What Factors Make Man and Woman Workers Happy?
Master thesis, M.Sc. in Human Resources Management Rijkuniversiteit Groningen, Faculty of Economics and Business
July 14, 2011 Martin William Student number: 2072963 Tuinbouwstraat 26A 9717 JJ Groningen tel.: +31 (0)65-‐468782 e-‐mail: mail.martinwilliam@gmail.com Supervisor:
Dr. P. H. van der Meer and
TABLE OF CONTENTS
TABLE OF CONTENTS
ABSTRACT 4
I. INTRODUCTION 4
II. LITERATURE REVIEW 7
Happiness and Satisfaction 7
Work-‐hours, Income, and Job Characteristics 8
Gender-‐based Differences 11
Dynamic of Gender Differences, Work-‐hours, Income, and Job characteristic 12
The Model 15
III. METHODOLOGY 17
Data and Sample 17
Measuring Dependent Variables 17
Measuring Independent Variables 18
Regression Analysis 21
Comparing Regression Analysis 22
IV. RESULT 23
Descriptives and Correlations 23
Regression Analysis 26
Differences Between Men and Women 28
V. DISCUSSION AND CONCLUSION 31
Discussion 31
Conclusion 35
APPENDIX 37
LIST OF FIGURES
LIST OF FIGURES
FIGURE 1. Happiness Chart 7
FIGURE 2. Relationship Model 15
LIST OF TABLE
LIST OF TABLE
TABLE 1. Descriptive Statistics and Independent Sample T-‐test 23
TABLE 2. Variables Correlations in Men Group 37
TABLE 3. Variables Correlations in Women Group 38
TABLE 4. Regression Coef^icients in Men Group DV: Job Satisfaction 39
TABLE 5. Regression Coef^icients in Women Group DV: Job Satisfaction 40
TABLE 6. Regression Coef^icients in Men Group DV: Life Happyfaction 41
TABLE 7. Regression Coef^icients in Women Group DV: Life Happyfaction 42
TABLE 8. Rank of Regression Coef^icient & t-‐test analysis 28
ABSTRACT
Happy workers are productive workers. The interest on well-‐being of workers has been the center of this thesis. Workers’ subjective well-‐being, which includes job satisfaction and life happiness, are believed to lead to many beneLits. Men and women may have different factors that inLluence their level of job satisfaction, and may have been inLluenced by different job
characteristics. According to the economic’s utility theory, work-‐hours and income affect workers’ level of satisfaction as well as other job characteristics. This thesis found that more factors inLluenced men’s job satisfaction than women’s, although there are hardly any different between man and women on effect size of those factors Only number of work-‐hours, preference to have more work-‐hours, and opportunity to learn new skill that were found have different effect size for men’s and women’s job satisfaction. The thesis found: 1) Women have higher job satisfaction than men, nevertheless both of them have the same happiness in life as a whole; 2) having a good interpersonal relations and good careers are the most important job
characteristics for men’s and women’s satisfaction; 3) there are hardly any differences on effect size of job-‐dimensions to men and women’s job satisfaction; 4) job is found only contribute a small part of life happiness as a whole. Further discussion provided.
Keywords: job satisfaction, gender differences, income, work-‐hours, job characteristics
I. INTRODUCTION
Happy workers are productive workers. A growing body of literature shares this notion. Happiness has been associated with more successes in professional life (Boehm &
Lyubomirsky, 2008), more and better ideas (Iverson, Olekalns, Erwin, 1998; Staw, Sutton, & Pelled, 1994), creativity (Baas, De Dreu, & Nijstad, 2008); positive interpersonal behaviors (Forgas, 2002); less prone to stress (Kubzansky, Sparrow, Vokonas, & Kawachi, 2001); better health (Watson, 1988; Kobasa, 1979; Blanch^lower, and Oswald, 2007) and better
performance (Cropanzano & Wright, 1999). In a broader context, happiness of the people is suggested to be a utility for a ^lourishing society (Veenhoven, 1998).
discussed euidamonia, or a state of living a good life. Eudamonia has since then commonly de^ined as happiness, although it real meaning is ^lourished life. Happiness were stemmed from 1950 ‘s psychologists who were more concern about pathology, stress, depression, and other negative emotions. The topic of happiness was started to be discussed mainly as a positive psychology movement. Nowadays, happiness has become the focus of a novel approach to study human. Happiness has also become the attention of study by economists who emphasize more on quantitative measurement in investigating it.
The bene^its of being happy has led happiness to be the primary foundation in designing policies in wider and narrower contexts such as between nations and within an organization. Bentham’s fundamental axiom of ‘greatest happiness for the greatest number’ served on the argument. Happiness is suggested that it should be viewed as a utility and as a goal for every policy design (as cited in Layard, 2011). The inextricable value of happiness has led
researchers to stumble on its determinants, including the determinants of happiness in the organizational setting. This kind of study will help the policy-‐makers and decision-‐makers in companies to build sustainable happiness for their employees, and to take the ripe fruit of it. This study is also aimed at delivering the same purposes.
Research on happiness should be carried on continuously and longitudinally, although many studies on what makes workers happy have been done by psychologists, sociologist, and economist, Happiness can be a proper evaluation criterium of social policies according to social dynamics. One social policy at one time may not be appropriate for another time. Therefore, research about ever-‐changing determinants of happiness should be repeated over time.
Schouteten (2001), job characteristics were, indeed, found to be the most salient aspect that affected the quality of working life, opinions about the work and its effects on the workers. The other aspect are workers’ characteristics and the ^it between the work and the workers’ characteristics. Therefore, it is bene^icial and relevant to study about which job
characteristics will affect happiness of the workers the most.
However, as van der Meer & Wielers (2011) have suggested, there are still rooms for examining gender differences in happiness and satisfaction at work. Particularly, when gender equality has become the main concern of todays world’s economy. Women’s achievements have become increasingly crucial for the national economy (Hyde & Kling, 2001). It implies that interests on their satisfaction has also become increasingly essential. Companies are trying to engage their employees. Because women and men have real
differences so that the efforts to improve the happiness of each gender should also be tailored according to their different needs.
Therefore to achieve the objective, we would like to answer the questions below: 1. Do men and women have a different level of job satisfaction?
2. What will affect men’s and women’s job satisfaction?
3. What will affect men’s and women’s job satisfaction differently?
This paper will give a contribution for policy-‐makers and decision-‐makers on how to
empower man and woman workers, and by then increasing their productivity by improving their happiness. We also hope this study can continuously monitor the happiness, and what should be done about it over time. By then, companies should focus on designing or
implementing policies that will affect most man and woman workers. Equality of work satisfaction may lead to more ^lourishing society. Applying what Bentham has said about the greatest happiness for the greatest number. It is an economic view that seen happiness and satisfaction as a utility. The research will also included psychological measures as a little research about job satisfaction has combined it with economic perspective.
II. LITERATURE REVIEW
We will discuss a body of literature related to happiness and satisfaction, and how gender differences can have an impact on happiness . We believe that differences in gender can be used as an explanation on differences between men’s and women’s responses to job
characteristics they experienced that in turn may be in^luencing their levels of happiness and satisfaction.
Happiness and Satisfaction
Happiness is a broad and a vague idea, which not yet can be de^ined narrowly. This is including the problems in measuring and analyzing it. Therefore, it is necessary to use a conceptual framework in studying happiness, which sometimes interchangeably used with satisfaction. Wright and Cropanzano (1997) has raised the issue that happiness and
satisfaction are two different constructs. The former one is more related to an affective state, while the latter represents an attitude, which can be more positive or negative towards on an attitude object (i.e. work, life, someone, etc).
FIGURE 1 Happiness Chart
HAPPINESS
Level One Level Two Level Three
momentary feelings judgement about feelings quality of life
joy, pleasure well-being, satisfaction flourishing, fulfiling
potential
more immediate more sensual and emotional
more reliably measured more absolute
more cognitive more moral and political
involving more cultural norm and values more relative
Three different senses of the term ʻhappinessʼ. Each level includes the content of the level below, plus some additional things (Nettle, 2005).
Kahneman, Walker and Sarin 1997; Kahneman 1999; Ryan and Deci 2001; as cited in Netttle, 2005). He suggests that happiness consists of three levels of de^inition. The ^irst level is temporary pleasure, or joy of having what we want or what we need. The second level, which also includes the ^irst one, is people’s evaluation on the feeling or emotion that they
experience at the ^irst level and how they feel satis^ied with it. This de^inition of happiness implies that people evaluate their ups and downs and re^lect on that experiences in overall measures (e.g. generally have more positive feelings, or sad most of the time). The ^inal de^inition, or the broadest de^inition is related to the quality of life as a whole. Maslow’s hierarchy of needs named it self actualization or the peak experience. To frame the research and construct meaning of it, we will use the second level de^inition as a framework. In this paper, we suggest that both de^initions can serve the same purpose in a sense that both constructs can be used to represent individual’s subjective well-‐being. We will use literature about life happiness and life satisfaction. because there is no clear cut between life happiness and life satisfaction, we then will use them interchangeably.
Life happiness or life satisfaction is considered to be strong and signi^icantly correlated to happiness at work or job satisfaction (Judge & Watanabe, 1993). It is not surprising that both of them are interrelated because most of workers’ lifetime is spent in work-‐related activities. Often, in this time of ef^iciency, employers would want to use them optimally. They sometimes ask workers to work in extended formal working-‐hours. We can deduce that this situation may strengthen the correlation of job-‐life satisfaction. There is also might be a reciprocal spillover effect between life. For example, work con^licts may affect the quality of
relationships with workers’ partners (Bakker, et al., 2009), or positive experiences in the family would be a buffer for psychological distress at work (Barnett, 1994). Therefore, general life happiness or satisfaction could be an important measurement that encompasses the satisfaction at work.
Work-‐hours, Income, and Job Characteristics
According to van der Meer & Wielers (2011) and Clarks (1998), there are six most popular aspects that are viewed by the worker themselves to be the determinant factors that are able to make them satis^ied at work. Those are the level of pay, hours of work, and job
aspects. People with less stress are happier than people with a higher level of stress. Higher motivation to work may indicate higher happiness as well.
On the other hand, economist contribution in happiness is explained by utility of income and work-‐hours. They are considered to be in^luential aspects for people’s job satisfaction. People will calculate their income with hours of work to be happy. The ^irst will be utility, and the second is considered to be the disutility of happiness. People may receive income that will make them motivated to put a certain amount of effort, and balances of it is the happiness itself.
Longer work-‐hours is associated with a higher level job stress in German medical staff
followed by growing happiness. It is suggested by later researches (Blanch^lower & Oswald, 2004; Clark et al., 2008) that the relative income is the one that may affect happiness more. It means people are happy and satis^ied because of their income rank higher than the others. By means of adaptation, habituation and increasing aspiration levels (B. S. Frey & Stutzer, 2002), the paradox is explained in such a way that absolute income affects happiness but with
diminishing marginal utility.
Future prospect, including job security, career advancement, and development possibilities, also affect people’s satisfaction (van der Meer & Wielers, 2011). The presence of those characteristics would bring positive effect to happiness. In Lazear & Gibbs (2009), worker’s satisfaction is also affected by implicit contract and ^irm-‐human speci^ic capital. Employee will be expected to have endured working relationship, as both of the employer and employee had invested on each other. In contrast, employee will be less satis^ied with hard contract, where job security is sparse. This longtime relationship represent job security, which fueled
employee self esteem, and in turn ful^ill human needs to be in certainty. This could increase people’s satisfaction on work. Another important aspect from future prospect is the chance to self development, including learning, training, and acquiring new skills. These will be
developed (soft and hard) skills and abilities that will affect the employee’s self esteem, security, and therefore, satisfaction. Another possibility is the development will increase the worker’s value, therefore may affect their income. The same explanation may be applied for the third features of job’s future prospect, career advancement.
Interpersonal relationship is seen by many psychologists as a source of happiness. Pryce-‐ Jones (2000), wrote in her book that happiness is not something people can do on their own, and they always need others. It is one of Abraham Maslow’s famous Hierarchy of Need, that humankind always need to belong to other people. Ful^illment of the needs may bring that person to higher needs, which in the end lead to happiness. In more speci^ic terms,
interpersonal relationship may also represent sources of help in the demanding job. It may reduce stress level, or becomes emotion buffer, resulting in more happy individuals.
Dif^iculty of the work including exhaustion, working environment, stress induced, exhaustion, physically unpleasant also found to be related with happiness and satisfaction. Working in such situations that are noisy, dirty, stressful, or time pressured may affect job satisfaction negatively. Karasek (1979) and Karasek & Theorell (1990) suggests that unbalanced work-‐ input and mandated work-‐output make the job become stressful. Therefore, heavy workload may not be stressful if it is accompanied by more latitude of job decision.
Job autonomy, People also tend to be more satisfy whenever they can contribute signi^icantly, and have an independence to manage their job. This is considered to be part of Job
Characteristic Model by Hackman & Oldham (1980). It implies that people are motivated (are happy) when they have autonomy and control over their job. Because autonomy are seen to have effect, respectively, in giving a sense of control, reducing uncertainties, reducing work stress, and therefore, increase satisfaction.
Gender-‐based Differences
Gender and Sex. There are differences between term ‘sex’ and ‘gender.’ Firstly discussed in the article of sexologist and psychologist, John Money (1955). Sex referring to the differences in biological differences between men and women, whereas gender, now become more popular term, is derived from the relationship between sex and behavior (Udry, 1994). Implicitly, gender differences could not be separated from sex differences. Differences in biological features (i.e. chromosomes, hormones) are found to be affected how men and women differ. While it is accepted that men have Y chromosome, while women does not, research has shown that it affects indirectly. They affect the development of gonadal hormones, which in turn affect the sex-‐differentiation of individuals. This includes
development of different body structure and different reproductive behavior, which is a part of gender-‐differences discussion (Udry, 1994; see also Kenrick, Trost, and Sundie, 2004). In sum, although many social theory often does not include this evolutionary perspectives, its contributions are still useful to support the social view of sex differences (i.e. gender). This perspective may give a contribution in explaining why men and women naturally behave and respond to the same situation differently. It was still debatable whether differences between gender is natural or nurtural.
Bandura clearly propose that children will successfully imitate when they have incentives. Developmental psychologist observed that boy and girl are rewarded and punished for a different reason and also girls are found to receive less recognition (Fels, 2004). It shapes the behavior and building normative measure on what women and men should behave.
Masculinity and femininity are shaped with their own distinguished trait. Masculinity is related to trait such as self reliant, willing to take risk, self-‐suf^icient, individualistic, competitive, ambitious, while femininity are largely related to providing resources (to children, lover, husband, sick parent, or even boss). People expect men and women to show different behavior. Women who speak as much as men or compete for high-‐visibility position are considered to be wrong (Fels, 2004). This expectation will based gender paradox
phenomenon, which is explaining that women are more satisfy with their job than men, although generally they have worse job conditions than men. Phelan (1994) suggested perceived (subjective) features of the work have a greater effect than the ‘real’ objective features (see also Mueller & Wallace, 1996). So even though women tend to have less good job, they may perceive it differently. The difference in perception may cause a difference in happiness and satisfaction at work. Perception could be affected by motivation and
expectation. Those relationship will be the hypotheses, and explained in the third part of this chapter.
Dynamic of Gender Differences, Work-‐hours, Income, and Job characteristic
Gender and work-‐hours. Number of work-‐hours is important variables in economic view as it is represented utility of happiness. Working women have more dual role con^lict than men. Women are expected to focus more in household matters while also have obligation for their job. This duality may reduce leisure time and bring higher pressure for them, producing higher level of stress. Women will appreciate less work-‐hours, as it means they can allocate more time for their spouse, children, or another household matters. Longer work-‐hours for women also found to make them more vulnerable than to indulge in unhealthy behavior, as there are more pressure experienced (O’Connor & Conner, 2011). In the other hand, men are traditionally expected to work and get some money for the household. Income are inline with the number of work-‐hours, as it is calculated from the number of hours an individual has worked. Therefore less work-‐hours will tend to produce less income, and therefore less satisfying what is needed for men according to their role.
Hypothesis 1c: women’s job satisfaction will be more affected by number of hours
compared to that of man’s
Gender and income. Different role expectations may also play a role in how men and women response to their income. We hypothesize that income may not be the priority for women compare to that of to men. This is because most society may still perceive men as the
breadwinner of the family. Wood^ield (2001) found that half of the female participants in her study perceived that their salary was not as important as other job characteristics. The
^indings also showed that male participants were less able to put off rewards and prioritize income.
She found that women are in a better position to accept their income level than men in the same roles. This is because women have more altruistic impulse than men. Women perceived themselves as not having responsibilities associated with family wage. In western countries, many of women have been socialized to be a good wife and good mother, instead of being ^inancially independence (Gilbert, 1993 as cited in Cinamon & Rich, 2002). However, measuring income is more than only the absolute amount. Regarding to Easterlin Paradox, men who tend to rate themselves as getting high income (subjective evaluation) may have more happiness affected by the relative income. Considering the population of this study, we propose that subjective income has played a more important role in their happiness. Relative income is a more signi^icant factor to happiness in developed countries, where an increase in absolute income brings less effect because it has a high standard of income. People in well developed countries tend to satisfy their basic needs, therefore there is a less effect on absolute income.
Hypothesis 2a: men’s job satisfaction will be affected by their subjective income level. Hypothesis 2b: women’s job satisfaction will be affected by their subjective income level Hypothesis 2c: women’s job satisfaction will be less affected by their subjective income
level compares to that of men’s.
Gender and future prospect. Career may increase satisfaction, and therefore happiness, because career may give a sense of self-‐achievement and recognition. Fels (2004) found that women faced more barriers to be ambitious since their childhood. Therefore, women tend to avoid being recognized for their achievement. Higher positions also mean bigger
willing to spend longer hours with their family, and therefore more ^lexible work hour is desirable. The expectation to be traditional household caretaker also helps them to buffer the effect of future prospects on their happiness. In Wood^ield (2001), women were also viewed as more ‘not concern’ to career compared to that of how men were viewed.
Hypothesis 3a: men’s job satisfaction will be affected by a job’s future prospect. Hypothesis 3b: women’s job satisfaction will be affected by a job’s future prospect. Hypothesis 3c: women’s job satisfaction will be less affected by a job’s future prospect
compare to that of men’s.
Gender and interpersonal relationships. Women ^ind more happiness in social interactions. Wood^ield (2001) suggests that women participants tended to give more efforts in social interactions than did men. Another explanation may come from women’s preference to tend-‐ and-‐befriend in facing unpleasant social interactions, contrasting with men that more ^ight-‐ or-‐^light behaviors. Much of the literature on sex roles and sex-‐role stereotypes leads one to expect signi^icant differences, with women being more interpersonally oriented than men (Bardwick, 1971; Oetzel, 1966; Hoffman, 1972; Sherman, 1971; Chafetz, 1974; Lynn, 1969 as cited in Balswick & Avertt, 1977). Therefore, we may suggest that relationships with others are more important for women than for men. It may includes relationships with supervisors and colleagues,
Hypothesis 4a: men’s job satisfaction will be affected by interpersonal relations Hypothesis 4b: women’s job satisfaction will be affected by interpersonal relations Hypothesis 4c: women’s job satisfaction will be more affected by interpersonal relations
compare to that of men’s
Gender and the level of difBiculty of work. Men may have ability to stand in the physically unpleasant situations than women. In neurobiological research, women were found to have lasting stress and more related to emotions, while men had shorter time in being in stressful conditions and half the rate of depression (Wang, et al., 2007). This ^inding is near-‐global-‐ wide, across many nations, cultures, and ethnicities (Nolen-‐Hoeksema, 1990; Weissman et al., 1996). So, women could be considered as more prone to stressful situations rather than men. Therefore, we hypothesize that they may be more affected by the dif^iculty of the work.
Hypothesis 5b: women’s job satisfaction will be affected by difLiculty of work Hypothesis 5c: women’s job satisfaction will be more affected by difLiculty of work
compared to that of men’s
Gender and job autonomy. Women may has a lower expectation for their job as suggested by occupational segregation analysis. This lower expectation may make them to be happier than do men, who have higher expectation on their job. Men are also said to have more willingness for proud and control, therefore they will be more affected by the presence of prestigious job and autonomy to do their job (Wood^ield, 2001). According to Fels (2004) men’s masculinity is related to self reliant, and self-‐independent traits which may lead men to expect more autonomy in their job.
Hypothesis 6a: men’s job satisfaction will be affected by job autonomy Hypothesis 6b: women’s job satisfaction will be affected by job autonomy
Hypothesis 6c: women’s job satisfaction will be less affected by job autonomy compared to
that of men’s
The Model
As we discussed the literature ^indings, we formulated the hypothesis and the model for this research. We believe that job characteristic, income and hours of work will affect job
satisfaction, and therefore affect life satisfaction as a whole. Different response on job satisfaction could be different for men and women, as they have certain characteristics that may be natural or nurtural.
From the model, utility of income and hours of work and job characteristic may mediated by gender differences, especially by how gender expect themselves in job market, and how they perceive job conditions they face. We propose that job satisfaction are mediating the effect of job characteristic, income, and work hours to life happiness. We also would like to see
III. METHODOLOGY
Data and Sample
We used survey data from LISS (Longitudinal Internet Studies for Social Sciences) panel data. LISS panel is a Dutch online survey conducted by CentERdata, an Institute for data collection and research supported by the Netherlands Organization for Scienti^ic Research (NWO). It comprises a large sample of the Dutch population. The data are available to be downloaded after registration in www.lissdata.nl. It consists of several core modules and assembled modules. We focused on the data related with happiness, satisfaction and job characteristics. These data were saved in several ^iles, according to the relevant modules. The modules used were: work and schooling, personality, income, value, politic and value, health, and
background variables dataset. This research used wave 1 that was conducted in April -‐ September 2008. We need to integrate all modules into one data set, using unique numbers for each respondent. The ^irst number of respondent after data set integration is 12,908 respondents. All answer that were analyzed comes from people that are working and paid for their employment. There are 5,124 respondents who are in paid employment. The respondent chosen also came from the most common working age group in the Netherlands. They were aged between 25 and 60 years-‐old at the time of ^illing in the questionnaire, reducing the number of respondents into 4,649. The respondents also had to work for at least 12 hours a week, which is an of^icial number of formal work. The ^inal number after considering all of these criteria are 3,214 respondents.
All important variables such as income and work-‐hours are made certain to be complete data. Therefore we replace all missing value in income with mean, and removing cases whose income is zero. We also ^ilter out cases with missing values in income standard and/or cases with not logical answer (look at income part in this section), therefore we got the ^inal ^igures of respondent is 2,160 respondent. We found a relatively balanced proportion of men and women from this number. There are 1,137 men, and 1,023 women.
Measuring Dependent Variables
(1993; as cited in Frey & Stutzer, 2002), the score is based on cognitive assessment about overall quality of their life or part of their life. Therefore we will use the items that directly ask the participant about how happy they are as a whole. LISS panel data has two measure of related to life as a whole: “in whole, how happy would you say you?,” (cp08a010) and “how satis^ied are you with the life you lead at the moment?” (cp08a011). Their answer will have 10-‐range answer from zero (not at all satis^ied) to ten (completely satis^ied). For life as a whole, satisfaction and happiness will be considered as a measure of subjective well-‐being and, indeed, are highly correlated (r= .836, p<.01). Therefore we combine them as a new variable, life happyfaction. We did it by ^inding mean between those two. The measure of reliability coef^icient supports this treatment, by showing very high result of α=.911.
Job satisfaction. In measuring job satisfaction, we also use self rated items, which point out the respondent’s satisfaction on their job. LISS panel data containing separate measure of different aspects of work. Therefore measuring the component of job satisfaction will lead us to the measurement of job satisfaction itself. On this measurement, respondent are asked how they satisfy with their job and aspects of it, such as on: type of work, atmosphere among colleagues, career, current work, and wages received (cw08a128, cw08a130-‐133). Average evaluation of respondent toward the aspects of their job will be considered as measurement of overall job satisfaction. The internal consistency of the scale is α=.814. The questionnaire also ask about satisfaction toward work-‐hours, but have been left out in this research as it has another measure of work-‐hours, and the item lowers the alpha if included.
Measuring Independent Variables
Income. Absolute income is measured by directly asked the their personal net income each month, there are several items in questionnaire that asked respondent to give information about their income, such as gross income, net income, imputed net income, and categorical income. Multiple questions about income is applied to enhanced the completeness of income data, as it was expected that many respondent will be uncomfortable revealing their personal income. This study will use imputed personal net income (nettoink_f), which contain more complete data as the missing values was replaced by calculation of gross income and/or by mid point of categorical income. Remaining missing values were replaced by the mean. Considering that happiness is to be concave in work-‐hours and income, natural logarithm of these two variables is used.
Another measure of income is related to subjective income, perceived position of their income compared to other’s income. The evaluation may include whether they perceived their income small, enough, or high compared to their needs. As people needs are in^luenced by their
comparison to others, measure of relative income based on aspiration may, indeed, showing one’s position in their group. To measure this kind of income, we use respondent’s evaluation on amount of net wages they ^ind very bad, bad, insuf^icient, suf^icient, good, and very good for their household situation (ci08a230-‐ci08a235). These six measures will be the cutting point to transform the absolute income into 7 categorical responses, which show the position of their absolute income compared to subjective standard they set. The higher the answer the higher the income position they perceived, and therefore the more able they ful^ill their
Interpersonal relations. The items will measured in whether respondent could get help from colleagues or not. This is involved help from supervisor or friends in the same level or
position. Getting help may indicate good relationship between respondent and other people in their work environment. The items used are “I get sufficient support in difficult
situations” (cw08a431) and “I get the appreciation I deserve for my work” (cw08432). The internal reliability of the scale allowing them to be combined as an average of interpersonal relation (α=.706).
DifBiculty of the job. On the work and schooling questionnaire there are several questions about the job dif^iculty. The question will measure the perceptive evaluation of respondent on how their work condition are, compared to what they expected. Because of the subjective nature of questionnaire, their answer may representing how their work condition met their expectations. From factor analysis, these questions are consists of three components, namely dangerous environment of the job, physically demands, and mental effort requirement. The ^irst component consists of three items, asking their perception whether their job involved getting dirty (cw08a413), is dangerous (cw08a414), and whether it involved hazardous substances (cw08a415). We found high internal consistency of α=.760. Therefore they could be combined as one measure of job dif^iculty related to dangerous environment. The second component is related to physical demand. This component have four item questions whether the job need high physical endurance or not. The ^irst item is “is your work physically
demanding?” (cw08a416). Followed by “do you need to lift heavy objects?” (cw08a417), “do you need to lift heavy objects?” (cw08a418), and “do you need to kneel or stoop?” (cw08a419). This job dif^iculty related to physical demand shows high reliability (α=.848). There are two items related to third component that concerns about mental process needed in the job. “Does your work require mental effort?” (cw08a420) and “do you need to work with a lot of
concentration?” (cw08a421) are then combined into job dif^iculty related to mental
requirement. The scale showing α=.707. There are consistent low number of missing values in each item, nevertheless they have been replaced by mean.
Job Autonomy. There are two items could be related to job autonomy. With one concerns on whether the respondent can set their own pace of the job (“Can you work at your own pace”;
cw08a412), and the other concerns on control they have in doing their job (“There is very little
much freedom given so they could manage their resources to complete the job. These two variables show relatives weak internal reliability to be combined (α=.502).
Control Variables. Besides all the main independent variable, we will also use variables that could be considered related to happiness to be controlled. They are: age (leeftijd), type of industries (cw08a402), health (ch08a004), marital status (woonvorm), number of children (aantalki), education level (oplcat), work ethic, and personality (Frey & Stutzer, 2002). Items asking directly about those variables are provided in the questionnaire. We make 15 dummies of type of industries to see effect of each industry. From the data about domestic status, we make three dummy variables: married-‐cohabitation, single, and children presence. This dummy variables could show the effect of marital status (being or not being with partner or spouse) and how presence of children affect job satisfaction. For measuring work ethic, there are four items: “you can only do what you feel like doing after you have done your
duty” (cv08a139), “if someone wants to enjoy life, he/she must be prepared to work hard for it” (cv08a140), “I feel happiest after working hard” (cv08a141), and “work should always come ^irst, even if it means having less leisure time” (cv08a142). To get measurement of personality, we used the big ^ive personality trait. There are 50 items (cp08a020-‐cp08a069) of the personality test, in which every trait are represented by 10 items. We combine those ten items to be one variables related to one trait. Therefore we got ^ive variables showing personality of the respondent. The reliability of all trait are very high (α = .774-‐.873). There are a few missing values in the control variables, and have been replaced by mean.
Regression Analysis
Comparing Regression CoefXicient
Using B values or regression coef^icient and its SE, we could do t-‐test analysis to check
whether the difference of the value is found signi^icant or not. We formulate the t with below formula:
IV. RESULT
Descriptive and Correlations
This section describes information about general statistic properties such as mean, standard deviation and correlation between the dependent variables and independent variables. The results are as it is shown in table 1, table 2 and table 3.
TABLE 1
Descriptive Statistics and Independent Sample T-testTABLE 1 Descriptive Statistics and Independent Sample T-testTABLE 1 Descriptive Statistics and Independent Sample T-testTABLE 1 Descriptive Statistics and Independent Sample T-testTABLE 1 Descriptive Statistics and Independent Sample T-testTABLE 1 Descriptive Statistics and Independent Sample T-testTABLE 1 Descriptive Statistics and Independent Sample T-testTABLE 1 Descriptive Statistics and Independent Sample T-testTABLE 1 Descriptive Statistics and Independent Sample T-testTABLE 1 Descriptive Statistics and Independent Sample T-test
Variable
Men
Men WomenWomen t-testt-testt-testt-test Variable
Mean SD Mean SD tt df sig
1 Happyfaction 7.65 1.09 7.72 1.15 -1.26 2158 0.208
2 Job satisfaction 7.28 1.21 7.41 1.21 -2.51 * 2158 0.012
3 (ln) work-hours 3.71 0.19 3.32 0.34 32.13 *** 1571 0.000
4 Prefer more work-hours 0.09 0.28 0.21 0.41 -7.90 *** 1802 0.000
5 Prefer less work-hours 0.61 0.49 0.44 0.50 7.72 *** 2126 0.000
6 Right work-hours 0.31 0.46 0.35 0.48 -2.17 * 2116 0.030
7 (ln) income 7.59 0.35 7.16 0.46 24.53 *** 1918 0.000
8 Missing income 0.04 0.19 0.04 0.19 0.20 2158 0.839
9 Subjective income 3.79 1.60 2.58 1.74 16.74 *** 2082 0.000
10 Opportunity to learn new skills 2.92 0.62 2.94 0.62 -0.77 2158 0.441
11 Having good careers 2.43 0.78 2.40 0.74 0.89 2150 0.373
12 Interpersonal relations 2.78 0.57 2.85 0.56 -2.90 ** 2158 0.004
13 Job difficulty - physical 2.29 0.55 2.26 0.57 1.13 2158 0.257
14 Job difficulty - mental 1.33 0.45 1.42 0.50 -4.52 *** 2065 0.000
15 Job difficulty - dangerous environment 2.58 0.51 2.73 0.37 -8.17 *** 2068 0.000
16 Work on own pace 2.63 0.56 2.52 0.62 4.10 *** 2071 0.000
17 Freedom to do work 3.00 0.67 3.01 0.66 -0.41 2158 0.680
Table 1 shows higher scores of job satisfaction than men (m=7.41, sd=1.21 and m=7.28,
sd=1.21). There are signi^icant differences between them with t(2158)=-‐2.51, p<.05.
Nevertheless the signi^icant difference is gone in life happyfaction scores. Although women scored higher than men (m=7.72, sd=1.15 and m=7.65, sd=1.09), equal scores of happyfaction can be assumed (t(2158)=-‐1.26, p>.05).
Lower scores of work-‐hours for women is found. Men scored m=3.71, sd=.19, while women only scored m=3.32, sd=.34. Lower score for women founded in income as well (m=7.59, sd=.35 and m=7.16, sd=.46). Women’s subjective income position is under men’s. They only score
m=2.58, sd=1.74, while men’s mean score is m=3.79, sd-‐1.60. T-‐test analysis results shows that
these inferiority of women in number of work-‐hours, (absolute) income, and subjective income are signi^icant (respectively t(1571)=32.13, p<.001; t(1918)=24.53, p<.001; and t(2082)
=16.74, p<.001).
Table 1 also give mean score for opportunity to learn new skill and good careers. There only slight different between men and women, and could be ignore based on t-‐test analysis result (t(2158)=-‐.77, p>.05; and t(2150)=.89, p>.05). Instead of being under men’s score again, women (m=2.85, sd=.56) have higher mean of interpersonal relationship than men (m=2.78,
sd=.57). Signi^icance at the 0.01 level is found for this differences (t(2158)=-‐2.90, p<.01).
Another signi^icant differences are found in job dif^iculty related to mental requirement and dangerous environment. Women feels higher mental requirement are needed to do their job compared to men (m=1.42, sd=.50 and m=1.33, sd=.45; with t(2065)=-‐4.52, p<0.001), as well as more presences of danger in their work environment (m=2.73, sd=.37 and m=2.58, sd=.51; with
t(2068)=-‐8.17, p <.001). Men have higher score of chance to set own work-‐pace compared to
women (m=2.63, sd=.56 and m=2.52, sd=.62), with signi^icance founded (t(2071)=4.10, p<.001). In table 2 and 3 correlations of variables are presented. As expected life happyfaction are correlated positively and signi^icantly with job satisfaction for both women and men. (r=.362,
p<.01 and r=.268, p<.01). All correlations between independent variables and life satisfaction
Insert table 2 about here Insert table 3 about here
Among independent variables, only preference to have more work-‐hours that found not to be signi^icantly correlated with men’s job satisfaction. In women group, there are more variables that are not correlated signi^icantly, they are: amount of work hours, willingness to have less work hours, job dif^iculty related to dangerous environment, and to mental demand (look at table 2 and 3).
Logarithm of work-‐hours found to have positive correlation with men’s job satisfaction (r=. 111, p<.01). In contrast, women’s job satisfaction are not signi^icantly correlated with amount
of work-‐hours (r=.008, p>.05). For subjective work hours aspiration, men respondent who prefer less work-‐hours and prefer more work-‐hours show negative correlation value (r=-‐.095,
p<.01; and r=-‐.057, p<.01, respectively). Having the right work-‐hours correlate positively and
show higher correlation than the other two (r=.136, p<.01). In women group, prefer less work-‐hours is not signi^icantly correlated with job satisfaction and even no correlation at all (r=.000, p>.05). In contrast, having right work-‐hours shows positive signi^icant correlated with job satisfaction (r=.117, p<.01), while prefer more work-‐hours also found to be signi^icant but in negative direction for women’s job satisfaction (r=-‐.137, p<.01).
Amount of absolute income are having signi^icant positive correlation (r=.225, p<.01), as well
as subjective income (r=.185, p<.01) to men’s job satisfaction. Meanwhile, absolute income and subjective income also correlate positively with women’s job satisfaction (r=.109, p<.01 and r=.086, p<.05).
Future prospect of a job also have signi^icant correlation; opportunity to learn new skills is
positively correlated with men’s job satisfaction (r=.387, p<.01), as well as with their career prospect (r=.361, p<.01). In the same manner, opportunity to learn new skill and career prospect also signi^icantly correlated to women’s job satisfaction. In respective, the correlation are r=.297, p<.01 and r=.325, p<.01.
478, p<.01). In both groups, interpersonal relation are showing the strongest correlation to job
satisfaction.
For job dif^iculty related to physical demand, men show signi^icant positive correlation (r=.
175, p<.01), while women show lower signi^icant correlation (r=.065, p<.05). The same pattern
also showed in job dif^iculty related to mental demand. Men shows signi^icant negative correlation r=-‐.117, p<.01, while women’s mental demand related job dif^iculty have no signi^icant correlation with their happiness. Job dif^iculty related to dangerous work
environment are signi^icantly correlated with men’s job satisfaction (r=.078, p<.01), and it is not correlates to women’s job satisfaction.
The last job characteristic are job autonomy. Both of chance to set own work pace and freedom to do work is signi^icantly correlate to men’s job satisfaction, with positive
correlations (respectively, r=.220, p<.01 and r=.292, p<.01). In women group the correlation between job satisfaction and ability to set own work pace is .077, p<.05. Freedom to do work is signi^icantly correlated with women’s job satisfaction as strong as r=.161, p<.01.
Regression Analysis
In the fourth model, which encompasses all variables into account, the regression analysis shows that 41.4% and 32.8% variance on job satisfaction could be explained by the
independent variables, respectively in men and women group. The explanation power of independent variables are lower in the ^irst three model. The most signi^icant increase is in the model three, after we took all job characteristics into account.
In men group, there are more independent variables that statistically signi^icant affecting job satisfaction compared to women group. Table 4 shows B value for each variable in each model.
