@prefix vivo: . @prefix edm: . @prefix ns0: . @prefix dcterms: . @prefix dc: . @prefix skos: . vivo:departmentOrSchool "Arts, Faculty of"@en, "Social Work, School of"@en ; edm:dataProvider "DSpace"@en ; ns0:degreeCampus "UBCV"@en ; dcterms:creator "Tang, Chen Yu"@en ; dcterms:issued "2009-03-04T21:44:25Z"@en, "2008"@en ; vivo:relatedDegree "Master of Arts - MA"@en ; ns0:degreeGrantor "University of British Columbia"@en ; dcterms:description """I examine the effects of two life stages --- the preschool stage (defined as having preschool children in the house) and the school-age stage (defined as having school-age children in the house) --- on the division of household labour in Sweden (N = 480), Germany (N = 689), and the United States (N = 465), using the 2002 International Social Survey Programme. I also examine time availability, relative resource, and gender ideology as the mediating factors in the relationship between life stages and the division of household labour. Initial analysis finds no variation in spouses’ relative frequency of housework participation when examined by the presence of children of either age group. Post-hoc analysis via separate examination of men’s and women’s reported housework hours finds stability in men’s number of hours of housework regardless of life stages across the three countries but variation in Swedish and U.S. women’s housework time across life stages and between countries. The results suggest that women spend more time in housework than men, regardless of life stages and country differences. However, while men’s time spent in housework is unaffected by the presence of preschool or school-age children, women’s time spent in housework may increase depending on the presence of children of these two age groups as well as the type of welfare regime of the country they live in. Neither in the initial nor in the post-hoc analysis are any of the proposed mediation effects established. Nonetheless, time, resource and gender ideology are shown to have differing predictive power on the division of household labour within and across countries."""@en ; edm:aggregatedCHO "https://circle.library.ubc.ca/rest/handle/2429/5524?expand=metadata"@en ; dcterms:extent "1128581 bytes"@en ; dc:format "application/pdf"@en ; skos:note "A COMPARATIVE STUDY OF DIVISION OF HOUSEHOLD LABOUR ACROSS FAMILY LIFE STAGES IN SWEDEN, GERMANY, AND THE UNITED STATES by Chen Yu Tang M.A., University of British Columbia, 2008 A THESIS SUMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS in The Faculty of Graduate Studies (Family Studies) THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) August 2008 ©ChenYuTang,2008 ABSTRACT I examine the effects of two life stages --- the preschool stage (defined as having preschool children in the house) and the school-age stage (defined as having school-age children in the house) --- on the division of household labour in Sweden (N = 480), Germany (N = 689), and the United States (N = 465), using the 2002 International Social Survey Programme. I also examine time availability, relative resource, and gender ideology as the mediating factors in the relationship between life stages and the division of household labour. Initial analysis finds no variation in spouses’ relative frequency of housework participation when examined by the presence of children of either age group. Post-hoc analysis via separate examination of men’s and women’s reported housework hours finds stability in men’s number of hours of housework regardless of life stages across the three countries but variation in Swedish and U.S. women’s housework time across life stages and between countries. The results suggest that women spend more time in housework than men, regardless of life stages and country differences. However, while men’s time spent in housework is unaffected by the presence of preschool or school-age children, women’s time spent in housework may increase depending on the presence of children of these two age groups as well as the type of welfare regime of the country they live in. Neither in the initial nor in the post-hoc analysis are any of the proposed mediation effects established. Nonetheless, time, resource and gender ideology are shown to have differing predictive power on the division of household labour within and across countries. TABLE OF CONTENTS Abstract ii Table of Contents iii List of Tables v List of Figures vi Chapter 1 1 Introduction 1 Chapter 2 7 Literature Review 7 What is Household Labour7 7 The Family Life Stages 9 Welfare Regimes in Sweden, Germany, and the United States 12 Time Availability, Relative Resource, and Gender Ideology 22 Chapter 3 26 Method 26 Sample 26 Measures 27 Dependent Variable: Division of Household Labour 27 Independent Variable: Family Life Stages 28 Mediating Variables: The Three Alternative Explanations 28 Time Availability 29 Relative Resources 29 III Gender Ideology .30 Controls 30 Chapter 4 34 Results 34 Descriptive Results 34 Test of Hypotheses 40 Post-Hoc Analysis 48 Descriptive Results 48 Test of Hypotheses 52 Chapter 5 55 Discussion 55 Findings Related to the First Two Hypotheses 56 Findings Related to the Third Hypothesis 61 Other Findings --- The alternative Explanations and Context 62 Limitations 68 Future Research 70 Conclusion 71 References 74 Appendices 80 Appendix A 80 Appendix B 81 Appendix C 84 iv LIST OF TABLES Table 1 Childcare, public parental leave, and social policies 16 Table 2 Female employment rate 18 Table 3 Description of mediating and control variables 31 Table 4 Descriptive statistics 39 Table 5 Regression results of the division of household labour on life stages 41 Table 6 Regression results of the mediation pathway “c’ “: the relationships between life stages and the division of household labour while accounting for mediators, and the pathway “b”: the relationships between mediators and the division of household labour 46 Table 7 Regression results of the mediation pathway “a”: the relationships between life stages and mediators 47 Table 8 Regression results of housework hours on life stages by gender 54 V LIST OF FIGURES Figure 1 The hypothesized relationship between life stages and the division of household labour 11 Figure 2 The hypothesized mediation relationship 24 Figure 3 Mean division of household labour score by gender in two life stages 33 Figure 4 Mean division of household labour along the life cycle 36 Figure 5 Mediation test illustration 43 Figure 6 Mean reported housework hours per week by gender along the life cycle 51 vi CHAPTER 1 Introduction Existing literature suggests that the division of household labour is related to family life stages, especially to the presence and age of children in the household. For example, although women are found to shoulder the majority of housework across life stages, the gender gap of housework contribution is the largest in childrearing years and the smallest in preparental and postparental stages (Rexroat & Shehan, 1987). Transition to parenthood typically has been found as the mark-off point of a time period in which housework allocation is more traditional and task-segregated (Rexroat & Shehan, 1987; Sanchez & Thomson, 1997; Shelton, 1992). On the one hand, this means that women spend more time doing housework during childrearing years, and that they do more when young children, especially preschool children, are present in the household than they do with older children (Brines, 1993; Presser, 1994; Shelton & John, 1993; South & Spitze, 1994; Waite & Goldscheider, 1992). On the other hand, this also means that women are mainly responsible for what is called “women’s work” or “female-typed” tasks, such as cleaning and cooking (Blair & Lichter, 1991; Orbuch & Eyster, 1997; Presser, 1994). These tasks are more mundane, routine, frequent, and less flexible than the tasks men are responsible for, such as mowing the lawn and doing household repairs. Compared to their female counterparts, men’s housework contribution declines proportionally upon entrance into parenthood, and their relative share of housework also decreases as the number of children in the family increases (Presser, 1994). This does not mean that they do less and their wives do more. Rather, wives’ time spent in housework increases more than does husbands’. Thus, in order to 1 understand how domestic tasks are arranged, it is essential to take into account the effects of various stages in the family life. The pattern of change in the division of domestic tasks, nevertheless, has been suggested to be an indirect outcome of family life stages. A couple of studies have tested three alternative explanations time availability, relative resource, and gender role ideology as rationales for the variation in men’s and women’s performance of household tasks along family life-course (Coltrane, 1992; Rexroat & Shehan, 1987). Simply put, the time availability explanation suggests that time spent on housework is negatively related to time spent on other competing demands (Coverman, 1985; England & Farkas, 1986; Hiller, 1984). Thus, the most readily used indicator in existing studies is time spent in paid jobs (i.e., Blair & Litcher, 1991; Bianchi, Milkie, Sayer, & Robinson, 2000; Nichols & Metzen, 1982; Pina & Bengston, 1987; Presser, 1994; Shelton, 1990). This explanation implies that women tend to perform more housework than men in childrearing years because they are less likely during this time period to be employed fuiltime and/or to have a paid job. In addition, the relative resource explanation assumes that housework is undesirable and people are, therefore, motivated to reduce it through negotiating their way out of housework via personal resources such as education, occupational prestige, and earnings (Blood & Wolfe, 1960; Brines, 1993). According to this explanation, whoever has the most resources has the most negotiating power. Since men are more likely to continue receiving job promotions and higher pay than women throughout childrearing years, they are more likely to successfully negotiate out of housework. It is, therefore, not surprising to find women shouldering more housework responsibilities with children in the household. Lastly, the gender ideology explanation suggests that the presence of more egalitarian gender role attitudes will result in a more equal division of domestic labour, while more traditional ones will result in a 2 less equal division (Brayfield, 1992; Huber & Spitze, 1983). Because both women and men tend to adhere to more traditional gender roles in parenthood (Nock, 1979; Roxroat & Shehan, 1987) as opposed to pre-parenthood, the division of household labour is less equal in this stage in life, with women being the primary caretaker and household manager while men being the primary breadwinner. However, while housework allocation has been found in the United States to vary along the life-course and such variations are suggested to be mediated by the time, the resources, and the gender ideology of the individuals involved, there are only a small number of studies that investigate the division of labour cross-nationally and a handful of studies that look at the relationship between life stages and the division of labour across countries. The suggestion that the three alternative explanations operate as mediating factors in such a relationship, nonetheless, has never been empirically examined. What we do know from the limited amount of cross-national research on housework arrangement, nevertheless, is that, like in the United States, the three alternative explanations have been found to affect the division of household labour in Sweden. For example, Calasanti and Bailey (1991) find that factors such as women’s high occupational prestige and men’s liberal attitudes and fewer working hours predict increased involvement of Swedish men in house cleaning, although the same factors are more powerful predictors for increased involvement among U.S. men. Also, Evertsson and Nermo (2007) find an indirect relationship between increased relative resource of Swedish women and their decreased participation in domestic labour. Despite the negative effect of relative resource on housework involvement as reported in these studies, Swedish women still shoulder the majority of the housework even when their 3 resources are equal to their husbands’. Furthermore, countries’ family and work policies affect how women balance domestic and market work (Kalleberg & Rosenfeld, 1990). For instance, Swedish woman many choose to trade their market involvement for fulfilling family responsibility because of greater opportunities for part-time employment and more flexible work arrangements. In addition, some literature supports the alternative explanations in the case of housework arrangements in Germany. For example, the division of labour is more likely to be equally shared when the wife has a better education, irrespective of the husband’s level of education (Lewin-Epstein & Stier, 2006). Women who spend longer hours than the normative time spent in the paid work decrease their housework independent of their spouse’s labour market situation. These findings not only lend support to the explanations of resource and time but also reveal the gender differences in the effects of these factors on housework participation. Moreover, women in Germany, resembling the findings from other studies regarding Sweden and the United States, also perform most of the housework even if they are economically independent. Lewin-Epstein and Stier also find that the presence of preschool children increases the time spent on housework by both spouses while not affecting the division of household chores; however, the presence of •school-age children does not result in more housework for either spouse. Cook (2004a), focusing on only fathers’ housework participation, links it to odds of second birth. In U.S. studies researchers have looked at how family composition, such as the age and presence of children, may affect the division of household labour. Cook’s study, however, suggests that German fathers’ division of household labour may affect family compositions, such as the odds of the family having a second child. 4 Due to a lack of cross-country comparisons in the relationship between life stages and the division of labour, this study affords an opportunity to explore the presence, direction, and rationale of such a relationship. The purpose of this study is three-fold. The first objective is to examine whether the variation in division of household labour depending on life stages is context-specific or cross-contextually established. The second objective is to investigate how life stages affect division of labour differently (or in similar ways) both within and across countries. The third objective is to test resource, time, and gender ideology as the mechanisms behind the relationship between life stages and housework sharing. These three objectives potentiate three possible kinds of contributions that the present study may offer to the body of literature on cross- national research of division of household labour. First, it may demonstrate the conceptual and empirical importance of the life-course approach to the study of division of labour across contexts. Second, it may explore the applicability of the three alternative theoretical frameworks in cross-national research of division of labour along the life-course. Third, it may illustrate the usefulness of certain context-specific characteristics as explanations for cross-country differences in the relationship between life stages and housework sharing. Speaking of context-specific characteristics, the question at this point is what guidelines could be used in selecting countries for comparison? Existing cross-national studies that examine life stages and housework sharing have adapted Esping-Andersen’s (1990) welfare regime types as a tool to provide contextually different countries for comparison in studies of the division of household labour across life stages (i.e., Drobni & Treas, 2006; Geist, 2005). The current study, therefore, examines the relationship between family life stages and the division of household labour, and with respect to the previously mentioned alternative explanations as mediators in the relationship between family life stages and division of labour in Sweden, Germany, and the 5 United States. Because these countries represent different welfare regimes first posited by Esping-Andersen, they provide different contexts in which this study may examine its hypotheses. Before I expound on the research questions and hypotheses of this study, I first further discuss some important concepts: household labour, family life stages, welfare regimes, and the three alternative explanations. This is accompanied by detailed outlines of research questions and hypothesized findings. Finally, I present data results, discuss findings, and conclude my cross national investigation of division of household labour across life stages. 6 CHAPTER 2 Literature Review What is household labour? Household labour has been defined in a variety of ways. In most studies, the definition of what composes housework is inferred by the way it is measured operationally rather than explicitly described in its conceptualization (Shelton & John, 1996). Dishwashing, cooking, and cleaning are a few examples of common operational terms that are used to define and measure housework. In contrast, childcare is often not included unless explicitly stated otherwise. Researchers have also used the amount of time spent on unpaid activities in the home as an indicator for housework participation. Blair and Lichter (1991) point out that “[asj a concept, household labour has two analytically distinct dimensions: number of hours and types of work tasks assigned to each spouse” (p. 92). Nevertheless, rarely has any study distinguished between these two dimensions in its empirical investigation. As mentioned earlier, most studies fail to provide a clear discussion of what housework is and depend on the way the concept is measured to speak for itself. As studies rely on mere operational measures of household labour without clear guidance from a well-outlined conceptual definition, the hour and the task dimensions often pair up in measurements which do not distinguish between the meanings they may carry. As a result, conclusions drawn from such analyses inevitably inherit questionable inferences. For example, “inequity” over the division of household labour between married couples seems to be of particular interest to researchers (e.g. Baxter, 1997; Bird, 1999; DeMaris & 7 Longmore, 1996). However, does “equal” suggest that both spouses spend the same amount of time performing unpaid activities in the home, regardless of what kind of tasks are involved in these activities? Or does it suggest that there is no gender segregation in the type of household tasks shared between the spouses? The fact that a spouse spends more hours on housework does not necessarily mean that there is segregation in the types of housework (Blair & Litcher, 1994). In order to make clear what “equal” denotes in the current study, it is defined as spouses’ assumed equivalent amount of responsibility in female-typed domestic tasks. When division of household labour is referred to as equal, it means that men bear as much responsibility for these tasks as women. However, researchers have employed different measures in assessing spouses’ housework responsibility. Previous studies on life-course transitions and the division of labour mostly adopt a measure of absolute time spent on housework (Coltrane & Ishii-Kuntz, 1992; Evertsson & Nermo, 2007; Rexroat & Shehan, 1987). Baxter, Hewitt, and Haynes’ study (2008) supports the value of hours of housework measure when examining life-course transitions and men’s and women’s contribution to housework separately. Cross-national studies of division of labour, nevertheless, mostly adopt a relative frequency measure (i.e. how often one spouse performs certain household tasks relative to the other spouse) (Cook, 2006; Diefenbach, 2002). Geist (2005) argues that a relative frequency measure is more robust to cross-national variation in social desirability across welfare regimes that can affect the report of overall time spend on housework. 8 The Family Life Stages As mentioned, in studies using data collected in the United States, family life stages have empirically been found to relate to housework allocation. Although women do more than men in all life stages, the discrepancy is even more apparent when young children are present in the house as opposed to older children. Besides being an empirical tool, however, family life stages are also a conceptually valuable instrument for studying the family, including division of household labour. In this section of the literature review, I present a discussion of the concepts of family life stages and what stages are included in the present study. Family life stages are also referred to as in terms of the family life cycle. The concept comes from a family developmental framework in which families are not viewed as static entities but as changing and evolving over time (Nock, 1979). It focuses on progress and change rather than cross-sectional descriptions. This kind of perspective is effectual and valuable because it allows for examination of the family dynamics in various points of the development of a family life-course (Hill & Rogers, 1964; Nock, 1979). Nock (1979) suggests that certain life events, such as marriage, birth of children, and children leaving home, bear important meanings for studies of the family because they significantly alter family members’ relationships to one another. Thus, one may view particular issues and problems in a family stage as the product of that corresponding stage in the family life cycle. However, Nock (1979) also points out that using a family life cycle approach to study families may pose a problem: actual causes of change in any issue of interest are not easily determined since there are simply too many dimensions involved in the family life cycle that may exert an influence on the issue under investigation. For example, when one examines role 9 changes in the family life cycle, it is difficult to determine whether the changes are associated with the maturation of children, chronological age, length of marriage, presence of children, or something else. In the case of change in housework performance, although not all possible causes can be identified in the scope of the present study, previous studies on household division of labour have provided three well-outlined explanations: individuals’ available time, relative resource, and gender ideology. In one way, these explanations help narrow this study’s examination of the antecedents of differences in housework performances between life stages. In another way, Nock’ s comment indicates the importance of examining in a more detailed way whether all three alternative explanations play an equally significant role in the differences of housework arrangements in the family life cycle, when another dimension of this cycle welfare policies, which have been indirectly connected to the division of household labour through female labour force participation (Drobni & Treas, 2006) --- also differ depending on the context in which housework arrangements are examined. Therefore, the present study investigates whether time, resources, and gender ideology mediate with the same strength the effect of life stages on housework allocation in Sweden, Germany, and the United States. These mediating relationships are discussed more thoroughly in a later section with Research Question 3. Before moving onto the next section, it is important to note that this study compares and contrasts two family life stages --- families with preschool (younger) children and families with school-age (older) children. Though previous studies have chosen to cover a more extensive range of life stages such as the empty-nest and the retirement stages, they do not fit the purpose of the current study. As Rogers (1973) points out, while researchers may directly use previously developed sets of life stages, it is more profitable for them to modify the stages examined in 10 order to meet the needs of the specific issues of their studies. Since I am to investigate division of household labour among married couples in different family life stages in three kinds of welfare regimes, the welfare policies become the guidance in selecting what life stages to be examined in this study. Welfare policies are different among social-democratic, conservative, and liberal welfare regimes in the provisions they offer to families with preschool children versus with school-age children. Therefore, I decide to examine how housework is shared differently between the stage with preschool children and the stage with school-age children. The first research question of the current study is: Research Question 1 Does division ofhousehold labour vary in Sweden, Germany, and the United States depending on thefamily life stages? Life Stages Division of Household Labour Figure 1. The hypothesized relationship between life stages and the division of household labour. I hypothesize that division of housework varies along the life-course. Women in all three countries have been found to perform most of the housework (Cook, 2006b; Fuwa, 2004; Geist, 2005). This may imply that regardless of country differences women are more responsive than men to domestic affairs and childcare responsibilities. In that case, women may spend more time doing housework in households with younger children as they creat more work to be done than 11 in households with older children. Therefore, I believe that the gap between a husband’s and a wife’s housework contribution is different depending on the life stages. Welfare regimes in Sweden, Germany, and the United States This study compares Sweden, Germany, and the United States because each of these countries represents a different welfare regime according to Esping-Andersen (1990). Thus, they provide different contexts for the comparison of division of household labour in the family life cycle mediated by the three alternative explanations. In this section I discuss the differences of these welfare regimes and their implications for the division of household labour. According to Esping-Andersen’ s (1990) typology, there are three models of welfare regimes --- social-democratic, conservative, and liberal --- that describe the kinds of relationships between the state, the market, and the family. Differences in welfare regime types not only entail structural and socio-political distinctions in welfare policies but also cultural distinctions. In fact, Pfau-Effinger (2004) argues that welfare policy and culture go hand-in-hand. Welfare policies are embedded in and justified by cultural values; cultural values are also magnified in welfare policies. She also argues that Esping-Andersen’s (1990) welfare regime typology already implicitly includes culture. Based on their structural, socio-political, and cultural differences, countries representing Esping-Andersen’ s welfare regimes have varied patterns of division of household labour corresponding to their regime type (Drobni & Treas, 2006; Geist, 2005). For instance, although there is persistent gender inequality in division of household labour regardless of welfare regimes, conservative countries have lower levels of equal sharing, regardless of time, resource, and gender ideology (Geist, 2005). In contrast, social-democratic countries have higher levels of equal sharing in housework. While less traditional gender roles are associated with 12 more egalitarian housework sharing in liberal and social-democratic regimes, the same association is not observed in a conservative regime by Geist. Furthermore, while men’s increased time spent at work is related to lower levels of egalitarian work sharing in social- democratic regimes, this is the case to a much greater extent in conservative countries. In other words, even though the existence of a certain relationship between the alternative explanations and division of household work may be present in one type of regime, it may not be present in another. Furthermore, even if this relationship appears to be present in both countries, it may not be of the same strength in its association. The question of the current study is then, whether there are also differences in the variation of housework arrangements by life stages between Sweden, Germany, and the United States, each representing a welfare regime by Esping-Andersen (1990). To help answer this question, the discussion below covers various differences among the three countries in their welfare systems and the implications of these differences. Sweden, a representative of the social-democratic welfare regime, views the state as part of the society, and thus welfare policies are made so as to help alleviate family responsibilities via governmental provisions (Sainsbury, 1996). Individuals are afforded certain benefits regardless of market position (Esping-Andersen, 1999). A social-democratic welfare regime also promotes gender equality; female employment is promoted as well as sharing between men and women of domestic work and childcare responsibilities (Geist, 2005; Haas, 1981). Germany, on the other hand, is a conservative welfare regime. Its policies are family oriented in the sense that women and mothers are encouraged to make family and domestic responsibilities their priority (Esping-Andersen, 1990, 1999; Mosesdottir, 2000). Employment 13 participation is, therefore, discouraged for women whereas a breadwinner role is promoted for men. On the one hand, families are viewed as self-sufficient entities. Family members are expected to look among themselves and to find assistance and provision to meet their needs (Brochorst, 1994); the government only intervenes when families cannot take care of themselves. On the other hand, a conservative regime has more well-established welfares programs compared to a liberal welfare regime (Orloff, 1996). Nevertheless, welfare policies are designed to maintain social and class status, including gender stratification, instead of promoting gender equality (Geist, 2005; Orloff, 1993). While Sweden and Germany represent the social-democratic and the conservative regime respectively, the United States represents a liberal welfare regime, in which individual responsibility and freedom are emphasized (Esping-Andersen, 1990). This regime supports the idea of gender equality yet does not actively promote an equal or gendered division of household labour (Geist, 2005). Although the United States is also characterized by a high rate of female employment, unlike Sweden it does not make policies to stimulate the female employment rate. Rather, women’s labour force participation is driven by the orientation of the market. The intrinsic differences of welfare regimes among Sweden, Germany, and the United States are practically reflected in the way their child care policies, parental leave policies, and school policies are made. (Refer to table 1, for a summary of these countries’ welfare policies.) For example, with regards to childcare policies, Sweden guarantees childcare coverage almost inclusively from when the child is born until five years of age (except the six-month period between when the child is eighteen months old until when he/she reaches three years of age), whereas there is virtually no childcare guarantees in either Germany or the United States for 14 children of the same age range. Furthermore, a good percentage (79%) of children in Sweden between three to seven years of age are in publicly funded childcare services. Even though the percentages of children younger and older than that age bracket enrolled in public childcare are smaller (32% for age zero to two; 34% for school-age), relatively speaking they are still much higher than those of children in Germany (2% for age zero to two; 4% for school-age) and the United States (1% for age zero to two; <1% for school-age). With regards to public parental leave policies, Sweden has not only legislated job protection and universal coverage for all employed women but also prolonged paid maternity leave of fifty-two weeks and twenty-six weeks of extended leave. On top of that, there are also parental leave benefits for fathers. Compared to Sweden, Germany provides less but still moderate levels of support through its parental leave policies. For instance, like Sweden, there are legislated job protection, universal coverage for all employed women, and paternity leave benefits. However, paid maternity leave is much shorter, as fourteen weeks is approximately a quarter of the paid leave in Sweden and extended leave is non-existent. Among the three countries examined in this study, the United States has the poorest parental leave policies with no legislated job protection, extended leave, or paternity leave benefits. Though paid maternity leave is available, it is only a meager six weeks and covers only twenty-five percent of employed women. 15 Table 1. Childcare, public parental leave, and school policies in Sweden, Germany and the United States (adapted from Gornick, Meyers, and Ross (1997)). Sweden Germany United States CHILDCARE POLICIES: Guaranteed child care >18 mo. None None coverage (0-2) Age- years Guaranteed child care 3-5 None None coverage (3-5) Age- years % Children (0-2) in 32 2 1 publicly funded child care % Children (3-school- 79 78 14 age) in publicly funded child care % Children in 34 4 <1 publicly funded after- school care PUBLIC PARENTAL LEAVE POLICIES: Legislated job Yes Yes No protection Paid maternity leave 52 14 6 (weeks) Coverage(% 100 100 25 Employed women) Extended leave 26 0 0 (weeks) Paternity leave Yes Yes No benefits SCHOOL POLICIES: Age of compulsory 7 6 6 school (Age-years) School-hours (Hours 30 25 33 per week) School-year (Days per 190 213 185 year) Continuous school- Yes No Yes day 16 Table 1. Childcare, public parental leave, and school policies in Sweden, Germany and the United States (adapted from Gornick, Meyers, and Ross (1997)). POLICY INDEX VALUES: Policies for mothers 61.9 34.1 17.1 with children <6 Policies formothers 55.5 32.6 57.0 with school-age children Furthermore, Eurostats (2007) shows that the female employment rate of the three countries examined in this study from 1996 to 2007. As presented in table 2, Sweden has the highest rate of female labour force participation, the United States has the second highest rate, and Germany has the lowest rate throughout the twelve years. Consideration of these statistics in conjunction with the coverage of employed women for public parental leave policies in the three countries reveals that because of Sweden’s universal coverage, Sweden also has the highest rate of women with access to its generous public parental leave benefits. Although Germany has the lowest rate of female employment among the three countries examined, because of its universal coverage it probably has the second highest rate of women with access to its moderate level of public parental leave benefits. The United States, on the contrary, probably has the lowest rate of women with access to its limited public parental leave benefits due to its low coverage. 17 Table 2. Female employment rate in Sweden, Germany, and the United States from 1996 to 2007 (Eurostats, 2007). 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 Sweden 68.1 67.2 67.9 69.4 70.9 72.3 72.2 71.5 70.5 70.4 70.7 71.8 Germany 55.3 55.3 55.8 57.4 58.1 58.7 58.9 58.9 59.2 60.6 62.2 64.0 U.S. 66.2 67.1 67.4 67.6 67.7 67.0 66.0 65.6 65.3 65.6 66.0 N/A Note: The female employment rate is calculated by dividing the number of women aged 15 to 64 in employment by the total female population of the same age group. The indicator is based on the EU Labour Force Survey. The survey covers the entire population living in private households and excludes those in collective households such as boarding houses, halls of residence and hospitals. Employed population consists of those persons who during the reference week did any work for pay or profit for at least one hour, or were not working but had jobs from which they were temporarily absent. Besides childcare and parental leave policies, school policies also show how welfare regimes represented by Sweden, Germany, and the United States differ from each other. School policies in both Sweden and the United States seem to work well with the schedule of a working parent: longer school hours per week and continuous school-days. These policies enable schooling to provide another form of childcare: keeping children at school while parents are free to go off to work if they wish. Note that Sweden’s age of compulsory school is seven years of age, one year later than Germany and the United States. At the same time, however, this probably does not affect parental employment as much because preschool children are well provided with childcare services. Germany, on the one hand, seems to have certain parental employment-supportive policies such as twenty-five hours of school per week and two hundred and thirteen days of school per year. On the other hand, its discontinuous school-day policy has been found to be very disruptive to parents’ employment (Gomick, Meyer, & Ross, 1997). In fact, schools in Germany routinely have a day off in the middle of the week, which causes 18 parents to take time off from work to care for their children. This inevitably discourages full-time employment, or even employment altogether, for women with school-age children. All in all, Sweden’s socio-democratic welfare regime provides the best support and help to families not only with preschool children but also with school-age children. Its childcare, parental leave, and school policies encourage mothers to remain in the paid labour force. Indeed, Stier and Lewin-Epstein (2001) found that employment continuity is the highest among countries providing support for working mothers. Also, paternity leave benefits encourage fathers to share childcare responsibilities, and probably domestic work as well. Germany’s welfare policies, however, confirm that the ideology behind the conservative welfare regime is one that encourages a domestic role for women. A lack of childcare policies accompanied by relatively well-established maternity leave policies and discontinuous school days are a demonstration of the regime’s belief that childcare and domestic work are not to be outsourced. Rather, they ought to be taken care of by women as domestic work, not paid employment, is women’s primary responsibility. The United States has the poorest welfare policy among the three countries examined, but this is no surprise for a liberal welfare regime. In the United States, there is virtually no childcare support and very limited, recipient-targeted maternal leave policies. The only aspect of its policies that may work towards the benefit of working mothers is school schedules. With respect to the impact of policies on family, I would also like to discuss one other point not covered in table 1 but important to the ramifications of regimes’ policies: namely, the matter of self-employment. The concern is that self-employed people would not have access to social policies. With little to no access to social policies, the effect of welfare policies is rather 19 diminished. Thus, based on the welfare policies of the country they are in, a sample with a considerable proportion of self-employed respondents probably should not be interpreted in the same way as a sample without a large proportion of its respondents in self-employment. I would, therefore, like to address the number of self-employed respondents and spouses and whether or not to control for self-employment in the method section. Based on the above review, the three countries’ welfare regimes and policies suggest implications as to how they could affect differences of housework arrangements in households with preschool children versus school-age children. This brings us to the second research question: Research Question 2 How does division ofhousehold labour across life stages differ among Sweden, Germany, and the United States? Because of differences in welfare policies between Sweden, Germany, and the United States, I examine whether these countries differ in division of housework across life stages. Firstly, I investigate whether having children, either children of preschool age or school age, exerts the smallest effect on housework sharing in Sweden, compared to the other two countries. Since in both life stages examined in this study the Swedish government is the most active in providing help to families and support to female employment, does this mean that neither preschool nor school-age children impact division of labour as much as they would in the other two countries? As reviewed earlier, a higher employment rate is related to less time for housework, more resources to negotiate one’s way out of work, and a more egalitarian gender 20 role attitude. Thus, if women are encouraged and enabled to remain in the paid labour force throughout their childrearing years, one might find that typically unequal housework arrangements would be more balanced with men and women sharing more equally throughout life stages. Secondly, I investigate whether the division of housework in Germany becomes more traditional and unequal, with women doing more as families proceed through life stages, relative to Sweden and the United States. This is because the German government provides little childcare support for preschool children and even less support for school-age children accompanied by a school schedule that disrupts parents’ work schedule. Thus mothers are discouraged from paid employment or at least full-time employment and encouraged to take care of the household. As female employment opportunities decline, women have more time to spend in the domestic sphere, less resources to negotiate their way out of housework, and may have a more traditional gender role attitude. Division of housework, therefore, may be more unequal as families proceed in their family life cycle. Thirdly, I investigate whether the U.S. housework arrangement is the opposite from that of the Germany sample. That is, housework sharing is more equal with school-age children in the household than with preschool children in the household. This is because the United States, though passive and lacking in providing typical public childcare facilities, has a school schedule that works in cooperation with parents’ working schedule. Therefore, mothers have more opportunities to rejoin the paid labour force or even obtain a full-time job when children are older. As a result, according to the relationship between the three alternative explanations and 21 housework, housework sharing might be more equal with older children present in the house than with younger children. To sum up the hypotheses for research question 2, I investigate three possible differences between the three countries in the division of housework through life stages. First, life stages have the smallest effect on the division of household labour in Sweden compared to Germany and the United States. Second, the school-age stage has more effect on the division of housework than the preschool stage in Germany compared to the other two countries in that it is shifted towards a more traditional and unequal arrangement. Third, having preschool children in the household makes the division of housework more traditional and unequal than having school-age children in the United States compared to the other two countries. Time availability, relative resources, and gender ideology Time availability, relative resource, and gender ideology are the three alternative explanations accounting for the change in the division of household labour across the life cycle. As stated earlier, time availability refers to the time left for housework after its competing tasks’ occupancy of their time share (Coverman, 1985; England & Farkas, 1986; Hiller, 1984). For example, those who work more in paid jobs are assumed to have less time available to spend on housework. This theoretical perspective assumes that people prioritize paid employment over housework. It also assumes gender-neutrality for housework allocation (Ishii-Kuntz & Coltrane, 1992); various household tasks are assigned to individuals who spend less time in paid work regardless of the nature of these tasks (i.e. male-typed or female-typed). It views unequal division of housework as the product of women having much more time outside of the labour market than men. When the gap between spouses’ work hours narrows, the time availability 22 explanation predicts a more egalitarian domestic work sharing. Thus in a more subtle manner it also assumes people’s willingness to do housework once they have time left outside of a paid job. It disregards the effect of personal preferences and cultural norms may have on housework arrangements. Relative resource refers to the resources one possesses to successfully negotiate his or her way out of housework (Blood & Wolfe, 1960; Brines, 1993). For instance, people with higher incomes, relative to their spouses, have more resources to “buy” their way out of housework. This perspective assumes that domestic tasks are undesirable, and, therefore, people try to decrease their share of them with resources (Brines, 1993). Under this framework, certain assets, such as education, occupational prestige, and income, are more valuable bargaining chips than others, such as professional expertise and personality, in the negotiation of housework arrangement. Unequal division of household labour is perceived to be the result of a discrepancy between husband’s and wife’s resources. With women’s increased resources, such as better education and higher income, relative to their husband’s, the relative resource explanation would predict a more equal share in domestic labour. However, by applying objective and outward standards of measurement in defining resources, it ignores the intrinsic meaning and value domestic work may possess. Gender ideology refers to the belief of gender equality. Whereas a more traditional gender role attitude would result in a less equal share of housework, a more egalitarian one would result in a more equal share (Brayfield, 1992; Huber & Spitze, 1983). This perspective assumes that people conduct themselves according to their internalized sex-appropriate values and behavior with which they are socialized since youth (Ishii-Kuntz & Coltrane, 1992). It 23 regards housework as possessing gender-related meanings. For example, men with traditional gender role attitudes have lower housework participation than men with egalitarian gender role attitudes because they see housework as women’s responsibility. With its focus on the individual and psychological level, the gender ideology explanation ignores the effect of the greater surround environment on people’s behaviors. Since the three alternative explanations, as mentioned earlier, have received some support in their relation to division of household labour, the present study is interested in whether they affect housework cross-nationally across life stages, especially whether they would mediate between life stages and division of household labour. This leads to the third research question of the present study: Research Question 3 Do all three alternative explanations mediate in the relationship between life stages and division ofhousehold labour in Sweden, Germany, and the United States? Life Stages Division of Household Labour Time Availability Relative Resources Gender Ideology Figure 2. The hypothesized mediation relationship. 24 Since existing research on division of household labour has often utilized time, resource, and gender ideology as explanations for housework arrangements, either with a life-cycle perspective or in cross-national comparisons, the present study expects these three explanations to exert some mediating power between life stages and division of household labour in Sweden, Germany, and the United States. However, whether all three explanations act as mediators in all three countries under investigation will be examined in this study. Even if they all mediate the relationship between life stages and housework, it is anticipated that the three explanations would not all mediate with the same strength across countries. Existing research indicates that some of the alternative explanations may be stronger determining factors for housework than the others in one welfare regime versus the other types of regimes. For example, less traditional gender role attitudes do not seem to be as effective in predicting a more egalitarian division of housework in conservative countries as they do in social-democratic and liberal countries (Geist, 2005). This might be due to the strong societal endorsement of the traditional gender role belief in conservative countries. As a result, even if individuals themselves hold an egalitarian gender role attitude, they would still subscribe to the socially acceptable behaviors, in which women’s priority is to attend the house while leaving men to be the breadwinners of the family. The present study does not go into details in investigating which mediating factor is the strongest or the weakest in which country. However, it is important to investigate and acknowledge the possibility that though time, resource, and gender ideology all mediate between the relationship of life stages and housework allocation, they do not all share equal importance in such a relationship across different welfare states. 25 CHAPTER 3 Method Sample Analysis of this study is based on data from the 2002 International Social Survey Programme (ISSP) concerning “Family and Changing Gender Roles” spanning 40 countries on different continents. However, the present study examines only three countries --- Sweden, Germany, and the United States --- because they respectively represent social-democratic, conservative, and liberal welfare regimes. The analysis includes married people only. It helps to exclude cohabiting and same-sex couples, who have been shown in the existing literature to exhibit different patterns of housework sharing from heterosexual married couples (Shelton & John, 1993, 1996; South & Spitze, 1994; Stafford, Backman, & Dibona, 1977). There are 480 respondents in the Sweden sample, 689 in the Germany sample, and 465 in the United States sample. Data from the Sweden sample is collected through separate postal survey with four reminders. The first two and the fourth are by mail while the third is by telephone. Data from the Germany sample is collected through self-completed questionnaires with an interviewer present while background variables are asked face-to-face. Data from the United States sample is collected through multi-state area probability sampling methods and in-person interviews coupled with self-administered questionnaires at the end of the interviews. 26 Measures Dependent Variable: Division ofHousehold Labour Division of household labour is the dependent variable. It is an index constructed on the basis of four items indicating who does (1) the laundry, (2) the caring for sick family members, (3) the household cleaning, and (4) meal preparations. Original ISSP answer categories for these items are “always me,” “usually me,” “about equally/both together,” “usually my spouse,” “always my spouse,” and “is done by a third person.” However, because the study is interested in how housework is divided between men and women, answer categories are genderized by assigning a value of 5 to “always me” and sequentially down to 1 for “always my spouse” when the respondent is female, and the reverse for a male respondent. In addition, in order to avoid the problem that missing responses may skew the index scores, the total score added up from a respondent’s response is divided by the number of questions he or she answers so that the index of division of labour ranges from 5 (indicating that women always do all the work) and 1 (indicating that men always do all the work). Since the index is meant to represent the division of household labour between spouses in their relative frequency of housework participation, the last response category from the original ISSP data set “is done by a third person” is excluded from the creation of the index. The item of doing “small repairs around the house” that is originally part of the housework questions in the ISSP is not included in the index because it is traditionally a male task while the focus of the study is to look at traditionally regarded female-typed tasks. The item “shopping for groceries” is not included either because it appears to be a neutral rather than a female-typed task (see Twiggs, McQuillan, and Ferree, 1999 for a discussion on female, male and neutral tasks). 27 Independent Variable: Family Life Stages Family life stage is the independent variable. There are two family life stages: with preschool/young children (less than 6 years of age for the Germany and the United States sample, and less than 7 years of age for the Sweden sample because of their different age for compulsory schooling) living in the house, and with school-age/older children (between 6 to 17 years of age for the Germany and the United States samples, and between 7-17 years of age for the Sweden sample) living in the house. Data for the categories with young and older children at home are obtained respectively from the items “number of people in household: kids younger than 5, 6 years old” and “number of people in the household: kids between 6, 7, to 17 years old.” Two dummy variables are created (the first one is 1=preschool, and O=otherwise; the second one is 1=school-age, and O=otherwise). The data sample includes only those who are married with children. Therefore, couples with preschool children are compared with those who do not have preschool children (those who have either school-age children or adult children, or both), while couples with school-age children are compared with those whodo not have school-age children (those who have preschoolers and/or adult children). Those who indicate “don’t know” and “no answer, refused,” are excluded from the study. Mediating Variables The mediating variables (table 2) cover the three alternative explanations in the relationship of life stages and division of housework. 28 Time availability. Time availability is measured by respondents’ and their spouses’ relative time available. This is done by subtracting wife’s number of hours worked weekly from husband’s number of hours worked weekly. If the respondent is male, the time availability measurement would be the respondent’s number of hours worked weekly subtracted by the spouse’s number of hours worked weekly, whereas if the respondent is female, the reverse subtraction is performed. In other words, I always subtract women’s number of hours worked weekly from men’s number of hours worked weekly. Scores range from -95 to 95. A positive score indicates that the husband works longer hours than the wife, while a negative one indicates that the wife works longer hours than the husband. The smaller the score the more it indicates the husband’s time available for housework, while the larger the score the more it indicates the wife’s time available for housework. Couple measure of relative time available is used instead of mere respondents’ time available because the outcome of housework performance is measured in couple’s relative contribution to it. The other variables involved in the study therefore are also examined in a couple measure as much as possible. Relative resource. Relative resource, by the same principle as time availability, is also a relative measure of the couple’s income. Data is obtained through the item “who makes the higher income.” Responses are categorized by a 7-point scale: (1) My spouse has no income, (2)1 have a much higher income, (3) I have a higher income, (4) we have about the same income, (5) my spouse has a higher income, (6) My spouse has a much higher income, and (7) I have no income. Values of 1 to 7 are assigned respectively to the seven response categories for male respondents. 29 Reverse value assignment is performed for female respondents. Consequently, the score range of the relative resource variable is 1 (indicating men as the sole income source) and 7 (indicating women as the sole income source). Gender Ideology. Unfortunately, the 2000 ISSP data do not provide couple measure in gender ideology. Gender ideology therefore using an additive scale based on four items. Each single item may score from 1 (strongly agree) to 5 (strongly disagree). The score of the last item “men ought to do more housework than they do now” is reversed. In order to avoid inaccurately skewed scores by missing responses, respondents’ total score is divided by the number of question they responded. The final scale thus ranges from 1 to 5. The higher the score the more egalitarian the gender role ideology is, whereas the lower the score the more traditional the gender role ideology is. Controls Control variables (table 3) include respondents’ educational level, spouses’ educational level, and age. These factors have been shown to affect housework arrangement (Batalove & Cohen, 2000; Blair & Lichter, 1991; Presser, 1994; Shelton & John, 1996; South & Spitze, 1994). In addition, this study also controls for respondents’ gender (for a discussion on wives’ and husbands’ housework reporting, see Press & Townsley, 1998) and self-employment. 30 Table 3. Description of mediating and control variables. Variables Description Time availability Relative time available Men’s number of hours worked weekly subtracted by women’s number of hours worked weekly Relative resources Relative income The item: Who makes more income? Gender Ideology Gender role attitudes Additive scale based on four items: Do you agree or disagree... . A preschool child is likely to suffer if his or her mother works • Family life suffers if the woman works • Men’s job is work; women’s job is to take care of the household . Men ought to do larger share of housework than they do now Controls Gender of respondent 0, male; 1, female Respondents’ college education 0, no college/university education; 1, incomplete or completed college/university education Spouses’ college education 0, no college/university education; 1, incomplete or completed college/university education Age Respondent’s age in years Respondents’ self employment 0, not self-employed; 1, self-employed Mean responses on the division of household labour scale by gender and country are shown in figure 3 for the preschool and school-age stages respectively. Comparing men’s and 31 women’s responses suggests that among both men and women, German respondents report the most traditional sharing of housework responsibility, whereas Swedish respondents report the most nontraditional. However, not all gender differences in reporting are statistically significant. These differences are significant in the United States only for the preschool stage (p 0.007). In the school-age stage, on the other hand, these differences are significant in both the United States and Germany (p < 0.00 1 for both the United States and Germany). Where there is a significant difference, men report a more nontraditional housework arrangement than do the women. Also, where there is a significant difference, the gap in gender reports of who perform the tasks is the greatest in the United States. Because the data is not actually obtained from both spouses of the same couple, it is impossible to determine whether the differences in reporting are reflections of the actual gendered difference in division of household labour or gendered differences in perceptions. However, I choose to control for gender of respondents in hope of obtaining a more accurate measure of division of labour in homes by life stages in this study’s regression analyses. Furthermore, as discussed earlier, self-employed people may not have access to social benefits in the same way as people who are not self-employed. Therefore, data including large numbers of respondents with self-employment may limit the explaining power of welfare policies on division of household labour. In the 2002 ISSP employed in this study approximately eight to fifteen percent of respondents are self-employed in each life stage by country. Preliminary analysis reveals that by controlling for respondents’ self-employment, the explained variance is increased without much compromising of the generalizability of the sample to the population. Consequently, even though I do not have large numbers of self-employed individuals in my sample, I still choose to control for respondents’ self-employment. 32 A. Preschool Stage Sweden Germany U.S. B. School-age Stage 0 0 4.5 4 X Dmale : 3 femaIe f2.5 _________ 0 4.5 0” x, 4 ‘I- U o. Gmale C 3 femaIe ‘- 2.5 Co o : 2 I I Figure 3. Mean division of household labour score by gender in two life stages. Married persons with preschool or school-age children in the household. 5 = Wife always responsible, 1 = Husband always responsible. Sweden Germany U.S. 33 CHAPTER 4 Results Descriptive Results I begin my analysis with a description of countries’ division of household labour at two points in time: (1) preschool stage, and (2) school-age stage. Figure 4 presents the mean traditional division of household labour scores for the three countries included in this study. Within-country comparisons firstly reveal that in both Sweden and Germany, compared to people with children, people without children have higher division of household labour scores in both life stages (with mean scores of 3.63 and 3.71 for with preschool and school-age children respectively, and with mean scores of 3.78 and 3.80 for without preschool and school-age children respectively in Sweden; with mean scores of 4.09 and 4.11 for with preschool and school-age children respectively, and with mean scores of 4.18 and 4.20 for without preschool and school-age children respectively in Germany). This indicates that people without children in these two age groups report a more traditional housework arrangement than their counterparts with children in the corresponding age group. In the United States, however, this pattern is only true for the preschool stage (with mean scores of 3.67 for with preschoolers and 3.76 for without preschoolers) whereas in the school-age stage, people with children, instead of people without children, have higher division of household labour scores (with mean scores of 3.82 for with school-age children and 3.73 for without school-age children). This indicates that, U.S. respondents report a more traditional housework arrangement if they are without preschool children or if they are with school-age children, compared to their counterparts within the same 34 life stage. Despite these differences in mean division of household labour score within life stages in each country, none of them are statistically significant. Within-country comparisons, secondly, reveal that in all three countries it is those with school-age children that have higher division of household labour scores than those with preschool children (with mean scores of 3.71 and 3.63 for with school-age and with preschool children respectively in Sweden; with mean scores of 4.11 and 4.09 for with school-age and with preschool children respectively in Germany; with mean scores of 3.82 and 3.67 for with school- age and with preschool children respectively in the United States). This indicates that, regardless of country, people who have school-age children in the household report a more traditional housework arrangement compared to people who have preschool children in the household. However, in no country is the difference in mean division of household labour score between the people with preschool children and the people with school-age children statistically significant. Furthermore, across-country comparisons reveal that division of household labour in Germany is significantly more traditional than Sweden (p = 0.001 and p < 0.001 for the preschool and the school-age respectively) and the United States (p = 0.001 andp < 0.001 for the preschool and the school-age respectively). 35 I 0 .0 3.8C) 0CD ‘ 3.6 : 3.4 ________________ 3.2 C 0 :€ 3 to I.. Figure 4. Mean division of household labour score along the life cycle. Married persons with and without preschool and school-age children in the household. 5 = Wife always responsible, 1 = Husband always responsible. In addition to division of household labour, the descriptive statistics for the rest of the variables in relation to country and life stages are reported in table 4. In all three countries, women’s reported time spent in paid jobs is lower than their husbands’. As previous studies suggest for a conservative welfare regime, German wives spend the least amount of time relative to their husbands in paid employment among the three countries examined for both life stages. Between-country gap in this regard is the largest between Germany and Sweden, where Swedish women’s reported time spent in paid work relative to their husbands’ is twice as much as their German counterparts (with Swedish women working 8.35 hrs/wk less than their husbands while German women working 17.04 hrs/wk less than their husbands in the preschool stage; with Sweden women working 5.10 hrs/wk less than their husbands while German women working 13.32 hrs/wk less than their husbands in the school-age stage). However, these differences are only significant in the school-age stage (p < 0.001) but not in the preschool stage. Even though 0 Without preschool fl With preschool O Without school-age • With school-age Sweden Germany U.S. 36 women in all three countries apparently spend even less time in paid work relative to their husbands when they are with preschool children than when they are with school-age children (with women working 8.35 hrs/wk, 17.04 hrs/wk, and 16.20 hrs/wk less than their husbands for Sweden, Germany and the United States respectively in the preschool stage; with women working 5.10 hrs/wk, 13.32 hrs/wk, and 10.43 hrs/wk less than their husbands for these three countries in the school-age stage), in no country are these differences across life stages in time availability scores statistically significant. Relative resource between husband and wife is more equal (in the sense that the spouses make about the same amount of income) in both life stages in the United States than it is in Sweden or Germany, even though in both the United States and Germany the husband makes more than the wife does whereas the opposite is true in Sweden. In the preschool stage, women’s income relative to their husbands’ in Germany is significantly lower than that in Sweden (p < 0.001) and in the United States (p = 0.024), while in the school-age stage, women’s income relative to their husbands’ is significantly higher in Sweden than that in Germany (p <0.001) and in the United States (p = 0.0 12). Moreover, when comparing across life stages, relative resource between husband and wife is almost the same in both Sweden and the United States, but in Germany husbands’ and wives’ incomes are significantly more equal with school-age than with preschool children (p = 0.004). Comparisons of the gender ideology scale between countries suggest that gender role attitudes are the most egalitarian in Sweden and the most traditional in Germany. However, although between-country differences are significant between Sweden and Germany (p < 0.001) and Sweden and the United States (p = 0.032) in the preschool stage, they are only significant 37 between Sweden and Germany (p = 0.011) in the school-age stage. When comparing across life stages, differences in gender ideology are only significant in Sweden. That is, Sweden has significantly more traditional gender role attitudes in the school-age stage than in the preschool stage (p = 0.020). Furthermore, it seems that Sweden has the highest percentages of people with a college degree, whether they have preschool or school-age children, whereas Germany has the lowest percentages among the three countries examined. 38 Ta bl e 4. D es cr ip tiv e st at ist ic so fS w ed en ,G er m an y, a n d th e U ni te d St at es .M ar rie d pe rs on s w ith e ith er pr es ch oo lo r sc ho ol -a ge c hi ld re n in th e ho us eh ol d. M ea ns , p er ce nt ag es ,a n d st an da rd e rr o rs in pa re nt he se s. Pr es ch oo l Sc ho ol -a ge V ar ia bl es Sw ed en G er m an y U SA Sw ed en G er m an y U SA M ea ns D iv isi on o fh ou se ho ld la bo ur sc al e 3. 63 4. 09 3. 67 3. 71 4. 11 3. 82 (0 .57 ) (0. 59 ) (0. 58 ) (0. 51 ) (0. 57 ) (0. 67 ) Ti m e av ai la bl e 8. 35 17 .0 4 16 .2 0 5. 10 13 .3 2 10 .4 3 (10 .68 ) (16 .81 ) (16 .92 ) (11 .58 ) (17 .00 ) (14 .89 ) R el at iv e re so u rc e 3. 37 2. 16 2. 85 3. 38 2. 69 2. 88 (1. 26 ) (1. 06 ) (1. 55 ) (1. 23 ) (1. 23 ) (1. 50 ) G en de ri de ol og y sc al e 3. 98 3. 23 3. 50 3. 65 3. 37 3. 42 (0. 70 ) (0. 88 ) (0. 98 ) (0. 83 ) (0. 80 ) (0. 94 ) A ge 35 .6 3 32 .8 9 32 .5 9 45 .3 3 42 .0 3 42 .9 9 (6. 21 ) (4. 81 ) (7. 96 ) (6. 28 ) (6. 25 ) (7. 81 ) Pe rc en ta ge s Se x o fr es po nd en t - m al e 45 .7 0 45 .3 0 30 .4 0 43 .5 0 52 .2 0 38 .2 0 Se x o fr es po nd en t- fe m al e 54 .3 0 54 .7 0 69 .6 0 56 .5 0 47 .8 0 61 .8 0 W ith co lle ge 40 .0 0 13 .2 0 26 .8 0 27 .5 0 17 .2 0 23 .5 0 S- Pw ith co lle ge 31 .4 13 .2 0 30 .4 0 27 .5 0 16 .7 0 26 .5 0 Se lf- em pl oy ed 8. 60 7. 50 14 .3 0 10 .7 0 7. 20 14 .7 0 W ith ki ds 7. 29 7. 69 12 .0 4 27 .2 9 26 .1 2 21 .9 4 N 35 53 56 13 1 18 0 10 2 N ot e: D iv isi on o fh ou se ho ld la bo ur sc o re ra n ge s fro m 1, w he re th e hu sb an d is al w ay s re sp on sib le , t o 5, w he re th e w ife is al w ay s re sp on sib le .T im e av ai la bi lit y sc o re ra n ge s fro m - 95 ,h us ba nd ha vi ng th e m o st tim e av ai la bl e fo rh ou se w or k, to 95 ,w ife ha vi ng th e m o st tim e av ai la bl e fo rh ou se w or k. R el at iv e re so u rc e sc o re ra n ge s fro m 1, w he re th e hu sb an d is th e so le br ea dw in ne r, to 7, w he re th e w ife is th e so le br ea dw in ne r. Th e ge nd er id eo lo gy sc o re ra n ge s fro m 1, m o st tr ad iti on al ge nd er at tit ud es ,t o 5, m o st eg al ita ria n ge nd er at tit ud es . w cD Test ofHypotheses To this point, I have considered the differences in mean scores of division of household labour by life stages across countries. Such a perspective, nonetheless, is incomplete because it does take in to account of how some factors such as the respondents’ age and education level may affect the relationship between division of household labour and life stages. Therefore in the following section I examine the relationship between life stages and division of household labour in the three countries by regression, controlling for the gender, the age, and the self- employment status of the respondents and the educational level of the respondents and their spouses. Regression results predicting division of domestic responsibility by life stages are presented in table 5. In all three countries neither of the life stages examined has a significant association with division of household labour. Not only does the comparison between couples with preschool and without preschool children reveal no significant association with housework sharing (p = 0.928 for Sweden; p = 0.748 for Germany; p = 0.29 1 for the United States) but also the comparison between couples with and without school-age children (p = 0.580 for Sweden; p = 0.477 for Germany; p = 0.50 1 for the United States), while controlling for gender, age, respondents’ and their spouses’ college education, and self-employment of the respondents. These data results do not seem to support the idea that sharing in domestic tasks varies according to life stages as proposed in research question 1. For the second research question, three patterns of relationships between life stages and division of household labour either across countries are proposed for examination. The first is that both the preschool and school-age stages have the smallest effect on housework sharing in 40 Sweden, compared to the other two countries. The second is that division of labour in Germany is more traditional and unequal with school-age children than with preschool children, compared to the other two countries. The third is that division of labour in the United States is more egalitarian and equally shared with school-age children than with preschool children, compared to the other two countries. Data results do not lend support to these proposed relationships since life stages are not significantly associated with housework sharing in the first place. Table 5. Regression results of division of household labour scores on life stages separately in Sweden, Germany, and the United States. Unstandardized coefficients and standard errors in parentheses. Division of Household Labour Variables Sweden Germany U.S. Independent variables Preschool 0.011 -0.031 -0.109 (0.122) (0.096) (0.103) School-age -0.04 1 -0.042 0.050 (0.073) (0.059) (0.075) Controls Gender 0.163 0.219** 0.620** (0.055) (0.045) (0.060) Age 0.007* 0.003 0.001 (0.003) (0.002) (0.002) With college -0.13 1 -0.149 -0.066 (0.069) (0.066) (0.076) Spouse with college -0.113 -0.032 -0.001 (0.068) (0.067) (0.074) Self-employed 0.081 -0.011 0.268* (0.082) (0.069) (0.087) Note: * indicates p <0.05; **indicatesp <0.001 41 For the third research question, let us consider the mediation relationships in figure 5 (a modified diagram of figure 2 by indicating regression pathways of a, b, c, and c’ for the purpose of illustrating the mediation tests performed in the study). Adopted from Baron and Kenny (1986), the total effect between the independent variable and the dependent variable is indicated by “c”, which is obtained in the present study by regressing housework hours on life stage while no alternative explanation (the proposed mediator) is used. Direct effect between the independent variable and the dependent variable through the mediating factor is indicated by “c’ “, which is obtained in the present study by regressing housework hours on life stage, considering an alternative explanation. Relationship between the independent variable and the mediating factor is indicated by the mediation pathway “a”, which is obtained in the present study by regressing an alternative explanation on life stage. Relationship between the mediating factor and the dependent variable is indicated by the mediation pathway “b”, which is obtained in the present study by regressing housework hours on an alternative explanation. In order to establish mediation according to Baron and Kenny (1986), four criteria must be met: (1) effect “c” must first of all be statistically significant, which means that life stage must predict housework hours (this establishes that there is an effect that may be mediated); (2) effect “a” must be statistically significant, which means that life stage predicts the alternative explanation; (3) effect “b” must be statistically significant, which means that the alternative explanation predicts housework hours; and (4) effect “c’ “must be either reduced to zero (for full mediation) or become a smaller effect (for partial mediation) when an alternative explanation is accounted for in the model. 42 CLife Stages Housework Hours C, Alternative Explanations: b Time Availability Relative Resources Gender Ideology Figure 5. Mediation test illustration. For testing the mediation effect of each alternative explanation, regression results “c” is presented in table 5, regression results of mediation pathways “c’ “ and “b” are presented in table 6, and regression results of mediation pathway “a” is presented in table 6. Because no significant relationships are found between life stages and division of household labour in any of the countries examined, the possibility that time availability, relative resource, and gender ideology operate as mediators between life stages and housework sharing are, by the definition of a mediating variable according Baron and Kenny (1986), reduced to null. However, this does not mean that results of such an analysis are entirely of no contribution to the understanding of relationships between life stages, division of labour, the three alternative explanations. Considering table 6, the regression results reveal that even though the three alternative explanations do not receive support as mediators in the relationship between life stages and division of household labour, they are significantly associated with division of household labour. For instance, firstly, time availability has significant associations with division of labour in a 43 Sweden (p <0.001) and the United States (p = 0.001). Though the effect size is larger in Sweden (b = 0.0 12) than in the United States (b = 0.008), the positive values of both of their unstandardized coefficients indicate that the less time women spend in paid work relative to their husbands, the more traditional division of household labour is. Secondly, relative resource is significantly associated with division of household labour in all three countries (p 0.002 for Sweden; p = 0.00 1 for Germany; p = 0.0 13 for the United States). While the effect size is the largest in Sweden and the smallest in the United States (b = -0.070 for Sweden, b = -0.055 for Germany, and b = -0.050 for the United States), the negative values of their unstandardized coefficients indicate that the higher women’s income is relative to their husbands’, the more egalitarian housework sharing is in all three countries. Thirdly, gender ideology is found to be significantly associated with division of household labour in Germany (p < 0.00 1). The negative value of the unstandardized coefficient (b = -0.171) indicate that the more egalitarian gender role attitudes are, the more equally housework is shared. Besides the relationships found between the three alternative explanations and division of household labour, significant relationships are also found between life stages and the alternative explanations (see table 7). The school-age stage is significantly related to time availability in Sweden (p = 0.027). Relative to their husbands’ time spent at work, Swedish women with school-age children spend about three and a half more hours per week in paid work than their counterparts without school-age children. Moreover, the preschool stage is significantly related to relative resource in Germany (p = 0.00 1). German women’s income relative to their husbands is lower if they have preschool children in the household than if they do not. Lastly, the school age stage is significantly related to relative resource in the United States (p = 0.034). U.S. 44 women’s income relative to their husbands’ is lower if they have school-age children in the household than if they do not. 45 Ta bl e 6. R eg re ss io n re su lts o ft he m e di at io n pa th w ay “ c ’ “ : th e re la tio ns hi ps be tw ee n lif e st ag es a n d di vi sio n o fh ou se ho ld la bo ur w hi le a c c o u n tin g fo rm e di at or s, a n d th e pa th w ay “ b” :t he re la tio ns hi ps be tw ee n m e di at or s a n d di vi sio n o fh ou se ho ld la bo ur .U ns ta nd ar di ze d co ef fic ie nt s a n d st an da rd e rr o rs in pa re nt he se s. In de pe nd en tV ar ia bl es D iv is io n o fH ou se ho ld La bo ur V ar ia bl es Sw ed en G er m an y U .S . Pr es ch oo l 0. 02 7 - 0. 01 5 0. 01 2 - 0. 07 7 - 0. 07 7 - 0. 05 7 - 0. 19 5 - 0. 10 4 - 0. 11 4 (0. 13 1) (0. 12 1) (0. 12 2) (0. 13 2) (0. 09 7) (0. 09 2) (0. 13 9) (0. 10 3) (0. 10 3) Sc ho ol -a ge - 0. 04 1 - 0. 05 3 - 0. 04 3 - 0. 03 2 - 0. 05 7 - 0. 02 7 - 0. 02 1 0. 03 9 0. 04 6 (0. 07 7) (0. 07 3) (0. 07 4) (0. 07 1) (0. 05 9) (0. 05 7) (0. 09 7) (0. 07 5) (0. 07 5) Co nt ro ls G en de r 0. 20 0* 0. 18 8* 0. 16 2* 0. 21 9* 0. 24 7* * 0. 26 6* * 0. 58 5* * 0. 64 3* * 0. 63 5* * (0. 06 4) (0. 05 5) (0. 05 6) (0. 06 3) (0. 04 6) (0. 04 4) (0. 08 3) (0. 06 2) (0. 06 1) A ge 0. 00 9* 0. 00 7* 0. 00 7* 0. 00 2 0. 00 3 0. 00 1 0. 00 3 0. 00 2 0. 00 1 (0. 00 4) (0. 00 3) (0. 00 3) (0. 00 4) (0. 00 2) (0. 00 2) (0. 00 4) (0. 00 2) (0. 00 2) W ith co lle ge - 0. 05 5 - 0. 12 9 - 0. 12 8 - 0. 12 4 0. 14 2* - 0. 10 0 - 0. 03 1 - 0. 07 0 - 0. 05 6 (0. 07 9) (0. 06 9) (0. 07 0) (0. 08 7) (0. 06 6) (0. 06 4) (0. 10 1) (0. 07 6) (0. 07 6) Sp ou se w ith co lle ge - 0. 12 1 - 0. 10 8 - 0. 11 2 - 0. 02 7 - 0. 05 8 - 0. 01 7 - 0. 00 8 - 0. 01 9 0. 00 7 (0. 07 5) (0. 06 7) (0. 06 9) (0. 09 3) (0. 06 7) (0. 06 5) (0. 09 7) (0. 07 4) (0. 07 4) Se lf- em pl oy ed 0. 17 2 0. 11 5 0. 07 8 0. 06 2 0. 01 3 0. 01 3 0. 26 3* 0. 25 5* 0. 24 8* (0. 09 2) (0. 08 2) (0. 08 2) (0. 08 5) (0. 07 0) (0. 06 7) (0. 11 3) (0. 08 7) (0. 08 8) M ed ia to rs Ti m e av ai la bi lit y 0. 01 2* * 0. 00 3 0. 00 8* (0. 00 3) (0. 00 2) (0. 00 3) R el at iv e re so u rc e s 0. 07 0* 0. 05 5* 0. 05 0* (0. 02 2) (0. 01 7) (0. 02 0) G en de r id eo lo gy - 0. 01 2 0. 17 1* * - 0. 05 7 (0. 03 8) (0. 02 6) (0. 03 2) No te: * jnd jca tes p < 0. 05 ;* * jnd jca tes p < 0. 00 1 0 , Ta bl e 7. R eg re ss io n re su lts o ft he m e di at io n pa th w ay “ a ” : th e re la tio ns hi ps be tw ee n lif e st ag es a n d m e di at or s. U ns ta nd ar di ze d co ef fic ie nt s a n d st an da rd e rr o rs in pa re nt he se s. Ti m e A va ila bi lit y Re la tiv e R es ou rc es G en de rI de ol og y V ar ia bl es Sw ed en G er m an y U .S . Sw ed en G er m an y U .S . Sw ed en G er m an y U .S . In de pe nd en tV ar ia bl es Pr es ch oo l - 2. 54 3 4. 14 0 6. 23 8 - 0. 10 4 0. 77 9* - 0. 39 3 0. 09 8 - 0. 09 2 - 0. 07 4 (2. 69 2) (4. 21 3) (3. 55 8) (0. 27 2) (0. 22 9) (0. 25 3) (0. 16 0) (0. 14 1) (0. 15 6) Sc ho ol -a ge 3. 50 7* 1. 60 8 2. 08 3 0. 01 4 - 0. 12 5 0. 38 9* - 0. 09 0 0. 16 4 - 0. 05 7 (1. 57 4) (2. 24 9) (2. 53 3) (0. 16 4) (0. 13 9) (0. 18 3) (0. 09 6) (0. 08 7) (0. 11 4) Co nt ro ls G en de r 0. 22 7 4. 98 3* . 59 97 * 0. 36 9* 0. 58 6* * 0. 69 8* * 0. 11 7 0. 27 8* * 0. 27 0* (1. 31 7) (1. 99 3) (2. 13 5) (0. 12 4) (0. 10 6) (0. 14 9) (0. 07 3) (0. 06 6) (0. 09 2) A ge 0. 26 4* 0. 02 8 0. 22 3* 0. 00 2 0. 00 3 0. 00 7 0. 01 4* * 0. 00 8* - 0. 01 1’ (0. 07 7) (0. 13 3) (0. 10 6) (0. 00 6) (0. 00 5) (0. 00 6) (0. 00 4) (0. 00 3) (0. 00 3) W ith co lle ge - 2. 29 2 - 2. 63 7 2. 80 5 0. 02 7 0. 05 3 0. 01 5 0. 26 7* 0. 20 7* 0. 17 7 (1. 62 0) (2. 76 6) (2. 62 7) (0. 15 5) (0. 15 6) (0. 18 6) (0. 09 1) (0. 09 8) (0. 11 6) Sp ou se w ith co lle ge 1. 00 9 4. 08 5 1. 44 1 0. 06 1 - 0. 30 2 - 0. 31 1 0. 22 1* 0. 16 3 0. 16 9 (1. 55 2) (2. 95 1) (2. 53 7) (0. 15 2) (0. 15 9) (0. 18 1) (0. 09 0) (0. 09 8) (0. 11 3) Se lf- em pl oy ed 2. 32 9 2. 00 3 0. 31 5 0. 27 6 0. 32 1 0. 18 8 - 0. 15 2 0. 04 4 0. 33 2* (1. 90 2) (2. 70 5) (2. 89 2) (0. 18 4) (0. 16 6) (0. 21 5) (0. 10 8) (0. 10 3) (0. 13 2) No te: * jnd jca tes p < 0. 05 ;* * jnd jca tes p < 0.0 01 Post-Hoc analysis As no relationship is found between life stages and division of household labour in any of the countries examined, I have the concern of whether the relative frequency measure of housework performance between the spouses employed in the initial analysis is sensitive enough to capture the variations of division of labour along the life-course. Therefore, a post-hoc analysis is conducted to determine if variations of housework sharing by life stages could be found with a different household labour measure reviewed earlier, that is, a measure that assesses the time expenditure for housework among men and women. The choice of this measure is based on the study by Baxter and colleagues (2008) in which women’s weekly housework hours are found to vary considerably in relation to transition to parenthood whereas men’s are not. This measure is obtained from the item “How many hours do you spend on housework?” in the ISSP 2002 data set. Response categories range from 0 (0 hours per week) to 95 (95 hours or more per week for Sweden and Germany; 96 hours or more per week for the United States). However, since the main goal of this post-hoc analysis is to investigate whether with the number-of-hour measure I could find variations of housework performance dependent on life stages, results of the mediation tests are only briefly stated in text. Descriptive Results Figure 6 presents the mean reported number of weekly hours spent in housework by gender and life stages in Sweden, Germany, and the United States. In the preschool stage, both Swedish and U.S. women with preschool children report spending more time on housework (with mean hours per week of 15.67 and 18.41 for Sweden 48 and the U.S. respectively) than women without preschool children do (with mean hours per week of 15.03 and 13.14 for Sweden and the U.S. respectively), whereas German women with preschoolers report spending less time (with mean hours per week of 19.33) than their counterparts without preschoolers (with mean 21.84 hrs/wk). However, in none of these countries are the differences in time spent between women with preschool children and women without statistically significant. On the contrary, reported housework hours in these three countries are all significantly lower (p = 0.008 for Sweden; p < 0.00 1 for Germany; p = 0.00 1 for the United States) for men with preschool children (with mean hours per week of 6.12, 5.04, and 6.59 for Sweden, Germany, and the United States respectively), compared to men without children of this age group (with mean hours per week of 9.41, 7.67, and 12.41 for Sweden, Germany, and the United States respectively). In the school-age stage, on the female side, only U.S women with school-age children spend significantly more (p = 0.038) time in housework (with mean hours per week of 16.76) than their counterparts without children of school-age (with mean hours per week of 12.98). Even though Swedish and German women exhibit an opposite pattern from the United States’ both Swedish and German women with school-age children report less time in housework (with mean hours per week of 14.41 and 19.95 for Sweden and Germany respectively) than their counterparts without school-age children (with mean hours per week of 15.35 and 22.20 for Sweden and Germany respectively) --- the differences in time spent between women with and without children of this age group are not statistically significant. A significant difference in mean hours per week between men with and without school-age children is only found in the U.S. sample as well (p = 0.03 1). U.S. men with school-age children spend less time in housework (with mean hours per week of 7.89) than their counterparts without school-age 49 children (with mean hours per week of 12.91). The Germany sample resembles this pattern (with reported mean hours per week of 6.56 for men with school-age children and 7.83 for men without school-age children), yet the difference between the two groups of men is not statistically significant. Neither is the difference significant for the two groups of men in the Sweden sample, although Swedish men with school-age children report more time in housework (with mean hours per week of 10.02) than their counterparts without children of this age group (with mean hours per week of 8.81). Overall the consistent finding is that women across life stages and countries spend more time on housework than men, and they do so more in the preschool stage than the school-age stage. Germany’s division of household labour is the most unequal between husbands and wives in both life stages among the three countries (with difference of mean hours per week of 14.29 in the preschool stage and 13.39 in the school-age stage), whereas Sweden’s division of household labour is the most equal (with difference of mean hours per week of 9.55 in the preschool stage and 4.39 in the school-age stage. Between-life-stage comparison reveals that the gendered gap between husbands’ and wives’ time spent in housework is the most resistant to change in Germany (with a 0.9 difference in gendered gap of mean hours per week between the preschool and the school-age stage) along the life-course, and the most susceptible to change in Sweden (with a 5.16 difference in gendered gap of mean hours per week between the preschool and the school-age stage). ANOVA tests reveal that both life stages have significantly different effects on men and women in all three counties (p < 0.001 for Sweden and Germany in both the preschool and the school-age stage; p = 0.006 and p = 0.009 for the United States in the preschool and the school-age stage respectively). Between-country differences in mean hours per 50 week are larger in the school-age stage than they are in the preschool stage, although neither of EJfemale without preschool female with preschool D male without preschool • male with preschool D female without school-age female with school-age D male without school-age • male with school-age Figure 6. Mean reported housework hours per week by gender along the life cycle. Married persons with preschool children or school-age children in the household. the life stages has statistically significant difference of effects on the countries. Refer to appendix A for descriptive statistics for the rest of the variables in relation to country, life stages, and gender. Essentially they are the same as what is discussed in the initial analysis. A. Preschool Stage Sweden Germany U.S. 4) 4) 25 4) 0. 20 U) 15 0 10 4) 4) E 4) 4) 25 4) 0. 20 U) 15 0 10 4) a) E B. School-age Stage Sweden Germany U.S. 51 Test ofHypotheses With regard to the first and the second research questions, I examine the relationship between life stages and housework hours in the three countries in regression. Regression results predicting housework hours by life stages are presented in table 8. Concerning the first research question, neither of the life stages examined has a significant association with number of hours for housework for the male respondents in all three countries. For the female respondents, on the other hand, the preschool stage is significantly associated with Swedish women’s housework hours (p = 0.0 12) while both the preschool and the school-age stage are significantly associated with U.S. women’s housework hours (p = 0.006 and p = 0.0 12 for the preschool and the school-age stage respectively). Therefore, the hypothesis regarding the division of housework being dependent on the life stages is partially supported. Concerning the second research question, the first hypothesized finding regarding both the preschool and the stage-age stage having the smallest effect on housework sharing in Sweden, compared to the other two countries, is partially supported. This is because the preschool stage has the smallest effect on housework time expenditure in Sweden (coefficients indicate that Swedish women with preschool children spend about six more hours per week on housework than their counterparts without preschool children, whereas U.S. women with preschool children spend nearly seven more hours per week on housework than their counterparts without preschool children, considering that men’s housework time is unaffected by the presence of preschool children). The school-age stage, nonetheless, has significant effects on housework time expenditure only in the United States (coefficients indicate that U.S. women with school-age children spend about four and a half hours more per week in housework than their counterparts 52 without school-age children while men’s housework time is unaffected by the presence of school-age children). The second hypothesized finding regarding the division of labour being more traditional and unequal in Germany with school-age children than with preschool children, compared to the other two countries, is not supported. The reason is that life stages are not significantly associated with housework performance in Germany in the first place. Thus I am not able to obtain significant effect of children in either life stage for comparison. The third hypothesized finding regarding the division of labour being more egalitarian in the United States with school-age children than with preschool children, compared to the other two countries, is supported. With U.S. men’s housework time unaffected by life stages, U.S. women with preschool children spend nearly seven more hours per week in housework than without preschool children whereas U.S. women with school-age children spend approximately only four and a half more hours per week in housework than without school-age children. Therefore, housework is more equally shared in the school-age stage and in the preschool stage. As mentioned earlier, the third hypothesis concerning the mediating effects of time, resource, and gender ideology in the relationship between life stages and housework sharing is not the main focus of the post-hoc analysis. Therefore, mediation results are not discussed in text in great details. However, results of the four effects such as those presented in figure 5 that are essential to the establishment of a mediation relationship are still included in this paper. Referring to figure 5, regression results “c” are presented in table 8. Regression results “c’” along with the regression results of mediation pathway “b” are presented in appendix B. Regression results of mediation pathway “a” are presented in appendix C. An examination of the 53 mediation results presented in the table and appendices mentioned above reveals that in no country is time, resource, or gender ideology established as a mediator through which life stages predict time spent in housework. Table 8. Regression results of housework hours on life stages by gender in Sweden, Germany, and the United States. Unstandardized coefficients and standard errors in parentheses. Housework Hours Sweden Germany U.S. Variables Men Women Men Women Men Women Independent Variables Preschool 2.367 6.082* -0.150 -0.912 -5.679 6.980* (3.896) (2.411) (1.692) (3.086) (5.200) (2.498) School-age 3.485 2.887 0.236 -1.212 -4.314 4.615* (2.336) (1.532) (1.012) (1.942) (3.590) (1.833) Controls Age 0.277* 0.183* 0.115 0.094 0.085 0.027 (0.091) (0.063) (0.038) (0.071) (0.102) (0.060) With college -0.806 2.512* 1.087 -3.713 -5.760 -0.754 (2.733) (1.216) (1.019) (2.434) (3.696) (1.866) Spouse with college 0.639 -0.402 -1.126 -1.752 4.171 -1.321 (2.643) (1.225) (1.392) (1.899) (3.809) (1.749) Self-employed -0.903 1.092 -0.266 7.107* -4.527 0.856 (2.400) (1.988) (1.121) (2.509) (3.835) (2.241) Note: * indicates p < 0.05 54 CHAPTER 5 Discussion The purpose of this study is to investigate the relationship between life stages and division of household labour across contexts. Specifically, this study examines the presence of this relationship in Sweden, Germany, and the United States, the effect directions of life stages on division of labour both within and across countries, and the three alternative explanations as the mediators through which variations of participation in domestic tasks are shaped by life stages. Past research using U.S. samples suggests that the life-course approach is an important perspective for the understanding of housework sharing. Housework sharing is never equal or mostly shouldered by men; women contribute more than men do in female-typed tasks. However, the presence of children in the house results in domestic work becoming more unequally shared, with women increasing their proportionate contribution more than men; this disparity is more pronounced with younger children present in the house as opposed to older children. Time, resource, and gender ideology are theorized as the factors shaping such changes across life stages. However, how much of the relationship between life stages and division of labour observed in the U.S. samples can be found in Sweden and Germany? The outcomes of the present study suggest evidence of the applicability of the life-course approach and the three alternative explanations to the study of division of housework not only in the United States but also in Sweden and Germany, whose contexts are different from the United States’. 55 Findings Related to the First Two Research Questions The first research question concerns the variation of division of household labour dependent on life stages, and the second research question concerns the differences between countries in their variation of division of labour by life stages. The second research question specifically investigates three between-country differences: (1) both life stages have the smallest effect on division of labour in Sweden, relative to the other two countries, (2) Germany has a more traditional division of labour in the school-age stage than in the preschool stage, relative to the other two countries, and (3) the United States has a more egalitarian division of labour in the school-age stage than in the preschool stage, relative to the other two countries. Both research questions are not supported when division of labour refers to the relative frequency of housework participation between husband and wife. However, the research questions receive partial support in Sweden and the United States when division of labour refers to the number of hours that husbands and wives spend on housework separately from one another. In this section I discuss the implications of the findings obtained from these two different measures of division of household labour. While the initial analysis does not reveal an association between the preschool stage and division of household labour in Sweden, such an association is found in the post-hoc analysis for women in Sweden. This reveals that while relative frequency of housework participation does not vary with the presence of preschool children in the household for Swedish couples, Swedish women’s number of hours of housework does. Compared to U.S. women, Swedish women with preschoolers spend less time on household tasks, indicating that the preschool stage has a smaller 56 effect on division of labour for Swedish women than for U.S. women in terms of their time expenditure on domestic chores. In the United States, even though neither life stage is associated with division of household labour in the initial analysis, such an association is observed in the post-hoc analysis. This reveals that while relative frequency of housework participation between U.S. spouses does not vary with the presence of preschool and school-age children, U.S. women’s number of hours of housework does. U.S. women with school-age children spend less time on household tasks than their counterparts with preschool children, indicating that they exhibit a more egalitarian housework arrangement in the school-age stage than in the preschool stage. Post-hoc findings in Sweden and the United States suggest that women shoulder more housework responsibility than men and that Swedish women do so more in the preschool stage while U.S. women do so more in both the preschool and school-age stages. The fact that the initial analysis does not find a relation between either life stage and division of household labour in these two countries does not negate the possibility of a difference in housework responsibility between women and men. For example, consider the following scenarios: a wife without preschool children who spends thirty hours on housework while her husband spends fifteen hours on the same tasks, and a wife with preschool children who spends sixty hours on housework while her husband spends twenty hours. In both cases, even though the women in the second scenario has a higher proportionate share of housework relative to her husband’s, the answer to the question “who does these housework tasks in the household” would still be “usually the woman” as the other response categories (“always the woman”, “equally shared”, “usually the men”, and “always the man”) would be deemed less fitting. Therefore, the lack of a 57 relationship between life stages and division of household labour in the initial analysis only implies the absence of variations in the relative frequency of who does housework based on the presence of children. When I first examine the relation between division of labour and life stages according to the measure of relative frequency of housework participation, there is no statistically significant variation. Nonetheless, when I re-examine the relation according to the measure of number of hours of housework, there are variations in that Swedish and U.S. women’s time expenditure increases with the presence of children. Stated differently, division of labour in Sweden and the United States does not vary significantly by life stages in men’s and women’s relative frequency of housework participation. This is probably because while women tend to always perform housework more frequently than men throughout the life-course, men also perform some of the household chores. Therefore, in terms of relative participation frequency, division of labour does not have statistically significant variation by life stages --- always “usually the woman”. However, division of labour does have statistically significant variation in Sweden and the United States in terms of women’s numbers of hours of housework separate from men’s. Both Swedish and U.S. women spend more hours on domestic tasks with the presence of children even though their husbands’ time expenditure on these tasks remains unchanged. In Germany, unlike in Sweden or in the United States, neither initial nor post-hoc analysis reveals an association between life stages and division of household labour. In other words, having children of a certain age group is irrelevant to housework arrangement in Germany, both in terms of couples’ relative frequency of housework participation and their individual number of hours of housework. German women not only always perform housework 58 more frequently than their husbands along the life-course but they also spend the same amount of time in housework regardless of the presence of children. In other words, division of labour in Germany is neither more nor less traditional in the school stage than in the preschool stage, regardless of whether division of labour refers to spouses’ relative frequency of housework participation or men’s and women’s individual number of hours spent on housework. Results related to the first two research questions also provide implications to the life- course perspective in relation to role changes. As mentioned before, parenthood has been found to be related to a more traditional division of labour in the United States (Rexroat & Shehan, 1987; Sanchez & Thomson, 1997; Shelton, 1992), as opposed to pre-parenthood. According to the life-course perspective, such a difference in housework arrangement between families with and without children is due to the effects that children may exert on family members’ relationships to one another (Nock, 1979). In other words, the presence of children changes the roles of both the husband and the wife in a family. Compared to before having children, husbands and wives adopt more traditional roles upon entrance into parenthood. While parenthood for fathers implies their increasing role as household providers, parenthood for mothers implies their increasing role as household caretakers. In this case, the life-course perspective predicts that the gender gap in housework time expenditure between the spouses widens with the presence of children. My U.S. results support this prediction: with mothers’ number of hours of housework increasing with both preschool and school-age children while fathers’ number of hours of housework being unrelated to the presence of children in both life stages, the gender gap in housework time expenditure between the spouses is bigger with than without children. In particular, my findings on the differential effects of the presence of children 59 on men’s and women’s housework participation in the United States are consistent with the differential effects observed in the study by Baxter, Hewitt, and Haynes (2008). However, while my U.S. results are consistent with the predictions of the life-course perspective with regards to housework participation, my Sweden and Germany results are not consistent with these predictions. A lack of support from the Sweden and Germany results for the first two hypotheses suggests that men and women in these two countries may adopt parental roles that are quite different from those adopted by their U.S. counterparts. U.S. mothers spend nearly seven more hours per week on housework with preschool children and approximately four and a half more hours per week on housework with school-age children. Nonetheless, Swedish mothers spend only six more hours per week on housework with preschool children, but they do not spend more time on housework with school-age children. In addition, German mothers do not spend more time on housework with children of either age group. This between-country difference in mothers’ housework time expenditure suggests that while parental roles for women in the United States may imply the shouldering of more housework responsibilities, it may not be the case in Sweden and Germany. Being a mother in Sweden or Germany may not imply that one has to clean the house, wash the dishes, and cook the meals, etc., although fulfilling these responsibilities seems to be part of what being a mother means in the United States. Furthermore, while in no country examined is men’s time spent on housework significantly related to the presence of children, results reveal that although Swedish and German fathers report spending more hours on housework, U.S. fathers report spending less hours on housework. This not only suggests that parental roles for men in the United States may not require them to fulfill more responsibilities for domestic chores but also suggests that being a father in the United States may denote that one should invest one’s time in activities other than housework. For example, a father 60 should spend more time in paid jobs in order to provide for his family, as the life-course perspective presented in the existing literature would suggest for the parental role of a man. On the contrary, being a father in Sweden and Germany does not seem to denote the same meaning as it does in the United States. The data suggests that the parental roles for fathers in Sweden and Germany may subscribe more to men’s increased responsibility in sharing domestic chores than the parental roles for fathers in the United States. In essence, therefore, Sweden and Germany may represent different subscriptions and expectations for the role of a parent than the United States. While men’s role as a parent may denote distinctly different responsibilities (that is, a family provider) from women’s role as parent (that is, a housework caretaker) in the United States, such a distinction between men’s and women’s parental roles may be of a much lesser degree in Sweden and Germany. Findings Related to the Third Research Question In no country do any of the alternative explanations of time, resource, and gender ideology receive support as mediating factors in the relationship between life stages and division of household labour under either the initial or the post-hoc analysis. This suggests that the inability to establish the alternative explanations as mediators is not, or at least not merely, related to the differences between the two divisions of labour measures. Whether division of labour is defined as the relative frequency of spouses’ housework involvement or the spouses’ weekly hours spent on domestic tasks, the alternative explanations do not mediate the relationship between life stages and division of household labour. 61 Other Findings --- The Alternative Explanations and the Context Even though the present study does not support the alternative explanations as mediators in the relationship between life stage and division of household labour, it finds that time, resource, and gender ideology shape division of labour in Sweden, Germany, and the United States. However, not all three explanations seem to be weighted equally. Time availability. The theory of time availability seems to be a better explanation for either a social-democratic or a liberal regime, but not a conservative regime: less time investment in paid work by Swedish and U.S. women relative to their husbands results in a more traditional division of housework. Compared to the United States, time availability has a stronger predictive power on division of household labour in Sweden. This means that, with equally higher relative time investment in paid work than their husbands, Swedish women are able to lower their relative frequency of housework participation more than U.S. women. Moreover, with equally lower relative time investment in paid work than their husbands, Swedish women also raise their relative frequency of housework participation more than their U.S. counterparts. According to the assumption of the time availability theory, this implies that with less relative time spent in the paid labour force, both Swedish husbands and wives are more willingly to invest their time in housework than U.S. husbands and wives. This implication corresponds to another assumption of the theory: gender-neutrality of housework allocation. Because routine household tasks such as cleaning and cooking probably have less gendered stigma in a social-democratic welfare regime than they do in a liberal welfare regime, Swedish individuals are more inclined than their U.S. counterparts to take on more responsibility for these household tasks upon lowered time expenditure in paid work relative to their spouses. 62 A previous study by Calasanti and Bailey (1991) finds that Swedish men’s fewer working hours predict their increased involvement in housework. The present study adds to the understanding of the relationship between time availability and division of household labour in Sweden. That is, while Swedish men’s time spent in paid job independent of their wives’ predict men’s housework involvement, their time spent in paid job relative to their wives’ predict their relative frequency of housework involvement. Moreover, while Calasanti and Bailey find men’s time spent in paid job independent of their wives’ a stronger predictor for men’s housework involvement in the United States than it is in Sweden, the present study finds men’s time spent in paid job relative to their wives’ a stronger predictor for men’s relative frequency of housework involvement in Sweden than it is in the United States. Furthermore, previous studies using U.S. samples have typically found an association between absolute hours of market work and division of household labour. For example, women’s longer work hours is associated with not only men’s increased housework time (Blair & Lichter, 1991; Brines, 1993) but also men’s increased proportional contribution to housework (Ishii Kuntz & Coltrane, 1992). Moreover, Bianchi and colleagues (2000) find individuals’ employment hours predict their spouses’ as well as their own time expenditure in housework: one’s work hours increase his or her spouse’s housework while decrease his or her own. However, compared to the effects of wives’ work hours on their husbands’ housework time, the effects of husbands’ work hours on their wives’ housework time are much smaller. Brayfield (1992) examining Canadian husbands’ housework participation in relation to their wives’ work hours relative to the husbands’, instead of wives’ absolute work hours independent of the husbands’, finds no association between employment and housework time. Nevertheless, the 63 present study finds that U.S. women’s relative employment time to their husbands’ is negatively associated with an egalitarian division of household labour. In addition to examining the relationship between time availability and the division of housework in Sweden and the United States, the present study also examines this relationship in Germany. A previous study by Lewin-Epstein and Stier (2006) finds that German women’s increased employment time predicts their decreased housework involvement independent of their husbands’ market situation. Geist’s (2005) comparative study of welfare regimes and domestic division of labour also finds a relationship between men’s increased employment time and less egalitarian division of household labour, and this is to a greater extent in conservative regimes than in social-democratic regimes. Nevertheless, the present study finds that the relative frequency of housework participation in Germany is unaffected by spouses’ relative time investment in paid work. A plausible explanation of this is that while in social-democratic and liberal regimes housework is less expected to be the women’s job, it is more so the case in the conservative regime. When the normative view of a society is that the domestic work is women’s responsibility while the paid employment is men’s, women may be less likely to put paid jobs before domestic responsibility. Therefore, whether or not women are employed outside of the home more than their husbands, they shoulder the majority of the housework. Consequently, time availability does not explain division of labour in the Germany sample. Perhaps, findings revealed in the previous and the present studies suggest that while spouses’ separatc employment time may be associated with division of housework, their relative employment time of one another may not be. 64 Relative resource. The theory of relative resource seems to be applicable to all three types of welfare regimes. This suggests that, regardless of context differences, some resources, such as relative income in the case of the present study, has a substitution effect for housework. Previous studies using only U.S. samples have observed a relationship between relative resource and the division of household labour. For example, some researchers find equal earnings between men and women predict equal sharing of household tasks (Blair & Lichter, 1991; Brayfield, 1992; Presser, 1994). Some other researchers find that the wife’s greater proportion of couple income predicts less housework for herself and more housework for her husband (Bianchi, Milkie, Sayer, & Robinson, 2000). In addition, the present study correspondents to the cross- national study by Geist (2005): across all types of welfare regimes, women’s higher relative income to their husbands’ increases the odds of equal sharing of household labour. Also, consistent with previous findings by Calasanti and Bailey (1991), Evertsson and Nermo (2007) and Lewin-Epstein and Stier (2006), resources predict the division of household labour not only in the United States but also in Sweden and Germany. In this regard, the different gender beliefs underlying different welfare regimes become irrelevant to the values and meaning of resources in relation to division of household labour. However, gender beliefs embodied in welfare policies still imply different values and meanings for resources as far as the strength of the relationship between relative resource and spouses’ relative frequency of housework participation is concerned. For example, the association between spouses’ relative income and their relative frequency of housework participation is stronger in Sweden than it is in Germany and the United States. This indicates that relative income is more effective in predicting an egalitarian division of household labour in Sweden than Germany and the United States. In order words, given equally higher income 65 relative to their husbands’, Swedish women are able to reduce their relative frequency of housework participation more than their German and U.S. counterparts. Conversely, Swedish women also encounter more increase in their relative frequency of housework participation than their German and U.S. counterparts with equally lower income relative to their husbands’. In this sense, a social-democratic welfare regime gives way to relative resource to influence housework arrangement more than a conservative or a liberal welfare regime. This seems to be a paradox in that, on the one hand, Sweden’s social-democratic welfare regime actively promotes gender equality more than the other two countries both in the paid labour force and in the domestic sphere. On the other hand, it is probably precisely because of such policies affording men and women equal opportunities in the paid labour force that higher relative income is entitled to more influencing power over housework sharing between the spouses regardless of who makes more. Therefore, when one spouse makes more than the other in a social-democratic regime, a greater difference in relative frequency of housework participation is observed compared to a conservative or a liberal regime. On the contrary, because neither a conservative nor a liberal regime is as active as a social-democratic regime in promoting gender equality and affording policies that reinforce beliefs concerning gender equality, higher relative income does not make as much difference in housework arrangement. Thus, housework sharing is only moderately more traditional when the husband makes more than the wife in a conservative or a liberal regime; conversely, housework sharing is only moderately more egalitarian when the wife makes more than the husband. Yet the differences between husband’s and wife’s relative frequency of housework participation are more dramatic when one spouse makes more than the other in a social-democratic regime. 66 Gender ideology. The theory of gender ideology seems to be the most limited in cross-context applicability among the three alternative explanations. It is only found to be significantly associated with the division of labour in Germany, a conservative welfare regime. The more egalitarian the gender role attitudes are in Germany, the more egalitarian division of household labour is. This suggests that even when macro-level factors, such as the societal values in a conservative regime, expect women to bear domestic responsibility, individual-level factors such as personal gender role attitudes may still play a role in affecting the division of household labour. A previous study by Geist (2005) suggests that less traditional gender role attitudes do not seem to be as effective in predicting a more egalitarian division of housework in conservative countries as they do in social-democratic and liberal countries. However, findings in this current study suggest otherwise: less traditional gender role attitudes are effective in predicting a more egalitarian division of housework in a conservative country but not in a social- democratic or a liberal country. This suggests that the macro-level factors are more powerful than the individual-level factors in predicting division of household labour in social-democratic and liberal countries, whereas both macro-level and individual-level factors are effective predictors of housework arrangement in conservative countries. It may seem surprising that the present study does not find support of the gender ideology explanation in its U.S. sample despite of an apparently well-documented relationship between gender role attitudes and the division of housework in the United States that is revealed in the existing research. For instance, Coltrane’s (2000) review of the studies from the 1990s show that both men’s and women’s egalitarian gender role attitudes are associated with equal sharing of household tasks, with women’s gender role attitudes being a stronger predictor of division of labour than men’s gender role attitudes. The review by Shelton and John (1996) also state that 67 most studies using proportional measures of men’s housework contribution find men’s egalitarian gender role attitudes associated with equal housework sharing. Bianchi and colleagues (2000) find that, on the one hand, wives’ egalitarian gender role attitudes are associated with a more equal division of household labour by their own lowered time spent in housework. On the other hand, husbands’ egalitarian gender role attitudes are associated with a more equal division of household labour not by their own housework time expenditure. Rather, it is by their wives’ less time spent in housework because wives married to husbands with traditional role attitudes tend to spend more time in housework. However, Brayfield (1992) finds that men’s gender role attitudes are not associated with their proportionate share of housework if both spouses are employed full-time. In addition, Shelton and John (1996) point out that studies that have weak measures of attitudes (i.e. attitudes are accessed by single item instead of multiple items) or housework (i.e. housework measured by who is responsible for housework produces less variance than when housework is measured by the number of hours of housework done by the wife, the husband, or both, as Ross (1987) discussed) may not find a relationship between gender role ideology and the division of household labour. Therefore, these studies raise the question of whether the fact that the present study does not observe an association between gender role ideology and the division of housework in the United States is due to the fact that most of the U.S. participants and their spouses in the present study have full-time employment and/or the way housework is measured in the present study. Limitations There are two major limitations I would like to address for the present cross-national study of division of labour along the life-course. One limitation is the difficulty in interpreting 68 differences found between countries. The other limitation is the possibly different meanings of the two measures for division of household labour. Concerning the first limitation, cross-national differences are not as readily explained as cross-national similarities, as Kohn (1987) extensively argued. To briefly summarize Kohn’ s argument , when cross-national similarities are observed, we can have more confidence that these similarities are not caused by any of the different factors between the two countries or any noncomparability resulting from methods of data collection because these would have produced differences, not similarities, across countries (adapted from Boss’s (1993) discussion of Kohn’s argument). Therefore, we can be more certain that our explanation accounting for the cross- national similarities are generalizable across countries. However, with cross-national differences it is almost impossible to disentangle competing theories or explanations merely by statistical means due to the classic problem of over-identification in cross-national research. With a good theory that predicts what differences should be observed under what kind of well-outlined condition, the observation of the predicted differences only constitutes support for that particular theory but cannot rule out other competing theories. Applying Kohn’ s argument on the findings of the present study means that, for example, even though I find welfare policies a good explanation for the differences of how some of the variables, such as time, resource, and gender ideology, are related to division of labour in one country but not in the other, the most defensible conclusion I may reach is that the different welfare policies are explanations supported by my data set. My data results have no power to discredit other competing explanations, such as how female labour force participation rates, country’s economic development, and the cultural gender norms may also explain the differences 69 observed in the application of time, resources, and gender ideology on division of household labour across Sweden, Germany, and the United States. With regard to the second limitation, the two measures of division of household labour employed in the study have different specifications as to what kind of housework they are to measure. For example, the initial division of household labour scale is constructed based on responses to four female-typed tasks --- doing laundry, caring for sick family members, cleaning, and cooking. However, the division of household labour measure used in the post-hoc analysis does not have any specification as to what kind of housework it refers to. Rather, it is a fairly general question of how many hours per week do the respondents spend on “housework”. The concept of “housework” is thus opened to individual respondents’ interpretations and decisions as to what is considered as housework. It is then possible that they respond on a range of different household tasks, including both male- and female-typed work. Therefore, result differences observed between the initial and the post-hoc analyses could be due to the fact that two different phenomena are being observed, as opposed to one phenomenon from different perspectives (i.e. frequency of involvement and time expenditure). Future Research Future cross-national research on division of household labour along the life-course should not only consider antecedents but also consequences of variations in housework arrangements. For example, what else besides welfare policies might explain variations across countries in division of labour by life stages? One might consider, for instance, how housework arrangements differ by life stages since female employment has been linked to division of housework (Drobni & Treas, 2006) and the rate of female labour force participation has been 70 shown to vary through the life stages (Stier & Lewin-Epstein, 2001). In addition, marital satisfaction on division of labour has been examined with U.S. samples (Suitor, 1991). It would contribute to the value of cross-national research on division of household labour by exploring how different (or similar) patterns in housework arrangements along the life-course between countries affect people’s well-being differently (or in similar ways). Furthermore, even though housework in the initial analysis of the present study refers to female-typed tasks only, housework in fact is not consisted of these tasks alone. There are also male tasks, such as yard work and home repairs, and gender-neutral tasks, such as taking out the trash and handling finances. In the 2002 ISSP data employed in the present study there is only one male task (i.e. small housework repairs) and one gender-neutral task (i.e. grocery shopping). Thus this study is unable to conduct a reasonable study on division of household labour encompassing and comparing between different types of household tasks. However, in order to obtain a fuller understanding of housework arrangements in various types of tasks, future studies may investigate the dynamics in differences of participation among female, male, and gender- neutral tasks in a life-course perspective. Across life stages, how do men and women participate more or less in one type of tasks versus the others? What are the implications of one doing more of one type of tasks than the other types in one life stage on his or her involvement in the same and/or different types of tasks in another life stage? Conclusion Results from the present study suggest that a life-course approach is useful in the examination of division of household labour cross-nationally. The usefulness of the life-course approach in examining housework arrangement seems to be dependent on the dimension of 71 housework performance measured (i.e. spouses’ relative frequency of housework participation or their number of hours spent on housework) as well as the context (i.e. social-democratic, conservative, and liberal welfare regime). Life stages predict division of household labour in Swedish and U.S. women’s number of hours for housework. However, in no countries examined do they predict division of housework in terms of couple’s relative frequency of housework participation. The three alternative explanations of time, resource, and gender ideology have varied predictive power on division of housework in different countries, but none of them are established as mediators in the relationship between life stages and housework sharing. Contextual characteristics such as welfare regimes are helpful in explaining how one alternative explanation may account for the division of household labour in one country but not in another. Findings in the relationships (or the absence of them) between the alternative explanations and housework sharing reveal both consistency and inconsistency with previous findings of the same kinds of relationships. Furthermore, this study provides a meaningful comparison between two different measures of the division of household labour in relation to the hypotheses. Furthermore, the life-course approach of the examination of the division of labour in Sweden, Germany, and the United States also suggests that parental roles for men and women may be different among these countries. A U.S. mother’s role may be more closely connected to women’s increased involvement in housework responsibilities, compared to their roles in pre-parenthood or post parenthood. A U.S. father’s role, on the other hand, maybe closely connected to men’s decreased involvement in housework responsibilities (which is possibly an indicator of their increased involvement in paid work), compared to their roles in pre-parenthood or post parenthood. Nevertheless, men’s and women’s roles as a parent in Sweden and Germany may 72 not be as different as their roles in pre-parenthood or post-parenthood, as far as their involvement in housework responsibilities is concerned. 73 References Baron, R.M., & Kenny, D.A. (1986). The moderator-mediator variable distinction in social psychology research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173-1182. Baxter, J. (1997). Gender equality and participation in housework: A cross-national perspective. Journal of comparative Family Studies, 28, 220-247. Baxter, J., Hewitt, B., & Haynes, M. (2008). Life course transitions and housework: Marriage, parenthood, and time on housework. Journal ofMarriage and Family, 70, 259-272. Bergen, E. (1990). The multidimensional nature of domestic labour: An investigation of husbands’ participation (Working Paper No. 36). Madison, WI: Center for Demography and Ecology. Blair, S. L., & Lichter, D. T. (1991). Measuring the division of household labour. Journal of Family Issues, 12, 91-113. Blau, D. M., & Robins, P. K. (1991). Child care demand and labour supply of young mothers over time. Demography, 28, 333-35 1. Blood, R. 0., & Wolfe, D. M. (1960). Husbands and wives. Glencoe, IL: Free Press. Bianchi, S. M., Milkie, M. A., Sayer, L. C., & Robinson, J. P. (2000). Is anyone doing the housework? Trends in the gender division of household labour. Social Forces, 79, 191- 228. Bird, C. E. (1999). Gender, household labour, and psychological distress: The impact of the amount and division of housework. Journal ofHealth and Social Behavior, 40, 32-45. Brayfield, A. (1992). Employment resources and housework in Canada. Journal ofMarriage and the Family, 54, 19-30. Brines, J. (1993). The exchange value of housework. Rationality and Society, 5, 302-340. Brines, J. (1994). Economic dependency, gender, and the division of labour at home. The American Journal of Sociology, 100, 652-688. Borchorst, A. (1994). Welfare states regimes, women’s interests and the EC. In Sainsbury, D. (Ed.), Gendering Welfare States. Thousand Oaks, CA: Safe, pp.24-44. Boss, P. (1993). Sourcebook offamily theories and methods: A contextual approach. New York: Plenum Press. 74 Calasanti, T. M., & Bailey, C. A. (1991). Gender inequality and the division of household labour in the United States and Sweden: A social-feminist approach. Social Problems, 38, 34-53. Coltrane, 5. (1996). Family man: Fatherhood, housework, and gender equity. New York: Oxford University Press. Coltrane, S. (2000). Research on household labour: Modeling and measuring the social embeddedness of routine family work. Journal ofMarriage and the Family, 62, 1208- 1233. Coltrane, S., & Ishii-Kuntz, M. (1992). Men’s housework: A life course perspective. Journal of Marriage and the Family, 54, 43-57. Coverman, 5. (1985). Explaining husbands’ participation in domestic labour. Sociological Quarterly, 26, 8 1-97. Cook, L. P. (2006a). “Doing gender” in context: Household bargaining and the risk of divorce in Germany and the United States. American Journal of Sociology, 112, 442-472. Cook, L. P. (2006b). Policy, preferences, and patriarchy: The division of domestic labour in East Germany, West Germany, and the United States. Social Politics, 13, 117-143. DeMaris, A., & Longmore, M. A. (1996). Ideology, power, and equity: Testing competing explanations for the perception of fairness in household labour. Social Forces, 74, 1043- 1071. Diefenbach, H. (2002). Gender ideologies, relative resources, and the division of housework in intimate relationships: A test of Hyman Rodman’ s theory of resources in cultural context. International Journal of Comparative Sociology, 43, 45-64. Drobni, S., Rlossfeld, H, & Rohwer, G. (1999). Dynamics of Women’s employment patterns over the family life course: A comparison of the United States and Germany. Journal of Marriage and the Family, 61, 133-146. Drobni, S., & Treas, J. (2006, Aug). “Household arrangements under welfare state regimes: Employment and the division of household labour in USA, Germany, and Finland”. Paper presented at the annual meeting of the American Sociological Association, Montreal Convention Center, Montreal, Quebec, Canada. Retrieved 2008-05-16 from http://www.allacademic.comlmetalp104653_index.html Esping-Andersen, G. (1990). The three worlds of welfare capitalism. Cambridge: Polity Press. Esping-Andersen, G. (1999). Socialfoundations ofpostindustrial economics. Oxford: Oxford University Press. 75 Estes, S. B., Noonan, M. C., & Maume, D. J. (2007), Is work-family policy use related to the gendered division of housework? Journal of Family and Economic Issues, 28, 527-545. Eurostats (2007). Female employment rate. Retrieved July 31, 2008 from http://epp.eurostat.ec.europa.eulportal/page?_pageid= 1996,39 140985&_dad=portal&_sch ema=PORTAL&screen=detailref&language=en&product=STRINDEMPLOI&root=ST RIND_EMPLOIJemploiIemO12 Evertsson, M., & Nermo, M. (2004). Dependence within families and the division of labour: Comparing Sweden and the United States. Journal ofMarriage and the Family, 66, 1272-1286. Evertsson, M., & Nermo, M. (2007). Changing resources and the division of housework: A longitudinal study of Swedish couples. European Sociological Review, 23, 455-470. Ferree, M. M. (1991). The gender division of labour in two-earner marriages. Journal ofFamily Issues, 12, 158-179. Folbre, N., & Hartmann, H. I. (1989). The persistence of patriarchal capitalism. Rethinking Marxism, 2, 90-96. Fuwa, M. (2004). Macro-level gender inequality and the division of household labour in 22 countries. American sociological Review, 69, 751-767. Fullerton Jr., H. N. (1999). Labour force participation: 75 years of change, 1950-98 and 1998- 2025. Monthly Labour Review, 12, 3-12. Gauthier, A. (1996). The state and the family. Oxford: Clarendon Press. Gershuny, J., & Robinson, J. P. (1988). Historical changes in the household division of labour. Demography, 25, 537-552. Hank, K., & Jurges, H. (2007). Gender and the division of household labour in older couples. Journal ofFamily Issues, 28, 399-421. Hersch, J. (1991). The impact of nonmarket work on market wages. The American Economic Review, 81, 157-160. Hiller, D. V. (1984). Power dependence and division of family work. Sex Roles, 10, 1003-10 19. Hochschild, A. R., & Machung, A. (1989). The second shift: Working parents and the revolution at home. New York: VikingIPenguin. Huber, J., & Spitze, G. (1983). Sex stratification: Children, housework, and jobs. New York: Academic. 76 Ishii-Kuntz, M., & Coltrane, S. (1992). Predicting the sharing of household labour: Are parenting and housework distinct? Sociological Perspectives, 35, 629-647. Jackson, S. (1992). Towards a historical sociology of housework: A materialist feminist analysis. Women’s Studies International Forum 15, 153-172. Kalleberg, A. L., & Rosenfeld, R. A. (1990). Work in the family and in the labour market: A cross-national, reciprocal analysis. Journal ofMarriage and the Family, 52, 23 1-246. Kohn, M. L. (1987). Cross-national research as an analytic strategy. American Sociological Review, 52, 713-73 1. Lennon, M. C., & Rosenfield, S. (1994). Relative fairness and the division of housework: The importance of options. American Journal of sociology, 100, 506-53 1. Marini, M. M., & Shelton, B. A. (1993). Measuring household work: Recent experience in the United States. Social Science Research, 22, 361-382. Mosesdottir, 2000. (2000). Pathways towards the dual breadwinner model: the role of the Swedish, German, and the American States. International Review of Sociology, 10, 189- 205. Nichols, S. Y., & Metzen, E. J. (1982). Impact of wife’s employment upon husband’s housework. Journal ofFamily Issues, 3, 199-2 16. Nock, S. L. (1979). The family life cycle: Empirical or conceptual tool? Journal ofMarriage and the Family, 41, 15-26. Noonan, M. C., Estes, S. B., & Glass, J. L. (2007). Do workplace flexibility policies influence time spent in domestic labour? Journal of Family Issues, 28, 263-288. Orloff, A. S. (1993). Gender and the social rights of citizenship: The comparative analysis of gender relations and welfare states. American Sociological Review, 58, 303-328. Orloff, A. S. (1996). Gender in the welfare states. Annual Review ofSociology, 22, 51-78. Pfau-Effinger, B. (2004). Culture and welfare state policies: Reflections on a complex interrelation. Journal of Social Policy, 34, 3-20. Press, J. E., & Townsley, E. (1998). Wives and husbands’ housework reporting: Gender, class, and social desirability. Gender and Society, 12, 188-2 18. Presser, H. B. (1994). Employment schedules among dual-earner spouses and the division of household labour by gender. American Sociological Review, 59, 348-364. 77 Pina, D. L., & Bengston, V. L. (1987). Time used for household work: A study of school-age children from single-parent, two parent and one-earner and two-earner families. Journal of Family Issues, 8, 212-225. Pina, D. L., & Bengston, B. L. (1995). Division of household labour and the well-being of retirement-aged wives. The Gerontologist, 35, 308-317. Rexroat, C., & Shehan, C. (1987). The family life cycle and spouses’ time in housework. Journal of Marriage and the Family, 49, 737-750. Robinson, J., & Godbey, G. (1997). Time for life. University Park, PA: Pennsylvania State University Press. Robinson, J., & Spitze, G. (1992). Whistle while you work? The effect of household task performance on women’s and men’s well-being. Social Science Quarterly, 73, 844-86 1. Ross, C. E. (1987). The division of labour at home. Social Forces, 65, 8 16-833. S ainsbury, D. (1994). Women’s and men’s social rights: Gendering dimensions of welfare states. D. Sainbury (Ed.), Gendering welfare states (pp. 150-169). London: Sage Publications. Sanchez, L., & Thomson, E. (1997). Becoming mothers and fathers. Parenthood, gender, and division of labour. Gender & Society, 11, 747-772. Shelton, B. A. (1992). Women, men and time: Gender differences in paid work, housework, and leisure. Westport, CT: Greenwood. Shelton, B. A., & John, D. (1993). Does marital status make a difference: Journal of Family Issues, 14, 40 1-420. Shelton, B. A., & John, D. (1996). The division of household labour. Annual Review of Sociology, 22, 299-322. South, S. J., & Spitze, G. (1994). Housework in marital and nonmarital households. American Sociological Review, 59, 327-347. Steven’s, D., Kiger, G., & Riley, P. (2001). Working hard and hardly working: Domestic labour and marital satisfaction among dual-earner couples. Journal ofMarriage and the Family, 63, 514-526. Stafford, R., Backman, E., & Dibona, P. (1977). The division of labour among cohabiting and married couples. Journal ofMarriage and the Family, 39, 43-57. 78 Stier, H., & Lewin-Epstein, N. (2001). Welfare regimes, family-supportive policies, and women’s employment along the life-course. American Journal of Sociology, 106, 173 1- 1760. Suitor, J. J. (1991). Marital quality and satisfaction with the division of household labour across the family life cycle. Journal ofMarriage and the Family, 53, 22 1-230. Thompson, L., & Walker, A. J. (1989). Gender in families: Women and men in marriage, work, and parenthood. Journal ofMarriage and the Family, 51, 845-871. Twiggs, J. E., McQuillan, J., & Ferree, M. M. (1999). Meaning and measurement: Reconceptualization measure of the division of household labour. Journal ofMarriage and the Family, 61, 7 12-724. Waite, L., & Goldscheider, F. K. (1992). Work in the home: The productive context of family relationships. S. J. South, & S. E. Tolnay (Eds.), The changing American family (pp. 267- 299). Boulder, CO: Westview. West, C., & Zimmerman, D. H. (1987). Doing gender. Gender and Society, 1:125-51. Wilkie, J. R., Ferree, M. M., & Ratcliff, K. S. (1998). Gender and fairness: Marital satisfaction in two-earner couples. Journal ofMarriage and the Family, 60, 577-594. 79 A PP EN D IX A D es cr ip tiv e st at is tic s o fS w ed en ,G er m an y a n d th e U ni te d St at es by ge nd er .M ar rie d pe rs on s w ith pr es ch oo lo r sc ho ol -a ge c hi ld re n in th e ho us eh ol d. M ea ns ,p er ce nt ag es ,a n d st an da rd e rr o rs in pa re nt he se s. Pr es ch oo l Sc ho ol -a ge Sw ed en G er m an y U. S. Sw ed en G er m an y U. S. . M en W om en M en W om en M en W om en M en W om en M en W om en M en W om en V ar ia bl es M ea ns H ou se w or k hr s (w kly ) 6. 12 15 .67 5. 04 19 .33 6. 59 18 .41 10 .0 2 14 .41 6. 56 19 .9 5 7. 89 16 .76 (2. 99 ) (7. 97 ) (2. 12 ) (13 .79 ) (3. 99 ) (15 .81 ) (8. 48 ) (7. 79 ) (5. 15 ) (11 .26 ) (9. 71 ) (12 .87 ) Ti m e av ai la bi lit y 8. 54 8. 22 12 .5 5 35 .0 0 16 .0 0 16 .30 5. 09 5. 10 16 .18 9. 15 13 .0 0 8.9 1 (10 .31 ) (11 .24 ) (14 .21 ) (15 .25 ) (15 .58 ) (17 .94 ) (11 .34 ) (11 .85 ) (16 .47 ) (17 .05 ) (15 .95 ) (14 .26 ) Re la tiv e re so u rc e 3. 19 3. 53 2. 13 2. 19 3.3 1 2. 65 3. 30 3. 44 2. 51 2. 88 2. 30 3. 22 (1. 11 ) (1. 39 ) (1. 25 ) (0. 88 ) (1. 70 ) (1. 46 ) (1. 07 ) (1. 34 ) (0. 99 ) (1. 43 ) (1. 15 ) (1. 58 ) G en de ri de ol og y 4. 08 3. 90 3.2 1 3. 25 3. 38 3. 55 3. 57 3.7 1 3. 23 3. 52 3.2 1 3. 55 (0. 60 ) (0. 78 ) (0. 82 ) (0. 95 ) (0. 95 ) (1. 01 ) (0. 76 ) (0. 88 ) (0. 78 ) (0. 80 ) (1. 04 ) (0. 86 ) A ge 38 .1 2 33 .53 34 .5 4 31 .5 2 36 .8 2 30 .7 4 47 .6 7 43 .5 3 43 .2 3 40 .7 1 44 .3 1 42 .1 7 (6. 93 ) (4. 75 ) (4. 90 ) (4. 35 ) (8. 23 ) (7. 19 ) (5. 49 ) (6. 29 ) (6. 74 ) (5. 40 ) (9. 78 ) (6. 24 ) Pe rc en ta ge s W ith ki ds 7. 17 7. 39 7. 12 8. 24 8. 90 14 .2 3 25 .5 6 28 .7 9 27 .8 9 24 .4 3 20 .4 2 22 .9 9 W ith co lle ge 25 .0 0 52 .6 0 12 .50 13 .8 0 35 .3 0 23 .1 0 22 .8 0 31 .1 0 19 .10 15 .10 33 .3 0 17 .50 SP w ith co lle ge 31 .2 0 31 .6 0 16 .70 10 .3 0 17 .60 35 .9 0 33 .3 0 23 .0 0 9. 60 24 .4 0 20 .5 0 30 .2 0 Se lf- em pl oy ed 18 .8 0 0. 00 12 .50 3. 40 11 .8 0 15 .40 12 .3 0 9. 50 7. 40 7. 00 17 .90 12 .7 0 N 16 19 24 29 17 39 57 74 94 86 39 63 N ot e: Ti m e a v a ila bi lit y sc o re ra n ge s fr om - 95 ,h us ba nd ha vi ng th e m o st tim e a v a ila bl e fo rh ou se w or k, to 95 ,w ife ha vi ng th e m o st tim e a v a ila bl e fo rh ou se w or k. R el at iv e re so u rc e sc o re ra n ge s fr om 1, w he re th e hu sb an d is th e so le br ea dw in ne r, to 7, w he re th e w ife is th e so le br ea dw in ne r. G en de r id eo lo gy sc o re ra n ge s fr om 1, m o st tr ad iti on al ge nd er a tt itu de s, to 5, m o st e ga lit ar ia n ge nd er at tit ud es . 00 0 APPENDIX B Regression results of the mediation pathway “c’ “: the relationships between life stages and housework hours by gender while accounting for the mediators, and the mediation pathway “b”: the relationships between the mediations and housework hours by gender. Unstandardized coefficients and standard errors in parentheses. Housework Hours Sweden Variables Men Women Independent Variables Preschool School-age Controls Age With college Spouse with college Self-employed Mediators Time availability Relative resources Gender ideology 0.502 2.307 1.418 4.734 6.109* 5.829* (3.486) (3.908) (3.917) (2.530) (2.425) (2.395) 3.421 3.405 3.520 1.955 2.893 2.486 (1.926) (2.350) (2.327) (1.617) (1.540) (1.533) 0.180 0.273* 0.305* 0.135 0.185* 0.165* (0.100) (0.092) (0.092) (0.078) (0.063) (0.063) 0.658 -0.650 -1.392 -1.280 2.462* -2.022 (2.599) (2.769) (2.740) (1.336) (1.233) (1.226) 0.050 0.470 -0.142 -0.400 -0.297 -0.494 (2.354) (2.685) (2.669) (1.341) (1.236) (1.231) 0.906 -0.955 -0.845 2.674 1.031 0.736 (2.157) (2.410) (2.394) (2.114) (2.001) (1.987) -0.030 0.128* (0.074) (0.054) 0.352 0.132 (0.914) (0.391) 2.409 -1.214 (1.369) (0.679) Note: *jndjcatesp <0.05; **jndjcatesp <0.001 81 APPENDIX B Regression results of the mediation pathway “c’ “: the relationships between life stages and housework hours by gender while accounting for the mediators, and the mediation pathway “b”: the relationships between the mediations and housework hours by gender. Unstandardized coefficients and standard errors in parentheses. Housework Hours Germany Variables Men Women Independent Variables Preschool -0.095 0.285 -0.015 -7.777 -1.847 -1.891 (1.546) (1.699) (1.669) (4.673) (3.246) (2.969) School-age 1.005 0.263 0.027 1.007 -1.210 -0.290 (0.927) (1.009) (1.001) (1.836) (1.993) (1.882) Controls Age 0.052 0.109* 0.126* -0.058 0.089 0.068 (0.053) (0.038) (0.038) (0.118) (0.073) (0.069) With college 1.002 1.409 0.805 -3.049 -3.539 -2.887 (1.075) (1.022) (1.009) (2.397) (2.526) (2.389) Spouse with college -1.081 -1.354 -1.196 -1.276 -2.015 -0.890 (1.320) (1.391) (1.370) (2.156) (1.986) (1.839) Self-employed 0.449 -0.824 -0.293 -0.706 6.594* 6.710* (1.225) (1.148) (1.118) (2.350) (2.600) (2.450) Mediators Time availability -0.044 0.161* (0.027) (0.047) Relative resources 0.760* -0.272 (0.356) (0.504) Gender ideology 1.591* 4.512** (0.487) (0.814) Note: *jndjcatesp <0.05; **jndjcatesp <0.001 82 APPENDIX B Regression results of the mediation pathway “c’ “: the relationships between life stages and housework hours by gender while accounting for the mediators, and the mediation pathway “b”: the relationships between the mediations and housework hours by gender. Unstandardized coefficients and standard errors in parentheses. Housework Hours U.S. Men WomenVariables Independent Variables Preschool School-age Controls Age With college Spouse with college Self-employed Mediators Time availability Relative resources Gender ideology -0.537 -6.158 -5.651 2.655 6.364* 6.836* (5.636) (5.423) (5.220) (2.577) (2.525) (2.505) -3.130 -3.828 -4.449 3.410 4.308* 4559* (4.236) (3.666) (3.605) (1.742) (1.843) (1.836) 0.274 0.050 0.073 -0.056 0.032 0.017 (0.159) (0.105) (0.103) (0.081) (0.060) (0.061) -5.403 -5.167 -5.649 -2.133 -0.739 -0.622 (4.874) (3.746) (3.708) (1.734) (1.876) (1.874) 7.361 3.869 4.579 -0.511 -1.868 -1.250 (4.595) (3.889) (3.848) (1.713) (1.793) (1.752) -7.254 -4.261 -5.545 4.172 1.249 0.826 (4.383) (3.870) (3.969) (2.136) (2.279) (2.243) -0.120 0.099* (0.106) (0.048) 0.821 -0.754 (1.131) (0.461) -1.501 -0.664 (1.486) (0.779) Note: *jndjcatesp <0.05; **jndjcatesp <0.001 83 APPENDIX C Regression results of the mediation pathway “a”: the relationships between life stages and mediators. Unstandardized coefficients and standard errors in parentheses. Time Availability Sweden Germany U.S. Variables Men Women Men Women Men Women Independent Variables Preschool -2.142 -2.939 -3.135 24.421* -0.411 7.703 (4.078) (3.737) (4.556) (8.730) (5.882) (4.644) School-age -2.479 -3.871 2.456 -0.261 -2.832 2.912 (2.242) (2.294) (2.707) (3.529) (4.411) (3.229) Controls Age 0.300* 0.25O* -0.091 0.012 -0.219 -0.258 (0.113) (0.111) (0.154) (0.228) (0.164) (0.144) With college 0.457 -3.206 -1.996 -5.293 10.197* 0.894 (3.018) (1.949) (3.205) (4.600) (4.978) (3.200) Spouse with college -1.532 1.843 4.193 6.617 -5.737 3.604 (2.696) (1.989) (3.988) (4.114) (4.760) (3.142) Self-employed 4.890 -1.375 12.813** l3.111* -5.728 3.704 (2.484) (3.101) (3.228) (4.370) (4.448) (3.894) Note: *jndjcatesp < 0.05; **jndjcatesp < 0.001 84 APPENDIX C Regression results of the mediation pathway “a”: the relationships between life stages and mediators. Unstandardized coefficients and standard errors in parentheses. Relative Resources Sweden Germany U.S. Variables Men Women Men Women Men Women Independent Variables Preschool 0.087 -0.414 O.627* 1.O19* 0.623 0.785* (0.318) (0.447) (0.271) (0.363) (0.377) (0.346) School-age 0.182 -0.264 -0.063 -0.189 -0.385 -0.352 (0.191) (0.273) (0.157) (0.227) (0.256) (0.254) Controls Age 0.007 -0.004 0.003 0.004 0.015* 0.005 (0.007) (0.011) (0.006) (0.008) (0.007) (0.008) With college 0.456* 0.270 0.384* 0.679* -0.306 0.064 (0.217) (0.224) (0.162) (0.288) (0.262) (0.259) Spouse with college 0.5 10* -0.179 0.404 0.733* 0.562* 0.722* (0.2 10) (0.226) (0.226) (0.225) (0.269) (0.244) Self-employed 0.193 0.320 0.264 0.376 0.120 0.280 (0.191) (0.355) (0.174) (0.299) (0.271) (0.315) Note: *jndjcatesp < 0.05; **jndjcatesp < 0.001 85 APPENDIX C Regression results of the mediation pathway “a”: the relationships between life stages and mediators. Unstandardized coefficients and standard errors in parentheses. Gender Ideology Sweden Germany U.S. Variables Men Women Men Women Men Women Independent Variables Preschool 0.394 -0.221 0.014 -0.158 0.08 -0.184 (0.206) (0.249) (0.199) (0.203) (0.266) (0.202) School-age -0.012 -0.192 0.132 0.200 -0.046 -0.063 (0.124) (0.152) (0.117) (0.130) (0.182) (0.149) Controls Age O.O12* O.017* 0.009* -0.006 -0.007 O.013* (0.005) (0.006) (0.004) (0.005) (0.005) (0.005) With college 0.223 0.303* 0.207 0.240 0.092 0.213 (0.142) (0.124) (0.120) (0.166) (0.189) (0.152) Spouse with college 0.328* 0.119 0.100 0.172 0.313 0.122 (0.137) (0.126) (0.168) (0.127) (0.195) (0.142) Self-employed -0.050 -0.371 -0.004 0.121 0.640* -0.043 (0.124) (0.197) (0.127) (0.171) (0.193) (0.181) Note: *jndjcatesp < 0.05; **jndjcatesp < 0.001 86"@en ; edm:hasType "Thesis/Dissertation"@en ; vivo:dateIssued "2008-11"@en ; edm:isShownAt "10.14288/1.0067009"@en ; dcterms:language "eng"@en ; ns0:degreeDiscipline "Social Work"@en ; edm:provider "Vancouver : University of British Columbia Library"@en ; dcterms:publisher "University of British Columbia"@en ; dcterms:rights "Attribution-NonCommercial-NoDerivatives 4.0 International"@en ; ns0:rightsURI "http://creativecommons.org/licenses/by-nc-nd/4.0/"@en ; ns0:scholarLevel "Graduate"@en ; dcterms:title "A comparative study of division of household labour across family life stages in Sweden, Germany, and the United States"@en ; dcterms:type "Text"@en ; ns0:identifierURI "http://hdl.handle.net/2429/5524"@en .