Different Contexts,Different Effects?

Abstract

This paper takes a comparative approach to the topic of work time and health, asking whether weekly work hours matter for mental health. We hypothesize that these relationships differ within the United States and Germany, given the more regulated work time environments within Germany and the greater incentives to work long hours in the United States. We further hypothesize that German women will experience greatest penalties to long hours. We use data from the German Socioeconomic Panel and the National Longitudinal Survey of Youth to examine hours effects on mental health score at midlife. The results support our initial hypothesis. In Germany, longer work time is associated with worse mental health, while in the United States, as seen in previous research, the associations are more complex. Our results do not show greater mental health penalties for German women and suggest instead a selection effect into work hours operating by gender.

Keywords

comparative cross-national Germany mental health schedules work stress work time worker health

Working time has been of great concern for over a century. Labor organizers and social critics have sought to protect workers from extended hours of labor and the loss of weekends, vacations, and, among the religious, a holy day of rest. Recently, organizations from across Europe launched the European Sunday Alliance to “promote work-free Sundays and decent working hours throughout Europe.” Its concern is that too much work hurts people’s health and psychological well-being by creating added stress, limiting leisure and family time, and reducing social cohesion (see Organization for Economic Co-operation and Development [OECD] 2004). General health concerns regarding work in Europe also continue to grow (e.g., Leschke, Watt, and Finn 2012).

In the United States, despite a growing interest in the health effects of working time, there is arguably a strong culture supporting long work hours (Williams 2001) and numerous structural incentives to extend work time. Jobs that provide the most pay, flexibility, and psychological enrichment are often those that demand the most hours (see Jacobs and Gerson 2004). While research using limited or cross-sectional samples from Europe and the United States links long hours with greater risk of health problems (Sparks et al. 1997; Wirtz and Nachreiner 2010), recent American research suggests that work hours may be related to health in complex ways. Rather than longer work hours always leading to worse mental health, for example, the health-promoting effects of hours flexibility or socioeconomic status may be at least as important as working hours. In a recent study of U.S. workers, those who worked the most hours were not worse off in levels of distress and mental or physical health scores than those who worked a standard 40-hour week, and this pattern did not differ significantly by gender (Kleiner and Pavalko 2010). Another study of U.S. men and women used a natural experiment to show that true scheduling flexibility, unconstrained by typical organizational disapproval, resulted in better psychological and emotional outcomes than controls (Moen et al. 2011). Inequitably distributed benefits to working long hours may underlie the complex association between work hours and health noted in U.S. research.

In Europe, however, the normative and regulatory landscape differs from that of the United States by providing comprehensive limitations on work hours, guaranteed leave, and minimum vacation time. Furthermore, Germany has enacted even more stringent regulations on working time. Germany’s strong welfare state and lower inequality, in comparison to the United States, allow for less reliance on the labor market to provide basic needs, including health insurance, and creates fewer incentives for long work hours (Bell and Freeman 2001). Work hours in Germany (and Europe) are, as a result, lower than in the United States (Demetriades and Pedersini 2008; OECD 2010) for both men and women. Together, the set of legal and policy reforms in effect within Germany reflect and reinforce these German norms against long work hours. Germans are more likely than Americans to desire fewer work hours (Bell and Freeman 2001; Reynolds 2004; Stier and Lewin-Epstein 2003), even though they already work shorter hours than Americans. In contrast, Americans are more likely than Germans to prefer to increase their work hours, even though they already work longer hours than Germans.

Given the substantially different contexts for working time in Germany and the United States, we ask whether working time in the United States and Germany has different relationships with mental health. Building upon prior American research, we compare German and American longitudinal data to investigate this issue. We expect that the current norm toward lower work hours in Germany operates to affect mental health differently in two distinct ways. First, because of the stronger welfare state and greater labor market regulation that sustain men’s and women’s lower work hours in Germany, we do not expect to see health penalties for working short hours. Second, because of the strong and highly regulated norm against long work hours, we expect that Germans who deviate from the norm by working particularly lengthy workweeks will experience worse mental health. This effect may be even greater for German women, for whom long work hours conflict with gendered divisions of labor and the dominant breadwinner/homemaker model and for whom work beyond part-time is unsupported by Germany’s traditional welfare state (Korpi 2000). We also expect that Germans with extreme schedules are unlikely to have sought them out (see Reynolds 2004) and that this lack of scheduling fit contributes to any mental health penalties they experience. Our investigation provides a much-needed comparative approach to the study of work time and health and expands the current understanding of relationships between work hours and mental health beyond the U.S. context.

Background

Theoretical Mechanisms Linking Work Time and Mental Health

Work time and mental health are theoretically linked through a wide array of mechanisms. At least three areas of research contribute to our understanding of these processes: the work strain approach, the resource approach, and the work–life approach.

The job strain literature, which provides the dominant approach to linking work time and health, considers the demanding nature of long work hours as a potential health hazard. In classic theories of work strain, work time is considered to be one of many potential job demands that may lead to stress and its accompanying psychological and physical symptoms. According to these theories, the health effects of such job stressors may be worsened, and potentially moderated, by undesirable characteristics of the job, such as low levels of control, or buffered by positive features of the job, such as complexity and other resources (Karasek 1979; Siegrist 1996; for an updated approach, see Bakker and Demerouti 2007). Overall, long work hours are expected to create varying levels of distress and stress-related illness.

In contrast, the resource-based approach, which encompasses much sociological literature on unemployment, focuses on the health-promoting resources tied to employment, including socioeconomic status, social networks, and benefits. Studies consistently find worse health among those who are unemployed or underemployed (Bambra and Eikemo 2009; Kessler, Turner, and House 1988; Kleiner and Pavalko 2010; Ross and Mirowsky 1995) and suggest further that prolonged unemployment may worsen future health prospects (Wadsworth, Montgomery, and Bartley 1999). The unemployed are particularly vulnerable to stressful life events (Kessler et al. 1988) and report more depressive symptoms (Burgard, Brand, and House 2007). As many workers who work part-time are underemployed and would prefer to work more, these mental health implications of lack of work are highly relevant for individuals working relatively few hours (Jacobs and Gerson 2004; Reynolds and Aletraris 2006).

A third stream of research highlights how work time affects balance between work and family or the balance between work and nonwork. Work is only one dimension of life, but the way it fits into one’s life overall has implications for mental health (Gareis and Barnett 2002) and other health dimensions (Tucker and Rutherford 2005). For instance, demanding work hours in Germany (Beham and Drobnič 2009) and the United States are associated with lower satisfaction with work–family balance and chronic feelings of time pressure regarding personal or family time (Roxburgh 2004, 2006). These challenges, in turn, are associated with distress (Roxburgh 2004, 2009).

Examining the relationship between working hours and mental health outcomes is complex, however, because work hours vary by gender, occupation, and income (Kalleberg, Reskin, and Hudson 2000), and those with health problems may be pushed out of the labor force or work reduced hours (Pavalko and Smith 1999; Stewart 2001). While studies from the United States, Europe, and Asia find a negative association between long hours and mental or physical health, these generally look within only one occupation or industry, one firm, or one gender (e.g., Kivimäki et al. 2011; Rosta and Aasland 2011; Trinkoff et al. 2006; Virtanen et al. 2012). Yet, social institutions, such as gender and social class, are also related to health (Bird and Fremont 1991; Keyes 2002; Schunck and Rogge 2012). Studies utilizing broader, nationally representative samples, however, are typically restricted to one point in time (e.g., Grosch et al. 2006; Wirtz and Nachreiner 2010), limiting the ability to model health symptoms or conditions that take time to develop.

Recent research from the United States using nationally representative longitudinal data to look across a range of occupations finds that part-time workers at midlife had worse health than full-time workers, due to their differential work and family circumstances. When controlling for gender, work characteristics, income, and work-limiting health limitations, researchers found a nonlinear pattern emerged in which moderately long hours, but not very long hours, were associated with worse mental health scores and greater distress as well as worse physical health (Kleiner and Pavalko 2010). Other research using nationally representative data suggests that Americans who work the longest hours have differential access to the health-promoting benefits of financial resources and hours flexibility (Golden 2008) and may, counterintuitively, be happier than those working fewer hours (Golden and Wiens-Tuers 2006).

Overall, the three approaches to work hours and mental health highlight very different processes, but evidence suggests that all three mechanisms—resources, job stress, and the fit between work and nonwork—are at play. We particularly expand upon this latter perspective, using a comparative approach to attend to the broad sociopolitical contexts in which work time is scheduled, experienced, and interpreted (see Levecque et al 2011).

The Importance of Examining Context

The context in which work is organized helps unpack the broader meaning of work time. Although the distressing consequences of long work hours and the health-promoting financial resources gained through employment might universally affect mental health, these processes are likely to be stronger in some countries than in others (e.g., Olafsdottir 2007). In particular, welfare state differences create unique levels of dependence upon the labor market (Bambra and Eikemo 2009; Bell and Freeman 2001; Levecque et al. 2011), and norms of working time that vary by country or region may systematically create greater or lesser fit for those working an undesirable number of hours (e.g., Berg et al. 2004; Gareis and Barnett 2002).

There are several underlying processes that currently push people to work more in the United States and to work less in Germany. Welfare state differences across the two countries and differences in laws regulating working time are two important distinctions that likely produce varying motivations to work and potentially dissimilar mental health consequences for work. Welfare state differences, noted previously, affect reliance on the labor market and reduce (Germany) or increase (United States) the importance of work hours to provide basic needs and health-promoting resources. Laws and trade union agreements concerning working time—which are extensive in Germany—reinforce and differentially reward or prohibit particular durations of work.

Overall, the German legal context is designed to dissuade people from working long hours, penalizing extremely long hours by making them illegal, withholding incentives to work long hours, and minimizing penalties to those who work less than full-time. Germans must abide by stringent European Union (EU) work time laws as well as additional national regulations. The EU working directive (2003) prohibits working time of greater than 48 hours per week (on average over a four-month period) and specifies minimum daily and weekly rest periods away from work, mandatory rest breaks on the job, and minimum weeks of annual and parental leave. The German Hours of Work Act of 1994 further outlines, in 26 sections, a wide range of working time laws.

Institutionalized practices and resources in Germany reinforce the norm favoring shorter work hours, and it is possible that by improving conditions for these workers, shorter work hours also help promote their mental health. For example, part-time employment in Germany has not limited access to benefits, such as health care, number of vacation days, or employment protection, to the same degree as in the United States, which may serve to protect these employees’ well-being and mental health (but see O’Reilly 1996). Furthermore, all employment is accompanied by a contract specifying the number of working hours required by the job. If an employee is asked to work more than those contracted hours, added time may be accounted for later, for example, by lengthening one’s contracted vacation time. Universal monetary rewards for long hours have been avoided in exchange for a more regulated environment that does not incentivize overtime work.

Laws within the United States, on the other hand, create a reward structure for overtime work. Until recently, Americans relied on work—typically, full-time work—to provide health care (Kalleberg et al. 2000). Laws do not limit the amount of time employees may be asked to work. Thus, the United States has no nationally required vacation time, maximum weekly or daily hours worked by adults, or required breaks (although some states, like New York and California, do regulate break periods). The only national law on working time in the United States relates to wages. For a nonsalaried employee (over half the U.S. workforce), every hour worked beyond 40 hours per week must be paid at 150% (“overtime pay”). Salaried employees do not typically get overtime pay for long work hours but have tended to receive more benefits from their employer, such as health insurance and retirement packages, educational benefits, sick days, and vacation days; and they often have greater control over their schedule than nonsalaried employees (Averett and Hotchkiss 1995). Thus, in the United States, long hours are often associated with greater rewards, although the choice to work those hours may be the decision of the employer, not the employee.

Where long work hours are less normative, there may be fewer perceived social and personal benefits for those who work extended hours. The legal context of working time should systematically affect the subjective “fit” of long hours in Germany versus the United States, contributing to lower mental health among Germans working longer hours. Those working non-normative hours may be subject to greater social disapproval and may hold internalized standards for themselves that they cannot reach, making long hours more stressful.

In Germany, lack of “fit” with long hours should be particularly magnified among working women. German family policies and welfare state arrangements provide little support for women’s labor force participation (Korpi 2000). Whereas the United States exemplifies a market-oriented family policy model, in Germany, traditional family policies support the dominant male-breadwinner/female-homemaker arrangement (Pfau-Effinger 2004), and provide little support for female labor force participation beyond part-time work (Boye 2011). Most employed German women work part-time and cite family reasons for doing so (Walwei 2014). The German context, however, may also create a heightened selection process, with high demand for part-time jobs among women and only women in unique or elite circumstances working long hours (Boye 2011).

Given the broader contexts of working time arrangements, including differences in the extent to which working time is gendered, we ask whether Germans experience (1) fewer mental health penalties to shorter schedules and (2) greater mental health penalties to longer schedules, compared to U.S. workers. We also tap into a key dimension of scheduling fit by asking (3) whether Germans who work long hours sought them out through contractual employment and (4) whether desired changes in hours helps explain any mental health effects. Finally, because German women may experience harsher consequences to long hours of work, we test for (5) gender differences in our key findings.

Data and Methods

Data

To compare relationships between work time and health in the United States and Germany, we analyzed longitudinal data from each country. We used American data from the National Longitudinal Survey of Youth (NLSY) to examine working time and mental health in the United States and to replicate existing research (Kleiner and Pavalko 2010) in a way that can be compared to parallel analyses in Germany.

The NLSY was a study launched by the Bureau of Labor Statistics in 1979. The initial sample was nationally representative and included 12,686 U.S. young adults, ages 14 to 22, in 1979. Two supplemental oversamples were dropped in the 1980s and 1990s, yielding an original sample of 9,964 individuals who continued to be followed into adulthood (Bureau of Labor Statistics 2004).

The oldest cohort members of the NLSY turned 40 in the late 1990s, and a “40-and-older” health module was added to the survey at this time. Respondents completed this battery of health questions once, in the survey wave between 1998 and 2004 that was closest to their 40th birthday. We refer to the survey when respondents completed the health module (at approximately age 40) as T₂. To maintain the temporal ordering of work hour effects on health, all independent variables and controls were drawn from the prior surveys, when respondents were approximately 38 (T₁). This design allowed us to examine the effect of work hours at approximately age 38 on the respondent’s health at age 40. We excluded respondents who were not currently working at T₁ in order to control for job characteristics, and to better compare workers in the United States and Germany.

Our analysis of the NLSY was limited to respondents who were working at T₁ and had valid responses to the 40-and-older health module. Of the 12,686 original respondents in the main sample, 8,399 were still in the sample in 1998 (an 84.3% retention rate).¹ Of these, 6,288 were working at T₁. We lost 14.46% due to missing data (9.03% missing on income), for a total sample size of 5,379.

The German data for our study come from the German Socio-economic Panel (SOEP). The SOEP is an ongoing, representative, longitudinal survey of German households (Wagner, Frick, and Schupp 2007), with currently around 20,000 individual respondents. It has been conducted annually since it was first administered in 1984. To ensure continuous national representativeness, the SOEP is regularly enlarged by inclusion of new subsamples. In 2002, the SOEP respondents were asked a range of health questions. Because the SOEP is a household survey, observations within households are nonindependent. We adjust for this nonindependence by estimating cluster-robust standard errors in the regression models (Rabe-Hesketh and Skrondal 2008).

In order to make the German data comparable to U.S. data, we restricted the sample to respondents who were between 38 and 44 years of age at T₂ (2002, when the health questions were fielded) and who had valid data on work hours and other important covariates at T₁, again defined as two years prior. Due to differences in the study designs, we enlarged the age spread in the German data slightly compared to the U.S. data to obtain a sufficient number of cases. Of the initial 24,576 respondents, 2,503 are working and between 36 and 42 years at T₁. After listwise deletion of missing values (15.7%), the final sample comprised 2,110 respondents. Most of the missing values were due to nonresponse on job satisfaction (9.15%).²

Measures

Mental Health

Mental health is often measured along a continuum, and we used the 12-Item Short Form Health Survey (SF-12) mental health score to capture the respondent’s overall level of mental health (Ware, Kosinski, and Keller 1996). The SF-12 mental summary score was weighted most heavily by self-reports on emotional experience, vitality, and role behaviors, including how often the respondent felt calm and peaceful, felt downhearted and depressed, had a lot of energy, accomplished less than he or she would like, did work or other activities less carefully than usual, or had physical health or emotional problems interfere with social activities. The SF-12 mental health score has been validated in a variety of populations—being able to discriminate, for example, those diagnosed with depression—and is widely used in both the United States and Germany (Gandek et al 1998; Ware et al. 1996). For both the United States and Germany, the SF-12 mental health scores from the SOEP and NLSY surveys have been standardized to have a mean of 50 and a standard deviation of 10 (Andersen et al. 2007; Bureau of Labor Statistics 2004). A lower score indicates worse health, and a higher score indicates better health. We have restandardized the SF-12 scores within each sample to facilitate cross-national comparisons.

Work Hours

Table 1 shows descriptive statistics for all variables used in the analysis. Except for percentage of each sample in self-employment, all means were significantly different across the two countries. Our key independent variable was number of hours the respondent worked per week, which we divided into six categories based on previous findings showing a nonlinear relationship between work hours and mental health (Kleiner and Pavalko 2010). We analyze reported hours worked, as opposed to hours one is contractually obligated to work. As shown in Table 1, 40 hours per week was the most commonly worked schedule in the American data (reported by 37% of workers). Although sectoral agreements within Germany frequently define normal working hours to be around 39 to 40 hours per week, in the German data, 40 hours was not the norm. Twenty-one percent of German respondents were working 30 to 39 hours, 21% working 40 hours, and 23% working >40 to 49 hours per week. Moderately long hours (>40–49) are more common in the German data (23% vs. 18%), while very long work hours (50–59 or 60 and greater) are more common in the American data (15% vs. 11% and 10% vs. 8%, respectively).

Table 1.

Descriptive Statistics for All Variables: NLSY (1996–2004, N = 5,379) and SOEP (2000–2002, N = 2,110).

	United States				Germany
Variable	M	SD	Min	Max	M	SD	Min	Max
SF-12 mental health score	50	10	−10.01	72.66	50	10	6.52	78.29
Own hours <30	.09	.29	0	1	.17	.38	0	1
Own hours 30–39	.10	.29	0	1	.21	.40	0	1
Own hours 40	.37	.48	0	1	.21	.40	0	1
Own hours >40–49	.18	.39	0	1	.23	.42	0	1
Own hours 50–59	.15	.36	0	1	.11	.31	0	1
Own hours 60+	.10	.30	0	1	.08	.27	0	1
No spouse	.39	.49	0	1	.24	.43	0	1
Spouse not working	.10	.30	0	1	.22	.41	0	1
Spouse working <40 hours	.10	.31	0	1	.24	.43	0	1
Spouse working 40 hours	.24	.43	0	1	.10	.30	0	1
Spouse working >40 hours	.16	.37	0	1	.20	.40	0	1
Female	.49	.50	0	1	.45	.50	0	1
Black	.28	.45	0	1
Hispanic	.18	.39	0	1
Migration background					.13	.33	0	1
Education (years)	13.36	2.39	7	20	12.32	2.49	7	18
Age	38.42	.57	37	39	38.97	1.98	36	42
Number of children	1.47	1.29	0	9	1.33	1.06	0	8
Child under age 6 at home	.21	.41	0	1	.25	.43	0	1
Log, household labor income	10.28	1.86	0	13.13	10.61	.56	6.46	12.47
Living in East Germany					.28	.45	0	1
Weeks worked per year	48.33	9.88	1	52
Managerial/professional	.30	.46	0	1	.20	.40	0	1
Technical/sales/administration	.27	.44	0	1	.39	.49	0	1
Service	.13	.34	0	1	.10	.30	0	1
Farming/craft	.13	.34	0	1	.19	.39	0	1
Operator/laborer	.17	.37	0	1	.12	.32	0	1
Government/civil service	.17	.37	0	1	.24	.43	0	1
Private/nonprofit	.74	.44	0	1	.67	.47	0	1
Self-employed	.09	.29	0	1	.09	.29	0	1
Job satisfaction	3.39	.69	1	4	3.01	.70	1	4
Job tenure (weeks)	324.28	292.84	1	1,477	459.82	358.13	0	1,394
Health limits work	.07	.25	0	1	.20	.40	0	1
Number of observations	5,379				2,110

Note: NLSY = National Longitudinal Survey of Youth; SOEP = German Socio-economic Panel; SF-12 = 12-Item Short Form Health Survey (Ware, Kosinski, and Keller 1996).

In examining work hours by gender (not shown but available upon request), we noted that nearly twice the percentage of German women work shorter hours (less than 40 hours), at 61%, compared to U.S. women (31%), while 28% of German men work these shorter hours, compared to only 8% of U.S. men. More Americans worked 40 hours (about 37% of the sample, with no gender difference), compared to 23% of German men and 18% of German women. Differences in long hours (>40) were most pronounced between German (21%) and American (36%) women. A similar percentage (57% vs. 59%, respectively) of American and German men reported these longer hours, but more of the American men (35%) worked >50 hours than the German men (28%).

Control Variables

As previous studies have shown (Jacobs and Gerson 2004; Kalleberg et al. 2010), work time may covary systematically with other job and familial characteristics. It is important to control for covariates, such as gender, presence of children in the household, spousal working hours, income, education, race, ethnicity, migration background, and job characteristics.

Both samples comprise fewer female respondents, 49% in the U.S. data and 45% in the German data. In the U.S. data, we differentiate between Hispanic and non-Hispanic respondents and between black and nonblack. In the German data, we distinguish first- or second-generation immigrants (13%) from nonimmigrants (87%). In the German data, we also distinguish between respondents living in East Germany (28%) and those living in West Germany (72%).

Since the effect of work time on mental health may depend in part on respondents’ familial situations, we control for the number of children, whether children under the age of six are present in the household, respondent’s age, and whether the respondent has a spouse (61% in the U.S. data and 76% in the German data). If the respondent had a spouse, we controlled for the spouse’s working hours, as reported by the respondent in the United States or by the spouse in Germany. As shown in Table 1, it was more common in the United States that both partners work. More than twice as many German spouses did not work. In the American data, most spouses worked 40 hours (24%), whereas working fewer hours (10%) and working longer hours (16%) was less common. The German data provided a different picture: A spouse working 40 hours was not the norm (10%), and approximately 24% of spouses worked fewer than 40 hours, while 20% worked more. These differences were driven in part by the stronger prevalence of the male-breadwinner model in Germany (Pfau-Effinger 2004). Nearly two thirds of the male respondents in the German sample (and less than a quarter of the female respondents) were married to a nonworking spouse or a spouse who worked less than 40 hours. In the United States, only one third of the male respondents (and less than a 10th of female respondents) were married to a nonworking spouse or a spouse who worked less than 40 hours (findings by gender not shown but available upon request).

Additionally, we control for socioeconomic status by measuring education, income, and job characteristics. Education is measured in years and is truncated at a minimum of seven years. Although educational systems vary considerably between the two countries, the average number of years spent in education differed by only one year. To control for the respondents’ financial situation, we included the log of annual household labor income in U.S. dollars or Euros. In the U.S. data, we also controlled for the number of weeks worked last year, to ensure that the less regulated, more variable vacation weeks were not strongly affecting the results. In contrast, because vacation weeks are legally regulated and do not depend on the hours worked in Germany, this type of control was not necessary (German Vacation Time Act 1963). Since job types and sectors may have different standards for work time and may vary in the strains they put on workers, we differentiate managerial and professional jobs (United States, 30%, vs. Germany, 20%); technical, sales, and administrative jobs (27% vs. 39%); service jobs (13% vs. 10%); farming and craft jobs (13% vs. 19%); and jobs as operators and laborers (17% vs. 12%) based on the International Standard Classification of Occupations. We also distinguish workers who are employed in the governmental and civil service sector (U.S., 17%, vs. Germany, 24%) from those in the private and nonprofit sector (74% vs. 67%) and those who are self-employed (9% vs. 9%). We included self-employed respondents because of the phenomenon of Scheinselbständigkeit (fictitious self-employment) in the German labor market, in which ostensibly self-employed workers are, in reality, dependent employees.

Job characteristics may also vary with work hours and mental health, so we included job satisfaction and job tenure as additional controls. In the NLSY, job satisfaction was measured using a 4-point scale, with higher values indicating higher job satisfaction (mean value 3.39). The German variable was measured on an 11-point scale, ranging from 0 (completely unsatisfied) to 10 (completely satisfied), but was recoded to a 4-point scale (0, 1, 2 = 1; 3, 4, 5 = 2; 6, 7, 8 = 3; 9, 10 = 4; mean value 3.01) to make it comparable to the U.S. data. Job tenure was measured in weeks in both the United States (mean value 324.28) and the German data (mean value 459.82).

Finally, to control for possible selection effects related to workers with health problems, we included a variable that indicates whether the respondents are limited in their work (U.S., 7%) or their daily activities (Germany, 20%) because of health impairments. The larger percentage of German workers reporting work-limiting health conditions was likely due to greater access to paid sick days and compensated disability-based reductions in work time or limitations on the type of work allowed.

Analytic Strategy

Our initial multivariate analyses proceed in two steps for each country. We first estimated a “baseline” model that captured the unadjusted relationship between respondents’ working time and mental health score in each country. Next, we added a set of controls to account for socioeconomic, occupational, and other sociodemographic characteristics that are likely to relate to both work hours and to health. These included education, age, number of children, presence of a child under the age of six at home, log of household earned labor income, occupational type, job sector, job satisfaction, job tenure, and whether the respondent’s health limited the amount or type of work one could perform. In the United States, race (black vs. not), ethnicity (Hispanic vs. not), and weeks worked per year were also included as controls. In Germany, immigrant status and East Germany residence were also controlled.

After examining work time and health within both countries using regression analysis, we descriptively investigated the extent to which work hours are beyond the expected schedule for each respondent. Within the United States, for two of the four waves, we were able to investigate whether the respondent worked any “extra” hours last week and, if so, how many. Within Germany, where detailed job contracts are the norm, we were able to investigate the extent to which typical working time was beyond the contracted hours and, if so, how many hours were generally worked beyond those specified in one’s contract. We also explored the extent to which the German respondents wish to change their hours.

Finally, for the German sample only, we included desired changes to hours in a final set of regression models predicting the relationship between hours and mental health. We also compared findings for German men and women in the relationship between work hours and health, with and without controlling for desired changes to hours.

Results

Table 2 presents the ordinary least squares regression coefficients for the unadjusted bivariate relationships between work hours and mental health in the United States and Germany (Models 1a and 2a). Overall, the relationships between work hours and mental health show distinct differences between the two countries. The most dramatic distinction is in regard to work time shorter than 40 hours per week. Specifically, we find that shorter work time is associated with significantly lower SF-12 mental health scores in the United States (compared to 40 hours per week). However, in Germany, this relationship does not hold.

Table 2.

Relationships between Work Hours and Mental Health in Germany (SOEP, 2000–2002, N = 2,110) and the United States (NLSY, 1996–2004, N = 5,379).

	Germany		United States
Variable	Model 1a	Model 2a	Model 1b	Model 2b
R works <30 hours/week	.01	1.09	−1.39**	−.19
R works <30 hours/week	(.69)	(.80)	(.50)	(.52)
R works 30–39 hours/week	.07	.25	−1.17*	−.30
R works 30–39 hours/week	(.69)	(.66)	(.49)	(.49)
R works 40 hours/week	Ref.	Ref.	Ref.	Ref.
R works >40–49 hours/week	−.81	−1.01	−.76*	−1.06**
R works >40–49 hours/week	(.68)	(.67)	(.39)	(.39)
R works 50–59 hours/week	.37	−.87	.112	−.34
R works 50–59 hours/week	(.81)	(.84)	(.42)	(.42)
R works 60+ hours/week	−1.09	−2.29*	.45	.05
R works 60+ hours/week	(.93)	(.97)	(.45)	(.48)
Constant	5.22***	33.21***	5.32***	42.79***
Constant	(.50)	(6.71)	(.22)	(9.07)
R ²	.002	.07	.003	.06
Observations	2,110	2,110	5,379	5,379

Note: SOEP = German Socio-economic Panel; NLSY = National Longitudinal Survey of Youth; R = respondent; Ref. = reference. Models 1a and 1b present unadjusted models. Models 2a and 2b control for gender, age, marital status and spousal working hours, children under six years old, number of children in household, racial-ethnic background (United States), migration background (Germany), education, log household income ($ or €), living in East Germany, weeks worked per year (United States), job type, sector, job tenure, job satisfaction, and whether health limits work at Time 1.

p ≤ .05, **p ≤ .01, ***p ≤ .001.

For working time longer than 40 hours, distinctions at the bivariate level are less dramatic, and the results point to some similarities. Both countries show a similar negative association between working time greater than 40 hours but less than 50 hours per week (β = −.81 in Germany, β = −.76 in the United States) and mental health, although this is statistically significant only in the U.S. sample. For even longer work hours, however, the patterns noticeably diverge (although there are no significant relationships in these bivariate models). In Germany, the strongest negative association with mental health (β = −1.09) is shown for work hours greater than 60 per week. In contrast, the coefficient for the longest work hours is positive (β = .45) within the U.S. data. These models, however, show a low R², indicating that very little of the variation in mental health is explained by hours alone, and no covariates.

Models 2a and 2b show results for adjusted models predicting work time and health, accounting for work, family, and economic controls and work-limiting health conditions at T₁.

As hypothesized, working very long hours (more than 60 hours per week) in Germany is significantly associated with worse mental health at T₂ (β = −2.29, p < .05). A separate Wald test indicated that this effect is significantly different from the United States (p ≤ .05). Hours of work more than 40 hours per week are also negatively associated with mental health, although not statistically significant at p ≤ .05. The coefficients for work hours less than 40 hours per week are positive and much larger than the unadjusted coefficients (Model 1a), although still not significant.

In contrast, while the negative relationship with long hours increases with progressively longer work schedules in Germany, in the United States, the negative relationship between long hours and mental health (compared to 40 weekly hours) declines as working time becomes increasingly prolonged (net of controls). Working between 40 and 49 hours per week has the strongest negative relationship with mental health (β = −1.06, p < .01), while the coefficient for work between 50 and 59 hours per week (β = −.34) is not statistically significant. Work at 60 hours per week or greater is not negatively associated with mental health (β = .05).

Although the unadjusted associations presented in Table 2 showed significant negative relationships between part-time work (compared to 40 hours per week) and mental health scores in the United States, the adjusted models show that much of this effect is accounted for by relevant covariates, such as socioeconomic status, work environment, and family status. In these models, the relationships between lower work hours and mental health are no longer significant, although they are still negative within the U.S. sample. Interestingly, the reverse effect is observed for Germany. In Table 2, the coefficients for part-time work were positive but close to zero. Controlling for additional factors, in Table 3, strengthens the positive relationship between hours of work and mental health for these shorter work hours.

Table 3.

Characteristics of Work Schedules in Germany (SOEP, 2000, N = 2,110) and the United States (NLSY, 1996–2002, N = 5,379).

	Germany	United States	Germany			United States
Usual Hours Worked	Average Hours	Average Hours	Working More Than Contract?^a	By How Much?^a	Desired Change?^b	Worked Extra Hours Last Week?^c	By How Much?^c
<30	18.0	18.5	32.7%	1.0	2.27	11.9%	8.8
30–39	35.7	34.4	24.8%	1.0	−1.33	12.2%	9.0
40	40.0	40.0	45.9%	1.4	−2.84	17.2%	6.9
>40–49	44.2	44.7	95.8%	5.0	−5.91	26.1%	9.0
50–59	51.4	52.5	96.2%	11.5	−9.87	21.2%	13.9
60+	64.3	72.1	96.9%	21.4	−17.13	20.3%	19.3

Note: SOEP = German Socio-economic Panel; NLSY = National Longitudinal Survey of Youth.

n = 1,787.

n = 2,074.

n = 2,141 (Time 1 = 1996 or 1998).

In a final set of analyses, we further investigate distinctions in the meaning of work hours in Germany versus the United States. Table 3 shows descriptive findings for working time within the two countries. First, for each country, we show the average work hours represented by each work time category. Second, for each country, we show the extent to which each work hour category represents work beyond contracted hours (Germany) or extra hours worked in the last week (United States). Finally, we show the average number of hours worked beyond one’s contracted hours and desired changes to work hours (Germany), and average number of “extra” hours worked in the last week (United States), by overall hours worked.

The first two columns of Table 3 allow us to assess whether the average hours worked within the United States and Germany differ systematically by hours category. As shown, hours worked within each general range are quite comparable between the two countries, with one exception. Notably, U.S. respondents who work the longest hours (>60 hours per week) report substantially more average hours than Germans with the same number of hours (>60 hours per week). Americans report working 72.1 hours, and Germans 64.3, a difference of 7.8 hours.

Table 3 also shows the percentage of respondents working longer hours than expected based on their work contracts (Germany) or usual hours worked per week (United States). There are several noticeable differences between the two countries. First, although working longer than contracted hours is common in Germany among workers reporting 40 or fewer weekly hours (between 25 and 46% of German respondents working ≤40 hours report typically working more than their contract), the average increase for these added weekly hours is approximately 1 hour per week. In the United States, the average number of “extra” hours worked is much larger—between seven and nine hours per week—but a smaller proportion of individuals working ≤40 hours reported any “extra” hours last week (between 12% and 17%).

Among respondents working long work hours (>40 hours per week), the differences across the two countries are more striking. Approximately 21% to 26% of Americans who usually worked long hours reported working “extra” hours in the past week, compared to 96% to 97% of Germans. Americans working >40 to 49 hours per week are most likely to report having worked extra hours in the past week, whereas Germans working 60 or more hours per week are slightly more likely to report working more than stated on their contract.

The number of additional hours worked also differs across the two countries. For work of 40 to 49 and 50 to 59 hours per week, Americans average more extra hours (nine and 14, respectively) than their German counterparts (five and 11.5, respectively). At 60+ hours per week, however, Germans working beyond their contracts average longer excess hours than Americans (21.4 vs. 19.3).

In a final set of regression models, focused solely on Germany, we examine whether desired changes (increases or decreases) to weekly work hours account for relationships between long hours and health. As a further step, we also examine gender differences in these patterns.³ The total sample size in these analyses is reduced by 26 cases due to missing data on how many hours the respondent would prefer to increase or decrease from their weekly workload.

Table 4 shows patterns of work time and mental health in Germany (in the slightly reduced sample), with long work hours predicting lower mental health (Model 1a). Model 2b shows the hours–health relationships after accounting for desired changes to work hours. Controlling for desired work hours reduces the effects of all coefficients, but particularly the 60+ hours category, resulting in a nonsignificant effect of long hours on mental health (from β = −2.05, p < .05, to β = −.57, ns). The estimated effect of desired hours indicates that wanting more weekly work hours is associated with better mental health (β = .10, p < .01), and conversely, wanting to reduce work hours, which applies to the majority of the respondents, is associated with worse mental health.

Table 4.

Regressions Predicting Mental Health in Germany, Adjusting for Desired Changes to Work Hours (SOEP, 2000–2002, N = 2,110).

	Full Sample		Women		Men
Variable	Model 1a	Model 1b	Model 2a	Model 2b	Model 3a	Model 3b
R works less than 30 hours	1.30	.60	2.02	1.07	2.14	1.00
R works less than 30 hours	(.81)	(.84)	(1.10)	(1.17)	(1.78)	(1.85)
R works 30–39 hours	.50	.30	1.19	.83	.01	−.11
R works 30–39 hours	(.67)	(.68)	(1.10)	(1.11)	(.85)	(.85)
R works 40 hours/week	Ref.	Ref.	Ref.	Ref.	Ref.	Ref.
R works >40–49 hours	−.83	−.51	−.25	.14	−1.24	−.97
R works >40–49 hours	(.67)	(.68)	(1.31)	(1.31)	(.77)	(.79)
R works 50–59 hours	−.50	.26	3.62^a	4.59^b*	−1.72^a	−1.07^b
R works 50–59 hours	(.85)	(.88)	(1.86)	(1.87)	(.97)	(1.02)
R works 60 or more hours	−2.05*	−.57	.52	2.25	−2.88*	−1.57
R works 60 or more hours	(.98)	(1.05)	(1.91)	(1.98)	(1.18)	(1.29)
Desire to increase/reduce (in hours)		.10**		.12*		.08
Desire to increase/reduce (in hours)		(.03)		(.05)		(.04)
Constant	33.81***	33.55***	33.20***	31.57**	3.96***	32.04***
Constant	(6.80)	(6.77)	(9.72)	(9.70)	(9.12)	(9.09)
R ²	.075	.079	.062	.067	.092	.095
Observations	2,074	2,074	930	930	1,144	1,144

Note: SOEP = German Socio-economic Panel; R = respondent; Ref. = reference. Models control for gender, marital status and spousal working hours, education, log household income ($ or €), living in East Germany, job type, job tenure, job satisfaction, and whether health limits work at Time 1. Standard errors in parentheses.

Significant gender differences for Models 2a and 3a.

Significant gender differences for Models 2b and 3b.

p ≤ .05, **p ≤ .01, ***p ≤ .001.

As shown in the last four columns of Table 4, we also examined whether the main findings for these analyses differ by gender and conducted Wald tests to formally assess significant gender differences in these coefficients. As shown in Models 2a and 3a, the relationship between work hours and mental health differs by gender. At 50 to 59 weekly hours, there is a positive but nonsignificant relationship between work hours and mental health among the women (β = 3.62, ns), whereas among the men, there is a negative relationship between working 50 to 59 hours and mental health (β = −1.72, p < .01). These coefficients are significantly different from each other (p < .01). Although there is no significant difference at 60+ weekly hours, the coefficient signs are again reversed for women as compared to men. Women at these levels of work are rare, however—34 women worked 50 to 59 hours per week, and only 26 women worked 60+ hours per week—so there is likely a lack of statistical power for the women.

Given the gender difference in the relationship between long hours and mental health, it is interesting to note that there was no gender difference in the relationship between desired work hour changes and mental health. This effect explains the negative relationship between 60+ work hours and mental health among men. Among women, however, desired changes suppress the extent of the positive relationship between hours and mental health, whereas controlling for desired change reveals a significant positive association between working 50 to 59 hours per week and mental health.

As a final step, we present descriptive findings for the German men and women by hours category (Table 5). Most notably, women in the shortest schedules seek smaller hours increases than men, and at all other levels of work, women prefer greater hours reductions than men, even when hours worked beyond their contracts are similar or less than men’s. While the NLSY does not include comparable data, other research finds that Americans’ desired changes to hours are quite similar by gender (Jacobs and Gerson 2004). The gender differences in desired hours changes within Germany shown in Table 5, however, do not produce stronger relationships between long hours and mental health among women (Table 4).

Table 5.

Comparison of German Men’s and Women’s Work Schedules (SOEP, 2000, N = 2,110).

	Average Hours		More Than Contract?^a		By How Much?^a		Desired Change^b
Usual Hours Worked	Men	Women	Men	Women	Men	Women	Men	Women
<30	16.41	18.08	25.0%	32.7%	1.38	.95	11.8^c	1.99^c
30–39	37.34^c	34.31^c	18.7% ^c	30.4% ^c	.34^c	1.55^c	−0.24^c	−2.30^c
40	40	40	49.6%	40.0%	1.46	1.37	−1.56^c	−4.86^c
>40–49	44.28^c	43.80^c	95.8%	95.8%	5.08	4.59	−4.92^c	−8.53^c
50–59	51.52	51.03	95.4%	100.0%	11.29	12.39	−9.22^c	−13.32^c
60+	64.44	63.81	98.0%	93.3%	21.90	19.73	−16.68	−19.52

Note: SOEP = German Socio-economic Panel.

n = 1,787.

n = 2,074.

Significant gender difference.

Discussion and Conclusions

Concerns regarding potential negative health effects due to long work hours are evident in many developed countries and are particularly growing within Europe. The present study examined relationships between work hours and mental health in two countries that differ strongly in their working time norms for men and women as well as in related legislation and sector agreements. The analyses modified and replicated existing U.S. research using nationally representative longitudinal data on Germany and further investigated national distinctions in the underlying meaning of working time in the two countries.

Overall, we found distinct differences in the patterned relationship between long work hours and mental health in Germany versus the United States. While the longest work hours in the United States (60+ hours per week) are not worse for mental health than working 40 hours per week, in Germany, these very long hours are significantly related to lower scores on the SF-12 mental health scale. Those who work long hours in Germany overwhelmingly report that they are working considerably more hours than agreed upon with their employer, and most indicate that they would prefer fewer hours even though it would reduce their income. Our findings suggest that working time norms and legislation reinforcing those norms, as well as control over hours and perceived fit of those hours, may each contribute to links between long work hours and mental health in Germany.

We also observed national differences in relation to shorter part-time work and mental health. In the United States, part-time employment tends to be stratified and to yield lower wages and no benefits. While U.S. workers who work part-time report worse mental health, in Germany, we fail to find a negative relationship between shorter hours of employment and mental health. We suspect that benefits accrued to German workers, such as health insurance, job protection, and vacation time, help to minimize negative mental health effects of these shorter work hours and allow for benefits, such as greater time for leisure and family life, to enhance mental health.

Our examinations of the two countries revealed further distinctions regarding the underlying norms of working time. Although very long work hours predict worse mental health in Germany but not the United States, we noted that the respondents working long hours in the United States report longer average hours. Upon further investigation, we found that nearly all respondents working beyond 40 hours per week in Germany report working in excess of their job contracts, although this was slightly more common among the most extreme number of working hours (60+ hours per week). Respondents who work these particularly long hours in Germany (60+ hours per week) averaged more than 21 weekly work hours beyond the working time agreed upon in their contracts. The United States does not require work hour expectations to be set out in a job contract, but we observed that working “extra” hours in the past week was reported most frequently among those working moderately long hours (>40–49) rather than among those working the longest hours.

These findings suggest that the distinct mental health patterns we observe in relation to long work hours in each of the two countries may be due to differences in the meaning, desirability, and voluntary nature of long hours (Berg et al. 2004). Because long work hours in the United States may be sought out and expected as part of working a salaried job (exempt from overtime), or else compensated at one-and-a-half-time pay, particularly long hours may be less likely to be perceived as problematic and disruptive to expectations. Other factors, such as overrepresentation of salaried workers in the 60+ weekly hour category who may have better health or positive selection of highly motivated nonsalaried employees into more demanding weekly work hours, might also underlie or contribute to this pattern. In Germany, where long work hours are discouraged, prohibited unless meeting particular requirements, and consequently often not compensated, workers may be more likely to hold idealistic expectations for personal or family time that are, consequentially, disrupted by long hours.

Finally, in further investigating the findings for Germany, we found that desired changes to weekly work hours explains the negative relationship between 60+ work hours and mental health. However, this finding held only for men in the sample. Among women, we found no negative relationship. We found a suppression effect of desired changes to hours, resulting in a positive relationship between 50 to 59 weekly work hours and mental health. Given the very small numbers of women working particularly long hours in Germany, and the particularly strong norms against long hours for women, our findings implicate a selection effect that operates by gender.

Our study has several important limitations. Perhaps most consequential is that our data only begin to tap into the meaning of particular working time arrangements in each country. A fine-grained examination of subjective preferences for hours, reasons for working long hours, and the resulting challenges and rewards of extended hours would more fully flesh out the mechanisms by which work hours relate to mental health. Unfortunately, this is beyond the scope of this study. Nonetheless, this line of research could likely be extended using a well-designed qualitative approach. Furthermore, as raised earlier in this discussion, possible health selection into work hours cannot be fully disentangled from health effects. Future research using more extensive longitudinal data should attend to this issue.

Our research suggests several policy-relevant findings. First, Germans working part-time are, surprisingly, not at a mental health disadvantage. We suspect that the regulatory environment, which protects these workers from stressors, such as job insecurity, health care insecurity, and difficulty taking time off, might provide some important clues for improving the overall mental health of part-time Americans. Second, German men working long hours are clearly at a mental health disadvantage. Many of the long work hours reported by Germans may be illegal under German law, and the discrepancy between legal rules and actual practices warrants greater oversight. Finally, we suspect that the negative mental health associations with long work hours in each country may be related in part to underlying patterns of mandatory overtime. We urge greater attention to employee health across regions that vary by mandatory overtime protection as well as inclusion of questions on mandatory overtime, and the voluntariness of work hours, in population surveys.

Footnotes

Acknowledgements

The authors thank Eliza Pavalko and Sheryl Skaggs for their valuable feedback, and Nate Breznau for methodological advice.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This project was supported in part by a research grant from the DAAD (German Exchange Service), with hosting provided by the Bremen International Graduate School of Social Sciences, which funded the lead author, Sibyl Kleiner, to conduct this research in Germany.

Notes

Author Biographies

Sibyl Kleiner is a postdoctoral fellow in sociology at the University of Calgary. Her research interests are centered on family and work, health, gender, and social psychology. She focuses primarily on the health effects of family and work inequalities. In addition to continued research on work time and health, she is currently studying subjective time pressure, commuting time, and workplace sex discrimination.

Reinhard Schunck is an assistant professor in the Department of Sociology at Bielefeld University. His research interests span life course, social stratification and inequality, and quantitative research methods. He is interested in understanding how social contexts shape stratification outcomes across the life course. His current research focuses on immigrant health and on the determinants and health consequences of work-related stress.

Klaus Schömann is the program head of System and Politics at the German Institute for Adult Education in Bonn and a professor of sociology at Jacobs University Bremen. With interests in labor markets and education research, he has applied a life course perspective to the processes of earnings attainment, further training selectivity, job satisfaction, and entry into retirement. His current work deals with social selectivity in training and outcomes of training.

References

Andersen

Hanfried H.

Mühlbacher

Axel

Nübling

Matthias

Schupp

Jürgen

Wagner

Gert G.

2007. “Computation of Standard Values for Physical and Mental Health Scale Scores Using the SOEP Version of SF-12v2.” Schmollers Jahrbuch 127(1):171–82.

Averett

Susan L.

Hotchkiss

Julie L.

1995. “The Probability of Receiving Benefits at Different Hours of Work.” American Economic Review 85(2):276–80.

Bakker

Arnold B.

Demerouti

Evangelia

. 2007. “The Job Demands–Resources Model: State of the Art.” Journal of Managerial Psychology 22(3):309–28.

Bambra

Clare

Eikemo

Terje A.

2009. “Welfare State Regimes, Unemployment and Health: A Comparative Study of the Relationship between Unemployment and Self-reported Health in 23 European Countries.” Journal of Epidemiology and Community Health 63(2):92–98.

Beham

Barbara

Drobnič

Sonja

. 2009. “Satisfaction with Work–Family Balance among German Office Workers.” Journal of Managerial Psychology 25(6):669–89.

Bell

Linda

Freeman

Richard

. 2001. “The Incentive for Working Hard: Explaining Hours Worked Differences in the US and Germany.” Labour Economics 8(2):181–202.

Berg

Peter

Appelbaum

Eileen

Bailey

Tom

Kalleberg

Arne L.

2004. “Contesting Time: International Comparisons of Employee Control of Working Time.” Industrial and Labor Relations Review 57(3):331–49.

Bird

Chloe E.

Fremont

Allen M.

1991. “Gender, Time Use, and Health.” Journal of Health and Social Behavior 32(2):114–29.

Boye

Katarina

. 2011. “Work and Well-being in a Comparative Perspective: The Role of Family Policy.” European Sociological Review 27(1):16–30.

10.

Bureau of Labor Statistics, U.S. Department of Labor. 2004. NLSY User’s Guide: A Guide to the 1979–2002 National Longitudinal Survey of Youth Data. Columbus, OH: Bureau of Labor Statistics and Center for Human Resource Research.

11.

Burgard

Sarah A.

Brand

Jennie E.

House

James S.

2007. “Toward a Better Estimation of the Effect of Job Loss on Health.” Journal of Health and Social Behavior 48(4):369–84.

12.

Demetriades

Stavroula

Perdersini

Roberto

. 2009. Working Time in the EU and Other Global Economies. Luxembourg: Office for Official Publications of the European Communities.

13.

Gandek

Barbara

Ware

John E.

Aaronson

Neil K.

Apolone

Giovanni

Bjorner

Jakob B.

Brazier

John E.

Bullinger

Monika

Kaasa

Stein

Leplege

Alain

Prieto

Luis

Sullivan

Marianne

. 1998. “Cross-validation of Item Selection and Scoring for the SF-12 Health Survey in Nine Countries: Results from the IQOLA Project.” Journal of Clinical Epidemiology 51(11):1171–78.

14.

Gareis

Karen C.

Barnett

Rosalind Chait

. 2002. “Under What Conditions Do Long Work Hours Affect Psychological Distress? A Study of Full-time and Reduced-hours Female Doctors.” Work and Occupations 29(4):483–97.

15.

Golden

Lonnie

. 2008. “Limited Access: Disparities in Flexible Work Schedules and Work-at-home.” Journal of Family and Economic Issues 29(1):86–109.

16.

Golden

Lonnie

Wiens-Tuers

Barbara

. 2006. “To Your Happiness? Extra Hours of Labor Supply and Worker Well-being.” Journal of Socio-economics 35(2):382–97.

17.

Grosch

James W.

Caruso

Claire C.

Rosa

Roger R.

Sauter

Steven L.

2006. “Long Hours of Work in the U.S.: Associations with Demographic and Organizational Characteristics, Psychosocial Working Conditions, and Health.” American Journal of Industrial Medicine 49(11):943–52.

18.

Jacobs

Jerry A.

Gerson

Kathleen

. 2004. The Time Divide: Work, Family, and Gender Inequality. Cambridge, MA: Harvard University Press.

19.

Kalleberg

Arne L.

Reskin

Barbara F.

Hudson

Ken

. 2000. “Bad Jobs in America: Standard and Nonstandard Employment Relations and Job Quality in the United States.” American Sociological Review 65(2):256–78.

20.

Karasek

Robert

. 1979. “Job Demands, Job Decision Latitude, and Mental Strain: Implications for Job Redesign.” Administrative Science Quarterly 24(2):285–308.

21.

Kessler

Ronald C.

Turner

J. Blake

House

James S.

1988. “Effects of Unemployment on Health in a Community Survey: Main, Modifying, and Mediating Effects.” Journal of Social Issues 44(4):69–85.

22.

Keyes

Corey L. M.

2002. “The Mental Health Continuum: From Languishing to Flourishing in Life.” Journal of Health and Social Behavior 43(June):207–22.

23.

Kleiner

Sibyl

Pavalko

Eliza K.

2010. “Clocking In: The Organization of Work Time and Health in the United States.” Social Forces 88(3):1463–86.

24.

Kivimäki

Mika

Batty

G. David

Hamer

Mark

Ferrie

Jane E.

Vahtera

Jussi

Virtanen

Marianna

Marmot

Michael G.

Singh-Manoux

Archana

Shipley

Martin J.

2011. “Using Additional Information on Working Hours to Predict Coronary Heart Disease.” Annals of Internal Medicine 154(7):457–63.

25.

Korpi

Walter

. 2000. “Faces of Inequality: Gender, Class, and Patterns of Inequalities in Different Types of Welfare States.” Social Politics: International Studies in Gender, State and Society 7(2):127–91.

26.

Kroh

2010. “Documentation of Sample Sizes and Panel Attrition in the German Socio-economic Panel (SOEP).” DIW Data Documentation 50:1–49.

27.

Leschke

Janine

Watt

Andrew

Finn

Mairéad

. 2012. Job Quality in the Crisis: An Update of the Job Quality Index. Working paper 2012.07. Brussels: European Trade Union Institute.

28.

Levecque

Katia

Van Rossem

Ronan

De Boyser

Katrien

Van de Velde

Sarah

Bracke

Piet

. 2011. “Economic Hardship and Depression across the Life Course: The Impact of Welfare State Regimes.” Journal of Health and Social Behavior 52(2):262–76.

29.

Moen

Phyllis

Kelly

Erin L.

Tranby

Eric

Huang

Qinlei

. 2011. “Changing Work, Changing Health: Can Real Work-time Flexibility Promote Health Behaviors and Well-being?” Journal of Health and Social Behavior 52(4):404–29.

30.

Olafsdottir

Sigrun

. 2007. “Fundamental Causes of Health Disparities: Stratification, the Welfare State, and Health in the United States and Iceland.” Journal of Health and Social Behavior 48(3):239–53.

31.

O’Reilly

Jacqueline

. 1996. “Labor Adjustments through Part-time Work.” Pp. 566–93 in International Handbook of Labor Market Policy and Evaluation, edited by Schmidt

Günther

O’Reilly

Jacqueline

Schömann

Klaus

. Cheltenham, UK: Edward Elgar.

32.

Organization for Economic Co-operation and Development. 2004. The OECD Employment Outlook. Paris: OECD Publications.

33.

Organization for Economic Co-operation and Develop-ment. 2010. OECD Factbook 2010: Economic, Environmental and Social Statistics. Paris: OECD Publications.

34.

Pavalko

Eliza K.

Smith

Brad

. 1999. “The Rhythm of Work: Health Effects of Women’s Work Dynamics.” Social Forces 77(3):1141–62.

35.

Pfau-Effinger

2004. “Socio-historical Paths of the Male Breadwinner Model: An Explanation of Cross-national Differences.” British Journal of Sociology 55(3):377–99.

36.

Rabe-Hesketh

Sophia

Skrondal

Anders

. 2008. Multilevel and Longitudinal Modeling Using Stata. College Station, TX: Stata Press.

37.

Reynolds

Jeremy

. 2004. “When Too Much Is Not Enough: Actual and Preferred Work Hours in the United States and Abroad.” Sociological Forum 19(1):89–120.

38.

Reynolds

Jeremy

Aletraris

Lydia

. 2006. “Pursuing Preferences: The Creation and Resolution of Work Hour Mismatches.” American Sociological Review 71(4):618–38.

39.

Roxburgh

Susan

. 2004. “‘There Just Aren’t Enough Hours in the Day’: The Mental Health Consequences of Time Pressure.” Journal of Health and Social Behavior 45(2):115–31.

40.

Roxburgh

Susan

. 2006. “‘I Wish We Had More Time to Spend Together . . .’: The Distribution and Predictors of Perceived Time Pressures among Married Men and Women in the Paid Labor Force.” Journal of Family Issues 27(4):529–53.

41.

Roxburgh

Susan

. 2009. “Parental Time Pressures and Depression among Married Dual-earner Parents.” Journal of Family Issues 33(8):1027–53.

42.

Ross

Catherine E.

Mirowsky

John

. 1995. “Does Employment Affect Health?” Journal of Health and Social Behavior 36(3):230–43.

43.

Rosta

Judith

Aasland

Olaf G.

2011. “Work Hours and Self-rated Health of Hospital Doctors in Norway and Germany: A Comparative Study on National Samples.” BMC Health Services Research 11(1):40.

44.

Schunck

Reinhard

Rogge

Benedikt G.

2012. “No Causal Effect of Unemployment on Smoking? A German Panel Study.” International Journal of Public Health 57(6):867–74.

45.

Siegrist

Johannes

. 1996. “Adverse Health Effects of High-effort/Low-reward Conditions.” Journal of Occupational Health Psychology 1(1):27–41.

46.

Sparks

Kate

Cooper

Cary

Fried

Yitzhak

Shirom

Arie

. 1997. “The Effects of Hours on Health: A Meta-analytic Review.” Journal of Occupational and Organizational Psychology 70(4):391–408.

47.

Stewart

Jennifer M.

2001. “The Impact of Health Status on the Duration of Unemployment Spells and the Implications for Studies of the Impact of Unemployment on Health Status.” Journal of Health Economics 20(5):781–96.

48.

Stier

Haya

Lewin-Epstein

Noah

. 2003. “Time to Work: A Comparative Analysis of Preferences for Working Hours.” Work and Occupations 30(3):302–26.

49.

Trinkoff

Alison M.

Rong

Geiger-Brown

Jeanne

Lipscomb

Jane

Lang

Gary

. 2006. “Longitudinal Relationship of Work Hours, Mandatory Overtime, and On-call to Musculoskeletal Problems in Nurses.” American Journal of Industrial Medicine 49(11):964–71.

50.

Tucker

Philip

Rutherford

Celia

. 2005. “Moderators of the Relationship between Long Work Hours and Health.” Journal of Occupational Health Psychology 10(4):465–76.

51.

Virtanen

Marianna

Stansfeld

Stephen A.

Fuhrer

Rebecca

Ferrie

Jane E.

Kivimäki

Mika

. 2012. “Overtime Work as a Predictor of Major Depressive Episode: A 5-year Follow-up of the Whitehall II Study.” PLoS ONE 7(1):e30719.

52.

Wadsworth

M. E. J.

Montgomery

S. M.

Bartley

M. J.

1999. “The Persisting Effect of Unemployment on Health and Social Well-being in Men Early in Working Life.” Social Science & Medicine 48(10):1491–99.

53.

Wagner

Gert G.

Frick

Joachim R.

Schupp

Jürgen

. 2007. “The German Socio-economic Panel Study (SOEP): Scope, Evolution and Enhancement.” Schmollers Jahrbuch: Journal of Applied Social Science Studies/Zeitschrift für Wirtschafts und Sozialwissenschaften 127:139–69.

54.

Walwei

Ulrich

. 2014. “Times of Change: What Drives the Growth of Work Arrangements in Germany?” Journal for Labour Market Research 47(3):183–204.

55.

Ware

John E.

Kosinski

Mark

Keller

Susan D.

1996. “A 12-item Short-form Health Survey: Construction of Scales and Preliminary Tests of Reliability and Validity.” Medical Care 34(3):220–33.

56.

Williams

Joan C.

2001. Unbending Gender: Why Family and Work Conflict and What to Do about It. Oxford, UK: Oxford University Press.

57.

Wirtz

Anna

Nachreiner

Friedhelm

. 2010. “The Effects of Extended Working Hours on Health and Social Well-being: A Comparative Analysis of Four Independent Samples.” Chronobiology International 27(5):1124–34.