This study examines and compares shared time for same-sex and different-sex coresident couples using large, nationally representative data from the 2003–2016 American Time Use Survey (ATUS). We compare the total time that same-sex couples and different-sex couples spend together; for parents, the time they spend together with children; and for both parents and nonparents, the time they spend together with no one else present and the time they spend with others (excluding children). After we control for demographic and socioeconomic characteristics of the couples, women in same-sex couples spend more time together, both alone and in total, than individuals in different-sex arrangements and men in same-sex couples, regardless of parenthood status. Women in same-sex relationships also spend a larger percentage of their total available time together than other couples, and the difference in time is not limited to any specific activity.
Research has shown that many people feel they do not have enough time with their spouses, that most would like to spend more time with their partners (Bianchi et al. 2006; Nomaguchi et al. 2005; Roxburgh 2006), and that couples try to coordinate their schedules (Hallberg 2003; Hamermesh 2000, 2002; Jenkins and Osberg 2004; Sullivan 1996; van Klaveren and van den Brink 2007). There is also a positive relationship between marital interaction, as defined by shared day-to-day activities, and global measures of marital well-being (Amato et al. 2007; Crawford et al. 2002; Hill 1988) and in time diary–based analyses of shared time and well-being (Flood and Genadek 2016; Sullivan 1996). However, this evidence comes solely from analyses of different-sex couples; studies general exclude those in same-sex cohabiting and married relationships.
Drawing on time diary and survey data from the American Time Use Survey (ATUS), a nationally representative and population-based data set, we extend the limited knowledge of same-sex couples’ relationships by examining how the amount and type of time same-sex couples spend together compare with different-sex couples in the United States. By leveraging large samples and pooling data from 2003 to 2016, we identify 631 individuals living with a same-sex partner. In addition to having a sizable sample of same-sex couples, the respondents’ 24-hour time dairies have rich information on the co-presence of others throughout the day. We use the reports on who was present during the activity to measure the total amount of time that couples spend together as well as subsets of this shared time, including time alone with a partner, time with a partner when others (excluding children) were present, and time with a partner and children for parents. We also analyze the type of shared time by looking at broad groups of activities done in the presence of a spouse or partner.
Our analysis of shared time for same-sex and different-sex couples provides insight into similarities and differences between same-sex and different-sex couples. Because the literature on couples’ shared time using nationally representative time-diary data has focused exclusively on different-sex couples, our investigation of same-sex and different-sex couples’ shared time allows us to investigate whether patterns documented in the literature for different-sex couples hold for same-sex couples. Understanding the extent to which same-sex couples are able to share time with a partner and how this compares with the average time different-sex couples share is timely as same-sex relationships gain acceptance in the United States (Umberson et al. 2015). We know that for different-sex couples, together time is important for both marital quality (e.g., Amato et al. 2007; Booth et al. 1986) and individual well-being (e.g., Flood and Genadek 2016; Sullivan 1996), but our knowledge about same-sex couples’ shared time is much more limited. Evidence suggests that men in same-sex couples are happier when with a partner than when they are not, and sharing time for women in same-sex couples is associated with lower stress (Flood and Genadek 2020). Thus, examining the amount and nature of same-sex couples’ shared time at the national level increases our understanding of same-sex relationships in the United States. Although our measure of shared time may be limited in its ability to tell us about couples’ relationship quality or satisfaction, it provides a glimpse into the daily lives of same-sex couples and their relationships.
Same-sex couples are absent in most of the extant research on couples’ shared time, yet it is reasonable to expect that shared time with a partner is desired by most couples regardless of sexual orientation. But to assume that same-sex couples will spend similar amounts of time together as different-sex couples, without considering the different potential time constraints, lacks consideration of the barriers and opportunities for shared time that may differ.
On the one hand, evidence suggests that shared time may be similar for same-sex and different-sex couples given similarities in terms of relationship quality and partnering patterns. Same-sex and different-sex couples in the United States have similar relationship quality (Kurdek 2005, 2006; Peplau and Spalding 2000). Same-sex couples, like different-sex couples, also have similar tendencies to partner with individuals like themselves in terms of race, education, and age (Jepsen and Jepsen 2002; Schwartz and Graf 2009). Together, these findings suggest that differences in shared time between same-sex and different-sex couples will be small.
The sparse literature on same-sex individuals’ time use suggests more similarity than difference in the availability of time to share with a partner, especially given the common broader context of work and family in which both same- and different-sex couples live (Carrington 1999). Interview-based research has shown that same-sex and different-sex couples have similar daily time use patterns and constraints on their time (Carrington 1999), although the research has not specifically considered time with a partner. Interview studies have also shown that allocation of housework is more equal among members of same-sex couples, and the amount of total time spent in housework does not vary between same- and different-sex couples (Kurdek 1993; Mock and Cornelius 2003; Solomon et al. 2005). These findings are similar to time diary–based results from the ATUS showing that men in same-sex relationships do more housework than heterosexual men, and women in same-sex relationships do slightly less housework than women in different-sex relationships (Martell and Roncolato 2016). Additionally, among parents, there is no difference in time spent with children (Prickett et al. 2015).
However, there are reasons to expect differences in the quantity and nature of shared time between same-sex and different-sex couples. First, on average, same-sex couples have different demographic and socioeconomic profiles than different-sex couples (Gates 2015), and the amount of time couples spend together is related to their couple-level characteristics, especially paid work, marital status, parenthood, and gender (Flood and Genadek 2016). Second, same-sex couples are in the minority and face discrimination, they lacked the ability to legally marry until recently, and their relationships may be less stable on average (Manning et al. 2016; Rostosky et al. 2007). We elaborate on how couples’ resources and time constraints as well as minority stress may contribute to same-sex and different-sex couples’ shared time.
Demographic and Socioeconomic Variation
The first demographic difference between same- and different-sex couples that could influence the amount—or reported amount—of shared time is gender. Although in theory, married individuals should report the same amount of time spent together, married women in different-sex couples generally report less shared time than do men (Flood and Genadek 2016; Freedman et al. 2012b). Thus, given gender differences in the perception of shared time with a partner among men and women in different-sex couples, we might expect that women in same-sex relationships will spend less time together than men in same-sex relationships. However, a gender-as-relational perspective (West and Zimmerman 2009) emphasizes that the way gender is enacted depends on the relational context; that is, the way gender matters for emotional intimacy and boundaries as well as shared time may be different for men and women in same- and different-sex couples. Qualitative studies of women in different-sex and same-sex relationships show that women place more importance on emotional intimacy than men (Kurdek 2006; Peplau and Fingerhut 2007; Umberson et al. 2015), and emotional intimacy is often characterized by a lack of boundaries between partners (Rubin 1990; Umberson et al. 2015). In addition, evidence from a small study of women surveyed daily for 14 days suggests that more time with a partner fostered more emotional intimacy (Milek et al. 2015). The literature is unclear about relationships among emotional intimacy, boundaries, and shared time, but to the extent that spending time with a spouse or partner is a way to foster emotional intimacy and minimize boundaries between members of a couple, we expect women in same-sex couples to share more time with a partner than men and women in different-sex couples and men in same-sex relationships.
Previous research on different-sex couples has shown that nonparents spend more time together than parents (Flood and Genadek 2016; Garcia Roman et al. 2017), and parents of school-aged children spend the least amount of time together—both in total and alone with one another (Dew 2009; Flood and Genadek 2016). Men and women in same-sex couples are less likely to have children compared with different-sex couples; this difference is especially large for men in same-sex relationships (Black et al. 2000; Gates 2015; Krivickas and Lofquist 2011). Even when same-sex couples have children, they are less likely than different-sex couples to have young children. Men and women in same-sex relationships are also more likely to have children from previous relationships than from their current relationships compared with individuals in different-sex relationships (Henehan et al. 2008). Given these patterns, on average, men in same-sex relationships may have more shared time than other couples, and parents in same-sex relationships may share more time than parents in different-sex relationships because their children are more likely to be older and require less attention.
Labor market participation has a major impact on the time different-sex couples spend together on a given day. The amount of time spent at work is negatively related to the amount of time spent together, and single-earner couples spend slightly more time together than dual-earner couples (Flood and Genadek 2016; Kingston and Nock 1987; Voorpostel et al. 2009). Likewise, individuals with higher incomes are more likely to synchronize their schedules (Hamermesh 2000) and share more leisure time together (Sevilla et al. 2012). Same-sex couples are more likely to be in dual-earner arrangements and have more income than different-sex couples (Black et al. 2000; Gates 2015; Krivickas and Lofquist 2011). Thus, same-sex couples may spend less time together than different-sex couples because both partners are more likely to be working, but this could be offset by greater incomes, which increase coordination and leisure time.
Same-sex couples are much less likely to be married than different-sex couples, which is a product of legislation. In some U.S. states, same-sex marriage was legalized as early as 2004, but it has been legal for same-sex couples to marry throughout the United States only since 2015. If marriage is a “greedy institution,” as some scholars have theorized, married couples may spend more time together, and exclusively alone together, than cohabiting couples (Coser and Coser 1974). Limited research partially supports this; married couples without children spend slightly more time together than cohabiting couples without children, and cohabiting couples with children spend slightly more time alone with one another than married couples with children (Flood et al. 2016). However, these previous findings are for different-sex couples only, and marriage and cohabitation often have different meanings for same-sex and different-sex couples, especially given the legal changes over the period of our study (Manning and Brown 2015; Ocobock 2018; Reczek et al. 2009).
The demographic and socioeconomic differences among male same-sex couples, female same-sex couples, and different-sex couples are important to consider because for different-sex couples, these factors are associated with varying average amounts of time spent together. Although men in same-sex relationships are the least likely to have children and thus may spend the most time with a partner, they are also the couple type most likely to have both members working and may spend less time together.
Minority Stress and Relationship Stability
The minority status of same-sex couples and resulting stress associated with their relationships (Cao et al. 2017; LeBlanc et al. 2018; Lehmiller and Agnew 2006; Meyer 1995; Reczek 2016a; Rostosky and Riggle 2017; Rostosky et al. 2007) are factors that may uniquely influence shared time for these couples. Shared time of same-sex couples may be negatively influenced by stress from discrimination (Otis et al. 2006), and the amount of time with a partner and the type of activities shared for same-sex couples may be directly related to social stigma and discrimination. Spending time together in public activities or with family members may be more stressful in the face of stigma for same-sex couples than for different-sex couples and therefore may be avoided (Reczek 2016b; Rostosky et al. 2007). Thus, discrimination may result in less shared time with a partner for individuals in same-sex couples; however, it is possible that public stigma could result in more shared time in activities that are more common at home than in public for coresiding same-sex partners.
Until recently, same-sex couples could not marry in the United States and lacked legal protections on joint investments (Herek 2006), and until 2016, the majority of same-sex coresidential couples were not married (U.S. Census Bureau 2016). Thus, if shared time is considered an investment in a relationship (Booth et al. 1986; Hill 1988), same-sex couples may invest less in their relationships by spending less time together than different-sex couples. Research in this area is limited, with some evidence supporting this lack of relationship-specific investment and other research finding no differences. Same-sex couple relationships are less stable than different-sex relationships in Europe (Lau 2012) and different-sex married relationships in the United States (Manning et al. 2016). Similarly, Joyner et al. (2017) showed that in young adulthood, men’s same-sex relationships have higher dissolution rates than female same-sex couples and different-sex couples. However, other research has indicated similar stability between same and different-sex nonmarried couples (Manning et al. 2016; Rosenfeld 2014). Furthermore, evidence suggests that same-sex couples more highly prioritize joint activities than different-sex couples, which may mean that same-sex couples spend more time together than different-sex couples (Haas and Stafford 2005). Given the emergent state of the literature on same-sex relationships, we do not have clear expectations about whether same-sex and different-sex couples make different investments in their relationships that would be evident in their time shared with a partner.
The analyses in this study use large, nationally representative data to estimate potential differences in shared time for same-sex and different-sex couples. Although we do not directly test the theoretical mechanisms that may produce differences, the previous literature suggests that even after demographic and socioeconomic differences among couples are controlled for, shared time differences may remain.
We use repeated cross sections of American Time Use Survey (ATUS) data (Hofferth et al. 2017) from 2003–2016 to examine the amount and nature of time same-sex couples spend together compared with different-sex couples. The ATUS is a time diary study of a nationally representative sample of Americans. ATUS sample members are invited to complete the survey two to five months after they exit the Current Population Survey (CPS). The CPS is a monthly household survey of the civilian noninstitutionalized population. One individual aged 15 or older per former CPS participating household was randomly selected to report their activities over one 24-hour period. ATUS data are collected using a computer-assisted telephone interview, and the respondents report the activities they engaged in over a 24-hour period from 4 a.m. of a specified day until 4 a.m. of the following 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 U.S. population. The data we examine include one daily dairy from one member of a couple on one randomly selected day. Because these are not couple-level data with diaries from both partners, we leverage the rich information on co-presence indicating who was with the respondent during each activity reported in the time diary. We examine time spent with a spouse or partner, drawing on information from the time diary about who else was present during activities.
In addition to time diary data, demographic and socioeconomic characteristics are available in the ATUS. The ATUS collects information on the respondent’s age and employment at the time of the survey. Other information, including race, educational status, and household income, was collected at the time of the CPS (two to five months earlier). Household members who were present at the time of the CPS and the ATUS have information from the CPS interview as well as some updated information at the time of the ATUS. The ATUS respondent is also asked about changes in the household roster, reporting who resides in the household at the time of the ATUS. This results in some household members who entered the home between the CPS and the ATUS, for whom a limited set of information is available. This has implications for spouse and partner characteristics, which we describe shortly and incorporate in our analysis.
The 2003–2016 ATUS data include daily time diaries of 181,335 civilians aged 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 restrict the sample to respondents who have a spouse or partner living with them at the time of the ATUS interview, referred to hereafter as partner, and both the respondent and partner must be aged 20 to 69. We use two strategies to identify couples following Kreider (2008). The first strategy uses relationship to the ATUS respondent (i.e., “householder couples”), and the second strategy uses self-identified partners at the time of the CPS (i.e., “additional couples”). First, the majority of respondents (n = 85,565) report a married or cohabiting partner in the household at the time of the ATUS interview (same as Prickett et al. 2015). Of the householder couples, 84,546 respondents and partners were also present at the time of the CPS interview, two to five months prior. For this set of partners, age and employment status were collected during the ATUS interview; race and education were collected during the CPS interview. Six partners were missing education information, and we impute these using multiple imputation with the partner’s characteristics and the respondent’s characteristics for all couples. A much smaller number of householder couples (n = 1,019) include a partner who moved into the household between the CPS and the ATUS interviews. For these partners, we know age and employment status, but we do not have race and education information. Thus, we impute race and educational status for the partners using multiple imputation with the partner’s characteristics and the respondent’s characteristics for all couples.
We identify our additional couples, who are coresident but do not list the partner as a spouse or unmarried partner directly, by leveraging the link to the CPS and self-reported partner data. A direct cohabitation question was added to the CPS in January 2007 that asked individuals, “Does [respondent] have a boyfriend/girlfriend or partner in the household?” This question improves measurement of coresident partners, especially for cohabiting couples (Kreider 2008). Using this method, we identify 855 more cohabiting couples between 2007 and 2016, 101 of which were same-sex couples. This is important because same-sex marriage was not legalized across the United States until 2016; we identify 21.7% of same-sex couples in our sample by using the additional couples method. For theses partners, we have race, education, and employment status from the CPS and age from the ATUS interview. However, employment status is missing for four of the partners, so we impute employment status for these cases.
We identify a total of 86,420 coresident cohabiting and married individuals.1 We use respondent and partner sex as reported at the time of the ATUS interview to classify couples as same- or different-sex couples.2 Our sample includes 631 same-sex couples, of which 356 are women in same-sex relationships, and 275 are men in same-sex relationships. The weighted representation of same-sex couples is 0.89% across all years and 1.0% for 2007 forward, when we are able to identify additional couples using the CPS pointer variables.3 Table 1 shows the couple-level demographic and socioeconomic characteristics for our analytic sample.4
We have a single-day time diary from one member of the couple, but by using the co-presence data from the respondent’s time diary, we create for our analysis four measures of shared time with a partner. Co-presence is captured by ATUS interviewer asking the respondent who was in the room during activities that took place at home, who accompanied the respondent during travel or during activities done at most other locations, or who was with them during the activity (U.S. Census Bureau 2016). Total shared time is a continuous measure of the total minutes per day spent during nonwork, nonsleep, and nonpersonal activities with one’s partner regardless of who else, if anyone, was present. We also consider three subcategories of total time with one’s partner. Exclusive time measures time spent alone with a partner when no one else was present. Family time indicates time spent with a partner and one or more children for couples with children under age 18 in the home. Partner and other time comprises all remaining shared time: time spent with a partner and one or more other people who, for parents, are not their children. These measures do not include time spent working, sleeping, grooming, or in personal care because “with whom” information is not collected for these activities in the ATUS during the entire 2003–2016 period.
After analyzing couples’ shared time, we also disaggregate the total shared time into different activities to better understand whether the nature of shared time varies among different-sex, men in same-sex, and women in same-sex couples. The shared activity information is based on the ATUS respondent’s report of what they were doing when they were with their partner (as described earlier), although it does not necessarily mean that both partners were doing the same activity.5 Nonetheless, at the very least, the partner was in the room with the respondent during the reported activity or, if the activity was outside the home, they were together during the activity. We consider meals, leisure, television, childcare, housework, travel, and other activities. Meals include work and non–work-related eating; leisure includes playing sports/exercising, socializing with others, reading, playing games, volunteering, attending religious events, and attending events such as sports, movies, and parties; television includes time watching television; childcare involves caring for children; housework includes activities such as meal preparation, cooking, cleaning, laundry, home repairs, and purchasing goods and services; travel includes all activity-related travel on the diary day; and other is a combination of all other possible shared activities on the day, including adult caregiving, education-related activities, and telephone calls.
Our focal independent variable is the relationship type, which indicates whether the respondent is a member of a different-sex couple, male same-sex couple, or female same-sex couple. We also include variables in our models to control for differences by marital status and for demographic and socioeconomic factors that have been shown to influence shared time, as we discuss previously. The education variable includes three codes: both members have at least a college degree, one member has at least a college degree but the other does not, or neither member of the couple has at least a college degree. Similarly, employment status indicates whether both members are employed, one member is employed, or neither of the partners is employed. To control for additional work characteristics that could influence time shared with a partner, we include a categorical variable, respondent detailed employment status, which indicates whether the respondent is in a full-time position, in a part-time position, or not employed. Union status indicates whether the partners are married or cohabiting as unmarried partners, which could include couples in formal civil unions or domestic partnerships, although these distinctions are not available in the data. The parental status variable identifies couples with children under the age of 18 in the household. For parents, the age of youngest child variable gives the age of the youngest child in the household in the following categories: age 0–5, age 6–13, and age 14–17. There is also a categorical variable for the number of children: one, two, and three or more children in the household. Finally, has nonresidential child indicates that the respondent has a child under the age of 18 who does not live in their home. Race is a couple-level variable indicating whether both members of the couple are White, one is White and the other is non-White, or both are non-White. The income variable is categorical and is based on household income. We have six categories for income: less than $25,000; $25,000–$49,999; $50,000–$74,999; $75,000–$149,999; $150,000+; and a missing information category. Income was asked at the time of the CPS, but there was a large amount of nonresponse with the variable prior to 2010, when the CPS started imputing values for missing cases. For the years prior to when the CPS began imputing the data, we code income into a missing category, which we control for in our models in order to retain as many same-sex couples as possible. The census region variable includes the four major regions (South, West, Midwest, and Northeast) of the country. We also include a four-category variable for period to indicate pre-recessionary (January 2003–November 2007), during recession (December 2007– June 2009), and post-recessionary periods (July 2009–June 2015) as well as the post-Obergefell period (July 2015–December 2016). Our period categories capture increases in time with family members during the recession (Morrill and Pabilonia 2015) as well as the legalization of same-sex marriage across the United States (Obergefell v. Hodges2015). Finally, we include indicators for whether the diary day was on a weekend and whether it was a holiday.
We first estimate the average amount of total time shared with one’s partner, exclusive time shared with a partner only, and time spent with a partner and others (who are not coresident children) by relationship status. For parents, we also estimate the amount of time shared with children and a partner. Descriptive analyses are followed by ordinary least squares (OLS) regression estimates of the daily time that individuals in same-sex and different-sex relationships spend with their partner while controlling for demographic and socioeconomic characteristics and diary day attributes.
Figure 1 shows the weighted averages of the amount of time spent with a partner on a daily basis during nonwork, nonpersonal care, and nonsleep activities for nonparents and parents, respectively. Different-sex couples who are nonparents spend 300 minutes per day of total shared time, which is in between that for female and male same-sex couples, although differences are not significant (panel a). However, differences in time with a partner between men and women in same-sex relationships are significant. Female same-sex couples spend 321 minutes together, on average, and men in same-sex couple arrangements spend more than an hour less together, averaging 255 minutes together per day. For exclusive shared time, men in same-sex couples spend about 17 minutes less per day than those in different-sex arrangements, but female same-sex couples spend about half an hour more time together per day than different-sex couples and nearly an hour more together than male same-sex couples. Partner and other time is the residual category after exclusive time is omitted from total time. We find that different-sex couples report 67 minutes with a partner and others, and men and women in same-sex relationships report less time shared with a partner and others (39 minutes for men and 50 minutes for women).
As shown in Fig. 1, and consistent with previous research, parents spend their shared time differently than nonparents. These large differences by parental status underscore our decision to analyze parents and nonparents separately. For total shared time, female same-sex couples with children spend more total time together than female same-sex nonparents. Mothers in same-sex relationships spend nearly 3 more hours together per day than fathers in same-sex relationships and 1.4 hours more together than different-sex parents (see Fig. 1, panel b). Fathers in same-sex couples spend significantly less time together than parents in different-sex arrangements. Female same-sex parents spend the most time alone together on average compared with other parents. Women in same-sex couples spend the most time together with a partner and child, an average of almost 2 hours more than men in same-sex relationships. Neither women nor men in same-sex parent couples spend statistically significant different amounts of time than different-sex couples with their partner and children or with a partner and others.
The observed differences in shared time between individuals in same-sex and different-sex relationships may be the result of the demographic and socioeconomic characteristics (shown in Table 1). Men in same-sex relationships are more likely to include a White and non-White member, and different-sex couples are the most likely to both be non-White. Only 10% of male same-sex couples are parents with a coresident child under age 18. Different-sex couples are the most likely to have children, and they are the most likely to have children under age 5 and to have three or more children. Both female and male same-sex couples are much more likely than different-sex couples to be both working and to both have college degrees, and men in same-sex relationships have higher incomes than other couples, on average.
Analytic Results for Shared Time
We perform OLS regression to see whether differences in shared time persist across couples after accounting for the variation in demographic and socioeconomic characteristics of the couples in the sample.6 Table 2 shows the regression results by parenthood status for each of the shared time measures. The results for nonparents, in the first three columns, show that women in same-sex couple arrangements spend about 49 minutes more per day together than different-sex couples (column 1, Table 2) and 52 minutes more alone together (column 2, Table 2). Men in same-sex relationships do not spend statistically significant different amounts of total or exclusive time together than different-sex couples without children. Moreover, the coefficients are small, and this finding is similar to the mean differences in total time and exclusive time shown in Fig. 1, panel a. Men in same-sex relationships do spend the least amount of time together with others, about 15 minutes less per day than different-sex couples. Wald tests confirm that the women in same-sex relationships spend significantly more time together than male same-sex couples as well as different-sex couples.
The four columns on the right side of Table 2 show the results for estimates of total time, exclusive time, family time, and partner and other time, respectively, for couples with children under age 18. The coefficient for men in same-sex relationships is negative for all types of shared time but is only statistically significant for total time, indicating that men in same-sex relationships who have children spend about an hour less time together per day than different-sex couples with children. The lack of statistical significance in exclusive and family time is likely due partly to the small sample size. The coefficients are similar in magnitude than the differences in the means, with men in same-sex couples spending about 25 minutes less exclusive time together (column 5, Table 2) and 32 minutes less with a partner and children (column 6, Table 2) than different-sex couples. As in the case of nonparents, mothers in same-sex arrangements spend more time together than parents in different-sex couples. Female same-sex couples with children spend more than 1 hour more together (68 minutes) in total when compared with different-sex parents. Similarly, female same-sex couples spend 36 minutes more exclusive time together than different-sex couples with children. The estimates for family time also indicate that women in same-sex relationships spend more time together with their children and partner than do different-sex couples, although this difference is not statistically significant. Wald tests again show that mothers in same-sex arrangements spend more time together than fathers in same-sex arrangements. We find no differences in partner and other shared time for parents by relationship type.
The coefficients on several demographic and socioeconomic characteristics for parents and nonparents are noteworthy. Cohabiting couples spend less time together than married couples; people in their 40s and 50s spend less time together than those older and younger; and dual-earner couples spend the least amount of time together. The results also show that couples where both partners are non-White spend the least amount of time together, and couples with higher incomes spend slightly less time together.
Analytic Results by Activity
To further explore the large differences between the shared time of women in same-sex couples and the different-sex and male same-sex couples, we estimate the total shared time spent in seven broad activity categories (meals, leisure, television, childcare, housework, travel, and other) as well as all activities eligible to be performed with a partner. The total predicted minutes shared with a partner in specific activities reported by the respondent are presented in Table 3 by relationship type separately for parents and nonparents. We estimate the minutes spent in each activity with a partner present using OLS regression (identical to models in Table 2). We then predict the amount of time spent in each activity with a partner by relationship status. For each activity, we present (1) the total number of minutes spent per day in the specific activity with the partner, (2) the total number of minutes spent in the activity, and (3) the share of time spent in the activity with the partner (row 1 / row 2 per activity). Note that the sum of seven broad activity categories is not 1,440 minutes (an entire day) because respondents are not asked who was with them during sleep, paid work, and personal care activities consistently over the period.
A large amount of shared time for respondents in either coresident same-sex or different-sex partnerships comprises leisure activities and television-watching. Women in same-sex relationships consistently spend both more time together (other than television-watching) and a larger share of the time spent in these activities together, regardless of parental status, in leisure, television-watching, and meals. This pattern of female same-sex couples spending more time together and greater shares of time together holds for all activities. Figure 2 graphically shows the activity breakdown for nonparents, with the total minutes shared on the horizontal axis and the share of the total time in each activity indicated in the bars. Relative differences in the share of time spent in various activities between different-sex couples and men in same-sex couples are generally quite small.
The general patterns for couples with children echo the patterns for nonparents: women in same-sex relationships spend more time together and greater shares of time together in all activities. Thus, although we see large differences in the amount of shared time for female same-sex couples compared with other couples, there is no one particular activity in which they spend more time together than other couples; rather, they just seem to spend more time together in most activities.
Perhaps female same-sex couples spend more time together because they spend more time in activities that lend themselves to spending time together. Recall that “with whom” information is not collected for work, sleep, and personal care activities across all years of the ATUS. Although the share of female same-sex couples’ total time together does not look different in terms of the share of time spent in specific activities compared with different-sex couples or men in same-sex couples, perhaps women in same-sex couples spend much more time in total (with or without a partner) in these seven activity groups. This would be evident in the “all eligible activities” category presented at the bottom of Table 3, which excludes time in work, sleep, and personal care. However, we find that among nonparents, women in same-sex couples spend less total time in eligible activities, but more of the time is shared with their partner. Fifty-four percent of time in these activities is shared for women who are nonparents and who are in same-sex relationships. Among parents, individuals in different-sex and women in same-sex couples have nearly identical total amounts of time in eligible activities, but 52% of time in these activities is shared for women who are parents and who are in same-sex relationships compared with only 40% for different-sex couples.
Despite a growing literature on couples’ shared time and experiences (Dew 2009; Flood and Genadek 2016; Mansour and McKinnish 2014; Sevilla et al. 2012; Sullivan 1996; Voorpostel et al. 2009), prior to this examination of shared time by coresident same-sex couples, little was known about time spent together among same-sex couples in the United States. We extend this growing literature by analyzing time with a partner among men and women in same-sex relationships. This is an important time to study same-sex families and partnerships given growing visibility and acceptance as well as major legal changes regarding same-sex marriage in the United States (Gates 2013; Manning et al. 2016). Comparing same-sex and different-sex couples’ shared time allows us to understand the extent to which expectations and evidence about how much time individuals spend with a partner hold across different relationship types. This is the first study, to our knowledge, to examine these differences using population-level data, in part because of a lack of data.
We analyze same- and different-sex couples’ shared time using different-sex couples’ time together as a benchmark because the focus in the literature to date has been on different-sex couples. We find important similarities and differences after accounting for demographic and socioeconomic characteristics of couples. Men in same-sex relationships look very similar in their shared time to individuals in different-sex partnerships on all types of shared time (total, exclusive), although male parents in same-sex relationships spend significantly less total time together than different-sex and female same-sex parents. Women in same-sex relationships, on the other hand, spend substantially more time with a partner and more time alone with a partner than individuals in different-sex relationships and men in same-sex relationships, regardless of parenthood status. For parents, we also find that women in same-sex couples spend more shared time with their children present than different-sex and male same-sex couples. We find almost no differences across relationship type for time shared with a partner and others. On average, the amount of partner and other time that couples share is small per day compared with total shared time, but the limited differences in the means between same-sex and different-sex couples are explained by the demographic and socioeconomic variables in our models.
Although same-sex couples rank joint activities highly for maintaining a relationship in comparison with different-sex couples (Haas and Stafford 2005), and they tend to spend about the same share of time in different types of activities, we find that women in same-sex couple arrangements spend much more time together. More time together for women in same-sex couple arrangements is not the result of simply having more time available to spend with a partner. In fact, women in same-sex relationships spend less time overall in activities in which we know who the respondent was with, but they spend more time with their partners. The greater amount of time that women in same-sex couples spend with a partner is not devoted to any one specific activity, as evidenced by the nearly identical shares of activity-specific time shown in Fig. 2. Whether housework is shared more evenly between individuals in same-sex couples is not our focus, but we do find that women in same-sex couples spend more time per day, on average, in housework with their partner in the room than individuals in other relationship types. Mothers in same-sex relationships also perform substantially more childcare with their partner present and spend more time together with their children than fathers in same-sex relationships and different-sex couples. These results are similar to research on the division of labor within same-sex households (Patterson et al. 2004; Rothblum 2010, 2017) and may suggest that with more equal gender roles in the household, these women are able to spend more time together.
Our work raises new questions about the relationship between couples’ shared time and relationship quality. We find small differences in shared time between men in same-sex couple arrangements and individuals in different-sex arrangements, which provides some evidence in support of the claim that relationship quality is similar for same- and different-sex couples (Kurdek 2005, 2006; Peplau and Spalding 2000) given the association between shared time and relationship quality (Amato et al. 2007). Our work, however, also shows large differences between women in same-sex couple arrangements and other couples after demographic differences, including gender, are accounted for. These results indicate that women in same-sex couples have more daily time with their partners co-present, on average. Whether these differences indicate that women in same-sex couples have greater relationship quality and compatibility compared with different-sex couples or men in same-sex couples (Baiocco et al. 2015; Balsam et al. 2008) is an open question. On the one hand, research suggests that relationship quality is similar between same-sex and different-sex couples (Kurdek 2004, 2005, 2006). On the other hand, there is evidence that spending more time together is suggestive of greater relationship quality (Amato et al. 2007) and may result in more enjoyable and less stressful time throughout the day (Flood and Genadek 2016, 2020; Sullivan 1996). The big question that emerges from our work is whether the greater time that women in same-sex couples spend together translates into higher relationship quality. Unfortunately, the ATUS lacks a measure of relationship quality to understand the impact of these time differences.
This study has limitations, and some of our results for same-sex parents should be interpreted cautiously. Despite the large sample sizes of the ATUS, our sample contains only 41 men who are parents and are in same-sex coresident partnerships. Although we do not see statistically significant differences between men in same-sex relationships and individuals in other relationship types, there is a statistical power issue that may be prohibiting our models from detecting some actual differences. A limitation of these data is that they are cross-sectional, with one diary day: we observe respondents’ behavior at only one moment in time and cannot assess change in shared time over an individual’s (or couple’s) life course. Nor do we have information on relationship duration. In addition, we can examine only coresident relationships using these data. Finally, we use reports of shared time from only one member of the couple, which is potentially problematic given discrepancies between survey-based responses to questions about the division of household labor and direct observation of same-sex couples’ daily lives (Carrington 1999). Fortunately, however, diary-based measures of time use—for example, usual hours of paid work and time spent reading to children—are less subject to social desirability bias than survey-based measures (e.g., Hofferth 2006; Robinson 1985). Furthermore, our results are unlikely to have been different if we had time diaries from the other member of the couple or both members of the couples. The limited research in the United States using diaries from couples shows that partners report similar amounts of shared time (Freedman et al. 2012a), although this finding may be due to the composition of the particular sample of older adults. Couple-level reports of time with a partner, if available, would still enhance our study; for example, we would be able to examine similarity in reporting within same-sex relationships and whether individuals report doing the same activities as their partners at the same time.
Given that this study is the first to investigate shared time among same-sex couples using a large nationally representative time diary study, it is intended to serve as a springboard for future research on couples’ shared time. This line of research is ripe for investigation, especially considering differences within same-sex couples; more nuanced examinations of the effects of minority stress and stigma on same-sex couples’ shared time are also warranted. Our analyses of broad activities of shared time do not suggest major differences in how couples spend their time by relationship type. Further investigation might include more specific types of activities and utilizing more “with whom” records to test the implication of stigma and discrimination on sharing time in social activities and family activities versus home-related or more isolated activities. Related to stigma, the ATUS data could also be used to look at the effects of same-sex marriage legalization on whom activities are done with and what same-sex couples are doing together.
This research was supported by the NICHD-funded Minnesota Population Center (P2C HD041023), the University of Colorado Population Center (P2C HD066613), and IPUMS Time Use (R01HD053654). In addition, Joan García-Román was supported by the Beatriu de Pinos Post-Doctoral Fellowship (2016 BP-00279) and is also supported by the I+D Project “Family strategies and demographic responses to the economic recession” (CRISFAM) (CSO2015-64713-R) and the CERCA Programme from the Generalitat de Catalunya.
We are not able to identify individuals in same- and different-sex relationships who live apart.
Despite concerns about the misidentification in many federal surveys of same-sex couples resulting from data errors, the concerns are smaller in the ATUS because for most, respondents’ and household members’ sex has been asked three times by the time of the ATUS. Authors’ investigations of the imputation and allocation of sex in the ATUS show just one of 86,420 respondents was imputed because of missing information.
The total percentage of same-sex couples is very similar to estimates from the Census Bureau from the American Community Survey (ACS), which range from 0.93% of couples in 2009 to 1.4% of couples in 2016 (U.S. Census Bureau 2016).
Characteristics for the individual respondents and partners can be obtained from the authors.
Based on separate time diaries from both members of couples in the Panel Study of Income Dynamics, Freedman et al. (2012a) showed 76% agreement between couples’ reports of activities with a partner. Similar work using the 2014–2014 United Kingdom Time Use Survey (Vagni 2019) showed about an 80% overlap in the time that couples reported being with a partner and doing the same activity.
Tobit models, propensity score matching models, and two-step models produced results very similar in magnitude and significance. These results are available upon request.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.