Time for Each Other: Work and Family Constraints Among Couples

Abstract

Little is known about couples’ shared time and how actual time spent together is associated with well-being. In this study, the authors investigated how work and family demands are related to couples’ shared time (total and exclusive) and individual well-being (happiness, meaningfulness, and stress) when with one's spouse. They used individual-level data from the 2003–2010 American Time Use Survey (N = 46,883), including the 2010 Well-Being Module. The results indicated that individuals in full-time working dual-earner couples spend similar amounts of time together as individuals in traditional breadwinner–homemaker arrangements on weekdays after accounting for daily work demands. The findings also show that parents share significantly less total and exclusive spousal time together than nonparents, though there is considerable variation among parents by age of the youngest child. Of significance is that individuals experience greater happiness and meaning and less stress during time spent with a spouse opposed to time spent apart.

There is evidence that couples try to coordinate their schedules (e.g., Hamermesh, 2002; Sullivan, 1996), that shared time is important for marital well-being (e.g., Daly, 2001; Gager & Sanchez, 2003; Milkie & Peltola, 1999), and that the quality of marital relationships is associated with the quality of parent–child relationships (e.g., Jekielek, 1998; Malinen et al., 2010). Yet the focus of most research on time spent with others has focused on parent–child time (e.g., Bianchi, 2000; Sayer, Bianchi, & Robinson, 2004); the examination of the time spouses spend together has been scarce. In this study, we addressed two specific questions: (a) How are the competing demands of work and family associated with couples’ shared time? and (b) How is couples’ shared time tied to well-being—that is, happiness, meaning, and stress?

Although couples’ joint engagement in daily activities is positively associated with marital well-being (Amato, Booth, Johnson, & Rogers, 2007; Booth, Johnson, White, & Edwards, 1985, 1986; Crawford, Houts, Huston, & George, 2002; Gager & Sanchez, 2003; Hill, 1988; White, 1983; Zuo, 1992), work and family demands, as well as the competing devotions they require (Blair-Loy, 2003), constrain the time couples can spend together. Indeed, studies suggest that time with a spouse may be sacrificed to manage work and parenting demands (Bianchi, Robinson, & Milkie, 2006; Claxton & Perry-Jenkins, 2008; White 1983; Wight, Raley, & Bianchi, 2008). Paid workers have reported that time with a spouse is limited, which affects their marital relationships (Bianchi et al., 2006; Nomaguchi, Milkie, & Bianchi, 2005; Roxburgh, 2006). Women's time with children, on the other hand, has remained stable over the past 50 years despite women's greater participation and hours spent in the labor market (Bianchi, 2000; Bianchi & Milkie, 2010; Sayer et al., 2004), yet parents’ marital satisfaction is lower than that of nonparents (e.g., Dew & Wilcox, 2011; Twenge et al., 2003). Therefore, a better understanding of couples’ shared time and how they feel about it is of great importance for the well-being of many Americans.

Despite the challenges contemporary couples face in finding time for one another, scholars have devoted only limited attention to the systematic analysis of couples’ shared time. The majority of the research on marital interaction is based on responses to stylized questions in which an individual reports how often or how long he or she did something with a spouse during a given reference period, typically a week, month, or year. Evidence suggests that time diary data are a better mechanism for capturing information on time spent with a spouse (Hamermesh, 2002, 2005), yet only a limited body of research has used this type of data to examine marital interaction. Previous research on couples’ shared time using time diary data has investigated leisure activities (Barnet-Verzat, Pailhé, & Solaz, 2010; Voorpostel, van der Lippe, & Gershuny, 2009), time spent alone with a spouse in any kind of activity (Dew, 2009), and total shared time as well as shared time in different types of activities (Kingston & Nock, 1987; Mansour & McKinnish, 2014). On the basis of our review of the literature, only two studies—Wight et al. (2008) and Bianchi et al. (2006)—have investigated both total time spent with a spouse and time alone with a spouse. Wight and colleagues did so in the context of nonstandard work arrangements, and Bianchi and colleagues conducted bivariate analyses of differences in shared time among working parents between 1975 and 2000.

In this study, we investigated the time couples spend together and associations with work and family demands, thereby contributing a thorough investigation to a sparse literature on an important topic. We build on a somewhat dated (Kingston & Nock, 1987) and limited set of knowledge about couples’ shared time (Barnet-Verzat et al., 2010; Bianchi et al., 2006; Dew, 2009; Mansour & McKinnish, 2014; Voorpostel et al., 2009; Wight et al., 2008) by bringing recent evidence from a large, nationally representative data set and by examining couples’ time spent alone together and their time spent together overall. Given the relevance of both work and family constraints in shaping time pressures and the competition between work and family in individuals’ daily lives, we investigated variation in couples’ shared time by work and life stage. We also examined how the actual experience of shared time is related to emotional well-being, using data from a 2010 module of the American Time Use Survey (ATUS; Hofferth, Flood, & Sobek, 2013), in which respondents provided momentary assessments of well-being during three randomly selected activities throughout the day. This research contributes to a rich literature that considers the relationship between global measures of marital interaction and marital well-being, in which well-being is typically divorced from specific time spent with a spouse.

Theoretical Perspectives and Previous Research

Work–family conflict assumes interdependency between work and family spheres and that conflict arises from incompatibility between demands in these domains (e.g., Eby, Casper, Lockwood, Bordeaux, & Brinley, 2005; Greenhaus & Beutell, 1985). One form of work–family conflict, time-based conflict, occurs when time demands required by one role make it difficult to meet the expectations of another role given the scarcity of time. The rise in dual-earner couples, especially those with children, means that the work–family interface may be a source of tension in individual lives and may influence how individuals and couples organize their lives after allocating time to paid work and parenting (Huston & Vangelisti, 1995).

Because work and family are “greedy” institutions (Coser & Coser, 1974), they are often in conflict. Both work and family have high demands and intensive time commitments (Moen & Roehling, 2005; Williams, 2000), requiring “devotion” from participants (Blair-Loy, 2003), and the blurring boundaries between work and home (Moen, Kelly, & Lam, 2013), forcing individuals to make constrained choices about how to allocate time and energy (Bird & Rieker, 2008; Jacobs & Gerson, 1998). At the same time, research on “intensive parenting” (Hays, 1996), “family devotion” (Blair-Loy, 2003), and “concerted cultivation” (Lareau, 2003) indicates heightened parenting demands. In short, the work–family conflict literature suggests that time-based conflict may be especially salient for couples with high work and parenting demands, thereby limiting the time spouses have available to allocate to one another. To better understand the implications of time demands on individual well-being, we need to examine variation among individuals with different responsibilities and how individuals feel when they are with their spouses.

Work, Parenting, and Couples’ Shared Time

Research on marital interaction shows that paid work limits the time couples spend together. Both husbands’ and wives’ employment—especially long work hours—are negatively associated with marital interaction (e.g., Amato et al., 2007; Booth, Johnson, White, & Edwards, 1984; Roeters & Treas, 2011; White, 1983; but see Roeters & Treas, 2011). From the time use literature we know that individuals in dual-earner couples spend less time with their spouse than individuals in single-earner couples (Kingston & Nock, 1987; Presser, 2000; Voorpostel et al., 2009) as do individuals in couples in which one member works on the interview (i.e., diary) day (Barnet-Verzat et al., 2010; Dew, 2009; Glorieux, Minnen, & van Tienoven, 2011; Polivka, 2008; Wight et al., 2008) and in couples with large differences in paid work hours (Mansour & McKinnish, 2014). For example, Kingston and Nock (1987) drew on a small sample of couples from 1981 and showed that dual-earner couples spend 30 minutes less in total spousal time than single-earner couples. In their comparison of individuals who worked standard versus nonstandard work hours, using the ATUS, Wight and colleagues (2008) considered only weekdays when respondents worked and found that evening workers spent less time with a spouse (both in total and alone with a spouse) than day workers. In an analysis of change between 1975 and 2003, Dew (2009) showed that the time spouses spend alone together was negatively associated with minutes spent in paid work on the diary day. Whereas evidence suggests that the relationship between paid work and couples’ shared time—either total time or joint leisure time—is negative, prior investigations are dated, have considered specific subsamples, and/or have analyzed leisure or time spent alone as a couple.

By conducting the present study we extend this literature by systematically examining how work constraints, operationalized in a variety of ways, are related to couples’ total shared time as well as time spent exclusively with one's spouse. The research cited above informed our first hypothesis:

Hypothesis 1: Greater work demands—both in terms of the respondent's and spouse's employment status as well as whether or not the respondent worked on the diary day—are associated with less total shared time with a spouse and less time alone with a spouse.

Although paid work is a major factor influencing how people spend their time, especially time with a spouse, parenting demands are another structural factor negatively associated with marital interaction (e.g., Helms-Erickson, 2001; Zuo, 1992). The literature consistently shows that parents spend less time with one another in leisure—and more generally—than nonparents (Barnet-Verzat et al., 2010; Dew, 2009; Glorieux et al., 2011; Hill, 1988; Mansour & McKinnish, 2014; Voorpostel et al., 2009). The transition to parenthood initially reduces leisure shared with a spouse (Claxton & Perry-Jenkins, 2008) as couples’ shared time shifts from leisure as a couple to leisure as a family (Huston & Vangelisti, 1995) and to more instrumental tasks following the birth of a child (Clements & Markman, 1996; Huston, McHale, & Crouter, 1986). This research indicates that the nature of parents’ shared time is different than that of nonparents, suggesting the importance of examining couples’ shared time more broadly as opposed to leisure specifically. This led us to our second hypothesis:

Hypothesis 2: Parents spend less time alone together and less time together in general than nonparents.

Additional research on how children shape time with a spouse has considered variation by the age of the youngest child. Compared to parents of younger children, parents of older children spend greater shares of time in leisure (Bittman & Wajcman, 2000; Voorpostel et al., 2009) and more time together without children (Barnet-Verzat et al., 2010; Kalmijn & Bernasco, 2001; Roeters & Treas, 2011). Differentiating between children of different ages makes sense both because children may demand less time from parents at older ages and because relationships between children's age and parents’ shared time are not necessarily linear (Bittman & Wajcman, 2000; Kalmijn & Bernasco, 2001; Roeters & Treas, 2011; Voorpostel et al., 2009). Although Kingston and Nock (1987) found no difference between couples with and without preschoolers in total time spent with a spouse outside of child care, most of the extant literature suggests that younger children (compared to older children) are negatively associated with couples’ shared time, suggesting our third hypothesis:

Hypothesis 3: Parents of younger children spend less time together than parents of older children.

Couples’ Shared Time and Well-Being

The bulk of the literature that has assessed married couples’ well-being relied on global assessments, asking, for example, how happy or satisfied people are with their marriages in general. Research shows a positive relationship between the global measure of marital interaction or frequency of shared time and marital stability (Booth et al., 1985, 1986; Hill, 1988), marital happiness (White, 1983; Zuo, 1992), and marital satisfaction (Amato et al., 2007; Crawford et al., 2002). By contrast, momentary assessments of well-being in which feelings are tied to specific activities tend to be more reliable and less biased (Kahneman & Krueger, 2006). The 2010 ATUS well-being data allowed us to connect subjective assessments of time to reports about the presence of a spouse, making ours the first of this kind since Sullivan's (1996) investigation into the enjoyment of time shared with a spouse.

Although most people feel that they do not have enough time with their spouses (Bianchi et al., 2006; Nomaguchi et al., 2005; Roxburgh, 2006), we do not know whether actually spending time together is beneficial to individual well-being. To date, data available to examine this topic have been extremely limited. To our knowledge, the only existing study that considered momentary well-being in activities with spouses used a small sample of British couple-level data from 1986 and found that spouses try to coordinate activities and that individuals find greater enjoyment in activities done with a spouse (Sullivan, 1996). Previous research on relationships between shared time and well-being based on both global and momentary assessments suggested our fourth hypothesis:

Hypothesis 4: Individuals are better off—that is, they experience greater happiness and meaning and less stress—during time spent with a spouse compared to time spent apart.

Couples’ Shared Time and Gender

Husbands’ and wives’ accounts of their marriages are often different, leading Bernard (1982) to suggest that every marriage has two marriages—“his” and “hers”—with men faring better than women on most dimensions. A gender perspective suggests that the expectations and demands of roles are different for men and women (West & Zimmerman, 1987). To be specific, expectations that workers will be unencumbered from caregiving responsibilities (Williams, 2000) and that parents—in particular, mothers—will invest heavily in their children's development (Hays, 1996; Lareau, 2003) together make the work–family balancing act problematic for anyone with both work and caregiving responsibilities. Indeed, research has found that women feel greater time pressure than men, especially when they hold multiple roles (Mattingly & Sayer, 2006; Roxburgh, 2002).

Actual differences in couples’ shared time are unexpected because husbands and wives, in theory, spend the same amount of time with one another, yet gender differences in time spent together have been identified previously, with women reporting less shared time with their husbands than men report with their wives (Amato et al., 2007; Claxton & Perry-Jenkins, 2008; Dew, 2009; Freedman, Stafford, Schwarz, Conrad, & Cornman, 2012; Gager & Sanchez, 2003; Mansour & McKinnis, 2014; Voorpostel et al., 2009; Wight et al., 2008; but see Kingston & Nock, 1987). The most compelling evidence about gender differences in reporting shared time is based on couple-level data from the Panel Study of Income Dynamics and indicates that, overall, women report spending about 20 minutes less per day with their husbands than men report spending with their wives, even when diaries show agreement between husbands’ and wives’ activities (Freedman et al., 2012). This implies different interpretations of what counts as shared time as opposed to differences in actual time spent together. Evidence shows that husbands would like to have more time with their wives, but women would like more quality time with their husbands rather than simply more time together (Roxburgh, 2006). Given a gender perspective and previous research on gendered interpretations of shared time, we expected that women will report less shared time with a spouse than men.

Data and Method

We use integrated ATUS data (Hofferth, Flood, & Sobek, 2013) to examine the time couples spend together, to analyze relationships between shared time and work and family demands (time-together analysis), and to compare subjective well-being during time couples spend together and apart (well-being analysis). The ATUS is a time diary study of a nationally representative sample of Americans. ATUS data are collected using a computer-assisted telephone interview during which the respondents recall and report the activities in which they have engaged over a 24-hour period—from 4:00 a.m. yesterday until 4:00 a.m. of the reporting day—as well as where, when, and with whom activities were done. Data are collected all days of the week, and weekends are oversampled. Sample weights correct for the survey design such that aggregating across different days of the week results in a representative picture of average time use among the population. Our results are based on pooled cross-sections from 2003 to 2010.

The Current Population Survey (CPS; http://www.census.gov/cps/), a monthly household survey of the civilian, noninstitutionalized population, is the sampling frame for the ATUS. Once a household has completed participation in the CPS it is eligible for inclusion in the ATUS sample 2–5 months later. For households selected to participate in the ATUS, one individual age 15 or older reports his or her activities over one 24-hour period. Although our data are from individuals, we treated the couple as the focal unit, including in our sample only individuals in couples in which at least one member is working for pay because these are the couples who will feel the constraints of both paid work and family demands. We fully acknowledge, however, that both single individuals and individuals in couples without an earner face other types of work and/or family demands. Our emphasis on the couple reinforces the idea that choices about how and with whom to spend time are made in the context of gendered roles and relationships (e.g., Moen, 2001).

In this article we present two sets of analyses, and we describe the measures, analytic strategy, and results in the sections that follow. The first analysis—the time-together analysis—focused on the time individuals spend with their spouse and relationships with paid work and family life stage demands. For the second analysis—the well-being analysis—we used individual reports of activity-level subjective well-being to assess how individuals feel about the time they spend with their spouse. A particular benefit of these activity-level measures is that they tie subjective well-being to specific activities and, in our case, to activities done with specific types of people. One limitation may be that the respondent must recall how he or she felt during the activities. Research in this area is very limited, although Kahneman et al. (2004) found little evidence of differences in recalled (like those used in this research) versus immediately provided assessments (via beeper studies) of subjective well-being.

The time-together analysis drew on a subsample of 2003–2010 ATUS respondents. The 2003–2010 ATUS data include daily diary entries of 112,038 civilians age 15 and older. Although the data may not typify any one respondent's daily activities, aggregations of the data are representative of the American population. We restricted our sample to married individuals with a spouse in the household at the time of the ATUS interview (N = 57,585) who were age 18–64 (N = 49,695) and who were in single- or dual-earner relationships (N = 46,883). In addition, we drew on linked CPS data to incorporate sociodemographic variables for ATUS respondents’ spouses, including race and education.

Analyses of time diary data typically focus on the total amount of time individuals spend in a given activity (e.g., work, leisure, sleep). The richness of the ATUS data extends beyond what people do, however, and can provide insight into patterns of social interaction, including marital interaction. Like other time use research, in our study we examined heterogeneity in how people spend time. In our time-together analysis we examined factors associated with both time with one's spouse as well as time with one's spouse and no one else.

For our well-being analysis, we used data from the 2010 well-being module of the ATUS, which was funded by the National Institute on Aging and collected momentary assessments of subjective well-being for up to three activities randomly selected from each respondent's time diary. The 2010 well-being module of the ATUS collected information from 12,829 respondents (38,085 activities). Restricting the sample to married individuals ages 18–64 in single- or dual-earner relationships yielded a sample of 18,487 activities. For each of up to three activities per person, respondents were asked to report how sad, tired, and happy they were during the activity; how much pain and stress they felt during the activity; and how meaningful the activity was to them. Our focus was on the happiness, stress, and meaningfulness of activities. Each momentary assessment is measured on a 7-point scale (0–6), with a 0 indicating that the respondent was not happy or did not experience stress during the activity and a 6 indicating that the respondent was very happy or very stressed during the activity. For meaningfulness, a 0 indicates that the activity was not very meaningful to the respondent and a 6 indicates that the activity was very meaningful.

Time-Together Analysis

Measures

We had two dependent variables for our time-together analysis. Total shared time was a continuous measure of the total minutes per day spent during non-work, non-sleep, and non-personal activities with one's spouse regardless of who else, if anyone, was present. Exclusive spousal time was a subset of total shared time, which includes the total minutes per day respondents spent with only their spouses. These measures did not include time spent working, sleeping, grooming, or in personal care because “with whom” information was not consistently collected for these activities in the ATUS during the entire 2003–2010 period.

Key independent variables

Couple-level work status combined single and dual-earner status with levels of employment. The four single-earner categories include (a) husband single earner, works full time; (b) wife single earner, works full time; (c) husband single earner, works part time; and (d) wife single earner, works part time. The four dual-earner categories include (a) husband and wife earners work full time (the reference category in the regression models), (b) husband works full time and wife works part time, (c) wife works full time and husband works part time, and (d) both work part time.

We controlled for two diary day work characteristics of the respondent. The workday measure indicated whether the respondent did any paid work on the diary day. Minutes in paid work indicated how many minutes the respondent worked for pay on the diary day.

Life stage, which was used to capture variation in the time demands of parents and nonparents, was coded into nine dichotomous variables based on the of the wife and the age of the respondent's youngest own child in the household. For individuals without children in the home, we differentiated between couples in which the wife was age 45 or younger and those in which the wife was over 45. The nine life stage categories include (a) no children and wife age 45 or under, (b) no children and wife over age 45, (c) youngest child in the household age 1 or under, (d) age 2, (e) ages 3–5, (f) ages 6–9 (reference), (g) ages 10–13, (h) ages 14–17, and (i) 18 or older.

Control variables

Age of the respondent was measured as a continuous variable, with respondents ranging from 18 to 64 years old. Race and education were coded identically for both husbands and wives. Race was coded as four dichotomous variables: (a) White, non-Hispanic (reference); (b) Black, non-Hispanic; (c) Other, non-Hispanic; and (d) Hispanic. Asians comprised 71% of the “Other, non-Hispanic” category in our sample. Education was coded into four dichotomous variables: (a) less than high school (the reference in the regression analyses), (b) high school degree, (c) some college, and (d) college degree or more. We also included controls for whether the diary day was a holiday or non-holiday (reference) and the year of ATUS participation as binary variables for 2003 (reference) to 2010.

Analytic Strategy

We first estimated the total time shared with one's spouse, on average, and exclusive spousal time (with one's spouse only). Weekdays and weekends were considered separately because of the ways in which the typical work week structures daily life. Descriptive analyses were followed by ordinary least squares (OLS) regression estimates of the relationship between couple-level work status, family life stage, and the daily time individuals spend with their spouse, as well as controls for individual and spousal characteristics. We used OLS regression despite having zeroes in our dependent variables (9% in total spousal time and 30% in exclusive spousal time) because recent evidence indicates that OLS produces less biased estimates than Tobit, especially as the number of zeroes in the dependent variable grows larger (Stewart, 2013).

Descriptive Results

Sample characteristics. A description of the full sample, as well as differences by gender and day of the week (weekday vs. weekend), is provided in Table 1. Dual earners with both husbands and wives working full time were the modal earner type among couples (44%). One quarter (26%) of the respondents were in a single-earner couple in which the husband works full time. The third most commonly represented type was that in which the husband works full time and the wife works part time (15%). The remaining 15% of the sample was distributed among part-time-earner families, female-breadwinner arrangements (7%), and couples in which women work full time and men work part time (2%). About one third of the respondents in the sample did not live with children; 12% of the sample was in a couple in which the wife is age 45 or under, and 22% of respondents were in a couple in which the wife is over age 45 (possibly with grown children). Parents of children age 2 represented 5% of the sample; the remaining parents in the sample were approximately evenly distributed, with roughly 10% each in the other categories based on the age of the youngest coresident child.

Table 1.

Weighted Means/Percentages of Couple-Level Characteristics, Work-Related Controls, Respondent and Spouse Demographic Characteristics, and Diary Day Characteristics (N = 46,883)

Variable		Weekdays					Weekends
	Couple-level characteristics
	Full sample	Men			Women		Men		Women
Work status
Single earner
Husband FT	26%	27%a			25%a		27%b		24%b
Wife FT	7%	7%a			8%a		6%b		7%b
Husband PT	2%	2%			2%		2%		2%
Wife PT	2%	1%a			2%a		1%b		2%b
Dual earner
Both FT	44%	44%			44%		44%		45%
Husband FT, Wife PT	15%	15%			16%		15%		15%
Wife FT, Husband PT	2%	2%a			3%a		2%		3%
Both PT	1%	1%			1%		1%		1%
Life stage
Nonparents
No children, wife ≤ 45	12%	12%			12%		12%		12%
No children, wife > 45	22%	22%a			24%a		21%b		23%b
Parents
Youngest child 1 or under	11%	12%			11%		11%		11%
Youngest child age 2	5%	5%			4%		5%		5%
Youngest child age 3–5	10%	10%a			10%a		11%		10%
Youngest child age 6–9	11%	11%			11%		11%		11%
Youngest child age 10–13	9%	9%			9%		9%		9%
Youngest child age 14–17	8%	8%			8%		9%		8%
Youngest child age 18+	11%	11%			11%		11%		11%
	Work-related controls
	Respondent	Spouse	Respondent	Spouse	Respondent	Spouse	Respondent	Spouse	Respondent	Spouse
Diary day work commitments
Engaged in paid work	59%		82%a,c		60%a,d		34%b,c		22%b,d
Minutes spent in paid work	270.19		420.86a,c		267.85a,d		112.52b,c		60.12b,d
	Demographic characteristics
	Respondent	Spouse	Respondent	Spouse	Respondent	Spouse	Respondent	Spouse	Respondent	Spouse
Age	43.29	43.63	44.03a	42.23a	42.68a	45.12a	43.83b	42.05b	42.48b	44.83b
Education
Less than high school	9%	10%	10%a	9%a	9%a	10%a	10%b	9%b	9%b	11%b
GED/HS degree	29%	28%	30%	28%	29%	28%	30%	27%	29%	28%
Some college	26%	26%	24%a	27%a	27%a	25%a	25%b	28%b	27%b	26%b
College/advanced degree	35%	35%	35%	35%	36%	36%	36%	36%	35%	35%
Race
White, non-Hispanic	73%	74%	73%	73%	74%	74%	72%b	72%b	74%b	74%b
Black, non-Hispanic	7%	7%	7%	6%	7%	7%	7%b	7%	6%b	7%
Other, non-Hispanic	5%	5%	5%b	6%	6%a	5%	5%	6%	6%	5%
Hispanic	14%	14%	15%	15%a	14%	14%a	15%	15%b	14%	14%b
Number of observations	46,883		10,831		10,987		12,395		12,670

Note. These data are based on the authors’ calculations from 2003–2010 American Time Use Survey (ATUS data) obtained from ATUS-X (Hofferth et al., 2013). The sample includes all married respondents between the ages of 18 and 64, with at least one spouse working. FT = full time; PT = part time.

^aCharacteristics of men who responded on weekdays were significantly different from those of women who responded on weekdays (p < .05).

^bCharacteristics of men who responded on weekends were significantly different from those of women who responded on weekends (p < .05).

^cCharacteristics of men who responded on weekdays were significantly different from men who responded on weekends (p < .05).

^dCharacteristics of women who responded on weekdays were significantly different from those of women who responded on weekends (p < .05).

Total and exclusive spousal time

The average number of minutes men and women spent together with their spouse on weekends and on weekdays, in both total and exclusive spousal time by work status and family life stage, is shown in Table 2. Total shared time was higher on weekends compared to weekdays. On weekdays, individuals spent just over 3 hours with a spouse (205 and 198 minutes according to men and women, respectively) and nearly 7 hours on weekend days (422 and 404 minutes for men and women, respectively). As would be expected, exclusive spousal time was lower, with individuals spending about 2 hours alone with a spouse on weekdays (116 and 112 minutes for men and women, respectively) and 3 hours on weekends (192 and 184 minutes for men and women, respectively).

Table 2.

Weighted Average of Total Shared Time and Exclusive Spousal Time on Diary Day (in Minutes), by Day of Week (N = 46,883)

Variable	Total shared time				Exclusive spousal time
	Weekday		Weekend		Weekday		Weekend
	Men	Women	Men	Women	Men	Women	Men	Women
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
Couple-level work status
Single earner	231	231	440a	423	127	123	190	184
Husband FT	215	220	441	428	107	104	171	163
Wife FT	262b	230	438a	403	175c	148	260	232
Husband PT	284	296	441	476	162	170	212	229
Wife PT	336	300	426	400	226	213	214	210
Dual earner	189b	180	411a	394	109	106	193d	183
Both FT	184b	176	409a	394	109	105	201d	190
Husband FT, Wife PT	194	184	413a	394	101	98	161	161
Wife FT, Husband PT	203	195	422	378	125	141	257d	189
Both PT	285	257	438	402	188	172	178	221
Life stage
Nonparents	221	220	443	435	186	180	330	316
No children, wife ≤ age 45	222	219	475	449	186	181	347	325
No children, wife > age 45	220	220	424	428	185	180	321	312
Parents	197b	186	411a	388	80c	74	123d	114
Youngest child 1 or under	222	211	471a	442	57	52	79	75
Youngest child age 2	207	195	452a	415	58	55	81	75
Youngest child age 3-5	193b	176	417a	388	62	54	86d	73
Youngest child age 6–9	190b	167	393a	367	70c	61	92	84
Youngest child age 10–13	192b	169	389a	364	78c	65	110	103
Youngest child age 14–17	179	178	388a	356	91	89	171d	143
Youngest child age 18+	193	207	379	383	134	132	225	218
Full sample	205b	198	422a	404	116	112	192d	184
Number of observations	10,831	12,395	10,987	12,670	10,831	12,395	10,987	12,670

Note. Data are based on the authors’ calculations from 2003–2010 American Time Use Survey (ATUS) data obtained from ATUS-X (Hofferth et al., 2013). The sample included all married respondents between ages 18 and 64, with at least one spouse working. FT = full time; PT = part time.

^aMen's and women's reports of total shared time on weekends is significantly different (p < .05).

^bMen's and women's reports of total shared time on weekdays is significantly different (p < .05).

^cMen's and women's reports of exclusive spousal time on weekdays is significantly different (p < .05).

^dMen's and women's reports of exclusive spousal time on weekends is significantly different (p < .05).

When we compared individuals who had different work arrangements, we found that most individuals in single-earner work arrangements spent more time with their spouses than individuals who were members of a dual-earner couple. Nevertheless, the differences were more nuanced depending on how much one or both members of the couple worked. For example, individuals in arrangements whereby both spouses worked part time spent more time with one another in both total and exclusive time on weekdays than individuals in single-earner arrangements whereby either the husband or wife worked part time.

In terms of family life stage, men's and women's total shared time was slightly U shaped on both weekends and weekdays, with total time together lowest among parents of school-age children (ages 6–9, 10–13, and 14–17) and higher among parents of children under 5 and over 18, as well as among those without coresident children. Exclusive spousal time was much lower for parents than nonparents on both weekends and weekdays, and family life stage differences in exclusive spousal time were more pronounced than differences in total shared time (see columns 5–8 of Table 2). Finally, a comparison of men's and women's reports of time with a spouse suggested that women may report less time with their husbands than husbands report with their wives.

Analytic Results

Our time-together analysis examined relationships between work and family demands and the time shared with one's spouse in total and exclusively with a spouse. Results from OLS regressions for weekday and weekend total shared time with a spouse are presented in Table 3, and results for exclusive spousal time separately for weekdays and weekends are shown in Table 4.

Table 3.

Ordinary Least Squares Models of Total Shared Time on Weekdays and Weekends, 2003–2010

Predictor	Weekday						Weekend
	Model 1		Model 2		Model 3		Model 4		Model 5		Model 6
Earner status (ref. = both FT)
Single earner
Husband FT	37.79	***	–4.32		–9.44	**	30.20	***	10.10	*	12.32	*
	(3.55)		(3.85)		(3.55)		(5.16)		(5.02)		(4.87)
Wife FT	66.44	***	25.06	***	15.16	*	26.24	**	2.94		1.89
	(6.60)		(6.60)		(6.55)		(9.12)		(9.06)		(8.88)
Husband PT	106.66	***	59.41	***	36.40	**	54.29	**	35.02		29.02
	(13.17)		(13.23)		(12.82)		(19.47)		(18.27)		(17.98)
Wife PT	136.49	***	94.08	***	68.49	***	19.59		1.65		–1.83
	(12.74)		(12.34)		(12.11)		(15.35)		(15.02)		(14.96)
Dual earner
Husband FT, Wife PT	11.53	**	3.92		–10.65	**	1.53		3.76		–1.34
	(3.77)		(3.68)		(3.58)		(5.75)		(5.53)		(5.37)
Wife FT, Husband PT	15.66		11.96		–2.64		–4.43		–3.61		–3.83
	(8.79)		(8.49)		(8.56)		(15.36)		(14.49)		(14.56)
Both PT	94.54	***	82.50	***	51.45	***	17.52		28.11		19.68
	(14.95)		(14.63)		(14.10)		(22.51)		(19.89)		(18.95)
Life stage (ref. = youngest child age 6–9)
No children, wife ≤ age 45	43.87	***	44.99	***	46.51	***	75.62	***	74.79	***	77.88	***
	(6.01)		(5.81)		(5.64)		(9.03)		(8.57)		(8.31)
Youngest child 1 or under	24.83	***	21.21	***	19.47	***	56.73	***	53.07	***	51.35	***
	(5.33)		(5.20)		(5.03)		(7.64)		(7.28)		(7.08)
Youngest child age 2	13.46	*	10.19		8.72		37.77	***	34.54	***	33.13	***
	(6.39)		(6.36)		(6.12)		(9.33)		(8.89)		(8.65)
Youngest child age 3–5	–0.14		0.15		–0.07		14.93	*	16.26	*	15.99	*
	(4.71)		(4.64)		(4.49)		(7.08)		(6.72)		(6.56)
Youngest child age 10–13	6.44		5.93		7.30		5.74		5.70		6.52
	(5.01)		(4.85)		(4.72)		(7.50)		(7.18)		(7.02)
Youngest child age 14–17	7.77		7.72		11.41	*	5.62		5.40		6.12
	(5.50)		(5.36)		(5.22)		(8.42)		(8.06)		(7.94)
Youngest child age 18+	30.49	***	31.14	***	37.51	***	22.76	*	20.84	*	21.64	*
	(7.13)		(6.87)		(6.70)		(10.33)		(9.92)		(9.76)
No children, wife > age 45	46.99	***	47.78	***	51.74	***	70.65	***	67.27	***	68.93	***
	(5.95)		(5.77)		(5.58)		(8.90)		(8.49)		(8.28)
Respondent work controls
Workday (ref. = non-workday)			–107.67	***					–159.57	***
			(4.29)						(4.26)
Minutes paid work					–0.25	***					–0.49	***
					(0.01)						(0.01)
Respondent characteristics
Female (ref. = male)	–9.04	**	–32.25	***	–48.27	***	–20.38	***	–38.73	***	–45.91	***
	(2.79)		(2.94)		(3.00)		(4.20)		(4.10)		(4.02)
Age	–1.01	***	–0.97	***	–1.13	***	–1.49	***	–1.34	***	–1.48	***
	(0.24)		(0.23)		(0.23)		(0.36)		(0.34)		(0.34)
Husband characteristics
Education (ref. = less than HS degree)
GED/HS degree	5.75		5.00		6.08		5.44		7.01		7.97
	(6.16)		(5.97)		(5.89)		(9.12)		(8.72)		(8.50)
Some college	8.70		8.14		8.46		4.97		8.85		5.59
	(6.34)		(6.14)		(6.06)		(9.46)		(9.09)		(8.89)
College/advanced degree	17.11	**	18.85	**	17.89	**	23.78	*	30.79	***	19.54	*
	(6.46)		(6.26)		(6.16)		(9.69)		(9.31)		(9.10)
Race (ref. = White)
Black, non-Hispanic	–15.41		–21.34		–19.18		–40.56	*	–41.64	*	–39.74	*
	(12.48)		(11.57)		(11.23)		(19.50)		(19.06)		(18.58)
Other, non-Hispanic	–9.55		–11.58		–11.42		–12.73		–7.80		0.40
	(7.86)		(7.68)		(7.32)		(13.05)		(12.40)		(12.12)
Hispanic	12.59		11.40		9.96		8.67		9.16		14.51
	(7.16)		(6.96)		(6.74)		(10.33)		(9.81)		(9.54)
Wife characteristics
Education (ref. = less than HS degree)
GED/HS degree	–3.59		–4.77		–5.57		3.89		–0.15		–2.26
	(6.77)		(6.60)		(6.49)		(9.90)		(9.37)		(9.07)
Some college	–11.19		–11.76		–11.69		–7.20		–8.25		–11.36
	(6.90)		(6.72)		(6.59)		(10.18)		(9.68)		(9.36)
College/advanced degree	–14.75	*	–12.17		–10.91		7.49		10.65		–2.41
	(7.10)		(6.92)		(6.79)		(10.52)		(10.03)		(9.70)
Race (ref. = White)
Black, non-Hispanic	–31.40	*	–27.33	*	–29.00	*	–61.44	**	–61.18	**	–56.76	**
	(12.43)		(11.60)		(11.30)		(20.11)		(19.74)		(19.15)
Other, non-Hispanic	–9.86		–7.73		–6.29		6.49		4.92		2.88
	(7.39)		(7.27)		(6.92)		(11.69)		(11.32)		(11.03)
Hispanic	–18.71	**	–17.58	**	–16.20	*	–0.42		–0.53		2.46
	(7.04)		(6.81)		(6.52)		(10.43)		(9.88)		(9.60)
Constant	202.99	***	306.51	***	336.45	***	434.44	***	484.97	***	500.89	***
	(12.39)		(13.08)		(12.60)		(18.80)		(17.79)		(17.38)
Model fit
F	27.11	***	44.76	***	68.90	***	17.68	***	59.60	***	139.27	***
N	23,226		23,226		23,226		23,657		23,657		23,657

Note. Numbers in parentheses are standard errors. Data are based on the authors’ calculations from 2003–2010 American Time Use Survey (ATUS) data obtained from ATUS-X (Hofferth et al., 2013). Models also control for year (2003 is the reference) and whether the diary was collected on a holiday (non-holiday is reference; results are available on request). ref. = reference category; FT = full time; PT = part time; HS = high school.

^*p < .05.

^**p < .01.

^***p < .001.

Table 4.

Ordinary Least Squares Models of Spousal Time on Weekdays and Weekends, 2003–2010

Predictor	Weekday			Weekend
	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6
Earner status (ref. = both FT)
Single earner
Husband FT	19.80***	–0.70	–4.18	11.71**	4.16	4.32
	(2.70)	(2.97)	(2.79)	(4.03)	(4.02)	(3.96)
Wife FT	41.21***	21.07***	15.17**	25.28**	16.53*	15.22*
	(5.28)	(5.22)	(5.18)	(7.85)	(7.82)	(7.72)
Husband PT	64.11***	41.11***	28.44**	30.85*	23.62	20.40
	(9.67)	(9.79)	(9.62)	(14.65)	(14.32)	(14.23)
Wife PT	98.57***	77.93***	64.05***	3.41	–3.33	–5.45
	(11.29)	(11.14)	(11.00)	(12.27)	(12.11)	(12.10)
Dual earner
Husband FT, Wife PT	10.30***	6.59*	–0.96	–3.60	–2.76	–4.78
	(2.90)	(2.86)	(2.83)	(4.31)	(4.30)	(4.24)
Wife FT, Husband PT	10.82	9.02	1.53	–1.86	–1.56	–1.62
	(8.14)	(8.06)	(8.14)	(12.60)	(12.33)	(12.38)
Both PT	66.64***	60.78***	44.77***	3.13	7.10	4.02
	(11.99)	(11.82)	(11.69)	(16.24)	(15.56)	(14.99)
Life stage (ref. = youngest child age 6–9)
No children, wife ≤ age 45	119.93***	120.47***	121.27***	243.23***	242.92***	244.17***
	(4.91)	(4.83)	(4.76)	(7.36)	(7.25)	(7.13)
Youngest child age 1 or under	–14.29***	–16.05***	–17.01***	–14.62***	–15.99***	–16.84***
	(3.08)	(3.09)	(3.08)	(4.20)	(4.18)	(4.18)
Youngest child age 2	–11.23**	–12.82***	–13.64***	–11.61*	–12.82**	–13.53**
	(3.46)	(3.47)	(3.42)	(4.73)	(4.70)	(4.69)
Youngest child age 3–5	–8.81**	–8.67**	–8.77**	–9.34**	–8.84*	–8.90*
	(2.84)	(2.85)	(2.82)	(3.57)	(3.59)	(3.60)
Youngest child age 10–13	7.05*	6.81*	7.49*	18.91***	18.90***	19.23***
	(3.17)	(3.13)	(3.11)	(4.18)	(4.18)	(4.19)
Youngest child age 14–17	25.75***	25.73***	27.60***	69.26***	69.18***	69.47***
	(3.74)	(3.71)	(3.68)	(5.64)	(5.60)	(5.57)
Youngest child age 18+	67.40***	67.71***	70.96***	133.71***	132.99***	133.25***
	(5.52)	(5.45)	(5.39)	(7.95)	(7.86)	(7.80)
No children, wife > age 45	112.61***	112.99***	115.02***	225.91***	224.65***	225.20***
	(4.60)	(4.53)	(4.47)	(6.80)	(6.71)	(6.62)
Respondent work controls
Workday (ref. = non-workday)		–52.40***			–59.91***
		(3.34)			(3.40)
Minutes paid work			–0.13***			–0.20***
			(0.01)			(0.01)
Respondent characteristics
Female (ref. = male)	–6.83**	–18.13***	–26.75***	–12.39***	–19.28***	–22.94***
	(2.18)	(2.35)	(2.41)	(3.31)	(3.33)	(3.30)
Age	–0.19	–0.17	–0.25	–0.26	–0.20	–0.26
	(0.19)	(0.19)	(0.19)	(0.30)	(0.29)	(0.29)
Husband characteristics
Education (ref. = less than HS degree)
GED/HS degree	7.07	6.70	7.24	4.19	4.78	5.23
	(4.45)	(4.43)	(4.42)	(7.18)	(7.07)	(7.03)
Some college	12.66**	12.39**	12.54**	1.25	2.71	1.51
	(4.67)	(4.63)	(4.61)	(7.46)	(7.37)	(7.33)
College/advanced degree	15.86**	16.70***	16.25***	3.24	5.87	1.49
	(4.92)	(4.88)	(4.85)	(7.72)	(7.61)	(7.57)
Race (ref. = White)
Black, non-Hispanic	–9.41	–12.30	–11.32	–25.59*	–26.00*	–25.25
	(9.11)	(9.33)	(9.33)	(13.03)	(13.03)	(12.92)
Other, non-Hispanic	–8.75	–9.74	–9.70	5.79	7.64	11.22
	(6.03)	(6.01)	(5.84)	(10.17)	(9.94)	(9.75)
Hispanic	2.51	1.93	1.18	–16.62*	–16.43*	–14.21*
	(5.57)	(5.52)	(5.46)	(7.30)	(7.20)	(7.09)
Wife characteristics
Education (ref. = less than HS degree)
GED/HS degree	1.61	1.03	0.60	7.53	6.01	4.98
	(4.94)	(4.93)	(4.90)	(7.45)	(7.32)	(7.26)
Some college	0.46	0.19	0.21	4.07	3.68	2.36
	(5.16)	(5.13)	(5.09)	(7.67)	(7.55)	(7.47)
College/advanced degree	–3.05	–1.79	–1.10	9.51	10.70	5.42
	(5.37)	(5.33)	(5.29)	(8.00)	(7.87)	(7.81)
Race (ref. = White)
Black, non-Hispanic	–9.12	–7.14	–7.91	–19.87	–19.78	–17.94
	(9.19)	(9.46)	(9.47)	(13.45)	(13.44)	(13.33)
Other, non-Hispanic	–7.76	–6.72	–5.94	–9.36	–9.95	–10.85
	(6.00)	(5.98)	(5.81)	(8.87)	(8.70)	(8.57)
Hispanic	–12.16*	–11.62*	–10.89*	–12.73	–12.77	–11.54
	(5.45)	(5.39)	(5.28)	(7.35)	(7.26)	(7.12)
Constant	59.43***	109.81***	127.18***	104.74***	123.71***	132.21***
	(9.37)	(10.10)	(9.89)	(14.65)	(14.46)	(14.30)
Model fit
F	78.37***	78.39***	82.48***	125.60***	128.10***	135.68***
N	23,226	23,226	23,226	23,657	23,657	23,657

Note. Numbers in parentheses are standard errors. Data are based on the authors’ calculations from 2003–2010 American Time Use (ATUS) data obtained from ATUS-X (Hofferth et al., 2013). Models also control for year (2003 is the reference) and whether the diary was collected on a holiday (non-holiday is the reference; results are available upon request). ref. = reference category; FT = full time; PT = part time; HS = high school.

^*p < .05.

^**p < .01.

^***p < .001.

Total shared time

Our expectation was that greater work demands would result in less time with a spouse (Hypothesis 1), but we found that the relationship is more complex. Compared to dual-earner couples in which both husband and wife work full time, individuals in most other arrangements spent more time with their spouse on weekdays (see Models 1–3 of Table 3). Differences in total shared time compared to both full-time-employed dual earners (Model 1) are largest among individuals in single-earner couples in which the husband or wife works part time (106 and 136 minutes, respectively), in which the wife works full time and the husband does not work (66 minutes), and among individuals in dual-earner couples in which both members of the couple work part time (94 minutes). On the other hand, differences were much more modest between full-time-working dual-earner couples (see Model 1 in Table 3) and traditional breadwinner couples in which the husband works full time and the wife does not work (37 minutes) and individuals in couples in which the husband works full time and the wife works part time (11 minutes). With controls for whether the respondents worked on the diary day (Model 2) and minutes worked (Model 3), many of the differences in the first model remained, although the magnitudes are reduced. Differences between individuals in full-time dual-earner couples and those in the two other most common arrangements (traditional male breadwinner and husband full-time/wife part-time work) were insignificant when controlling for diary day paid work (Model 2) and changed direction in Model 3 when we controlled for time spent in paid work.

Weekend earner status differences in shared time with a spouse were less pronounced compared to weekdays (Models 4–6 in Table 3). Individuals in single-earner arrangements, except those in which the wife works part time and the husband does not work, spent more time with their spouse on weekends than individuals in dual-earner couples in which both members work full time (Model 4). However, when controlling for weekend diary day paid work (Model 5) and time spent working on the diary day (Model 6), only individuals in single-earner couples in which the husband works full time spent more time together than individuals in dual-earner couples where both members work full time, and the difference was only 10 minutes.

The relationship between family life stage and an individual's time shared with a spouse observed in the Table 2 descriptive results—that total shared time was lower among parents compared to nonparents—generally persisted in multivariate models (see Table 3), supporting Hypothesis 2, which posited that parents spend less time together than nonparents. Men and women without coresident children spent more time with their spouse on weekends and weekdays than parents whose youngest coresident child is over age 1. Nonparents spent about 45 minutes more per day with their spouse than parents of children ages 6–9 on weekdays (Models 1–3 in Table 3) and just over an hour more on weekends (75 minutes when the wife is under age 45, and 70 minutes when the wife is over age 45; Model 4 in Table 3).

In Hypothesis 3 we posited that parents of younger children spend less time together than parents of older children. Again, we found that the relationship is a bit more complex than hypothesized. Parents of children age 1 or younger spent about 20 minutes more with their spouse on weekdays (Models 1–3 in Table 3) than parents whose youngest children are ages 6–9 and nearly an hour (56 minutes in Model 4) more on weekends compared to parents of children ages 6–9. On weekdays and weekends, parents whose youngest children are 2 and under and who are age 18 and over, as well as parents of children ages 3–5 (weekends only), shared greater amounts of time together than couples whose youngest children are between ages 6 and 9 years. In short, parents of school-age children (6–17) appeared to be the most constrained in finding time to share with one another on both weekends and weekdays.

In addition to differences in total shared time by earner status and family life stage, we found evidence of gender differences, as expected. Women reported less time with their husbands than husbands reported with their wives net of work and family arrangements (see Table 3; 9 minutes on weekdays in Model 1 and 20 minutes less on weekends in Model 4), which is consistent with previous research. On both weekdays and weekends, differences between women and men are larger in magnitude when we accounted for diary day work characteristics (see Models 2–3 and 5–6).

Exclusive spousal time

Table 4 contains estimates of exclusive spousal time (time spent with a spouse and no one else) on weekdays and weekends. On weekdays, individuals in all types of earner arrangements, except for those in which the wife works full time and the husband works part time, spent more time alone with their spouse than individuals in both full-time-working couples (Model 1). Controlling for whether (Model 2) and how much (Model 3) the respondent works on the diary day accounted for differences between individuals in both full-time-working dual-earner couples and traditional male-breadwinner couples (i.e., husband works full time), as did the work time control in couples in which the husband works full time and the wife works part time (Model 3). Differences in exclusive spousal time on weekends were largest between individuals in couples in which both the husband and wife work full time and individuals in dual-earner couples in which both members work part time (66 minutes) and in single-earner couples in which the husband works part time (64 minutes) and the wife works part time (98 minutes), which is consistent with Hypothesis 1, that greater work demands are associated with less time together.

On weekends, quite a different story about work commitments and exclusive spousal time emerged. Individuals in single-earner couples (except those in which the wife works part time) spent more time alone together than individuals in full-time-working dual-earner couples (Model 4). However, when we controlled for whether the respondent works (Model 5) and how much he or she works (Model 6), we found that only individuals in single-earner couples in which the wife works full time spent more time alone together on weekends than dual-earner couples in which both work full time, a difference of about 16 minutes, suggesting that work demands matter most on days that people work (as would be expected), which for many people are weekdays.

We also found that mothers and fathers spent less time in exclusive spousal time on weekdays and weekends than nonparents (see Table 4). Nonparents spent nearly 2 hours more alone together on weekdays (in Model 1, 119 minutes for those in couples where the wife is under 45 and 112 minutes when the wife is over 45) than parents whose youngest children are 6–9 and nearly 4 hours more alone together on weekends (Model 4), providing support for Hypothesis 2, which posited that nonparents spend more time together than parents. In contrast to the total spousal time results (see Table 3), in which parents of school-age children spent the least amount of time together, parents’ time alone together was most limited among the parents of the youngest children, which is as expected (Hypothesis 3). On weekdays, parents of infants spent 14 fewer minutes alone together than parents whose youngest children were ages 6–9; parents of children ages 14–17 spent 25 more minutes together, and parents of adult children spent 67 more minutes together. Although parents of children under age 5 spent less time alone together on weekdays than parents whose youngest children are ages 6–9 on, differences between parents of children under age 5 were not significant (results available on request). On weekends, the patterns are very similar, especially for parents of children age 5 and under. Differences between parents of children ages 6–9 and parents of older children were roughly twice as large on weekends (Model 4) compared to weekdays (Model 1). Parents of children ages 14–17 spent an hour (69 minutes in Model 4) more alone with their spouse than parents of children ages 6–9; the difference was 133 minutes for parents with coresident adult children (age 18+; Model 4).

Regarding men's and women's reports of time spent with a spouse and no one else, we noted persistent differences that are consistent with those observed in the total spousal time models. Women consistently reported less time alone with their husbands than men reported with their wives. Including work controls yielded larger differences between men and women, with women reporting 18 fewer minutes with their husbands when controlling for workday (Model 2) as opposed to only 6 minutes without the control (Model 1). The differences are similar (12 minutes in Model 4 and 19 minutes in Model 5) on weekends even though exclusive spousal time was higher than on weekdays.

Well-Being Analysis

Despite a positive association between marital interaction and happiness (e.g., Amato, Johnson, Booth, & Rogers, 2003), suggesting that time shared with a spouse is important for an individual's well-being, there is limited evidence regarding the quality of the time that couples actually spend together (but see Sullivan, 1996). The time-together analyses described above showed differences in total shared time and exclusive spousal time at various life stages, with parents generally spending less time with their spouse than nonparents. Also, although paid work was negatively associated with an individual's time with his or her spouse (either in total shared or exclusive spousal time), being in a dual-earner couple was not necessarily a hindrance to sharing time with a spouse after accounting for daily work commitments. In the analyses we report next, which are based on the 2010 well-being module of the ATUS, we addressed our fourth hypothesis, which posited that individuals experience greater happiness and meaning and less stress during time spent with a spouse compared to time spent apart.

Measures

Recall that with the ATUS time diaries respondents were asked how they felt during three randomly selected activities (see DATA AND METHOD section for more information). We focused on three activity-level measures of affect (or subjective well-being): (a) happiness, (b) meaningfulness, and (c) stress. The variables happy, meaningful, and stress were measured on a 7-point scale (0–6), and we recoded them into dichotomous variables indicating that the respondent was very happy (5–6), the activity was very meaningful (5–6), and whether the respondent experienced any stress during the activity (> 0) because there is substantial heaping within the 7-point scale.

Analytic strategy

We performed individual fixed effects and activity-level logit analyses to compare well-being during activities with and apart from spouses. The individual fixed effect method is a within-person analysis that contrasts well-being during time spent with and without a spouse for individuals who reported at least one activity with their spouse and one activity without their spouse. The within-person analysis is possible because respondents reported affect information for three activities; thus, we analyzed whether individuals are happier, or more stressed, and whether activities are more meaningful when they are with their spouse compared to when they are not with their spouse. In this application, fixed effect models fully account for the individual's characteristics, and we included a broad indicator for the activity performed when affect was measured for the individual. With the fixed effects analysis we have greater confidence in the causal impacts of spousal presence on activity-level well-being. However, despite the appeal of the within-person fixed effect model, the sample of respondents includes only those with variation in affect (the dependent variable), for example, one (or more) activity for which the respondent reported being very happy and one (or more) activity for which the respondent reported being less than very happy. Estimates for the independent variable of interest—with spouse—were calculated on the basis of the respondents with sampled activities performed both with and without the spouse. This requirement of the fixed effect model—to have variation on the dependent variable coupled with the limited number of observations on the dependent variable (only three activities were selected for respondents to report their affect)—may result in sample selection bias.

As a robustness check of the individual fixed effect analyses we also estimated well-being using activity-level logit models with the affect measures (happy, meaningful, stress) as dependent variables and the presence of the spouse during the activity as the key independent variable. This approach allowed us to retain all cases even if the respondent had no variation in well-being (the dependent variable). Because most individuals contributed more than one activity to the analysis, we clustered the standard errors on the person. All models estimated include the full set of controls used in the time-together analyses.

Descriptive Results

Descriptive statistics on men's and women's subjective well-being rating of activities with a spouse (any activities with a spouse, including alone with a spouse) and not with a spouse are shown in Table 5. The top panel of the table shows the average activity-level happiness, meaningfulness, and stress ratings on the 0-to-6 scale. The bottom panel shows the proportion of activities for which respondents reported being very happy, felt the activities were very meaningful, and experienced stress during the activities. The dichotomous measures shown in the lower panel were used in the multivariate analyses. The bivariate results suggest that well-being was enhanced when respondents were with their spouses. During more than 60% of the activities with a spouse, both men and women reported being very happy, compared to half of the activities not done with a spouse. Meaningfulness was also rated more highly during activities done with a spouse, whereas the percentage of activities in which respondents experienced stress was higher when the spouse was not present (just over 50% of activities).

Table 5.

Weighted Average Happiness, Meaningfulness, and Stressfulness During Activities When Married People Are With Their Spouse and Not With Their Spouse

Variable	Not with spouse		Not with spouse
	Men	Women	Men	Women
	(1)	(2)	(3)	(4)
Rating (0–6)a
Happy	4.62b	4.64c	4.19	4.26
Meaningful	4.47b	4.61c	4.17	4.31
Stress	1.05b	1.17c	1.43	1.57
Dichotomousd
Very happy (5–6)	0.62b	0.63c	0.49	0.51
Very meaningful (5–6)	0.59b	0.63b	0.54	0.57
Any stress (≥ 1)	0.42b	0.42c	0.53	0.53
n	2,573	2,704	4,264	5,185

Note. Data are based on the authors’ calculations from 2010 American Time Use Survey (ATUS) well-being module data obtained from ATUS-X (Hofferth et al., 2013).

^a“Happy,” “Meaningful,” and “Stress” were reported by respondents for the given activity on a 0–6 scale.

^bSubjective well-being rating for “with spouse” for men is significantly different than the rating for “not with spouse” (p < .05).

^cSubjective well-being rating for “with spouse” for women is significantly different than the rating for “not with spouse” (p < .05).

^d“Very Happy,” “Very Meaningful,” and “Any Stress” are dichotomous measures created from the 0–6 scale; “Very Happy” indicates that “Happy” was reported as a 5 or 6, “Very Meaningful” indicates that “Meaningful” was reported as 5 or 6 and “Any Stress” indicates that “Stress” was reported from 1–6. “With Spouse Only” is a subset of “With Spouse” and includes activities during which the only other person present was the respondent's spouse.

Analytic Results

The estimates of subjective well-being during activities done with one's spouse versus apart are given in Table 6. The upper panel of the table shows the individual fixed effect logit results (the focus of our discussion). The coefficients shown are odds ratios, indicating the odds of the respondent rating the activity as very happy, very meaningful, or as having experienced any stress when with their spouse compared to not with their spouse. The fixed effects analyses included only the respondents who had variation in the dependent variables (e.g., at least two different happiness ratings) across the three randomly selected activities and who had activities selected when they were with their spouse and without. The activity-level logit analyses, including the same controls as in the time-together analyses, confirmed the results of the fixed effects analyses with the full sample of individuals. The differences in coefficients between the activity-level logit and fixed effects models were expected because the sample in the activity-level logit was larger and composed of different people than the fixed effects analysis.

Table 6.

Odds Ratios of Well-Being While Married People Are With Their Spouse Compared to When Not With Their Spouse

Predictor	Very happy	Very meaningful	Any stress
	Model 1	Model 2	Model 3
Individual fixed effects logita
With spouse (ref. = not with spouse)	1.81***	1.52***	0.79**
	(0.13)	(0.11)	(0.06)
N activities	6,240	6,831	5,821
N respondents	2,092	2,289	1,950
χ 2	403.6	388.1	445.5
Log likelihood	–2,082	–2,306	–1,908
Activity-level logitb
With spouse (ref. = not with spouse)	1.46***	1.36***	0.84*
	(0.11)	(0.10)	(0.06)
Female (ref. = male)	1.04	1.10	1.03
	(0.08)	(0.09)	(0.08)
N activities	14,726	14,726	14,726
N respondents	4,962	4,962	4,962
Pseudo R2	.052	.052	.098
χ 2	289.3	279.6	419.2

Note. Numbers in parentheses are standard errors. Data are based on the authors’ calculations from 2010 American Time Use Survey (ATUS) well-being module data obtained from ATUS-X (Hofferth et al., 2013). Reference category (ref.) is not with spouse during activity.

^aIndividual fixed effect models include seven category activity measure as control variables.

^bThe activity-level logit models include control variables for couple-level work status, life stage, age, husband's education and race, wife's education and race, the seven-category activity measure, weekend, and holiday diary days. The models allow for nonindependent errors within groups defined by the individual respondent.

^*p < .05.

^**p < .01.

^***p < .001.

We expected to find greater well-being (more happiness, more meaningfulness, and less stress) during activities with a spouse compared to activities done apart (Hypothesis 4), and we found support for this hypothesis across our three dependent variables. The fixed effects analyses, accounting for unobserved individual characteristics and indicating within-person variation, showed that activities performed with the spouse elicited greater happiness and more meaning, and were less stressful than activities apart from the spouse, controlling for the specific activity performed. Specifically, the respondents were 1.8 times as likely to report being very happy when they were with their spouse than when they were not (Model 1 in Table 6). Women and men found activities with their spouse 1.5 times as meaningful as activities performed without their spouse (Model 2 in the upper panel of Table 6). Finally, the presence of a spouse during an activity was associated with less stress, with respondents 21% less likely to report stress when with a spouse in the fixed effects model compared to during activities done without a spouse.

Discussion

Recall that we began this article by situating this research in the work–family conflict literature that proposes time-based conflict occurs when the demands of multiple roles conflict with one another. To the extent that being a worker, parent, and spouse are roles in conflict with one another, there is clear evidence of parents’ commitment to children despite heightened work demands (e.g., Bianchi et al., 2006; Raley, Bianchi, & Wang, 2012) and a suggestion in the literature that limiting time with a spouse may be one strategy for managing competing time-based demands (e.g., Bianchi et al., 2006; Wight et al., 2008). However, limiting time with a spouse as a way to alleviate time-based conflict between work and parenting roles may create other problems to the extent that interaction with a spouse is positively associated with marital well-being (Amato et al., 2007; Booth et al., 1985, 1986; Crawford et al., 2002; Gager & Sanchez, 2003; Hill, 1988; White, 1983; Zuo, 1992). In this study we brought together these separate strands of work by using nationally representative time diary data from the ATUS, and a new module on well-being, to consider how work and parenting demands are associated with couples’ shared time and how couples’ shared time is related to well-being.

This study makes several contributions to our understanding of how work and family demands shape spouses’ total shared time and their exclusive spousal time (with one another but no one else) as well as their subjective well-being during the time spent with their spouses (compared to time spent apart). First, our examination considered both total spousal time and exclusive spousal time on weekends and weekdays, and our results underscore the importance of such distinctions. Most other studies examine total spousal time, exclusive spousal time, or shared leisure specifically, and generally only on either weekends or weekdays or just controlling for day of the week. Making these distinctions allowed us to better understand the ways in which work and family demands matter for time shared with a spouse.

Second, the literature shows that individuals in dual-earner couples spend less time together than those in single-earner arrangements, which is of particular significance given that the dual-earner family is the dominant form in the 21st century (e.g., Jacobs & Gerson, 2001). Yet the simple distinction between single- and dual-earner couples masks important differences within these two categories, and the ability to consider more finely grained distinctions within single- and dual-earner categories is a real strength of the ATUS data. By categorizing individuals according to both the respondent and his or her spouse's level of employment, we shed light on the experiences of individuals with different work arrangements, who might be expected to experience different time-based conflict. To be specific, on weekdays, controlling for daily, time-based work constraints, we found that couples’ shared time is similar once we accounted for diary day paid work for individuals in the three most common family forms—full-time-working dual-earner couples, male-breadwinner couples, and dual-earner couples in which the husband works full time and the wife works part time—representing the vast majority (85%) of the married population ages 18–64. Nonetheless, individuals in full-time-working dual-earner couples still spent less time together compared to individuals in the less common female-single-earner and/or exclusively part-time-earner arrangements. In part, these persistent differences may be related to differences in spouses’ diary day work commitments, which we could not analyze with these data. On weekend days, when work and family are less likely to directly conflict, individuals in both full-time-working, dual-earner couples fare similarly to most other couples save for individuals in full-time male-breadwinner arrangements, who spend about 10 more minutes together, and individuals in full-time female-breadwinner arrangements, who spend about 15 minutes more together, exclusively. The single-earner advantage in shared time with a spouse on weekends may result from the ability of at-home spouses to complete chores around the home or run errands while their spouses are working for pay, thereby freeing up time for couples to spend together.

Third, we observed considerable variability in shared time with a spouse and in particular exclusive spousal time by family life stage, with significant differences between parents and nonparents. Such finely grained analyses would not be feasible if it were not for the ability to pool the large, annual ATUS samples. Like previous research (e.g., Dew, 2009; Kalmijn & Bernasco, 2001), we found that parents spent less time together than nonparents, although parents still spent 3 hours per day together, on average, on weekdays. The most substantial difference between parents and nonparents was in exclusive spousal time, with parents averaging less than half as much time alone together as nonparents on weekends and weekdays. Even among parents, however, we found noteworthy differences—consistent with Bittman and Wajcman (2000)—that likely reflect the changing time demands of children as they age. For example, parents of children under age 2 spent more time together in total, but less time alone together, than parents of elementary school–age children (6–9) on both weekends and weekdays. Similarly, parents of coresident children age 18+ spent more time together (both total and exclusive) on weekends and weekdays than parents whose youngest children were of elementary school age. In short, our evidence suggests that the time-based demands of children change as they age and that exclusive spousal time gradually increases as children age, although longitudinal data are required to understand how patterns of interaction with a spouse map onto the changing demands of children as they age.

Fourth, our chronicling of persistent yet modest gender differences in the amount of time spent with a spouse as was expected. That women reported less total shared time with a spouse and less exclusive spousal time than did men is consistent with much of the literature (e.g., Dew, 2009; Freedman et al., 2012; Mansour & McKinnis, 2014; Voorpostel et al., 2009; Wight et al., 2008). Unfortunately, with our data we cannot explain this difference. Couple-level data, as opposed to individual-level data, are required to better understand why husbands and wives report being with one another at the same time and discrepant reports of shared time (e.g., Freedman et al., 2012). Perhaps what constitutes shared time for husbands and wives differs on the basis of who is reporting the shared time, which would be in line with a gender perspective and evidence about greater time-based conflict for women (Roxburgh, 2002).

Finally, this study makes an important contribution to the literature because we used new, unique data from the ATUS well-being module to examine the link between time with a spouse and well-being. Ours is the first study in nearly two decades to examine the association between well-being during activities performed with a spouse as opposed to considering associations between global measures of marital interaction and marital well-being. Using British data from 1986, Sullivan (1996) found that couples reported greater enjoyment during simultaneous activities than activities done apart. Our results are consistent with Sullivan's; when men and women were with their spouses, they reported being happier, finding more meaning, and experiencing less stress. This suggests the relevance of shared time with a spouse for married individuals’ well-being, which is consistent with research using global as opposed to momentary measures of interaction and well-being. Our findings are robust, as evidenced by consistent results using two different modeling strategies—one leveraging within-individual differences and the other comparing all activities with a spouse compared to those without a spouse—and controlling for the actual activity performed and other respondent and spouse characteristics.

Despite the significance of this research, it is not without limitations. First, the ATUS data are cross-sectional, so the differences we observed between parents with children of different ages suggest that parents may spend more time together as their children age, although longitudinal data are necessary to examine changes in couples’ shared and exclusive spousal time allocation as children age. Second, ATUS data are collected at the individual level, and thus we did not have data from couples, so we relied on one member of the couple's report about the time spent with his or her spouse. Couple-level data are required to better understand gender differences in shared and exclusive spousal time; perhaps there are certain activities that men report doing with their wives that women do not interpret as being done with their husbands, or vice versa. With couple-level data, one could assess the overlap between couples’ reports of shared activities and the presence of others (e.g., Freedman et al., 2012; Lesnard, 2008). Finally, marital survivorship is a concern. Parents of older children have either survived the early years of parenthood and are still married, or we were observing remarriages; unfortunately, we are not able to distinguish between these groups in the ATUS. To the extent that couples stay together because they spend time together and vice versa, selection must be acknowledged.

Despite the limitations of these data, our work shows the importance of considering individuals’ time both in full, and exclusively with their spouse, as well as how shared time is related to work and family demands and individual well-being; it also suggests directions for future research. A comparison of total shared time with a spouse and exclusive spousal time over time in the United States as well as cross-nationally would help us understand the extent to which the patterns we observed are emergent and/or specific to the United States or common across time and place and the extent to which work and family life stage operate similarly in different contexts. Similarly, considering couples’ shared time among older individuals is also of considerable interest, given that this is a period when work demands are typically diminishing and family caregiving demands may be increasing. Further understanding of time spent with a spouse in the presence of other people is also warranted because the bulk of shared time for parents is in the company of others. Supplemental analyses (not shown, results available on request) suggest that parents find time with a spouse and children particularly meaningful, and this should be investigated more thoroughly. Future research might also focus on differentiating between the activities individuals do with their spouses that promote momentary well-being and consider the relationship between spousal time and global assessments of well-being. In short, future work should consider couples’ shared time, because such investigations have the potential to strengthen our understanding of large-scale social and demographic patterns.