The relationship between job crafting and job satisfaction mediated by the work-life balance of the employee : the moderating effect of exhaustion

Final Version Master Thesis

Name:

Alida Martina Louwen, Alies

Student Number:

10677585

First supervisor:

Eloisa Federici

Second supervisor:

Corine Boon

Topic:

Job Crafting

Date of Submission:

23-06-2017

Programme:

MSc. Business Administration

Specialisation:

Leadership and Management

Institution:

Amsterdam Business School

University of Amsterdam

The relationship between Job Crafting and Job Satisfaction mediated by the

Work-Life Balance of the employee. The moderating effect of Exhaustion.

2

Statement of Originality

This document is written by Alies Louwen, student of the University of Amsterdam, 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

The present study aims to uncover the way daily job crafting influences daily job satisfaction with a mediating effect of work-life balance. The interaction effect of exhaustion on the relationship between daily job crafting and work-life balance is also predicted. Building blocks for the argumentation found in current literature are the job demands-resources model, role theory and conservation of resources theory. A sample of 111 Dutch employees filled in a quantitative diary study for three, four or five consecutive working days in combination with one general survey. It is predicted and partly found that job crafting was positively associated with work-life balance and that work-life balance was positively associated with job

satisfaction. Furthermore, a weak interaction effect of exhaustion has been found. It is concluded that job crafting can have an important effect on work-life balance, and work-life balance in turn can affect job satisfaction. Implications for managerial practice and future research are provided.

4

Table of content

Introduction……… p. 5 Literature Review………... p. 8 - Job Crafting………. p. 8 - Job Crafting and Job Satisfaction……… p. 12 - Hypothesis 1……… p. 14 - Job Crafting and Work-Life Balance……….. p. 15 - Hypothesis 2a……….. p. 19 - Hypothesis 2b……….. p. 19 - Hypothesis 2……… p. 21 - Exhaustion……… p. 22 - Hypothesis 3………. p. 25 - Research Question……… p. 25 - Control Variables………. p. 25 Methodology……….. p. 27 - Procedure………. p. 27 - Sample………. p. 28 - Measures……….. p. 30 Results……… p. 33 - Descriptive Statistics……… p. 33 - Regressions……….. p. 41 Discussion……….. p. 52 Conclusion………. p. 58 Reference List……… p. 59 Appendix A – Measures……… p. 71

5

Introduction

Organizations nowadays are shifting responsibility downward: there is an increasing desire for employees that are proactive, assertive in their actions, seek for improvements themselves and go beyond the task requirements (Den Hartog & Belschak, 2012). This proactive

behaviour concerning the employees’ work tasks or social interactions at work can be found in the concept of job crafting. This concept can be defined as ‘the physical and cognitive changes individuals make in the task or relational boundaries at their work’ (Wrzesniewski & Dutton, 2001, p. 179).

Although job crafting is a relatively new concept in the world of research and practice, studies already provided evidence for the relationship between job crafting and several job outcomes (Tims, Bakker & Derks, 2013). Job crafting behaviours of employees may lead to outcomes that are rather positive, such as resilience, thriving, work engagement, performance, and well-being (Tims & Bakker, 2010; Berg, Dutton & Wrzesniewski, 2008; Tims, Bakker & Derks, 2013; Wrzesniewski, LoBuglio, Dutton & Berg, 2013). Furthermore, job crafting has been recognized to be important for one’s job satisfaction (Tims, Bakker & Derks, 2013; Tims & Bakker, 2010; Tims, Bakker & Derks, 2012; Leana, Appelbaum & Shevchuk, 2009). The relationship between job crafting and job satisfaction is explained in this study with the job demands-resources model (Demerouti, Bakker, Nachreiner & Schaufeli, 2001). This model states that job characteristics can be divided in resources and demands. In this study we expect this relationship because with job crafting behaviours the job demands and resources can be altered, which will have a positive influence on job satisfaction (Sousa-Poza & Sousa-Poza, 2000; Babin & Boles, 1996). With examining this relationship contributions to current literature are made. Previous research used several models to explain job crafting, making use of different dimensions and not all the dimensions used proved to have relationships with the job outcomes (Tims, Bakker & Derks, 2013). This suggests that

6

different dimensions of job crafting can have a different impact on the job outcomes, which makes the distinction between the dimensions important to include instead of using a general term for job crafting. Another contribution to current literature is the use of a daily diary study. Research suggests that job crafting is daily employee behaviour and is described as a continuous and everyday activity (Wrzesniewski & Dutton, 2001; Berg, Wrzesniewski & Dutton, 2010; Petrou, Demerouti, Peeters, Schaufeli & Hetland, 2012).

However, still much work needs to be done to disentangle the mechanisms through which job crafting affects job satisfaction. In this study we focus on the role of work-life balance as a mechanism, arguing that job crafting can indeed contribute to a well-functioning work-life balance of the employee which in turn can influence job satisfaction. Work-life balance is defined as ‘the satisfaction and good functioning at work and at home’ (Clark, 2000, p. 751). Current literature suggests that there are unresolved issues about what influences and shapes a good functioning work-life balance (Guest, 2002). While there is much research done about the effects of lack of work-life balance, too little is yet known about how to improve it (Jones, Burke & Westman, 2013). Work-life balance is proposed as a mediator with use of the role theory. Role theory suggests that individuals behave in ways that are predictable depending on the social identities and situation (Biddle, 1986). The

argumentation behind this theory is that job crafting could influence the roles of an individual concerning work and life which influences the balance between them. When a high level of balance across the system of roles is achieved, the positive outcome of job satisfaction can occur. This study proposes to unravel this mechanism through which job crafting affects job satisfaction.

In this study exhaustion is added as a moderator for the relationship between job crafting and work-life balance. Building on Conservation of Resources theory (Hobfoll, 1989), exhaustion is studied for an interaction effect. This theory states that people strive to

7

retain, protect and build resources and that what is threatening to them is the potential or actual loss of these valued resources (Hobfoll, 1989). In response individuals can capitalize on internal or external resources available in the environment, which can cause exhaustion

(Hobfoll, 1989). When an employee is exhausted, the possessed resources might be low. This way the acquiring of resources with use of job crafting behaviours will therefore not be that successful and in turn will not improve conditions for a good functioning work-life balance. The contribution of adding exhaustion into the research model is expanding on the theory of under which conditions job crafting can actually translate into increased work-life balance. Overall, this study examines the relationship between job crafting and job satisfaction with a mediating effect of work-life balance. The moderating effect of exhaustion is also included for the relationship between job crafting and work-life balance. The research design is a daily diary study.

Figure 1: Conceptual Model

The concept of job crafting is first discussed, followed by the argumentation of the suggested relationship between job crafting and job satisfaction. After this argumentation the suggested mediating effect of work-life balance is explained. At last the proposed interaction effect of exhaustion is explained with use of current literature. The contribution to theory and practice

H2a + Daily Job Crafting

▪ Seeking Resources ▪ Seeking Challenges ▪ Reducing Demands

Daily Job Satisfaction H2

+ H1

+

Work – Life Balance

▪ Negative spillover from work to life ▪ Positive spillover from work to life ▪ Negative spillover from life to work ▪ Positive spillover from life to work

H2b +

Exhaustion H3

8

is explained for every suggested relationship and used variable. A discussion and conclusion are present with practical implications, limitations and theoretical contributions included. The reference list and appendix are included.

9

Literature Review

Job Crafting

To examine the experience of employees about their work researchers have used the design of jobs as a starting point for decades (Berg, Wrzesniewski & Dutton, 2010; Tims, Bakker & Derks, 2013). Research focused on the top-down approach: this approach is about managers who are designing and altering the jobs for their employees (Berg, Wrzesniewski & Dutton, 2010; Wrzesniewski, LoBuglio, Dutton & Berg, 2013). Traditionally when

researchers examined what composes the experience of a job, they focused on external characteristics of the job itself or on the individual determinants (Wrzesniewski & Dutton, 2001). All these perspectives minimize the role that employees play in the composition of their own job (Wrzesniewski & Dutton, 2001). However recently the role that employees play in the design of their job has been recognized by introducing the concept of job crafting (Berg, Wrzesniewski & Dutton, 2010; Petrou et al., 2012). The focus of job crafting is on the proactive changes employees make to their own job boundaries, whereas job design focuses on the managers creating and enforcing the features of employees’ jobs. Job crafting can be defined as ‘the physical and cognitive changes individuals make in the task or relational boundaries at their work’ (Wrzesniewski & Dutton, 2001, p. 179), and can be seen as an action. Employees engage in job crafting because motivation for this proactive behaviour to alter their jobs arises from individual needs (Wrzesniewski & Dutton, 2001). The first need to fulfil is to maintain control over their job, this in order to prevent estrangement from their work (Braverman, 1974; Wrzesniewski & Dutton, 2001). A second need is to create a positive self-image in their work. And the third need to fulfil is a basic human need: the need for connection to others (Baumeister & Leary, 1995; Wrzesniewski & Dutton, 2001).

There are several definitions and operationalisations of job crafting. One approach to define and examine job crafting is to divide job crafting into three aspects of an employee’s

10

job: aspects related to tasks, relationships or cognitions (Tims, Bakker & Derks, 2012;

Wrzesniewski & Dutton, 2001). Employees proactively may alter the aspects of their job that are task-related, such as the content of a task or the amount of tasks (Tims, Bakker & Derks, 2012; Wrzesniewski & Dutton, 2001). Aspects related to relationships at work, for example with colleagues or customers, can also be changed in their intensity and amount (Tims, Bakker & Derks, 2012; Wrzesniewski & Dutton, 2001). Employees can also alter the cognitions they have about certain aspects of their job, the way how one sees the job. This way the meaning of their work can be enhanced, which is the aim of this dimension of job crafting (Tims, Bakker & Derks, 2012; Wrzesniewski & Dutton, 2001). Another approach of defining and examining job crafting can be found in the job demands-resources model (Demerouti, Bakker, Nachreiner & Schaufeli, 2001), which will be used in this study. This model states that job characteristics can be divided in resources and demands. Job resources are aspects of the job that can be physical, social, organizational or psychological and relate to the motivational process (Petrou et al., 2012). These aspects can be functional in achieving the work goals, stimulate personal growth and learning or reduce job demands and the associated psychological or physiological costs. Examples of job resources are performance feedback, autonomy and social support (Bakker & Demerouti, 2007). Job demands on the other hand are related to impaired health. Job demands are aspects of the job that require skills and effort, cognitive and emotional, and therefore are associated with certain physiological and

psychological costs (Petrou et al., 2012). Examples of job demands are role ambiguity, high work pressure and emotional demands (Bakker & Demerouti, 2007). Overall, job resources are aspects of the job that enable the employee to achieve their goals at their work, while job demands can put strain on the employee.

There are three different dimensions in the job demands-resources model regarding job crafting where an employee can engage into altering their job. The first dimension is

11

‘seeking resources’, which is help-seeking behaviour at work with the goal to gain specific resources (Lee, 1997). Examples are information seeking, asking feedback or advice from managers and colleagues and seeking learning opportunities (Petrou et al., 2012). Seeking resources is a possibility to mobilize job resources in order to cope with job demands (Tims & Bakker, 2010). The second dimension of the model is ‘seeking challenges’. A positive

interpretation of stressors results in the perception of challenges (Petrou et al., 2012).

Challenge stressors can have positive implications, for example increased job satisfaction and commitment and decreased turnover (Petrou et al., 2012). Behaviours of seeking challenges include taking on more responsibilities or trying to acquire new tasks (Petrou et al., 2012). A pitfall is that a challenge becomes too demanding and is not seen as a challenge anymore by the employee. The third dimension is ‘reducing demands’ and contrary to the other two dimensions reducing demands sheds light on the negative side of job crafting (Petrou et al., 2012). It includes behaviour with the aim of minimizing the demanding job aspects, mentally or physically, or reducing the time pressure and workload. When the job demands have become overwhelming for the employee, he or she may lower the job demands in a proactive way (Tims, Bakker & Derks, 2012). Examples of reducing demands are reducing the

workload, task avoidance and procrastination (Demerouti & Bakker, 2014). Reducing

demands might be a good way to deal with possible health problems but at the same time it is possible to have a negative effect on the job performance.

12

Job Crafting and Job Satisfaction

Although job crafting is a relatively new concept in the world of research and practice, studies already provided evidence for the relationship between job crafting and several job outcomes (Tims, Bakker & Derks, 2013). One of these outcomes includes job satisfaction. The

relationship between job crafting and job satisfaction is studied already and shown to be present (Tims, Bakker & Derks, 2013; Tims & Bakker, 2010; Tims, Bakker & Derks, 2012; Leana, Appelbaum & Shevchuk, 2009).

In this study we include the relationship between job crafting and job satisfaction. Job satisfaction is how employees feel about their job and different aspects of the job, whether they like or dislike their job, and it is an attitudinal variable (Spector, 1997). When an

employee is not satisfied, he or she will show reluctant behaviour and will not meet his or her responsibilities and will ultimately quit the job (Fuming & Jiliang, 2007). Different factors can influence the job satisfaction of the employee. These factors are environmental effects, job characteristics and personality variables (Agho, Mueller & Price, 1993). In this study we focus on job characteristics, which can potentially be influenced by job crafting behaviours of the employees. Other relevant job characteristics that showed to be important determinants of job satisfaction are, among others, interesting work, independent work and good relationships with managers and colleagues (Sousa-Poza & Sousa-Poza, 2000).

In this study it is proposed that job crafting would be positively related to job satisfaction because by altering the job demands and resources it might affect several determinants of job satisfaction. For example, when the employee is taking on more tasks (e.g. seeking challenges) or is seeking for learning opportunities (e.g. seeking resources) it will likely result in interesting and independent work. Independent work is an important determinant for job satisfaction (Sousa-Poza & Sousa-Poza, 2000), thus job crafting can potentially affect this positive outcome. Further explanation of this relationship involves the role of job resources. Job resources such as co-worker involvement, social and supervisory

13

support, learning opportunities, performance feedback and job control have a positive effect on job satisfaction (Sousa-Poza & Sousa-Poza, 2000; Babin & Boles, 1996). When the proactive behaviour of job crafting occurs the employee can seek for these resources or alter them, which will have a positive effect on job satisfaction. Furthermore, it is found that perceived flexible working hours (Scandura & Lankau, 1997) and job enrichment (Loher, Noe, Moeller & Fitzgerald, 1985) also positively influence job satisfaction. Both can be achieved with job crafting behaviours of the employee.

In this study a positive relationship is expected between job crafting and job

satisfaction because of the argumentation given earlier. Although different research is done which provides evidence for this relationship (Tims, Bakker & Derks, 2013; Leana,

Appelbaum & Shevchuk, 2009), this study makes a contribution by using the job demands-resources model with the three dimensions measured separately. Previous research used the job demands-resources model but with different dimensions and not all the used dimensions proved to have a relationship with the job outcomes (Tims, Bakker & Derks, 2013). This suggests that different dimensions of job crafting can have a different impact on the job outcomes, which makes the distinction between the dimensions important to include instead of using a general term for job crafting.

The other important contribution of this study is the use of a daily diary study. Research suggests that job crafting is daily employee behaviour and is described as a continuous and everyday activity (Wrzesniewski & Dutton, 2001; Berg, Wrzesniewski & Dutton, 2010; Petrou et al., 2012). This study uses a daily aggregate measure for a

consecutive three, four or five days for job crafting and job satisfaction. Since job crafting is part of organizational behaviour, there are several reasons why this should be studied at day-level. The first reason is that it is a very suitable research design to measure job crafting, since proactive behaviours show intra-individual variation over time (Sonnentag, 2003) and

14

proactivity includes a situational dimension and is therefore not quite stable (Ohly & Fritz, 2010). It is also in line with the affective events theory which states that job features influence the employee’s affect (Weiss & Cropanzano, 1996; Petrou et al., 2012). This means that the job characteristics on a daily level affect the mood of the employee which can result into proactive behaviours on a daily basis (Ohly & Fritz, 2010). The second reason is that a daily diary study eliminates recall biases (Bolger, Davis & Rafaeli, 2003; Ohly, Sonnentag, Niessen, & Zapf, 2010; Petrou et al., 2012). The retrospective recollection of employees’ experiences are frequently biased by semantic memory (Beal & Weiss, 2003; Petrou et al., 2012) and with the use of a daily diary study this bias will be reduced. The third reason is that it enables the research team to control for general individual inclinations and therefore

evaluate the effect of day-level predictors to day-level results (Daniels & Harris, 2005; Petrou et al., 2012). In this study an aggregate measure will be used for multiple days, showing the average of the daily variables across the week. This type of measurement is preferred because repeated real-time measures are helpful for developing precise individual summary variables, because some people are consistently less stable in affect, self-esteem or other variables (Fisher & To, 2012). This variability is shown to be not very accurate when described retrospectively by an individual (Solhan, Trull, Jahng & Wood, 2009). Therefore evaluating the desired indices from repeated real-time reports are preferred (Fisher & To, 2012).

Overall, this study proposes a direct positive relationship between job crafting and job satisfaction. With the aim to fill the research gaps, the hypothesis is as follows:

H1: Daily Job Crafting, which consists of the dimensions seeking resources, seeking

challenges and reducing demands, has a direct positive relationship with Daily Job Satisfaction.

15

Job Crafting and Work – Life Balance

A good perceived work-life balance of the employees is important for both employees and organizations and receives more attention nowadays in research, practice and media. Most of the research confirms that a lack of work-life balance has negative consequences for the effective functioning of the employees and their well-being (Guest, 2002). When there is a lack of life balance, there can be serious consequences in terms of personal- and work-related problems for the employee (Hobson, Delunas & Kesic, 2001). For the employee problems can occur such as increased levels of stress and stress-related illness, higher rates of family strive and violence and a lower life satisfaction (Hobson, Delunas & Kesic, 2001). Other consequences of a lack of work-life balance for the employee is reduced productivity, more absenteeism and decreased job satisfaction (Hobson, Delunas & Kesic, 2001), these consequences influence the organization as well.

These consequences of an imbalance are rather negative and therefore there is a need to prevent this. Before prevention can be discussed, knowing what the components are of work-life balance and what influences this balance is important. An imbalance can be explained with the use of positive and negative spillover (Grzywacz & Marks, 2000). This means that factors of one domain, work or life, are positively or negatively influenced by a spillover of the other domain (Grzywacz & Marks, 2000). There are four different spillovers which can cause the imbalance: a positive or negative spillover from work to life and a positive or negative spillover from life to work (Grzywacz & Marks, 2000). A negative spillover from work to life means that the job of the employee reduces the energy for activities at home or that the stress from work makes the employee irritable at home. The work-life balance can be influenced because of developments at work with this type of

spillover (Guest, 2002), like a co-worker that suddenly quits his or her job and therefore more tasks are assigned to the employee. A positive spillover from work to family means that aspects of work help to deal with issues at home or that work gives the employee energy and a

16

nice mood at home (Grzywacz & Marks, 2000). A negative spillover from life to work means that the employees’ responsibilities from their personal life interfere with the effort they can devote to their job or that personal problems are distracting at work. The work-life balance can be influenced because of developments at the personal life concerning this type of spillover (Guest, 2002). An example may be that the employee moved further from the workplace. At last, a positive spillover from life to work means that the communication in their personal life helps them to deal with struggles experienced in their job or that the home situation makes the employee feel ready for work. These four spillovers can potentially cause an imbalance (Grzywacz & Marks, 2000). Another component besides developments at work or personal life that can influence the work-life balance is about individuals and their lives that increases the need to achieve work-life balance (Guest, 2002). This could be a change in the cognitions of the employee which put more importance to the work-part of their life. Research showed that the most common influence when an imbalance is present is due to the developments at work (Guest, 2002). When employees perceive an imbalance between their work and life, approximately 9 out of 10 perceive their work-part of life as the dominant part (Guest & Conway 1998; Guest & Conway, 2000).

More specifically, research indicated that the structure of work has a strong impact on the life of the employee (Higgins, Duxbury & Irving, 1992). A sense of work-life balance, or imbalance, is the result of the demands and pressures of work that begin to dominate the life of employees (Guest, 2002). These demands and pressures of work can be found in longer working hours, the growth of evening and weekend work and results in more exhaustion and less scope for quality time. Stressful situations at work can influence the health of the

employee, and the timing of working hours can restrict the ability to participate in the employees’ personal life (Emslie & Hunt, 2009). A correlation is shown between working hours and the perceptions of imbalance between work and life (Guest, 2002; White, Hill,

17

McGovern, Mills & Smeaton, 2003; Biswas & Hassan, 2009). It is found that the perceived control of work schedules increases the work-life balance (Tausig & Fenwick, 2001). Individuals with higher job flexibility have more favourable work-family balance (Hill, Hawkins, Ferris & Weitzman, 2001). Employees with high perceived job flexibility are able to work longer hours before workload negatively impacts their work-family balance. Also the quality of work tasks and participation in decision-making within the organization result in lower levels of life conflict and are thus positively related to a good functioning work-life balance (Crompton & Lyonette, 2006). Employees who perceive less imbalance reported several conditions which are important for achieving a well-managed and good functioning work-life balance. These conditions are autonomy, more scope for direct participation and a friendly climate (Guest, 2002).

In this study we suggest that daily job crafting has a positive relationship with the perceived work-life balance of the employee, because it is expected that the employee can alter the resources of the job or handle the demands in such a way that the conditions for a good functioning work-life balance will be met. With the job crafting behaviour of seeking resources employees can increase their scope for direct participation and a friendly climate, which are factors that encourage a good work-life balance (Guest, 2002). Employees can for example ask for feedback from colleagues and their supervisors which will help to increase the direct participation and can influence the work climate. When engaging in seeking challenges, the employee can take on more responsibility or try to acquire more tasks. Therefore, with more responsibility the employee can acquire a stronger sense of autonomy, which is a good indicator for perceived balance between work and life by the employee (Guest, 2002). When engaging in reducing demands the employee can try to reduce the workload by changing the workhours or schedules, which in turn is favourable for a good work-life balance (Guest, 2002).

18 Contributions to current literature are made concerning the use of the variable life balance. Previous studies mostly focused on life conflict (Guest, 2002) and work-family balance (Greenhaus, Collings & Shaw, 2003). To get new and broader insights about the impact of different and more diverse factors, using work-life balance instead of work-life conflict or work-family balance is recommended (Greenhaus, Collings & Shaw, 2003). Furthermore, research using different methods is needed on work-life balance according to current literature (Sturges, 2012), the recommendation includes that other research methods than qualitative methods should be used. Regarding the sample a more diverse workforce is recommended with more diversity in type and size of the organizations of the employees (Sturges, 2012). The sample is also recommend to become more diverse regarding

demographics of the employees like age, family situation and tenure because these factors could potentially have an influence (Sturges, 2012; Carlson, Grzywacs & Zivnuska, 2009). We expect that the different dimensions of job crafting can result in a decrease in demands of work and in an increase of the indicators for a well-managed and good functioning work-life balance. Current literature suggests that there are unresolved issues about what influences and shapes a good functioning work-life balance (Guest, 2002). While there is much research done about the effects of lack of work-life balance, too little is yet known about how to improve it (Jones, Burke & Westman, 2013). Some studies mention the causes of work-life balance, but their focus is on the psychological or micro-sociological level (Fleetwood, 2007). Therefore, they tend to ignore causal factors at work (Fleetwood, 2007). We suggest that job crafting can potentially influence factors at work, since the employee alters the job resources and demands.

Overall, this study proposes a direct positive relationship between job crafting and work-life balance. With the aim of filling the stated research gaps regarding work-life balance, this study hypotheses the following:

19

H2a: Daily Job Crafting, which consist of the dimensions seeking resources, seeking

challenges and reducing demands, has a positive relationship with Work-Life Balance.

Furthermore, we also expect a positive relationship between work-life balance and job satisfaction. The relationship between work and life can have an important effect on job satisfaction and life satisfaction (Adams, King & King, 1996). Furthermore, a good

functioning balance between work and life leads to positive feelings about the job (Moore, 2007; Hobson, Delunas & Kesic, 2001). From the negative perspective, a conflict in work-to-life is linked to job dissatisfaction (Anderson, Coffey & Byerly, 2002). When there is a well-functioning work-life balance there is none to very little conflict, so this is not a hindering factor that could affect the employee’s job satisfaction. We also suggest this relationship because the determinants of job satisfaction can be influenced through work-life balance. As mentioned above, the most important determinants of job satisfaction are interesting work, independent work and good relationships with managers and colleagues (Sousa-Poza & Sousa-Poza, 2000). For example, a good functioning work-life balance leads to positive development and growth within work and non-work areas (Kalliath & Brough, 2008). Positive development and growth can potentially lead to independent and interesting work, which are important determinants of job satisfaction (Sousa-Poza & Sousa-Poza, 2000). Therefore, the following hypothesis is proposed:

H2b: Work – Life Balance has a positive relationship with Daily Job Satisfaction.

20

Building on role theory, we propose that work-life balance acts as a mediator in the

relationship between job crafting and job satisfaction. Role theory focuses on an important feature of social life, namely the characteristic behaviour in patterns or roles (Biddle, 1986). By presuming that persons are members of social positions, including expectations for own behaviour and those of others, it explains roles. Role theory explains that individuals behave in various ways that are predictable depending on the situation and social identities (Biddle, 1986). Role is behaviour referring to expectations from the individual associated with a position in the social system where he or she is in (Biddle, 1986).

Job design describes how jobs, tasks and roles are structured (Tims & Bakker, 2010), thus roles are also included in the design of jobs and characteristics of the jobs. As described before, job crafting behaviours alter the characteristics of the job (Wrzesniewski & Dutton, 2001). Therefore, with use of job crafting behaviour the employee can change the job characteristics and thus the roles involved. Job crafting behaviours are embedded in role theory because employees with the same jobs enact their roles in different ways (Tims & Bakker, 2010). They will perform a slightly different set of tasks even though the jobs are the same. Several outcomes of job crafting behaviour are correlated with role ambiguity and conflict: autonomy, task variety, feedback from others and participation in decision-making are negatively correlated with role ambiguity and role conflict (Jackson & Schuler, 1985). Even role conflict and role ambiguity can be demands (LePine, Podsakoff & LePine, 2005; Tims & Bakker, 2010). Therefore we propose that job crafting behaviour has an effect on the roles of the individual.

This study proposes that work-life balance gives an explanation of why job crafting and job satisfaction are related with use of role theory. Work-life balance is related to the multiple roles individuals have both in the work field and in the personal field (Allen, 2001). Role theory predicts that multiple life roles result in interrole conflict as individuals

21

experience difficulty performing each role successfully because of conflicting demands (Allen, 2001), which can in turn result in increased stress (Biddle, 1986). Work-family role strain is the result of the influence of demands and resources derived from several sources, work-related and personal. Whereas the demands of multiple roles can result in role strain, resources may reduce role strain by empowering individuals to cope with these demands (Allen, 2001). Work-life balance can therefore give an explanation why job crafting and job satisfaction are related because the roles at work that can be altered by job crafting with use of demands and resources have an influence on the roles concerning work-life balance (e.g. various personal life roles and work roles). For example, work schedules and work orientation may produce pressures to participate extensively in the work role or the family role

(Greenhaus & Beutell, 1985), and this can be altered through job crafting. This balance in roles is important in achieving job satisfaction because people who maintain a higher level of balance across their entire system of roles can enjoy a range of positive outcomes (Marks & MacDermid, 1996). These outcomes include role ease and lower measures on role strain (Marks & MacDermid, 1996). The outcome role strain is negatively related to job satisfaction (Um & Harrison, 1998). Thus, job crafting influences the different roles of an individual concerning work and life which influences the balance between them. When a high level of balance across the system of roles is achieved, positive outcomes occur which will lead to job satisfaction. Therefore work-life balance is the explanation of why job crafting and job satisfaction have a relationship, and the following hypothesis is proposed:

H2: Work – Life Balance mediates the positive relationship between Daily Job Crafting,

which consists of the dimensions seeking resources, seeking challenges and reducing demands, and Daily Job Satisfaction.

22

Exhaustion

This study suggests that exhaustion of the employee has a moderating negative effect on the positive relationship between daily job crafting and work-life balance. Exhaustion is a state of intensive physical, emotional and cognitive strain (Demerouti, Bakker, Nachreiner &

Schaufeli, 2001). Previous research provided evidence on the relationship between work-life balance and (emotional) exhaustion (Keeton, Fenner, Johnson & Hayward, 2007). It is

proposed that the pressures and demands at work, reflected in longer working hours and more exhaustion, leave less time for the personal life of the employee (Guest, 2002) and therefore might lead to increased absenteeism (Anderson, Coffey & Byerly, 2002; Hughes &

Bozionelos, 2007). A sense of work-life balance, or imbalance, can be the result of the demands and pressures of work that begin to dominate the life of employees (Guest, 2002). Energy levels need to be taken into account in the context of high demands because they may be linked to the capacity of coping with pressures of competing demands (Guest, 2002). This coping with pressures of competing demands is important in achieving a well-functioning work-life balance. Also, the intrusion of work demands into personal life was related with reports of (emotional) exhaustion for employees (Hughes & Bozionelos, 2007). This is proposed because the arousal associated with high job demands will be directed internally with the consequence of exhaustion. Therefore, increasing job demands produce exhaustion (Van Yperen & Hagedoorn, 2003) and job demands have an influence on work-life balance (Guest, 2002).

Current research also provided information about exhaustion and job crafting. Job demands such as high work pressure, emotional demands and role ambiguity may lead to sleeping problems and exhaustion (Bakker & Demerouti, 2007). Job demands were primarily and positively related to exhaustion (Bakker & Demerouti, 2007) and challenging demands are related with exhaustion (Crawford, LePine & Rich, 2010). Conditions where job demands

23

were high and job resources were low resulted in the highest levels of exhaustion

(Xanthopoulou, Bakker, Demerouti & Schaufeli, 2006). But employees who proactively craft their own jobs can decide for themselves how to meet their job demands and therefore they can potentially reduce job strain including exhaustion (Bakker & Demerouti, 2007). Thereby high levels of job resources protect the employees from exhaustion because having access to larger pools of resources allows the employees to handle the job demands and guard

themselves from strain (Tims, Bakker & Derks, 2013). An adequate job resources pool has been found to relate positively to protect employees from experience of exhaustion (Schaufeli & Bakker, 2004). Also, poorly designed jobs or chronic job demands exhaust employees’ mental and physical resources and may therefore lead to depletion of energy which is a state of exhaustion (Bakker & Demerouti, 2007). Therefore job crafting can be useful for

preventing exhaustion. Seeking resources and seeking challenges were associated with low exhaustion and reducing demands seemed to predict exhaustion positively (Petrou, Demerouti & Schaufeli, 2015). Employees who reduced their demands reported higher exhaustion. This in turn leads to more reducing demands (Petrou, Demerouti & Schaufeli, 2015). Therefore, exhaustion and reducing demands strengthen each other and are related over time. This can be explained by the following: exhausted employees put less effort in their tasks (Banks,

Whelpley, Oh & Shin, 2012) and therefore they increase their workload (Van Eerde, 2000) and intensify their exhaustion. All together this means that the dimensions of job crafting seeking challenges and seeking resources can result in a decrease in exhaustion and reducing demands in an increase.

Building on Conservation of Resources theory (COR) (Hobfoll, 1989) it is suggested

that the proposed relationship between job crafting and work-life balance is negatively moderated by exhaustion. COR states that people strive to retain, protect and build resources and that what is threatening to them is the potential or actual loss of these valued resources.

24

There are different behaviours that can occur when people attempt to offset resource loss or to gain resources (Hobfoll, 1989). Employees can capitalize on internal or external resources available in the environment. When employees invest their important resources, the attempt is to translate them into more important resources (Hobfoll, 1989). When an employee invests their time and energy to acquire important resources like money, with the loss of the resources time and energy, exhaustion can occur. When time is limited there is less time to sleep and when energy is limited the employee can feel fatigued. In relation to the suggested moderating effect, there is an offset in resources between the variables. When an employee is exhausted this could influence the effect the job crafting behaviours will have on the work-life balance. Because when an employee is exhausted the possessed resources might be low. This way the acquiring of resources with use of job crafting behaviours will therefore be not that

successful. The low level acquired resources with use of job crafting behaviour will therefore not improve conditions for a good functioning work-life balance like autonomy, scope for direct participation or a friendly work climate (Guest, 2002).

Another explanation for the suggested moderating effect involves work pressure. Previous research has shown that exhaustion leads to more work pressure over time for the employee (Demerouti, Bakker & Bulters, 2004). Work pressure can be seen as a job demand (Bakker & Demerouti, 2007) and a negative work to life spillover. This means that work pressure can have a negative influence on work-life balance. Therefore we propose that under high levels of exhaustion the positive relationship between job crafting and work-life balance will be weakened.

The empirical links between job crafting and exhaustion remain largely unexamined (Petrou, Demerouti & Schaufeli, 2015). The contribution of adding exhaustion into the model is expanding on the theory of under which conditions job crafting can actually translate into increased work-life balance. Therefore the hypothesis will be as follows:

25

H3: Exhaustion negatively moderates the positive relationship between Daily Job Crafting,

which consists of the dimensions seeking resources, seeking challenges and reducing demands, and Work – Life Balance such that this relationship is weaker for higher levels of exhaustion.

Research Question

The overall research question of this study is:

What is the mediating effect of Work – Life Balance on the relationship between Daily Job Crafting, which consists of the dimensions seeking resources, seeking challenges and reducing demands, and Daily Job Satisfaction? And what is the moderating effect of Exhaustion on the relationship between Daily Job Crafting and Work-Life Balance?

Control variables

Three control variables were included in this study because they can influence the variables in the model according to literature. These control variables are working hours, age and

education. The first control variable included is working hours. In regard to work-life balance, long working hours are a constant source of negative work to life spillover (White, Hill, McGovern, Mills & Smeaton, 2003). This is supported by the fact that employees who worked longer hours were more likely to report an imbalance between work and life (Guest, 2002). It has been shown that the numbers of hours worked contributes directly to feelings of work and life conflict (Sturges & Guest, 2004). The strongest predictor of work-life balance is the amount of hours worked (Keeton, Fenner, Johnson & Hayward, 2007).

Another control variable incorporated is the age of the respondents. With regards to

work-life balance the variable age can have some influence. The demands coming from the personal life might be very low for someone who is young in comparison with older

26

employees (Guest, 2002), this can be due to (the lack of) a partner, children or elderly parents. Furthermore, young people emphasise achievement of work-life balance more than their predecessors (Kugelberg, Lewis & Smithson, 2001). It has been suggested that the

relationship between work and life may be even more important to young employees than it is to other groups of workers (Sturges & Guest, 2004). Young people want to develop and manage their careers on their own terms and an important part of this is the balance between work and non-work aspects of their lives (Loughlin & Barling, 2001). Although young people seek for work-life balance more than older people, their concern with career success results in feelings of pressure to work longer hours. In turn they can experience dissatisfaction with their own work-life balance (Sturges & Guest, 2004). Younger individuals perceive more work-life imbalance, work-life balance increases with age (Tausig & Fenwick, 2001). The third control variable included is education. Higher educated persons perceive more work-life imbalance (Tausig & Fenwick, 2001). Job crafting activities are also explored of employees working routine jobs instead of highly educated knowledge workers

(Wrzesniewski & Dutton, 2001). Employees in low-autonomy jobs may also discover job-crafting strategies (Wrzesniewski & Dutton, 2001) although the majority of most research samples regarding job crafting are highly educated (Tims, Bakker & Derks, 2013).

27

Methodology

This section contains the explanation of the procedure of the research, information about the sample and the measures used in this study are explained.

Procedure

A quantitative study of explanatory nature was performed in order to answer the suggested research question. A deductive approach was used in this study, this means that existing theory is used to develop the suggested hypotheses and these are tested with use of the acquired data to add knowledge and new insights to existing theory (Eisenhardt & Graebner, 2007).

The data collection instruments for this study were two different surveys. These surveys were handed out with use of e-mail to the respondents. In the first email information was present about the aim of the study, explanation about the other surveys, explanation that there is a raffle of vouchers as an incentive which can be expected when all the surveys are filled in, contact information and the link to the first survey. The surveys were administered online with use of Qualtrics. The respondents could fill in the surveys with use of a mobile phone, computer or tablet. Firstly a survey was handed out with the aim to measure general variables which are low in fluctuation like work-life balance and demographics. The

following week the participants were emailed a daily questionnaire for five consecutive days. This survey had the same content of questions everyday of that week. The participants were instructed to fill in the surveys right before leaving the office or at the latest before going to sleep. After that time the survey closed and could not be filled in anymore. Therefore the research team checked if the surveys were completed and if this was not the case a reminder was send to the particular respondent. These daily surveys were used to measure the variables daily job crafting, daily job satisfaction and daily exhaustion. Both the surveys were in Dutch.

28

Before the surveys were send by email to the respondents a pilot test was performed. For both surveys individuals were asked if the questions were clear, how much time the survey took them and if there were any mistakes or unclarities. The comments given by the individuals are taken into consideration, discussed by the research team and incorporated.

Three members of the research team, consisting of four researchers, collected the data. Every member of the research team had a different focus in research, which resulted in

combined surveys with more variables and items than necessary for this study. Therefore the gathered data useful for the other members of the research team were excluded from the dataset. With use of this dataset the suggested relationships were tested by doing multiple regression analyses .

Sample

The population of interest for this study was the Dutch workforce, consisting of employees who work at least four days per week. Freelancers were excluded because this could influence the variable daily job crafting drastically. Respondents were gathered with use of the non-probability convenience sampling technique. This means that the acquired respondents were ready and available to fill in the surveys (Fink, 2003). Respondents were reached by

Facebook, LinkedIn, personal phone number, personal email-address or in person. The amount of Dutch employees that were invited was 133. Out of these invited respondents 127 filled in the daily diary surveys partly or all of them. The response rate therefore would be 95%, however there were respondents who filled in the survey just one or two days of the required five days. Only the respondents who filled in the daily diary surveys for three or more days were included in the dataset, which are in total 111. Therefore the effective response rate would be 83%. The sample consisted of 111 Dutch employees. From the 111 employees, 80 filled in five daily diary surveys (N =80), 20 filled in the surveys for

29

four days (N = 20) and 11 filled in three of the five surveys (N = 11). From the 111 employees, 105 filled in their age. The age is ranging from 21 years till 63 years, with an average age of 34 years (SD = 13.04). Looking at the sample, 60% are women and 40% are men. Concerning working hours per week according to contract, the sample is ranging from 12 hours to 82 hours with an average of 37 hours (SD = 7.49). From the 111 respondents, 101 filled in their working hours per week according to contract. From the 111 respondents, 106 filled in their education level. These highest achieved education levels range from high school to PhD, but the most common levels were HBO with 37.8%, a Masters University Degree with 34.2% and MBO with 12.6%. The sectors of work field of the respondents are diverse but the most common sectors are Medical and Care with 18.0%, Service with 17.1% and Educational for 12.6%. Concerning tenure in the organization and tenure with the current supervisor the response is very low. For tenure in the organization 54 respondents from the 111 responded and for tenure with the current supervisor only 48 respondents responded. The range of tenure in the organization is from one year to 40 years, with an average of 11 years (SD = 10.56). The range of tenure with the current supervisor ranges from one year to 19 years, with an average of five years (SD = 3.70). Due to the low response rate for the tenures they are not included in the data analysis. All together diversity can be found in age (e.g. slightly more women), gender, sector and working hours. For the highest achieved education level the higher educated respondents are slightly overrepresented. All together the sample can be described as a highly educated group, diverse in age and sector which on average work nearly full-time and with slightly more females than men.

30

Measures

To measure the relationships suggested in this study several measurements were used. The measurements work-life balance, daily job crafting with the three dimensions, daily job satisfaction and daily exhaustion will be discussed in this section.

Work-life Balance

In this study the variable work-life balance is included in the general survey and consists of four dimensions: negative and positive spillover from work to life, and negative and positive spillover from life to work. All four the spillovers are measured with use of the items of Grzywacs & Marks’s (2000) scale. All the items were translated into Dutch with use of back-translation.

The negative spillover from work to life consists of four items and this scale has a Cronbach’s alpha of 0.77, which indicates a reliable scale (α > 0.7). An example item is ‘Job worries or problems distract you when you are at home’. The positive spillover from work to life is measured with the use of three items and this scale has a Cronbach’s alpha of 0.78 which indicates a reliable scale. The items consisted of statements like the following: ‘The things you do at work help you deal with personal and practical issues at home’. The third spillover, negative from life to work, consists of four items. This scale has a Cronbach’s alpha of 0.77, which indicates a reliable scale. One example of the four items includes

‘Responsibilities at home reduce the effort you can devote to your job’. The last type of spillover is positive from life to work and consists of four items. This scale has a Cronbach’s alpha of 0.72, which indicates a reliable scale. An example item is ‘Talking with someone at home helps you deal with problems at work’.

Not a single item of the all the different scales was removed since it would not increase the reliability of the scale. The corrected item-total correlations are all above 0.3,

31

which indicates that all the items have a good correlation with the total score of the scale. The respondents could answer the items with a five-points Likert scale ranging from ‘never’ to ‘all the time’ for all the spillovers. All the items can be found in Appendix A.

Daily Job Crafting

In the daily diary survey job crafting is measured with the three dimensions separately: seeking resources, seeking challenges and reducing demands. All the items for these three dimensions originate from Petrou, Demerouti, Peeters, Schaufeli and Hetland (2012). The items were altered to a daily-level measure by adding ‘today’. All the items were translated into Dutch with use of back-translation.

Seeking resources was measured with use of four items and this scale has a

Cronbach’s Alpha of 0.80. An example of an item of seeking resources includes ‘I have asked others for feedback on my job performance’. The dimension seeking challenges was measured with use of three items and this scale has a Cronbach’s Alpha of 0.94. An example item is ‘I have asked for more tasks if I finish my work’. Reducing demands was measured with use of four items and this scale has a Cronbach’s Alpha of 0.82. An example item is ‘I have tried to ensure that my work is emotionally less intense’.

The Cronbach’s Alpha for the three dimensions were all sufficient, they all indicated a reliable scale (α > 0.7). Not a single item of the scale was removed since it would not increase the reliability of the scale. The corrected item-total correlations are all above 0.3, which indicates that all the items have a good correlation with the total score of the scale.

Respondents indicated whether to what extent the items applied to them using a five-point Likert scale ranging from ‘does not apply to me’ to ‘totally applies to me’. All the items can be found in Appendix A.

32

Daily Job Satisfaction

In this research daily job satisfaction was measured by the single-item scale of the previous studies of Kunin (1955) and Kunin (1998). The item included in the daily diary survey for measuring daily job satisfaction is the following: ‘Please indicate how satisfied you are with your job today’ (Appendix A). Respondents indicated to what extent they were satisfied with their job that particular day using a five-point Likert scale ranging from ‘very dissatisfied’ to ‘very satisfied’. The item was translated into Dutch with use of back-translation.

Exhaustion

In this study the variable daily exhaustion is measured using the four-item scale of McNair, Lorr and Droppleman (1971), but due to the length of the daily diary survey we chose to use only two items of this scale. The measurement included two items where the respondents had to answer to which extent they felt ‘exhausted’ and ‘fatigued’ (Appendix A). Respondents indicated to what extent the two items applied to them for that particular day using a five-point Likert scale ranging from ‘not at all’ to ‘very much’. The Cronbach’s alpha of this scale is 0.87. This indicates a reliable scale (α > 0.7). Not a single item of the scale was removed, since there are only two items. Both items were translated into Dutch with use of back-translation.

33

Results

In this paragraph the results were interpreted from the tests that were executed. Descriptive statistics of the sample are discussed followed by analyses of several regression analyses. The hypotheses were tested with use of the analysis of the outputs, and therefore supported or rejected.

Descriptive Statistics

After the data of the respondents were collected, the data of this sample was exported to the program IBM SPSS version 24 for the analysis. Before testing the actual hypothesized model of this study the dataset was cleaned up by finding errors and adjust or delete these errors. Respondents who filled in the daily diary survey for less than three days were excluded listwise from the analysis. It was found that there were 16 respondents out of 127 who had to be excluded (N = 111). The aggregated scale means were computed for the variables of the daily diary survey. The means, standard deviations, correlations and reliabilities of the variables of the sample (N = 111) are displayed in Table 1. These variables include daily job crafting, work-life balance, daily job satisfaction, exhaustion and the control variables working hours, age and education.

34

35

The correlations were computed with the scale means of the variables and are

displayed in Table 1. Several correlations are found with the control variables age, education and working hours. Also tenure in the organization has a significant correlation but because of the small sample (N = 59) this is not taken into consideration in further analyses. Dimensions of job crafting behaviours are correlated with each other: seeking resources is significantly correlated with seeking challenges r(126) = 0.47, p < 0.01. This means that when an

employee involves in the job crafting behaviour of seeking resources, it is likely that he or she will also involve relatively more in seeking challenges. Furthermore, a significant correlation is found between the daily job crafting dimension seeking resources and a positive spillover from life to work, r(114) = 0.32, p < 0.01. This may indicate the presence of the suggested relationship in hypothesis 2a under certain circumstances. Also, there were correlations found among the spillovers and between several spillovers and job satisfaction, the latter could indicate the presence of hypothesis 2b. Furthermore, exhaustion has significant correlations with the two negative spillovers. This could indicate a presence of the moderating effect. While examining the three dimensions of job crafting, there is no significant correlation with job satisfaction or the different spillovers from work-life balance (p > 0.05) with the

exception of positive spillover from life to work. Furthermore, the correlations were also very small which could indicate the absence of relationships among these variables which are proposed in hypothesis 1.

A principal axis factoring analysis (PAF) was conducted on the scales for the

dimensions of daily job crafting and the spillovers from work-life balance. For the dimensions

of daily job crafting the Kaiser-Meyer-Olkin measure verified the sampling adequacy for the

analysis (KMO = 0.743). The Bartlett’s test of sphericity χ² (55) = 770,348 with p-value of 0,000 indicated that this test is sufficient for this scale. Three components had eigenvalues over one, as Kaiser’s criterion (Eigenvalues = 3.866; 2.442; 1.416). This indicates that three

36

factors can be obtained which in this case are the dimensions seeking resources, seeking challenges and reducing demands. In combination these factors explained 61.01% of the variance. Examination of the scree plot revealed levelling off after the third factor in

agreement with Kaiser’s criterion. Therefore, three factors were retained and rotated with an Oblimin with Kaiser normalization rotation. Table 2 shows the factor loadings after rotation. It shows that items that cluster on the same factor represent the same factor, factor 1

represents seeking challenges, factor 2 represents reducing demands and factor 3 represents seeking resources.

For the dimensions of spillover from work-life balance the Kaiser-Meyer-Olkin measure verified the sampling adequacy for the analysis (KMO = 0.735). The Bartlett’s test of

sphericity χ² (91) = 514,289 with p-value of 0,000 indicated that this test is sufficient for this scale. Four components had eigenvalues over one, so Kaiser’s criterion is met (Eigenvalues = 3.643; 2.522; 1.706; 1.135). This indicates that four factors can be obtained which are the four different spillovers from work-life balance. In combination these four factors explained 51.41% of the variance. In agreement with Kaiser’s criterion the scree plot revealed levelling off after the fourth factor. Thus four factors were retained and rotated with an Oblimin with Kaiser normalization rotation. Table 3 shows the factor loadings after rotation which shows that items that cluster on the same factor represent the same factor. The clustering of the items show that factor 1 represents negative life to work spillover, factor 2 positive life to work spillover, factor 3 positive work to life spillover and factor 4 negative work to life spillover. One item of negative work to life spillover ‘Problemen of zorgen over werk leiden je af wanneer je thuis bent’ shows high cross-loadings on negative life to work spillover as well. It could be due to the content of the item since it is about problems at work that cause

37 Table 2

38 Table 3

39

After the correlations and factor analysis the normality was checked for the variables. The variables were examined on skewness and kurtosis which are shown in Table 4. The only variable that shows some concern regarding the normality is job satisfaction. Job satisfaction is slightly negatively skewed as showed in Figure 1. This implies that the majority of the respondents reported relatively high levels of job satisfaction. The absence of a normal distribution can be explained through literature. It is found that job satisfaction contributes to the intention to leave which in turn has a strong linkage with turnover (Tett & Meyer, 1993). This may imply that people who are dissatisfied with their job have stronger intentions to leave their current job and try to find another job they are more satisfied with. Thus, those employees who left their job or switched to a more satisfying job as a consequence of job dissatisfaction could be more rare and therefore difficult to include in the sample and are therefore possibly underrepresented. It is chosen not to transform the variable job satisfaction, since it is only very slightly negatively skewed and it is generally recognized that with a large sample size the effects of skewness are trivial (Games, 1984). Transforming the data to eliminate skewness merely adds an extra layer of obfuscation (Games, 1984).

40 Table 4

41

Regressions

After getting insights concerning the data by investigating correlations, factor loadings and normality the hypotheses were tested by doing multiple regressions. For executing this Process is used which is an external macro developed by A.F. Hayes (2012) in combination with the program SPSS. For this study Model 7 of Process is used to test the hypothesized relationships (Hayes, 2012). This model is ran three times, one for each of the dimensions of job crafting. These regressions were performed to examine the three dimensions of job crafting and their potential relationship with job satisfaction, direct or indirect through the four spillovers of work-life balance. The suggested interaction effect of exhaustion is examined after controlling everything for working hours, age and education.

Regression Seeking Resources as Independent Variable

In hypotheses 1 it is proposed that the dimension seeking resources of daily job crafting has a direct positive relationship with daily job satisfaction. After examining the correlation it is found to be non-significant. This could imply the absence of a relationship between the

variables. Using regression this is tested and the results can be found in Table 5. The results of the regression for the direct relationship between seeking resources and job satisfaction did not support the relationship since it was found to be non-significant (β = 0.085, p > 0.05). The bias-corrected and accelerated confidence interval included zero (BCa95 = [-0.13, 0.30]),

which indicates that seeking resources has no relationship with job satisfaction. Therefore, hypothesis 1 is rejected for the dimension seeking resources.

In hypothesis 2a, 2b and 2 the mediating effect of work-life balance is proposed for the relationship between seeking resources and job satisfaction. Hypothesis 2a states the relationship between seeking resources and work-life balance, hypothesis 2b states the relationship between work-life balance and job satisfaction and hypothesis 2 is the overall

42

mediating effect of work-life balance on this relationship. The regression is performed with the four spillovers of work-life balance analysed separately. Examining the results of the regression analyses for hypothesis 2a, displayed in Table 5, showed that there is no significant relationship for three of the four spillovers of work-life balance. Negative spillover from work to life showed no significance (B = -0.12, p > 0.05), this is also the case for positive spillover from work to life (B = 0.13, p > 0.05) and negative spillover from life to work (B = -0.04, p > 0.05). However, the results of the regression for positive spillover from life to work show that there is a significant relationship between the job crafting dimension seeking resources and this kind of spillover (B = 0.36, p < 0.01). Concerning hypothesis 2a, it is supported under the conditions that the job crafting behaviour contains of the dimension seeking resources which in turn influences the positive spillover from life to work concerning the work-life balance of the employee.

For hypothesis 2b the results of the regression are also examined, a positive

relationship between work-life balance and job satisfaction is proposed. The results can be found in Table 5. For the dimension seeking resources the different spillovers are evaluated whether they are related to job satisfaction. A positive spillover from work to life (B = 0.14, p > 0.05), a negative spillover from life to work (B = -0.16, p > 0.05) and a positive spillover from life to work (B = 0.11, p > 0.05) are not shown to be significant. However, the negative spillover from work to life is significantly related to job satisfaction (B = -0.26, p < 0.05). Therefore hypothesis 2b is supported for the negative spillover from work to life in relation to job satisfaction the dimension seeking resources of job crafting. This implies that the higher the negative spillover from work to life the lower the job satisfaction of the employee. The indirect mediating effect of work-life balance on the relationship between job crafting and job satisfaction will now be discussed, also known as hypothesis 2. It is proposed that work-life balance has a positive mediating effect on this relationship. The results are

43

shown in Table 5. For the negative spillover from work to life the coefficient is B = 0.02, but is found non-significant because the confidence interval included zero (CI = -0.03; 0.14). The coefficient of the indirect effect of positive spillover from work to life is found to be B = -0.04, but is not found to be significant according to the confidence interval (CI = -0.12; 0.00). For the negative spillover from life to work the coefficient is B = 0.00, and it is

non-significant since zero is not excluded from the confidence interval (CI = -0.03; 0.07). In hypothesis 3 the moderating effect of exhaustion is proposed. It is suggested that exhaustion negatively moderates the positive relationship between job crafting and work-life balance. First the results are discussed with the negative spillover from work to life. Based on these results, the regression coefficient of XM is c3 = -0.07, but this coefficient is not found to be significant t(107) = -0.70, p > 0.05. For the positive spillover from work to life, the

regression coefficient of c3 = -0.27 is found to be significant t(107) = -2,11, p < 0.05. Whilst analysing the related effect at different percentiles of the moderator, it shows that the effect is not significant in all the different levels of the moderator. However, the effect shows to be slightly weaker at higher levels of positive spillover from work to life in comparison with lower levels. Due to these results it can therefore be stated that it is not a very strong interaction. For the negative spillover from life to work the regression coefficient is c3 = -0.05, but is found non-significant t(107) = -0.45, p > 0.05. The regression coefficient of the last spillover, positive spillover from life to work, for seeking resources is c3 = 0.03. This is also not found to be significant t(107) = 0.28, p > 0.05.

44 Table 5

45 Table 6