HPWS and their relationship with well-being and performance : the mediating role of job crafting

(1)

Master Thesis

High Performance Work Systems and Their Relationship with Employee

Well-being and Performance: The Mediating Role of Job Crafting

Name: Josephine Selchow

Study Programme: MSc in Business Administration – Leadership and Management track University: University of Amsterdam (UvA)

Student number: 10828419

Professor: dr. C.T. Boon

(2)

Statement of Originality

This document is written by student Josephine Selchow who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(3)

Abstract

Given that work contexts are changing and jobs are becoming increasingly demanding, employees need to find ways to deal with these changes. Job crafting represents a promising behaviour that allows employees to actively shape their jobs according to their personal needs and abilities. Consequently, there is a need to better understand how job crafting behaviour among employees can be actively initiated by organisations. This study aims to examine whether high performance work systems are related to job crafting, as well as whether the latter additionally explains the relationship between high performance work systems and employee well-being and performance. The results of our regression analysis show that high performance work systems are positively related to most job crafting behaviours. In addition, several job crafting behaviours are related to performance and job satisfaction. One type of job crafting also mediates the relationship between high performance work systems and performance, although, we could not find mediation between high performance work systems and employee well-being. These findings contribute to the understanding of the antecedents and outcomes of job crafting, as well as the understanding concerning the link between high performance work systems and employee outcomes.

Keywords: job crafting; high performance work systems; well-being; stress; performance; job satisfaction

(4)

List of Tables and Figures

Tables

Table 4.1 Means, standard deviations, correlations, and reliability scores

Table 4.2 Results of HPWS as predictor of well-being (stress and job satisfaction) and performance

Table 4.3 Results of HPWS as predictor of increasing structural job resources and increasing social job resources

Table 4.4 Results of HPWS as predictor of increasing challenging job demands and decreasing hindering job demands

Table 4.5 Results of job crafting as predictor of well-being (stress and job satisfaction) and performance

Table 4.6 Results of job crafting as a mediator in the relationship between HPWS and job satisfaction

Table 4.7 Bootstrapping results of job crafting as a mediator in the relationship between HPWS and job satisfaction

Table 4.8 Bootstrapping results of job crafting as a mediator in the relationship between HPWS and stress

Table 4.9 Bootstrapping results of job crafting as a mediator in the relationship between HPWS and performance

Figures

Figure 1 Research model

(6)

Index of Appendices

Appendix A Questionnaires (managers and employees) Appendix B Emails (managers and employees)

(7)

1. Introduction

In recent years, work contexts have changed and jobs have become increasingly complex. These developments have occurred owing to various changes, such as technological breakthroughs, the transformation from a manufacturing to a service-oriented economy and the increasing importance of employee knowledge, resulting in rising demands concerning the cognitive abilities of workers (Grant, Fried, Parker, & Frese, 2010). Furthermore, organisations nowadays depend on their employees more than ever before, as the quality of the human capital pool has become crucial to the success of each organisation (Lu, Wang, Lu, Du, & Bakker, 2014).

In order to address these developments, a new concept has recently emerged from job design literature, which is referred to as job crafting (Wrzesniewski & Dutton, 2001). Whereas job design typically describes a top-down approach, i.e. managers designing the jobs for employees, job crafting refers to a bottom-up approach, i.e. employees taking the initiative to proactively craft their own jobs (Berg, Wrzesniewski, & Dutton, 2010; Demerouti, 2014). Job crafting is a new answer to changing job conditions and their unique constellation (Demerouti, 2014; Petrou, Demerouti, Peeters, Schaufeli, & Hetland, 2012). It can be defined as employees proactively shaping their jobs by altering the job characteristics, i.e. job demands and job resources (Bakker & Demerouti, 2007; Tims & Bakker, 2010). The emergence of the job crafting concept is in line with Frese and Fay (2001), who have highlighted the importance of employees taking personal initiative in shaping their job characteristics.

Due to its recent emergence, job crafting research remains in its infancy (Demerouti, 2014). Existing work has focused on the predictors (e.g. Bakker, Tims, & Derks, 2012; Wrzesniewski & Dutton, 2001) and outcomes (e.g. Chen, Yen, & Tsai, 2014; Tims, Bakker,

(8)

& Derks, 2013) of job crafting behaviour. Situational and individual predictors of job crafting include decision latitude (Wrzesniewski & Dutton, 2001), task complexity (Demerouti, 2014), proactive personality (Bakker, Tims, & Derks, 2012) and work engagement (Lu et al., 2014).

However, it has rarely been considered to date how job crafting can be initiated on purpose and how this affects employees. Tims and Bakker (2010) emphasised a need in this field to identify how organisations are able to stimulate employees to engage in job crafting. In line with this, Demerouti (2014) pointed out that organisations could indeed initiate and support job crafting behaviour. Tims and Bakker (2010) proposed tailored feedback as an organisation’s human resource management (HRM) practice that might be able to foster job crafting, although they did not test this assumption. Therefore, this paper aims to further develop the proposition by Tims and Bakker (2010) by examining whether HRM practices are related to job crafting among employees. Rather than using an individual HRM practice, we will use a bundle of HRM practices, frequently referred to as high performance work systems (HPWS) (Takeuchi, Lepak, Wang, & Takeuchi, 2007). This choice is in line with current strategic HRM literature and builds upon the assumption that HRM practices do not work individually and in isolation but rather together as a bundle that influences employees (Jiang, Lepak, Han, Hong, Kim, & Winkler, 2012). We expect HPWS to be associated with job crafting given that HPWS are able to foster proactive behaviour (Arefin, Arif, & Raquib, 2015). Job crafting represents a proactive kind of work behaviour (Parker, Bindl, & Strauss, 2010) and thus we expect HPWS to also have a positive influence on job crafting.

Furthermore, this paper takes into account two proven outcomes of job crafting, namely employee well-being and performance (Bakker, Tims, & Derks, 2012; Tims, Bakker, & Derks, 2013). Well-being will be divided into job satisfaction and stress. Scholars have developed a consensus that HPWS lead to increased individual performance (Ang, Bartram,

(9)

McNeil, Leggat, & Stanton, 2013). However, it is still discussed which costs are associated with this link at the same time. In fact, well-being outcomes have been discussed frequently and controversially in the context of HPWS (e.g. Ehrnrooth & Björkman, 2012; Karatepe, 2013; Van de Voorde, Paauwe, & Van Veldhoven, 2012). Whereas some scholars support a positive effect of HPWS on employee well-being (e.g. García-Chas, Neira-Fontela, & Castro-Casal, 2014), others demonstrate a negative effect (e.g. Kroon, Van de Voorde, & Van Veldhoven, 2009). However, as a device for employees to proactively change the work characteristics of their job, job crafting may provide an explanation concerning why HPWS have a positive indirect effect on employees’ well-being. This paper thus argues that HPWS may be designed in a way that fosters job crafting among employees, whereby job crafting behaviour may eventually lead to the increased performance and well-being of employees. As a result, this paper will examine the following research question:

Does job crafting mediate the relationship between HPWS and employee well-being and performance?

With this research question, the paper aims to extend previous work on job crafting and HPWS by adopting a theoretical model that contributes to the understanding of how organisations may initiate job crafting behaviour among employees by using HPWS. Additionally, this paper aims to contribute to the issue of how HPWS may be able to lead to mutual gains for both the employees and the organisation, with several scholars arguing that an increased well-being and performance of employees is not only beneficial for the employees themselves but also for organisational performance (Ehrnrooth & Björkman, 2012; Van de Voorde, Paauwe, & Van Veldhoven, 2012). Finally, we want to contribute to the understanding of job crafting as a specific type of proactive behaviour and additionally provide an example of how HPWS might affect this specific type of proactive behaviour.

(10)

The research model is illustrated in figure 1 and suggests that job crafting mediates the relationship between HPWS and employee well-being and performance.

Figure 1: Research Model

Following, the study will draw an overview of job crafting, HPWS, employee well-being and performance and develop the hypotheses. The job-demands-resources (JD-R) model will be used as the main theoretical foundation. Subsequently, we will describe the methodology and results and end with a discussion of the results.

2. Theoretical Background

2.1 Conceptualisation of Job Crafting

According to Wrzesniewski and Dutton (2001), job crafting behaviour can be defined as employees shaping either the task boundaries of the job or the relational boundaries of the job, or indeed both simultaneously. First, shaping task boundaries can be achieved both physically and cognitively. A physical manner refers to changing either the form or number of activities that an employee undertakes while doing his/her job, while a cognitive manner refers to modifications made to change how an employee sees the job. Second, altering relational boundaries refers to employees having the freedom to decide with whom to interact while doing the job. By engaging in one of these behaviours, employees actively modify the job as well as the social environment in which they execute their work

HPWS Job crafting Employee satisfaction Employee performance Employee stress + + + -

(11)

(Demerouti, 2014). While the conceptualisation of Wrzesniewski and Dutton (2001) represents one possible definition of job crafting, there is an alternative stream of research that frames the conceptualisation and definition of job crafting in the JD-R model (Bakker & Demerouti, 2007; Bakker, Tims, & Derks, 2012; Tims, Bakker, & Derks, 2012). The JD-R model categorises all job characteristics into two groups: job demands and job resources. Using this frame, job crafting can be defined as employees changing their job demands and job resources to balance them with their personal abilities and needs (Demerouti, 2014). However, job crafting as a proactive behaviour also has its boundaries, given that changes can only be attained within the context of their prescribed jobs, which are characterised by prescribed expectations, tasks and positions in the organisational hierarchy (Berg, Wrzesniewski, & Dutton, 2010).

In this study, we will proceed with and further explain the conceptualisation of job crafting as framed in the JD-R model for the following reasons. First, Wrzesniewski and Dutton’s (2001) conceptualisation is limited to shaping task and relational boundaries of the job. However, recent scholars have emphasised the possibility of other forms of job crafting such as self-initiated skill development (Lyons, 2008). In order to also capture these forms of job crafting, we will thus make use of the JD-R conceptualisation of job crafting (Tims, Bakker, & Derks, 2012). Second, this choice is in line with recent scholars in this field (e.g. Tims & Bakker, 2010).

2.1.1 Job Crafting and the JD-R Model

As already mentioned, employees craft their job by making changes to their job resources and job demands. Job resources refer to those job characteristics that positively contribute to the employees by supporting them in achieving their work goals, reducing the job demands, the physiological and psychological costs associated with it, as well as fostering personal development and growth (Bakker & Demerouti, 2007; Tims & Bakker,

(12)

2010). Examples of job resources are autonomy, feedback and skill variety (Tims, Bakker, & Derks, 2014). Job resources comprise a motivational process, which states that job resources have motivational potential and thus are able to animate employees to achieve their goals (Demerouti, Bakker, & Fried, 2012). In other words, if employees have enough resources available, they are more motivated, which might subsequently lead to higher work engagement as well as advantageous work outcomes (Ângelo & Chambel, 2014) such as goal attainment (Schaufeli, Bakker, & Van Rhenen, 2009). Besides fostering goal attainment, job resources also motivate because they increase employees’ growth, learning and development. For example, feedback motivates learning among employees and thereby increases the job competence of employees (Schaufeli, Bakker, & Van Rhenen, 2009).

On the other hand, job demands refer to those job characteristics that are associated with certain costs for the employee (Tims, Bakker, & Derks, 2013). They refer to “those physical, social or organisational aspects of the job that require sustained physical or mental effort and are therefore associated with certain physiological and psychological costs” (Tims & Bakker, 2010:3). Job demands include a variety of different job characteristics such as high work pressure and interactions with clients that are emotionally demanding (Tims & Bakker, 2010). Research on job demands has found contradicting results, with one stream of scholars referring to merely negative effects of various job demands on employees such as increased exhaustion, cynicism and decreased work engagement (Demerouti et al. 2012; Van den Broeck, De Cuyper, De Witte, & Vansteenkiste, 2010). These effects support the underlying theory of the JD-R model, which is referred to as a health impairment process of job demands. The health impairment process states that chronically high job demands may exhaust employees’ resources, thus potentially leading to health problems (Demerouti et al.,

2012). However, the other stream describes that several job demands simultaneously lead to positive effects on employees. For example, scholars have found that cognitive demands are

(13)

positively related to vigour and dedication (Bakker, Demerouti, & Schaufeli, 2005). These contradicting results prompted scholars to develop a distinction of job demands, which will be described in chapter 2.1.2.

Next to the motivational process of job resources and the health impairment process of job demands, an interaction effect also exists between job resources and job demands, which is also referred to as the “buffer effect” of job resources. This process assumes that job resources are able to buffer the negative effects of job demands on employees’ well-being (Bakker & Demerouti, 2007). Thus, a high level of job resources can actually outweigh the effects of high job demands for example on employee burnout (Bakker, Demerouti, & Euwema, 2005).

2.1.2 Dimensions of Job Crafting

Building upon the aforementioned distinction of job resources and job demands, job crafting can be conceptualised into four different dimensions (Tims, Bakker, & Derks, 2012): two dimensions focus on crafting job demands, whereas the other two focus on crafting job resources.

In order to address the described contradicting effects of job demands on employees, researchers such as Van den Broeck and colleagues (2010) employ a distinction between two types of job demands, namely, challenging and hindering job demands.

First, job crafters increase challenging job demands (Tims, Bakker, & Derks, 2012). Challenging job demands describe those demands that lead to both positive and negative outcomes (Van den Broeck et al., 2010). Employees perceive them as positive because these demands are advantageous for their personal development and progress if they engage in them (Tims, Bakker, & Derks, 2013). Challenging job demands include time pressure and workload (Tims & Bakker, 2010). Although employees need to engage in extra efforts to meet the requirements of challenging job demands, employees like to engage in them due to

(14)

the positive effect on their own progress and development (Tims & Bakker, 2010). However, challenging job demands also exhaust employees’ energy due to the higher effort necessary, thus positively relating to ill-health. Thus, challenging job demands can be seen as a double-edged sword: on the one side, they are positively related to well-being such as job satisfaction, although on the other side, they are also related to ill-health such as stress and anxiety (Van den Broeck et al., 2010).

Second, job crafters decrease hindering job demands owing to the negative effects that such demands have on employees (Tims, Bakker, & Derks, 2012). Hindering job demands are such demands that hinder employees in their personal development and functional ability and thus only have negative consequences for employees. Examples of hindering job demands are role ambiguity, role conflict and job insecurity (Tims & Bakker, 2010). Employees react negatively when being confronted with hindering job demands (LePine, Podsakoff, & LePine, 2005). While they may initially be able to handle hindering job demands by increasing the use of job resources, if hindering job demands remain for a longer period, employees need to engage in other coping reactions such as withdrawal from that work (Schaufeli & Taris, 2005). Hindering job demands are positively related to ill-health, such as emotional exhaustion and anxiety (Van den Broeck et al., 2010).

The two dimensions of job crafting that are focused on job resources are divided into increasing structural job resources and increasing social job resources. Increasing structural job resources refers to increasing the variety of resources available, the opportunity for development and job autonomy. Increasing social resources refers to increasing the amount of feedback, social and supervisor support (Tims, Bakker, & Derks, 2012). Job resources represent an important tool to deal with job demands. Moreover, they are positively related to advantageous work outcomes such as work engagement and commitment (Tims & Bakker, 2010).

(15)

According to job crafting theory, employees can craft all four dimensions of job crafting (e.g. Tims, Bakker, & Derks, 2012). The job crafting process includes making changes regarding job characteristics to achieve a certain goal and strive for a future-focused state. In this sense, job crafting can be seen as a specific kind of proactive behaviour (Parker, Bindl, & Strauss, 2010; Slemp & Vella-Brodrick, 2014).

Employees engage in job crafting for the following two reasons. First, employees craft to achieve goals and envisioned states in the future. The more that these goals are central to the identity or values of the employees, the more motivated they will be in attaining these goals and their envisioned future (Demerouti, 2014). Second, employees craft their jobs because they want to achieve a positive work identity and greater meaning of work. Indeed, by crafting their jobs, employees can achieve another purpose of their work and experience their job in a new and different way (Wrzesniewski & Dutton, 2001).

2.1.3 Antecedents and Outcomes of Job Crafting

While research on job crafting remains scarce, it is consistent in supporting mainly positive effects of job crafting on employees and organisations. For instance, Bakker and colleagues (2012) found that job crafting acts as a tool for employees to remain engaged in their work, as it was positively related to in-role performance and work engagement. However, hindering job demands was not supported to increase work engagement. In addition, Tims, Bakker and Derks (2013) found that crafting job resources positively affects work engagement and negatively affects burnout. Finally, Lu and colleagues (2014) found that employees who craft their work achieve a better person-job fit.

Nevertheless, job crafting research is still in its infancy, which is particularly noticeable if we consider the antecedents of job crafting. Research has identified several individual and situational predictors such as decision latitude (Wrzesniewski & Dutton, 2001), task complexity (Demerouti, 2014), proactive personality (Bakker, Tims, & Derks,

(16)

2012) and work engagement (Lu et al., 2014). These are very specific and narrow antecedents and they do not offer an answer to the question of whether and how organisations can actively initiate job crafting. However, in line with Demerouti (2014), we argue that organisations might be able to actively foster job crafting behaviour. Building upon the proposition of Tims and Bakker (2010) to use HRM practices for this purpose, we argue that HPWS might work as an important antecedent of job crafting. In the following, we will describe HPWS and explain why and how they might relate to job crafting.

2.2 High Performance Work Systems

In order to understand the relationship between HPWS and job crafting, it is necessary to first consider how HPWS have developed from the strategic HRM literature and how they are defined.

2.2.1 High Performance Work Systems and Strategic HRM

Human resource management (HRM) describes the overall process of managing employment relations in an organisation. Recently, there has been a switch to strategic HRM, stating that HRM undertakes strategic decisions regarding its employee management (Boxall & Purcell, 2000). Within strategic HRM, an approach to combine individual HRM practices to systems has evolved, building upon the assumption that HRM practices do not work individually and in isolation but rather together as a bundle that influences employees (Jiang et al., 2012). However, HRM systems significantly vary throughout the literature. They differ in their names as well as in the number and kind of HRM practices used in one system (Guthrie, 2001; Truss, 2001). Scholars have used names such as high involvement work systems (Guthrie, 2001), high performance work systems (Becker & Huselid, 1998) and high commitment work systems (Xiao & Björkman, 2006). Moreover, there are different kinds of HRM practices included. In fact, Delaney and Huselid (1996) highlighted that there

(17)

are rarely two studies measuring the same set of HRM practices. In this study, we decided to focus on HPWS as a conceptualisation of HRM systems. HPWS usually include HRM practices such as selective recruitment and staffing procedures, extensive employee involvement and training, extensive benefits, incentive compensation and performance management systems (Takeuchi, Chen, & Lepak, 2009; Takeuchi et al., 2007).

HPWS have been frequently used for studies in the field of strategic HRM (e.g. Sun, Aryee, & Law, 2007; Takeuchi et al., 2009) and thus they are suitable for this study. Furthermore, this paper chose to focus on HPWS as perceived by employees. This choice is in line with current literature stating that managers’ perceptions of HPWS deviate from what

employees actually perceive (Liao, Toya, Lepak, & Hong, 2009). Indeed, employee perceptions of HPWS follow managers’ HPWS implementation and thus are more predictive of employees’ attitudinal and behavioural outcomes (Kehoe & Wright, 2013), i.e. job

crafting, well-being and performance. Thus, the way in which employees perceive HPWS is crucial in determining the effect of HPWS on job crafting and eventually on well-being and performance.

However, organisations often make crucial mistakes regarding their HPWS. Either their HRM practices only consist on paper and are thus poorly understood or even out-dated, or, the implementation itself is not executed properly. For instance, line managers who are responsible for the implementation of HRM practices might differ in the extent to which they implement these practices consistently. However, if not implemented consistently, HPWS cannot lead to the intended employee outcomes (Sikora & Ferris, 2014), e.g. to proactive behaviours of employees (Arefin, Arif, & Raquib, 2015). This would not only be disadvantageous because organisations do not accomplish the intended employee outcomes, but also because this could negatively affect organisational performance. Many scholars have emphasised the importance of human factors in the causal link between HPWS and

(18)

organisational performance (Takeuchi et al., 2007). In fact, employee behaviour has been considered as an important mediator in the HPWS – organisational performance link (Huselid, 1995).

In the following, we will elaborate on the relationship between HPWS as perceived by employees and job crafting.

2.2.2 High Performance Work Systems and Job Crafting

HPWS can be seen as a tool to foster desired employee behaviour (Truss, 2001). They provide employees the motivation to perform by offering extensive training, development opportunities and merit-based promotion opportunities. As a result, employees feel that the organisation is supportive and invests in them (Kehoe & Wright, 2013; Sun, Aryee, & Law, 2007). To draw an example, employees might feel that the organisation highly values them given that they select and recruit new employees in a rigorous way. Furthermore, performance appraisal systems might give the employee the feeling of being valued and appraised appropriately for their effort (Takeuchi et al., 2007). Using social exchange theory, employees may subsequently feel that they need to reciprocate behaviour that benefits the organisation. Social exchange can be defined as “favo[u]rs that create diffuse future obligations, not precisely specified ones, and the nature of the returns cannot be bargained about but must be left to the discretion of the one who makes it” (Blau, 1964:93, as qtd. in Takeuchi et al., 2007). In the context of the employer-employee relationship, social exchange theory can be defined as a norm of reciprocity. According to this norm, employees would not harm their employer, from whom they have received certain benefits (Sun, Aryee, & Law, 2007). Thus, an employee may decide to work harder to demonstrate the organisation that it was worth investing in the employee. One way of working harder may be actively searching for new and additional challenging tasks to engage in. If employees choose to actively search for new challenges, they thereby increase

(19)

their workload and their challenging job demands. However, it is important to emphasise that HPWS can only cause reciprocal behaviours of employees if the HPWS implemented by managers are also perceived and recognised by employees. Otherwise, employees might not feel that they need to return the favours the organisations offer them. To summarise, we suggest that employees increase their challenging job demands if they perceive the presence of HPWS.

HPWS also motivate employees to actively define their own work roles by empowering them to do so. Perceptions of empowerment are created through measures such as high levels of autonomy and performance-based rewards (Evans & Davis, 2005). Job autonomy is reflected by characteristics such as job discretion, employees’ ability to select methods to complete tasks and the freedom to schedule their work (Tai & Liu, 2007). Autonomy can be seen as an important job resource that HPWS offer employees. For instance, if employees are confronted with hindering job demands, autonomy may enable employees to proactively decrease these demands to reduce perceived strain. This behaviour would also be in line with self-regulation theory (Lord, Diefendorff, Schmidt, & Hall, 2010), which argues that employees who set a goal and strive for its attainment compare their actual goal attainment state against their anticipated goal attainment state. If there is a discrepancy between the current and anticipated goal attainment state, employees will make changes, e.g. in their behaviour (Lord et al., 2010). If employees who strive for a goal realise that they are imposed to certain hindering job demands, which hinder them in attaining their goal as planned, they might use the autonomy provided by HPWS to decrease these hindering job demands and increase their ability to attain their goals. Again, it is important to emphasize that this employee behaviour can only occur if employees actually perceive that HPWS offer them autonomy. Thus, we expect employees to decrease their hindering job demands in the presence of HPWS.

(20)

Building upon the statements above, structural resources such as autonomy might help employees to handle hindering job demands and attain their goals. Indeed, job resources are essential to deal with high job demands (Tims & Bakker, 2010). For example, research has found that social resources (e.g. social support) diminish the relationship between burnout and demands (Leiter & Maslach, 1999). Although HPWS provide employees with resources such as autonomy and feedback, these resources are limited and other important resources might also be needed. Moreover, as described above, HPWS motivate employees to engage in additional challenging job demands. Although challenging job demands are positive for employees’ growth and development (Tims, Bakker, & Derks, 2013), they also

bear negative effects such as stress and anxiety (Van den Broeck et al., 2010). As a result, employees might need to increase their job resources to deal with and buffer the negative effects of an increased number of challenging job demands.

This would be in line with the conservation of resources (COR) theory (Hobfoll, 1989) which assumes that employees invest in resources to deal with situations that are threatening or to prevent negative consequences (Ângelo & Chambel, 2014). In general, COR theory assumes that employees engage in attaining resources that they do not have, try to retain and protect the resources that they possess, as well as to foster resources by positioning themselves in a way that allows their optimal use (Ângelo & Chambel, 2014, Hobfoll, 1989). As a result, this paper suggests that employees actively engage in increasing both structural and social job resources in the presence of HPWS.

Hypothesis 1: HPWS are positively related to job crafting (all four dimensions).

2.3 Well-being

According to Van de Voorde and colleagues (2012), well-being can be categorised into the following three types of being: happiness, health, and relationship-related

(21)

well-being. This study will focus only on happiness and health as types of well-being, as this is also in line with other job crafting studies (e.g. Tims, Bakker, & Derks, 2013).

First, happiness well-being refers to the subjective experiences of employees and the functioning at work. Employee satisfaction would be an example to describe this type of well-being (Van de Voorde et al., 2012). Satisfaction will also be examined in this paper, given that it represents the most commonly used way of operationalizing well-being (Tims, Bakker, & Derks, 2013).

Scholars have identified a positive relationship between challenging job demands and job satisfaction (e.g. Cavanaugh, Boswell, Roehling, & Boudreau, 2000; Podsakoff, LePine, & LePine, 2007). Challenging job demands represent positive demands that are related to goal attainment and work motivation (Cavanaugh et al., 2000). They increase the effort of employees and thus may result in higher satisfaction when a goal has been attained (Tims & Bakker, 2010). This is in line with goal-setting theory, which builds upon the assumption that employees have certain difficult goals. If employees perform and achieve these goals, they eventually feel satisfied and are willing to engage in new challenges (Locke & Latham, 2002). Therefore, this paper argues that increasing challenging job demands is positively related to job satisfaction.

On the other hand, hindering job demands represent stress to employees, disabling them from achieving their goals (LePine, Podsakoff, & LePine, 2005). However, in line with goal-setting theory, employees measure their satisfaction based upon their goal attainment. Goals can be seen as an object or outcome that is used as a point of reference to evaluate satisfaction. Indeed, achieving or exceeding a goal leads to higher satisfaction, whereas not attaining a goal leads to increased dissatisfaction (Locke & Latham, 2002). If employees actively engage in decreasing their hindering job demands, they might be able to better focus on their tasks (Tims, Bakker, & Derks, 2013) and thus increase their chance of goal

(22)

attainment, which might eventually lead to higher satisfaction. In addition, hindering job demands are related to the health impairment process. However, if employees decrease their hindering job demands, they might also be able to reverse the health impairment process and thus they might become more satisfied. For these reasons, we propose that decreasing hindering job demands is positively related to happiness-related well-being, i.e. satisfaction.

Research has found that increasing both structural and social job resources is positively related to job satisfaction (Tims, Bakker, & Derks, 2013). According to COR theory, maximising job resources is positively related to well-being, as a large amount of resources works as a protection against resource exhaustion. Indeed, many resources help employees in dealing with job demands (Nielsen & Abildgaard, 2012). Moreover, having enough resources is crucial in terms of attaining goals (Schaufeli, Bakker, & Van Rhenen, 2009). As described above, goal-setting theory encompasses that goal attainment leads to higher satisfaction. Thus, we propose that employees will engage in increasing both their structural and social job resources to achieve their goals. Building upon the aforementioned theories and findings, we suggest that increasing both structural and social job resources is positively related to satisfaction as part of happiness-related well-being.

Second, health-related well-being can be conceptualised into two dimensions, namely stressors and strain. On the one hand, stressors such as workload and work intensification display events or situations that cause stress. On the other hand, strain can be defined as the response to stressors. Examples of responses include stress or burnout (Van de Voorde et al., 2012). This paper will focus on stress as an outcome, given that it also functions as a predictor of burnout (Halbesleben & Buckley, 2004).

Research has found that the presence of challenging job demands is positively related to stress (Van den Broeck et al., 2010). However, as employees proactively choose to engage in these challenges when crafting their challenging job demands, they might not necessarily

(23)

perceive higher levels of stress. They see the challenge that they are facing as opportunity to grow and develop (LePine, Podsakoff, & LePine, 2005) and thus may not react negatively towards a higher workload. In line with this, the effort-reward-imbalance (ERI) model (Siegrist, 1996) argues that stress is a result of an imbalance between efforts and rewards. Efforts represent characteristics such as job demands and rewards represent career opportunities or monetary benefits (Bakker & Demerouti, 2007). Therefore, if efforts are high but rewards are low, employees perceive higher levels of stress. However, on the other hand, if employees perceive having a balanced relationship between rewards and efforts, perceived stress levels may be lower. The latter symbolises the characteristics of challenging job demands. Employees excel effort but simultaneously see the opportunities for rewards. Thus, employees perceive a well-balanced relationship between efforts and rewards. As a result, we expect no relationship between increasing challenging job demands and stress.

The presence of hindering job demands imposes stress on employees (LePine, Podsakoff, & LePine, 2005). As explained above, hindering job demands are related to the health impairment process. However, if employees choose to decrease these hindering job demands, they may also decrease the effects of the health impairment process and their stress levels. We thus expect decreasing hindering job demands to be negatively related to stress.

Structural and social job resources are an important source for employees. Indeed, employees who have many resources are able to diminish the strains (Crawford, LePine, & Rich, 2010) that come from hindering job demands, for instance (Bakker & Demerouti, 2007). This is in line with the aforementioned interaction effect (buffer effect) between job demands and job resources, as proposed by the JD-R model. According to this effect, job resources buffer the negative outcomes of job demands on employees (Bakker & Demerouti, 2007). As a result, increasing social and structural job resources might help employees to buffer the negative effects of existing job demands and subsequently to decrease their stress

(24)

levels. Aside from the buffer effect, job resources are additionally crucial in their own right, given that they help employees to achieve their goals (Bakker & Demerouti, 2007), which in turn might also reduce stress levels of employees. We thus expect increasing structural and social job resources to be negatively related to stress.

Hypothesis 2a: Job crafting (all four dimensions) is positively related to employee satisfaction.

Hypothesis 2b: Job crafting (increasing structural and social job resources, decreasing hindering job demands) is negatively related to employee stress. However, increasing challenging job demands is not related to employee stress.

2.4. Performance

The performance of employees can be defined as an employee’s level of productivity on several job-related outcomes or behaviours, examined relative to his/her colleagues (Babin & Boles, 1998). Several studies have provided evidence that employees have the highest job performance in challenging and resourceful work settings because these environments increase their work engagement (Bakker, Tims, & Derks, 2012). Moreover, job crafting has been found to be predictive of colleague-rated in-role performance (Bakker, Tims, & Derks, 2012).

Increasing challenging job demands is a behaviour in which employees engage proactively. They like to engage in challenging tasks because it helps them to progress in their career and development. Similarly, LePine, Podsakoff, and LePine (2005) found that challenging job stressors are positively related to performance. Here, goal-setting theory can again explain why challenging job demands have a positive influence on performance. Goal-setting theory argues that difficult and specific goals motivate employees to higher performance (Locke & Latham, 2002). However, according to expectancy theory, this is only the case if employees think they actually can achieve these goals (Grant & Shin, 2012).

(25)

By increasing challenging job demands, employees actively choose their difficult goals and thus may believe that they can also attain these goals. As a result, this paper hypothesises that increasing challenging job demands is positively related to performance.

Hindrance stressors or hindering job demands are negatively related to performance (LePine, Podsakoff, & LePine, 2005) because they distract employees from focusing on completing their tasks, reduce their belief in attaining a goal and thus reduce employees’ motivation (Grant & Shin, 2012; LePine, Podsakoff, & LePine, 2005). However, if employees engage in decreasing their hindering job demands and thus are able to focus again on their tasks and goal attainment, this may result in higher performance. This behaviour would again be supported by self-regulation theory, assuming that employee make changes to their behaviour, i.e. decreasing hindering job demands, in order to increase their performance and eventually achieve a goal (Lord et al., 2010). This paper thus expects decreasing hindering job demands to be positively related to performance.

Both social and structural job resources are a predictor of performance. They are essential in attaining work goals (Bakker & Demerouti, 2007). In addition, as explained above, the JD-R model assumes job resources to have a motivational effect on employees, which increases their work engagement and leads to other positive work outcomes (Ângelo & Chambel, 2014). Thus, if employees craft their social and structural job resources, they have more resources available and should be more engaged in their work (Bakker, Tims, & Derks, 2012). Additionally, goal-setting theory states that the correlation between difficult goals and high performance is stronger when employees receive feedback (Grant & Shin, 2012). As a result, this paper proposes that increasing both structural and social job resources is positively related to performance.

Hypothesis 2c: Job crafting (all four dimensions) is positively related to employee performance.

(26)

2.5 The Mediating Role of Job Crafting

Previous scholars have agreed on the positive effect of HPWS on employee performance (Ang et al., 2013). Although previous studies have also found a relationship between HPWS and well-being, their results are contradictory. One stream emphasises positive well-being outcomes of HPWS (e.g. García-Chas et al., 2014), stating that HPWS increase job satisfaction because they offer benefits such as opportunities for growth, high levels of control and participation in decision-making processes, which are important to employees’ satisfaction (García-Chas et al., 2014; Guest, 2004). However, another stream

emphasises a negative relationship between HPWS and well-being (Kroon et al., 2009). It argues that HPWS lead to higher stress levels because employees feel that they need to reciprocate the benefits that HPWS provide them by working more (Kroon et al., 2009).

However, we will focus on the positive view of HPWS and argue that HPWS may have an advantageous effect on employee well-being, i.e. increased satisfaction and reduced stress, because job crafting acts as a mediator in this relationship. Indeed, HPWS may provide employees the opportunity to craft their own jobs and thus to proactively create their own positive outcomes. Building upon previous statements, HPWS might empower and motivate employees to proactively shape and craft their jobs. If employees craft their jobs, they might be able to optimize their work environment and thus to increase their well-being and performance. Based upon this argumentation, we expect HPWS to have a positive indirect effect on both well-being and performance. Put briefly, we propose that job crafting mediates the effect of HPWS on employee well-being and performance. Regarding stress, we have already proposed that increasing challenging job demands would not be related to stress. Thus, we assume that only the remaining three job crafting dimensions will mediate the relationship between HPWS and stress.

(27)

Hypothesis 3a: Job crafting (all four dimensions) mediates the relationship between HPWS and employee satisfaction.

Hypothesis 3b: Job crafting (increasing structural and social job resources, decreasing hindering job demands) mediates the relationship between HPWS and employee stress. However, increasing challenging job demands does not act as a mediator.

Hypothesis 3c: Job crafting (all four dimensions) mediates the relationship between HPWS and employee performance.

3. Research Method

3.1 Sample

In order to collect data, we used a purposive sampling technique, i.e. we chose those participants who were best able to represent a heterogeneous sample. The sample collected for the purpose of this thesis comprised employees who are working in organisations located in Germany. The German sample consisted of 79 employees and 46 managers corresponding to response rates of 82.29% and 95.83%.

However, the questionnaire used in this thesis has already been utilised for prior thesis projects in the Netherlands. We additionally used the data from these prior thesis projects to increase our sample size. The existing data set comprised a set of 190 employees and 88 managers who are working in organizations in the Netherlands.

Overall, our final sample involved 269 employees and 134 managers. As we are focusing on predicting employees’ behaviours with our hypotheses, we will only discuss the employee sample in more detail in the following. The final employee sample included 105 (39.5%) male and 161 (60.5%) female employees, while three gender entries were missing. The age ranged from 14 to 71 with a mean of 36.26 (SD = 12.593). Tenure varied greatly between the employees, ranging from less than one year to 42 years (mean = 8.8, SD =

(28)

9.01). About half of the employees (58.73%) have at least a higher vocational education degree. In addition, 78.65% of the employees have worked for at least one year together with their supervisor.

3.2 Procedure

We used an existing questionnaire for the purpose of data collection. The questionnaire comprised two different sub-questionnaires: the first sub-questionnaire was addressed to managers, whereas the second sub-questionnaire was addressed to the subordinates of each manager. Depending on the availability and agreement to participate, either one or two subordinates of each manager participated in this survey. To distribute the questionnaires, we sent an email to all participants, which included a hyperlink to the online questionnaire. In order to match both manager and subordinate questionnaires for our subsequent analysis, all participants additionally received an individual code in this email, which they had to enter in the online questionnaire. A week after sending the email invitations to participate in the survey, we sent a reminder to those participants who had not completed the survey. We used a cross-sectional time horizon as we measured results at one point in time. After the data collection was completed, we matched the answers of managers and their subordinates and eventually removed the codes to ensure anonymity for the next steps.

3.3 Measures

3.3.1 Translation

In order to collect data in Germany, we first needed to translate the existing Dutch questionnaire into German, which we realised as described in the following. We first translated the Dutch survey into German and subsequently compared it with a second

(29)

translation made by a German assistant professor of the human resource department at the University of Amsterdam.

3.3.2 Variables

The questionnaire was also used for prior thesis projects and thus it includes variables that are not relevant for this thesis. This part will only focus on variables that were included in our research model. Aside from job crafting, all variables were measured on a 7-point scale ranging from 1 (I strongly disagree) to 7 (I strongly agree).

HPWS were measured using a 15-item scale developed by Kehoe and Wright (2013).

We measured employee perceptions of HPWS. Questions are for instance: “Associates in this job have a reasonable and fair complaint process” and “pay raises for associates in this job are based on job performance”.

Job crafting was measured using a 21-item scale, as used by Tims, Bakker and

Derks (2012). The scale reflects the four different job crafting behaviours, i.e. increasing challenging job demands (5 items), decreasing hindering job demands (6 items), increasing structural job resources (5 items) and increasing social job resources (5 items). Questions are for example: “I make sure that my work is mentally less intense” and “I ask others for feedback on my job performance”. We used a 5-point scale ranging from “never” to “always” to collect responses.

Performance of the employees as rated by the manager was measured using a 4-item

scale, adapted from Van Dyne and LePine (1998). Examples of questions are: “This subordinate meets performance expectations” and “the particular subordinate fulfils the responsibilities specified in his/her job description”.

In order to measure job satisfaction, this paper utilised a 3-item scale, as used by Cammann, Fichman, Jenkins and Klesh (1979). An example question is: “All in all, I like working here”.

(30)

Finally, stress was measured using a 4-item scale, as utilised by Motowidlo, Packard and Manning (1986). Questions include for example: “I feel a great deal of stress because of my job” and “I almost never feel stressed at work”.

3.3.3 Control Variables

The control variables used in this study are gender, age and country. Aside from country, these variables were measured with questions allowing either open answers or single answers from a pre-defined list. The questions are: “What is your gender?” and “What is your age in years?” Country was not measured in the survey, given that the data collection

for this thesis only took place in Germany. In addition, data added from prior thesis projects were only collected in the Netherlands. Therefore, we could manually add a country code to the data set.

4. Data Analysis and Results

In the following, we will describe the steps undertaken to analyse the data set. Subsequently, we will interpret the results of the analysis and support or reject our hypotheses.

4.1 Data Analysis

We examined the collected data set by using IBM SPSS Statistics, version 23. In order to test our hypotheses, we initially conducted the steps as described in the following. First, we cleaned the data set to ensure that we only analysed data that is useful for our analysis. Second, we re-coded all counter-indicative items that were part of our analysis. As a third step, we examined the reliability of the scales, which is also depicted in table 4.1. The reliability values of the four job crafting behaviours are: 0.796 for increasing structural job resources, 0.837 for decreasing hindering job demands, 0.775 for increasing social job

(31)

resources and 0.764 for increasing challenging job demands. Moreover, the reliability values of the other scales are: 0.862 for HPWS, 0.748 for stress, 0.841 for job satisfaction and 0.874 for employee performance. All reliability values are above 0.7 and thus scale reliability can be expected (Pallant, 2007). As a result, we did not remove any items from the existing scales. As a fourth step, we ran a normality check by analysing kurtosis and skewness. All job crafting items were approximately normally distributed. HPWS, performance, stress and job satisfaction deviated from a normal distribution. However, owing to the relatively large sample size in our analysis (269 employee and 134 manager questionnaires), we expected kurtosis and skewness not to cause a meaningful difference in our analysis (Tabachnick & Fidell, 2001). Fifth, we computed scale means by averaging all items belonging to one scale. Finally, we checked all variables for multicollinearity. For all variables, variance inflation factor (VIF) values were below 10 and tolerance values scored higher than 0.10, thus indicating that no multicollinearity issues would arise in the subsequent analysis (Field, 2009).

After having completed these steps, we tested the hypotheses. We used hierarchical regression to test the direct effects and the SPSS macro of Hayes (2013) to test the mediation effects.

(32)

Table 4.1: Means, Standard Deviations, Correlations, and Reliability Scores Variables Number of items M SD 1 2 3 4 5 6 7 8 9 10 11 1. Age 1 36.26 12.59 - 2. Gender (0=male, 1=female) 1 .61 .49 -.013 - 3. Country (0=Netherlands, 1=Germany) 1 .29 .45 .266** .104 - 4. HPWS 15 4.3 .98 .088 -.036 -.105 (.862)

5. Increasing structural job

resources 5 3.73 .66 .127* .010 .367** .162** (.796) 6. Increasing social job

resources 5 2.78 .75 -.137* .006 .082 .254** .522** (.775) 7. Increasing challenging

job demands 5 3.07 .75 .030 .010 .263** .184** .679** .535** (.764) 8. Decreasing hindering job

demands 6 2.16 .75 .112 .088 .527** .037 .191** .280** .182** (.837)

9. Stress 4 3.9 1.17 .133* .034 -.003 -.013 .049 .089 .053 .048 (.748)

10. Job satisfaction 3 5.83 .99 .061 .046 .032 .253** .275** .188** .256** -.220** -.109 (.841)

11. Performance 4 5.95 .80 .024 .132 .188** -.023 .125 .112 .251** .030 -.123 .252** (.874) ** Correlation is significant at the 0.01 level (2-tailed).

(33)

4.2 Results

In the following, we will present the results of our hypotheses testing. We proceeded according to the recommendations of Baron and Kenny (1986) who proposed four conditions that must be met to support a mediation effect (Baron & Kenny, 1986). First, there must be a significant relationship between the independent and the dependent variable. Second, it is required that the independent variable is significantly related to the mediator variable (hypothesis 1). Third, the mediator and dependent variable have to be significantly related (hypotheses 2a, 2b, 2c). Finally, it is important that the relationship between independent and dependent variables is weaker and/or no longer significant when we include the mediator. If all these conditions have been met, we would be able to confirm our proposed mediation effects (hypotheses 3a, 3b, 3c).

4.2.1 Direct Effects

In order to test the prerequisites for a mediation effect, we conducted hierarchical regression analysis. In the first step of every hierarchical regression analysis we entered the control variables age, gender and country. In the second step, we additionally entered the particular independent variable(s). This stepwise procedure ensures that we can measure the effect of the independent variable on the dependent variable independent of the effects of demographic control variables.

The first prerequisite requires a significant relationship between HPWS and both well-being and performance. An overview of our results can be found in table 4.2. The results of hierarchical regression analysis demonstrate a positive relationship between HPWS and job satisfaction (β = .271, p < .01). However, there was no significant relationship between HPWS and stress (β = -.031, p > .05); instead, the control variable age was positively related to stress (β = .149, p < .05). In addition, we could not confirm a direct

(34)

relationship between HPWS and performance (β = -.001, p > .05). Only one control variable was significantly related to performance. Country (β = .190, p < .01) had a positive relationship with performance. As a result of these hierarchical regression analyses, the first condition for a mediation effect is only supported for the relationship between HPWS and job satisfaction.

In order to test the second condition for a mediation effect, which is a significant relationship between HPWS and job crafting, we again conducted hierarchical regression analysis. As is visible in tables 4.3 and 4.4, a direct effect was found between HPWS and most job crafting behaviours. HPWS were significantly and positively related to increasing structural job resources (β = .213, p < .01). In addition, the control variable country was positively related to increasing structural job resources (β = .368, p < .01). HPWS were also positively related to increasing social job resources (β = .317, p < .01). Again, some control variables were positively related to the dependent variable. Country (β = .134, p < .05) and age (β = -.157, p < .05) demonstrated a significant relationship with increasing social job resources. Next, HPWS were positively related to increasing challenging job demands (β = .227, p < .01). In addition, the control variable country was positively related to increasing challenging job demands (β = .276, p < .01). Finally, there was no significant relationship between HPWS and decreasing hindering job demands (β = .096, p > .05). However, the control variable country was significantly related to decreasing hindering job demands (β = .530, p < .01).

As a result of these hierarchical linear regressions, we could partially support hypothesis 1. HPWS are positively related to increasing structural and social job resources, as well as increasing challenging job demands. However, we could not confirm a significantly positive relationship between HPWS and decreasing hindering job demands and thus we needed to reject hypothesis 1 regarding decreasing hindering job demands. In

(35)

Table 4.2: Results of HPWS as Predictor of Well-being (Stress and Job Satisfaction) and Performance

Variables

Job satisfaction Stress Performance

R2 _B _{SE B} _β _R2 _B _{SE B} _β _R2 _B _{SE B} _β Step 1 .007 .022 .048* Age 0.005 0.005 .064 0.014 0.006 .149* -0.002 0.004 -.032 Gender 0.092 0.129 .045 0.098 0.150 .041 0.149 0.110 .093 Country 0.027 0.142 .012 -0.116 0.165 -.046 0.318 0.119 .190** Step 2 .078** .023 .048 Age 0.003 0.005 .032 0.014 0.006 .152* -0.002 0.004 -.032 Gender 0.109 0.125 .053 0.096 0.151 .040 0.149 0.111 .093 Country 0.105 0.138 .048 -0.127 0.167 -.050 0.317 0.120 .190** HPWS 0.276 0.063 .271** -0.037 0.076 -.031 -0.001 0.054 -.001

** Significant at the 0.01 level (2-tailed). * Significant at the 0.05 level (1-tailed).

(36)

Table 4.3: Results of HPWS as Predictor of Increasing Structural Job Resources and Increasing Social Job Resources

Variables

Increasing structural job resources Increasing social job resources

R2 _B _{SE B} _β _R2 _B _{SE B} _β Step 1 .145** .032* Age 0.002 0.003 .045 -0.009 0.004 -.157* Gender -0.036 0.079 -.027 -0.014 0.096 -.009 Country 0.526 0.087 .368** 0.219 0.105 .134* Step 2 .189** .129** Age 0.001 0.003 .019 -0.012 0.004 -.195** Gender -0.028 0.077 -.021 -0.000 0.091 .000 Country 0.567 0.085 .397** 0.289 0.101 .177** HPWS 0.143 0.039 .213** 0.243 0.046 .317**

Table 4.4: Results of HPWS as Predictor of Increasing Challenging Job Demands and Decreasing Hindering Job Demands

Variables

Increasing challenging job demands Decreasing hindering job demands

R2 _B _{SE B} _β _R2 _B _{SE B} _β Step 1 .072** .278** Age -0.002 0.004 -.035 -0.002 0.003 -.029 Gender -0.025 0.094 -.016 0.048 0.084 .031 Country 0.450 0.103 .276** 0.873 0.092 .530** Step 2 .122** .287 Age -0.004 0.004 -.062 -0.002 0.003 -.041 Gender -0.015 0.091 -.010 0.052 0.083 .034 Country 0.499 0.101 .306** 0.894 0.092 .542** HPWS 0.174 0.046 .227** 0.075 0.042 .096

(37)

Table 4.5: Results of Job Crafting as Predictor of Well-being (Stress and Job Satisfaction) and Performance

Variables Job satisfaction Stress Performance

R2 _B _{SE B} _β _R2 _B _{SE B} _β _R2 _B _{SE B} _β Step 1 .007 .022 .050* Age 0.005 0.005 .064 0.014 0.006 .149* -0.003 0.004 -.044 Gender 0.092 0.129 .045 0.098 0.150 .041 0.174 0.109 .110 Country 0.027 0.142 .012 -0.116 0.165 -.046 0.299 0.117 .183* Step 2 .195** .036 .117** Age 0.006 0.005 .080 0.016 0.006 .167* 0.000 0.004 -.008 Gender 0.136 0.117 .067 0.097 0.151 .041 0.195 0.106 .123 Country 0.233 0.163 .108 -0.211 0.209 -.083 0.349 0.150 .214* Increasing structural job resources 0.215 0.130 .142 -0.021 0.167 -.012 -0.191 0.119 -.164

Increasing social job

resources 0.216 0.101 .163* 0.160 0.129 .103 0.093 0.093 .087 Increasing challenging job demands 0.156 0.108 .117 0.022 0.139 .014 0.302 0.097 .292** Decreasing hindering job demands -0.507 0.094 -.386** 0.069 0.120 .045 -0.138 0.088 -.133 . ** Significant at the 0.01 level (2-tailed).

(38)

line with this, the second prerequisite for a mediation effect does not account for decreasing hindering job demands.

In order to test hypotheses 2a, 2b and 2c and the third prerequisite for a mediation effect, we again conducted hierarchical regression analyses. Our results, as depicted in table 4.5, reveal that both increasing social job resources (β = .163, p < .05) and decreasing hindering job demands (β = -.386, p < .01) were significantly related to job satisfaction. However, the relationship between decreasing hindering demands and job satisfaction was negative and thus opposite to what we expected. In addition, we could not confirm a relationship between the two job crafting behaviours of increasing structural job resources (β = .142, p > .05) and increasing challenging job demands (β = .117, p > .05) and the dependent variable job satisfaction. Therefore, hypothesis 2a was supported for one type of job crafting, i.e. increasing social job resources, and rejected for all other job crafting behaviours (increasing structural job resources, increasing challenging job demands, decreasing hindering job demands).

Considering the dependent variable stress, we could not find a significant relationship with any of the job crafting behaviours. This contradicts our propositions for all job crafting behaviours aside from increasing challenging job demands, as we predicted that increasing challenging job demands would not be related to employee stress. As a result, we could confirm hypothesis 2b regarding increasing challenging job demands. However, hypothesis 2b was rejected for increasing structural job resources, increasing social job resources and decreasing hindering job demands.

Finally, increasing challenging job demands was positively related to performance (β = .292, p < .01). By contrast, no relationship was found between the three remaining job crafting behaviours (increasing structural job resources, increasing social job resources and decreasing hindering job demands) and performance. Therefore, hypothesis 2c was only

(39)

supported in terms of increasing challenging job demands. Considering increasing structural job resources, increasing social job resources and decreasing hindering job demands, hypothesis 2c was rejected.

4.2.2 Mediation Effects

As described in the findings of the first prerequisite, we only found a direct relationship between HPWS and job satisfaction. However, Baron & Kenny’s (1986) propositions for testing mediation effects assume that a direct relationship between independent and dependent variable is necessary to test the final prerequisite and mediation, respectively. According to Baron and Kenny (1986), we thus only needed to test the fourth prerequisite for the relationship between HPWS and job satisfaction.

However, in contrast to Baron & Kenny (1986), recent scholars have rejected the assumption that a direct relationship between independent and dependent variable is necessary to support mediation. Instead, they argue that an indirect effect may still occur if no significant relationship between the independent and dependent variable exists (Hayes, 2009). Building upon this assumption, we not only tested the mediation effect of job crafting in the relationship between HPWS and job satisfaction but also tested whether an indirect effect exists in the relationship between HPWS and the dependent variables stress and performance. For this purpose, we used a bootstrapping method, as proposed by Hayes (2009). This procedure is also in line with Zhao, Lynch and Chen (2010), who state that it is only necessary to test whether an indirect effect occurs to decide whether mediation exists. In this sense, the analysis of a direct relationship between the independent and the dependent variable is only necessary to determine which kind of mediation exists.

Building upon Baron and Kenny (1986) and recent scholars (Hayes, 2009) we thus continued with the mediation testing as described in the following. We first tested the fourth prerequisite of Baron and Kenny (1986) regarding job satisfaction. Subsequently, we

HPWS and their relationship with well-being and performance : the mediating role of job crafting

Master Thesis

High Performance Work Systems and Their Relationship with Employee

Well-being and Performance: The Mediating Role of Job Crafting

Statement of Originality

Abstract

Table of Contents

List of Tables and Figures

Index of Appendices

1. Introduction

2. Theoretical Background

3. Research Method

4. Data Analysis and Results