Abstract

Most research on the stability of adult relationships has focused on coresidential (cohabiting or married) unions and estimates rates of dissolution for the period of coresidence. Studies examining how the stability of coresidential unions differs by sex composition have typically found that same-sex female couples have higher rates of dissolution than same-sex male couples and different-sex couples. We argue that the more elevated rates of dissolution for same-sex female couples are a by-product of the focus on coresidential unions. We use data from the National Longitudinal Study of Adolescent to Adult Health to compare rates of dissolution based on the total duration of romantic and sexual relationships for same-sex male couples, same-sex female couples, and different-sex couples. Results from hazard models that track the stability of young adult relationships from the time they are formed demonstrate that male couples have substantially higher dissolution rates than female couples and different-sex couples. Results based on models restricted to the period of coresidence corroborate the counterintuitive finding from earlier studies that female couples have the highest rates of dissolving coresidential unions. This study underlines the importance of comparisons between these couple types for a better understanding of the role that institutions and gender play in the stability of contemporary relationships.

Introduction

Scholarly and media attention to same-sex relationships (i.e., relationships between partners of the same biological sex) has risen dramatically in recent years, reflecting increases in their visibility and validation. The prevalence of same-sex cohabitation in the United States doubled in the last decade, and in 2016, 7.3 % of millennials reported that they were lesbian, gay, bisexual, or transgender (Gates 2017; Lofquist et al. 2012). The share of American adults who consider a same-sex cohabiting couple to be a family grew to such an extent that by 2010, approximately two-thirds of Americans included one or more types of same-sex couple (e.g., a same-sex female couple with children) in their definition of family (Powell 2017; Powell et al. 2010). The legal environment supporting marriages of same-sex couples has also fundamentally shifted, as evidenced by the 2015 U.S. Supreme Court ruling that declares marriage to same-sex couples a constitutional right in the United States. Against this backdrop, a recent Institute of Medicine report (IOM 2011) called for more research on how the health and well-being of adults and children differ across a variety of living arrangements, including households with same-sex couples. Given that health and well-being are associated with the stability of romantic and sexual relationships, it is essential to compare same-sex relationships with different-sex relationships in terms of stability and identify factors that explain any differences.

Comparisons of same-sex and different-sex couple stability are motivated by public health concerns as well as the insight that such comparisons provide on the importance of social and legal institutions for the outcomes of contemporary relationships. Contrasting relationships between two men and two women, in particular, reveals how gender influences relationships (e.g., Blumstein and Schwartz 1983). Yet, prior studies comparing these couple types have tracked the stability of only select samples of couples (e.g., coresidential relationships). Overall, same-sex couples (both male and female) have higher rates of dissolution than different-sex couples (e.g., Lau 2012). This finding is consistent with a minority stress perspective, articulating the unique challenges same-sex couples face because of the stigmatized status of their relationship (Frost and Gola 2015). Most research has shown that relationships with two women have higher rates of dissolution than relationships with two men (e.g., Andersson et al. 2006). The finding that same-sex female relationships are the least stable (compared with different-sex and same-sex male relationships) belies decades of theory and research suggesting that women are, on average, more relational than men. It begs the question of whether patterns found in previous studies are an artifact of differential selection of lesbians and gays into more established relationships.

Data from the National Longitudinal Study of Adolescent to Adult Health (Add Health) offer an unprecedented opportunity to compare the stability of same-sex and different-sex relationships for a cohort of young adults that came of age when state courts and the media were beginning to debate the rights of same-sex couples. Importantly, these data allow us to track the stability of relationships from the time they are formed. The fourth wave of Add Health (2007–2008) obtained detailed information from young adult respondents on their most recent romantic and/or sexual relationship, including the biological sex of their partner and the duration of their relationship. We compare the rates at which same-sex male, same-sex female, and different-sex relationships dissolve from their start (e.g., first date or sexual involvement) and then show how patterns differ during the period of coresidence. Although a growing number of surveys permit an examination of same-sex couple stability, their sample sizes require that researchers pool a broad age range of respondents in same-sex relationships. Because these respondents are, on average, older and in relationships of longer duration, these surveys represent same-sex couples that are hardier. Thus, they may not provide the most accurate barometer of the challenges that same-sex couples face, particularly those composed of young adults.

Theories of Same-Sex Relationship Stability

Theories on the stability of same-sex relationships have evolved from decades of research comparing them with different-sex relationships. These theories assume that the correlates of relationship stability are, for the most part, similar for same-sex relationships and different-sex relationships (Moore and Stambolis-Ruhstorfer 2013). Differences in the stability of same-sex and different-sex relationships are therefore a function of differences in their rewards, alternatives, and/or barriers (see Lau 2012). A fundamental way in which same-sex couples have differed from different-sex couples is in their right to legally marry (Herek 2006). To the extent that marriage engenders enforceable trust and relationship investments (Cherlin 2004), same-sex couples are destabilized by their low rates of marriage. Prior to the fourth wave of Add Health (2007–2008), same-sex couples were entitled to civil union or domestic partner benefits in six states and legal marriage only in Massachusetts. Few same-sex coresidential couples (i.e., roughly one-fifth) in the United States were in legally recognized relationships (domestic partnerships, civil unions, marriages) when the Add Health fourth wave was fielded (Badgett and Herman 2013; Gates 2009); the share of legally recognized unions among young adults at this time was likely smaller.

Same-sex couples not only have fewer legal barriers to dissolution but experience more stress in their day-to-day lives (Meyer 2003). According to the minority stress perspective, sexual minorities (e.g., gays, lesbians, and bisexuals) and members of other socially disadvantaged groups are exposed to unique stressors (e.g., interpersonal and institutional discrimination) that can diminish the quality of their romantic relationships (Doyle and Molix 2014, 2015; Frost and Meyer 2009). This perspective also highlights couple-level stressors that partners must navigate (LeBlanc et al. 2015). For example, same-sex couples must negotiate and manage the visibility of their relationship among friends, family members, and coworkers. Although different-sex couples must also decide when and how to introduce partners to third parties, they are less likely to anticipate and experience disapproval of their relationship (Frost and Gola 2015; Rostosky et al. 2007). Men and women in same-sex relationships appear to be similarly exposed to most minority stressors; however, men in same-sex relationships experience more stress related to violence and harassment, whereas women in these relationships experience more stress due to family reactions (Rothblum 2009; Todosijevic et al. 2005). To the extent same-sex couples are more likely than different-sex couples to anticipate disapproval, they may be less likely to invest in their relationship by moving in together or integrating partners into their social networks—investments that potentially stabilize relationships (Lau 2012).

An obvious way in which same-sex and different-sex relationships differ is in their sex composition. According to the gender-as-relational perspective, the dynamics of relationships “unfold in different ways for couples with two men in comparison to couples with two women, or couples with one man and a woman” (Umberson et al. 2015a, b:9). Typically, same-sex relationships reinforce gendered tendencies, and different-sex relationships challenge these tendencies (Umberson and Kroeger 2016; Umberson et al. 2015a, b). In a recent qualitative study, Umberson and colleagues (2015a, b) compared gay, lesbian, and heterosexual couples and found oppositional patterns of emotional intimacy and autonomy in male and female couples, reflecting the fact that partners in these couples were more concordant in relationship preferences than partners in different-sex couples. Female partners in same-sex couples spoke of sharing their thoughts and feelings and attending to each other’s emotional needs to enhance emotional intimacy and minimize boundaries. Male couples discussed expending considerable effort respecting boundaries and providing emotional space. For many different-sex couples, the differential value that partners placed on emotional intimacy and autonomy was a source of contention (Umberson et al. 2015a, b). If efforts to enhance intimacy and minimize boundaries promote the well-being of partners and bind them, we would expect female couples to have lower rates of dissolution than both male couples and different-sex couples.

Research on Same-Sex Relationship Stability

Empirical comparisons of same-sex and different-sex couples stability are limited in scope (Table 1). Fewer than 20 papers, chapters, books, or reports contrast same-sex and different-sex couple stability, and the bulk focus on coresidential unions. Many studies used data collected decades ago; consequently, they cannot depict contemporary patterns. These studies combined individuals from a wide age range and thus failed to distinguish the relationships of younger respondents. Further, the population-based samples were drawn primarily from European countries, specifically Norway, Sweden, Britain, and the Netherlands (e.g., Andersson et al. 2006; Kalmijn et al. 2007; Lau 2012; Noack et al. 2005). Given the challenges in identifying sufficient numbers of same-sex couples in random samples, early U.S. studies were based on convenience samples (Blumstein and Schwartz 1983; Kurdek 1998, 2004). Some recent U.S. studies relied on population-based samples (i.e., Badgett and Herman 2013; Balsam et al. 2008; Gates 2006; Manning et al. 2016; Rosenfeld 2014; Weisshaar 2014). Although U.S. studies based on national survey data—namely, How Couples Meet and Stay Together (HCMST) and the Survey of Income and Program Participation (SIPP)—provide new insights into the stability of same-sex relationships, their samples include a wide range of ages and relationship durations (two weeks to 50 years in the HCMST). Because the HCMST and SIPP samples are left-truncated, they are disproportionately composed of couples that are more stable (Guo 1993).

Many of the studies included in Table 1 show that same-sex coresidential unions (cohabiting and formalized) have greater instability than different-sex married unions (Andersson et al. 2006; Balsam et al. 2008; Blumstein and Schwartz 1983; Gates 2006; Kalmijn et al. 2007; Kurdek 1998; Lau 2012; Manning et al. 2016). It is difficult to generalize from these findings because the contrast group—married unions—is more stable than both same-sex and different-sex cohabiting unions (Manning et al. 2016). To date, no evidence exists that legalized same-sex relationships are less stable than legalized different-sex relationships. Analyses using U.S. survey data (HCMST) indicate that same-sex and different-sex couples in marriages or marriage-like relationships (e.g., civil unions and domestic partnerships) have similar rates of dissolution (Rosenfeld 2014). In contrast, analyses based on registry data in England and Wales suggest that same-sex registered partnerships are more stable than different-sex marriages (Ross et al. 2011). Aggregate-level analysis of U.S. administrative data indicates that dissolution rates are higher among different-sex marrieds than same-sex couples in legally recognized unions (Badgett and Herman 2013). These differences may stem in part from the selection of individuals who form same-sex legally recognized relationships (e.g., older and more-educated individuals).

Given that same-sex unions have only recently been legally recognized in the United States, it is critical to highlight studies that are more inclusive in their samples. Besides, marriage occurs late in the romantic relationships of young adults (Sassler 2010) and late in young adulthood (Manning et al. 2014). Ten of the studies listed in Table 1 included cohabiting relationships in their samples (Balsam et al. 2008; Blumstein and Schwartz 1983; Gates 2006; Kalmijn et al. 2007; Kurdek 1998, 2004; Lau 2012; Manning et al. 2016; Rosenfeld 2014; Weisshaar 2014). Only one of these studies additionally included romantic and/or sexual relationships that did not involve coresidence (Rosenfeld 2014). European studies focusing on cohabiting relationships have found that same-sex couples are less stable than different-sex couples and that differences in the stability of these couples persist in models including a raft of control variables (Kalmijn et al. 2007; Lau 2012). Older U.S. studies similarly found greater instability on the part of same-sex cohabiting couples relative to different-sex cohabiting couples (Blumstein and Schwartz 1983; Kurdek 1998, 2004). In contrast, studies based on more recent U.S. data found that same-sex and different-sex unmarried couples have similar dissolution rates. Relying on data from SIPP, Manning et al. (2016) found that same-sex and different-sex cohabiting couples share comparable dissolution rates. Descriptive analyses of HCMST indicate that same-sex unmarried couples experience higher rates of dissolution than different-sex unmarried couples during the first five years of the relationship but lower rates thereafter (Rosenfeld 2014).1

Considering the role of gender, most of the studies listed in Table 1 (nearly all which are coresidential) document lower rates of dissolution among couples with two men than among couples with two women (Andersson et al. 2006; Blumstein and Schwartz 1983; Carpenter and Gates 2008; Gates 2006; Noack et al. 2005; Office for National Statistics 2013; Rosenfeld 2014; Ross et al. 2011). In only a few studies were similar rates of dissolution reported for male and female couples (Kurdek 2004) or lower rates of dissolution reported for female couples (Kalmijn et al. 2007; Lau 2012). At the same time, there is evidence of gender differences in the formation of coresidential unions, with fewer gay men than lesbian women classified as married or cohabiting at the time of interview (Carpenter and Gates 2008; Rosenfeld 2014). Coresidential unions with two men appear to be more select with respect to stabilizing factors (e.g., commitment) than unions with two women (Lau 2012).

In sum, the stability of same-sex couples appears to rest in part on whether they are in a legally married, formalized (registered partnership, domestic partnership), cohabiting, or dating relationship. The finding that legalized same-sex relationships are just as stable as legalized different-sex relationships, if not more stable, suggests that marriage has a protective effect on same-sex relationships—or alternatively, that more stable same-sex couples gravitate toward marriage. The greater instability of same-sex couples in dating and cohabiting relationships is consistent with a minority stress perspective that emphasizes the challenges of being in a stigmatized relationship. Gaps in the stability of same-sex and different-sex unmarried relationships appear to be weakening over time and in some recent studies are not significant. Most prior studies on same-sex couple stability found that female couples are less stable than male couples, a finding that runs counter to the gender-as-relational perspective. Given their emphasis on emotional intimacy and minimizing boundaries between partners, female couples should be more stable than male couples. Again, the findings of prior studies examining same-sex couple stability are potentially biased by differential selection into more established relationships.

Current Investigation

This study examines the stability of the most recent romantic and/or sexual relationship reported by young adult Add Health respondents during the fourth wave (2007–2008). Tracking the stability of relationships from the point of their initiation removes the bias posed by differential selection into more established relationships. We distinguish three groups of relationships based on respondent’s own sex (marked by the interviewer) and the sex of their most recent partner (marked by the respondent): different-sex, same-sex female, and same-sex male. As in previous research, we do not distinguish different-sex couples according to the respondent’s sex. We compare the stability of these three couple types from the point of initiation and during the period of coresidence. This strategy helps address whether patterns in previous studies are due to sample selection.

We test two specific hypotheses. Our first hypothesis is that male couples will have higher rates of dissolution than different-sex couples. This hypothesis is motivated by theory and research suggesting that male couples are uniquely exposed to minority stressors, such as harassment, and that they place a greater premium on personal space and autonomy. Our second hypothesis is that gaps in the rates at which male couples and different-sex couples dissolve will be smaller during the period of coresidence. This hypothesis assumes that male couples transitioning to coresidence are more select with respect to factors that stabilize relationships (e.g., relational orientation, age, and socioeconomic status) than are different-sex couples transitioning to coresidence. It is less clear how female couples will compare with different-sex couples in terms of stability. Female couples, like male couples, are exposed to minority stressors that are potentially destabilizing. However, their efforts to enhance partner well-being and merge could be stabilizing.

To test these hypotheses, we estimate hazard models that contrast the dissolution rates of young adult male and female couples with those of different-sex couples before and after including variables that are associated with relationship stability (e.g., demographic heterogamy) and have been found to distinguish same-sex and different-sex couples (Jepsen and Jepsen 2002; Rosenfeld and Kim 2005; Schwartz and Graf 2009). We then contrast the dissolution rates of couples during the period of coresidence. We supplement these analyses with a number of sensitivity tests. For instance, Add Health’s collection of detailed information on sexual behavior and sexual identity allows us to examine the sensitivity of results to inconsistent reporting. These sorts of comparisons are the hallmark of demographic studies on sexual minorities (Black et al. 2000; Lau 2012).

Data and Sample

Add Health is a longitudinal school-based study (Harris et al. 2009). Using rosters from each school, Add Health selected a nationally representative (core) sample of 12,105 adolescents in grades 7 to 12 to participate in the first in-home interview completed in 1995. In 2001 and 2002, Add Health reinterviewed 15,197 of the Wave 1 in-home respondents. Between 2007 and 2008, Add Health conducted the fourth wave of in-home interviews with 15,701 respondents (80.3 % response rate) who were between the ages of 24 and 32.

Add Health used audio computer-assisted self-interviewing (ACASI) and partner rosters to identify all the partners with whom respondents had ever experienced a “romantic or sexual relationship” that eventuated in pregnancy, cohabitation, or marriage. Respondents were then asked to identify any additional partners with whom they had a romantic or sexual relationships since 2001. Toward this end, Add Health asked respondents to provide the initials of each romantic/sexual partner and subsequently report their gender, age, and race/ethnicity. ACASI not only maximizes privacy but also allows for more complicated skip patterns (Paik 2015). The partner rosters produce little discrepancy in the number of different-sex partners reported by men and women (results not shown).

Add Health targeted the most recent partner (the focal partner) for detailed questions about the relationship, including its entire length.2 At the fourth wave, 14,800 respondents had valid weights; 14,346 of these respondents reported having had a romantic and/or sexual relationship. Twenty-four of these respondents failed to provide information necessary to classify their most recent relationship (e.g., sex of their partner). Among the 14,322 respondents with such information, 170 men and 175 women were classified as having a same-sex partner; 6,509 men and 7,468 women were classified as having a different-sex partner. We further restrict the sample to 14,005 respondents who supplied information on the duration of their entire relationship. Our final sample includes 169 men and 166 women with a same-sex partner, and 6,339 men and 7,331 women with a different-sex partner.

Dependent Variable

Relationship Stability

We estimate the rates at which couples dissolve their most recent relationship following its formation using life table methods. These methods require information on whether the respondent defined the relationship as current at the time of interview (i.e., censored) and the duration of the relationship at the time of dissolution (if dissolved) or interview (if current).3 To obtain information on relationship duration, Add Health asked respondents, What is “the total amount of time that you (have been/were) involved in a romantic or sexual relationship with {initials}?” Respondents were allowed to report the time in days, months, and/or years; we use this information to compute the duration of the relationship in months. Note that this measure of duration represents the total amount of time they have been a couple. For instance, for respondents who coresided with a partner, this represents the time from dating through coresidence. We measure the length of the coresidential relationships on the basis of beginning and end dates that respondents gave for coresidence and marriage. Because few respondents contributed years of exposure beyond the seventh year, we limit the period of risk by censoring respondents whose relationships extended beyond seven years at 84 months.4

Independent Variables

Individual Characteristics

Our models include a number of relevant variables, including age at Wave 4.5 Race of respondent was collected from the first wave of the study and recoded to a series of dummy variables (non-Hispanic black, Hispanic, and non-Hispanic other vs. the omitted reference category of non-Hispanic white). Based on the household roster at Wave 1, we establish whether respondent was living with both biological parents in adolescence. Family socioeconomic status (SES) is based on a measure developed by Bearman and Moody (2004) that incorporates information on parental education and occupation from Wave 1. As a measure of adolescent context, we use the Wave 1 school administrator’s report of whether the school was in a rural, urban, or suburban area. Relationship history variables include whether respondents were ever married prior to the relationship and their logged number of lifetime sex partners (excluding their focal partner if the most recent relationship was sexual).

Heterogamy

We use respondents’ reports on partner characteristics to create measures of whether they were a different race (using the aforementioned categories) and the age difference (absolute value) between partners.

Ultimate Status of Relationship

To conduct some of the sensitivity checks, we create a categorical variable capturing the eventual status of the most recent relationship at the time of dissolution or Wave 4 interview based on several questions. The categories for ultimate status are hookups (“only having sex”); dating casually (i.e., “dating frequently, but not exclusively” or “dating once in a while”); dating seriously (“engaged” or “dating exclusively”); cohabiting; and married. We also create a time-varying covariate for marriage that is included in a model of stability for the coresidential period, as discussed later.

Parenthood Within the Relationship

For descriptive purposes, we distinguish whether respondents had a biological child with their most recent partner.

Consistency of Reports at Wave 4

Previous studies have suggested that some respondents inadvertently select the wrong sex when answering questions about their partner (Black et al. 2000). Prior to the section in which respondents provided information on recent partners, Add Health respondents were asked about their sexual identity and how many male and female sex partners they had during their lifetime. Based on this information, we identify a subset of respondents who were consistent in their reports of sexual orientation across two different sections of the questionnaire. Respondents in same-sex relationships are defined as consistent if they indicated they have ever had at least one same-sex sex partner and were not “100 % heterosexual (straight).” Respondents in different-sex relationships are defined as consistent if reported they were 100 % straight and had no same-sex sex partners in their lifetime. Albeit not a perfect flag, our indicator of consistent reporters enables us to examine how results differ when potentially inaccurate reporters are excluded from the sample.

A recent study suggested that differences in the wording of questions used to define the end of marriage may lead to different conclusions about demographic disparities in marital dissolution (Tumin et al. 2015). Add Health asked respondents to provide the end dates of their relationships, even if they had earlier reported that the relationship was current. The questions for the end dates offered respondents the option of selecting “relationship not ended.” Respondents are flagged as having an inconsistent report on the status of the relationship if they selected current but then later provided an end date of the relationship that preceded the interview by more than one month. Other respondents are flagged as having an inconsistent report if they failed to select current but then later selected relationship not ended when asked to provide a relationship end date.

Analysis Plan

We first present descriptive statistics on the dependent and independent variables for the three groups of respondents: male and female respondents with different-sex partners (different-sex couples), female respondents with female partners (female couples), and male respondents with male partners (male couples). We also display their cumulative estimates of dissolution based on life table methods. Next, we present results from two sets of hazard models of dissolution, juxtaposing zero-order and multivariate effects. The first set of these models estimates the relative hazards of dissolution from the start of the relationship, and the second set estimates the hazards of dissolution from the start of coresidence. Results from these models test whether significant differences exist in the hazard rates of dissolution for male and female couples in comparison with different-sex couples. As a supplement, we show how differences in the relative hazard rates of dissolution for male and female couples vary according to a series of sensitivity tests.

Results

Table 2 presents descriptive statistics that are adjusted for survey design effects. Recall that our sample consists of young adult respondents who reported a relationship at Wave 4 and that we focus on their most recent relationship. We indicate the level of significance for differences between same-sex and different-sex couples. Table 2 first displays whether the relationship dissolved prior to Wave 4 and its duration (the number of months since the partners became romantically or sexually involved). Male couples are significantly more likely than different-sex couples to have dissolved their relationship (38 % vs. 19 %, respectively); however, female couples do not differ significantly from different-sex couples in this respect (22 % vs. 19 %, respectively). Male and female respondents in same-sex relationships reported durations that are significantly shorter than those reported by respondents in different-sex relationships (26 months and 35 months, respectively, vs. 59 months). These same statistics for the period of coresidence (for respondents who coreside) reveal a different pattern of results: male couples are similar to female couples in terms of their likelihood of dissolving during the period of coresidence (roughly 20 % dissolve) and the average duration of their coresidential relationship (approximately 40 months). Different-sex couples are less likely than same-sex couples to have dissolved during the period of coresidence (12 %) and have coresided for a significantly longer period (more than 55 months).

Contrasts of the demographic variables for same-sex and different-sex couples indicate some significant differences. Male respondents with male partners are less likely than respondents with different-sex partners to have been married prior to the relationship and reported a greater number of prior sexual partners, on average. In contrast with respondents with different-sex partners, male respondents (but not female) with same-sex partners are more likely to be in an interracial relationship. Male and female respondents in same-sex relationships differ more from their partners in terms of age than do respondents in different-sex relationships. Similar shares of respondents in same-sex and different-sex relationships indicate that they were in a hookup or casual dating relationship; however, male respondents in same-sex relationships are most likely to indicate that they were dating seriously. Respondents in different-sex relationships are most likely to have coresided (73 %), followed by female respondents in same-sex relationships (65 %), and male respondents in same-sex relationships (48 %). A small share of male and female respondents in same-sex relationships married their partner (5 % and 11 %, respectively) versus 46 % of respondents in different-sex relationships.

Although almost two-fifths of respondents in different-sex relationships reported having a birth with their partner, virtually no male respondents in same-sex relationships and less than one-tenth of female respondents in same-sex relationships did so.6 Births for female couples are concentrated among the subgroup that was inconsistent in their reports of sexual orientation (results not shown).7

Figure 1 shows results from life table analyses of dissolution for the three groups of couples. The x-axis shows the number of years since the start of the relationship, and the y-axis shows the weighted proportion of relationships that have dissolved by a given year. This figure reveals that male couples are most likely to dissolve, and different-sex couples are least likely to do so; female couples fall between these other two groups but are closer to different-sex couples. These results support our first hypothesis that male couples would have higher rates of dissolution than different-sex couples.

Figure 2 shows results from life table analyses of dissolution for the period of coresidence for relationships that transition to cohabitation or marriage. It is like Fig. 1 except that the x-axis shows the number of years since the start of the coresidential relationship. Figure 2 reveals that male and female couples continue to have greater rates of dissolution than different-sex couples during the period of coresidence. However, male couples and female couples share similar likelihoods of dissolving coresidential relationships. In general, these patterns are consistent with the notion that coresidence attracts partners with stabilizing characteristics to the relationship and that it creates a barrier to dissolution. As predicted by the second hypothesis, percentage differences between male couples and different-sex couples in the likelihood of dissolution are smaller during the period of coresidence.

We present the results from design-adjusted Cox models of dissolution in Table 3. All four models capture differences between the three different couple types using two separate indicator variables for male and female couples, with different-sex relationships serving as the omitted reference group. Results from Model 1 show that at the zero-order level, both male and female couples have a significantly higher monthly hazard rate of dissolution than different-sex couples. Specifically, female couples have a hazard rate that is 63 % higher (i.e., 63 % = 0.63 × 100 = 1.63 – 1), or 1.63 times (p < .05) the hazard rate of different-sex couples. Male couples have a hazard that is 241 % higher, or 3.41 times (p < .001) the hazard of different-sex couples. Results from Model 2 demonstrate that the hazard ratios for male and female couples diminish in magnitude after we control for individual and relationship characteristics. With the inclusion of the control variables, the hazard ratios for female and male couples are 1.48 and 3.10 (p < .001), respectively. In fact, the female couple ratio falls out of significance. The unusually large hazard ratio for male couples is consistent with the first hypothesis.8

Several other variables have significant associations with the dissolution rate of most recent relationships at a p < .05 or higher level of significance. Black respondents have a higher hazard of dissolution than do white respondents. The hazard of dissolution decreases with age at interview. Higher values of family SES are associated with a higher hazard of dissolution in the multivariate models. Individuals have a higher hazard of dissolving their relationship if they accumulated more sex partners prior to its formation. As expected, differences in both the race and age of partners translate into higher rates of dissolution. The declines in the hazard ratios for male and female couples across Models 1 and 2 are largely driven by number of prior sex partners and heterogamy by race and age.

Results from Model 3 show hazard ratios at the zero-order level for the period of coresidence. Many of the same covariates that are significantly associated with the dissolution of the entire relationship are associated with the dissolution of the coresidential relationship (i.e., whether black, age, number of sex partners, and heterogamy). Consistent with Fig. 2, male and female couples have more than twice the monthly hazard rate of dissolution (at a p < .05 and p < .001 level, respectively) as different-sex couples (2.46 for male couples and 2.66 for female couples). Results from Model 4 show, once again, that the hazard ratios for male and female couples decrease with the inclusion of control variables and decline in their levels of significance. Specifically, the hazard ratios for male and female couples are 1.90 and 2.09 (p < .01), respectively. The declines in the hazard ratios across Models 3 and 4 are once again due to the inclusion of information on number of prior sex partners and heterogamy by race and age. The findings from these models are consistent with the second hypothesis that predicted gaps between male couples and different-sex couples would be smaller during the period of coresidence. Models that include a time-varying indicator of legal marriage (results not shown) suggest that marriage is associated with greater stability. In these models, female couples continue to have a higher hazard rate, and the ratio is reduced in magnitude (1.70) yet still significant. The hazard ratio for male couples is also diminished in magnitude (1.39).

Sensitivity Tests

To investigate the robustness of these findings, we estimate Model 2 of Table 3 for five additional samples. Figure 3, a forest plot, displays a pair of 95 % confidence intervals for the hazard ratios of dissolution based on the models for each sample. The first interval of each pair shows the likely values of hazard ratios corresponding to female couples; the point estimate of this ratio is marked by a circle. The second interval shows the likely values of the hazard ratios corresponding to male couples, with the point estimate marked by a square. Values for point estimates greater than 1 indicate a higher hazard of dissolving the relationship relative to different-sex couples, and values less than 1 reflect a lower relative hazard. Intervals that include 1 (i.e., intervals that cross the y-axis on the left) indicate that the relative hazard ratio fails to significantly differ from 1; a hollow marker also indicates that the hazard rate for same-sex couples fails to significantly differ from the hazard rate for different-sex couples.

To place these results in perspective, we first redisplay the results from Table 3 (labeled Most Recent Relationships). Again, the point estimates are 1.48 for female couples and 3.10 (p < .001) for male couples (vs. different-sex couples). The next five sets of intervals are based on models that restrict the sample in different ways. First, we estimate a model that removes respondents who were inconsistent in their reports of partnering. The estimates for consistent reporters (the second set of intervals) are slightly further away from 1 than the estimates for the broader sample, but the patterns of significance are the same. Next, we estimate a model that instead removes respondents who were inconsistent in their reports of whether the relationship dissolved. Once again, the results barely change.

Third, we remove respondents who selected the response category “only having sex with partner.” Excluding these respondents increases the hazard ratios slightly and produces a significant difference between female couples and different-sex couples. Additional analyses (not shown) reveal that hookups are the least stable of relationships on the ultimate status spectrum for different-sex couples (but less distinct among same-sex couples). Therefore, removing this group increases the instability of female couples relative to different-sex couples. Fourth, we exclude from the sample both hookups and casual dating relationships so that the sample comprises only serious dating relationships and coresidential unions. The patterns for this model are like those for the model excluding hookups, suggesting this added restriction is inconsequential.

Finally, we estimate the model only for respondents who eventually cohabit or marry. Restricting the sample to those who coreside alters the ratios (and sample sizes) for male and female couples dramatically, reducing the gap between male and female couples. Precisely, female couples have 2.20 times the hazard of dissolution as different-sex couples, and male couples have 1.90 times the hazard of dissolution as different-sex couples. For the coresidential sample, female couples differ significantly from different-sex couples, but male couples do not. This pattern is consistent with the results in Model 4 of Table 3 and offers additional support for the second hypothesis.

Discussion

Differences in the stability of same-sex and different-sex relationships vary across prior studies, reflecting their period, country, and selection of relationships. One of the most consistent patterns in prior studies is arguably the most perplexing: female couples are less stable than both male couples and different-sex couples. As Blumstein and Schwartz (1983) first noted, this finding runs counter to the notion that women are more relational than men. However, samples used in previous studies comparing the stability of same-sex and different-sex couples have primarily included relationships that were ongoing at the time of first observation and that had attained a specific status (coresiding or legally recognized). Data from Add Health gave us the unusual opportunity to track the stability of young adult same-sex and different-sex relationships from the date they were formed and to differentiate the period of coresidence. This strategy improves on prior work that analyzed the stability of relationships that were ongoing at the time of the initial interview and typically coresidential by allowing us to assess the bias posed by differential selection into coresidential relationships.

Our first hypothesis was that male couples would have higher rates of dissolving their most recent relationship than different-sex couples. The rationale for this hypothesis is that compared with different-sex couples, male couples place greater emphasis on individual autonomy and that they are uniquely exposed to minority stressors. Hazard analyses for our sample of young adult relationships demonstrated that male couples had much higher monthly rates of dissolution than different-sex couples. These patterns were observed across several different specifications of the analytic sample, including when the sample was purged of respondents with inconsistent reports on sexual orientation. Female couples also exhibited higher rates of dissolution than different-sex couples, but the differences were explained in multivariate models.

We also hypothesized that gaps between male couples and different-sex couples in rates of dissolution would be smaller during the period of coresidence than during the entirety of the relationship. This second hypothesis assumes that male partners who coreside are relatively more select than different-sex couples in terms of stabilizing characteristics. Indeed, models distinguishing the period of coresidence for our sample of young adult relationships offered a different picture of how sex composition shapes couple stability. Female couples exhibited significantly higher dissolution rates than different-sex couples of dissolving coresidential relationships, but male couples did not. The dissolution rates of female couples were only slightly higher than those of male couples, but power for detecting significant differences was greater for female couples. These results demonstrate the need to consider differential selection into coresidential relationships when assessing relationship stability, especially for male couples.

The finding that female couples were the least stable of the three types of young adult relationships during the period of coresidence is consistent with many prior studies focused on coresidential unions, but it is somewhat at odds with the notion that female couples are more relational than other couple types. To the extent that women are more relational, female couples may monitor their relationships more closely (Umberson and Kroeger 2016). As documented in prior studies, married women are ultimately more likely than married men to initiate separation (Amato and Previti 2003). In addition, issues specific to coresidence may pose special challenges for female couples (Lavner 2017). Female couples may experience more stress in response to concealment of the relationship by a partner or hostility (or rejection) from family members (Todosijevic et al. 2005). Male couples may be more likely to compartmentalize such experiences (Rostosky et al. 2007). Furthermore, qualitative studies continue to suggest that same-sex couples, particularly lesbian couples, are distinct in terms of the value they place on autonomy, equality, trust, and disclosure (e.g., Dunne 1997; Weeks et al. 2001). According to Giddens (1992), these are defining features of pure (or confluent love) relationships. Pure relationships are not structured by gender role norms or by a lifetime commitment but are defined collaboratively by partners and subject to ongoing renegotiation. These relationships are inherently unstable because they last only as long as both partners remain satisfied with them, or in the words of Giddens (1992:187), “good until further notice.” We believe more nuanced attention to the nature of same-sex relationships is warranted.

This study contributes to a groundbreaking body of research that has brought same-sex couples to the forefront of social science research. Importantly, it provides innovative analyses of couple stability for a cohort overlooked by prior studies comparing same-sex and different-sex stability. Studies have rarely compared these relationship types for a specific age or cohort (with the exception of Lau 2012), presumably reflecting the challenges posed by small sample sizes in studies of atypical family arrangements (Cheng and Powell 2005). Most of Add Health’s respondents were born between 1976 and 1982, and consequently, were between the ages of 24 and 32 at the time of the fourth wave. Add Health’s large sample and recent fielding resulted in ample numbers of same-sex male and female relationships for a recent birth cohort of young adults. Prior studies considering same-sex couples have presented results based on much smaller samples, even after pooling observations from different surveys (i.e., Black et al. 2000).

Our sample and measurement is, however, limited in some respects. Because it may be challenging for respondents to provide the initials of most recent partners who were casual, our sample of most recent relationships is likely biased toward more meaningful relationships. Our focus on most recent relationships means that we are excluding prior and concurrent relationships, most of which have ended. We cannot address the bias posed by this restriction, given that we know the total duration of most recent relationships but not prior relationships. The fact that male respondents with same-sex partners (and not female respondents with same-sex partners) accrue more lifetime sex partners than both male and female respondents with different-sex partners is consistent with the results pertaining to the stability of most recent relationships.

Another limitation is our measurement of stability. The estimation of rates of dissolution requires information on the length of the relationship and whether it was current at the fourth wave. It may have been challenging for respondents to determine the length of their entire relationship, especially if it had dissolved. As in prior studies, we measure dissolution as a dichotomous event, defining relationships as dissolved if respondents did not classify them as “current” at the time of the fourth wave. Like most prior studies on dissolution, we cannot capture the fluidity of ending a relationship with a dichotomous construct (Agnew et al. 2006; Le et al. 2010). It is reassuring to note that we continue to find elevated rates of dissolution for same-sex male couples when we exclude the types of relationships for which our measure of dissolution could be most problematic: hookups and casual dating relationships. We also obtain similar patterns when we use an indirect indicator of whether the relationship has dissolved (i.e., Fig. 3).

Our focus on a narrow age band of young adults means that our findings cannot be generalized to other birth cohorts or periods of the life course. Given the growing acceptability of LGBT relationships, we would expect fewer differences in the outcomes of same-sex and different-sex relationships the more recent the cohort of young adults. We might also expect differences to be less pronounced at older ages as LGBT individuals gain more practice with these relationships. As shown in Table 2, and consistent with national statistics, very few same-sex relationships became marriages. Thus, we cannot examine whether marriage has the same protective effect for same-sex couples as it does for different-sex couples. Nor are we able to identify sufficient numbers of couples with children to assess stability for this important subgroup (Moore and Stambolis-Ruhstorfer 2013; Tornello and Patterson 2015).

Finally, this study does not identify the mechanisms producing relatively greater instability for men in same-sex relationships that were not coresidential. Add Health includes items capturing the experience of minority stress at Wave 4, but they were not asked at prior waves, so they cannot be measured prospectively. Another mechanism that is traditionally related to relationship stability and that we cannot examine prospectively is sexual exclusivity. Prior studies based on recruitment samples consistently find that same-sex male couples are more likely to be sexually nonexclusive than other couple types (Adam 2006; Blumstein and Schwartz 1983; Kurdek 1998; Rothblum 2009; Solomon et al. 2005). Yet, some couples, particularly same-sex male couples, are shown to have agreements about the acceptability of sexual non-exclusivity (Adam 2006; Blumstein and Schwartz 1983; Parsons et al. 2012; Weeks et al. 2001). Future research should use prospective data to identify the mechanisms that underlie relationship stability for same-sex male and female couples (Eeden-Moorefield et al. 2011).

Prior studies on the stability of same-sex relationships are dominated by samples that are coresidential, European, or decades old. We demonstrate that it is critical for researchers to address the differential selection of lesbians and gays into more established relationships when comparing the stability of same-sex and different-sex relationships. Our work offers new insights by focusing on young adults who came of age in the United States, but it does not address how patterns differ spatially. A new feature of the American legal climate is the recent legalization of same-sex marriage in every state on June 26, 2015. It is important for data collections to keep pace with this change and include questions on same-sex marriage. The Add Health sample is moving beyond the typical age at first marriage for different-sex couples, and new assessments of the lives and well-being of new cohorts of young adults will be necessary. Despite calls for new population-based surveys that include measurement of sexual orientation and same-sex couple dynamics (Gates 2011; Manning 2015), as of this date there appear to be no large-scale quantitative data collections of young adults under way that will inform new research on the implications of marriage for same-sex couples.

Acknowledgments

This research was supported in part by the Center for Family and Demographic Research, Bowling Green State University, which has core funding from the Eunice Kennedy Shriver National Institute of Child Health and Human Development (P2CHD050959). We appreciate the insightful comments that Jenifer Bratter provided on an early draft of this article presented at the 2012 annual meeting of the Population Association of America. We also thank Barbara Prince for editorial suggestions. This research uses data from Add Health, a program project directed by Kathleen Mullan Harris and designed by J. Richard Udry, Peter S. Bearman, and Kathleen Mullan Harris at the University of North Carolina at Chapel Hill, and funded by Grant P01-HD31921 from the Eunice Kennedy Shriver National Institute of Child Health and Human Development, with cooperative funding from 23 federal agencies and foundations. Special acknowledgment is due to Ronald R. Rindfuss and Barbara Entwisle for assistance in the original design. Information on how to obtain the Add Health data files is available on the Add Health website (http://www.cpc.unc.edu/addhealth). No direct support was received from Grant P01-HD31921 for this analysis.

Notes

1

The multivariate left-truncated hazard models in these recent U.S. studies do not specify interactions between the covariates and time. In the case of HCMST, the risk period extends several decades (e.g., Rosenfeld 2014). We suspect that differences between same-sex and different-sex couples would be larger in models that begin exposure to the risk of dissolution at the time of relationship formation and include a shorter risk period (e.g., the first seven years) because the same-sex couples at risk of dissolution would be relatively less select.

2

It was rare for respondents, including males with same-sex partners, to have more than one current partner at the fourth wave. In such cases (roughly 5 % of the entire sample), Add Health administered a set of rules for choosing the focal partner.

3

Respondents who indicated their relationship was not current and reported that their marriage ended with the death of a spouse were treated as censored.

4

We extended the models to cover the entire period of risk (not shown), not just the first seven years. The results were generally similar regardless of whether we censored relationship dissolution at seven years.

5

The models control for age at the time of interview rather than age when the relationship began. In this sample of most recent relationships, age when the relationship began is a proxy indicator of duration.

6

We also investigated whether male and female couples had an adopted child by examining the household roster. This was an extremely rare event for these couples.

7

Men were more likely than women to have consistent reports of sexual orientation: 93 % of men in same-sex relationships and 92 % of men in different-sex relationships were consistent, compared with 78 % of women in same-sex relationships and 78 % of women in different-sex relationships. We suspect the elevated rates of inconsistency for women are partly a reflection of their greater sexual fluidity (Savin-Williams et al. 2012).

8

We constrain differences between same-sex and different-sex couples in hazard rates to be proportional. We failed to find any evidence that their rates were not proportional.

References

References
Adam, B. D. (
2006
).
Relationship innovation in male couples
.
Sexualities
,
9
,
5
26
. 10.1177/1363460706060685
Agnew, C. R., Arriaga, X. B., & Goodfriend, W. (
2006, July
).
On the (mis)measurement of premarital relationship breakup
. Paper presented at the Conference of the International Association for Relationship Research,
Crete, Greece
.
Amato, P. R., & Previti, D. (
2003
).
People’s reasons for divorcing gender, social class, the life course, and adjustment
.
Journal of Family Issues
,
24
,
602
626
. 10.1177/0192513X03024005002
Andersson, G., Noack, T., Seierstand, A., & Weedon-Fekjær, H. (
2006
).
The demographics of same-sex marriage in Norway and Sweden
.
Demography
,
43
,
79
98
. 10.1353/dem.2006.0001
Badgett, M. V. L., & Herman, J. L. (
2013
).
Patterns of relationship recognition by same-sex couples in the United States
. In Baumle, A. K. (Ed.),
International handbook on the demography of sexuality
(pp.
331
362
).
New York, NY
:
Springer
.
Balsam, K. F., Beauchaine, T. P., Rothblum, E. D., & Solomon, S. E. (
2008
).
Three-year follow-up of same-sex couples who had civil unions in Vermont, same-sex couples not in civil unions, and heterosexual married couples
.
Developmental Psychology
,
44
,
102
116
. 10.1037/0012-1649.44.1.102
Bearman, P. S., & Moody, J. (
2004
).
Suicide and friendship among American adolescents
.
American Journal of Public Health
,
94
,
89
95
. 10.2105/AJPH.94.1.89
Black, D., Gates, G. J., Sanders, S., & Taylor, L. (
2000
).
Demographics of the gay and lesbian population in the United States: Evidence from available systematic data sources
.
Demography
,
37
,
139
154
. 10.2307/2648117
Blumstein, P., & Schwartz, P. (
1983
).
American couples: Money, work, and sex
.
New York, NY
:
William Morrow & Co.
.
Carpenter, C., & Gates, G. J. (
2008
).
Gay and lesbian partnerships: Evidence from California
.
Demography
,
45
,
573
590
. 10.1353/dem.0.0014
Cheng, S., & Powell, B. (
2005
).
Small samples, big challenges: Studying atypical family forms
.
Journal of Marriage and Family
,
67
,
926
935
. 10.1111/j.1741-3737.2005.00184.x
Cherlin, A. J. (
2004
).
The deinstitutionalization of American marriage
.
Journal of Marriage and Family
,
66
,
848
861
. 10.1111/j.0022-2445.2004.00058.x
Doyle, D. M., & Molix, L. (
2014
).
How does stigma spoil relationships? Evidence that perceived discrimination harms romantic relationship quality through impaired self-image
.
Journal of Applied Social Psychology
,
44
,
600
610
. 10.1111/jasp.12252
Doyle, D. M., & Molix, L. (
2015
).
Social stigma and sexual minorities’ romantic relationship functioning: A meta-analytic review
.
Personality and Social Psychology Bulletin
,
41
,
1363
1381
. 10.1177/0146167215594592
Dunne, G. A. (
1997
).
Lesbian lifestyles: Women’s work and the politics of sexuality
.
Basingstoke, UK
:
Macmillan Press
.
Eeden-Moorefield, V., Martell, C. R., Williams, M., & Preston, M. (
2011
).
Same-sex relationships and dissolution: The connection between heteronormativity and homonormativity
.
Family Relations
,
60
,
562
571
. 10.1111/j.1741-3729.2011.00669.x
Frost, D. M., & Gola, K. A. (
2015
).
Meanings of intimacy: A comparison of members of heterosexual and same-sex couples
.
Analyses of Social Issues and Public Policy
,
15
,
382
400
. 10.1111/asap.12072
Frost, D. M., & Meyer, I. H. (
2009
).
Internalized homophobia and relationship quality among lesbians, gay men, and bisexuals
.
Journal of Counseling Psychology
,
56
,
97
109
. 10.1037/a0012844
Gates, G. J. (
2006
).
Characteristics and predictors of coresidential stability among couples
(CCPR Working Paper No. CCPR-069-06).
Los Angeles
:
California Center for Population Research, University of California Los Angeles
.
Gates, G. J. (
2009
).
Same-sex spouses and unmarried partners in the American Community Survey, 2008
.
Los Angeles, CA
:
Williams Institute
. Retrieved from http://williamsinstitute.law.ucla.edu/wp-content/uploads/Gates-ACS2008FullReportSept-2009.pdf
Gates, G. J. (
2011, July
).
Recommendations for improving measurement of intimate partner relationships
. Paper presented at the Counting Couples, Counting Families Conference, National Center for Family and Marriage Research,
Bethesda, MD
.
Gates, G. J. (
2017
,
January
11
).
In US, more adults identifying as LGBT
.
Social Issues
.
Washington, DC
:
Gallup
. Retrieved from http://www.gallup.com/poll/201731/lgbt-identification-rises.aspx
Giddens, A. (
1992
).
The transformation of intimacy: Sexuality, love and eroticism in modern societies
.
Stanford, CA
:
Stanford University Press
.
Guo, G. (
1993
).
Event-history analysis for left-truncated data
. In Marsden, P. V. (Ed.),
Sociological methodology
(pp.
217
243
).
Washington, DC
:
American Sociological Association
.
Harris, K. M., Halpern, C. T., Whitsel, E., Hussey, J., Tabor, J., Entzel, P., & Udry, J. R. (
2009
).
The National Longitudinal Study of Adolescent Health: Research design
.
Chapel Hill
:
Carolina Population Center, University of North Carolina at Chapel Hill
.
Herek, G. M. (
2006
).
Legal recognition of same-sex relationships in the United States: A social science perspective
.
American Psychologist
,
61
,
607
621
. 10.1037/0003-066X.61.6.607
Institute of Medicine (IOM)
. (
2011
).
The health of lesbian, gay, bisexual, and transgender people: Building a foundation for better understanding
.
Washington, DC
:
National Academies Press
.
Jepsen, L. K., & Jepsen, C. A. (
2002
).
An empirical analysis of the matching patterns of same-sex and opposite-sex couples
.
Demography
,
39
,
435
453
. 10.1353/dem.2002.0027
Kalmijn, M., Loeve, A., & Manting, D. (
2007
).
Income dynamic in couples and the dissolution of marriage and cohabitation
.
Demography
,
44
,
159
179
. 10.1353/dem.2007.0005
Kurdek, L. A. (
1998
).
Relationship outcomes and their predictors: Longitudinal evidence from heterosexual married, gay cohabiting, and lesbian cohabiting couples
.
Journal of Marriage and the Family
,
60
,
553
568
. 10.2307/353528
Kurdek, L. A. (
2004
).
Are gay and lesbian cohabiting couples really different from heterosexual married couples?
.
Journal of Marriage and Family
,
66
,
880
900
. 10.1111/j.0022-2445.2004.00060.x
Lau, C. Q. (
2012
).
The stability of same-sex cohabitation, different-sex cohabitation, and marriage
.
Journal of Marriage and Family
,
74
,
973
988
. 10.1111/j.1741-3737.2012.01000.x
Lavner, J. A. (
2017
).
Relationship satisfaction in lesbian couples: Review, methodological critique, and research agenda
.
Journal of Lesbian Studies
,
21
,
7
29
.
Le, B., Dove, N. L., Agnew, C. R., Korn, M. S., & Mutso, A. A. (
2010
).
Predicting nonmarital romantic relationship dissolution: A meta-analytic synthesis
.
Personal Relationships
,
17
,
377
390
. 10.1111/j.1475-6811.2010.01285.x
LeBlanc, A. J., Frost, D. M., & Wight, R. G. (
2015
).
Minority stress and stress proliferation among same-sex and other marginalized couples
.
Journal of Marriage and Family
,
77
,
40
59
. 10.1111/jomf.12160
Lofquist, D., Lugailia, T., O’Connell, M., & Feliz, S. (
2012
).
Households and families: 2010
(2010 Census Briefs No. C2010BR-14).
Washington, DC
:
U.S. Census Bureau
.
Manning, W. D. (
2015
).
Family formation processes: Assessing the need for a new nationally representative household panel survey in the United States
.
Journal of Economic and Social Measurement
,
40
,
171
193
. 10.3233/JEM-150419
Manning, W. D., Brown, S. L., & Payne, K. (
2014
).
Two decades of stability and change in age at first union formation
.
Journal of Marriage and Family
,
25
,
247
260
. 10.1111/jomf.12090
Manning, W. D., Brown, S. L., & Stykes, J. B. (
2016
).
Same-sex and different-sex cohabiting couple relationship stability
.
Demography
,
53
,
937
953
. 10.1007/s13524-016-0490-x
Meyer, I. H. (
2003
).
Prejudice, social stress, and mental health in lesbian, gay, and bisexual populations: Conceptual issues and research evidence
.
Psychological Bulletin
,
129
,
674
697
. 10.1037/0033-2909.129.5.674
Moore, M. R., & Stambolis-Ruhstorfer, M. (
2013
).
LGBT sexuality and families at the start of the 21st century
.
Annual Review of Sociology
,
39
,
491
507
. 10.1146/annurev-soc-071312-145643
Noack, T., Seierstad, A., & Weedon-Fekjær, H. (
2005
).
A demographic analysis of registered partnerships (legal same-sex unions): The case of Norway
.
European Journal of Population
,
21
,
89
109
. 10.1007/s10680-005-3626-z
Office for National Statistics (ONS)
. (
2013
).
Civil partnerships in the UK, 2012
(Statistical Bulletin).
London, UK
:
ONS
.
Paik, A. (
2015
).
Surveying sexualities: Minimizing survey error in study of sexuality
. In DeLamater, J., & Plant, R. F. (Eds.),
Handbook of the sociology of sexualities
(pp.
93
107
).
Cham, Switzerland
:
Springer International Publishing
.
Parsons, J. T., Starks, T. J., Gamarel, K. E., & Grov, C. (
2012
).
Non-monogamy and sexual relationship quality among same-sex male couples
.
Journal of Family Psychology
,
26
,
669
677
. 10.1037/a0029561
Powell, B. (
2017
).
Changing counts, counting change: Toward a more inclusive definition of the family
.
Journal of the Indiana Academy of the Social Sciences
,
17
(
1
),
1
14
.
Powell, B., Bolzendahl, C., Geist, C., & Steelman, L. C. (
2010
).
Counted out: Same-sex relations and Americans’ definitions of family
.
New York, NY
:
Russell Sage Foundation
.
Rosenfeld, M. J. (
2014
).
Couple longevity in the era of same-sex marriage in the United States
.
Journal of Marriage and Family
,
76
,
905
918
. 10.1111/jomf.12141
Rosenfeld, M. J., & Kim, B. (
2005
).
The independence of young adults and the rise of interracial and same-sex unions
.
American Sociological Review
,
70
,
541
562
. 10.1177/000312240507000401
Ross, H., Gask, K., & Berrington, A. (
2011
).
Civil partnerships five years on
.
Population Trends
,
145
,
172
202
. 10.1057/pt.2011.23
Rostosky, S. S., Riggle, E. D. B., Gray, B. E., & Hatton, R. L. (
2007
).
Minority stress experiences in committed same-sex couple relationships
.
Professional Psychology: Research and Practice
,
38
,
392
400
. 10.1037/0735-7028.38.4.392
Rothblum, E. D. (
2009
).
An overview of same-sex couples in relationships: A research area still at sea
. In Hope, D. A. (Ed.),
Contemporary perspectives on lesbian, gay, and bisexual identities
(pp.
113
139
).
New York, NY
:
Springer
.
Sassler, S. (
2010
).
Partnering across the life course: Sex, relationships, and mate selection
.
Journal of Marriage and Family
,
72
,
557
575
. 10.1111/j.1741-3737.2010.00718.x
Savin-Williams, R. C., Joyner, K., & Rieger, G. (
2012
).
Prevalence and stability of self-reported sexual orientation identity during young adulthood
.
Archives of Sexual Behavior
,
41
,
103
110
. 10.1007/s10508-012-9913-y
Schwartz, C. R., & Graf, N. L. (
2009
).
Assortative mating among same-sex and different-sex couples in the United States, 1990–2000
.
Demographic Research
,
21
(
article 28
),
843
878
. 10.4054/DemRes.2009.21.28
Solomon, S. E., Rothblum, E. D., & Balsam, K. F. (
2005
).
Money, housework, sex, and conflict: Same-sex couples in civil unions, those not in civil unions, and heterosexual married siblings
.
Sex Roles
,
52
,
561
575
. 10.1007/s11199-005-3725-7
Todosijevic, J., Rothblum, E. D., & Solomon, S. E. (
2005
).
Relationship satisfaction, affectivity, and gay-specific stressors in same-sex couples joined in civil unions
.
Psychology of Women Quarterly
,
29
,
158
166
. 10.1111/j.1471-6402.2005.00178.x
Tornello, S. L., & Patterson, C. J. (
2015
).
Timing of parenthood and experiences of gay fathers: A life course perspective
.
Journal of GLBT Family Studies
,
11
,
35
56
. 10.1080/1550428X.2013.878681
Tumin, D., Han, S., & Qian, Z. (
2015
).
Estimates and meanings of marital separation
.
Journal of Marriage and Family
,
77
,
312
322
. 10.1111/jomf.12149
Umberson, D., & Kroeger, R. A. (
2016
).
Gender, marriage, and health for same-sex and different-sex couples: The future keeps arriving
. In S. M. McHale, V. King, J. Van Hook, & A. Booth, (Eds.),
Gender and couple relationships
(pp.
189
213
).
Cham, Switzerland
:
Springer International Publishing
.
Umberson, D., Thomeer, M. B., Kroeger, R. A., Lodge, A. C., & Xu, M. (
2015a
).
Studying same-sex relationships: Challenges and opportunities?
Journal of Marriage and Family
,
77
,
96
111
.
Umberson, D., Thomeer, M. B., & Lodge, A. C. (
2015
).
Intimacy and emotion work in lesbian, gay, and heterosexual relationships
.
Journal of Marriage and Family
,
77
,
542
556
. 10.1111/jomf.12178
Weeks, J. B., Heaphy, B., & Donovan, C. (
2001
).
Same sex intimacies: Families of choice and other life experiments
.
New York, NY
:
Routledge
.
Weisshaar, K. (
2014
).
Earnings equality and relationship stability for same-sex and heterosexual couples
.
Social Forces
,
93
,
93
123
. 10.1093/sf/sou065
Wiik, K. A., Seierstad A., & Noack, T. (
2012
).
Divorce in Norwegian same-sex marriages 1993–2011
. (Discussion Papers No. 723).
Statistics Norway, Research Department
. Retrieved from https://www.ssb.no/a/publikasjoner/pdf/DP/dp723.pdf