Value creation in the firm : the effect of relational models, social dispositions and self-control on OCBO

Value Creation in the Firm: The Effect of Relational

Models, Social Dispositions and Self-Control on OCBO

Master Thesis

MSc. in Business Administration – Strategy Track Amsterdam Business School, University of Amsterdam

Student: Kevin Segeren Student no.: 10871721

Statement of Originality

This document is written by Student Kevin Segeren 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.

Acknowledgements

First, I would like to thank my thesis supervisor, Flore Bridoux, for all her help, useful suggestions and valuable feedback throughout the process of writing this thesis. Secondly, I would like to thank my family, my friends and my girlfriend for their support and encouragement. Finally, and above all, I would like to thank my parents for everything that they have done for me in my life.

Table of contents

2.1 Value creation of a public good nature and stakeholder-firm relationships ... 10

2.2 Organizational citizenship behavior by employees ... 11

2.3 Relational models ... 12 Communal sharing ... 13 Authority ranking ... 14 Equality matching ... 15 Market pricing ... 15 2.4 Social dispositions ... 16 Self-regarding ... 17 Reciprocator ... 18 Competitor ... 18 Altruist ... 19 2.5 Self-control ... 20 3. Research questions ... 22 4. Theoretical Framework ... 24

4.1 Relational models and OCBO ... 24

4.2 Social dispositions, relational models and OCBO ... 26

Communal sharing ... 27

Equality matching ... 28

Authority ranking ... 29

Market Pricing ... 30

4.3 Self-control and OCBO ... 31

5. Methodology ... 34

5.1 Research design ... 34

5.2 Research sample and data collection ... 35

5.3 Operationalization of constructs and variables ... 37

Dependent variable - OCBO ... 37

Independent variable – Relational Models ... 37

Moderator – Social Disposition ... 38

Moderating moderator – Self-Control ... 39

Control variables... 40

5.4 Merging the questionnaires ... 41

6. Results ... 42

6.1 Reliability analysis ... 42

Moderator – Social Disposition ... 45

Moderating moderator – Self-Control ... 48

Control variables... 48 6.2 Descriptive statistics ... 49 6.3 Correlations analysis ... 52 6.4 Normality analysis ... 54 OCBO ... 54 Relational models ... 54 Social dispositions ... 55 Self-Control ... 55 6.5 Analyses ... 55

Hypothesis 1: Relational models and OCBO ... 55

Additional analyses relational models ... 58

Hypothesis 2: Social dispositions, relational models and OCBO ... 63

Results: Self-control ... 70

Additional analysis: Direct effect of self-control ... 71

7. Discussion and conclusion ... 73

7.1 Main findings and contributions to theory ... 73

Main findings about Relational Models ... 73

Additional findings ... 77

7.2 Implications for management practice ... 78

7.3 Limitations and recommendations for future research ... 79

8. Conclusion ... 83

9. References ... 84

Appendix A – Graphs and Tables ... 96

Appendix B – Questionnaires ... 98

English questionnaire ... 98

Abstract

Getting employees to perform organizational citizenship behavior directed at the organization (OCBO, which is extra-role behavior from which the organization benefits) is an important managerial challenge, as it is good for the organization but individual employees are not (directly) rewarded for it. Previous theory papers have proposed that some relational models (schemata that people use to organize all social interaction) are better suited for value creation in organizations than others. Using a survey, I empirically test how four different relational models affect OCBO. Furthermore, I investigate the effect of social dispositions (individual preferences for the allocation of resources between

themselves and others) and self-control (the ability to control your own emotions and behavior) on this relationship.

Results indicate that some relational models are indeed better suited for OCBO than others, and that individuals whose social disposition more closely resembles the relational model they use to frame their relationship with their employer, generally perform more OCBO. Self-control did not provide any significant results, but a number of suggestions for future research are provided.

The results contribute to a better theoretical understanding of the conditions under which OCBO is performed. Furthermore, a number of ways are suggested how management practice can use these results to positively influence the amount of OCBO that employees perform. In light of the results and limitations of this research, several suggestions for future research are provided which should shed further light on the relationships between relational models, social dispositions, self-control and OCBO.

1. Introduction

Value creation in organizations often takes the form of a public good dilemma (Cabrera & Cabrera, 2002), in which individuals can benefit from the outcome of joint value creation, regardless of whether they contributed to it (Dawes, 1980). For example, organizational citizenship behavior (OCB) is good for value creation in firms, but is costly for the individual employee as it requires time and effort and is not (directly) rewarded (Organ, 1997). Getting employees to perform OCB despite its public good nature is thus an important managerial challenge. In this thesis, I will focus on organizational citizenship behavior at the organization (OCBO).

The stakeholder literature suggests that relationships are crucial to get stakeholders to create value for organizations (Freeman, 2000; Freeman & Phillips, 2002), and employees are the stakeholders who are ultimately responsible for creating value within the firm.

Relationships exist between individuals, but employees also perceive their connection with the firm as a relationship (Mossholder, Richardson, & Settoon, 2011) because they tend to ascribe human characteristics to collectives such as firms (Morrison & Robinson, 1997; Sluss & Ashforth, 2008).

These relationships are framed using four relational models according to Fiske's (1992) relational models theory: communal sharing (in which everybody is treated as the same) equality matching (in which balance and equal contributions are important), authority ranking (in which hierarchical rank is important) and market pricing (in which the focus is on

individual payoff). In practice, a mix of these four relational models is adopted, in which one or more models are dominant (Bridoux & Stoelhorst, forthcoming). It has been argued that firms can also influence which relational models individuals use to frame their relationship with the firm, for example though HR policies (Mossholder et al., 2011) and that some of

& Stoelhorst, forthcoming). But, these theoretical arguments have not been tested yet and little attention has been paid so far to how the different relational models affect OCBO.

Furthermore, it is likely that stakeholders have an individual preference toward one of the relational models because of their social disposition, i.e., their preference regarding the allocation of resources between themselves and others (Fehr & Fischbacher, 2002). Four social dispositions can be distinguished; self-regarding individuals (who care about their own payoff), reciprocators (who care about equality), competitors (who care about maximizing their own payoff compared to others) and altruists (who care about others‟ payoff) (Bridoux & Stoelhorst, forthcoming). Individual social dispositions may to a greater or lesser extent match the adopted relational models (Bridoux & Stoelhorst, forthcoming). When the individual‟s social disposition matches the dominant relational model used to frame the relationship with the firm, an employee should perceive the relationship as harmonious. If, however, there is a mismatch, this may lead to negative emotions and possible deterioration of the relationship between the employee and the firm (Boer, Berends, & van Baalen, 2011). Although this indicates the importance of a good match between individual social disposition and relational models, little is known about how a match or mismatch between relational models and social dispositions affects OCBO.

Finally, individuals may vary in their reactions to a mismatch between the adopted relational model and social disposition based on the level of self-control. The self-control literature shows that individuals high in self-control, which is “the exertion of control over the self by the self” (Muraven & Baumeister, 2000, p. 247), are better able to perform behavior that is beneficial in the long run over behavior that provides short term benefits (Muraven & Baumeister, 2000). Self-control is a trait that has a positive influence on performance, interpersonal relationships, as well as on many other areas of life (Tangney, Baumeister, &

particularly interesting to study the effect of self-control when there is a mismatch between a stakeholders‟ social disposition and the firms‟ dominant relational.

Using a survey, this thesis contributes to theory in three main ways. First, it shows how the different relational models affect OCBO. Secondly, I pay attention to the moderating effect that social dispositions have on the relationship between relational models and OCBO. Finally, I look at the role of self-control, and particularly the role that self-control plays when the adopted relational model does not match an individual‟s social disposition.

Besides contributing to theory, this thesis is valuable for management practice. First, it provides insight in how relational models (which can be influenced by the firm) affect OCBO. Secondly, it helps to understand how social disposition and self-control (both traits) can be taken into account in the value creation process.

In the remainder of this thesis, I will provide a review of the relevant literature (2),

followed by the research gap and the research questions that I wish to address (3). After that I will present the theoretical framework, leading to a number of hypotheses (4), followed by an explanation of the methodology (5), the results of the analyses (6), a discussion (7) and, finally, a conclusion (8).

2. Literature review

In this part, I will discuss the main findings of the existing literature related to the topic described in the introduction. I will start with the general challenges of value creation in public good settings, followed by an overview of the relational models and social dispositions literature. After that, I will conclude by discussing self-control.

2.1 Value creation of a public good nature and stakeholder-firm relationships

Knowledge work is becoming more and more important for economic welfare (Powell & Snellman, 2004). In knowledge work, the primary source of production is knowledge, which resides within individuals and is dispersed among different employees (Cabrera & Cabrera, 2002). Because this knowledge is used to come to a joint outcome, it is difficult to identify each stakeholder‟s individual contribution (Alchian & Demsetz, 1972). The central task of the firm is to coordinate activities, by getting stakeholders to jointly create value. The challenge here is that the goals of individual stakeholders may not be aligned with other stakeholders‟ and the firms‟ goals (Grant, 1996).

As a consequence, stakeholders involved in joint value creation often face a public good dilemma: contributing to value creation is best for the firm, but it is costly for them because it requires time and effort (Cabrera & Cabrera, 2002). Because it is difficult to identify each stakeholder‟s individual contribution to the public good, stakeholders who aim to maximize their own payoff have an incentive not to contribute (Alchian & Demsetz, 1972). By doing so, stakeholders can benefit from the public good (the outcome of production) without

contributing to it. On the other hand, if too many individuals decide not to contribute, there will not be a public good to benefit from (van Dijk & Wilke, 1995).

(Simon, 1955). According to this view, everybody would try to maximize his or her own payoff by not contributing to the public good. Yet, in reality, stakeholders sometimes do.

Following the logic of stakeholder theorists, contributing to the public good occurs because humans are not (always) self-interested (Bosse, Phillips, & Harrison, 2009). They argue that a more complex view of human behavior should be taken into account when looking at these public good situations (Bosse et al., 2009). Stakeholder theory focuses on cooperation between the firm and stakeholders (those who can influence and/or are influenced by a firm) as the driver of value creation (Freeman, 2000).

Of particular interest here is instrumental stakeholder theory, which places an emphasis on the development of relationships between the firm and its stakeholders (Jones, 1995). It is argued that the firm will gain competitive advantage if it is able to develop relationships based on trust and cooperation with its stakeholders (Donaldson & Preston, 1995; Hillman & Keim, 2001; Jones, 1995).

2.2 Organizational citizenship behavior by employees

As discussed, the development of relationships is important to get stakeholders to jointly create value according to stakeholder theory. While a firm has many stakeholders to consider, primary stakeholders are essential for the survival of the organization (Clarkson, 1995). Employees are an important group of primary stakeholders (Berman, Wicks, Kotha, & Jones, 1999; Hillman & Keim, 2001), and they are responsible for creating value within the

organization. That is why here, I will focus on employees when discussing stakeholders.

Getting employees to perform behavior that contributes to achieving organizational goals has long been identified as important (e.g., Katz ,1964). Katz (1964) argues that firms can only survive if employees perform beyond the job behavior, next to carrying out the

defined “organizational citizenship behavior” (OCB) by Organ (1988) More precisely, OCB is behavior that is not formally rewarded, but that does contribute to value creation in the firm (Organ, 1997).

Previous research discusses a wide range of benefits that OCB has for the organization. Some of the advantages are more effective employees (Yen & Niehoff, 2004), lower employee turnover (Chen, Hui, & Sego, 1998; N. P. Podsakoff, Whiting, Podsakoff, & Blume, 2009) and lower employee absenteeism (N. P. Podsakoff et al., 2009). On a more general level, different studies have indicated that OCB positively contributes to

organizational performance (P. M. Podsakoff, MacKenzie, Paine, & Bachrach, 2000).

The literature distinguishes between two types of OCB. The first type of OCB is directed at specific individuals within the organization (OCBI) and the second type of OCB is directed at the entire organization, and is more about doing what is right for the system (OCBO) (Ann, Organ, & Near, 1983). Here, I will focus on OCBO, because OCBI can originate from

individual relationships (for instance: helping a coworker because he is a personal friend), whereas OCBO is directed at the entire organization.

OCBO is sensitive for the public good dilemma because it does contribute to value creation in the firm but it is costly for individual employees to perform OCBO because it is not (formally) rewarded. Getting employees to perform OCBO is thus an important

managerial challenge.

2.3 Relational models

As discussed, relationships are important to get stakeholders to create value for the organization according to stakeholder theory. Traditionally, stakeholder theory distinguishes between a transactional and a relational approach to manage stakeholder relationships (Jones,

based on unique social interactions while the transactional approach can more easily be imitated by competitors (Hillman & Keim, 2001).

The behavioral approach to instrumental stakeholder theory further develops stakeholder theory by adopting a more realistic, complex view of human behavior, relationships and stakeholder behavior (Bosse et al., 2009). Bridoux & Stoelhorst (forthcoming) integrate relational models theory (Fiske, 1992) into the behavioral approach to instrumental

stakeholder theory. By doing this, they provide a more realistic view of human relations than the simple distinction between either a transactional or relational approach.

According to relational models theory (Fiske, 1992), individuals use four relational models as schemata to organize all social interaction, including social interaction in organizations. Individuals use these relational models to frame relationships with other

individuals, but they also attribute human characteristics to firms, therefore they perceive their connection with the firm as a relationship as well (Morrison & Robinson, 1997; Mossholder et al., 2011). Usually, not only one relational model is used for a relationship, but a

combination of the four (Fiske, 1992).

The four relational models are communal sharing, authority ranking, equality matching and market pricing (Fiske, 1992).

Communal sharing

In communal sharing (CS), little attention is paid to individual identity; the focus is on group identity and solidarity (Fiske, 1992). In CS, others in the group are treated as equal, and as a result, individual welfare is less important than the welfare of the group (Haslam & Fiske, 1999). In CS settings, resources do not belong to individuals but are pooled, and individuals contribute to these resources as best as they can, while they take what they need

CS setting can benefit from collective resources simply by being part of the group, and that it does not depend on an individual‟s contribution.

Because individuals in a CS setting have a strong sense of collective commitment, they wish to be similar to others and be part of the group (Fiske, 1992). Intimacy and idealism provide the motivation for sharing in CS relationships (Boer et al., 2011). In work settings, this means that the mere fact that someone is part of the organization, entitles him to common resources, and all resources are shared with each other, regardless of how much someone contributes. On the other hand, individuals feel the joint responsibility to get work done in a CS relationship (Fiske, 1992), which makes it unlikely that free-riding occurs in public good settings.

Authority ranking

In authority ranking (AR) relationships, social interaction is organized around some kind of hierarchy (Fiske, 1992). This hierarchy is linear, and higher up is considered as better. Higher-ranked individuals appropriate more possessions and have more power than lower-ranked individuals, who obey their superiors (Fiske, 1992). Differences in hierarchy in AR relationships can be based on, for example, differences in expertise, age or formal power (Boer, van Baalen, & Kumar, 2004). An individual‟s identity is based on the individual hierarchical rank, which is contrary to CS relationships – in which individuals get their identity from the non-hierarchical collective to which they belong (Fiske, 1992).

AR relationships correspond with the traditional, hierarchical organization, in which the manager has authority over lower level employees and earns a higher salary. Another typical example of an AR relational model can be found in the army, in which roles, interaction and power are based on functional hierarchy (Fiske, 1992).

Equality matching

Equality matching (EM) is based on balance; and reciprocity is at the core of it. “The idea is that each person is entitled to the same amount as each other person in the relationship, and that the direction and magnitude of an imbalance are meaningful” (Fiske, 1992, p. 691). Thus, a strong emphasis is placed upon equality in EM relationships.

Although they are distinct models, EM and CS are frequently used together to explain relationships because they tend to occur simultaneously in some cultures (Haslam & Fiske, 1999). However, contrary to CS, asymmetries between individual contributions in a

relationship are an important aspect of EM. In practice, this means that it is relevant how much each individual owes another to restore balance in EM (Fiske, 1992). The balanced exchange in EM can manifest itself in material (e.g., borrowing and repaying money) and non-material exchanges (e.g., members of a group having an equal say in decision making).

Perhaps paradoxically, the underlying principle of EM can lead to unequal outcomes in social interactions (Fiske, 1992). Because balance is the desired equilibrium in EM

relationships, individuals aim to restore imbalances; this means that unequal distributions are meaningful (Fiske, 1992). If an individual is unable to repay debt, the relationship can develop into a relationship in which one or more individuals hold power over the other(s) (Fiske, 1992).

Market pricing

Market pricing (MP) is about proportionality in social relationships, and interactions revolve around a single metric (Fiske, 1992). An example of a MP transaction is that of a cost-benefit analysis, in which someone makes a decision to do something based on whether the perceived benefits outweigh the costs. The metric does not necessarily have to be money (Fiske, 1992).

Mossholder et al. (2011) propose that a MP culture in organizations leads to more

transactional, short-term relationships and lower involvement. For example, when a firm only offers employees temporary contracts and employees are always at the risk of getting fired if performance targets are not achieved, it is more likely that individuals feel less committed to the organization.

Relational models and OCBO

Firm behavior, such as incentive systems and appraisal structures, influences the relational climate and the relational models stakeholders adopt to frame their relationship with the firm (Mossholder et al., 2011). This is important for firms to consider, because that means that firm behavior can influence stakeholder behavior.

Many other authors have found mechanisms for organizing social interaction, which are similar to these relational models. Examples are Ouchi's (1979) market, bureaucracy and clan, which is very similar to Adler's (2001) market, hierarchy and community. More closely related to the traditional dichotomy in stakeholder theory are Clark & Mills' (1979) communal relationships and exchange relationships.

Considering four relational models for stakeholder management rather than the two mechanisms typically used in stakeholder theory (transactional / relational), or the

mechanisms discussed by other authors, could shed new light on how to get employees to perform OCBO.

2.4 Social dispositions

Furthermore, Bridoux & Stoelhorst (forthcoming) argue that stakeholders have an individual preference, or a social disposition, for a certain relational model. This means that an employee‟s social disposition matches the relational model he or she uses to frame the

disposition is important, because it determines how harmonious an individual perceives the relationship with the firm. These social dispositions therefore play an important role in the value creation process.

The fact that most stakeholders are not only interested in maximizing their own payoff, but that they also consider the payoff of others (Bridoux & Stoelhorst, forthcoming; Henrich et al., 2001), is in line with the idea of social preferences (Fehr & Fischbacher, 2002). These social preferences, or social dispositions, are relatively stable personality traits (Bogaert, Boone, & Declerck, 2008; Eisenberg et al., 1999). Although several social dispositions can be distinguished, I will treat four social dispositions, because most people can be classified in one of these categories. These social dispositions are self-regarding individuals, reciprocators, competitors and altruists (Bridoux & Stoelhorst, forthcoming).

Self-regarding

Self-regarding individuals perceive maximizing their own outcome as rational behavior, and do not contribute to the public good if they do not think that this will increase their personal payoff (Bogaert et al., 2008). This does not mean that self-regarding individuals do not cooperate, but their motivation to do so comes from the expectation that cooperation will increase their own payoff, rather than from their regard for other individuals‟ outcome (Fehr & Gintis, 2007).

Self-regarding individuals perceive their ambition to maximize individual outcomes as rational and intelligent and they expect others to behave rationally too (Bogaert et al., 2008; Kelley & Stahelski, 1970). They perceive self-regarding behavior as a sign of strength, whereas they perceive cooperation for the sake of cooperating or equality as weak and unintelligent (Smeesters, Warlop, Van Avermaet, Corneille, & Yzerbyt, 2003).

Reciprocator

Reciprocators care about fairness, possibly even at their own expense (Fehr &

Fischbacher, 2002). In practice, reciprocators care more about the wellbeing of the collective than self-regarding individuals; this means that they perceive cooperative behavior as rational if it increases joint value creation (Bogaert et al., 2008). Yet, reciprocators prefer equal outcomes over larger joint outcomes, and particularly over larger unequally distributed joint outcomes (Eek & Gärling, 2006). Because self-regarding individuals are not only concerned about their own outcomes, it is more likely that reciprocators contribute in public good dilemmas than self-regarding individuals.

Reciprocators not only care about the equality and fairness of the outcome, they will also change their behavior in response to the behavior of others. Reciprocators will cooperate if they expect this cooperation to be returned by the other party (Bogaert et al., 2008), or not cooperate when they do not expect the other party to cooperate (Eek & Gärling, 2006). In response to other people‟s actions, they may even become less cooperative than the party that did not behave fair in their perception (Kelley & Stahelski, 1970). This uncooperative and unforgiving reaction to unfairness stems from reciprocators‟ strong desire for fairness and equality (Van Lange, 1999).

Competitor

Competitors and self-regarding individuals are often treated as a single group (Eek & Gärling, 2006; Kelley & Stahelski, 1970). However, competitors and self-regarding

individuals have different motivations to either cooperate or not. Self-regarding individuals seek to maximize their own outcome, while competitors care about maximizing their own payoff compared to others (Van Lange, 1999). Following this logic, self-regarding individuals will cooperate if this will increase their own outcomes, while competitors will cooperate if

Altruist

Altruism is described as unconditional kindness, meaning that altruists cooperate because they care about increasing other individuals‟ payoffs, regardless of the other person‟s actions (Fehr & Fischbacher, 2002). Pure altruism is very uncommon, for that reason reciprocators and altruists are often grouped together, for instance as „prosocials‟ (e.g., Van Lange, 1999).

To illustrate the logic of social dispositions, consider the following, hypothetical, situation. An employee has to divide one hour of its time between two different projects. Project A is his own project; project B is the project of a coworker. Depending on how much time the employee spends on project B, the coworker would spend an equal amount of time working on project A. The employee has the choice to a) spend an hour on his own project, b) spend an hour working on the coworkers project, c) divide time equally between the two projects, or d) spend 45 minutes on his own project and 15 minutes on the coworkers‟ project. The payoffs for the employee and coworker are listed in Table 1.

Option Time spent on project A Time spent on project B Own payoff Coworkers‟ payoff

a 60 minutes 0 minutes $ 50 $ 0

b 0 minutes 60 minutes $ 0 $ 50

c 30 minutes 30 minutes $ 40 $ 40

d 45 minutes 15 minutes $ 60 $ 20

Table 1: Social dispositions

In this situation, a self-regarding individual would choose option d, because it maximizes his or her own payoff. A reciprocator would choose option c because he would perceive it as „fair‟ because the equal time spent on each other‟s projects leads to equal outcomes. A competitor would choose option a, because it maximizes his own payoff compared to the other. And finally, an altruist would choose option b, because it maximizes the coworkers‟ outcome. Only reciprocators take the process into account because they attach much value to fairness, while the other groups are primarily concerned with the outcome.

2.5 Self-control

As discussed, individual differences play an important role in stakeholder behavior. According to Cohen & Morse (2014), the distinction between what is right and wrong, and finding a balance between self-interest and the interest of others depends on a motivational element (consideration of the needs and interests of others) and identity element (how one values morality), as well as an ability element (the regulation of one‟s own behavior).

Whereas the motivational and identity element show strong similarities with social disposition, the ability element is not yet discussed. This ability element is about the

regulation of one‟s own behavior, also referred to as self-control, i.e., “the exertion of control over the self by the self” (Muraven & Baumeister, 2000, p. 247). Self-control is an important trait to consider when investigating OCBO, because employees do not only have to identify the appropriate behavior, but should also be able to act upon it. Self-control has proven to be beneficial in many areas of life (Tangney et al., 2004). Individuals who have high self-control are better able to perform behavior that is beneficial in the long run over behavior that

provides short-term benefits (Muraven & Baumeister, 2000).

The terms self-control and self-regulation are either used interchangeably (e.g., Alquist & Baumeister, 2012) or self-control is seen as deliberately exercising self-regulation (e.g., Tangney et al., 2004). Here, I will make no such distinction, and treat control and self-regulation as the same, consisting of both the deliberate and non-deliberate exertion of control over the self.

Self-control can be seen as both a trait and as a resource. Self-control as a trait is relatively stable over time, and individuals with high self-control benefit from a wide range of

advantages over those who have low self-control (Tangney et al., 2004). Examples are higher grades in school, fewer impulse control problems, better psychological adjustment and better

interpersonal relationships (Tangney et al., 2004). Individuals with low self-control, on the other hand, are more likely to engage in a wide range of damaging habits such as

procrastination, cigarette smoking, alcoholism, excessive spending, overeating and violence (Sayette & Griffin, 2004).

Self-control can also be treated as a resource that can get depleted (ego depletion) by exercising self-control (making it more difficult to exercise self-control on subsequent tasks that require self-control) and replenished again (for example by performing unrelated tasks or by resting) (Muraven & Baumeister, 2000). Here, I will focus on self-control as a trait, because it is not subject to ego depletion, and will therefore be influenced less by external factors than self-control as a resource.

3. Research questions

As discussed, OCBO has been identified as essential for firms to function well, but it is also vulnerable to a public good dilemma. Instrumental stakeholder theory stresses the importance of relationships in these situations. Relational model theory suggests that stakeholders frame these relationships using four relational models, rather than the two usually discussed in the stakeholder literature (transactional and relational). However, little attention has been paid to how relational models affect stakeholders‟ contributions to value creation. Bridoux & Stoelhorst (forthcoming) identified this gap in a theory paper, and propose that certain relational models are more suitable for value creation in firms than others. Here, I will study these relationships empirically by specifically looking at how the different relational models used by employees affect OCBO. OCBO positively contributes to organizational performance and little research has been done in this area, therefore it is an interesting area to look into. I focus on employees because they are important primary stakeholders, who are responsible for value creation in firms.

Research question 1: Do relational models affect OCBO?

Furthermore, Bridoux & Stoelhorst (forthcoming) suggest that individuals are predisposed towards a relational model. This depends on the stakeholder‟s social disposition. Yet, it is unclear how certain combinations of relational models and social dispositions affect OCBO, specifically when there is a mismatch between the relational model and social disposition.

Research question 2: Do social dispositions moderate the effect that relational models

have on OCBO?

Although a lot of research has been done on the benefits of self-control, only little is known about the benefits of self-control on OCBO. Most research on self-control in firms

focuses on self-control at the unit or firm level, instead of individual self-control (e.g., Osterloh & Frey, 2000).

Given the many advantages of control, and the lack of research on the effect of self-control on OCBO, it is a promising stream of research to look into. Particularly interesting is the role of self-control when there is a mismatch between the stakeholder‟s social disposition and the relational model used to frame the relationship with the firm.

Research question 3: Does self-control moderate the moderating effect that different

4. Theoretical framework

In the literature review, I highlighted the importance of firm-stakeholder relationships in the value creation process and, more specifically, OCBO. Next to this, I discussed individual differences in the consideration for others‟ payoffs (social disposition) and self-control. After this, I identified a number of specific gaps in the literature and corresponding research

questions. In this section, I will address the relationships between relational models, OCBO, social dispositions and self-control and formulate several hypotheses. A graphical

representation can be found in the conceptual model at the end of this section (Figure 1).

4.1 Relational models and OCBO

According to Bridoux & Stoelhorst (forthcoming), stakeholders contribute most to team production if they frame the relationship with the firm in terms of CS, followed by

respectively EM, AR and MP. I expect the same order to apply for the amount of OCBO performed.

First of all, I expect that most OCBO occurs in CS relationships. Individuals view the welfare of the group as more important than their individual welfare in CS relationships (Haslam & Fiske, 1999). A CS relationship motivates individuals to perform OCBO because it contributes to the welfare of the group (value creation and firm performance), which is perceived as more important than the price paid by the individual (time and effort without a direct reward in return). Following the same logic, individuals in CS relationships are more likely to contribute in a public good dilemma because the public good is considered as valuable.

Secondly, I expect stakeholders to perform more OCBO in EM relationships than in AR or MP relationships. EM relationships revolve around fairness and equality (Fiske, 1992).

perform OCBO if they believe that they are treated well by the organization, or if they believe that they are obliged to do so because of the benefits they receive from the organization (Mossholder et al., 2011). However, I expect EM to lead to less OCBO than CS, because in EM employees will only perform OCBO as long as they expect their contributions to be equal to the effort they (will) receive from the firm. Stakeholders in CS, on the other hand, are motivated to perform OCBO because they care about the welfare of the collective (the firm).

Finally, I expect that more OCBO is performed in AR than in MP. In MP, the relationship is based on prices and contracts, and OCBO is, by definition, behavior that is not formally rewarded and that extends beyond job specifications. For this reason, OCBO is unlikely to occur in MP settings. AR, on the other hand, can positively influence OCBO because it causes obedience and respect for authority, and OCBO is a way to show this (Hongyu, Mingjian, & Qiang, 2011).

Furthermore, Bridoux & Stoelhorst (forthcoming) argue that EM is a better coordination mechanism than AR, but they do not expect the difference between stakeholder motivation to contribute to be significant between the two models. However, I expect EM to lead to more OCBO than AR because perceived fairness (which is at the core of EM relationships) has shown to correlate positively with OCB (Organ & Ryan, 1995).

Hypothesis 1a: Employees who frame their relationship with the firm as CS perform

more OCBO than those who frame their relationship as EM, AR or MP.

Hypothesis 1b: Employees who frame their relationship with the firm as EM perform

less OCBO than those who frame their relationship as CS, but more than those who frame

Hypothesis 1c: Employees who frame their relationship with the firm as AR perform

less OCBO than those who frame their relationship as CS or EM, but more than those who

frame their relationship as MP.

4.2 Social dispositions, relational models and OCBO

Bridoux & Stoelhorst (forthcoming) propose that individuals are predisposed to frame their relationships as CS, EM, AR or MP, depending on their social disposition. They argue that self-regarding individuals tend to frame relationships as MP, reciprocators as EM, competitors as AR and altruists as CS. An individuals‟ social disposition indicates how individuals prefer to distribute payoffs between themselves and others (Fehr & Fischbacher, 2002). In this section, I will discuss how different social dispositions will influence the effect that a relational model has on OCBO. There are two sides to the underlying idea. First of all, dispositional characteristics play a role. For example; altruists care about other people‟s payoff regardless of their own payoff, while self-regarding individuals seek to maximize their own payoff. Therefore I expect altruists, in general, to perform more OCBO than

reciprocators, competitors and self-regarding individuals.

Secondly, it is important how well an individual‟s social disposition fits the adopted relational model. Although little is known about the match between relational model and social disposition, findings from studies on person-organization fit show that a match between individual personality and organizational culture and values is an important predictor of task- and contextual performance (Borman & Motowidlo, 1997; Goodman & Svyantek, 1999). Moreover, personality shows a stronger correlation with contextual performance (of which OCBO is part), than overall performance (Borman & Motowidlo, 1997). In line with these findings, I expect that individuals, whose social disposition matches the relational model they use to frame the relationship with the firm with, are likely to perform more OCBO. Or, to put

it differently: if individuals feels that their values correspond with the firm, they are more motivated, feel more competent and are more likely to perform extra effort (such as OCBO) (Benkhoff, 1996; Chatman, 1989; Organ & Ryan, 1995)

Hypothesis 2a: Employees perform more OCBO when there is a match between the

relational model they use to frame their relationship with the firm and their social

disposition, than when there is a mismatch.

Next to the simple distinction between match and mismatch, I also expect that different combinations of relational models and social dispositions lead to different performed amounts of OCBO. From here on, I will discuss how I expect the different social dispositions to affect OCBO for different relational models.

Communal sharing

As discussed, I expect individuals to perform most OCBO when there is a match between their social disposition and the relational model they use to frame the relationship. Since altruists are predisposed to framing relationships as CS, I expect altruists to perform most OCBO in a CS relational model. Furthermore, I expect altruists to perform more OCBO than individuals with any other social disposition in any relational model, because they care about others‟ payoffs.

I also expect reciprocators to perform a lot of OCBO when they frame their relationship with the firm as CS. Reciprocators expect the firm and others within the firm to contribute as much as they can to the common good, because that is central to CS relationships. On the other hand, reciprocators care about equality, meaning that they want to maximize joint outcomes as long as they perceive others to make equal contributions. However, CS relationships are not based on equality, but on contributing what you can, and taking what you need. This

contribute, but feel like they do not receive the same in return, will perceive this as unfair. Kelley & Stahelski (1970) show that the strong desire for equality may lead to

overcompensation in cooperators. As a result, reciprocators may react even more selfish, in order to punish others whose behavior they perceive as unfair; therefore I expect reciprocators to perform less OCBO than altruists in CS contexts.

Unlike altruists and reciprocators, self-regarding individuals and competitors care primarily about their own outcome; therefore I expect them to perform less OCBO. In CS contexts, this would mean that self-regarding individuals are motivated to perform OCBO if they expect this to maximize their own outcome. Yet, I still expect self-regarding individuals to perform more OCBO than competitors, because competitors optimize their payoff compared to others by not performing OCBO, because that way they can benefit from the positive outcome of OCBO without contributing to it. Self-regarding individuals however, care about maximizing their own payoff, that means that they will perform OCBO if that leads to increased

individual payoff.

Hypothesis 2b: When individuals frame their relationship with the firm as CS, (1)

altruists perform more OCBO than reciprocators, self-regarding individuals and

competitors; (2) reciprocators perform more OCBO than self-regarding individuals and

competitors; (3) self-regarding individuals perform more OCBO than competitors.

Equality matching

The norms in an EM relational model are about reciprocity; you help others and expect the same in return. As discussed, reciprocators exhibit behavior similar to what is at the core of this relational model, and are predisposed to frame relationships as EM. Because a good person-organization fit is important for contextual performance (such as OCBO), I expect reciprocators to perform most OCBO when they frame their relationship with the firm as EM.

However, because reciprocators expect an equal reward in return for their contribution, I expect altruists – who aim to maximize others‟ payoffs, without expecting something in return - to perform more OCBO than reciprocators.

Similar to what is discussed in CS contexts, I do not expect competitors to perform a lot of OCBO in EM contexts. The underlying argument for this is that competitors seek to

maximize their own payoff compared to others, but will only receive similar rewards in return for their contribution, which decreases their payoff compared to others. For example, if a competitor has found a method to get things done more efficiently, sharing this with the entire workforce will decrease his or her competitive advantage over coworkers. On the other hand, self-regarding individual may decide to share the information, as long as the increased joint efficiency also leads to higher individual payoff.

Hypothesis 2c: When individuals frame their relationship with the firm as EM, (1)

altruists perform more OCBO than reciprocators, self-regarding individuals and

competitors; (2) reciprocators perform more OCBO than self-regarding individuals and

competitors; (3) self-regarding individuals perform more OCBO than competitors.

Authority ranking

Similar to the relational models discussed earlier, I expect altruists to perform most OCBO in AR contexts because they seek to maximize the payoffs for others. After altruists, I predict that competitors perform most OCBO in AR contexts. Competitors are motivated by competition and status, which is central to AR contexts. Superiors value beyond the job activities (such as OCBO) (Borman & Motowidlo, 1997), therefore it is likely to contribute to status and position on the hierarchical ladder in the organization if it is noticed.

to perform OCBO. However, competitors are more motivated by increasing their outcomes over others (Van Lange, 1999), which is central to AR contexts, therefore I expect them to perform more OCBO. Finally, I expect reciprocators not to fare well in AR contexts, because individuals are entitled to privileges in AR relationships based on their rank in the

organization, rather than their contribution. This principle is central to AR relationships and does not correspond with reciprocators‟ concern for equality. In addition to this, reciprocators tend to behave even more uncooperatively in situations which they perceive as unequal (Kelley & Stahelski, 1970), therefore I predict that reciprocators will perform least OCBO in AR contexts.

Hypothesis 2d: When individuals frame their relationship with the firm as AR, (1)

altruists perform more OCBO than competitors, self-regarding individuals and

reciprocators; (2) competitors perform more OCBO than self-regarding individuals and

reciprocators; (3) self-regarding individuals perform more OCBO than reciprocators.

Market pricing

Again, I expect altruists to perform most OCBO, because their goal is to maximize the payoffs for others, even though rational comparisons by ratios is at the core of MP

relationships (Fiske, 1992). This is because altruists‟ contributions to other‟s wellbeing are unconditional (Fehr & Fischbacher, 2002). Perhaps paradoxically, I expect self-regarding individuals to perform most OCBO after altruists in MP contexts. Because self-regarding behavior (maximizing individual payoff) is appropriate behavior in MP relationships, regarding individuals are predisposed to frame their relationships as MP. As a result, self-regarding individuals feel comfortable in MP contexts and therefore perform more OCBO. This is in line with Organ & Ryans (1995) findings that job satisfaction, perceived fairness and leader supportiveness are antecedents of OCBO.

Finally, I expect competitors to perform more OCBO than reciprocators in MP relationships. Reciprocators are unlikely to perform much OCBO in MP contexts for a number of reasons. First of all, they will not receive an equal compensation in return for performing OCBO in MP relationships (as rewards in MP relationships revolve around a single metric, and OCBO is not formally rewarded). Secondly, reciprocators do not feel comfortable in a MP relationship, which leads to lower OCBO (Organ & Ryan, 1995). This is because maximizing individual payoff is appropriate behavior, rather than reciprocation. The third argument is in line with the previous argument; when reciprocators contribute, they expect equal contributions from the other party in a relationship. Maximizing individual payoff is appropriate behavior in MP relationships, but perceived as unfair by reciprocators. Consequently, reciprocators will respond by behaving even more selfish (Van Lange, 1999). Competitors, on the other hand, are not very likely to perform OCBO in MP relationships either. They care about maximizing their own payoff compared to others, and OCBO is not (formally) rewarded. However, reciprocators care about equality, and competitors care about maximizing payoff compared to others. The outcome of this is that competitors may perform OCBO if they expect this to contribute positively to their position in the firm.

Hypothesis 2e: When individuals frame their relationship with the firm as MP, (1)

altruists perform more OCBO than self-regarding individuals, competitors and

reciprocators (2) self-regarding individuals perform more OCBO than competitors and

reciprocators; (3) competitors perform more OCBO than reciprocators.

4.3 Self-control and OCBO

In the previous part, I explained why I expect some combinations of relational models and social dispositions to lead to more OCBO than others. As discussed in the literature review, it is interesting to investigate the role of self-control, specifically because OCBO is costly but

has many advantages for both the individual and the firm. Furthermore, individuals who have high self-control are better able to exert control over their own behavior, and they are able to perform behavior that is beneficial in the long run rather than the short run (Muraven & Baumeister, 2000).

I expect individuals who have high self-control to perform more OCBO, because it is costly for the individual in the short term (it takes time and effort to perform OCBO), but beneficial for both the individual and the firm in the long term (e.g., better organizational performance and better individual performance appraisals (N. P. Podsakoff et al., 2009)). More specifically, I expect self-control to play an important role when there is a mismatch between the individuals‟ social disposition and the adopted relational model because

individuals who have high self-control, are better able to regulate their behavior in such a way that there is a better fit between themselves and their environment (Rothbaum, Weisz, & Snyder, 1982).

Hypothesis 3: When there is a mismatch between the relational model a stakeholder

uses to frame his relationship with the firm and the stakeholder’s social disposition, stakeholders who are high in self-control perform more OCBO.

5. Methodology

In this section, I will discuss the design of my research. I will start with an overview of the overall research design, followed by an explanation of the used sample, the method data collection and the operationalization of the used variables.

5.1 Research design

The three main goals of this research are to empirically test 1) whether relational models influence OCBO, 2) whether social disposition moderates this effect and 3) whether self-control moderates this moderating effect.

Because the variables that are tested are well known, and all relationships are hypothesized, this is explanatory research (Saunders, Lewis, & Thornhill, 2009). More specifically, I have conducted quantitative research among employees who work at various organizations. Because it is important to collect a relatively large amount of data to be able to test the hypotheses formulated earlier, I jointly collected data with a student who was conducting a similar research. We decided to use a self-administered questionnaire, which is internet-mediated. The reason to use this method of collecting data is that it enabled us to collect a large amount of data from many respondents from a large geographic area in an economical way (Saunders et al., 2009; Wright, 2005). Next to this, internet-mediated self-administered questionnaires can be filled out anonymously, which limits the chance that respondents feel the need to give socially desirable answers (Duffy, Smith, Terhanian, & Bremer, 2005). Furthermore, because I have well-defined variables, a questionnaire enables me to collect standardized quantitative data (Wright, 2005), which makes it easy to compare respondents‟ answers and analyze them using statistical software. The questionnaire was developed and distributed using the online software Qualtrics.

A major disadvantage of collecting data through self-administered online-mediated

questionnaires is that it is convenience sampling, which means that it is not random (Saunders et al., 2009). Next to that, the surveys are self-administered, thus respondents cannot ask a researcher questions if something is not clear (Saunders et al., 2009)

Furthermore, respondents may be more susceptible to fatigue when filling out an online-mediated questionnaire than when an interviewer is present, this is particularly the case for longer questionnaires (Duffy et al., 2005). This may have negative consequences for the quality of responses as respondents may put less effort in reading and answering the questions thoroughly (Duffy et al., 2005).

5.2 Research sample and data collection

Because I am studying the effects of relational models, social disposition and self-control on organizational citizenship behavior by employees, this research is conducted among

employees in general, regardless of the type of organization they work at. It is, however, important that respondents have a strong sense of the organizational culture because this influences the relational model they use to frame their relationship with the firm (Haslam, 2004). Therefore, I am looking for employees who spend considerable time at the

organization. I decided to conduct the research among people who work at least 24 hours a week.

Because it is impracticable to collect data from the entire population, a sample is used in this research. In order to collect all the necessary data, filling out the questionnaire will take a significant amount of time. To increase the chance that potential respondents fill out the questionnaire, we personally approached potential respondents in our network and also asked them to distribute the questionnaire among others. This was done by sending out e-mails to

message asked people to fill out the survey followed by a request to forward the survey to colleagues, relatives, friends and acquaintances who work a minimum of 24 hours per week.

As a result, the main sampling technique was convenience sampling, with some responses coming from snowball sampling. These methods of collecting data are not random and it is therefore unlikely that the results can be generalized to the entire population. Although this could lead to problems, there are two reasons to believe that the research is designed in such a way that the collected data is very much usable. First of all, I am not interested in drawing conclusions about the composition of the population (e.g., which percentage of the population is self-regarding), but rather to find enough variety in the (behavior and attributes of)

respondents so that I can investigate the effect on OCBO. Secondly, I included several control variables, which may shed light on potential limitations of not using a random sampling technique in this research.

Because I collected data with a student with an academic background in the United Kingdom, data was collected among Dutch speaking employees as well as English speaking employees. For this reason, two versions of the survey were designed (Dutch and English).

The target sample size was 200 respondents, which was chosen because this is feasible within the timeframe of this research, yet a big enough number to draw meaningful conclusions from the data. All data was collected between April 30th and May 18th, 2015. In total, 363 people started the survey, 21 people did not complete the first part of the survey (control questions) and were therefore excluded. 22 individuals indicated that they work 24 hours per week or less, and were also excluded. 94 respondents did not answer the questions about the relational models, which is the independent variable that is used for all hypotheses in this research, therefore they were excluded as well. After cleaning up the data, 226 respondents remained.

5.3 Operationalization of constructs and variables

In this section, I will discuss the operationalization of my variables. First, I will discuss the independent variable (relational models) and the dependent variable (OCBO). After that I will discuss the moderator (social disposition) followed by the moderator of the moderator (self-control). Finally, I will describe the control variables that were used.

Dependent variable - OCBO

The dependent variable is OCBO. OCBO is organizational citizenship behavior directed at the organization (rather than OCBI, which is directed at other individuals). The scale used in my survey was developed by Lee & Allen (2002). Respondents are presented with 8 statements about certain behavior at the workplace. For each of these statements they are asked to indicate how often they engage in these activities (7-point Likert scale from “Never” to “Always”). This scale was originally developed in English, and was translated by myself and other students participating in similar research using the usual procedure of translation and back translation.

Independent variable – Relational models

The independent variable in this study is the relational model. As discussed, individuals use relational models to frame relationships with others, but also with nonhuman entities such as a firm (Morrison & Robinson, 1997; Mossholder et al., 2011). To measure which relational model an individual uses, an edited version of the Organizational Culture Assessment Instrument (OCAI) by Cameron & Quinn (2011).The OCAI is a widely used instrument to assess organizational culture. Three of the four categories in the original version of the OCAI closely resemble the relational model (i.e., market is similar to market pricing, clan is similar to communal sharing, hierarchy is similar to authority ranking). A version that includes the fourth relational model (equality matching) instead of adhocracy was developed and pretested

items were changed. In the edited version that was used in this survey, respondents are

presented with 32 statements (7 categories, each consisting of 4 items) and are asked to rate to what extent their employment organization fits each statement (a 6-point Likert scale from “Not like my organization at all” to “Completely like my organization”). I translated the final scale from English to Dutch with the help of other students using the usual procedure of translation and back translation.

Moderator – Social disposition

I hypothesize that an individual‟s social disposition moderates the effect that relational models have on OCBO. To measure social disposition, I use Schwartz‟ portrait value

questionnaire (Schwartz, 2003), in which respondents are presented with several descriptions of persons, for which they are asked to indicate how much they are similar to them (6-point Likert scale from “Not like me at all” to “Very much like me”). Male respondents were provided with descriptions of male persons, while female respondents were provided with descriptions of female persons, apart from the personal pronoun, all descriptions were the same.

Schwartz offers a theory of 10 cross-cultural human values, which individuals use to pursue goals (Schwartz, 1992, 1994, 2003; Schwartz et al., 2001). The 10 values are classified in two dimensions; self-enhancement vs. self-transcendence and openness to change vs conservatism (Schwartz, 2003). The former dimension consists of four values that closely resemble the four social dispositions discussed in this thesis. Self-enhancement consists of the values power and achievement, which are about pursuing self-interest. Self-transcendence consists of

universalism and benevolence, which are about the welfare of others (Schwartz, 2003).

More specifically, power is about wealth, authority and status (Schwartz, 1994, 2003), which closely resembles the social disposition competitor. Achievement is about personal success

(Schwartz, 1994; Schwartz et al., 2001), which is similar to the social disposition self-regarding.

On the other hand of the spectrum, universalism is about protecting and helping everyone, including non-human entities such as the environment (Schwartz, 1994; Schwartz et al., 2001), which is similar to what altruists care about. Benevolence is more about caring for the welfare of individuals with whom one interacts frequently (Schwartz, 1992, 1994), like reciprocators do.

The scale consists of 17 items. One item (item 5 - “It is important to him/her to be rich. He/she wants to have a lot of money and expensive things”) is categorized under the power value by Schwartz (2003), but based on the theory, I decided that this item more closely resembles the social disposition self-regarding. An exploratory factor analysis confirms this choice (reported on page 47). The rationale behind this is that it does not have to do with outperforming others (as is characteristic of competitors), but only about personal

achievement (which corresponds with self-regarding). For this reason, I will treat this item as if it is an achievement item. This scale was originally developed in English. For the Dutch version, I used the translation by the European Social Survey(“European Social Survey,” 2010)

Moderating moderator – Self-control

Self-control is measured using the brief self-control measure developed by Tangney et al. (2004). This measure presents respondents with 13 statements for each of which they are asked to indicate to what extent it reflects how they typically are (5-point Likert scale from “Not at all” to “Very much”). This scale was originally developed in English. For the Dutch version, I used the Dutch translation by Helmerhorst, de Vries Robbé, & de Vogel (2011).

Control variables

The following variables were included in the survey to control for: gender, age, nationality, highest achieved level of education, working hours per week, occupation, organizational size (“How many people work at your employment company (yourself included)?”),

organizational tenure (“How long have you been working at the company you are currently employed with?”), and days at the employers‟ premises (“How many days a week do you work at your employer‟s premises”).

There are various views on the effect that the abovementioned variables have on OCBO. In a review article, P. M. Podsakoff et al. (2000) state that demographic variables are usually not related to OCB. On the other hand, several theoretical and empirical studies indicate that these variables do influence OCB. For example, Kidder & Parks (2001) propose that there may be differences between men and women in the amount of extra-job behavior they perform. Turnipseed & Murkison (2000) argue that (national) culture may have an effect on OCB. Jahangir, Akbar, & Haq (2004) propose a theoretical framework in which age is an antecedent of OCB. A positive relationship between level of education and OCB was found by Ng & Feldman (2009) as well as between organizational tenure and OCB (Ng & Feldman, 2010). Finally, Redman, Snape, & Ashurst (2009) found no significant negative relationship between homeworking and OCBO in knowledge work, but arguments can be made that little physical presence can hinder employees in the ability to perform OCBO.

Because of contradicting findings on the effect of the mentioned variables, I will include them as controls in this research. Another reason why it is interesting to include these control variables is that the majority of studies have focused on the antecedents of OCB in general, and very little is known about the relationship between these variables and OCBO.

5.4 Merging the questionnaires

Because two versions of the questionnaire were used, data from the Dutch and the English version had to be combined in one dataset. This did not lead to any problems, as the Dutch version was a translation of the English version. One exception was the question about the highest level of education. In Dutch, the tertiary education system is divided between MBO (middle-level applied education), HBO (universities of applied sciences) and WO (bachelor and master levels at research universities). In the Dutch questionnaire, high school was followed by MBO as a separate category, followed by HBO or WO Bachelor, and finally WO master or higher. In the English version the distinction was made between “college or

university bachelor” and “university bachelor or higher”. To combine both surveys, MBO was added to the “college or university bachelor” group, as it is part of the Dutch tertiary education system.

6. Results

In this section, the results of the statistical tests are presented. I will start with a reliability analysis for all variables that consist of multiple items, including an overview of the variables I constructed (6.1). After that, descriptive statistics are presented to give an impression of the data (6.2). Before testing the hypotheses, a correlations analysis (6.3) and normality analysis (6.4) is performed. Finally, the findings of the analyses are presented (6.5).

6.1 Reliability analysis

Dependent variable - OCBO

The scale for OCBO consists of eight questions, with a corresponding Cronbach‟s alpha of .868, which indicates good internal consistency. No questions could be removed to improve internal consistency. I calculated an average OCBO score per respondent, which is the variable I will use for testing all hypotheses.

Independent variable – Relational Models

To assess the variance in variables, I performed a principal components analysis on the data for relational models. To determine whether this is an appropriate analysis, I first calculated the overall KMO value (.889) as well as the individual KMO values, which were between .771 and .940 (Table A1 in Appendix A). These KMO values show adequacy of sampling. Furthermore, Bartlett‟s Test of Sphericity is statistically significant, indicating that a principal components analysis is appropriate.

The principal components analysis does not clearly show four components (see Table A2 in Appendix A), as would be expected judging from the literature. Using to the Guttman-Kaiser criterion (use variables that have an eigenvalue of at least 1 (Yeomans & Golder, 1982)), seven factors should be included, but a closer inspection of the scree plot shows a

components (see Figure A1 in Appendix A). Hence, different ways of interpreting these results would lead to a different number of components retained.

I consequently performed a principal components analysis with a fixed number of four factors, which is the number of relational models the questionnaire is designed for to measure. An inspection of the rotated components matrix shows that these four factors explain 52.45% of variance. However, most questions about the CS and EM relational model both load on the first factor. This can be due to a number of reasons, such as the quality of the questions, the fact that EM and CS are actually not two distinct models in reality or that both occur simultaneously. I will discuss this in more detail in the discussion, but removing questions does not change this. The rotated components matrix can be found in table 2, with in bold the value of the component the questions load strongest on. The Cronbach‟s alphas did show high levels internal consistency for all four relational models (CS: .832; AR: .781; EM: .860; MP: .869), and removing questions would not significantly improve these values. Although the principal components analysis does not provide the expected results, I decided to keep the literature as starting point and I do not remove questions from the results, as this does not seem to solve the problem.

I used the items about relational models to calculate a number of variables. First of all, I calculated the mean score per relational model per respondent, so each respondent ends up with a continuous value between 1 and 6 for each of the four relational models. After that, I calculated the dominant relational model per respondent by identifying for which relational model each respondent scored highest. Because 12 respondents report a mixed dominant relational model (i.e., their highest scores are equal for two or more relational models), they were excluded from analyses that involve dominant relational models. Ultimately, 214 respondents were included in all analyses involving dominant relational models.