THE BENEFITS OF SELF-EMPLOYMENT

THE BENEFITS OF

SELF-EMPLOYMENT

Does self-employment really lead to more satisfaction

and happiness?

Master Thesis, MSc HRM

University of Groningen, Faculty of Economics and Business

June 12, 2020

ZOË NICOLAI

S3205371

Keizersmantel 28

7908 XB Hoogeveen

+31634866445

z.i.a.nicolai@student.rug.nl

Supervisor

THE BENEFITS OF SELF-EMPLOYMENT

Does self-employment really lead to more satisfaction and happiness?

ABSTRACT

The number of self-employed workers has been growing over the years, despite the fact that entrepreneurship can be a very uncertain process. Major benefits are that it brings more freedom, autonomy and flexibility. Consequently, self-employed workers have been found to have a higher job satisfaction than paid workers. Whether their subjective well-being is also higher is not yet clear. Long and irregular hours could disturb the work-life balance of the self-employed, which has consequences for their being. This paper adds to existing literature by looking further into the subjective well-being of the self-employed. The dataset of the 6th _{European Working Conditions Survey was used to}

test the hypotheses in Stata by applying regression and mediation analysis. The results show that there is no significant difference between the subjective well-being of self-employed workers and paid workers. Work-life balance has been found to be a partial mediator in this relationship.

1. INTRODUCTION

Current literature suggests that self-employed workers have a higher job satisfaction than workers who work under an employment contract (paid workers), even though they often earn less and work more hours (Binder & Coad, 2013). Twenty years ago, Blanchflower and Oswald (1998) found that half of the self-employed workers included in their research were very satisfied with their job, while only one in three paid workers felt like this. Quite a lot of these participants mentioned they wanted to start working for themselves, but that constraints kept them from doing it (Blanchflower & Oswald, 1998). Nowadays, though, work forms are changing, the gig economy is growing, and more people do start working for themselves (Kelliher, Richardson & Boiarintseva, 2010). These developments warrant a greater attention to self-employment in research.

It is not surprising that the number of self-employed workers is increasing, as these workers have been found to have a higher job satisfaction as a result of motivating working conditions. This can be explained using the self-determination theory, which suggests that humans have a basic need for autonomy, competence and relatedness (Deci & Ryan, 2000). When these needs are satisfied, motivation, performance and well-being will be increased. Since self-employment comes with a lot of freedom, autonomy and flexibility, it can be more attractive than working in an organisation (Parasuraman & Simmers, 2001). Autonomy seems to especially be playing an important explaining factor for the higher job satisfaction of the self-employed (Benz & Frey, 2008).

The high amount of control and flexibility would also suggest that self-employed workers report a higher subjective well-being than paid workers. They are often able to do their work whenever and wherever they want to, which can help structure their workday effectively for their own situation. This means that the self-employed worker would be able to reduce their work-life conflict and stress-levels, which increases their well-being (Parasuraman & Simmers, 2001). At the same time, the self-employed risk working long and irregular hours. Despite the fact that they have a lot of control over their work, the high level of job involvement and responsibility for their own success makes it hard for them to switch off their work completely. Consequently, their work-life balance could suffer from being self-employed (Binder & Coad, 2013; Kelliher et al., 2019; Parasuraman & Simmers, 2001).

It is important to note that job satisfaction and subjective well-being are not independent of each other. Rather, job satisfaction is one of the many factors that can influence the subjective well-being of a worker (Diener, Suh, Lucas & Smith, 1999). For example, Bowling, Eschleman and Wang (2010) defined subjective well-being as ‘life satisfaction, happiness, the presence of positive affect, and the absence of negative affect’. They go on to suggest that job satisfaction is related to all of these aspects as well. Subjective well-being, in turn, also seems to have a positive influence on job satisfaction (Diener et al., 1999).

In summary, the working conditions of the self-employed do not necessarily mean that everything in their lives will be balanced perfectly. There are trade-offs between the positive and negative aspects of working on one’s own (Parasuraman & Simmers, 2001). It is clear, however, that job satisfaction increases when becoming self-employed. The question at hand is whether subjective well-being will increase in the same way job satisfaction does. To find an answer to this, paid workers and self-employed workers will be compared in this paper. Additionally, this paper will look into ‘autonomy’ and ‘work-life balance’ as explaining factors for the level of job satisfaction and subjective well-being of self-employed workers. Therefore, the following questions will be examined:

Is the difference in job satisfaction between paid workers and self-employed workers bigger than the difference in their subjective well-being? If so, can this be explained by their level of work-life balance?

This paper will add to existing literature by looking further into the subjective well-being of self-employed workers. So far, very little is known about this topic, as a lot of research into the positive outcomes of work has only focussed on paid workers. The papers that did look into the subjective well-being of self-employed workers were not all convinced about the results (Kara & Petrescu, 2018). It is important to further look into this topic as well-being can significantly improve one’s life (Diener and Ryan, 2009). By including work-life balance as a mediator between self-employment and subjective well-being, this paper will offer novel insights into the line of research.

2. THEORY AND HYPOTHESES

Workers who are self-employed have a lot of autonomy, flexibility and control over their job (Parasuraman & Simmers, 2001). This means they themselves can decide how and when they do their job. However, being self-employed also requires a lot of a worker. For example, self-employed workers earn less while working longer hours (Binder & Coad, 2013). In addition, it has also been found that being self-employed can lead to higher levels of stress and grief, as entrepreneurship is an emotionally demanding and uncertain process (Shir, Nikolaev & Wincent, 2019). Self-employed workers cannot know what the future will bring, and they are personally responsible for their income. They have to make judgements about whether a certain action will bring an opportunity for profit (McMullen & Shepherd, 2006). This could affect their well-being in a negative way, as the continual search for income opportunities might disturb their work-life balance.

As self-employed and paid workers will be compared in this research, it is beneficial to define both. Self-employed workers will be defined as workers who own an enterprise and are their own boss (Parasuraman & Simmer, 2001). In addition, self-employed workers have a right on the residual claim of the enterprise. They are working for their own account and risk (Stephan & Roesler, 2010). Paid workers will be defined here as workers who work part-time or full-time under a contract of employment (Business Dictionary, 2020). These workers have a paid income by their boss.

2.2 Job satisfaction

There is a lot of research that supports the finding that self-employed workers have a higher job satisfaction than paid workers. Their job satisfaction increases no matter the reason why workers become self-employed (Josten & Vlasblom, 2017). This is an important finding, as satisfaction is positively associated with productivity, well-being and performance (Bradley & Roberts, 2004; Kara & Petrescu, 2018). In this paper, job satisfaction (JS) will be defined as ‘affective orientations on the part of individuals towards work roles which they are presently occupying’ (Vroom, 1964). The level of job satisfaction seems to be determined by the difference between an individual’s expectations of a job and the actual experiences of said job (Cooper & Artz, 1995). When the gap between expectations and experiences is high, job satisfaction is low. Thus, it seems that for self-employed workers, the actual reality of their work is close to their expectations. This could be because of selection, where the workers who become self-employed have a higher level of self-efficacy in the first place (Bradley & Roberts, 2004). This means that workers only develop interests in activities in which they perceive themselves to be successful and competent, which could increase their job satisfaction. However, there are more explanations for the high satisfaction rate of the self-employed, of which the most prominent research will be discussed next.

design of the job also seems to play a role in determining the level of job satisfaction. Benz and Frey (2008) explain this using the term ‘procedural utility’: workers do not only value the outcomes of their work, but also the conditions and processes that lead to these outcomes. This also seems to be the case with self-employment: even though self-employed workers may have lower pay and more workload than paid workers, the way they are able to do their job makes being self-employed worth it (Benz & Frey, 2008). It offers greater autonomy and flexibility, more skill utilization and a higher perceived job security (Hundley, 2001).

Job dimensions that workers especially value are discussed in the self-determination theory by Deci and Ryan (2000). This theory suggests that autonomy, competence and relatedness are basic human needs that need to be fulfilled in order to be satisfied with a job. By becoming self-employed, workers get a lot more freedom in their work and their autonomy in decision-making increases. The fact that these workers are their own boss will probably also increase their feeling of competence, as they are able to apply their talent, skills and knowledge wherever they can and want to (Hundley, 2001). Lastly, because self-employed workers are solely responsible for the success of their enterprise, the relatedness to and active engagement with their job increases (Hilbrecht & Lero, 2014). As being self-employed satisfies all three basic human needs, their job satisfaction increases considerably (Deci & Ryan, 2000).

Overall, the high level of autonomy of self-employed workers experience has been found to be the most prominent explaining factor for the difference in job satisfaction between self-employed workers and paid workers (Josten & Vlasblom, 2017). Therefore, autonomy will be included in this study as a mediator between self-employment and job satisfaction. It will be defined as having the freedom and independence to determine one’s own course of action. In accordance to the abovementioned, the following will be hypothesized:

H1: Job satisfaction is higher for self-employed workers than for paid workers.

H2: The difference in job satisfaction between self-employed workers and paid workers can be explained by the level of autonomy they have in their job.

2.3 Subjective well-being

Compared to job satisfaction, the findings on the subjective well-being of self-employed workers are less clear and relatively limited. However, it is evident that self-employment can influence the subjective well-being of a worker positively as well as negatively. Yet, the overall finding is that the self-employment is primarily positively related to a worker’s subjective well-being (Sevä, Vinberg, Nordenmark & Strandh, 2016). Moreover, over twenty years ago Blanchflower and Oswald (1998) already found evidence that the self-employed are generally happier than paid workers.

emotions (Wiklund, Nikolaev, Shir, Foo & Bradley, 2019). High levels of subjective well-being have been found to significantly improve health, work and income, social relations and societal benefits of happiness (Diener and Ryan, 2009). As subjective well-being is a very broad concept, there are many factors that can influence it. Hence, certain aspects of self-employment cause self-employed workers to report a higher subjective well-being than paid workers. For instance, self-employed workers have favourable working conditions, where workers have very high autonomy and control as they are the owner of the enterprise (Nordenmark, Vinberg & Strandh, 2012; Stephan & Roesler, 2010). Consequently, high levels of autonomy have been found to decrease burn-out complaints, especially combined with having an interesting job (Pot & Smulders, 2019). Hence, self-employed workers often experience lower stress levels than paid workers (Parasuraman & Simmers, 2001). According to Karasek (1979), these high levels of control also enable them to deal with higher potential job demands. Therefore, the self-employed are less likely to find their work mentally straining (Andersson, 2008).

Then again, self-employed workers are more involved with and responsible for their work, and thus risk working longer and more irregular hours than paid workers. This is also called the paradox of self-employment: despite having more control over their work, the self-employed can find it harder to switch off their work entirely (Binder & Coad, 2013; Kelliher et al., 2019; Parasuman & Simmers, 2001). They are solely responsible for the success of their company, monthly income and eventual profit, which is why they are continually pursuing income opportunities (Hilbrecht & Lero, 2014; McCullen & Shepherd, 2006). As a result, boundaries between work and personal life might become disturbed, which can lead to stress. Without recovery, these elevated stress levels could lower well-being and even result in a burnout (Pot & Smulders, 2019). Besides that, the fact that entrepreneurship is filled with insecurity may further increase stress-levels (Bradley & Roberts, 2014; Shir et al., 2018). Because the future is hard to predict, self-employed workers often do not have the security of making profit and earning an income each month (McMullen & Shepherd, 2006).

and even burn-out complaints (Pot & Smulders, 2019). Understandably, this could decrease one’s positive outlook on life.

The fact that self-employed workers have a lot of control over when, where and how they do their work indicates that they have the opportunity to design their workday in such a way that it does not interfere with their life outside of their work. However, they also risk working longer hours than paid workers. Research done by Hughes and Parkes (2007) suggests that the amount of control a worker has over their work can partly make up for the negative effect of the longer hours one makes in terms of work-life balance. The job demand-control model by Karasek (1979) supports this by suggesting that when a worker can make a lot of decisions by themselves, they are better equipped to deal with high job demands such as long working hours. Therefore, in spite of their long and irregular hours, it will be assumed that self-employed workers have a higher subjective well-being than paid workers. As a result of the abovementioned, the following is hypothesized:

H3: Subjective well-being is higher for self-employed workers than for paid workers.

H4: The difference in subjective well-being between self-employed workers and paid workers can be explained by their level of work-life balance.

2.4 Differences in SWB and JS

Job satisfaction and subjective well-being are two concepts that are very much related. In fact, job satisfaction is one of the many factors that influences subjective well-being (Diener et al., 1999). Therefore, it is assumed that the subjective well-being of the self-employed would be higher than that of paid worker, as their job satisfaction is also higher. This has also been supported by research: self-employed workers overall have been found to be happier and thus to have a higher subjective well-being than paid workers (Blanchflower & Oswald, 1998; Sevä et al., 2016; Stephan & Roesler, 2010). However, the difference in the subjective well-being of self-employed workers and paid workers might not be as prominent as the difference in their job satisfaction. An important indicator for this is that the work-life balance of the self-employed seems to be worse than that of paid workers. Their higher job demands and workload cause them to work longer during irregular working hours. They most likely also work outside of set working hours and on weekends (Baines & Gelder, 2003). This leads to a disturbed work-life balance, as the different aspects of one’s life are no longer segmented (Sirgy, Efraty, Siegel & Lee, 2001). This can be particularly hard for parents, who need to juggle between these income opportunities, long hours and being readily accessible to their children and other domestic responsibilities (Hilbrecht & Lero, 2014; Kelliher et al., 2019). When a work-life balance is disturbed like this, all life domains can be negatively affected (Allen et al., 2000). Paid workers will not likely have a disturbed work-life balance like this, as they often only work when they are actually at work and during set working hours. Therefore, the following hypothesis is posed in this paper:

To conclude, the mediation models that will be examined in this paper are portrayed in Figure 1 and 2. In Figure 1, all paths (A, B and C) are predicted to be positive. In Figure 2, path A is predicted to be negative, while path B and C are predicted to be positive. For there to be a full mediation effect, however, the C paths should not be significant.

Figure 1. Mediation model 1

Figure 2. Mediation model 2

3. METHODS

In order to examine the hypotheses posed in this paper, empirical research was done using an existing dataset. Relevant questions were selected with which the concepts were operationalised. The data from the dataset was tested using Stata.

3.1 Data collection

The dataset of the 6th _{European Working Conditions Survey (EWCS) was used to test the}

hypotheses. This survey includes information about the working lives of the respondents, with topics such as working conditions, satisfaction, health, work-life balance, pay and well-being. The concepts of this paper are related to these topics and were thus operationalised using relevant questions from the survey. The 6th _{EWCS took place in 2015 in thirty-five countries and includes a total of 43.850 worker}

respondents. This sample represented the entire employed population of the included countries.

3.2 Research population

Selected from the sample of the 6th _{EWCS were all workers from the age of 25 until 65 that live}

in the EU15. The total population existed of 19706 workers, however after deleting missing answers on the relevant variables 15458 observations remained. The age range of 25-65 was chosen specifically to filter out student workers. Students, or younger workers, most likely have a different perspective on and

B A

C

Self-Employment Subjective Well-Being Work-Life Balance

Autonomy

A B

C

experience with work, as their job does not play as big of a role in their life yet. Besides that, they often have a temporary or zero-hour contract, which also contributes to them having a different working experience than the average worker such as a higher job insecurity (WRR, 2020). The EU15, the first 15 EU countries (Austria, Belgium, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg, The Netherlands, Portugal, Spain, Sweden, United Kingdom), was selected here because these countries have similar working conditions, perspectives and working hours (Eurofound, 2015). In addition, their economies are integrated with each other and have a similar structure as a result of working together for a long time (Hansson & Olofsdotter, 2013).

3.3 Measurement of variables

With the intention of testing the hypotheses, the concepts discussed in the theory section were operationalised. To do this, questions and statements from the EWCS were selected based on their relevance to the concepts. The selected questions and the corresponding answer options/scales are disclosed in Appendix A. In addition, some variables were constructed by combining several statements, for which all variables met the minimum requirement of scale reliability. The specific outcomes can be found in Appendix B, along with more precise information about how the variables were formed.

Two questions were used to make a distinction between self-employed workers and paid

workers. The first question examines directly whether the worker is self-employed or an employee, and

the second question looks at employment contract. Based on these questions the eventual distinction was made between self-employed workers, permanent workers and temporary workers. To measure job

satisfaction, one question and a statement was used. The question asks directly whether the worker is

satisfied with their working conditions or not, while the statement examines how the worker feels about their job. Autonomy was measured using the question whether workers are able to choose or change the order of their tasks, work methods and speed of their job. Subjective well-being was measured using a series of statements about how the worker has been feeling in the past two weeks. Lastly, a question and a series of statements were used to operationalise work-life balance. The questions examine whether the worker is able to fit in their working hours with their commitments outside of work, while the series of statements examines if there have been conflicts between the participant’s work and non-work life over the last 12 months.

3.4 Analysis

The program Stata was used to test the hypotheses. Before any tests could be done, however, the dataset was prepared by selecting the population and the relevant questions, removing missing answers and recoding variables. Appendix B shows how the variables were recoded and how the dataset was prepared. After that, descriptive statistics (mean, standard deviation, correlation) were done to describe the research population and their scores on the main and control variables. Correlation was used to have a first indication whether and how the concepts are related to each other. To test the actual hypotheses posed in this paper, multiple analyses were done. Firstly, one-way anova was done to test the predicted differences between self-employed and paid workers. After that, regression analyses were done on ‘subjective well-being’ and ‘job satisfaction’, which also already gave an indication of mediation effects. Next, structural equation models were used to see the examine the exact effect of the mediators. Lastly, the nonlinear combinations of estimators were used to explicitly test H5. Within these analyses, country-fixed effects were controlled and clustered variances at country levels were used to further control for differences between countries.

4. RESULTS

Descriptive statistics and correlations were done first to get some early indications about the variables. The findings for the main variables can be found in Table 1 and those for the control variables in Appendix C. Consequently, to see the differences in variables between employment types, means and correlations were also examined per employment type. These findings can be found in Table 2.

A few things become clear from Table 1. In this table we can see that the different employment types correlate with most of the main variables, except for subjective well-being. This could indicate that the subjective well-being of workers might not be influenced by the type of employment they have. In addition, the correlations also already show signs of mediation. The different employment types are correlated with autonomy and autonomy is also correlated with job satisfaction. And while not all employment types are correlated with subjective well-being, they are correlated with work-life balance and work-life balance is correlated with subjective well-being. This could indicate a full mediation effect of work-life balance.

TABLE 1

Descriptive Statistics of Main Variables

Variable Mean S.D. Correlation

S-E P.C. T.C. JS1 JS2 SWB Autono my WLB1 WLB2 Self-employed 0.135 0.342 1 Permanent 0.764 0.425 -0.711* 1 Temporary 0.101 0.301 -0.132* -0.603* 1 JS1 3.115 0.689 0.019* 0.020* -0.050* 1 JS2 3.960 0.905 0.074* -0.022* -0.054* 0.456* 1 SWB 4.455 0.992 0.010 -0.017* 0.013 0.381* 0.349* 1 Autonomy 0.721 0.369 0.186* -0.076* -0.104* 0.128* 0.184* 0.081* 1 WLB1 3.082 0.766 -0.066* 0.066* -0.019* 0.308* 0.161* 0.247* 0.109* 1 WLB2 3.567 0.899 -0.136* 0.103* 0.093 0.274* 0.102* 0.354* -0.022* 0.468* 1 N = 15458 *p < 0.05 TABLE 2

Correlation Between Main Variables Sorted by Employment Type

Variable N Mean S.D. Correlation

WLB2 3.618 0.870 0.299* 0.114* 0.383* 0.007 0.460* 1 Temporary contract 1559 (10.09%) JS1 3.013 0.741 1 JS2 3.815 1.041 0.456* 1 SWB 4.493 1.019 0.280* 0.263* 1 Autonomy 0.606 0.409 0.138* 0.149* 0.034 1 WLB1 3.085 0.803 0.289* 0.196* 0.209* 0.113* 1 WLB2 3.592 0.940 0.238* 0.118* 0.279* -0.203* 0.461* 1 *p < 0.05

4.2 Mean Variable Differences Between Employment Type

To see whether the differences found in Table 2 are significant, one-way anova was used. Firstly, the anova indicated that there is a significant effect of employment type on job satisfaction for both the first job satisfaction measure (F (2, 15455) = 20.37, p < 0.001) and the second job satisfaction measure (F (2, 15455) = 58.18, p < 0.001). This means that the experienced job satisfaction by self-employed workers is higher than that of both workers with a permanent contract and workers with a temporary contract, which supports H1. Secondly, there was also a significant effect found of employment type on the level of autonomy of workers (F (2, 15455) = 329.20, p < 0.001). Self-employed workers seem to have a higher level of autonomy than workers with a permanent contract and with a temporary contract. This seems to support H2. Thirdly, there was also a significant effect found of employment type on the level of work-life balance workers experience for both the first work-life balance measure (F (2, 15455) = 39.61, p <0.001) and the second work-life balance measure (F (2, 15455) = 146.58, p < 0.001). Workers with a permanent contract and workers with a temporary contract both experience a higher level work-life balance than self-employed workers, which seems to support H4. Lastly, there was no significant effect found of employment type on the subjective well-being of workers (F (2, 15455) = 2.25, p = 0.105). This is an interesting indication, as this does not support H3. Then again, it indicates that H5 might be supported by the findings.

4.3 Mediation Analysis

Regression analyses were used to identify which effects the independent variables (employment types, mediators, control variables) have on the dependent variables ‘job satisfaction’ and ‘subjective well-being’. These analyses also already gave an indication of any mediation effects. In addition to the regression analyses, structural equation models were used to see what the exact mediation effects are of ‘autonomy’ and ‘work-life balance’.

Appendix D. Looking at the model overall, R2 _{seems to increase from model 1 through model 4, which}

means that every model explains more of the variation in the dependent variable ‘job satisfaction’. Looking at the direct effect of the independent on the dependent variable, also called path C of the mediation model (model 1), there only seems to be a significant direct effect of employment type on job satisfaction with workers who do temporary work using the first job satisfaction measure. The regression analysis using the second measure, which can be found in Appendix D, indicated a significant direct effect for both employment types. These effects were all negative, which means that in comparison to self-employed workers, paid workers (both permanent and temporary) have a lower job satisfaction. So, path C of the mediation model is partly significant. Looking at path A (model 2) of the mediation model, there is a significant effect of employment type on the mediator ‘autonomy’. As can be seen in model 2, while controlling for the control variables, autonomy seems to decrease for paid workers when comparing it to self-employed workers. Lastly, path B of the mediation model (model 4) has also been found to be significant with both measures of job satisfaction. This path measures the effect of the mediator ‘autonomy’ on the dependent variable ‘job satisfaction’, while controlling for the independent variable ‘employment type’ and the control variables.

Table 4 shows the direct, indirect and total effects of employment type on job satisfaction, with autonomy as the mediator. From theseresults it can be concluded that there is a partial mediating effect of autonomy in the relationship between employment type and job satisfaction. This is in line with the regression analysis. Comparing the total effects with the indirect effects, about 20-25% of the total effect is explained by the indirect effect (the mediator autonomy), which means that H2 is supported.

TABLE 3

Regression Analysis of Employment Type on Job Satisfaction (1) Mediated by Autonomy1

Dependent variable

JS 1 Autonomy JS 1 JS 1

Model 1 Model 2 Model 3 Model 4

Household nr 0.008* 0.002 0.018* 0.004 Age 0.001** 0.000 -0.002 0.001 Working hours 0.001 0.000 -0.002** 0.001 Competence 0.003 0.008 0.005 0.010 Education level 0.033* 0.002 -0.001 0.006 Relatedness -0.004 0.002 0.062* 0.005 Income 0.024* 0.005 0.117* 0.010 More hours 0.009 0.009 -0.093** 0.025 Less hours -0.003 0.006 -0.221* 0.018 R Squared 0.029 0.112 0.049 0.131 St. Error or Estimate 0.680 0.348 0.673 0.643 N = 15458 *p < 0.001 **p < 0.05

1 _{Country fixed effects and variances clustered by countries}

TABLE 4

Direct, Indirect and Total Effects of Employment Type on Job Satisfaction (1) Mediated by

Autonomy1

Direct effects Indirect effects Total effects

B S.E. Z B S.E. Z B S.E. Z

Job satisfaction 1 Autonomy 0.237 0.023 10.08* 0 0.237 0.023 10.08* Permanent contract -0.149 0.049 -3.01* -0.048 0.006 -7.78* -0.197 0.047 -4.17* Temporary contract -0.149 0.048 -3.11* -0.062 0.008 -7.57* -0.211 0.050 -4.22* N = 15458 *p < 0.05

1 _{Country fixed effects and variances clustered by countries}

Table 5 and 6 show the findings of the regression analysis and direct, indirect and total effects of the relationship between employment type and subjective well-being, where work-life balance is used as a mediator. Only the findings using the first measure of work-life balance are included here. The findings for the second measure can be found in Appendix E.

Firstly, looking at Table 5, R2 _{increases with each model, so more of the variation in subjective}

type on subjective well-being (model 1), which means path C of the mediation model is not significant. This means that there is no significant difference in the subjective well-being between workers, hence H3 is not supported. Looking at path A (model 2), employment type seems to have a significant negative effect on work-life balance. Consequently, looking at path B (model 4), while controlling for both employment type and the control variables, work-life balance has a significant mediation effect on the relationship between the employment type and subjective well-being of a worker. However, this is only significant for permanent workers.

The results from Table 6 show that there is a partial mediating effect of work-life balance on the relationship between employment type and subjective well-being. There are significant direct effects, indirect effects and total effects in this mediation analysis. The indirect effect is not a major portion of the total effects, about 15-20%, but the mediation effect is still there. Hence, H4 is supported.

TABLE 5

Regression Analysis of Employment Type on Subjective Well-Being Mediated by Work-Life Balance (1)1

Dependent variable

SWB WLB 1 SWB SWB

Model 1 Model 2 Model 3 Model 4

Children -0.080** 0.019 -0.010 0.017 More hours -0.119** 0.027 -0.029 0.026 Less hours -0.250* 0.024 -0.164* 0.017 R Squared 0.029 0.194 0.092 0.240 St. Error or Estimate 0.979 0.688 0.946 0.866 N = 15458 *p < 0.001 **p < 0.05

1 _{Country fixed effects and variances clustered by countries}

TABLE 6

Direct, Indirect and Total Effects of Employment Type on Subjective Well-Being Mediated by Work-Life Balance (1)1

Direct effects Indirect effects Total effects

B S.E. Z B S.E. Z B S.E. Z

Subjective well-being Work-life balance (1) 0.216 0.015 14.71* 0 0.216 0.015 14.71* Permanent contract -0.101 0.026 -3.82* -0.014 0.006 -2.48* -0.115 0.028 -4.11* Temporary contract -0.062 0.036 -1.72 -0.017 0.007 -2.43* -0.080 0.035 -2.27* N = 15458 *p < 0.05

1 _{Country fixed effects and variances clustered by countries}

4.4 Differences in Job Satisfaction and Subjective Well-Being of Self-Employed Workers and Paid Workers

Within the structural equation model, differences between equations can be tested. This was done using the nonlinear combination of estimators, of which the results are shown in Table 7. The purpose of this was to see whether the difference in the job satisfaction of the self-employed and paid workers is significantly bigger than the difference in their subjective well-being. Table 7 exists of 4 different parts. The first line (1) is the estimator done with the first measure of job satisfaction and the second line (2) with the second measure of job satisfaction. The relevant control variables were included in line 3 and 4. In the table, workers with a permanent contract are used as a reference group to compare with self-employed workers.

self-employed and (permanent) paid workers is significantly bigger than the difference in their experienced job satisfaction. With these results, H5 is supported.

TABLE 7

Test of Differences Between Job Satisfaction and Subjective Well-Being1

B S.E. Z 1 (JS1) 0.084 0.070 1.20 2 (JS2) 0.291 0.080 3.62* 3 (JS1 + Control) 0.131 0.582 2.25** 4 (JS2 + Control) 0.327 0.064 5.12* N = 15458 *p < 0.001 **p < 0.05

1 _{Country fixed effects and variances clustered by countries}

5. DISCUSSION AND CONCLUSION

The aim of this paper was to add to current literature on self-employed workers by finding an answer to the research questions ‘Is the difference in job satisfaction between paid workers and

self-employed workers bigger than the difference in their subjective well-being? If so, can this be explained by their level of work-life balance?’ using several hypotheses. The short answer to these questions is

yes, the difference in job satisfaction between self-employed workers and paid workers is bigger than the difference in subjective well-being. This can be explained by their level of work-life balance. The answers to these questions will be elaborated on in this section by discussing the hypotheses, along with the implications of these findings.

5.1 The Findings

Compared to paid workers, self-employed workers have more opportunities to apply their talent, skills and knowledge (Hundley, 2001). Concluding, self-employed workers have a higher job satisfaction than paid workers, which can be partly explained by their higher level of autonomy.

The third and fourth hypotheses were used to test if self-employed workers and paid workers differ in their subjective well-being and whether this difference could then be explained by their work-life balance. From the one-way anova and the regression analysis it became apparent that there is no significant direct relationship between employment type and the level of subjective well-being, thus self-employed workers do not significantly differ from paid workers in terms of their subjective well-being. As the subjective well-being of a worker is a very broad concept, it can be influenced by many aspects of work. The paradox of self-employment may explain here why the subjective well-being of self-employed workers is not necessarily higher than that of paid workers (Kelliher et al., 2019). Despite having more control and freedom in their work, the self-employed work long and irregular hours. From the results it also became clear that there is a significant indirect effect in the relationship between employment type and subjective well-being when work-life balance and control variables are taken into account. Work-life balance was found to have a partial mediating effect on this relationship: it explains about 15-20% of the total effects. This means that the lack of difference in subjective well-being between self-employed workers and paid workers can be partly explained by the lower level of work-life balance of self-employed workers. It also supports current literature, which already indicated that work-life balance is associated with well-being (Greenhaus et al., 2003).

The last, and most important, hypothesis was used to examine whether the difference in job satisfaction between self-employed workers and paid workers was bigger than the difference in their subjective well-being. The outcomes of the nonlinear combination of estimators test supported this hypothesis, which means the answer to the first research question is ‘yes’. This is not surprising when looking at the earlier findings, as there was no significant difference in the subjective well-being of the different types of workers in the first place. Hence, the difference in job satisfaction is understandably bigger. The second research question can also be answered with a ‘yes’, as the work-life balance of paid workers has been found to be significantly higher than that of self-employed workers. In addition, as mentioned earlier, work-life balance is a partial mediator for the relationship between employment type and subjective well-being. For this reason, it is likely that the lack of significance in difference of subjective well-being is influenced by their level of work-life balance.

5.2 Theoretical Implications

The aim of this paper was to add to current literature about (the subjective well-being of) self-employed workers. This was successfully done by indicating that self-self-employed workers do not seem to have a significantly higher subjective well-being than paid workers with either a permanent or temporary contract. This is partly due to their lower level of work-life balance. In addition, support was found for research that indicated that self-employed workers have a higher job satisfaction than paid workers, which is partly due to their high levels of autonomy. However, the mediation effects found in this study were only partial. In other words, job satisfaction and subjective well-being are influenced not only by the mediators tested here, but also by other variables that have not been included. So, while this paper contributed to current literature by finding significant (mediation) results, there is still a lot to discover when it comes to the subjective well-being of self-employed workers.

The fact that only partial mediation found in this study could indicate that the theoretical model might not be completely correct. Looking at the results, it seems that job satisfaction has a high correlation with both work-life balance and subjective well-being. Hence, instead of the two separate relationships that were tested here, it could be that all variables are connected in some way. For example, self-employment might influence work-life balance and subjective well-being via job satisfaction, which means that one big theoretical model could be more appropriate. This seems logical when taking into account the research by Diener et al. (1999), which indicated that job satisfaction is one of the factors that influences subjective well-being. Future research might offer new insights into these potential theoretical models.

An interesting finding from the regression analyses was that the effect of employment types increased when taking into account the control variables. One would think that this effect would decrease, as more variables were taken into account. It seems that the control variables only intensify the effects that self-employment has on the dependent variables, which suggests some hidden effect of the control variables. For instance, the majority of self-employed workers included in this study were men. Men have been found to dedicate more time on work than on family when being self-employed, while women dedicate more time on their children when they are self-employed (Baines & Gelder, 2003; Craig, Powell and Cortis, 2012). Parasuraman and Simmer (2001) found gender to be significantly related to job involvement among self-employed workers. Hence, control variables like gender could have played a bigger role in the examined relationships than initially anticipated.

5.3 Practical Implications

research might help self-employed workers in becoming more aware of their work-life balance. If they recognise that it might be disturbed, they can make changes in their current working style so that their (subjective) well-being will not be endangered any further. In addition, given the importance of the self-employed workers for the economy, organisations like the government could use this study and other literature to help and encourage self-employed workers. For example, they could offer self-employed workers a helping hand in improving their work-life balance, such as by compiling a guide to become aware of their working style and how to improve it to ensure work-life balance. Not only would these organisations then promote entrepreneurship (by using programmes like the Entrepreneurship Action Plan 2020 by the European Commission), but they could also then help these self-employed workers preserve their subjective well-being. This, however, will only be possible if research into this topic will be continued.

5.4 Strengths and Limitations

As for any study, there are some strengths and limitations that need to be discussed. The first limitation of this study is that the research population is only representable for the fifteen countries that were selected, as they have a similar economic and social climate. Countries with similar climates might get the same results, however it is likely that countries outside of the EU and Europe will score differently on the survey. According to Bosma and Schutjens (2011), entrepreneurial attitudes differ between countries as it is a component of a culture. Hence, the results found in this paper will not be representable for every country. Secondly, the findings were based on an already existing dataset. The questions used to measure variables were set and could not be changed according to the specific definitions that were used in this paper. Hence, the selected questions and their answer scales might not have fully connected with the exact definitions that were used. The findings in this study could therefore have been slightly different if a questionnaire was made based on the exact definitions of the concepts that were discussed in the theory section.

The major strength of this paper is that it adds to the current literature into self-employed workers, especially into their subjective well-being. Thus far, existing studies have been scarce and unconvincing (Kara & Petrescu, 2018). This study used a big dataset with a sample of workers from thirty-five countries, which represents the entire employed population of these countries. The eventual research population was also big: it existed of fifteen of these countries, which means this study included around 15000 workers. These workers were assumed to be representable for the entire population of these fifteen countries.

5.5 Future research

6. REFERENCES

Allen, T. D., Herst, D. E., Bruck, C. S., & Sutton, M. (2000). Consequences associated with work-to-family conflict: a review and agenda for future research. Journal of occupational health

psychology, 5(2): 278.

Andersson, P. (2008). Happiness and health: Well-being among the self-employed. The Journal of

Socio-Economics, 37(1): 213-236.

Baines, S., & Gelder, U. (2003). What is family friendly about the workplace in the home? The case of self‐employed parents and their children. New Technology, Work and Employment, 18(3): 223-234.

Benz, M., & Frey, B. S. (2008). Being independent is a great thing: Subjective evaluations of self‐ employment and hierarchy. Economica, 75(298): 362-383.

Binder, M., & Coad, A. (2013). Life satisfaction and self-employment: a matching approach. Small

Business Economics, 40(4): 1009-1033.

Blanchflower, D. G., & Oswald, A. J. (1998). What makes an entrepreneur? Journal of labor

Economics, 16(1): 26-60.

Bosma, N., & Schutjens, V. (2011). Understanding regional variation in entrepreneurial activity and entrepreneurial attitude in Europe. The Annals of regional science, 47(3): 711-742.

Bowling, N. A., Eschleman, K. J., & Wang, Q. (2010). A meta‐analytic examination of the relationship between job satisfaction and subjective well‐being. Journal of Occupational and Organizational

Psychology, 83(4): 915-934.

Bradley, D. E., & Roberts, J. A. (2004). Self‐employment and job satisfaction: investigating the role of self‐efficacy, depression, and seniority. Journal of small business management, 42(1): 37-58.

employed. BusinessDictionary.com. Accessed January 12, 2020, via: http://www.businessdictionary.com/definition/employed.html

Cooper, A. C., & Artz, K. W. (1995). Determinants of satisfaction for entrepreneurs. Journal of Business

Venturing, 10(6): 439-457.

Craig, L., Powell, A., & Cortis, N. (2012). Self-employment, work-family time and the gender division of labour. Work, employment and society, 26(5): 716-734.

Cromie, S. (1987). Motivations of aspiring male and female entrepreneurs. Journal of Organizational

Behavior, 8(3): 251-261.

Diener, E., Suh, E. M., Lucas, R. E., & Smith, H. L. (1999). Subjective well-being: Three decades of progress. Psychological bulletin, 125(2): 276.

Deci, E. L., & Ryan, R. M. (2000). The" what" and" why" of goal pursuits: Human needs and the self-determination of behavior. Psychological inquiry, 11(4): 227-268.

Diener, E., Lucas, R. E., & Oishi, S. (2002). Subjective well-being: The science of happiness and life satisfaction. Handbook of positive psychology, 2: 63-73.

Diener, E., & Ryan, K. (2009). Subjective well-being: A general overview. South African journal of

psychology, 39(4): 391-406.

Eurofound (2015). Sixth European Working Conditions Survey. Accessed 22 February, 2020.

European Commission (2017). Internal market, industry, entrepreneurship and SMEs. Accessed 4 June, 2020 via http://ec.europa.eu/growth/smes/promoting-entrepreneurship/action-plan/.

Greenhaus, J. H., Collins, K. M., & Shaw, J. D. (2003). The relation between work–family balance and quality of life. Journal of vocational behavior, 63(3): 510-531.

Hansson, Å. M., & Olofsdotter, K. (2013). FDI, taxes and agglomeration economies in the EU15. Applied Economics, 45(18): 2653-2664.

Hughes, E. L., & Parkes, K. R. (2007). Work hours and well-being: The roles of work-time control and work–family interference. Work & Stress, 21(3): 264-278.

Hundley, G. (2001). Why and when are the self‐employed more satisfied with their work?. Industrial

Relations: A Journal of Economy and Society, 40(2): 293-316.

Josten, E., & Vlasblom, J. D. (2017). Maakt zzp'er worden tevreden?. Tijdschrift voor

Arbeidsvraagstukken, 33(3).

Kara, A. & Petrescu, M. (2018). Self-Employment and its Relationship to Subjective Well-Being.

International Review of Entrepreneurship, 16(1): 115-140.

Karasek Jr, R. A. (1979). Job demands, job decision latitude, and mental strain: Implications for job redesign. Administrative science quarterly: 285-308.

Kelliher, C., Richardson, J., & Boiarintseva, G. (2019). All of work? All of life? Reconceptualising work‐life balance for the 21st century. Human Resource Management Journal, 29(2): 97-112. McMullen, J. S., & Shepherd, D. A. (2006). Entrepreneurial action and the role of uncertainty in the

theory of the entrepreneur. Academy of Management review, 31(1): 132-152.

Nordenmark, M., Vinberg, S., & Strandh, M. (2012). Job control and demands, work-life balance and wellbeing among self-employed men and women in Europe. Vulnerable Groups &

Inclusion, 3(1): 18896.

Parasuraman, S., & Simmers, C. A. (2001). Type of employment, work–family conflict and well‐being: a comparative study. Journal of Organizational Behavior: The International Journal of

Industrial, Occupational and Organizational Psychology and Behavior, 22(5): 551-568.

Pot, F., & Smulders, P. (2019). Arbeidskwaliteit moet een indicator voor brede welvaart zijn. METEN

VAN WELVAART: 42.

Sevä, I. J., Vinberg, S., Nordenmark, M., & Strandh, M. (2016). Subjective well-being among the self-employed in Europe: macroeconomy, gender and immigrant status. Small Business

Shir, N., Nikolaev, B. N., & Wincent, J. (2019). Entrepreneurship and well-being: The role of psychological autonomy, competence, and relatedness. Journal of Business Venturing, 34(5): 105875.

Sirgy, M. J., Efraty, D., Siegel, P., & Lee, D. J. (2001). A new measure of quality of work life (QWL) based on need satisfaction and spillover theories. Social indicators research, 55(3): 241-302. Sirgy, M.J., & Lee, D.J. (2018). Work-life balance: An integrative review. Applied Research in Quality

of Life, 13(1): 229-254.

Stephan, U., & Roesler, U. (2010). Health of entrepreneurs versus employees in a national representative sample. Journal of occupational and organizational psychology, 83(3): 717-738.

Wiklund, J., Nikolaev, B., Shir, N., Foo, M. D., & Bradley, S. (2019). Entrepreneurship and well-being: Past, present, and future.

WRR. (2020). Het betere werk. De nieuwe maatschappelijke opdracht. Accessed on 14 February, 2020.

7. APPENDIX

Research population Using the logged Country

Q2b. Starting with yourself, how old are you?

Self-employed or paid employee

Q7. Are you working as an employee or are you self-employed?

1. An employee 2. Self-employed Job satisfaction

Q88. On the whole, are you very satisfied, satisfied, not very satisfied or not at all satisfied with working conditions in your main paid job?

1. Very satisfied 2. Satisfied

3. Not very satisfied 4. Not at all satisfied

Q90. The following statements are about how you feel about your job. For each statement, please tell me how often you feel this way... (1=always, 5=never)

B. I am enthusiastic about my job Autonomy

Q54. Are you able to choose or change... (1=yes, 2=no)

A. Your order of tasks B. Your methods of work C. Your speed or rate of work Subjective well-being

Q87. Please indicate for each of the five statements which is the closest to how you have been feeling over the last two weeks. (1=all of the time, 6= at no time)

A. I have felt cheerful and in good spirits B. I have felt calm and relaxed

C. I have felt active and vigorous D. I woke up feeling fresh and rested

Work-life balance

Q44. In general, how do your working hours fit in with your family or social commitments outside work?

1. Very well 2. Well

3. Not very well 4. Not at all well

Q45 How often in the last 12 months, have you...? (1=always, 5=never)

A. Kept worrying about work when you were not working

B. Felt too tired after work to do some of the household jobs which need to be done C. Found that your job prevented you from giving the time you wanted to your family

Control variables

Gender

Q2a. (Gender coded)

Household

Q1. Including yourself, can you please tell me how many people live in this household? Q3b. How old is he/she? (household member)

Q3c. What is this person’s relationship to you?

1. Spouse//partner 2. Child

Employment contract (paid workers only)

Q11. What kind of employment contract do you have in your main job?

1. Contract of unlimited duration 2. Contract of limited duration

3. A temporary employment agency contract Education (competence)

Q106. What is the highest level of education or training you have successfully completed?

1. Early childhood education 2. Primary education

3. Lower secondary education 4. Upper secondary education

5. Post-secondary non-tertiary education 6. Short-cycle tertiary education

7. Bachelor or equivalent 8. Master or equivalent 9. Doctorate or equivalent Relatedness

Competence

Q64. Which of the following statements would best describe your skills in your own work?

1. I need further training to cope well with my duties 2. My present skills correspond well with my duties 3. I have the skills to cope with more demanding duties Income

Q100. Thinking about your household’s total monthly income, is your household able to make ends meet…?

1. Very easily 2. Easily 3. Fairly easy

4. With some difficulty 5. With difficulty 6. With great difficulty Work pressure

Q46. Over the last 12 months, how often have you worked in your free time to meet work demands?

1. Daily

2. Several times a week 3. Several times a month 4. Less often

5. Never Job security

Q89. To what extent do you agree or disagree with the following statements about your job?(1= strongly agree, 5=strongly disagree)

H. If I were to lose or quit my current job, it would be easy for me to find a job of similar salary. Health

Q75. How is your health in general? Would you say it is…

1. Very good 2. Good 3. Fair 4. Bad 5. Very bad

Actual vs. Desired working hours

Q24. How many hours do you usually work per week in your main job?

APPENDIX B: Preparation of the Dataset

Before doing tests in Stata, the dataset needed to be prepared. This was done by recoding and renaming the variables, selecting the population and deleting irrelevant variables from the dataset. Lastly, missing variables were deleted.

Recoding the Variables

Most variables that were used in the tests were recoded. If this wasn’t necessary, the variable was just named according to what it represented. The recoding of these variables will be shown here.

Gender

For gender there was a dummy variable made, where 0 represents male and 1 represents female.

Household

To get more insight into the household, the variable ‘partner’ and variable ‘children’ was made. For partner, 0 represents being single and 1 represents having partner. The variable children looked into whether the participant had a child who was younger than twelve years old, as these children often need more attention from their parents. Here, 0 represents having no children (or older ones) and 1 represents having children who are younger than twelve years old.

Employment type

The variable employment type is generated using Q7 and Q11. Here, 1 represents being self-employed, 2 represents having a permanent contract and 3 represents having a temporary contract.

Actual working hours vs. preferred working hours

To check whether the participants wanted to work a similar number of hours than they currently have, less or more there were two variables made. The first variable signifies if a participant wants to work more hours and the second variable signifies if the participant wants to work less hours. For both a dummy variable was made, where 0 represents ‘no' and 1 represents ‘yes’.

Work-life balance

all statements were taken together and then divided by three so the variable represents an average of the three statements.

Work pressure

For work pressure Q46 was used. To be used as a variable, the question scale was recoded so 1 represented a low work pressure and 5 represented high work pressure.

Autonomy

The question scale of Q54abc was also recoded so that 0 represented low autonomy and 1 represented high autonomy. However, since there will be three statements used from this question, the Cronbach’s alpha was used to check whether they could be made in one variable. It was 0.7604, which meets the requirement of 0.7. Therefore, one variable was made, where all statements were summed up and then divided by three.

Relatedness

To control for relatedness, Q61a was used. However, the question was recoded so that 1 represented low relatedness and 5 represented high relatedness.

Health

Q75 was used to control for the health of the participants. The question was recoded so that the variable health represented 1 for very bad health and 5 for very good health.

Subjective well-being

For subjective well-being, five statements of Q87 were used (a until e). These five statements were first recoded so that 1 represented low subjective well-being and 6 represented high subjective well-being. After this was done, Cronbach’s alpha was used to check if the five statements could be made into one variable. And this was the case: Cronbach’s alpha was 0.8703, so high enough to sum up the results of all five statements, and then divide it by five to make it into one variable that represents subjective well-being.

Job satisfaction

Job security

To control on whether the participant has a strong chance to find another job quickly, Q89h was used. This question was then recoded so the variable ‘job security’ represented ‘strongly disagree’ with 1 and ‘strongly agree’ with 5.

Income

Lastly, to check whether participants are able to make ends meet with their current household income, Q100 was used. This question was recoded so that 1 represents ‘with great difficulty’ and 6 represents ‘very easily’.

The variables ‘household number’, ‘age’ and ‘education’ were only renamed accordingly.

Selecting the Research Population and Variables

To select the research population, the option ‘drop if’ was used so that only the following countries remained in the dataset: 1 (Austria), 2 (Belgium), 7 (Denmark), 9 (Finland), 10 (France), 11 (Germany), 12 (Greece), 14 (Ireland), 15 (Italy), 18 (Luxembourg), 20 (The Netherlands), 22 (Portugal), 26 (Spain), 27 (Sweden), 28 (United Kingdom). In addition, only workers from the age of 25 until 65 remain. When this was done, all irrelevant variables to this research were deleted from the database.

Deleting Missing Variables

Appendix C: Descriptive Statistics Control Variables TABLE C1

Summary of Control Variables

Variable Mean Std. Dev. Min Max

Gender 0.501 0.500 0 1 Age 44.361 10.311 25 65 Household number 2.759 1.294 1 10 Working hours 37.071 11.411 1 120 Competence 2.137 0.627 1 3 Education level 4.997 1.779 1 9 Work pressure 1.856 1.089 1 5 Relatedness 3.634 1.546 0 5 Health 4.040 0.745 1 5 Job security 1.746 0.723 1 3 Income 4.032 1.262 1 6 Partner 0.686 0.464 0 1 Children 0.310 0.462 0 1 More hours 0.122 0.327 0 1 Less hours 0.324 0.468 0 1 N = 15458 TABLE C2

Correlation of Control Variables

Gender Age Household

Relatedness -0.016 -0.089* 0.0079 0.011 -0.024* 0.102* -0.051* 1.000 Health -0.014 -0.215* 0.0494* 0.028* 0.029* 0.149* -0.010 0.112* 1.000 Job security -0.005 -0.1237 -0.011 -0.013 -0.024* 0.042* 0.004 -0.024* 0.049* Income -0.041* 0.0667 -0.057* 0.090* -0.028* 0.283* 0.095* 0.155* 0.168* Partner -0.071* 0.0448 0.513* 0.024* -0.025* 0.017* 0.025* 0.015 0.022* Children 0.015 -0.3448 0.570* -0.011 -0.019* 0.085* 0.047* 0.025* 0.098* More hours 0.089* -0.0556 -0.000 -0.399* 0.029* -0.092* -0.049* -0.091* -0.042* Less hours -0.048* 0.0180 0.013 0.381* 0.004 0.136* 0.196* -0.009 -0.034* Job security

Income Partner Children More hours Less

Appendix D: Mediation Analysis Job Satisfaction Measure 2 TABLE D1

Regression Analysis on Job Satisfaction (2) Mediated by Autonomy1

Dependent variable JS 2 JS 2 JS 2

Model 1 Model 3 Model 4

Step and variables B SE B SE B SE Intercept 4.155* 0.028 3.836* 0.052 3.063* 0.139 Main effects Employment Type Permanent Temporary -0.225* -0.336* 0.031 0.045 -0.145** -0.226* 0.038 0.040 -0.295* -0.322* 0.033 0.033 Autonomy 0.368* 0.040 0.323* 0.028 Control Gender 0.087** 0.022 Household nr 0.017** 0.004 Age 0.002** 0.001 Working hours 0.003 0.002 Competence -0.007 0.011 Education level 0.010 0.014 Relatedness 0.077* 0.007 Income 0.091* 0.015 More hours 0.013 0.026 Less hours -0.188* 0.029 R Square 0.057 0.077 0.119 St. Error or Estimate 0.880 0.870 0.850 N = 15458 *p < 0.001 **p < 0.05

TABLE D2

Direct, Indirect and Total Effects of Employment Type on Job Satisfaction (2) Mediated by

Autonomy1

Direct effects Indirect effects Total effects

B S.E. Z B S.E. Z B S.E. Z

Job satisfaction 2 Autonomy 0.323 0.027 11.58* 0 0.323 0.028 11.58* Permanent contract -0.295 0.033 -8.85* -0.066 0.014 -4.86* -0.361 0.032 -11.21* Temporary contract -0.322 0.033 -9.83* -0.085 0.016 -5.46* -0.407 0.039 -10.36* N = 15458 *p < 0.05

Appendix E: Mediation Analysis Work-Life Balance Measure 2 TABLE E1

Regression Analysis on Subjective Well-Being Mediated by Work-Life Balance (2)1

Dependent variable WLB 2 SWB SWB

Model 2 Model 3 Model 4

Step and variables B SE B SE B SE Intercept 3.362* 0.121 3.205* 0.081 1.329* 0.168 Main effects Employment Type Permanent Temporary 0.072** 0.070** 0.031 0.028 -0.164* -0.186** 0.031 0.057 -0.138* -0.102** 0.024 0.036 WLB 2 0.406* 0.025 0.315* 0.028 Control Gender -0.181* 0.013 -0.062** 0.016 Household nr -0.016 0.009 0.019 0.010 Age 0.007* 0.001 0.003** 0.001 Working hours -0.013* 0.001 0.004** 0.001 Education level -0.017** 0.008 -0.009 0.007 Work pressure -0.264* 0.015 -0.003 0.013 Health 0.241* 0.014 0.423* 0.017 Job security 0.023** 0.009 -0.045** 0.012 Income 0.085* 0.011 0.075* 0.012 Partner -0.081* 0.016 0.027 0.019 Children -0.082** 0.024 0.019 0.017 More hours -0.014 0.029 -0.050 0.024 Less hours -0.260* 0.016 -0.136* 0.021 R Square 0.315 0.157 0.274 St. Error or Estimate 0.745 0.912 0.847 N = 15458 *p < 0.001 **p < 0.05

TABLE E2

Direct, Indirect and Total Effects of Employment Type on Subjective Well-Being Mediated by Work-Life Balance (2)1

Direct effects Indirect effects Total effects

B S.E. Z B S.E. Z B S.E. Z

Subjective well-being Work-life balance (2) 0.315 0.028 11.10* 0 0.315 0.028 11.10* Permanent contract -0.138 0.024 -5.82* 0.023 0.009 2.58* -0.115 0.028 -4.11* Temporary contract -0.102 0.036 -2.85* 0.022 0.009 2.39* -0.080 0.035 -2.27* N = 15458 *p < 0.05