narcissistic individual could have different needs than others (e.g. higher preference for admiration) regardless of the hierarchy (Latham & Ernst, 2006; Griffin, Ledbetter, & Sparks, 2014). Nevertheless, the theory does give valuable insight of how someone's preference derived from a particular need (Griffin, Ledbetter, & Sparks, 2014). The Maslow's theory is important since an employer should first take the employee's needs into consideration. Secondly, the employer has to create an environment which will motivate the employee based on these needs. Last but not least, the employer should control the employee’s actions by support or punishment (Latham & Ernst, 2006).
The expectancy theory of Vroom explains the relationship between the effort someone put into work, the performance which derives from the efforts and eventually the reward which they receive from the effort and performance (Lunenburg, 2011). The theory consists of three essentials, which are expectancy, instrumentality and valence (Lunenburg, 2011). The expectancy is based on that the effort will eventually result in a high, agreeable performance. This is based on the probability that the individual will achieve the
performance (by education or attracting the right individual). Instrumentality is that the performance will be rewarded which is also based on the perceived probability that the individual expect the performance will lead to a reward. Valence is that the reward is valued positive by the individual. Hereby the preference of the reward plays an important role. A
12 promotion for example is valued differently by individuals. Therefore employers should make use of a total reward strategy (Lunenburg, 2011).
There are different ways to categorize rewards. The organization can choose between financial as well as non-financial rewards to motivate employees. The type of rewards should be linked to the intrinsic and/or extrinsic motivation of the individual (Armstrong & Murlis, 2007). Armstrong & Murlis (2007) have shown another way to categorize rewards, which is the categorization of individual and group rewards. Others have conceptualized rewards as either intrinsic or extrinsic rewards (Guzzo, 1979). The point is that there are different ways to categorize and conceptualize to research rewards and neither one is good or bad.
One of the most used management control tools to align the interest of the employee and the employer are financial rewards. Chiang and Birtch (2007) make use of two categories for financial rewards, which each have two subcategories. The first category is whether it is a direct cash based reward or an indirect reward. An example of a direct cash-based reward is the basic salary. The most important elements are that it is monetary and you receive it directly. An example of an indirect financial reward is a benefit such as a health insurance. The second category is whether it is a goal or a non-goal based financial reward. The goal-oriented financial reward is characterized by a variable pay. You have to meet the pre-specified targets to be awarded (based on individual performance for example). The non-goal oriented financial reward is characterized by a fixed pay. A simple example of this is the basic salary (Chiang, F.T., & Birtch, 2007).
Chiang and Birtch (2007) make use of three categories for non-financial rewards. The first category is support rewards. Support rewards consist of six rewards which are work-life balance, working environment, job location, organizational support, flexible benefits and alternative work arrangements. These rewards help the employee to support their work. The second category is extrinsic rewards and consists of seven rewards: fair treatment, team spirit, relationships, management style, workload, authority/power given and job
title/status. The extrinsic rewards have a big variety and therefore it is better to divide it in subcategories. Wingen (2015) has divided this category into subcategories including power and relationship rewards. I have chosen to divide it in these categories as well since this give more detailed information about the preferences. The power rewards include rewards which give power/control to the employee like authority, the amount of work or job
13 activities and job title/status. Relationship rewards, also as the categorization of Wingen (2015), are based on the internal and external relationships of the employee (Chiang, F.T., & Birtch, 2007). Other relationship rewards are social recognition, praise, loyalty and job security (Chiang, F.T., & Birtch, 2007). The third category is intrinsic rewards which consist of nine rewards: job satisfaction, nature of work, challenge, use skills opportunities, job
responsibility, job variety, autonomy, learning opportunities and chance to accomplish. These non-financial rewards are based on the intrinsic motivation of the individual and are based on job satisfaction and challenge (Chiang, F.T., & Birtch, 2007).
Chiang and Birtch’s (2007) model/theory do not categorize rewards based on individual and team based rewards. Team based rewards are often introduced because the organization want that the employees corporate in an effective way instead of focusing on the individual. Individual rewards harm the team in two ways. Firstly, the individual will only focus on his own actions and results. Secondly, higher level employees will focus on their employees as individuals rather than what the purpose of the team is and what the
individual can do for the team (Armstrong & Murlis, 2007). Increasing complexity in today’s world may lead to an increase in more cooperation between the divisions of the company and therefore this category is also added to the thesis.
2.3 The Independent Variable: Narcissism
The American Psychiatric Association (2013) describes narcissism as a pattern of an
unrealistic sense of superiority, where the individual is seeking admiration and has a lack of empathy for others. The Association consider someone to have a narcissistic personality disorder (NPD) if the individual has five or more of the following characteristics: (1) has a grand amount of self-importance, (2) seeks high amount of success, power, brilliance, beauty or love, (3) thinks himself/herself as a special person who can only be understood by other people with a above average status, (4) has a high amount of admiration, (5) the fact of having the right of something, (6) exploitative: achieve his/her own goals by taking advantage of others, (7) a low level of empathy, (8) envious or felt envious by others and (9) arrogance (American Psychiatric Association, 2013). However, NPD is the extreme form of narcissism. It is possible that people have certain narcissistic characteristics without having NPD (Young, Du, Dworkis, & Olsen, 2016).
14 effects at work. Narcissistic people have a high level of authority and superiority which can be very useful for functions that emphasize these competences. Other positive
characteristics are self-confidence and elegance (Young, Du, Dworkis, & Olsen, 2016). A recent study has found for example that narcissistic CEOs have higher earnings-per-share and share price than non-narcissistic CEOs (Olsen, Dworkis, & Young, 2014). Furthermore, narcissistic people seek status. Since the will to succeed is high, these characteristics can be useful for climbing the hierarchical ladder to get to the admired functions (Young, Du, Dworkis, & Olsen, 2016). However, narcissistic people also have characteristics which can be harmful for the organization or colleagues. These characteristics can be described as
egocentric, easily affected, and continually seeking for recognition and superiority even if this harms others. As described before they have a lack of empathy and also like
competition. These negative characteristics can lead to take undesired actions such as fraud or lying. Additionally, it has also implications for their motivation. Extrinsic motivation like money rewards could not motivate these people in the right way since these rewards lack recognition and praise. Finally, the risk approaches of these people are higher than non-narcissistic people since these people focus on their self. Takings risks in expense of others (agency theory) is considered less important (Young, Du, Dworkis, & Olsen, 2016).
Recent studies have examined the effects of narcissism on businesses. A recent study found a relationship between narcissism and counterproductive work behavior (Grijalva & Newman, 2015). Another study found a positive relationship between CEO narcissism and fraud (Rijsenbilt & Commandeur, 2013). These characteristics and attitudes of narcissistic people have also a serious impact on the management control tools. MCS were developed in the 1950s, while the workforce between then and now are different. The increase in narcissistic employees in the workforce has also led to different work attitudes and personalities (Young, Du, Dworkis, & Olsen, 2016). Twenge J. M., Konrath, Foster, Campbell, & Bushman (2008) found a percentage of 17 or higher for narcissistic college students between the period 1995 and 2006. This confirms that the new generations tend to be more narcissistic than the older generations (Twenge J. M., Konrath, Foster, Campbell, & Bushman, 2008). The focus in this thesis is for this reason on students (see more in the chapter methodology). The current management control system does not take these changes into consideration and could be ineffective in guiding these people (Young, Du, Dworkis, & Olsen, 2016).
15
2.4 Hypothesis Development
The research question, as stated in the introduction, is as follow: “does narcissism influences the reward preferences of employees?”. The purpose of this research question is to find an alignment of interest between the principal and agent through the management control tool rewards. In this paragraph hypotheses will be developed to examine whether the contingency variable narcissism influences the reward preferences of employees. It is
expected that people with high level of narcissism prefer different rewards than people with low level of narcissism.
In the reward section we have seen Maslow’s theory of needs. It is expected that people with high level of narcissism differ in safety needs since they are comfortable with risk (Young, Du, Dworkis, & Olsen, 2016). Also the esteem needs is expected to differ because narcissistic people seek more admiration than others (American Psychiatric Association, 2013). The hypotheses in this thesis are therefore developed with the opinion of Griffin, Ledbetter, & Sparks (2014) that Maslow’s hierarchy does not have to be met in sequence. Narcissism also seems to have implications for the theory of Vroom. The
expectancy factor is affected by the fact that narcissistic people have an overestimated view of themselves (Dworkis, Nair, & Soderstrom, 2015). The hypotheses in this section are in order to maximize the valence for high narcissistic people.
Nowadays financial rewards are often used to motivate people. It is questionable whether money rewards will motivate narcissistic people since it lack recognition and praise (Young, Du, Dworkis, & Olsen, 2016). However, money rewards are important for every individual because it covers the baseline needs1. Previous literature has analyzed that these rewards could be enhanced by combining it with recognition rewards such as public
communication (Dworkis, Nair, & Soderstrom, 2015). These kinds of rewards are especially used in big companies with important functions (CEO rewards). Also question 20 of the NPI-40 model (Appendix A) or the NPI-16 model (Appendix B) implicate that narcissistic
individuals like to show-off their accomplishments. The following hypothesis is therefore expected:
Hypothesis 1: People with high narcissism prefer recognition rewards (such as public communication) more than people with low level of narcissism.
16
According to Young, Du, Dworkis, & Olsen (2016) narcissistic people are risk takers. Non-goal rewards are used for stability because these rewards are fixed. However, it is expected that narcissistic people prefer more goal rewards because of their higher risk-attitude and secondly this is a way to differentiate themselves of others. Furthermore, according to Dworkis, Nair, & Soderstrom (2015) narcissistic people have an overestimated view over them self. This overestimated view will lead to a higher preference of goal rewards since these people think that they are capable to get these rewards.
Hypothesis 2: People with high narcissism prefer goal rewards more than people with low narcissism.
Young, Du, Dworkis & Olsen (2016) stated that narcissistic people love to operate on their own instead of a team. It is also known that narcissistic people are egocentric.
Therefore it is expected that narcissistic people want to focus on their self. Group rewards will focus also on the interests of others. Next to this it is also expected that narcissistic people prefer individual rewards more than group rewards from an admiration perspective. Question 7 of the Narcissistic Personality Inventory-40 in Appendix A shows that narcissistic people want to be the centre of attention. Group rewards will put others also in attention. Therefore the following hypothesis is expected:
Hypothesis 3: People with high narcissism prefer individual rewards more than people with low narcissism.
The American Psychiatric Association (2013) described narcissistic people as people who seek high amount of success, power and brilliance. Next to this a high amount of admiration is preferred (Young, Du, Dworkis, & Olsen, 2016). Power rewards such as authority (e.g. decision rights), important and broader job activities (e.g. managing instead of simple administration tasks) and job title (e.g. promotion) rewards will fit to narcissistic people since it provide a certain degree of power and an admired job position. Since the will to succeed is high, power rewards can be useful for climbing the hierarchical ladder to get to the admired functions (Young, Du, Dworkis, & Olsen, 2016). Authority (Question 12 of NPI-40) is also valued by these people and therefore power rewards can be very useful (for example increasing the amount of work or job activities).
17
Hypothesis 4: People with high narcissism prefer power rewards more than people with low narcissism.
Intrinsic rewards consist of manifold rewards. Rewards like job satisfaction and learning opportunities seem to be not different for narcissistic people than non-narcissistic people. These rewards are not triggered or impacted by specific characteristics of narcissistic people2.
In the section of power rewards it is expected that narcissistic people tend to love power rewards and therefore autonomy rewards seems also to be important for narcissistic people.
Nature of work, use specific skills and job responsibility rewards seems to be important for narcissistic people. In the theory section of narcissism it is explained that narcissistic people have some skills and competences for specific functions which emphasize these. Narcissistic people prefer hierarchical organizations because they believe they will outperform their colleagues and get to rise in the company (Zitek & Jordan, 2016).
Therefore it is expected that rewards to use their specific skills to climb in the hierarchy are important. The nature of work is also important since narcissistic people would prefer a job which matches their image (as praiseworthy people e.g.). Narcissistic people have a high level of authority (Young, Du, Dworkis, & Olsen, 2016). Therefore it is expected that they prefer responsibility since responsibility goes hand in hand with authority.
Narcissistic people are comfortable with risk. Next to this, their standards are higher which gives them the opportunity to differentiate themselves from others and satisfy their own needs (Young, Du, Dworkis, & Olsen, 2016). Therefore it is expected that people with high level of narcissism prefer rewards with challenges. This could also be explained by the fact that narcissistic people have an overestimated view over themselves (Dworkis, Nair, & Soderstrom, 2015).
The implications above show that most intrinsic rewards could be useful to motivate narcissistic employees. Therefore the following hypothesis is expected:
Hypothesis 5: People with high narcissism prefer intrinsic rewards more than people with low narcissism.
18 The following table summarized the hypotheses of this thesis:
Independent Variable Dependent Variable Relationship
H1 Narcissism Recognition rewards Positive
H2 Narcissism Incentive rewards Positive
H3 Narcissism Individual rewards Positive
H4 Narcissism Power rewards Positive
H5 Narcissism Intrinsic rewards Positive
Table 1: The hypotheses of the thesis3
3 There are no hypothesis drawed for the categories relationship and support rewards because of no or mixed
19
3 Research Methodology
In this section I will explain the selection of the sample. Next to this I will explain the measurement of the dependent variable which is rewards and also a principal component analysis will be conducted. Furthermore, the measurement of the independent variable narcissism will be explained. Finally, the control variables of this research will be explained.
3.1 Method and Sample Selection
The data for the thesis will be generated with help of cross-sectional surveys. There are a couple of reasons why the survey approach has been preferred. First of all, the focus of this thesis is the people’s opinion or preferences. With help of surveys narcissism and rewards are measurable. Second of all, the surveys are anonymous which gives the respondents the opportunity to answer the questions honestly. Last but not least, surveys are efficient in terms of time and money and it is possible to reach to the participants online as well as personal (Debois, 2016).
The University of Amsterdam’s policy regarding surveys has certain limits. It is forbidden to send out surveys to random people. Therefore an alternative way of collecting respondents has been employed which is a partly random and a partly non-random
sampling method. The surveys are collected by sending the surveys to friends, surveying fellow students of the master Accountancy and Control of the University of Amsterdam and surveying students on my previous university: The Hague University of Applied Science. The respondent are in some way affiliated with the researcher. Random and non-random sampling has advantages and disadvantages. For example, a disadvantage of non-random sampling is that it could lead to difficulties in generalizing the results (Bryman, 2012). The data has been collected by a self-completion survey online (Google Forms) and in some cases a print-out version of the survey. A self-completion survey approach has been chosen since the questions to measure narcissism are very personal. For the same reason all of the surveys were handed-in anonymously. Although it was a self-completion survey, participants were able to contact the researcher for further explanation.
According to Bryman (2012, p. 714) population could be defined as: “The universe of units from which a sample is to be selected”. The population where the sample is picked from is students and/or recent graduates (graduated in the last three years). The reason for choosing students is because students are from the generation where narcissism has
20 increased. Therefore it is expected that the range between low and high narcissistic traits will be high. Also economical or business back grounded students were preferred (but not necessary) because of their knowledge regarding management control tools. Also
interesting is that studies show that business students score higher on narcissism than other disciplines (Westerman, Bergman, Bergman, & Daly, 2010).
The sample size is the amount of participants in a sample. The minimum sample size for a research depends on situational factors (Bryman, 2012). The desired amount of participants (in agreement with the supervisor) for this research was a minimum of a three digit number of participants. In total, 102 students and recent graduates have filled in the survey. However, one participant did not completely fill in the first section which measures narcissism. Therefore one participant is excluded from the research which leaves a total of 101 participants.
3.2 Rewards
The dependent variables in this thesis are the various types of rewards. The financial and non-financial rewards will be measured based on the theory/paper of Chiang & Birtch (2007). This model consists of various reward categories and even more detailed
subcategories. However, it does not contain all kind of categories such as individual and group rewards. Therefore an additional theory has also been used which is the theory of Armstrong & Murlis (2007).
The questions to measure these dependent variables will be based on the theories used in the theory section. For instance, In order to measure the authority rewards the following statement will be put forward to the participant: “I prefer rewards which gives me more authority (such as decision rights for the policy of the company)”. The participants have to give their preferences questions on a five-point likert scale. This scale gives the participants the opportunity to agree or disagree with statements. Furthermore, the five-point likert scale does not force the participants to choose for one option like the four-five-point likert scale (Bertram, 2009). Although, the likert scale have also some disadvantages such as participants answering in a socially desirable way (Bertram, 2009). In order to minimize these disadvantages the anonymity of the participants will be kept. In Appendix C the questionnaire of the rewards can be found.
21 In Appendix D the rotated component matrix can be found with all items of the survey loading on a specific component (eigenvalue > 1) with an absolute value higher than 0.5. The measurement of the five dependent variables (of the hypotheses) will be explained with help of the matrix. Also a reliability analysis has been done with the Cronbach’s Alpha with a minimum limit of 0.7 (Field, 2009).
The component recognition rewards has in the matrix two sub-variables which are recognition and public communicated rewards. However, the Cronbach’s Alpha for these rewards is 0.661 which is beneath the acceptable limit. These rewards will, to be sure, analyzed individually in further analysis.
Goal rewards will be measured by the single variable goal rewards. Goal and non-goal rewards load on the same component since these rewards are the opposite of each other. However, non-goal rewards were only included in the survey to clarify the opposite of goal rewards for the participants and will be left out in answering the relevant hypothesis. Individual rewards will also be measured by the single variable individual rewards. Individual and group rewards load on the same component since these rewards are also the opposite of each other. However, group rewards were only included in the survey to clarify the opposite of individual rewards for the participants and will be left out in answering the relevant hypothesis.
Power rewards, in line with the theory and survey, include the following
sub-variables: job title / status rewards, authority rewards and more important and broader job activities as reward. Also job satisfaction loads mainly on power rewards. However, job satisfaction is not included in the hypothesis and will be left out by further analyses. The Cronbach’s Alpha for power rewards is 0.732.
Intrinsic rewards, in line with the theory, include the following sub-variables: nature of work, skills, learning opportunities, responsibility, challenge and autonomy rewards. The sub-variable learning opportunities is not included in the hypothesis and will be left out for further analyses. The Cronbach’s Alpha for intrinsic rewards is 0.718.
3.3 Narcissism
The Narcissistic Personality Inventory (NPI) model of Raskin and Terry (1988) is one the most used models to measure narcissism (Ackerman, Witt, Donnellan, Trzesniewsk, Robins, & Kashy, 2011). The model consists of 40 questions which measure various components of the
22 participants such as authority, self-sufficiency, superiority, exhibitionism, exploitativeness, vanity, and entitlement (Raskin & Terry, 1988; Ackerman, Witt, Donnellan, Trzesniewsk, Robins, & Kashy, 2011). Each part consists of two options. For instance, the participant have to choose in part one between “I have a natural talent for influencing people” or “I am not good at influencing people”. The participant score points based on these questions (on a scale of 40 points). The full questionnaire can be found in Appendix A (Raskin & Terry, 1988), which is retrieved from www.personality-testing.info. Although, the length of the model can negatively influence the use (Ames, Rose, & Anderson, 2006). The participants could for example fill in the questionnaire fast since it takes too much time. Therefore an alternative model of the NPI-40 will be used, which consists of 16 questions. These questions have been taken out of the original model and has been tested valid for measuring narcissism (Ames, Rose, & Anderson, 2006). The NPI-16 can be found in Appendix B.
In the preliminary analysis the results of the NPI-16 will be grouped into two groups, low and high level of narcissistic people. This will be done like the analysis of Dworkis, Nair, & Soderstrom (2015) by taking everyone who scores under or the median as people with low level of narcissistic people and everyone above the median as people with high level of narcissistic people. Tests will be done to look if the means of the two groups differ
significantly from each other.
A Factor Analysis or Principal Component Analysis has not been conducted since the NPI test is a widely used model to measure narcissism. An analysis is superfluous since these analyses have been conducted already in the past. The same applies for the Cronbach’s Alpha of the NPI-16. Ames, Rose and Anderson (2005) have conducted several tests which showed that the Cronbach’s Alpha was around the acceptable cut-off point of 0.700 (Field, 2009). Interesting to note is that Ames, Rose and Anderson (2005) found an average Cronbach’s Alpha around 0.800 for the NPI-40 which has 24 additional items.
3.4 Control Variables
Control variables are in place to exclude other variables impact the relationship between narcissism and rewards (Spector & Brannick, 2011). In this research the following variables are controlled: gender, nationality, education level, kind of education and grade point average (GPA).
23 difference between the preferences of male and female. Also males score higher on
narcissism than females (American Psychiatric Association, 2013). Therefore gender (male and female) will be controlled in this research.
Gunkel and Marjaana (2006) have researched the reward preferences in two
different countries, which are the United States of America and Germany. According to their research the preferences in these two countries differ (Gunkel, Lusk, & Wolff, 2006). Briefly, nationality does matter in this research and should be controlled for since the students could come from various countries with different preferences.
Another control variable is the type of education (MBO / HBO / University) and education level (Bachelor / Master / PhD / Other) of the individual. Employees who are higher educated earn more money than less or non-educated employees. Next to this higher educated employees also profit from other benefits such as a health insurance and a pension benefit (Baum, Ma, & Payea, 2013). It is important to control on this variable since higher educated employees expect to receive these benefits more than non-educated employees. They are more entitled for these rewards.
The faculty should be controlled for since jobs vary in their tasks. An individual who is educated to become a general officer is expected to have a certain power over the people beneath him. The faculties of The University of Amsterdam and The Hague University of Applied Science (Economics, Business, Finance & Marketing / Management & Organization / Technology, Innovation & Society / Public Management, Law & Safety / Social Work & Education / Medicine / Science / Other) has been chosen to categorize this variable. Lastly, the GPA of students is another control variable since a student with high grades could expect to receive more rewards than average or low scoring students. The categorization of GPA (on a scale of 10 with a minimum of 5.5) is as follow: 5.5 – 7; 7 – 8.5; 8.5 – 104.
24
4 Results
In this section the data of the research will be analyzed and explained. The descriptive statistics will be presented to provide further information about the background of the participants. A preliminary analysis without the control variables will be done to see if there is a difference between high narcissistic and low narcissistic participants. Next to this, the correlations between the variables will be presented. Lastly, the data will be analyzed with help of a regression analysis in order to answer the hypotheses of the theory section.
4.1 Descriptive Statistics
As stated in table 2 participants scored an average of 6.2 (with a standard deviation of 3.02) at a scale of 16 on the Narcissistic Personality Inventory. The individual scores vary between a minimum of 0 to a maximum of 13. The skewness of the distribution is 0.098 which
indicates an almost symmetrical distribution. Normally, the distribution is skewed to the right. This could be explained by the majority of men involved in this sample since males score higher on narcissism than females (American Psychiatric Association, 2013). Europeans score on average 15.0 (with a standard deviation of 6.3) on the Narcissistic Personality Inventory on a scale of 40 (Foster, Campbell, & Twenge, 2003). This is approximately in line with the average of this research (6.2 / 16 * 40 = 15.5).
Participants
Scale (1 -
…) Mean Stdev Q1 Q2 Q3 Min Max NPI 101 16 6,20 3,02 4,00 6,00 8,00 0 13 H1 Recognition 101 5 3,24 1,11 2,00 3,00 4,00 1 5 H1 Publicly Communicated 101 5 3,10 1,03 2,00 3,00 4,00 1 5 H2 Goal 101 5 3,82 0,82 3,00 4,00 4,00 2 5 H3 Individual 101 5 3,71 0,95 3,00 4,00 4,00 1 5 H4 Authority 101 5 3,58 0,85 3,00 4,00 4,00 1 5 H4 Activities 101 5 3,78 0,80 3,00 4,00 4,00 2 5 H4 Job title / status 101 5 4,16 0,80 4,00 4,00 5,00 2 5 H5 Autonomy 101 5 3,67 0,74 3,00 4,00 4,00 2 5 H5 Challenge 100 5 3,95 0,81 4,00 4,00 4,00 1 5 H5 Nature of work (Personality fit) 101 5 4,25 0,84 4,00 4,00 5,00 1 5 H5 Skills 101 5 4,21 0,80 4,00 4,00 5,00 1 5 H5 Responsibility 101 5 3,80 0,88 3,00 4,00 4,00 1 5
Table 2: Descriptive statistics of the dependent and independent variables
Table 2 contains also descriptive statistics of the dependent variables rewards. Interestingly, rewards for improving the nature of the work to the personality of the
individual and rewards for using employee’s high improved skills are scoring on average the highest. The goal rewards have been preferred in general to a high extent. This finding is in
25 line with the increasing narcissism under students which results in an attitude of risk taking (Dworkis, Nair, & Soderstrom, 2015; Campbell & Foster, 2007).
There is a slight majority of male participants in the sample. In table 3 the scores of gender on narcissism has been computed. In line with the theory of The American
Psychiatric Association males scores higher (based on the mean and quartiles) on narcissism than females. Also the preferences of rewards have been computed for males and females (Appendix F). However, there were no major differences in the preferences. The difference between male and female on narcissism is the reason to control for this variable.
NPI and Gender Participants
Scale (1 -
…) Mean Stdev Q1 Q2 Q3 Min Max Females 42 16 5,36 2,86 3 5,5 7 0 11 Males 59 16 6,80 3,01 4,5 7 9 1 13
Table 3: Gender and narcissism
Many business participants were involved (42%) because of the researcher’s background. The disciplines of the other participants vary (management, technology, medicine, science et cetera). There are no faculties included which have to be analyzed differently such as a military education. In table 4 narcissism based on business and non-business students has been computed. In line with Westerman, Bergman, Bergman & Daly (2010) business students score higher on narcissism (based on the mean and quartiles) than students of other faculties.
Faculty: Participants
Scale (1 -
…) Mean Stdev Q1 Q2 Q3 Min Max Business Students 42 16 6,86 3,12 5 7 9 0 13 Non-Business Students 59 16 5,73 2,88 4 6 8 0 13
Table 4: Faculty and narcissism
The type of education divers from HBO to university and to a lesser extent students or recent graduates with a MBO background. In Appendix F the type of education has been computed with the different reward variables. Higher educated participants (HBO,
University) seem to prefer more goal rewards than lower educated participants (MBO). This is logic since higher educated people have in general a higher probability to accomplish their goals. This also applies for intrinsic rewards. This could also be explained by the fact that higher educated people have more freedom to choose an education which matches their intrinsic motivation (a lower educated person can for example not choose to become a
26 doctor). There are no major differences for the other variables.
The participants have in total eleven different nationalities. However, the greater part (94.1%) is Dutch. Therefore nationality will not be controlled in this research.
The participants have also provided their GPA. Participants with a GPA of 8.5 – 10 seem to prefer individual rewards more than people with a GPA of 5.5 – 7 or 7 – 8.5. This is logically since for example higher performing participants would rather work individually than working with others who would decrease his performance. Also recognition, power and intrinsic rewards are more preferred than people with a GPA of 5.5 – 7 or 7 – 8.5. Lastly, participants have put their detailed education level which is bachelor, master and other. The mixed results in preferences of this control variable have led to the decision to leave it out.
4.2 Preliminary Analyses and Correlation Matrix
The relationship between narcissism and rewards will be preliminary analyzed with help of an analysis of variance’s (ANOVA). This will be done by comparing the reward preferences of people with high narcissism against the reward preferences of people with low narcissism. Every participant who has scored beneath or equal to the median (6) is categorized as an individual with low narcissism. Logically, scores above the median has been categorized as high narcissism. An ANOVA analysis has been chosen because it provides statistics in detail (such as the means in table 5) in comparison to a regression analysis. In table 5 the results of the ANOVA are presented. All hypotheses, in line with the expectations, have a higher mean for participants with high narcissism in comparison with participants with low narcissism. However, only the hypotheses for recognition and goal seem to be significant to a certain level.
Means
Reward Low Narcissism High Narcissism F p-value
H1 Recognition 3,04 3,45 3,566 0,062
H2 Goal 3,67 3,98 3,642 0,059
H3 Individual 3,63 3,80 0,722 0,398
H4 Power 3,81 3,88 0,283 0,596
H5 Intrinsic 3,92 4,03 0,86 0,356
27 In table 6 the correlations are presented of all variables which are going to be used in the analyses. In line with the previous ANOVA analyses narcissism correlates with recognition (r = 0.216, p < 0.05) and goal rewards (r = 0.203, p < 0.05), but not with the other rewards (except for the sub-variable responsibility rewards). Next to this, gender does have an impact on narcissism (r = 0.237, p < 0.05). As expected non-business students negatively correlate with narcissism, however this is not significant.
Logically, sub-variables of a specific reward correlate with the other sub-variables of that specific reward. Power rewards, which consist of (9) authority, (10) more important and broader job activities and (11) job title/status do correlate with each other (9 & 10: r = 0.473, p < 0.01; 9 & 11: r = 0.389, p < 0.01; 10 & 11: r = 0.518, p < 0.01). There are also correlations between variables of different categories because they are in some way related or measure the same underlying factor. For example, goal rewards are correlated with challenge rewards (r = 0.286, p < 0.01).
Some of the control variables do correlate with rewards. The type of education has a significant positive correlation (r = 0.214, p < 0.05) with responsibility rewards. This indicates that a higher education, for example university in comparison with HBO, is associated with the preferences of more responsibility. Higher educated people are in general more capable and could expect therefore to do job activities with more responsibility. The same applies for GPA and responsibility rewards (r = 0.232, p < 0.05) since higher performing people could expect to be more in charge than lower performing people. There are also
relationships which are not straightforward such as the negative significant relationship between the type of education and publicly communicated rewards (r = -0.198, p < 0.05). This indicates that lower educated people prefer more publicly communicated rewards than higher educated people.
The significant relationship between the control variable gender and the independent variable narcissism could indicate in the analyses that the relationship between narcissism and rewards are impacted by characteristics of male and female. However, the relationship is weak (0.237).
28 Correlations 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 1. NPI 1,000 2. Recognition ,216** 1,000 3. Publ.Communicated 0,114 ,487*** 1,000 4. Goal ,203** 0,159 0,149 1,000 5. Individual 0,086 0,126 ,263*** 0,148 1,000 6. Authority 0,137 ,228** ,198** 0,000 ,241** 1,000 7. Activities 0,140 0,157 0,140 0,163 0,041 ,473*** 1,000
8. Job title / status 0,039 0,140 ,219** 0,191* 0,163 ,389*** ,518*** 1,000
9. Autonomy 0,078 0,074 -0,083 0,104 -0,041 ,348*** ,295*** 0,173* 1,000 10. Challenge 0,163 0,055 -0,023 ,286*** 0,131 ,326*** ,380*** ,322*** ,469*** 1,000 11. Nature of work -0,091 0,092 0,113 0,190* 0,081 0,159 0,172* ,289*** ,251** ,232** 1,000 12. Skills 0,146 0,024 0,090 ,309*** 0,067 0,117 0,113 ,320*** 0,153 ,330*** ,567*** 1,000 13. Responsibility ,205* 0,067 0,160 0,122 0,154 ,380** ,268** ,255* ,323** ,491** 0,179 ,310** 1,000 14. Gender ,237** -0,060 0,005 0,022 0,019 0,021 -0,063 -0,053 0,102 0,118 -0,314*** 0,062 0,073 1,000 15. Type of education 0,118 0,034 -0,198** 0,033 -0,035 0,028 0,010 -0,127 0,180* 0,160 0,011 0,065 ,214** ,246** 1,000 16. GPA 0,148 0,065 0,065 0,067 0,028 -0,079 0,090 -0,052 -0,098 -0,052 0,081 0,188* ,232** -0,017 0,169* 1,000 17. Faculty -0,195* -0,077 -0,063 -0,102 -0,178* 0,005 0,077 0,117 -0,040 -0,062 ,234** 0,105 -0,174* -,386*** -,296*** 0,085 1,000
*. Correlation is significant at the 0.1 level (2-tailed). **. Correlation is significant at the 0.05 level (2-tailed). ***. Correlation is significant at the 0.01 level (2-tailed).
29
4.3 Regression Analyses
The hypotheses of this research will be answered with a regression analysis. The
independent variable in the regression analysis is narcissism and the dependent variables are the rewards. The relationship has been controlled on the variables gender, type of education, GPA and faculty. In the table beneath the standardized betas are presented and if it is significant on a two-tailed level.
H1 H2 H3 H4 H5
Recognition Goal Individual Power Intrinsic
NPI 0,25** 0,194* -0,02 0,178 0,136 (0.022) (0.078) (0.856) (0.109) (0.208) Gender -0,157 -0,078 0,02 -0,053 -0,074 (0.161) (0.493) (0.858) (0.643) (0.512) Type of education -0,001 0 -0,103 0,023 0,221** (0.990) - (0.348) (0.853) (0.044) GPA 0,036 0,027 0,051 -0,031 0,04 (0.731) (0.801) (0.633) (0.775) (0.703) Faculty -0,104 -0,103 -0,225* 0,08 0,044 (-0,358) (0.368) (0.051) (0.487) (0.699) R square 0,08 0,051 0,049 0,032 0,069 Adjusted R Square 0,03 0 0 0 0,018
* Significant at 0.1 level (2-tailed). ** Significant at 0.05 level (2-tailed). *** Significant at 0.01 level (2-tailed)
Table 7: The standardized betas and significance
In the hypothesis section the preferences of narcissistic people and rewards have been discussed. It is important to note that all of the hypotheses, the expected relationship between narcissism and a specific reward, are directional and therefore could actually answered by one-tailed tests (in contrast to the table above). To give a complete overview one-tailed as well as two-tailed significance will be mentioned.
In hypothesis 1 it was expected that narcissistic people prefer recognition rewards since these people value recognition and praise of others (Young, Du, Dworkis, & Olsen, 2016). In the preliminary analysis high narcissistic people, in line with the expectations, preferred recognition rewards significantly more than low narcissistic people. After
conducting a regression analysis the relationship reveals a standardized beta of 0.250. The significance on a two-tailed level is 0.022 and 0.011 on a one-tailed level. The control
30 variables gender, type of education, GPA and faculty have a standardized beta coefficient of -0.157; -0.001; 0.036 and -0.104. However, none of these control variables are significant. The results above would indicate that recognition rewards will be effective to motivate narcissistic people.
An extra analysis has been conducted on publicly communicated rewards which are a form of recognition. The regression analysis on publicly communicated rewards reveals a standardized beta coefficient of 0.107 which is not significant (p-value = 0.314 on a two-tailed level and 0.157 on a one-two-tailed level). This could also be explained by the Dutch culture of the sample. The Dutch culture is characterized by not naming and shaming of people (Speklé & Verbeeten, 2014). Interestingly, the type of education has a beta
coefficient of -0.292 (p-value = 0.007), which indicates that a higher education type (MBO / HBO / University) results in a lower preferences of publicly communicated rewards.
In hypothesis 2 it was expected that narcissistic people prefer goal rewards since these people are higher risk takers, have an overestimated view over them self and are able to differentiate themselves from others with these rewards (Dworkis, Nair, & Soderstrom, 2015; Young, Du, Dworkis, & Olsen, 2016). The regression analysis reveals a beta of 0.194 with a p-value 0.078 on a two-tailed level after controlling for the variables. The one-tailed test is significant on a 0.05 level (p-value of 0.039). The analysis reveals that, in line with the expectations, high narcissistic people prefer goal rewards more than low narcissistic people. Hypothesis 3 regarding individual rewards it was expected that narcissistic
individuals love to operate on their own (Young, Du, Dworkis, & Olsen, 2016). The regression analysis reveals a negative standardized beta of 0.020 which would mean that higher
narcissistic individuals will prefer to work and get rewarded less individually. However, the relationship is not significant (p-value of 0.856 on a two-tailed level and 0.428 on a one-tailed level). This could be explained by the fact that narcissistic individuals seek recognition or approval from others (Young, Du, Dworkis, & Olsen, 2016). If an individual is operating alone, the only validation the person could get is from his supervisor/boss. If the person is working in a group the individual could get validation from his supervisor/boss and the other group members. The control variable faculty is significant with a p-value of 0.051 with a beta of -0.225 which means that non-business students prefer less to work individually than business students.
31 prefer these rewards since it gives them success, power, brilliance and admiration
(American Psychiatric Association, 2013; Young, Du, Dworkis, & Olsen, 2016). The regression analysis of power rewards, which consist of authority, activities and job title rewards,
reveals a positive standardized coefficient of 0.178. The significance of the relationship is 0.109 on a two-tailed level and 0.055 on a one-tailed level. None of the control variables are significant with the power rewards. The analysis reveals that, based on a one-level test, high narcissistic people prefer power rewards more than low narcissistic people. A more detailed analysis of the individual power rewards can be found in the next paragraph (4.4 additional analyses).
In hypothesis 5 it was expected that narcissistic people prefer intrinsic rewards for various reasons (such as a high level of authority (Young, Du, Dworkis, & Olsen, 2016)) The regression analysis for intrinsic rewards reveals a beta of 0.136 which means that higher narcissistic individuals prefer more intrinsic rewards. However, the relationship is not significant (p-value of 0.208 on a two-tailed level and 0.104 on a one-tailed level). The type of education has a standardized beta of 0.221 which means that higher educated people prefer more intrinsic rewards than lower educated people (p-value 0.044 on a two-tailed level). This could be explained by the fact that higher educated people are more able to choose the education they want and therefore is more intrinsic motivated (e.g. a person studying on MBO level cannot choose to become a doctor).
The R square and the Adjusted R square are presented in figure 7. An additional table is presented below which consist of the R square and Adjusted R square without control variables (the direct relationship of NPI and rewards). As stated in the table NPI explains almost for every reward (except for individual rewards) a part of the variance. After adding the control variables the Adjusted R square decreased for recognition, goal,
individual and power rewards. For example, the adjusted R square decreased from 2.9% to 0% for goal rewards. This could be explained by two factors. It could be that the
combination of the control variables gender, type of education, GPA and faculty turned out that they do not explain more of the data. This is partly understandable. For example, the gender of a participant does not explain if they would prefer individual rewards or not. Another explanation is that the amount of participants in the research is limited. This results in a limited amount of degrees of freedoms.
32 Recognition Goal Individual Power Intrinsic
R square 0,056 0,04 0,001 0,021 0,02 Adjusted R Square 0,046 0,029 0 0,011 0,009 Table 8: The R square and Adjusted R square of NPI and rewards
4.4 Additional Analyses
In this section an additional analyses will be conducted for the individual rewards of power and intrinsic rewards.
Power rewards consist of three individual rewards which are authority, more
important and broader activities and job title rewards. The regression analysis for authority rewards reveals a beta of 0.163 which means that higher narcissistic individuals prefer more authority rewards than lower narcissistic individuals. The two-tailed level test of the
relationship is not significant (p-value 0.141), but the one-tailed test is on a 10% significance level (p-value 0.071).
More important and broader activities as reward has a standardized beta of 0.166 with a significance level of 0.133 on a two-tailed test and 0.067 on a one-tailed test. This would indicate, based on the one-tailed test, that high narcissistic people prefer these rewards more than low narcissistic people.
The job title rewards have a standardized beta of 0.097 which would indicate that high narcissistic people prefer these rewards more than low narcissistic people. However, the relationship is not significant (p-value 0.378 based on two-tailed test and 0.189 on a one-tailed test).
The analysis reveals that some individual power rewards, based on a one-tailed significance level, are preferred to some extent.
33 Authority Activities Job Title
NPI 0,163 0,166 0,097 (0.141) (0.133) (0.378) Gender -0,013 -0,064 -0,052 (0.908) (0.576) (0.648) Type of education 0,075 0,036 -0,06 (0.497) (0.742) (0.589) GPA -0,066 0,057 -0,062 (0.538) (0.593) (0.566) Faculty 0,024 0,081 0,09 (0.834) (0.479) (0.437) R square 0,031 0,039 0,026 Adjusted R Square 0 0 0
* Significant at 0.1 level (2-tailed). ** Significant at 0.05 level (2-tailed). *** Significant at 0.01 level (2-tailed)
Table 9: The standardized betas and significance of the individual power rewards
Intrinsic rewards consist of five rewards which are autonomy, challenge, nature of work, skills and responsibility rewards. Autonomy has in line with the expectation a positive beta of 0.099. However, the relationship is not significant (p-value of 0.363 on a two-tailed level and 0.182 on a one-tailed level). The type of education has for autonomy rewards a beta of 0.183 with a p-value of 0.095 which means that higher educated people prefer rewards which gives them more to say. This seems logically since educated / knowledgeable people in general have more sensible to say in comparison with lower educated people. Rewards with a high challenge have a beta of 0.146, which is in line with the
expectation that higher narcissistic people prefer higher challenges. The significance level on a two-tailed level is 0.181 and 0.091 on a one-tailed level. None of the control variables are significant.
The nature of work has a beta of -0.052. The beta is in the opposite direction since it was expected that narcissistic people prefer jobs which suit to their personality. The control variable gender has a beta coefficient of -0.309 with a significant p-value of 0.005. This means that males in comparison with females prefer fewer rewards that fit their personality.
The specific high skills have a beta of 0.118, which is in line with the expectation that high narcissistic people prefer jobs which emphasize specific skills of the individual. These
34 skills could be used to climb the hierarchical ladder. However, the relationship is not
significant (p-value of 0.274 on two-tailed level and 0.137 on a one-tailed level). None of the control variables are significant.
The responsibility rewards have a beta of 0.158, which is in line with the expectation that high narcissistic people prefer rewards which give them more responsibility. The significance of the relationship is 0.133 on a two-tailed level and 0.067 on a one-tailed level. The GPA of the students has a beta of 0.174 (p-value: 0.090) which reveals that a higher GPA results in a higher preference of responsibility rewards. Once again this seems logic since higher performing people expect to be more in charge.
The analysis reveals that some intrinsic rewards such as challenge and responsibility are significant to a certain level based on a one-tailed test.
Autonomy Challenge Nature of Work Skills Responsibility
NPI 0,099 0,146 -0,052 0,118 0,158 (0.363) (0.181) (0.616) (0.274) (0.133) Gender 0,08 0,057 -0,309*** 0,016 -0,067 (0.482) (0.610) (0.005) (0.883) (0.539) Type of education 0,183* 0,154 0,127 0,145 0,149 (0.095) (0.161) (0.229) (0.182) (0.158) GPA -0,09 -0,155 0,043 0,135 0,174* (0.398) (0.145) (0.672) (0.198) (0.090) Faculty 0,053 0,028 0,083 0,156 -0,156 (0.639) (0.803) (0.451) (0.169) (0.156) R square 0,057 0,065 0,128 0,075 0,13 Adjusted R Square 0,005 0,013 0,08 0,025 0,082
* Significant at 0.1 level (2-tailed). ** Significant at 0.05 level (2-tailed). *** Significant at 0.01 level (2-tailed)
