Relational models and value creation : exploring the effect of relational models on organizational citizenship behaviours and discovering the mediating role of trust and the moderating role of social dispositions

(1)

E l i n e R e y n a e r s ( 6 0 5 2 4 5 2 )

Relational models and value creation: Exploring

the effect of relational models on organizational

citizenship behaviours and discovering the

mediating role of trust and the moderating role of

social dispositions.

!

(2)

(3)

Value creation is one of the most important objectives of firm strategy discussed in the strategy literature (e.g., Bowman & Ambrosini, 2000; Brandenburger & Stuart, 1996; Bridoux, Coeurderoy & Durand, 2011; Rumelt, 2003). Ideas of how to maximize value creation have differed over time (Rumelt, 2003), although the concept is regularly connected to firm performance (Bowman & Ambrosini, 2000; Brandenburger & Stuart, 1996). Recently, the role of human capital within the process of value creation has received increasing attention within the strategy literature (Cabrera & Cabrera, 2002; Strikwerda & Stoelhorst, 2009). Several researches have advocated that human resources are an essential driver of value creation (Bridoux, Coeurderoy & Durand, 2011; Coff, 1997; Coff, 1999; Oostrom, 2000). Having valuable resources is not per definition sufficient to achieve value creation, what matters is the motivation of employees to effectively use these resources (Bowman &

(6)

Ambrosini, 2000; Bridoux, Coeurderoy & Durand, 2011). Employees’ motivation to create value manifests itself among others as organizational citizenship behaviour (OCB). OCB refers to behaviour of human resources within the company that goes beyond role requirements without being recognized by the formal reward system (Bolino, Turnley & Bloodgood, 2002). Given its importance, the present study will focus on OCB to capture employees’ contributions to value creation.

One theory that acknowledges the importance of human resources for organizational performance is instrumental stakeholder theory (Bridoux & Stoelhorst, forthcoming). It is concerned with the enhancement of firm performance through the management of stakeholders (Bridoux & Stoelhorst, forthcoming). More specifically, it aims to find ways to manage stakeholders’ behaviour in such a way that it positively affects the firm’s performance (Bridoux & Stoelhorst, forthcoming). In contrast to established theories concerned with human resources in the strategy field, like transaction cost economics and agency theory, instrumental stakeholder theory takes issue with the assumption of opportunism (Jones, 1995). Instrumental stakeholder theory focuses on relations with stakeholders from a moral perspective, and claims that it is not opportunistic behaviour that provides the organization with a competitive advantage (Jones, 1995). Opportunism can result in malfunctioning relationships, especially when it goes hand in hand with a lack of reliable information. Organizations should therefore engage in trusting and cooperative behaviour in order to gain a competitive advantage (Jones, 1995).

The moral perspective proposed by instrumental stakeholder theory provides opportunities for the individual focus that is advocated for in the microfoundational approach, which has recently gained interest in the strategy field. Following the recent focus on microfoundations,

(7)

there is an emphasis on the role of individual-level factors and the way in which these influence the organization (Foss, Felin & Ployhart, 2015). One theory that could contribute to advance our knowledge of what is going on at the individual level is proposed by Fiske (1991), and is called relational models theory. Relational models theory centres on the idea that people are fundamentally sociable (Fiske, 1992). People accordingly use four relational models that determine their behaviour in social interactions (Fiske, 1992). Empirical support for the existence of the four relational models already exists (e.g. Haslam & Fiske, 1999; Vodosek, 2009), but existing literature has not yet been able to expand its scope much beyond the empirical justification of the models’ existence (Fiske, 1992; Haslam, 2004). Perhaps more importantly, relational models theory has not yet been connected to the strategic literary field. Some exploratory research has been done (e.g., Bridoux & Stoelhorst, forthcoming), but no empirical study has been conducted to statistically determine possible correlations between relational models and OCB.

The current study contributes to the strategy field by conducting such a research. The potential influence of the different relational models proposed by Fiske (1991) on organizational citizenship behaviour is tested by use of an online survey. The possible mediating influence of trust in the organization and the potential moderating role of social dispositions are included to look deeper into the relationship between relational models and OCB. The introduction of relational models within strategic management enriches and possibly expands theories build upon microfoundations and around stakeholders. The results contribute to the theory for example by showing that only specific models directly affect OCB. Also, trust does not always mediate the relationship between relational models and OCB, and social dispositions are only of importance when the most egoistic relational model is the one dominating the relationship. From a more practical point of view, this study

(8)

contributes to management functions by providing organizations with valuable information about the practical implications of the relational models that are used within their organization. Some relational models apparently lead to higher levels of OCB than others, which makes it worth considering to implement these where possible. It is furthermore suggested by the results that depending on which of the four organizational models occurs, organizations should adapt their selection criteria and organizational structure to allow them to optimize value creation in each situation.

This study looks into the possible relation between relational models and OCB by first reviewing the existing literature that discusses the relevant variables. A theoretical framework will follow in which expectations will be stated in relation to the proposed research questions. Then, the research design will be outlined after which the results will be discussed. The thesis will end with a conclusion that summarizes the results and makes recommendations for future research.

2. LITERATURE REVIEW

Even though the way in which it can best be achieved varies across different theories and schools of thought (e.g. Brandenburger & Stuart, 1996; Peteraf & Barney, 2003; Stoelhorst & Bridoux, 2008), the importance of value creation for organizational existence and success is accepted in each and every one of them. Value creation is specifically interesting to look at in the light of relational models since human resources are a necessity for the creation of value (Bridoux, Coeurderoy& Durand, 2011). Dealing with human resources implies dealing with knowledge (Soliman & Spooner, 2000) and behaviours (Medsker, Williams & Holahan, 1994), and relational models theory provides a deeper insight into relationships between

(9)

individual employees and organizations (Fiske, 1992). It is however first of all important to properly explain the theory behind the relational models described by Fiske (1991).

2.1 Relational models theory

According to relational models theory, people are fundamentally sociable (Fiske, 1992). The core idea of the theory is that moral judgment and behaviour depend on the type of relationship that people perceive themselves and their relational partner to be involved in (Rai & Fiske, 2011). Relational models theory stresses that social relationships are the construct that determines behaviour. This contradicts existing theories, which approach the socio-relational context as a potential bias for behaviour (Rai & Fiske, 2011). According to Fiske (1991), people use only four relational models to give meaning to all their relationships. These relational models are mental schemes, which determine the self-representations held by people (Fiske, 2001). Each model triggers different behaviour within social interactions, because they lead to different types of self-representations (Bridoux & Stoelhorst, forthcoming). Each self-representation then leads to different motivations, resulting in different behavioural rules (Bridoux & Stoelhorst, forthcoming). Even though complex relationships might lead to the employment of multiple models at the same time to navigate the different interactions, one of the models tends to dominate the relationship (Rai & Fiske, 2011). The four relational models proposed by Fiske (1991) are communal sharing, authority ranking, equality matching, and market pricing.

2.1.1 Communal Sharing

Communal sharing (CS) includes a relation of unity with a collective identity (Bridoux & Stoelhorst, forthcoming). People within such a relationship are perceived as being undifferentiated, while they experience differences with those who are not participating

(10)

within the relationship (Rai & Fiske, 2011). These differences with external groups often make participants within the relationship feel better than outsiders, which leads them to ignore individual differences within the relationship (Fiske, 1992). Those involved in the relationship treat each other in the same way based on shared aspects, and altruistic behaviour comes naturally as this feels like the right thing to do (Fiske, 1992). The relationship exists out of affective bonds and shared social values, and involves identity based trust (Mossholder, Richardson & Settoon, 2011). It shows similarities with the idea of clan control provided by Ouchi (1979), which emphasizes the importance of shared norms and values within the organization. An example of a CS relationship will be a family.

2.1.2 Authority Ranking

The second model is Authority Ranking (AR), which involves a relationship of asymmetric differences that exists within a hierarchical order of status (Bridoux & Stoelhorst, forthcoming). Such a relationship exposes relative individual positions within a hierarchy, since individuals are ranked along a particular social dimension (Rai & Fiske, 2011). The important factor within such a relationship is whether a person is above or below the other person (Fiske, 1992). Both options have their positive sides. Higher ranked individuals have amongst other things certain privileges, but lower ranked individuals possess a certain level of protection (Fiske, 1992). Thinking along the line of AR however in practice implies that people in higher ranks will be treated as superior to the lower ranked individuals. When people share a similar status, this type of relational model brings them to perceive themselves as being the same person, making the ordering perfectly linear (Fiske, 1992). AR can be related to the concept of hierarchy provided by Williamson (1975), which indicates that trust is based upon the status that is granted to an individual. An example of an AR relationship will be the vertical differentiation within the population of a military base.

(11)

2.1.3 Equality Matching

The third model is called Equality Matching (EM) and has reciprocity at its core (Bridoux & Stoelhorst, forthcoming). People will repay their relational partners similarly to what they receive from them in order to maintain a balance within their relationship, since they strongly value equality (Rai & Fiske, 2011). The main concern with EM type of relationships is maintaining this balance while keeping track of possible imbalances (Fiske, 1992). People compare themselves constantly with the other party involved to make sure the other receives the same amount as they do, and always know exactly how much inequality exists as well as how they can restore this (Fiske, 1992). They relate to each other based on equality, and build upon knowledge-based trust (Mossholder, Richardson & Settoon, 2011). An example of an EM relationship would be a car-pool agreement between two people, in which the two members have to provide the same amount of rides in their cars.

2.1.4 Market Pricing

The fourth and last model is called Market Pricing (MP). This model has clear economic roots since it centres on self-interest while people attend to ratios and rates (Fiske, 1992). Within a MP relationship, people use a single value to measure many differing factors. What matters most is the position in proportion to others (Fiske, 1992). Social value is thus determined by ratios and the relationship depends upon calculus-based trust (Mossholder, Richardson & Settoon, 2011). These relations use rates to compare while using a common metric (Rai & Fiske, 2011). The MP relational model can be linked to Williamson’s (1975) concept of market and transactional theories because of its reliance on ratios. An example of an MP relationship would be the relationship between an agent and a principal as described in agency theory (Eisenhardt, 1999).

(12)

2.1.5 Relational models and microfoundations

The four models describe the influence of individual relationships on behaviour, which implies that relational models theory goes a step further than the organizational forms provided by Ouchi (1980). Ouchi (1980) describes three organizational forms that influence organizational efficiency. Organizations function best through the hierarchy, market, or clan control depending on the level of performance ambiguity and goal incongruence (Ouchi, 1980). Compared to relational models, these organizational forms appear to be at the macro level of the organization. Relational models theory, on the other hand, could provide an explanation at the micro level. In line with the recent literature on microfundations in the strategy field (e.g. Foss, Felin & Ployhart, 2015), Coleman (1990) proposed a theory that explains the additional value of such a micro level. He stressed that there actually is no tangible macro level, since causality is always influenced by effects at the micro level (Coleman, 1990). Applying this perspective to the current subject of discussion implies the following: Ouchi (1980) describes organizational forms, which have a certain effect on organizational performance. In reality, the causal influence might actually be a result of the underlying relational models and their implications for employee behaviour, all happening at the micro level. This in turn allows organizational performance to improve or worsen due to the behaviour they evoke. This approach is in line with the ‘microfoundations at levels’ argument provided by Felin, Foss and Ployhart (2015). This states that microfoundations are about explaining the outcomes at a lower level than the explained phenomenon itself without denying the possible causal influence that can simultaneously occur at a higher level (Felin, Foss & Ployhart, 2015). This argument does however emphasize that the lower level phenomena, like in this case the relational models, should be involved in order to completely explain a causal relation (Felin, Foss & Ployhart, 2015).

(13)

The three organizational forms of Ouchi (1980) are well known and often discussed within strategy literature (Barney & Hesterly, 1996), but no study has tried to explain their influence by looking at relational models. This study will therefore contribute to the field by trying to determine whether a causal effect between relational models and individual-level value creation exists, in order to contribute to the macro-oriented existing literature from a microfoundational perspective.

2.2 Value creation as organizational citizenship behaviours

Value creation has received a lot of attention within the strategy field (e.g., Bowman & Ambrosini, 2000; Brandenburger & Stuart, 1996; Rumelt, 2003). To achieve value creation, it is important that employees in the organization are willing to contribute to the organizational goals (Bridoux & Stoelhorst, forthcoming). Employees however often face a collective good dilemma in which the pursuit of their individual goal conflicts with the pursuit of the organizational and thus collective goals (Ostrom, 2000; Bridoux, Coeurderoy & Durand, 2011). One form of value creation that depends very much on the behaviour of employees is organizational citizenship behaviour (OCB).

The concept of OCB can be defined in a variety of ways. The first definition, upon which the later ones build, was provided by Organ (1988). He stated that OCB is “ 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). A more updated definition of OCB, that nevertheless still has similarities with the former by Organ (1988), comes from Bolino, Turnley & Bloodgood (2002). They define OCB as “employee behaviours that go beyond role requirements, that are not directly or explicitly

(14)

recognized by the formal reward system, and that facilitate organizational functioning” (Bolino, Turnley & Bloodgood, 2002, p. 505).

Much research concerned with OCB has shown that OCB influences organizational effectiveness (e.g. Mackenzie, Podsakoff & Ahearne, 1998; Podsakoff et al., 2000; Walz & Niehoff, 1996; Koys, 2001). OCB has for example proven to enhance managerial productivity, to improve the availability and usability of resources, to enhance the stability of organizational performance, and to enhance the ability to adapt to environmental changes (Podsakoff et al., 2000). Podsakoff and MacKenzie (1994) concluded both employees’ tolerance of inconvenience in their jobs and their involvement in the organizational life significantly contributed to organizational performance. Walz and Niehoff (1996) furthermore described how the willingness to help colleagues contributed to workgroup performance within an organization.

OCB takes multiple forms like helping behaviour and organizational loyalty (Podsakoff et al., 2000). Two main forms that have been discussed often within the literature about OCB are OCB-O and OCB-I (Podsakoff et al., 2000; Skarlicki & Latham, 1997). OCB-O involves behaviour that benefits the organization as a whole (Williams & Anderson, 1991). This includes for example adhering to informal rules or giving notice when you are unable to attend (Podsakoff et al., 2000). OCB-I, on the other hand, is often referred to as altruism and involves behaviour that is beneficial for specific individuals (Williams & Anderson, 1991). Such personal influence in the end also contributes to the organizational well being (Williams & Anderson, 1991). OCB-I for example concerns the helping of colleagues or having a personal interest in colleagues (Podsakoff et al., 2000).

(15)

Considering the aim to study the impact of relational models on value creation, it seems interesting to use both OCB-O and OCB-I to measure value creation in the context of this study. The connection between relational models and OCB has never been researched before. As stated earlier, many studies have looked at the influence of OCB on organizational performance and possible antecedents leading to OCB (Podsakoff et al., 2000). Bolino, Turnley and Bloodgood (2002) have already signalled the lack of deeper insights in the theoretical mechanisms that could explain the empirically confirmed conclusions about OCB as an operationalization of organizational performance. Relational models could provide an explanation, since relationships are inherent to the concept of OCB. OCB-O is mainly built upon the relations between employees and the organization, while OCB-I is concerned with relations between employees that in the end benefit the organization (Podsakoff et al., 2000). The using of relational models theory might thus shed new light on the antecedents of OCB, in general, and OCB-O and OCB-I, in particular.

2.3 Trust in the organization

Studies about both OCB-O and OCB-I have revealed multiple variables that influence these specific categories of OCB (e.g. Feather & Rauter, 2004; Podsakoff et al., 2000). Besides organizational commitment and organizational identification (Feather & Rauter, 2004), trust in the organization is often recognized as having a significant influence on organizational behaviours (e.g. Mengue, 2000; Wech 2002; Dirks & Ferrin, 2001). Considering the incorporation of a specific form of trust in each of the relational models, it seems interesting to research how trust in the organization mediates the relationships between the relational models and OCB-O and OCB-I.

(16)

Trust has been operationalized and studied in many different ways (Mayer, Davis & Schoorman, 1995; Rousseau et al., 1998). When discussing trust in the organization from an employee perspective, an important aspect is the perceived trustworthiness of the organization (Mayer, Davis & Schoorman, 1995). Trust is based on the expectations about how another person, or in the present study how the organization, will behave (Lieberman, 1981; Rousseau et al., 1998). Two conditions must be present for trust to exist (Rousseau et al., 1998). The first is risk, which has a reciprocal relationship with trust. Risk is needed for trust to exist, and the existence of trust in its turn leads to risk taking (Rousseau et al., 1998). The second condition is interdependence, meaning that reliance upon another party is essential in order to succeed (Rousseau et al., 1998; Sheppard & Sherman, 1998). The form and depth of interdependence are believed to determine the amount of trust (Sheppard & Sherman, 1998).

The four relational models discussed all include both risk and interdependence, albeit in different ways and levels of intensity. MP implies a risk of receiving insufficient returns on the invested behaviour, EM has a risk of poor coordination and unbalanced reciprocation, and CS has the risk of misinterpreting others’ needs or experiencing empathetic inaccuracy (Mossholder, Richardson & Settoon, 2011). Mossholder, Richardson and Settoon (2011) do not discuss the last type of relationship, which is AR. Since an AR relationship is based upon the hierarchical positions of the trustor and the trustee, it can be argued there exists a risk when the performed behaviour can be observed poorly. This type of argument relates to the idea of agency theory that the hierarchy should only be preferred over the market when the behaviour within the organization can be observed sufficiently (Barney & Hesterley, 1996). Not being able to do so might lead to the occurrence of the wrong people in the wrong places of the organization, which can be potentially harmful for individuals within the organization as well as for the organization as a whole.

(17)

While MP can be argued to frame interdependence more as a burden due to its fundamental element of self-interest, CS relies on strong interdependence since this is needed to establish the relationship which is based upon affective and emotional bonds (Mossholder, Richardson & Settoon, 2011). EM approaches the level of interdependency of CS, but deals with this by using reciprocity as a measure for the appropriate behaviour. This means that the trusted party will be rewarded or punished depending on and in accordance with his or her own behaviour (Mossholder, Richardson & Settoon, 2011). AR is once again not included in the analysis of Mossholder, Richardson and Settoon (2011), but can be argued to involve interdependence since the hierarchy that lies at the core of this relational model can only exist when both parties fulfil their position within the hierarchy.

Despite the differences for each model, relational models theory in general builds upon what is referred to as relational trust. Such trust exists based on frequent interactions between two parties (Rousseau et al., 1998). The level of trust depends on the available information from within the relation itself. Such information is formed by previous interactions. Due to the emotion incorporated in these relationships as a result of frequent and longer-term interaction, relational trust is also referred to as affective trust and identity-based trust (Coleman, 1990). The different types of trust incorporated into the different relational models (Mossholder, Richardson & Settoon, 2011) could mean that trust in the organization differs depending on the relational model that is used. Sheppard and Sherman (1998) share this point of view by suggesting a similarity between their view of trust and the relational models theory proposed by Fiske (2001). Trust is in their view the acceptance of risk, which is associated with the amount of interdependence that is inherent to a relationship (Sheppard & Sherman, 1998). Sheppard and Sherman (1998) describe how trust varies as a result of the nature of the relationship. This brings up the question whether there could be a mediating role for

(18)

employees’ trust in the organization to explain the effect of relational models on organizational citizenship behaviours.

Investigating the potential influence of trust within the relationship however leads to yet another consideration. The focus of this study lies upon employees of organizations, and it is important to keep in mind that a workforce is composed of heterogeneous employees (Bridoux, Coeurderoy & Durand, 2011). These different types of employees might have different motives to cooperate. Acknowledging the heterogeneity of the workforce means taking into consideration that employees might experience different levels of trust in the organization for the same relational model due to moderating influences of their differing social dispositions.

2.4 Social dispositions

Agency theory is one of the most established theories discussing how cooperation between two parties should be approached (Barney & Hesterley, 1996). It starts from the assumptions that people are boundedly rational, self-interested, and potentially opportunistic (Barney & Hesterley, 1996; Eisenhardt, 1989). Agency theory solves the so called agency problems through the use of contracts, which can be both behaviour- or outcome-based depending on the observability of the work that is actually being done and the complexity of the environment (Eisenhardt, 1989). It however assumes that all people comply with the assumptions of the basis of agency theory, which restricts the applicability to a homogeneous workforce of self-interested individuals. With the rise of a microfoundational approach to organizations, the idea of a homogeneous workforce is nowadays being questioned (Bridoux, Coeurderoy & Durand, 2011), with more and more attention for individual-level factors (Felin, Foss & Ployhart, 2015). As Coff (1997) pointed out, the firm should be approached as

(19)

a nexus of contracts. This means that value creation happens at the level of organizational stakeholders, amongst which employees. This once again emphasizes the importance of considering relational models for the achievement of value creation, but also emphasizes the need to take into account differences between individuals.

Referring to the literature in behavioural economics, Ostrom (2000) already pointed out that not all individuals are self-regarding. A second important group of people can be described as being reciprocators. These reciprocators value the joint payoff more than self-regarding individuals, and they take the fairness of the distribution of the payoff more into account (Bridoux, Coeurderoy & Durand, 2011; Ostrom, 2000). Depending on their personal traits, which are believed to be stable, these different types of individuals engage in different types of behaviours (Bridoux, Coeurderoy & Durand, 2011). Individuals who are self-regarding aim to maximize their personal payoffs and are therefore generally unwilling to contribute to a common good (Bridoux, Coeurderoy & Durand, 2011; De Cremer & Van Lange, 2001; Fehr & Gächter, 2002). They are not concerned with the payoff of others, at least as long as these do not influence their own payoffs (Bridoux, Coeurderoy & Durand, 2011). Reciprocators on the other hand are concerned with their own payoff as well as the payoff of others who are involved, which means they essentially are willing to contribute to a common good (Bridoux, Coeurderoy & Durand, 2011; De Cremer & Van Lange, 2001; Fehr & Gächter, 2002). Within relationships, self-regarding individuals are only willing to cooperate if this contributes to their own payoff while reciprocators always cooperate at the beginning of the relationship (Fehr & Gächter, 2002). Reciprocators however stop cooperating when the other party involved is not cooperating either (Fehr & Gächter, 2002).

(20)

These differences between self-regarding individuals and reciprocators implicitly point towards different levels of trust, depending on the personal and apparently stable traits included in both types of individuals (Bridoux, Coeurderoy & Durand, 2011). Self-regarding individuals find it difficult to trust others, since they themselves cannot be trusted to contribute to the public good (De Cremer et al., 2001). Reciprocators on the other hand are willing to trust others at least in the beginning of the relationship. When others disregard their trust, they are however no longer willing to trust (De Cremer et al., 2001). They will furthermore stop acting in a trustworthy manner themselves, which can be expected to have implications for the relationship. It therefore seems plausible that the social dispositions of individuals can influence the potential effect of relational models on trust in the organization.

Former research has studied social dispositions (e.g., Ostrom, 2000; Bridoux, Coeurderoy & Durand, 2011; Fehr & Gächter, 2002) and has related them to the relational models proposed by Fiske (1991) (e.g., Haslam, 2004; Bridoux & Stoelhorst, 2005). De Cremer (2001) has more specifically drawn conclusions about the influence of social dispositions on trust, but no research has yet empirically studied the possible influence of social dispositions on any reactions to the relational models. Considering the suspected effect of social dispositions on trust and the incorporated aspect of trust within each of the relational models, it is interesting to include social dispositions in the study.

2.5 Research questions

There is a strong need to investigate the implications of relational models for value creation when looking at the theory discussed. Studying relational models and their possible effects on OCB will contribute to the macro-oriented existing literature about value creation from a microfoundational perspective. It will shed new light on the antecedents of OCB by looking at

(21)

the possible contribution of the individual level of relational models. The inclusion of both trust as a mediator and social dispositions as a moderator will take specific individual differences into account, thereby further specifying its results. Besides these theoretical contributions, this study will also contribute to the ability of managers to recognize and perhaps frame the relational models used by their employees. Understanding the relation between relational models and organizational citizenship behaviours, including possible effects of trust and social dispositions, could therefore contribute to the maximization of value creation. The current study aims to answer the following questions:

RQ1: How are the relational models as described by Fiske (1991) related to OCB-O and OCB-I?

RQ2: Does trust in the organization mediate this relationship?

RQ3: Is the influence of relational models on trust in the organization moderated by social dispositions of employees?

3. THEORETICAL FRAMEWORK

The following section will further describe the conceptual model, which will result in several hypotheses regarding the research questions. Expectations about the impact of each relational model on OCB-O and OCB-I, as well as the influence of both trust in the organization and social dispositions will be discussed below. The hypotheses that will be discussed are depicted in the following conceptual model:

(22)

Figure 1: Conceptual model

3.1 Relational models theory and OCB

Fiske (1991) proposed four relational models that operate as mental schemas on which individuals rely to determine the type of relationship they are engaged in and thus the appropriate behaviour. Rai and Fiske (2011) stated that even though multiple models can be present simultaneously, one model tends to dominate the relationship. This dominant model determines behaviour (Rai & Fiske, 2011), which is why the dominant model will be the main focus in the current study. As mentioned before, OCB is defined as the willingness of employees to make an extra effort that serves the wellbeing of the organization without receiving any explicit or direct reward for this benevolent behaviour (Bolino, Turnley & Bloodgood, 2002). Two types of OCB that have been researched specify such behaviour further. OCB-O involves behaviour that benefits the organization as a whole, while OCB-I digs deeper and looks at the overspill effect of the willingness of colleagues to individually help each other on the wellbeing of the organization (Williams & Anderson, 1991). OCB-O will be discussed first in the light of the relational models of Fiske (1991).

(23)

3.1.1 Relational models and OCB-O

The four relational models of Fiske (1991) include different levels of social focus. CS, AR, and EM type of relationships all imply the consideration of relational partners. In the case of CS, there is a focus on the relational partner since he or she is also responsible for the identity of the group (Fiske, 1991). The relational partners focus on their similarities in order to maintain the relationship. Both the organization and its employee are thus focused on what they have in common, and differentiate themselves from others based on these similarities. With AR, there is a social focus in the sense that the different positions within the hierarchy define the relationship (Fiske, 1991). A focus on the relational partner forms an understanding of the different roles within the relationship. The organization and the employee recognize each other’s positions within the hierarchy, and this focus determines their behaviour (Fiske, 1991). For EM, the social focus is different. There is a need for social focus in order to maintain the reciprocity on which the relationship is based. The organization needs to maintain a social focus on the employee in order to make sure it repays the employee in a similar way, and the employee needs to maintain a social focus on the organization to do the same (Fiske, 1991). These three relational models are expected to lead to OCB-O due to the social focus on the relational partner. MP on the other hand includes a purely self-interested vision (Fiske, 1991). MP relationships require a focus on the self in order to make sure the outcome is optimal for the self. The relational partner is not considered as an end in itself but rather as a means to reach something else that one values, and is expected to have the same egoistic focus (Fiske, 1991). This implies that the organization only focuses on itself when making a decision in relation to its employees, and the employees do the same (Fiske, 1991). This suggests that MP relationships will result in less OCB-O than CS, AR, and EM type of relationships.

(24)

H1: Relationships that are predominantly related to the MP relational model will lead to lower levels of OCB-O than relationships that are predominantly related to CS, AR, or EM relational models.

3.1.2 Relational models and OCB-I

OCB-I goes one step further than OCB-O. It includes beneficial results for specific individuals, which could then have a beneficial effect on the organization as a whole (Williams & Anderson, 1991). OCB-I in other words includes a spill over effect of the individual relationship between two employees on the organization as a whole. Cohesion has proven to be a significant predictor of engagement in OCB-I (Podsakoff et al., 2010; Shin & Choi, 2010). Shin and Choi (2010) found that the internal coherence of a group of individuals stimulates the engagement in OCB-I. When group members possess similar knowledge or affect, this increases their willingness to participate in extra-role behaviour (Shin & Choi, 2010). Such extra-role behaviour in the end is beneficial for the organization as a whole. Podsakoff et al. (2000) demonstrated a similar positive effect of cohesion on OCB-I. Cohesiveness within a group of employees led to, for example, altruism, courtesy, and conscientiousness, which are all elements that are part of OCB-I (Podsakoff et al., 2000). Relationships include a group of individuals, so the former results can be applied to relational models theory. This leads to the expectation that those relational models that include higher levels of cohesion between the relational partners will lead to higher levels of OCB-I than those relational models that include lower levels of cohesion.

When a person labels a relationship as CS, there is strong cohesion between the members who are included in the relationship (Rai & Fiske, 2011). This cohesion is formed by some similarities between the group members, and differences with those who are excluded from

(25)

the relationship (Bridoux & Stoelhorst, forthcoming). Cohesion is less present in relationships defined as AR or EM (Rai & Fiske, 2011), but these relational models are still characterized by adherence (Fiske, 1992). In the case of AR, a hierarchy between the involved individuals forms the relationship (Fiske, 1992). This implies that the relational partners depend on each other to keep the hierarchy intact. In the case of EM, reciprocity lies at the core of the relationship (Fiske, 1991). This implies that the input of one relational partner always influences the input of the other relational partner. The fourth and last model, which is MP, lacks any form of social coherence. MP is characterized by self-interest accompanied by ratios and rates (Fiske, 1992), which makes the main goal within such a relationship to be better than the other individuals involved. These observations lead to the following hypotheses:

H2a: Relationships that are predominantly related to the CS relational model will lead to higher levels of OCB-I than relationships that are predominantly related to the AR, EM, or MP relational models.

H2b: Relationships that are predominantly related to the MP relational model will lead to lower levels of OCB-I than relationships that are predominantly related to CS, AR, or EM relational models.

3.2 The mediating role of trust

Based on earlier studies it has been noticed in the former chapter that the different relational models proposed by Fiske (1991) include different levels of trust. Risk and interdependence are the two conditions that form the basis of trust (Rousseau et al., 1998), and especially this second condition can vary in form and depth (Sheppard & Sherman, 1998). Sheppard and

(26)

Sherman (1998) conclude that the level of interdependence and thus the level of trust varies depending on the characteristics of a relationship. The higher the amount of interdependence between the relational partners, the higher the amount of trust that can result from the relationship. Relational models theory describes four mental schemas or relational models, each with its own characteristics (Fiske, 1991). All models include risk, but the level of interdependence differs per relational model. Relational models are expected to have different impacts on both OCB-O and OCB-I through trust, as explained by the fluctuating levels of interdependence (Sheppard & Sherman, 1998). Trust in other words is expected to mediate the relationship between the four relational models and OCB-O and OCB-I.

3.2.1 The mediating role of trust for OCB-O

People involved in CS relationships acknowledge and embrace certain similarities between themselves and their relational partners (Fiske, 1991). At the same time, they emphasize the differences with those excluded from the relationship to stress the relational cohesion even further (Fiske, 1991). The relationship functions through the social norms inherent to the relationship and the willingness of the relational partners to behave according to those norms. Both the organization and the employee trust each other enough to rely on social norms instead of explicit rules for the existence and continuance of their relationship. This high level of trust between the members thus explains the impact of the CS relational model on OCB-O.

The AR relational model is characterized by a different level of interdependence. It functions on a hierarchical basis and it is accepted within the relationship that people in higher ranks are treated as superior to those in lower ranks (Fiske, 1992). The relational partners need each other in order for the relationship to exist, which implies that the relational partners are interdependent. The hierarchical aspect is believed to be important for the development and

(27)

existence of trust. Because a hierarchy exists, there is a clear structure. This idea is supported by agency theory, which proposes the use of the hierarchy as a controlling mechanism within an organization to determine appropriate rewards when the outcome of the behaviour is difficult to measure (Eisenhardt, 1989). Within the hierarchy, the places of the different individuals are clear and accepted by the relational partners involved (Fiske, 1992). It is therefore also commonly known what can be expected from the individuals involved in the hierarchy, and those involved are aware of the things they are expected to do. Such a clear role description is important for the functioning of the organization, and essential for its survival (Floyd & Lane, 2000). Well-defined roles furthermore lead to predictable behaviour, which results in a fluent transfer of information throughout the organization and increased trust within the organization (Floyd & Lane, 2000). It also has a positive effect on employee wellbeing (Floyd & Lane, 2000), which increases the willingness to do something extra for the sake of the organization. Within an AR relational model, the employee trusts the hierarchy to function appropriately due to the clear role descriptions resulting from the use of the hierarchy, and the AR relational model therefore has a positive impact on OCB-O through this level of trust.

Relationships in line with the EM relational model are based upon reciprocity (Fiske, 1992). The relational partners strive for equality, meaning that one partner will repay the other similarly to what he or she has received from him or her (Rai & Fiske, 2011). The presence of interdependence is essential within this setting, meaning that the EM relational model strongly relies on trust. Assuming that the organization is capable to encourage its employees to behave in a way that is desirable for the organization, this behaviour will be reinforced by the concept of reciprocity (Berg, Dickhaut & McCabe, 1995). Reciprocity is enforced when a positive social history develops, and this can be expected to be the case when an EM

(28)

relationship holds (Fiske, 1991). If the development is negative, the relational partners will most likely stop contributing to protect themselves from any harm. This assumption implies that the level of trust increases as the EM relationship continues (Berg, Dickhaut & McCabe, 1995). Within an EM relationship, relational partners will monitor each other to make sure they will get what they deserve. Due to the development of social history, the trust in the relational partner increases with the continuing of this tit-for-tat type of relationship (Berg, Dickhaut & McCabe, 1995). The organization and the employee will thus monitor each other, and the development of social history means they will know what they can expect from each other. An EM relationship will thus have an impact on OCB-O through the trust that is included in the reciprocity mechanism on which the relational model relies.

People involved in MP relationships are, as has been stated, driven by self-interest (Fiske, 1991). Like in agency theory, this does not only imply that people act accordingly to their own self-interest, but they will expect their relational partners to do the same (Eisenhardt, 1999). The relational partner is therefore expected to also be driven by self-interest. Agency theory takes into account that everyone is self-interested, which is why monitoring is necessary to prevent abuse of the reward system (Eisenhardt, 1999). It is therefore likely that those involved in a MP relationship will try to prevent any form of interdependence in order to have optimal control to ensure the most positive outcome for him- or herself. This lack of interdependence implies that trust cannot exist beyond calculative trust, since a level of interdependence is one of its basic requirements (Sheppard & Sherman, 1998). A lack of trust is therefore expected to explain a lower engagement in OCB-O for individuals who are involved in a MP relationship.

(29)

These considerations lead to hypothesize that the relational models CS, AR, and EM have a positive impact on OCB-O through trust, while the MP relational model is expected to have a negative influence on OCB-O through trust:

H3A: The mediating effect of trust leads to more OCB-O for relationships that are predominantly related to the CS, AR, or EM relational model, while the mediating effect of trust leads to less OCB-O for relationships that are predominantly related to the MP relational model.

3.2.2 The mediating role of trust for OCB-I

OCB-I goes deeper than OCB-O by including relationships at the individual level. For this type of OCB, the focus changes from the relationship between the organization and an employee to the relationship between two individual employees (Bolino, Turnley & Bloodgood, 2002). OCB-I is concerned with the overspill effect of the implications of the relationship between the organization and the individual employee on the relationship between two individual employees. Employees will use their relationship with the organization to give meaning to their relationship with another employee, which then determines the engagement in OCB-I. The organization and employee both trust each other enough to rely on social norms within a CS relationship. When employees feel they can rely on such norms when it comes to the organization, they will probably feel comfortable to rely on these same norms in their relationships with other employees of the same organization. A CS relationship thus also has an impact on OCB-I through trust. EM relationships have a similar result even though the way towards it is different. Employees that characterize their relationship with the organization as EM will have trust in a system of reciprocity. Because the organization indeed repays them similarly to what they provide, they will expect other

(30)

employees of the organization they work for to do the same. The relational partner or partners who will be returning the favour will furthermore compensate for the lack of recognition by the formal reward system, which makes it more appealing for employees to engage in OCB-I. EM relationships are therefore also expected to lead to OCB-I through the trust in the reciprocity mechanism that has been developed by the relationship with the organization.

For AR relationships, trust plays a different role. The relationship between the organization and the employee shows the employee that the hierarchical model can be trusted to guard the functioning of the organization. It provides clear structure and distinction between functions and rewards, from which both the employee and the organization benefit in their relationship. The experience from the relationship between the employee and the organization is however likely to be reapplied to the relationship with other employees, meaning one will trust the hierarchy to arrange properly what is needed for his or her colleagues. Therefore, people will feel fewer urges to engage in supporting behaviour for the direct wellbeing of co-workers through for example helping behaviour. Interdependence between colleagues is no longer an issue, since the hierarchy is expected to make sure that every relational partner gets what he or she deserves. AR relationships thus do not lead to OCB-I due to a lack of trust at the individual level. The importance of self-interest results in a similar lack of trust for MP relationships. Because an employee experiences the organization as being self-interested in a MP relationship, he or she will also believe their colleagues to be self-interested in their functioning. Trusting such a colleague would be considered unwise and even dangerous from a MP point of view, which is why OCB-I is unlikely to appear. These differences between the relational models lead to the following hypothesis:

(31)

H3B: For OCB-I, trust is a stronger mediator for relationships that are predominantly related to the CS or EM relational model than for relationships that are predominantly related to the AR or MP relational model.

3.3 The moderating role of social dispositions

As stated before, individual differences have to be considered as a factor of impact when looking at possible effects of relational models. With the rise of the microfoundational approach, the idea of a homogeneous workforce is being criticized (Bridoux, Coeurderoy & Durand). Attention is nowadays directed towards individual-level factors, with the idea that value creation happens at the level of organizational stakeholders (Coff, 1997). Acknowledging that groups of people are heterogeneous implies it has to be taken into account that the same models might not have the same implications for each individual employee. De Cremer et al. (2001) for example found a moderating effect of self-monitoring and accountability on trust. High self-monitors feel more need to comply with the dominant social norm, which is why they will be more willing to resolve conflicts by contributing to the common good (De Cremer et al., 2001). Conditions of high accountability have proven to convince those with low trust that others do contribute to the common good (De Cremer et al., 2001). This makes those low in trust to be more willing to contribute as well.

Former studies have made the distinction between self-regarding individuals and reciprocators (e.g. Bridoux, Coeurderoy & Durand, 2011; Ostrom, 2000). Self-regarding individuals are only concerned with their own pay-offs and are therefore not interested in contributing to a common good, while reciprocators do contribute since they are also concerned with the pay-off of others (Bridoux, Coeurderoy & Durand, 2011). Both types of individuals value different things, which in the end results in different behaviours (Bridoux, Coeurderoy &

(32)

Durand, 2011). Schwartz (1994) developed his values theory along this line of thinking. He describes ten basic values that can be captured along two dimensions. The first dimension represents the tension between self-enhancement and self-transcendence, while the second describes the tension between change and conversation (Schwartz, 1994). Especially the first dimension is important for this study. Self-enhancement can be described as having the tendency to overestimate one’s potential (Paulhus, 1998), which occurs most likely among people with high self-interest and self-esteem (John & Robins, 1994). Self-enhancement can therefore be linked to individuals who are labelled as self-regarding in former literature. Self-transcendence on the other hand occurs both internally and externally (Reed, 1991). It takes place internally through the engagement in introspective activities, and externally through concern about others (Reed, 1991). Especially the external form links the concept of self-transcendence to the type of individuals that are also defined as reciprocators.

Individuals high on self-enhancement have been found to most importantly value power, hedonism, and achievement (Schwartz, 1994). Power includes the social status of the individual as well as the amount of control over others and resources, while achievement stands for the importance of personal success and admiration (Schwartz, 1994). Hedonism involves the valuation of pleasure and gratification for oneself. Individuals high on self-transcendence tend to value especially universalism and benevolence (Schwartz, 1994). Universalism stands for the tendency to strive for social justice and tolerance, while benevolence implies the promotion of the welfare of others (Schwartz, 1994).

De Cremer et al. (2001) in addition found that self-regarding individuals find it more difficult to trust others than reciprocators, since they do not engage in trustworthy behaviour themselves. This result makes sense when looking at the different characteristics that are

(33)

commonly found amongst the different type of individuals. Power, hedonism, and achievement describe a need for control in order to gain personal success, pleasure, and gratification (Schwartz, 1994). Individuals high on self-enhancement will therefore find it more difficult to develop trust in the organization, which implies that the positive effect of the CS, AR, and EM relational model on trust will be smaller. In contrast, the higher employees score on self-enhancement, the stronger the urge for the self-interested behaviour that is inherent to the MP relational model, leading to lower levels of trust in the organization.

H4: The higher employees score on self-enhancement, the smaller the positive effect of CS, AR, and EM relational models on trust in the organization, and the stronger the negative effect of the MP relational model on trust in the organization.

The characteristics that are commonly found in individuals high on self-transcendence will have an opposed effect. Because they value social justice, tolerance, and the welfare of others, these individuals will find it easier to develop trust in the organization (Schwartz, 1994). These characteristics will strengthen the effect of the relational models that allow trust in the organization to exist. This can be expected to be the case for CS since these relationships rely upon social norms and values (Fiske, 1991). This social basis of the CS relationship requires, as explained before, trust in the relational partner in order for the relationship to exist. Taking Schwartz (1994) his values theory into account, it appears that the characteristics of CS relationships are very much in line with the characteristics of individuals who score high on self-transcendence. This is why the values that relate to self-transcendence are expected to enforce the effect of CS on trust in the organization. The same effect is expected for the AR model. Since these relationships function through the functionality of a hierarchy (Fiske, 1991), the relational partners need to trust in the functionality of the hierarchy in order for the

(34)

relationship to exist. AR relationships as explained earlier also imply a social focus on the relational partner, because this social focus on the relational partner informs people about their own position within the relationship as well. Those people who value social justice, tolerance, and the welfare of others are expected to be more comfortable with a social focus on their relational partner (Schwartz, 1994), and are therefore expected to have more trust in the organization when they are involved in an AR type of relationship. EM relational models are as explained build on the concept of reciprocity (Fiske, 1991). Within these relationships, an individual trusts on the assumption that he or she will receive back what he or she has provided the relational partner with (Fiske, 1991). Every action is in that sense met with a similar reaction, in order to retain a balance in the relationship (Fiske, 1991). Social justice seems to be at the basis of this type of relationships, as well as the welfare of others and tolerance, albeit in perhaps a more modest amount. This suggests that individuals who score high on self-transcendence can easily comply with the characteristics of the EM relational model, which would imply that for people who score high on self-transcendence the effect of trust in the organization will be enforced when they are involved in an EM type of relationship.

For MP, the expectation is slightly different. The characteristics of a person high on self-transcendence are expected to lower the negative effect of the MP relational model on trust in the organization. The values that are important to those who score high on self-transcendence are in conflict with the characteristics of MP. Self-transcendence emphasizes on social justice, tolerance, and the welfare of others (Schwartz, 1994), while MP relationships function on the basis of ratios, rates, and an egoistic perspective (Fiske, 1991). The social values are believed to provide some counterweight for the self-interested and economic nature on which the MP

(35)

relational model relies. Therefore, the negative effect of MP on trust in the organization is expected to be lower when individuals score high on self-transcendence.

H5: The higher employees score on self-transcendence, the larger the positive effect of CS, AR, and EM relational models on trust in the organization, and the smaller the negative effect of the MP relational model on trust in the organization.

4. RESEARCH DESIGN

This section explains how the research has been conducted in order to answer the proposed questions. It will start by explaining the chosen research methods, after which the scales that have been used to measure the variables will be discussed.

4.1 Research method: A survey

Testing the hypotheses required an explanatory study, since the study aimed to establish a causal relationship between variables (Saunders & Lewis, 2012). A survey has been spread online to collect the data needed to answer the research questions. Surveys are suited to collect data about the same topics from a relatively large amount of respondents (Saunders & Lewis, 2012). Because the survey exists out of standardized questions, it is well suited for explanatory research (Saunders & Lewis, 2012). A downside, however, is that the collected data is not as detailed as would be possible when using other strategies such as interviews (Saunders & Lewis, 2012).

In order collect data from as many respondents as possible in a limited amount of time, the survey has been administered online. People can participate whenever it suits them and at a location that is comfortable for them, which are both advantages of the use of an online

(36)

survey (Saunders & Lewis, 2012). This distant way of collecting participants however also implies that people can forget more easily about the request to participate, or they might misjudge their fit with the participant profile (Saunders & Lewis, 2012). Given the small timeframe and the large reach, an online survey is however believed to be the best method for data collection regarding this study.

The personal networks of the researcher and the other members of the thesis group have been used for distribution of the survey, leading to a convenience sample. Data has been collected collectively since this helped to reach a larger amount of respondents in the limited amount of time that was available for writing the thesis. The survey was written in Dutch, since this is the native language of most people in the social network of the thesis group. People understand questions easier and proceed faster when the questions are asked in their native language.

4.2 Sample

Participants had to be collected in a limited amount of time. Therefore, there has been made use of a non-probability sampling method (Saunders & Lewis, 2012). Specifically, there has been made use of self-selection. This sampling method implies that a participant profile is distributed, which aims to attract participants that fit the required profile (Saunders & Lewis, 2012). These prospective participants should recognize themselves in the given description, leading to their participation in the survey. This selection method asks for clear selection criteria (Saunders & Lewis, 2012). First, since the thesis is concerned with employees, an important criterion was that only people who were working in an organization at the time of answering the survey could participate. Second, only those people who were seventeen years or older at the time of answering could participate, since this is legally the minimum age for

(37)

people to work in the Netherlands. Third, participants needed to possess sufficient knowledge of the Dutch language since the survey was in Dutch. These selection criteria were made clear in the participant profile, so that only those people who adhered to these criteria would select themselves to participate in the survey.

The search for new participants was planned to continue until at least 200 participants had completed the survey. After a period of two weeks, 272 respondents had participated in the survey. A number of 224 participants worked more than 24 hours, and 203 of them had completely finished the survey, which determined the final sample size.

4.3 Operationalization of the variables

The survey included the variables relational models, trust in the organization, OCB-O, OCB-I, and social dispositions (self-enhancement and self-transcendence). Besides these, organizational commitment was included since fellow group members needed this variable for their analysis. Organizational commitment has been excluded from the data analysis of this thesis since it was not included in any of the research questions or hypotheses. Last but not least, the survey included several control variables. The control variables included in the survey were gender, age, highest level of education, the amount of hours that the employee is working for the organization, type of employment, the amount of employees working in the organization, years of employment in the current organization, and the amount of hours that are actually spend at the organization. The survey that was administered is included in the Appendix.

Gender, age, and highest level of education are generally included as control variables in many studies (Saunders & Lewis, 2012). They are however of specific interest for this study

(38)

for the following reasons. Gender was included as a control variable since former research has suggested that women are more likely to score high on self-transcendence, while men are more likely to score high on self-enhancement (Schwartz & Rubel, 2005). Age was included because older employees have shown to engage in more OCB than younger employees, since older employees perceive OCB more to be in-role behaviour than younger employees (Wanxian & Weiwu, 2007). Level of education was taken into account since employees with higher educational levels appear to engage more in OCB than employees with lower educational levels (Lambert, Hogan & Griffin, 2008). This difference was caused by differences in job type that often result from a difference in educational level. Those with higher levels of education have shown to be more likely to end up in supervising functions (Lambert, Hogan & Griffin, 2008; Stack, 2002). These potentially different job characteristics were also reason to include type of employment as a control variable.

The other control variables are included because this study is concerned with relational models between employees and the organization. The amount of hours that the employee is working for the organization is of importance since frequent interaction with the organization is needed to allow for the relationship to exist. Therefore, respondents who are employed less then 24 hours per week were excluded from the survey. However, when an employee works more than 24 hours per week for an organization, this does not necessarily imply that the employee spends these hours at the organization. Employees might be able to do most of their job at home or might be travelling for the entire time, and this can be expected to influence their relationship with the organization. It could also deprive them of the opportunity to perform the types of OCB measured in this study. Therefore, the amount of hours that are actually spent at the organization was also included as a control variable. Years of employment in the current organization is of similar importance, since the enduring

Relational models and value creation : exploring the effect of relational models on organizational citizenship behaviours and discovering the mediating role of trust and the moderating role of social dispositions