– a multifirm study The relationship between personality traits, management control systems, and organizational commitment

(1)

1

The relationship between personality traits, management control systems, and

organizational commitment – a multifirm study

Master’s Thesis 15 credits

Department of Business Studies Uppsala University

Spring Semester of 2020

Date of Submission: 2020-06-03

Emil Sirén Gualinga Dan Lennartsson

Supervisor: Shruti Kashyap

(2)

ACKNOWLEDGEMENT

The realization of this thesis would not have been possible without the help of several persons.

First, we would like to thank our supervisor Shruti Kashyap for her invaluable advice. We highly appreciate her engagement in guiding us through each stage of the process.

We would also like to thank James Sallis for invaluable input regarding the statistical methods for this thesis.

Additionally, we would like to thank all members of our seminar group, from whom we received important feedback and shared interesting discussions.

Finally, we would like to thank all the organizations and managers that participated in this study. This study would not have been possible without your participation.

Thank you!

Emil Sirén Gualinga

Dan Lennartsson

(3)

ABSTRACT

Management control research has often focused on finding the right fit between management control systems and the unique situational context of organizations, but few studies have examined the role of individual personality traits. Therefore, the purpose of this study was to provide a better understanding about the relationship between the Big Five personality traits, management control systems and organizational commitment. We addressed the research gap regarding how the interaction between the Big Five personality traits and management control systems may affect organizational commitment. To answer the research questions, data was collected using a web survey, which was analyzed in statistical software. In total, 103 responses from managers across 30 Swedish companies, were obtained. Our findings indicate that the interaction of the personality trait conscientiousness and results control, and the interaction of the personality trait agreeableness, and cultural control may affect organizational commitment.

Based on these findings, theoretical and practical implications are discussed.

Key words: management control systems, organizational commitment, Big Five personality traits, results control, cultural control

(4)

1. INTRODUCTION

1.1 Background and Purpose Statement

As the environments in which organizations operate become more uncertain, organizations are looking to find new ways of organizational structures, procedures, and management control systems in order to cope better. Management control systems are important tools that organizations use in order to implement their strategy and reach their objectives by mitigating agency conflicts and providing the guidance necessary for employees to perform in accordance with organizational interests (Merchant and Van der Stede, 2003). Literature indicates that organizational commitment among employees is important in order to ensure that organizational goals are met and has been found to be positively related to employee performance as well as financial performance of firms (Riketta, 2002; Steyrer, Shiffinger and Lang, 2008). Previous research indicates that organizational commitment of employees can be affected by the type of management control systems organizations use, although the current understanding is limited (Ouchi, 1979; Kleine and Weissenberger, 2014). Thus, understanding how management control systems affect organizational commitment could have important implications for the design of management control systems.

Literature also suggests that the personal characteristics (Big Five personality traits) of employees may influence organizational commitment, but the current understanding is limited (Erdheim, Wang and Zickar, 2006; Tziner et al., 2008; Farrukh, Ying and Mansori, 2017).

Research has found the relationship between personality traits and organizational commitment to vary across different types of organizations and different types of professions. This raises the question if there is a pre-dispositional effect of certain personality traits in developing organizational commitment at management level. Understanding the relationship between personality traits and organizational commitment among managers could have important implications for the hiring process of managers.

Another factor that could be important in determining organizational commitment among managers is the interaction of their personality traits and the types of management control system they are exposed to. Whereas discussion about the effectiveness of management control has often focused on environmental factors, it has also been suggested that individual personality may affect how individuals respond to different aspects of management control systems (Chenhall, 2003). Findings from previous research has shown some indications that the type of control exercised over individuals may affect individuals differently, based on their

(6)

2 individual personality traits (Meadows, 1980), but we are not aware of any study that has studied specifically the interaction effect of management control systems and personality traits on organizational commitment. Understanding how the interaction of different personality traits and different aspects of management control systems affect organizational commitment could have important implications for the design of management control systems and hiring processes.

In this paper we aim to provide a better understanding about the relationship between the Big Five personality traits, management control systems and organizational commitment.

This addresses the research gaps identified above, and leads us to the following research questions:

1. What is the relationship between the Big Five personality traits and organizational commitment among managers?

2. What is the relationship between management control systems and organizational commitment among managers?

3. How does the interaction of the Big Five personality traits and management control systems affect organizational commitment among managers?

This study focuses on both top and middle managers because of their important role in ensuring that the organization's objectives and strategies are met (Nutt, 1987; Simons, 1990;

1994; Herzig and Jimmieson, 2006). We contribute with new findings regarding how organizational commitment is affected by the interaction of personality traits and management control systems. We look specifically at how the relationship between the Big Five personality traits and organizational commitment is moderated by management control systems. To answer the research questions, we created hypotheses based on previous literature and deductive reasoning, which we tested statistically. In total, we obtained 103 questionnaire responses from managers across 30 Swedish companies. Some of the main findings of this study were that the personality traits agreeableness and conscientiousness were particularly important personality traits in regard to explaining the organizational commitment of managers under different types of management control systems.

(7)

3

1.2 Outline of Thesis

The remainder of this thesis is structured as follows:

Chapter 2: Literature review and hypotheses. In this chapter key concepts such as organizational commitment, management control systems, and personality traits are defined.

Additionally, previous research concerning relationships between organizational commitment and personality traits, as well as relationships between management control systems and organizational commitment are presented. The hypotheses are constructed throughout the chapter and summarized at the end of the chapter.

Chapter 3: Method. This chapter describes the research design of the thesis and explains the variables included in the study. Moreover, the questionnaire design is explained as well as the process of data collection and data analysis. The regression models are explained in order to facilitate understanding of the data analysis for the reader.

Chapter 4: Results. The results obtained from data analysis are presented. The results from the hypothesis tests are presented in tables, as well as in text. Additionally, other results related to the research questions are presented.

Chapter 5: Discussion. In this chapter we interpret the results and compare them to previous findings. The chapter includes discussion regarding theoretical and practical implications, and directions for future research.

Chapter 6: Conclusion. Concluding remarks about the findings from this study in relation to the purpose of this study.

2. LITERATURE REVIEW AND HYPOTHESES

In this chapter we define the theoretical concepts that will be used in this study, and we review previous literature related to our research questions. First, with reference to prior literature in the field, we define the management control systems, organizational commitment, and the Big Five personality traits. Based on prior literature, hypotheses are constructed throughout the chapter, and are then summarized at the end of the chapter.

2.1 Management Control Systems

Conceptualization of management control systems. To understand this study, it is important to understand what is meant by management control systems. Over the years, management control systems have been conceptualized in several different ways (Simons,

(8)

4 2000; Merchant and Van der Stede, 2003; Anthony and Govindarajan, 2007). Merchant and Van der Stede (2003) conceptualize management control within organizations to include informal and formal controls, which may be applied at multiple hierarchical levels within the organization. Because we look at both top and middle management, and both formal and informal control systems, the framework provided by Merchant and Van der Stede (2003) is used in this study as a basis for conceptualizing management control systems. This framework of management control system may be understood to build on four types of control mechanisms:

results control, action control, cultural control, and personnel control. These control systems are needed in order to guide the employees in the right direction and ensure that objectives are met. In this study, we focus on results control and cultural control. Under this categorization, cultural control is more of an informal control system, whereas results control is more of a formal control system (Chenhall, 2003). Thus, this study includes both a formal control system and an informal control system.

Results control and cultural control.Result control is considered a formal management control system (Chenhall, 2003) and focuses on evaluating the outcome of employees’ actions (Jaworski and MacInnis, 1989; Hansen, Otley and Van der Stede, 2003; Merchant and Van der Stede, 2003; Widener, 2004). Cultural control is considered an informal control system (Chenhall, 2003) and emphasizes creating a common purpose among employees in the organization (Rockness and Shields, 1984; Flamholtz, Das and Tsui, 1985; Abernethy and Brownell 1997; Sandelin, 2008).

Previous literature has suggested that management control systems may be most effective when formal and informal control is combined (Malmi and Brown, 2008). In this thesis we include results control and cultural control, to represent formal control and informal control, respectively. These management control systems, together with the Big Five personality traits will be hypothesized as having an interaction effect on organizational commitment. We look specifically at how the relationship between the Big Five personality traits and organizational commitment is moderated by cultural control and results control. Action control and personnel control are excluded in order to limit the number of questions in the questionnaire.

Results control. In this study, results control refers to controls that evaluate the outcome of employee’s actions. This definition is consistent with previous definitions (Jaworski and MacInnis, 1989; Hansen, Otley and Van der Stede, 2003; Merchant and Van der Stede, 2003 and Widener, 2004). Jaworski and MacInnis (1989) refer to results control as output control which is exercised by evaluating individuals by the result obtained relative to the set target.

Results control is the cornerstone of budgetary controls (Hansen, Otley and Van der Stede,

(9)

5 2003). It provides financial data that is reported for external purposes. Common characteristics are operational controls based on variances from budgeted standards and reward systems tied to financial performance. Newer forms of results control also include non-financial controls (Widener, 2004).

Cultural control. In this study cultural control refers to control exercised by communicating the organizations missions, core values and informal codes of conduct. This definition is consistent with previous literature. Originally introduced by Ouchi (1979), cultural control is described with emphasis on employee selection, corporate missions, visions, values and codes of conduct. Cultural control is characterized by socialization that aims to create a collective goal congruence among the organizational members. Basic core values and organizational purpose is communicated to provide direction. The shared belief and understanding within the organization is expected to result in attainment of organizational objectives (Ouchi, 1979; Hansen, Otley and Van der Stede, 2003; Widener, 2007).

Contingency theory of management control. The contingency theory approach to management control research became popular in the mid 1970’s (Otley, 1980). The underlying logic is that there is no single best management control system, but rather, the appropriate management control system depends on the context and the unique circumstances of an organization (Chenhall, 2003). Research has focused on identifying specific circumstances, often related to the external environment, in which specific management control systems are appropriate (Otley, 1980). Whereas early research often focused on financial measures, non- financial measures have increased in importance (Otley, 2016). Furthermore, Otley (2016) argues that the context of many organizations has changed, and therefore research from the past may not stand true in the present day. In this study we focus on the interaction between management control systems and personality traits, rather than situational factors related to the external environment. Since the effectiveness of management control system depends on the unique circumstances of the organization, we argue it could also depend on the personality traits of the individuals in the organization.

2.2 Organizational Commitment

Porter et al. (1974) defines organizational commitment as the strength of an individual's identification with and involvement in a particular organization. Studies have shown that organizational commitment is positively related to several beneficial aspects such as work motivation (Strumpf and Hartman, 1984) job performance (Riketta, 2002), job satisfaction (Mosadehgrad, Ferlie and Rosenberg, 2008), low employee turnover (Porter et al., 1974), and

(10)

6 financial performance (Steyrer, Shiffinger and Lang, 2008). Thus, having a high organizational commitment among employees is likely important for a company to thrive and meet their objectives.

Several conceptualizations of organizational commitment have been made (Buchanan, 1974; Porter et al., 1974; Mowday, Steers and Porter, 1979). In this study, we will use the definition provided by Porter et al. (1974). Porter’s definition of organizational commitment is characterized by: 1) a strong belief in and acceptance of the organization's goals and values, 2) a willingness to exert considerable effort on behalf of the organization, and 3) a definite desire to maintain organizational membership.

According to Merchant and Van der Stede (2003), the purpose of management control systems is to guide employees in the right direction, and that otherwise employees might act in self-interest or perform poorly because they do not know what to do. Davis, Schoorman and Donaldson (1997) propose that employees with high levels of organizational commitment are less likely to act in self-serving manners, and more likely to act in the best interest of the organization. This leads us to the following assumption:

Assumption: Employees with high levels of organizational commitment are more likely to act in the best interest of the organization, compared to those with low levels of organizational commitment.

It is important to keep in mind that Porter’s (1974) definition of organizational commitment is similar to affective commitment proposed by Allen and Meyer (1990), since they both relate to value congruence and identification with the organization. Allen and Meyer (1990) propose a model of three different types of organizational commitment: affective commitment, normative commitment, and continuance commitment, whereas Porter et al.

(1974) propose a single, broader measure of organizational commitment. Although the definition of affective commitment by Allen and Meyer (1990) varies slightly from the definition we use here in this study, some of the hypotheses in this study will be based on previous research that studied affective commitment.

Personality traits and management control systems: antecedents of organizational commitment? Previous research has found significant relationships between some of the Big Five personality traits and organizational commitment. However, the studies have found mixed results (Erdheim, Wang and Zickar, 2006; Tziner et al., 2008; Choi, Oh and Colbert, 2015;

Farrukh, Ying and Mansori, 2017). Furthermore, literature suggests that management control

(11)

7 systems may influence organizational commitment (Kleine and Weissenberger, 2014; Verburg et al., 2018). Yet, the current understanding about the effects of cultural control and results control on organizational commitment is limited.

Caldwell, Chatman and O’Reilly (1990) found that an emphasize on sharing organizational values was positively related to commitment based on identification and shared values. Likewise, Kleine and Weissenberger (2014) found that cultural control had a positive effect on organizational commitment. Because cultural control aims to create a shared belief in the organization's core values, it seems reasonable to suggest cultural control affects organizational commitment positively. This leads us to the following hypothesis:

Hypothesis 1. Cultural control has a positive relationship to organizational commitment.

Literature has suggested that high levels of evaluation and control is likely to cause people in the organization to deviate from the organization’s goals (Ouchi, 1979). Caldwell, Chatman and O’Reilly (1990) found that the relationship between formal reward systems and commitment to the organization’s values was negative. Furthermore, Ogbonnaya, Daniels and Nielsen (2017), found that profit-related compensation was negatively (albeit weakly) correlated with organizational commitment. These findings suggest that results control affects organizational commitment negatively. This leads us to the following hypothesis:

Hypothesis 2. Results control has a negative relationship to organizational commitment.

However, Kleine and Weissenberger (2014) found that results control had no significant effect on organizational commitment. The ambiguous findings form previous research could possibly be explained by their survey design. Kleine and Weissenberger (2014) based the indicator of organizational commitment on top managers perception of their employee’s organizational commitment, which might have been biased. In this study, the measure of organizational commitment will be based on managers reports of their own organizational commitment.

Researchers have mostly looked at the relationships between the variables in separate studies. They have not focused on how the interaction between management control systems and personality traits may affect organizational commitment. In the next section, we define the Big Five personality traits, and construct hypotheses regarding bivariate relationships as well as the interaction effects of management control systems and personality traits.

(12)

8

2.3 The Five-factor Model of Personality

The five-factor model of personality, also known as the Big Five personality traits, consists of five personality traits: conscientiousness, openness, agreeableness, neuroticism, and extraversion (Cook, 2016, p. 168). The model has been further developed by Costa and McCrae and has reached widespread acceptance for classifying personal traits on a global level (Cook, 2016, p. 168). Although several different models of personality constructs have been proposed, the five-factor model is one of the most used because of its comprehensibility and robustness (Tokar, Fischer and Subich, 1998). Previous, studies have found several significant relationships between the five personality traits and various work-related aspects, including organizational commitment (Erdheim, Wang and Zickar, 2006; Tokar, Fischer and Subich, 1998). Descriptions of low scores and high scores of each personality trait are illustrated in Table 1.

Table 1. Description of the Big Five personality traits

Big Five trait Low score High score

Neuroticism Calm Anxious

Emotionally stable Unstable

Extraversion Reclusive Sociable

Submissive Dominant

Openness Conventional Original

Inflexible Flexible

Agreeableness Irritable Good-natured

Cold Warm

Conscientiousness Undisciplined Disciplined

Undependable Responsible

Note. This table is a modified version of a table illustrated by Tokar, Fischer and Subich (1998).

Previous research regarding the relationship between the Big Five personality traits and organizational commitment has found mixed results. This could partly be explained because different definitions of organizational commitment have been used. Erdheim, Wang and Zickar (2006), Choi, Oh and Colbert (2015) and Farrukh, Ying and Mansori (2017) studied the relationship between the Big Five personality traits and affective commitment, continuance

(13)

9 commitment and normative commitment, whereas Tziner et al. (2008) studied the relationship between the Big Five personality traits and organizational commitment as a single measure. In this study, we use a single measure of organizational commitment, although our definition is close to that of affective commitment. Thus, our hypotheses will partly build on previous findings regarding the relationship between big five personality traits and affective commitment. Furthermore, previous research has studied the relationship among different types of professions, whereas this study looks at the relationship specifically at management level.

The hypotheses regarding the relationships between the Big Five personality traits, results control, cultural control, and organizational commitment are presented next.

Conscientiousness. Conscientiousness captures how individuals relate to and strive for achievements. Characteristics such as diligence, perseverance and sense of duty are associated with this trait (Bozionelos, 2003). McCrae and Costa (2008, p. 274) describe people with high levels of conscientiousness as hardworking, purposeful, and disciplined while low level of the trait denotes individuals that are laid-back, unambitious, and weak-willed. Moreover, conscientious people are organized, effective in carrying out tasks, and have a high level of self-control (Rothmann and Coetzer, 2003). They tend to be active in planning processing rather than taking impulsive actions (Barrick, Mount, and Judge, 2001). Furthermore, this trait refers to people that are persistent, dependable, responsible, and careful (Rothmann and Coetzer, 2003). Employees with a high level of conscientiousness tend to be good organizational citizens. However, employees with low levels of conscientiousness are described as non- productive (Hogan and Ones, 1997, p. 850).

Previous research regarding the relationship between conscientiousness and organizational commitment has found mixed results. Tziner et al. (2008) expected a positive relationship between conscientiousness and organizational commitment but found that the relationship was negative. According to Tziner et al. (2008), the results could be explained by the compensation system in the organization they studied. The compensation system was based on rank, which could possibly have caused negative feelings for individuals with high levels of conscientiousness, which they propose might have led to lower organizational commitment.

However, Erdheim, Wang and Zickar (2006), Choi, Oh and Colbert (2015) and Farrukh, Ying and Mansori (2017) found a positive relationship between conscientiousness and affective commitment. Overall, the findings seem to suggest that individuals with high levels of conscientiousness are more likely to develop high organizational commitment. This leads us to the following hypothesis:

(14)

10 Hypothesis 3. Conscientiousness has a positive relationship to organizational commitment.

Previous research has not yet established how results control affects the organizational commitment of individuals with high levels of conscientiousness. Nevertheless, previous studies in similar research areas indicate that the interaction of formal controls and conscientiousness may affect organizational commitment negatively. Fong and Tosi (2007) conducted a study regarding the interaction of conscientiousness and monitoring and found that individuals with low levels of conscientiousness exerted more effort when monitored, whereas highly conscientious individuals did not increase their effort to the same extent when monitored.

Although the concepts in their study were not identical to results control and organizational commitment, they share some similarities. In our study, we defined organizational commitment as defined by Porter (1974), which includes a willingness to exert considerable effort on behalf of the organization. Likewise, results control could be perceived as a form of monitoring. This indicates that the relationship between conscientiousness and organizational commitment should be lower when results control is high. Additionally, an older study by Meadows (1980) suggest that more formal environments were associated with frustration for persons with a high need for achievement. Since a high need for achievement is typical of conscientious individuals, this could be an indication that individuals with high levels of conscientiousness are likely to respond negatively to formal control systems such as results control. Thus, the relationship between conscientiousness and organizational commitment seems likely to be lower when results control is high. To test this relationship, we formulate the following hypothesis:

Hypothesis 4. Results control has a negative moderating effect on the relationship between conscientiousness and organizational commitment.

Openness. Openness is characterized by flexibility of thought, inventiveness, active imagination, aesthetic sensitivity, attentiveness to feelings, a preference for variety, intellectual curiosity, and independence of judgement (Bozionelos, 2003). Moreover, a high level of openness is associated with willingness to question authorities and stronger positive and negative emotions than closed individuals. Individuals with a low level of openness tend to be conventional and conservative, prefer the familiar to the novel and express low emotional response (Rothmann and Coetzer, 2003).

(15)

11 Previously, Tziner et al. (2008) found a positive relationship between openness and organizational commitment. However, Erdheim, Wang and Zickar (2006) advised not to hypothesize the relationship between openness and organizational commitment because it can be a double-edged sword, which causes employees to feel emotions stronger, and could lead to either higher or lower organizational commitment. Furthermore, Farrukh, Ying and Mansori (2017) expected a positive correlation between openness and affective organizational commitment but found that the correlation was negative. This, they suggest, could be because individuals with high levels of openness value switching jobs. Therefore, the relationship between openness and organizational commitment will not be hypothesized in this study.

However, it seems logical that high level of results control would inhibit the intellectual curiosity and independence that characterize high level of openness. Thus, it seems likely that individuals with high levels of openness will have relatively lower organizational commitment in organizations with high levels of results control, compared to individuals with lower levels of openness. To test this relationship, we formulate the following hypothesis:

Hypothesis 5. Results control has a negative moderating effect on the relationship between openness and organizational commitment.

Since openness can act as a double-edged sword (Erdheim, Wang and Zickar, 2006), it seems likely that employees with high levels of openness are likely to strongly react either positively or negatively to cultural control. Thus, effect of cultural control on individuals with high levels of openness may vary highly from person to person. The effect is likely to be dependent on other factors which makes it difficult to hypothesize a positive or negative interaction effect. Therefore, the interaction effect of openness and cultural control is not hypothesized.

Agreeableness. High level of agreeableness is characterized by generosity, honesty, and modesty whereas low levels of agreeableness is characterized by competitiveness, selfishness, aggression, and arrogance (Rothmann and Coetzer, 2003; McCrae and Costa, 2008, p. 274).

High level of agreeableness is also associated with altruism, friendliness and individuals tend to be engaged in their work when it has a correlation to maintenance and enhancement of their feelings of personal worth (Bozionelos, 2003). Sympathy and helpfulness toward others and the belief that others will be equally helpful is typical for agreeable individuals. Furthermore, cooperativeness is high in individuals with a high level of agreeableness suggesting that high

(16)

12 levels of agreeableness is important in occupations where teamwork is relevant (Judge et al., 1999).

Previously, Erdheim, Wang and Zickar (2006), Choi, Oh and Colbert (2015) and Farrukh, Ying and Mansori (2017) found a positive correlation between agreeableness and affective organizational commitment. Tziner et al. (2008) also found a positive correlation between agreeableness and organizational commitment. The inherent nature of the agreeableness trait, together with previous findings, suggest that individuals with high levels of agreeableness are more likely to develop high organizational commitment. This leads us to the following hypothesis:

Hypothesis 6. Agreeableness has a positive relationship to organizational commitment.

Furthermore, cultural control is based on communication, sharing the same values, and creating a feeling of group coherence and organizational purpose (Ouchi, 1979; Hansen, Otley and Van der Stede, 2003; Widener, 2007; Sandelin, 2008). Since individuals with high levels of agreeableness value cooperativeness and friendliness (Bozionelos, 2003), it seems reasonable to suggest that they will be more committed to organizations that focus on cultural control and less committed to organizations with less cultural control. To test this relationship, we formulate the following hypothesis:

Hypothesis 7. Cultural control has a positive moderating effect on the relationship between agreeableness and organizational commitment.

Neuroticism. Neuroticism is associated with emotional instability (Rothmann and Coetzer, 2003). A high level of neuroticism implies a variety of distressing emotions, for example fear, anger, dejection, shame, and sadness (Costa and McCrae, 2011). Low levels of neuroticism are associated with traits such as calmness and stability and the ability to handle stressful situations easier without becoming upset (Rothmann and Coetzer, 2003). Neuroticism is also characterized by excessive worry, pessimism, low confidence, and negative emotions Bozionelos (2003). Malouff et al. (1990) found that neuroticism was negatively related to the tendency to be goal oriented. Furthermore, individuals with a high level of neuroticism tend to exhibit self-pity and look at themselves as victims (Barrick and Mount, 1991).

In previous research Erdheim, Wang and Zickar (2006), Choi, Oh and Colbert (2015) and Farrukh, Ying and Mansori (2017) found a negative correlation between neuroticism and

(17)

13 affective organizational commitment, whereas Tziner et al. (2008) found no correlation between neuroticism and organizational commitment. The differences between their findings could possibly be explained by the differences in organizational settings of the studies or the different aspects of organizational commitment measured. However, the characteristics of worry, pessimism, low confidence, and tendencies of negative emotions imply that individuals with high levels of neuroticism are less likely to develop organizational commitment compared to those with low levels of neuroticism. To test this relationship, we formulate the following hypothesis:

Hypothesis 8. Neuroticism has a negative relationship to organizational commitment.

Furthermore, the low confidence and negative relation to goal orientation which is typical for neurotic individuals suggest that individuals with high levels of neuroticism will be less likely to have high organizational commitment if they are subject to high levels of results control. This suggests that highly neurotic persons are less likely to be committed in organizations with high levels of results control. To test this relationship, we formulate the following hypothesis:

Hypothesis 9. Results control has a negative moderating effect on the relationship between neuroticism and organizational commitment.

Extraversion. Typical characteristics of extraversion are sociability, assertiveness, social dominance, ambition, tendencies towards action (Bozionelos, 2003). Other facets are excitement-seeking and positive emotions (Costa and McCrae, 2011). Extraverts place high value on close interpersonal relationships and enjoy the company of others, while introverts are characterized as being quiet, reserved and less interpersonally effective (Watson and Clark, 1997, p. 776). Introverts, or individuals with low levels of extraversion tend to be internally oriented, focusing on internal, subjective experiences, own thoughts, feelings, and perceptions (Watson and Clark, 1997, p. 767-769). It has been found that extraversion is a predictor of job performance in jobs characterized by social interaction, such as sales and police work (Rothmann and Coetzer, 2003). Additionally, extraversion positively relates to a preference for intrinsically motivating job features (Furnham, Forde and Ferrari, 1999).

Erdheim, Wang and Zickar (2006), Choi, Oh and Colbert (2015) and Farrukh, Ying and Mansori (2017) previously found a positive correlation between extraversion and affective

(18)

14 organizational commitment, whereas Tziner et al. (2008) found no significant correlation between extraversion and organizational commitment. The different findings could possibly be explained by the differences in organizational settings of the studies or the different aspects of organizational commitment measured. Based on the characteristics of extraversion, such as ambition, self-promotion, and the positive approach, it seems likely that individuals with high levels of extraversion are more likely to develop higher organizational commitment compared to those with low levels of extraversion. The tendency to be dominant suggests a preference to have a central position in their workplace which would encourage higher organizational commitment. This leads us to the following hypothesis:

Hypothesis 10. Extraversion has a positive relationship to organizational commitment.

Furnham, Forde and Ferrari’s (1999) findings that extraverts prefer intrinsically motivating job features imply that extraverts are likely to have higher organizational commitment in organizations in which they find personal purpose and fulfil higher-level psychological needs. Additionally, extraverts enjoy social interaction and value personal relationships, which indicates that they will respond positively to a focus on creating shared values, purpose, and group coherence. Since individuals with high levels of extraversion value social interaction and personal relationships, it seems reasonable to suggest that they will be more committed to organizations that focus on cultural control and less committed to organizations with less cultural control. To test this relationship, we formulate the following hypothesis:

Hypothesis 11. Cultural control has a positive moderating effect on the relationship between extraversion and organizational commitment.

2.4 Summary of Hypotheses

In the previous section, we hypothesized the relationship between management control systems and organizational commitment, the relationship between personality traits and organizational commitment, and lastly, the moderating effect of management control systems on the relationship between personality traits and organizational commitment. The hypotheses are summarized in Table 2. The statistical methods used to test the hypotheses are presented in the next chapter, along with the research methods used in this study.

(19)

15 Table 2. Summary of hypotheses

Hypothesis

H1 Cultural control has a positive relationship to organizational commitment.

H2 Results control has a negative relationship to organizational commitment.

H3 Conscientiousness has a positive relationship to organizational commitment.

H4 Results control has a negative moderating effect on the relationship between conscientiousness and organizational commitment.

H5 Results control has a negative moderating effect on the relationship between openness and organizational commitment.

H6 Agreeableness has a positive relationship to organizational commitment.

H7 Cultural control has a positive moderating effect on the relationship between agreeableness and organizational commitment

H8 Neuroticism has a negative relationship to organizational commitment.

H9 Results control has a negative moderating effect on the relationship between neuroticism and organizational commitment

H10 Extraversion has a positive relationship to organizational commitment.

H11 Cultural control has a positive moderating effect on the relationship between extraversion and organizational commitment.

3. METHOD

In this chapter we present the methodological choice of the study and address the approach applied to answer the research questions. We elaborate the choice of variables and the indicators used to measures these, and the statistical methods used to test relationships between the variables. Furthermore, we outline how the data was collected and discuss the reliability and validity of the study.

3.1 Research Design

In this study we aimed to provide a better understanding about the relationship between the Big Five personality traits, management control systems and organizational commitment.

To answer the research questions, we quantified the relationships between the personality traits, management control systems and organizational commitment with a deductive approach.

(20)

16 In order to test the hypotheses, a quantitative research strategy was used. A quantitative research strategy is typically used for deductive studies and testing theories (Bryman and Bell, 2011, p. 27). The choice of a quantitative strategy was based on the problematization and objectives of the study. In order to answer the research questions, it was necessary to use a strategy that enabled us to gather data from multiple organizations with different levels of result control and cultural control, as well as their managers having to consist of a wide array of different personalities. A qualitative strategy such as interviewing different companies would not have provided the broad data necessary to answer the research questions.

We used a cross-sectional design, using self-completion questionnaires for data collection. A cross-sectional design is appropriate when studying quantitative relationships between two more variables (Bryman and Bell, 2011, p. 53). Furthermore, Bryman and Bell (2011, p. 53-54) suggest that a cross-sectional design should be used in research that requires variation of data. To answer our research questions, we needed data with variation in personality traits and organizational commitment of respondents, as well as variation in the management control systems in the organizations they work in.

In this study, using self-completion questionnaires for data collection was considered the most appropriate method in sense of feasibility and appropriability, since it can be sent to many potential respondents by email. Data was collected through a web survey that was sent out through email. The questionnaire was designed in the survey administration application Google Forms, because of its ease of use. Using a web survey reduced the risk of error in the processing of data (Bryman and Bell, 2011, p. 663)

3.2 Variables

Below is an outline of how the concepts used in our hypotheses were operationalized.

Indicators for each concept were selected and designed based on existing literature and their relevance to our research questions. Bryman and Bell (2011, p. 154) recommend using several questions to measure a concept. We based our variables on indicators that had previously been found to be reliable (Mowday, Steers and Porter, 1979; Jaworski and MacInnis, 1989;

Donnellan et al., 2006; Kleine and Weissenberger, 2014), and that were in line with the concepts used in this study. Altogether, we had eight variables that consisted of 37 questions.

These were measured with a 1 to 5 Likert scale, in which 1 represented do not agree at all, and 5 represented fully agree (in Swedish). Additionally, we included a question about the respondent’s management level.

(21)

17 Organizational commitment. We based our operationalization of organizational commitment on Porter et al.’s, (1974) definition of organizational commitment. Eight questions were selected out of a 15-item questionnaire that was originally designed by Mowday, Steers and Porter (1979).

Big Five personality traits. We based our personality trait indicators on a scale developed by Donnellan et al. (2006). Donnellan et al. (2006) named this scale the Mini-IPIP as they developed a short form of the 50-item IPIP-scale (International Personality Item Pool – scale) developed by Lewis Goldberg. Even if the Mini-IPIP scale only consists of 20 questions, its results have been found to be consistent with the 50-item IPIP-scale which implies that it is a valid indicator for each personality trait. We included all five personality traits, which made up one variable each (openness, conscientiousness, extraversion, neuroticism, and agreeableness).

Four questions were included for each personality trait.

Results control. The items we chose for the variable results control were based on a five- item scale originally developed by Jaworski and MacInnis (1989), as well as an adapted version of this construct developed by Kleine and Weissenberger (2014). Five questions were included for this variable.

Cultural control. The items we chose for the variable cultural control were based on Kleine and Weissenberger (2014), who based their construct on Ouchi’s (1979) clan control theory and Widener’s (2007) belief system. Four questions were included for this variable.

Control variable. Organizational commitment can be explained by various factors (Allen and Meyer, 1990). To limit the number of questions in the questionnaire, only one control variable was included. Since both top managers and middle managers were surveyed, a question about management level was included in the questionnaire, to test if management level would have any effect on the results. Furthermore, this control variable was used to exclude respondents that were neither top managers nor middle managers.

3.3 Questionnaire Design

Translation of items. All our questions were based on questionnaires originally written in English, which we then translated to Swedish. Translated questionnaires do not necessarily possess the same properties as the original ones (Behling and Law, 2000). One criterion for evaluating the translation technique is informativeness, or the degree to which a translation technique provides an equivalent semantic meaning as the original questions.

(22)

18 To meet the informativeness criteria, we carried out multiple translations from English to Swedish, which we then compared to the original questions in English. We determined the most accurate Swedish translations, which we then chose for the questionnaire. More advanced techniques such as parallel blind tests have been proposed to provide a higher degree of informativeness. However, it is also relevant to consider the practicality of the technique (Behling and Law, 2000). Considering the scope of our study and time constraints, we chose not to opt for a parallel blind test or similar techniques, which are difficult and time-consuming.

Another relevant criterion is security, the degree to which readers can compare the original questions with the translations (Behling and Law, 2000). To meet the security criteria, we have included all questions and the corresponding translations to Swedish in Appendix A.

Order of items. Literature indicates that the order of the questions may influence, and cause biased answers (Dillman, Smyth and Christian, 2014). A general guideline is to group questions according to the topic covered. Therefore, we grouped together questions regarding the individual (Big Five personality) traits together, and grouped questions related to the organizational context together (organizational commitment and management control systems).

Furthermore, as recommended by Dillman, Smyth and Christian (2014) we paid careful attention to the choice of first question, in order to reduce the likelihood that the respondent quits halfway through.

Pre-testing questionnaire. Converse and Presser (1986) argue questionnaires often include questions that are too complicated or too demanding to answer. Instead, they suggest a simple language should be used. To make sure our questions were simple enough, we conducted pilot studies in three rounds, with 13 persons responding, in total. We asked our test respondents questions regarding testing task difficulty and adjusted our questions accordingly. Converse and Presser (1986) also suggest pre-testing for meaning and for variation. Within the scope and time constraints of this study, it did not seem feasible to test whether the respondents would interpret our questions in a consistent way to what we intended.

Besides testing the questions, it is also important to test aspects of the questionnaire, such as the flow, respondent interest and order of questions (Converse and Presser, 1986). Because those aspects can be quite subjective and difficult to measure, we asked the test respondents for feedback regarding those aspects and considered their answers when designing the questionnaire.

(23)

19

3.4 Sample Selection

The study surveyed top and middle managers in medium-sized companies. Bryman and Bell (2011, p. 176) argue that for the findings of a study to be generalizable, the sample must be representative of the population. Because the number of managers and the variance in their personality traits, as well as the variance in the management control systems they are exposed to was unknown, we made use of convenience sampling, which is the most common sampling method in business and management research (Bryman and Bell, 2011, p. 191). We reached out to medium-sized companies that were randomly selected from the retriever business database.

In total, 30 companies participated in this study. The companies that participated in the study were instructed to send the survey further to their managers. Thus, the respondents were not based on random selection, but selected by the contact person at the company. Consequently, the sampling method may have imposed unavoidable limitations on the generalizability of this study.

Our choice to collect data from medium-sized companies was based on the characteristics of this study, which focuses both formal and informal control systems. Compared to larger companies, medium-sized enterprises rely less on formal management control systems and are characterized by an interplay of formal and informal systems (Pesalj, Pavlov and Micheli, 2018). Furthermore, large companies were not included because it may be more difficult to reach out to managers in larger companies. Similarly, small companies (under 50 employees) were not included, because they may not have many managers in their organization.

3.5 Data Collection

The contact persons at the companies that participated in this study sent the surveys further to top and middle managers. Anonymity and confidentiality were reassured, both for the organization, and for the individual respondents. In our collected dataset, individual responses were not distinguishable from the rest. Thus, the privacy principle was respected (Bryman and Bell, 2011, p. 137). Furthermore, anonymity in survey studies has been found to reduce respondent bias (Krumpal, 2013). To increase the response rate, we sent out reminders to the companies. For anonymity reasons, neither IP-addresses nor e-mail addresses were collected with the survey. This means there was no way to control that the same respondent did not answer the questionnaire multiple times.

(24)

20

3.6 Data Analysis methods

The data obtained was analyzed in R and in SPSS. The questionnaires included responses that ranged from 1 to 5 on a Likert-scale. Reverse questions were re-coded, so that for instance, a value of 5 became 1, as recommended by Bryman and Bell (2011, p.240). This made it possible to combine the questions to a composite index. The different questions in the questionnaire were combined in order to create unified mean values for each variable and person. This was done by summing the score of an individual within each variable, and then dividing the sum by the number of questions within that variable. Because the items were combined into composite indexes, the variables were treated as interval scale, which enabled the use of parametric statistical methods (Berntson et al., 2016, p. 38). In total, 37 questions were combined into 8 separate variables: organizational commitment, results control, cultural control, neuroticism, extraversion, agreeableness, conscientiousness, and openness.

Significance levels. The relationships that we tested were considered significant if the p- value was under 0.05. However, since we used a convenience sample which entails certain sample errors, attention was also paid to p-values under 0.10. This is consistent with Manderscheid (1965) and Dahiru (2008) who suggest that the significance level should be decided by reasoning instead of looking blindly at standard levels.

Pearson’s correlation. To test the relationship between variables, we used correlation tests. Pearson’s correlation is used to examine the relationship between variables that use interval scales. Pearson’s correlation coefficient ranges from 1 to -1. A correlation coefficient of +1 implies a perfect positive relationship, a coefficient of -1 implies a perfect negative relationship and a correlation coefficient of 0 implies there is no relationship (Bryman and Bell, 2011, p. 347).

Regression models.In addition to the Pearson’s correlation, we used multiple regression analyses with organizational commitment as a dependent variable. Although causal relationships are difficult to prove (Bryman and Bell, 2011, p.346), we found it unlikely that organizational commitment would affect the organization’s management control systems or personality traits.

To test interaction effects, we used multiple hierarchical regression models. Interaction effects, or moderating effects occur when the effect of one independent variable depends on the level of another independent variable (Pallant, 2011, p. 265). Equations 1, 2 and 3 illustrate the hierarchical regression models used in this study. In the equations below, Y represents organizational commitment, whereas X represents personality traits, and Z represents

(25)

21 management control systems (see Table 3). The regression coefficient for each variable is represented by B. In the third step, the interaction term B3XnZnis included to show the interaction effect of the two independent variables. If the regression coefficient of B3XnZn is significant, there is a significant interaction effect (Fairchild and MacKinnon, 2009). Multiple regressions were carried out separately for each hypothesis. Furthermore, the change in R-squared between the second step and third step was tested through F-tests to establish if adding the interaction term increased the explained variance in organizational commitment (Pallant, 2011, pp. 165- 166). Additionally, the interaction effects were plotted (see Appendix B) on the basis of the procedures recommended by Cohen and Cohen (2013, ch. 7).

First step: Simple linear regression

Y = B0+B1Xn+ error (1)

Second step: regression model with third variable included

Y = B0 + B1Xn + B2Zn+ error (2)

Third step: regression model with interaction term included

Y = B0 + B1Xn + B2Zn + B3XnZn + error (3)

Table 3. Notation of variables in the regression models

Variable

Y Organizational commitment Xn Big Five Personality traits Zn Management control systems XnZn Interaction term

3.7 Reliability and Validity

Reliability. Internal reliability refers to the degree that the items in a construct relate to the same thing and thus, measure the same underlying attribute (Bryman and Bell, 2011, pp.

158-159; Pallant 2011, p. 6). Internal reliability can be measured by Cronbach’s alpha with a recommended minimum alpha of 0.7. Cronbach’s alpha values were calculated for all the items, which are shown in Table 5. All the variables except neuroticism had high levels of internal reliability. Neuroticism had an alpha value of 0.64, which indicates a relatively low reliability

(26)

22 (Pallant, 2011, p. 6). However, these measures were developed by Donnellan et al. (2006) who based on empirical findings, argue that the indicators are reliable and useful for survey research.

Internal validity. Internal validity refers to the extent to which the causality relationships measured are true (Bryman and Bell, 2011, p. 42). Our research used a cross-sectional design, which means the population is observed at a specific point in time. This means the cause and effect relationships of the independent and dependent variables used in this study are difficult to prove (Bryman and Bell, 2011, p. 56). Furthermore, there may be other variables that were not included in the study that may also affect organizational commitment.

Measurement validity. Measurement validity deals with the question if a measurement used in a study really measures what it is intended to measure (Bryman and Bell, 2011, p. 159).

The indicators in this study were based on theoretical concepts that had priorly been operationalized, which makes comparison to previous studies easier. The indicators used to measure organizational commitment included a limited number of questions from the original questionnaires, which could have resulted in a lower validity. Furthermore, there was no way to ensure that the measurements really measured what they intended to measure. Because of the relatively low number of questions per indicator, it is possible that the indicators only partially captured what they intended to measure, while other elements may have been omitted.

Replicability. To enable researchers to replicate findings of other researchers, the procedures of studies conducted must be reported in detail (Bryman and Bell, 2011, pp. 41-42).

To foster replicability, we have delivered a high level of transparency on the procedures used to conduct this study. The questions used for data collection are displayed in appendix A and the method chapter is delineated in such a way that a replication of this study should be readily enabled.

External validity. In comparison to case studies or in-depth interviews, the methods used for this study imply better generalizability (Bryman and Bell, 2011 p. 61). However, the respondents were not based on random sampling, which also implied limitations with generalizability of this study. Moreover, a larger sample size could have provided better generalizability (Bryman and Bell, 2011, pp. 185-187).

Method limitations.As we explained in previous sections, the methods used in this study imply limitations regarding the generalizability of the findings. Due to difficulties data collection based on random sampling a convenience sample had to lay ground to the collection of data. Hence, the generalizability of this study is limited. Nevertheless, the data obtained showed that the respondents consisted of a wide spectrum of personalities and worked under

(27)

23 different types of management control systems. Thus, the results should serve as indications of broader trends.

3.8 Chapter summary

In this chapter we presented the methodology and research design used to answer the research questions. The thesis had a quantitative strategy and examined the relationship among eight variables with organizational commitment as the dependent variable. Indicators of concepts were constructed based on previous studies that had used similar indicators.

Questionnaires were sent out to managers in medium-sized companies. The collected data was analyzed through quantitative methods including Pearson's correlation, simple regression, and multiple hierarchical regression. The data analysis is presented in the next chapter.

4. RESULTS

In this chapter the results and data analyses of the data collected are presented. First, descriptive statistics are presented, followed by hypotheses testing. The hypotheses are tested through correlation analyses and multiple regression models. Additionally, other results that are relevant in relation to the research questions are presented.

4.1 Descriptive Statistics

In total, 30 companies participated in the study. Out of those, 25 companies reported that they sent out the questionnaire to 127 managers. However, five companies did not report how many managers they sent out the questionnaire to. Thus, the sample size was probably larger than 127. In total, 108 questionnaire responses were obtained, out of which 52 respondents answered that they were part of top management, whereas 51 responded that they were middle managers. Five responses were excluded from analysis, since they responded that they were neither top managers nor middle managers.

(28)

24 Table 4. Descriptive sample statistics

Number of companies 30

Sample size 127+

Respondents 108

Non-response Unknown

Invalid responses 5

Valid responses 103

Table 5. Descriptive variable statistics

Variable Questions Mean Standard

deviation

Median Alpha

Organizational commitment 8 3.92 0.55 3.87 0.79

Cultural control 4 3.80 0.76 3.76 0.82

Results control 5 3.36 1.04 3.4 0.88

Conscientiousness 4 3.82 0.73 4 0.71

Openness 4 3.60 0.71 3.5 0.73

Agreeableness 4 4.06 0.65 4 0.77

Neuroticism 4 2.27 0.68 2.25 0.64

Extraversion 4 3.33 0.73 3.25 0.76

Table 5 gives an overview on the descriptive statistics of the variables in this study. The mean value of organizational commitment was 3.92 which indicates that the general level of organizational commitment was rather high among the respondents. Furthermore, the reported levels for cultural control (mean = 3.79) were higher than the reported levels for results control (mean = 3.36). Moreover, we can see that the strongest personality traits of the respondents were agreeableness and conscientiousness, whereas the weakest personality trait was neuroticism.

To control the internal reliability of each variable, Cronbach’s alpha was calculated. The results show that all measures except neuroticism had a high, or acceptable levels of reliability (Pallant, 2011, p. 6). Alpha values were above 0.7 for all variables except neuroticism which had a value of 0.64.

(29)

25

4.2 Data Analysis and Hypotheses Testing

In this section, the data analyses that were used to answer the research questions are presented. First, the correlation between variables is presented in a correlation matrix (Table 6), then the tests conducted to test the hypotheses are presented. Lastly, other results that are relevant in relation to the research questions, as well as unexpected results are presented.

Correlation matrix. A requirement to perform a regression analysis is that there is no multicollinearity between the variables. Pearson’s correlation was calculated for the correlations between all variables used in this study. Table 6 shows that the correlation between all variables was under 0.9, which means there was no multicollinearity between the variables (Pallant, 2011, p.150- 151).

Table 6. Correlation matrix (Pearson’s)

Variable 1 2 3 4 5 6 7 8

1. Organizational commitment

1

2. Cultural control 0.62** 1

3. Results control 0.43** 0.55** 1

4. Conscientiousness 0.13 0.18 0.07 1

5. Openness 0.08 0.18† 0.03 -0.17† 1

6. Agreeableness 0.18† 0.35** 0.03 0.06 0.35* 1

7. Neuroticism -0.31** 0.04 0.05 -0.21* -0.13 -0.13† 1

8. Extraversion -0.02 -0.03 0.034 -0.22* 0.32** 0.11 -0.06 1 Two-tailed significance levels: ** p < 0.01, * p < 0.05, † p < 0.1

Findings. To answer the research questions, we conducted multiple statistical tests, in addition to the hypothesized relationships. Correlations were found between cultural control and organizational commitment (r = 0.62, p < 0.01) and between results control and organizational commitment (r = 0.43, p < 0.01). Although both cultural control and results control were positively correlated to organizational commitment, a multiple regression analysis showed that when both cultural control and results control were included in the equation, results control did not have a significant relationship to organizational commitment (B = 0.07, NS).

Furthermore, correlations were found between agreeableness and organizational commitment (r = 0.18, p < 0.10), and neuroticism and organizational commitment (r = -0.31, p < 0.01).