Relationships within organizations:
The effect of different relational models on firm
performance
Master Thesis
MSc. in Business Administration – Strategy Track Amsterdam Business School, University of Amsterdam
Student: Christine van Rijs Student no.: 10884416
Date of submission: 29 January 2016 Supervisor: Dr. F.M. Bridoux
2
Statement of Originality
This document is written by Christine van Rijs who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
3 ABSTRACT
Since organizations are made up by humans and they would not exist without them, it is interesting to study what drives them to contribute (or not) to value creation in favor of the firm. We all know that people are different, and that they might have different reasons to contribute or not. It might be difficult for managers to motivate diverse individuals, and to get them to contribute to the collective, namely the firm. In this research a particular form of value creation stands central, namely organizational citizenship behavior: behavior that is not written in the contract, and thus not per se individually rewarded. It is interesting to study the reasons why individuals would contribute to this or not. By considering four, rather than two approaches to stakeholder management, I wanted to shed new light on people’s motivation to contribute. I did this by applying relational models theory, as developed by Fiske. I thought that different relational models would lead to different levels of organizational citizenship behavior towards the organization (OCBO). Moreover, my expectations were that organizational commitment would count as a mediator in the relationship between some relational models and OCBO. Lastly I argued that social dispositions of individuals might also play a role in the relationship between relational models and organizational commitment. To underpin my expectations I surveyed 186 individuals who are currently employed. Results partly supported my argumentation; some relational models indeed lead to more OCBO than others. Moreover, I found out that affective organizational commitment mediates in some relationships. Lastly, for some relational models a match between the relational model and an employees’ social disposition seems to positively affect commitment to the organization. This research contributes to management practices, since it sheds new light on the drivers of OCBO. In a world where there is a focus on the individual, it is interesting to study what drives and motivates them.
4
TABLE OF CONTENTS
1. INTRODUCTION ... 6
2. LITERATURE REVIEW ... 9
2.1. Stakeholder theory ... 9
2.2. Relational models theory ... 11
2.3. Employees as stakeholders and OCBO ... 14
2.4. Organizational commitment as a mediator ... 16
2.5. Social dispositions as a moderator ... 17
2.6. Literature gap and research question ... 19
2. THEORETICAL FRAMEWORK ... 21
1. Relational models, employees and OCB ... 21
2. Relational models, organizational commitment and OCBO ... 24
3. Relational models, social dispositions and the influence on organizational commitment ... 26
4. METHODOLOGY ... 28 4.1. Research Design ... 28 4.2. Sample ... 31 4.3. Measures ... 32 4.3.1. Independent variables... 32 4.3.2. Dependent variable ... 34 4.3.2. Mediating variables... 35 4.3.3. Moderating variable ... 35 4.3.4. Control variables ... 36 5. RESULTS ... 38 5.1. Reliability Analysis ... 38 5.2. Descriptive Statistics ... 40 5.3. Correlations Analysis ... 43 5.4. Normality Analysis ... 48
5
5.5. Regression Analysis ... 49
6. DISCUSSION ... 76
6.1. Relational models and OCBO ... 76
6.2. Relational models, organizational commitment and OCBO ... 78
6.3. Relational models, personal values and organizational commitment ... 79
6.4. Control variables ... 80
6.5. MANAGERIAL IMPLICATIONS ... 81
6.6. LIMITATIONS AND AVENUES FOR FUTURE RESEARCH ... 82
8. CONCLUSION ... 83
6 1. INTRODUCTION
Humans play an important role within organizations, especially in the knowledge economy where knowledge may be the primary source of competitive advantage (Grant, 1996). What is more, without humans, organizations would not even exist (Felin & Foss, 2005). Despite the important role of these individuals within organizations, little attention has been yet paid to them in the strategic organization (Felin & Foss, 2005). In fact, there has been an increasing focus on collective conceptualizations such as routines and structures; the so-called ‘macro theories’. These theories take the organization for granted, without explaining how collective action emerges and how action and organization might take on new forms (Felin, Foss & Ployhart, forthcoming). Felin and Foss (2005) have criticized the lack of focus on the individual in strategy literature and point out the need for micro-foundations in strategy. After all, organizational performance depends on employees and their personal behaviour (Foss, 2007).
Another stream in the strategy literature is the stakeholder theory, which highlights the importance of relationships to get stakeholders creating value for the organization (Freeman 2000; Freeman & Phillips 2002). Stakeholders are those who can affect the firm or be affected by it (Freeman, 2000). It is argued that the firm could gain a competitive advantage if it is able to develop relationships with its stakeholders, based on trust and cooperation (Jones 1995; Donaldson & Preston 1995). The stakeholder literature defines two approaches to govern value creation namely the ‘transactional’ approach based on economic logic, and the ‘relational’ approach which is based on moral premises. The theory argues that the relational one outperforms the transactional one.
The concept of relationships between individuals is furthermore widely discussed in the work of Fiske (1991, 1992, 2004, 2012). According to Fiske (1992) people are sociable, which
7
means that “people organize their social life in terms of their relations with other people” (Fiske, 1992, p. 689). He argues that people use just four relational models in their interaction with others.
My research aims to shed new light on how to get stakeholders to contribute to value creation for the firm by considering these four relational models of Fiske, rather than the two approaches explained in stakeholder literature. By applying relational models theory on stakeholder theory I try to give a better explanation of stakeholder behaviour and what drives them to contribute to value creation.
Since value creation is a broad term and could be interpreted in different ways (what is ‘value’ exactly, and how is it created?) I focus on a more narrow term, namely organizational citizenship behaviour (OCB). OCB is defined as “individual behaviour that is discretionary, not directly or explicitly recognized by the formal reward system, and that in the aggregate promotes the effective functioning of the organization” (Organ, 1988, p.4). It is a particular form of value creation because of two reasons: 1) the behaviour is not per se captured in the job description and 2) it is not directly compensated by the reward system. Based on existing literature about these topics I think it is fruitful to study the impact of different relationships on this particular type of value creation.
Furthermore, the contribution of an individual to OCB might also depend on how committed he/she is to the organization. It is namely argued in literature that more commitment leads to more contribution of OCB (Williams & Anderson, 1991). Lastly I take individual differences into account (measured by the personal values instrument as developed by Schwartz, 2012), to see if this also might impact the relationship between relational models and organizational commitment.
8
By using a survey, the contributions to theory are threefold: first, results show how the four relational models affect OCB. Secondly, the mediating effect of organizational commitment on this relationship is studied. Finally the role of personal values in the relationship between relational models and organizational commitment is researched.
Besides contributing to theory, this research also has implications for management practices. It shows how different relational models (which can be influenced by the firm) may affect employees’ OCB: relationships framed as communal sharing will for example lead to more contribution to OCBO, than relationships framed as market pricing. If managers would like to increase the contribution of employees to OCB, they then could try to reshape the type of relationship. Furthermore, results show that when the social disposition of the employee and the type of relationship with the firm are similar, employees feel more committed to the organization (since they can identify themselves with it). Managers can thus strive to create a relationship that fits with the employees’ social disposition, in order to make them create value for the firm. This might be a difficult task however, since organizations/departments may consist of employees with diverse characteristics. It is then difficult to decide which relational model you would like to create.
9 2. LITERATURE REVIEW
In the following paragraphs the main literature regarding the topics introduced above will be discussed, since that are the concepts that will be at the centre of this thesis. I will start with explaining stakeholder theory and the future research possibilities that will be researched in this thesis. After that I will discuss in greater detail the relational model theory of Fiske because that will be the main construct of the thesis. Then I will discuss a particular type of stakeholders (employees) where the focus will lie on in my thesis, and a particular type of value creation (organizational citizenship behaviour towards the organization). For the mediating effect between relational model theory and organizational citizenship behaviour I will make use of organizational commitment, which also will be explained. Finally there will be a paragraph about social dispositions, which will count as a moderator in the research.
After the literature review the gap that I have found will be mentioned. Finally the research question will be formulated.
2.1. Stakeholder theory
Stakeholders are those who can affect the firm or be affected by it (Freeman, 2000). Stakeholder theory argues that they all have a stake in the firm and that the purpose of the firm should thus be to serve the interests of all these parties involved (Berle & Means, 1932; Freeman & Reed, 1983). The theory also focuses on the role of stakeholders when explaining firm performance, rather than describing it in monetary terms (Freeman, 2000). Donaldson and Preston (1995) argue that stakeholder theory has been used in different ways, and that there is confusion about its nature and purpose. They make a distinction between three types of uses: descriptive/empirical, instrumental and normative. In the first one stakeholder theory is used to describe something, for example how
10
firms operate. The instrumental one is used to identify the connections, or the lack of connections, between stakeholder management and the accomplishment of corporate objectives such as growth. This is particularly an approach to deal with the participants in the value-adding process. The normative type of stakeholder management is based on the fundamental notion that the purpose of an organization is to serve the stakeholders (Buono & Nichols, 1985; Donaldson & Preston, 1995).
The emphasis will be on the instrumental type of use, because I will research stakeholders’ contribution to value creation, and not the way in which the organization serves its stakeholders.
Kotter and Heskett (1992) furthermore observed that all successful companies share a stakeholder perspective. They stated that: “almost all their managers care strongly about people who have a stake in the business” (Kotter & Heskett. 1992, p. 59). According to Donaldson and Preston (1995) the alternative to the stakeholder theory, namely the shareholders perspective where the purpose is to maximize shareholder value, is morally untenable. Few stakeholders would like to contribute to the mission of maximizing shareholder value, especially if the shareholders bear no responsibility of other stakeholders’ interests (Campbell & Yeung, 1991). It may be difficult to work as a motivated team if only one groups’ interest is served.
Stakeholder theory argues that cooperation between the firm and its stakeholders drives economic welfare (Freeman 1984, 2000; Jones, 1995;). The theory studies therefore relationships between stakeholders of a firm and between the firm and its stakeholders. They also look at the impact of these relationships on stakeholder behaviour (Freeman, 1984; Freeman et al., 2010; Jones & Wicks, 1999; Phillips, 2003). It is argued that the way in which stakeholders frame their relationship with the organization is essential in affecting stakeholders’ behaviour (Mossholder, Richardson & Settoon 2011).
11
Despite the vision of a firm as a cooperative enterprise, where relationships are the unit of analysis, a detailed explanation of what drives stakeholders’ cooperation is missing (Bridoux & Stoelhorst, forthcoming). The theory has so far proposed two different approaches to stakeholder management, namely an arm’s-length ‘transactional’ approach based on economic logic and a ‘relational’ approach, based on moral premises. Stakeholder literature predicts that the ‘relational’ approach outperforms the ‘transactional’ one when it comes to value creation. To give a better explanation of stakeholder behaviour and what drives them to contribute to value creation, I would like to consider four rather than two models. These four models are derived from psychological mechanisms discussed in the relational models theory of Fiske (1991, 1992, 2004, 2012).
2.2. Relational models theory
Relational models theory is a theory developed by Fiske (1991, 1992, 2004, 2012) to explain social relationships. He argues that individuals are fundamentally sociable and that they organize their social life in terms of their relations with other people (Fiske, 1992). The theory identifies four different relational models of which people in all cultures make use, not only to explain their own behaviour but also to understand and anticipate on others’ actions. It is a theory developed in the sociology field, which still remains under explored in recent strategy literature. The models explained in the theory are mental schemata which trigger different behaviours in social interactions, because different needs and motivations are related to each model. Each model requires different rules of behaviour.
According to Fiske (1992), relational models theory posits that social features in social situations mostly can be described in just four fundamental structures. “Whatever the context and content, whatever the substance and surface form of the interaction, people’s primary frames of
12
reference in social life are the same four elementary relational models” (Fiske, 1992 p. 690). Considering these four relational models rather than the two approaches to stakeholder management of stakeholder theory could give new insights in what drives stakeholders to contribute to value creation.
The first relational model is communal sharing, which is characterized by unity and community. People treat all members as equivalent, focusing on commonalities. Little attention is been paid to the individual identity; solidarity and the identity of the group are more important (Fiske, 1992). Actors in these relationships are therefore more likely to cooperate and act in a way that benefits the collective (e.g. Dutton, Dukerich & Harquail, 1994; Dukerich, Golden & Shortell, 2002). Resources are all shared with each other, and it does not matter how much each individual contributes to it.
Authority ranking, the second model, is a linear ordering where everyone’s rank can be compared with everyone else’s. The higher up in the hierarchy, the better it is. It is thus a relationship of asymmetric differences. They are classified for example in ‘leaders’ and ‘followers’ (Fiske, 1992). These relationships are similar to the traditional form of the organization, where hierarchy is important and where people higher up in rank earn more money and control those who have a lower position.
Equality matching is a one-to-one relationship based on equality. People are primarily concerned about whether the relationship is in balance (Fiske, 1992). In a relationship referred to equality matching people may be distinct but they are equal. This type of relationship is often used together with CS, since in some cultures they appear simultaneously (Haslam & Fiske, 1999). But an important distinction between the two is the “tit-for-tat” element in EM relationships. Where people in CS relationships will contribute to the collective no matter how much others contribute,
13
people in EM relationships want something in return. They want the relationship to be in balance and will only contribute if others do so too.
Social relationships that are organized as market pricing, the last relational model, are structured like the rational numbers (Fiske, 1992). People in this type of relationship attend to ratios and rates. What matters in this type of relationship is how a person stands in proportion to others. Individuals want to engage in relationships that offer the best cost-benefit ratios (Murnighan, 1994). According to Mossholder et al. (2011), these types of relationships are more likely to be short-term and there is less involvement (in contrast with the other models). People see themselves as individuals and are independent of others. They will contribute as less as possible, to get the highest pay-off (note the difference with CS and EM relationships).
Several management scholars used Fiske’s (1992) relational models theory to explain the effects of relationships within organizations on organizational outcomes (Mossholder, Richardson & Settoon, 2011). In these scholars it is already argued that CS leads to higher levels of cooperation than the three other models (e.g., Mossholder et al., 2011; Sheppard & Sherman, 1998). An MP relationship would contribute the least to team production, because stakeholders are then focused on their personal payoffs. People in an MP relationship conduct a cost-benefit analysis to find out how to get the highest payoffs in return for the lowest contribution (Fiske 1991; 1992). In EM and AR relationships stakeholders are not only interested in their own welfare but also in the firms’ welfare (Bridoux & Stoelhorst, forthcoming). There is however no empirical evidence yet to underpin these argumentations.
In my research I would like to find out how individuals frame their relationship with the organization (either CS, EM, AR or MP) and how this influences OCB. I do acknowledge that the organization itself does not behave, humans do. However, people can ascribe some human
14
characteristics to collectives and often perceive organizations as ‘acting’ (Bridoux & Stoelhorst, forthcoming).
2.3. Employees as stakeholders and OCBO
In stakeholder theory, an important issue is the actual definition of the stakeholder (Freeman & Reed, 1982). The term is used in many different ways, and is often too general. Donaldson and Preston (1995) refer to stakeholders as: “persons or groups with legitimate interests in procedural and/or substantive aspects of corporate activity” (p. 67). They thus cover a large group, which can have different interests. Freeman and Reed (1982) on the other hand propose two definitions of a stakeholder: a wide sense and a narrow sense. With the wide sense they mean any identifiable group or person who can affect a firm’s objectives or can be affected by those objectives. The more narrow sense covers any identifiable group or person on which the organization is dependent for its survival. In my research it would be difficult to include all possible types of stakeholders in this research. What I am particularly interested in, are the relationships between stakeholders and the firm and the effect of these relationships on stakeholders behaviour towards the firm. Will they contribute to value creation for the firm or not? The focus will be on employees, since they are the key actors in creating value for the firm (Bridoux & Stoelhorst, 2015).
A particular form of value creation by employees is organizational citizenship behaviour (OCB). OCB is defined as “individual behaviour that is discretionary, not directly or explicitly recognized by the formal reward system, and that in the aggregate promotes the effective functioning of the organization” (Organ, 1988, p.4). Several studies have tested the correlation between OCB and organizational effectiveness, and the results are supportive (Karambaya, 1991; Podsakoff & MacKenzie, 1994). It is a particular form of value creation because of two reasons:
15
1) the behaviour is not per se captured in the job description and 2) it is not directly compensated by the reward system. OCB is thus performed by employees as a personal choice and the behaviour goes beyond formal role requirements (Smith, Organ & Near, 1983). According to Katz (1964), “an organization which solely depends upon its blue-prints of prescribed behaviour is a very fragile social system” (p. 132). This type of behaviour thus seems essential to make an organization function.
OCB can further be divided in two broad categories, namely OCBO and OCBI. OCBO behaviour benefits the organization in general, while OCBI behaviours immediately benefits specific individuals and indirectly contributes to the organization as a whole (Williams & Anderson, 1991). OCBI, or Altruism, seems to correspond with the forms of helping behaviour which is studied by social psychologists in experimental or non-work contexts (e.g. Berkowitz, 1972). It is strongly influenced by a ‘mood’, and is often situational (Smith, Organ & Near. 1983). OCBO on the other hand, or Generalized Compliance, is more defined as an impersonal sort of conscientiousness, where doing things ‘right and proper’ are for the sake of the organization rather than for specific persons (Smith, Organ & Near, 1983).
In my research I will focus on the OCBO behaviour, because I am interested in the contribution of employees to value creation for the firm as a whole rather than helping specific individuals.
Furthermore, I propose to investigate the mediating effect of organizational commitment on the relationship between relational models theory and OCBO. I expect that each relational model lead to different levels of organizational commitment. In addition, the literature suggests that organizational commitment is an antecedent of OCBO (Williams & Anderson, 1991). In the following paragraph I will explain the concept organizational commitment and I will explain in
16
more detail why I think it is fruitful to investigate the mediating effect of organizational commitment on the relationship.
2.4. Organizational commitment as a mediator
There are many trends in defining organizational commitment (Mowday, Steers & Porter, 1979), but the focus in this research will be on defining organizational commitment as an attitude. This attitude arises when “the identity of the person is linked to the organization” (Sheldon, 1971, p. 143) and is also described as “an individual’s attachment to, or identification with, an organization and willingness to exert additional effort to maintain organizational goals and values” (Brashear et al, 2003, p. 192). The individual thus identifies itself with the organization and wishes to maintain a member in order to help accomplishing these goals (Mowday, Steers & Porter, 1979). Meyer and Allen (1991) make a clear distinction between three types of organizational commitment and each type of commitment represents another psychological state. Commitment can reflect a) a desire (affective commitment), b) a need (continuance commitment) and c) an obligation (normative commitment). The affective type of commitment is where people are (emotionally) attached to the organization and this one could be linked to the quotes about organizational commitment above. Normative commitment on the other hand, is when employees feel it as an obligation to remain with the organization (Meyer & Allen, 1991). Marsh and Mannari (1977) state that employees with lifetime commitment are those who “consider it morally right to stay in the company, regardless of how much status enhancement or satisfaction the firm gives him over the years” (p.59). Wiener (1982) also describes this type of commitment, and states that employees feel it as right and moral to meet organizational goals and interests. The last type described by Meyer and Allen (1991), continuance commitment, is not taken into account in this
17
research. This is more referred to as a side bet. People for example remain with an organization because there are costs associated with terminating it (Meyer & Allen, 1991).
Wiener (1982) furthermore argues that organizational commitment is responsible for behaviours that: represent personal offers made in favour of the organization; do not primarily depend on punishments or reinforcements and; reflect a personal engagement to the organization. Since these are all characteristics that may be used to describe OCB, support is provided that organizational commitment is an antecedent of OCB (Williams & Anderson, 1991).
Empirical research on this subject is done by O’Reilly and Chatman (1986). They go deeper in detail by examining the basis of individual’s psychological attachment to organizations. The dimensions included compliance, identification and internalization. The identification dimension for example was shown to be a significant predictor of self-reports of OCBO behaviour.
The short review above indicates that there is both theoretical and empirical support for organizational commitment as an antecedent of OCB. More empirical research could be done however, to measure the effect of organizational commitment on OCB (Williams & Anderson, 1991).
2.5. Social dispositions as a moderator
As already mentioned in the introduction, individuals, of which firms are made up, may have different needs and preferences. The question is thus what is the best approach to get organizational members to contribute to value creation? Because of these individual differences, they may prefer different types of approaches. There is however little theory about different approaches to stakeholder management and how these approaches may affect firm performance (Harrison et al., 2010).
18
It is argued in earlier literature that by treating stakeholders fairly, they will be more likely to contribute to firm performance (Donaldson & Preston, 1995; Freeman, 1984; Harrison et al., 2010; Jones, 1995). But then how do you explain the success of two companies with similar business models, who treat their employees very differently? It seems that these successes show that employees may differ in their motives to contribute to value creation (Bridoux & Stoelhorst, 2014). These different types of motives may be caused by individuals’ social preferences. According to Eisenberg et al. (1999), individuals have different social preferences, which are relatively stable personality traits. These traits are not affected by the dynamics of specific social interactions (Bridoux & Stoelhorst, forthcoming).
Research in stakeholder theory has so far focused on one or two social dispositions, namely reciprocity and self-interest (Bridoux & Stoelhorst, 2014). Many other social dispositions have been identified, however researchers typically focus on four of them because the majority of people can be categorized into these four groups. The first disposition seeks to maximize payoffs, regardless to other payoffs, which suits with the pursuit of self-interest. People are motivated by individual rewards and see themselves as independent of others.
The second one is about maximizing the difference between their own payoffs and that of others. They are called competitors and are motivated by gaining or maintaining an advantage over others (Messick & McClintock, 1968; Van Lange, 1999).
The third social disposition is labelled reciprocators, and these people value fairness. They maximize the joint payoffs for themselves and others, as long as it is perceived as fair. In a study done by Schouten et al. (2005), reciprocators showed emotional reactions that arose because of a violation of the equality norm. This indicates that reciprocators care about equality.
19
The last group, altruists, is often studied together with reciprocators, so little is known about their motives. What is known, is that people in this group are ready to give up their own welfare in favour of others. They seek to maximize the payoffs for others, without concerning their personal payoffs (Bridoux & Stoelhorst, forthcoming).
Bridoux and Stoelhorst (forthcoming) justly ask in their research the question: “what brings individual stakeholders to represent their relationship with the firm as CS, AR, EM or MP?” (p. 9). They propose that both the social dispositions of the stakeholder, as well as the representation of the firm might affect the salience of a particular relational model. They argue that each relational model implies a different self-representation. CS implies an altruistic representation of the self, EM could be linked to reciprocators, AR to competitors while self-interested individuals could be linked to MP. This will be explained in more detail in the theoretical framework.
Since social dispositions are stable personality traits, and the representation of the firm might be different than the one of the stakeholder himself, there might evolve some friction. Actions that are not appropriate according to the relational model that actors are using are then evaluated as immoral and generate moral emotions such as guilt and shame (Rai & Fiske, 2011). When that occurs, I expect that the level of organizational commitment decreases, which has a negative impact on an employees’ performance of OCB.
2.6. Literature gap and research question
Stakeholder literature has so far focused on two approaches to stakeholder management, to explain the behaviour of stakeholders in organizations. In my research I would like to consider the four relational models of Fiske to try to give a better explanation of what drives stakeholders to contribute to value creation, and to find out if these models can also influence their decision to not
20
behave in favour of the firm. By looking at one specific mechanism, namely the level of organizational commitment of employees, I would like to go deeper and see if this mechanism holds for certain relational models. It is already argued that organizational commitment has an impact on OCB performance, but little is known about the relationship between relational models and the impact on organizational commitment.
Furthermore, I want to take the employees’ social dispositions into account to see if this also influences their behaviour.
In order to address the gaps in the literature the following elements will be researched in this thesis: 1) the relationship between the four relational models and OCBO; 2) whether organizational commitment is the mechanism that underlies this relationship and; 3) whether social dispositions moderate this mechanism.
This results in the following research questions:
1. Do the four relational models lead to different levels of organizational citizenship behaviour toward the organization?
2. Are the four relational models affecting organizational citizenship behaviour toward the organization through organizational commitment?
3. Is this mediating effect moderated by social dispositions?
Figure 1. Conceptual Model
Relational Models - Communal Sharing - Equality Matching - Authority Ranking - Market Pricing - Normative Commitment - Affective Commitment Organizational Citizenship Behaviour - Organization - Social dispositions
21 2. THEORETICAL FRAMEWORK
This paper examines if the way in which employees perceive their relationship with the firm has an impact on their performance of organizational citizenship behaviour. Moreover, it is argued that for certain relational models organizational commitment counts as a mediator in this relationship, while for other models it is not. In addition, I expect the social disposition of the employee to influence the eventual organizational citizenship behaviour performance caused by the perception of the relationship with the firm. It is argued for example, that when there’s a match between the employees’ social disposition (personal values) and the perceived relationship with the firm, this has a positive impact on the OCB performance.
1. Relational models, employees and OCB
In a relationship based on communal sharing, people are treated equally and there is a focus on commonalities. There is an emphasis on group identity, rather than on individual identity (Fiske, 1992) which results in paying more attention on welfare for the whole group than for the individual itself. Relationships in an authority ranking context are based on asymmetry between people (Fiske, 1992). People in this type of relationship often use spatial order and magnitude order to distinguish themselves. They are for example classified in ‘leaders’ and ‘followers’ and speak of superiors as ‘greater’ (Fiske, 1992). In the equality matching model, there is an emphasis on equality in the relationship. Individuals in this model care about whether the relationship is in balance (Robinson, Kraatz & Rousseau, 1994). According to Fiske (1992) “Equality matching is like using a pan balance: People know how to assemble actions on one side to equal any given weight on the other side” (p. 691). Finally, in a market pricing context, individuals seek to
22
maximize their own pay-offs at the lowest possible cost. So they engage in relationships that offer the best cost-benefit ratios (Murnighan, 1994).
In general, several management scholars used Fiske’s relational models theory to explain the effects of relationships on organizational outcomes (e.g. Mossholder et al., 2011). They have already argued that stakeholders who frame their relationship with the firm as CS will contribute more to value creation, followed by respectively EM, AR and MP (Bridoux & Stoelhorst, forthcoming). Actors in a CS relationship identify themselves with a collective and are more likely to cooperate and behave in a way that benefits the collective (e.g. Dutton, Dukerich & Harquail, 1994; Dukerich, Golden & Shortell, 2002). They see themselves as a member of a community (the firm) and remain loyal even though it is not individually rewarding. Stakeholders in a CS relationship will do whatever is necessary to get the work done because it is perceived as “our work” (Fiske, 1991 p. 66). Consequently, I expect that the higher employees score on a relationship framed as CS, the higher their contribution will be to OCBO. OCBO, as opposed to OCBI, is the contribution of employees to value creation for the firm as a whole rather than helping specific individuals. It is not per se part of the job description, difficult to measure and not directly rewarded by the system. This argumentation leads to the first hypothesis:
H1a: The higher employees score on a relationship framed as CS, the higher their contribution will be to OCBO.
23
In sharp contrast, stakeholders who frame their relationship with the firm as MP are motivated by self-interest and will only seek to maximize their own pay-offs. They see themselves as individuals and independent of others, and will try to get the highest pay-off for the lowest contribution (Fiske, 1991; 1992). They are driven by monetary incentives, and will not contribute to value creation if they are not personally rewarded or punished for the fact that they do not contribute (Ellemers et al., 2004; Haslam & Ellemers, 2005). For this reason, I expect that the higher the score is on an MP relationship, the lower their contribution will be to OCBO. This leads to the following hypothesis:
H1b: The higher employees score on a relationship framed as MP, the lower their contribution will be to OCBO.
Employees in an EM and AR relationship are not only interested in their own welfare but also in that of the firm. Other than in MP, the level of self-representation is interpersonal, which means that that they are willing to bear some personal costs to benefit the firm (Bridoux & Stoelhorst, forthcoming). The level of self-sacrifice however is lower than that of a CS relationship, because employees do care about personal pay-offs. Based on the argumentation above, I expect that employees framing their relationship as EM and AR both are willing to contribute to OCBO. This leads to the two following hypotheses:
H1c: The higher employees score on a relationship framed as EM, the higher their contribution will be to OCBO.
24
H1d: The higher employees score on a relationship framed as AR, the higher their contribution will be to OCBO.
I would now like to compare the relational models, to see which relational model is most positively related to OCBO performance. This means I also have to change the type of the independent variable: I will look at the dominant models and how this influences OCBO. Since employees in relationships framed as CS prefer the collective above their own pay-offs, and contribute to the collective even though it is not individually rewarding, I expect this type of relationship to contribute more to OCBO than the other ones. Relationships framed as EM and AR are interpersonal, which means that they are willing to sacrifice in favour of the collective, but they also care about their personal pay-offs. People in MP relationships are self-interested and will not contribute if they are not individually rewarded. This leads to the following hypotheses:
H1e: Relationships with CS as the dominant model are more positively related to OCBO than AR or EM relationships.
H1f: Relationships with AR or EM as the dominant model are more positively related to OCBO than MP relationships.
2. Relational models, organizational commitment and OCBO
I would now like to go a little bit deeper by looking at one mechanism in the relationship between the way in which employees frame their relationship with the firm and OCBO, namely organizational commitment. Like already mentioned earlier, there is both theoretical and empirical research provided that organizational commitment is an antecedent of OCB (Williams &
25
Anderson, 1991). Considering the concepts used to characterize affective organizational commitment such as psychological attachment, identification with the firm and personal sacrifice made in favour of the organization, I expect affective organizational commitment to function as a mediator in relationships framed as CS. In the CS relationship individuals see themselves as belonging to one community, and they have a strong identification with it. Furthermore, I expect individuals framing their relationship as AR to also feel committed to the organization, but feel this as an obligation to remain in the organization. This is referred to as the normative organizational commitment. People in an AR relationship emulate, defer to, or obey their superiors (Fiske, 1992). Simple commands of superiors can be very powerful, and the highest ranking person decides for all subordinates, or delegates decision authority when needed. I expect employees in this type of relationship to remain at the organization because they feel they ought to do so.
Individuals framing the relationship as EM care about balanced reciprocity; it is a one-for-one exchange (Fiske, 1992). They will only contribute to the collective if they expect all other employees to do the same. I expect them to feel affective committed to the organization as long as they feel treated fairly. According to Meyer and Allen (1991), normative commitment can develop not only through remaining loyal (as expected in AR relationships) but also “through the receipt of benefits that create within the employee a sense of obligation to reciprocate” (p.539). I therefore expect that employees who frame their relationship as EM score high on both the affective as the normative type of organizational commitment. Since individuals who frame their relationship as MP care about their personal pay-offs and are not primarily concerned with the collective, I do not expect organizational commitment to count as a mediator in the relationship.
26
H2a: Affective organizational commitment mediates more strongly the relationship between relational models and OCBO in CS relationships than it does in AR, EM and MP.
H2b: Normative organizational commitment mediates more strongly the relationship between relational models and OCBO in AR relationships than it does in CS, EM and MP.
H2c: Both affective and normative organizational commitment mediate more strongly the relationship between relational models and OCBO in EM relationships than in MP.
3. Relational models, social dispositions and the influence on organizational commitment According to Bridoux and Stoelhorst (forthcoming), the salience of a particular relational model can be influenced by the firm, but is also affected by social dispositions, since each relational model implies a different representation of the self. Each individual may have different needs and preferences, which are relatively stable personality traits (Eisenberg et al., 1999). The four types of social dispositions where researchers typically focus on could be linked to the different relational models of Fiske (1991): reciprocators, who value fairness and equality, can be linked to the EM model; self-regarding individuals, who are motivated by individual rewards and see themselves as independent of others, can be linked to the MP model; competitors, who are motivated by gaining and maintaining an advantage over others, are expected to link with the AR model (if in the superior role) and finally; altruists could be linked to the CS model, although little is known about the motives of altruists since this group is often studied together with reciprocators. Bridoux and Stoelhorst’s (forthcoming) argumentation for the latter link is based on the readiness of altruists to give up their own welfare in favour of others. This corresponds with Fiske’s (1992) statement that in CS relationships a “person does not need to give something in order to get something in return” (p. 693). People in this relationship seek to maximize the payoffs for others,
27
without concerning their personal payoffs. Based on these linkages I assume that reciprocators frame their relationship with the firm as EM, self-regarding individuals as MP, competitors (superiors) as AR and altruists as CS.
In my research I will measure the employees’ personality traits with Schwartz ten basic values theory (1994). These values can be summarized along two dimensions: self-enhancement vs. self-transcendence and openness to change vs. conservation. This research will focus on the first dimension, where the power and achievement values are opposed to universalism and benevolence values. Self-enhancement is often referred to as self-regarding, which emphasizes the pursuing of self-interest through controlling people, power values and achievement values. Self-transcendence on the other hand is labelled as other-regarding (or reciprocity), which emphasizes serving others’ interests and care for those who are close (benevolence values) and showing acceptance, tolerance and concern for all people regardless of group memberships, which are the universalism values (Gomà-i-Freixanet, 1995). Compared with the four social dispositions that are mentioned above; competitors (AR relationships) are expected to score high on power values, reciprocators (EM relationships) on benevolence values and altruists (CS relationships) on universalism values.
If the social disposition of the employee fits with the relational model in which the employee shapes its relationship with the firm, employees will perceive the relationship as harmonious because they understand the firm’s behaviour and have the same motivations and expectations (Boer, Berends & Van Baalen, 2011; Connelley & Folger, 2004; Giessner & Van Quaquebeke, 2010). Moreover, Meyer, Allen and Smith (1993) argue that when the expectations of the employees are consistent with their experiences, they ought to develop a stronger affective attachment to the organization than when it is less satisfying. However, when the firm behaves in
28
a way that does not fit the relational model, and the employee feels that this is done by purpose, they perceive this mismatch as a transgression which will bring up negative emotions (Fiske, 1991). Summarizing the above, I expect that the commitment to the organization is higher when there is a match between the relational model and the personal values of the employees. This leads to the last hypotheses:
H3a: The higher employees’ score on universalism, the stronger the positive impact of a CS relationship on affective organizational commitment.
H3b: The higher employees’ score on benevolence, the stronger the positive impact of an EM relationship on affective organizational commitment.
H3c: The higher employees’ score on power, the stronger the positive impact of an AR relationship on normative organizational commitment,
4. METHODOLOGY
In this chapter the research method is explained. Firstly, the research instrument is presented which is followed by a description of the sample and the data collection method. After that follows a detailed explanation of the variables and the way in which they are measured. See the appendix for the complete questionnaire in Dutch.
4.1. Research Design
The technique used in this paper to test hypotheses and to answer the research question was an electronic survey, which was conducted among employees of diverse firms. The survey was used to collect data on the topics described in the theoretical framework. The survey furthermore
29
contained several other variables, since I have been collecting data with two other students. These other variables were not used in the present research. We collaborated to get as many (reliable) respondents as possible. The research is mostly explanatory; we would like to measure first the direct effect of the four relational models on OCBO, then the indirect effect of commitment on this relationship and lastly the conditional effect of the personal values to understand the conditions under which an effect operates. A survey is a good method to collect this type of information (Saunders, Lewis & Thornhill, 2009). Moreover, I think the evolution and change of these topics and relations through time are at least as important, but due to time constraints we were not able to execute a longitudinal study. Instead we had to choose for a cross-sectional study.
Using the online survey as a research method allows us to collect large amounts of data. Moreover, the data can easily be analyzed and interpreted (Saunders et al., 2009). It saves costs and people filling in the questionnaire could choose by themselves to fill it in or not, which will make the outcomes more reliable (Lefever, Dal, & Matthíasdóttir, 2007). Other advantages of online surveys are the lower transmitting time and the reduction of processing costs. This is because the raw data can be automatically transformed into computer readable form (Kiesler & Sproull, 1986).
There are however several issues with collecting data online, such as how the survey affects responses of the people who use it. Other issues are the loss of participants who not use Internet and the possibility that questions are misunderstood or interpreted in different ways (Kaplowitz, Hadlock & Levine, 2004; Fricker & Schonlau, 2002; Groves, 1989). Another issue with collecting data through a survey, either electronic or on paper, is that the participants are asked to think about past behavior, while this information can be more accurately collected through for example observation (Schwartz, 1999). Moreover, the questions asked in a survey bias the answers and
30
judgments of respondents (Schwartz, 1999). To make a research of higher quality and more representative, multiple research methods could be combined, but this also involves higher costs (Andrews, Nonnecke & Preece, 2003) and it takes more time. An important factor in determining which research method to use for this thesis is time. Since electronic surveys can be distributed very fast, and the response cycles are short (Andrews et al.; Yun & Trumbo, 2000) I decided to only make use of this research method to collect data.
We created the survey questionnaire in Qualtrics, and asked respondents through e-mails and social media platforms like Facebook and LinkedIn to fill in the survey. I have sent e-mails to my contacts of which I knew they are working. We wanted to gather information coming from diverse organizations because we hope to measure different relational models. However, there are respondents working for the same company who participated at the survey since I asked some people to send it to a few colleagues. Still I think the sample is diverse and represents the population.
Considering the survey design, we decided to ask mostly closed-ended questions. Too many open-ended questions could namely increase attrition rates (Crawford, Couper & Lamais, 2001). The questions of the measurement instruments were originally in English, but they have been translated in Dutch, because we think employees are more willing to fill in the survey when it is in their native language. The fact that we only collected data from Dutch employees immediately points out a limitation of our research: we were not able to measure international differences.
31 4.2. Sample
The sample consisted of Dutch employees that worked at least 24 hours per week for their organization. To measure hopefully different relational models it was necessary to approach employees from diverse companies. To collect data we have made use of a convenience sampling technique which means that the subjects are selected that are both easily accessible and willing to participate in the study (Teddlie & Yu, 2007). The technique is widely used and is fast, easy and usually the least troublesome and expensive (Lunsford & Lunsford, 1995). A disadvantage of this is that the population which is selected does not represent all the members of the population since all members do not have the same chance of being selected. It can thus be assumed that the sample does not represent the target (Lunsford & Lunsford, 1995). The random sampling technique, as opposed to convenience sampling, represents a better part of the population, but is very difficult to achieve. It takes more time and is often combined with high costs (Lunsford & Lunsford, 1995). We therefore chose to use a sampling technique which is not random.
To enhance the validity of the research, we wanted to have at least 250 respondents. There are always individuals who for example fill in just a few questions, and these respondents have to be taken out afterwards. When we closed our survey, there were 300 individuals who filled in our survey. Since the people who worked less than 24 hours for a company were not allowed to finish the survey, I deleted those. After that, the respondents who only filled in the control questions were eliminated. Furthermore, I deleted those who only filled in the questions measuring OCBO, those who only filled in the OCBO and OC items and the respondents that only filled in the RM items. There are now 186 respondents left.
32 4.3. Measures
This chapter explains the different measurement instruments for the variables that were tested. Wherever possible, I have made use of scales that have been published in earlier research. These research instruments have already been validated, and will be discussed in more detail below. For the independent variables we had to develop a new scale, which was based on the Organizational Culture Assessment Instrument of Cameron and Quinn (1999). We had to do a pretest to find out if it was a valid scale to measure the independent variables. The results of this pretest are explained below.
4.3.1. Independent variables
The new scale we developed to measure the relational model that employees use to frame their relationship with the firm was based on the Organizational Culture Assessment Instrument (OCAI) of Cameron and Quinn (1999). This original instrument is an accurate assessment of organizational culture, and has been used in many scholarly research and a variety of industry sectors (Cameron & Quinn, 2011). It is designed to help identifying the current culture of an organization. The instrument consists of six items, which describe some of the fundamental manifestations of organizational culture (Cameron & Quinn, 2011). Each item has four alternatives. Organizational members are asked to divide 100 points among these four alternatives (A, B, C and D), based on which is most similar to their organization. If for example A is very similar while B & C are somewhat similar and D is not similar at all, you might give 55 points to A, 20 to both B & C and 5 points to D (Cameron & Quinn, 2011). All four of the organizational cultures represent a letter in each question. In the end, the total score of all questions per letter relate to one of the organizational cultures (hierarchy, market, clan or adhocracy).
33
We wanted to use the instrument, not to measure the organizational cultures, but to capture the relational models. The CS relational model is argued to have a lot in common with the “clan” culture (Bridoux & Stoelhorst, forthcoming), the AR model with the hierarchical culture and the market culture with the MP model. The EM model was difficult to link to a certain culture, which could cause trouble in the measurement. To find out if the OCAI could be turned into a valid measurement of the relational model an employee uses to frame its relationship with the firm, a pretest has been done. Two strategy scholars familiar with relational models which adapted the OCAI to fit the relational models. The dimension “strategic emphasis” of the OCAI was replaced by relationships with external stakeholders, because it was difficult to determine which model relates to which strategic emphasis. Unfortunately the results of the new dimension were not satisfactorily either. The items were pretested on Mechanical Turk, which is a platform established by Amazon in 2005. Its original goal was to “crowd-source” labor intensive tasks, but it also provides the possibility to distribute research. The items were pretested between the 16th and 20th of April 2015.
In the pretest, a first control question was inserted to eliminate respondents who a) did not work for an organization (students or unemployed while looking for a job); b) were self-employed or; c) worked without being paid. The respondents left got a few more control questions before they came at the point where they had to rate their organization. The survey was used to measure 9 dimensions, and each dimension was described through four views (each description represents a different relational model). The respondents were asked to rate how well their organization fits the description. Note that this is another way of computing the models than used in the OCAI, where respondents had to divide 100 points. The 7-point answer scale ranges from “Completely false description” to “Completely true description”. A number of changes were made to the items
34
that exhibited the poorest fit, based on exploratory factor analyses, confirmatory factor analyses and computations of internal reliability. Eventually we decided to eliminate the following dimension: “Relationship with external stakeholders”, because results of the pretest has shown that this dimension did not work satisfactorily. The number of usable responses was 106 (21 were eliminated after the control questions).
After the results of the pretest the questions were translated in Dutch, since all respondents to the survey had Dutch as their native language. We did a parallel translation, which means that more people translated the same items so that they could be compared. The translations were sent to our supervisor (dr. F. Bridoux) and dr. J.W. Stoelhorst and they provided us with feedback. Both the questionnaires are included in the Appendix.
Each relational model was measured on 8 dimensions, namely dominant characteristics, leadership, management, decision making, management of employees, the glue that holds the organization, success criteria and justice principles. The degree of how similar the organization was to a specific relational model was measured using a 6-point Likert scale ranging from 1 “Completely unlike my organization” to 6 “Completely like my organization”. The items were presented randomly in the survey.
4.3.2. Dependent variable
Firstly I wanted to discover if there was a direct effect of the relational models on the dependent variable OCBO. OCBO was measured using the scale developed by Lee and Allen (2002). These authors used items which were selected from a pool created by earlier OCB scales (Lee & Allen, 2002). There were 8 OCBO items selected. These items are measured on a 7-point Likert scale (1=never, 7=always), to measure how often employees engage in this behavior.
35 4.3.2. Mediating variables
The next step was to understand the mechanisms by which the relational models causally influenced the dependent variable. These mechanisms were the mediating variables in my research: affective and normative organizational commitment. They belong to the three-component model for organizational commitment and refer to the emotional attachment of an employee to the organization (Meyer & Allen, 1991). All the three types of commitment consisted of 6 items, which were measured on a 7-point Likert scale (1=fully disagree, 7=fully agree), adapted from Meyer, Allen and Smith (1993).
4.3.3. Moderating variable
The third test of this research was done to measure the conditional effect. So the main question was: when are the relational models and OCBO related? We wondered if the effect was stronger when there was a match between the employees’ characteristics and the way in which the relationship with the firm was framed. The last variable measures therefore the respondents’ personal values.
Schwartz distinguishes ten basic values, which include all the core values that are recognized in cultures around the world (Schwartz, 2012). To measure a person’s values he developed the portrait value questionnaire (PVQ) which consists of 40 items. Each item describes the goal or aspiration of a hypothetical person and points implicitly to the importance of a value. The respondents are asked to rate how much each person looks or does not look like you. It is a 6-point Likert scale ranging from 1 (“not like me at all”) to 6 (“very much like me”). In our research we only measured four of the ten basic values, namely universalism, benevolence, power and
36
achievement, so we eliminated the items that measured the other value domains. This resulted in a questionnaire of 17 items. We did not change the measurement scale.
4.3.4. Control variables
The research controlled for the respondents’ background and personal information; gender, age, education and profession. Moreover, it controlled for the hours respondents spend working at the office (instead of working from home), the organizational size and the years the employee works for that organization. Respondents who worked less than 24 hours a week were immediately eliminated from the survey after the control questions.
We included gender as a control question, since women are argued to be more socially-oriented and cooperative, while men are more individually socially-oriented. According to Van Lange et al. (1997) the self-interested behavior is more common among men than among women. In addition, they argue that value orientation may change through time, so age can also be an important factor in the analysis. It is argued for example that especially young people tend to behave more self-interested while when getting older, individuals tend to behave more as reciprocators (Van Lange et al., 1997). Education is also important to take in consideration when analyzing the data. Already a positive relationship have been found between the level of education and OCB (Ng & Feldman 2009). Profession and the hierarchical position within the organization also might influence the outcomes of the research, especially when the level of normative commitment is measured. This type of commitment might not count for respondents managing their own company. It is namely argued that an employee could feel committed to the organization in a normative way, because he/she has a moral obligation to reflect loyalty and duty (Wiener, 1982). I do not expect respondents who manage their own company, or work for themselves without employees feel normative committed. Unfortunately it is difficult to measure the
37
profession and hierarchical level, since these are both captured in one question. The question in the survey was “What is your profession?”. Respondents could choose between several options, but there might also have been overlap in the answers. There were for example options like “Manager” and “Company owner”, but also “Ambachtslied”. A respondent who filled in “Ambachtslied” could also have been a company owner. It was therefore difficult to make a distinction between the professions and hierarchical levels. I still wanted to take these results in my analyses though, therefore I categorized them. This will be explained in the next chapter.
Next to gender, age, education and profession, the hours an employee physically spends at the office could influence his/her commitment towards the organization, and also the relational model used to frame the relationship with the firm. Unfortunately it is difficult to insert this as a control variable, since the amount of hours working in the office should be compared to the total working hours of the respondent. This comparison can not be done because the respondents only had two options: 1) work less than 24 hours per week and 2) work more than 24 hours per week. The people who choose the first option were immediately eliminated from the questionnaire. This means all the reliable respondents work > 24 hours per week, but we did not know the exact amount. Measuring how many hours the respondents physically spent at the office in comparison with their total working hours could have been done by asking a percentage instead of a number. Based on the argumentation above I decided to eliminate the “Hours spent at work” as a control variable. Another control variable that might influence the organizational commitment is the organizational tenure. It is namely argued in earlier literature that it has a positive relationship with OCB (Ng & Feldman, 2010). Last but not least I have included size of the organization as a control variable. According to Stevens, Beyer and Trice (1978), the size may have a negative influence on organizational commitment. When looking at managers however, a large organization may require
38
bigger investments on the field of coordination, innovative behavior and control (Baldridge & Burnham (1975), which increases commitment (Stevens et al., 1978). Moreover, I think that the organizational size might influence the way in which employees frame their relationship with the firm.
5. RESULTS
This section contains the results of the research. First, a reliability analysis has been done for the variables that consisted of multiple items. After that follows a chapter with the descriptive statistics to give an overview of the data and to explain the different categories for the control variables. Subsequently the correlation matrix is shown, which is followed by a normality test for OCBO, the averages of the relational models, organizational commitment and the age. In the chapter that follows the hypotheses will be tested by doing a linear regression analysis, a test for mediation and a test for moderation.
5.1. Reliability Analysis
Since I have made use of variables that contain multiple items, both a Cronbach’s alpha test and exploratory factor analyses had to be done to find out if these items could measure one topic. The results of these tests determine if the scales show high internal reliability and validity. If this is the case, the items can be combined to measure one variable. A Cronbach’s alpha outcome above .7 means it is reliable (Pallant, 2007). The outcome of the KMO & Bartlet test (the exploratory factor analysis) has to be >.6.
First the Cronbach’s alpha of the relational models was measured, which gave the following outcomes: CS .8; AR .79; EM .82 and MP .87. They all appear to be reliable. After these
39
tests exploratory factor analyses have been done for each of the relational models apart with the items. The outcomes of the KMO Bartlet test were as follows: CS .85; AR .84; EM .87 and MP .91. It appears to be valid if the outcome is >.5. This is the case for all the relational models. But I also wanted to assess the variance in the variables, to see if there was multicollinearity between the four relational models. Therefore I performed a principal components analysis. The analysis did not clearly show four different components (see Table A in Appendix A). According to the Guttman-Kaiser criterion, variables with an eigenvalue of at least 1 can be used as factors (Yeomans & Golder, 1982). When looking at Table A1 in Appendix A, this would mean that there are 7 factors. The scree plot however shows a clear inflection point, which would lead to three components.
After that I ran a principal component analysis with four fixed factors, which represents the amount of the relational models. The rotated components matrix showed that the four factors explain 51.29% of variance. But the table also shows that the questions measuring CM and EM both load on the first factor. There can be multiple explanations for this, for example that the two models are not really that different from each other, but this will be explained in more detail in the discussion. I decided not to remove items but to keep the four relational models as they were.
Subsequently, a Cronbach’s alpha has been done to measure the reliability of the OCBO items. The outcome was .89, which means there is internal reliability.
A Cronbach’s alpha and a principal component analysis have been done for the mediating variables, namely affective and normative organizational commitment. The Cronbach’s alpha of affective commitment was .9 and normative commitment .85. The results of the principal component analysis are shown in Table A2 and A3 n Appendix A. It shows indeed two factors. Table A3 however also shows that the item “De werkgever verdient mijn loyaliteit” correlates with
40
both the types of organizational commitment, and even more belongs to the factor for affective commitment (.489) than with the one it was originally used for (normative commitment; .472). I decided to let the item in and use it to measure normative commitment.
Lastly the reliability and the validity of the personal value items were measured. The Cronbachs alpha’s were as follows: UNI .79; BEN .55 and POW .76. The Cronbach’s alpha of benevolence was too low and by deleting one item (“Het vergeven van mensen die […]) the outcome would have been .71 which would have make it reliable. However, I rather keep all the items in, since there are only four items in total to measure benevolence. The principal component analysis gave very different results that are shown in Table A4 of Appendix A. There are not really four clear factors, and a lot of scores are negative. It will become difficult to do further analyses with the personal values but I did not want to create other factors based on the scores in the factor analyses. The results of this analysis are too complicated to create other factors.
5.2. Descriptive Statistics
This research contains categorical as well as continuous variables. Almost all the control variables are categorical, except the one for age, which is continuous. The categorical control variables –
gender, education, profession, years working for the organization and organizational size – had
to be recoded into dummy variables. Like mentioned in the former chapter, the variable that measures profession also contains hierarchical levels. Respondents who filled in for example “ambachtslied” might also be owner of their own company. Because I am more interested in the hierarchical level than in the industry/sector, I have clustered the professions into three categories: 1) director; 2) management function and 3) employee. Appendix A Table B1 shows which professions belong to which category. Table B2 shows the 25 respondents who filled in the option
