CAIRN-INT.INFO : International Edition

Same-sex unions in Europe have rarely been studied because they are difficult to identify in the statistical data. With the opening of civil partnerships and marriage to same-sex couples, more is now known about union formation in this population group, although union stability is still poorly documented. Does it differ from that of different-sex couples? Using data from the international Generations and Gender surveys, the author compares the separation risks of cohabiting gay and lesbian couples in six European countries with those of different-sex couples, controlling for the specific features of each.

1The first scientific studies of union stability and duration were conducted in the 1950s. At that time, the system of values in Western societies was changing, with a progressive secularization of unions, an increase in divorce, and a decline in fertility. Against this backdrop of societal shifts, researchers became interested in the factors of marital longevity. The growth in second marriages and repartnering, which followed the increase in divorce, as well as new couple lifestyles (non-marital cohabitation, living apart together, stepfamilies), then prompted the development of theoretical frameworks to explain these changes, such as that of the ‘second demographic transition’ (Van De Kaa, 1987; Lesthaeghe, 2010).

2Focusing on couples formed by a man and a woman, most of these studies ignored same-sex couples. Some did not even mention their existence (Becker, 1981; Oppenheimer, 1988), while others specified that their observations applied to heterosexual couples only (Brines and Joyner, 1999). The first analyses of same-sex unions, based on small-scale survey data or qualitative studies, date back to the 1980s and primarily concern the United States (Blumstein and Schwartz, 1983; Kurdek, 1998).

3One factor behind this omission is that civil union and marriage were opened to gay and lesbian couples only very recently—starting in the 1990s in pioneer countries such as Denmark and Norway, and some years later in other Western countries. Moreover, these couples represent a small minority of the population and cannot always be clearly identified or counted using available data sources. With no legal existence and no visibility in statistical data, they have been difficult to study.

4Today, same-sex couples may marry, have children, and achieve social recognition. Yet the factors underlying the stability or instability of their unions are largely unexplored. The sociodemographic characteristics of different-sex couples differ from those of same-sex couples, which may affect their partnership dynamics. While certain characteristics favour the longevity of different-sex couples, they may act differently upon same-sex couples. This study therefore raises two key research questions. First, is the separation risk for same-sex couples higher than that of different-sex couples? Second, what are the factors associated with this risk?

5Drawing on data from the European Generations and Gender Survey (GGS) for several countries (Belgium, France, Germany, the Netherlands, Norway, and Sweden), respondents’ partnership histories are reconstituted to compare the longevity of same-sex unions (n = 385) and different-sex unions (n = 36,090) formed since the 1980s.

6After describing the main theories and the factors influencing union stability, we give a synopsis of the few quantitative studies on same-sex unions. We then examine the GGS’s definition of a union as well as the partnership histories of individuals in cohabiting unions. A Cox model is used to analyse separation risks, taking account of individual and union characteristics.

I – Two types of unions with specific characteristics

1 – A literature related to the shifting legislative context

7Until the late 1990s, most sociodemographic research on union stability concerned heterosexual unions only. [1] A few researchers then started to include homosexual unions (Andersson et al., 2006; Lau, 2012; Wiik et al., 2014; Manning et al., 2016) in response to changes in the social and legislative context affecting both types of unions. These studies often compare the longevity of same-sex unions to that of different-sex unions. They find a higher separation risk for same-sex couples, but given the structural effects specific to these unions, it is important to qualify this result.

8Among cohabiting different-sex couples, married partners’ expectations are different from those of unmarried partners. People in non-marital unions tend to be more individualistic and to seek a more egalitarian relationship, particularly in the sharing of domestic tasks and income, thus challenging the ‘traditional’ gender role models (Brines and Joyner, 1999; Weisshaar, 2014). For married couples, the division of tasks is more specialized, so the partners are more interdependent. Divorce costs may hinder separation, especially for women, whose living standard drops more sharply than that of men after a divorce (Andreß et al., 2006). However, no direct evidence of a causal effect of union formalization on separation risk has been found (Blumstein and Schwartz, 1983; Brines and Joyner, 1999). Although the proportion of unmarried couples is increasing in many Western countries, their separation risk remains higher than that of couples who are married or in a civil partnership (Brines and Joyner, 1999; Blackwell and Lichter, 2004; Lyngstad and Jalovaara, 2010).

9Because same-sex couples have not always been able to marry, the legislative and normative context surrounding these unions is different from that of different-sex unions. In some countries, homosexual unions have been recognized only in the last 2 decades and in a variety of forms (civil partnership, marriage, sometimes both). Most cohabiting same-sex couples have not been able to register their union from the outset (Banens, 2010; Rosenfeld, 2014). If this non-formalization affects same-sex and different-sex couples similarly, it might largely explain same-sex couples’ higher separation risk.

10Furthermore, for same-sex couples, being able to formalize their union does not mean they will necessarily do so. Registration rates vary across countries. In Norway and Sweden they are low (around 5 per 100,000 inhabitants), while in the United Kingdom they are high (30 per 100,000 inhabitants) (Banens, 2010). Alongside formalization, levels of commitment between partners can also be expressed in different ways. That homosexual couples live together less frequently than heterosexual couples reflects differences in partnership norms between the two types of unions (Rault, 2018; Régnier-Loilier, 2018b). In his study of union stability in the United Kingdom, Lau (2012) showed that for unmarried cohabiting same-sex couples, the risk of separation is slightly higher than for equivalent different-sex couples; married heterosexual unions are the stablest. But the national context may also play a role, since certain studies in the United States yield different results, with some finding an equal separation risk for both same-sex and different-sex unmarried couples (Manning et al., 2016).

2 – Presence of children

11Most studies of union longevity also identify the presence of children as a protection against separation (Levinger, 1965). A sense of moral obligation towards the children may reduce the risk of breakup. However, the increase in separations of parental couples and the growing number of stepfamilies suggest this moral commitment has been weakening. Childless couples also have specific characteristics that may be associated with higher separation risks. They generally have less conservative family values, are often unmarried, or have no strong religious attachment (Lyngstad and Jalovaara, 2010).

12Few same-sex couples have children, gay couples especially, as they face both legal and biological obstacles to parenthood, and their access to adoption is more limited than for different-sex couples. Lesbian couples have more children on average, whether from a previous relationship or via artificial insemination, and so are more often faced with this moral duty to stay together for the children’s sake (Gross et al., 2014). Gay couples with children appear to be less stable than lesbian couples with children (Wiik et al., 2014). Children from a previous relationship seem to be associated with a higher separation risk, while children born in the current union tend to increase union stability (Lyngstad and Jalovaara, 2010; Wiik et al., 2014).

3 – Heterogamy among same-sex couples

13Homosexual people have fewer opportunities to meet than heterosexual people, who often find a partner in their place of work or education (Blackwell and Lichter, 2004; Bozon and Rault, 2012). As some studies have shown, differences in age or socioeconomic profile between same-sex partners are greater than those between different-sex partners, which may affect separation risk, since heterogamy is associated with greater union instability (Andersson et al., 2006; Kalmijn et al., 2007; Verbakel and Kalmijn, 2014).

II – Research hypotheses

14Research on union duration thus suggests separation risks are generally higher for same-sex unions than for different-sex unions (H1). The GGS allows us to broaden the study to other countries of Europe and to test this hypothesis.

15Marriage (or civil partnership) is more common among different-sex couples than among same-sex couples and has a much longer history, so it may not be experienced in the same way by both. Moving in with a partner, without necessarily formalizing the union, may be an important expression of partnership commitment for same-sex couples, given that they more frequently live apart together than do different-sex couples. Union formalization may thus be less symbolic for same-sex couples and may incur different separation risks. Our second hypothesis is that the stabilizing effect of marriage (or civil partnership) is weaker for same-sex couples than for different-sex couples (H2).

16Parenthood is another factor liable to influence separation risk, but its effect on same-sex couples has not yet been clearly demonstrated. It is difficult for these couples to become parents, and their children (if any) were often born in a previous union, so they represent less of an obstacle to separation. Childless same- and different-sex couples may therefore be similar. Our last hypothesis is that childless couples have similar separation risks (H3).

III – Data and background

1 – Data

17The GGS is a series of longitudinal surveys (three survey waves at 3-year intervals) conducted in a number of developed countries to gain new insights into the causes and consequences of demographic changes. These surveys cover the population living in ‘ordinary’ (non-collective) households and aged 18–79 at the time of the first survey wave. [2] They provide information on individuals’ partnership histories, with questions on respondents’ current and previous cohabiting unions. We selected six countries for our study: Belgium, France, Germany, the Netherlands, Norway, and Sweden, which were chosen because the ‘partner’s gender’ variable was included in the GGS questionnaires, making it possible to study same-sex unions. Furthermore, the relative tolerance of homosexuality in these countries limits the proportion of individuals potentially unwilling to report that they live or have lived with a same-sex partner, even though some may still be fearful of social stigma. The geographical proximity of these countries does not signify cultural homogeneity, however, and the country effect is controlled for in our models. That said, several national surveys had to be combined to obtain a large enough sample of cohabiting same-sex couples; they form a very small minority of all unions, and few are included in national surveys representative of the general population. Only the first wave is analysed in this study.

2 – Background

18Same-sex couples have progressively gained legal recognition in several European countries, and those who so wish can now register their partnership (Table 1). The symbolic significance of this official status and the proportion of couples concerned vary across countries.

19In the GGS, only civil partnerships registered in France (pacte civil de solidarité [PACS]) can be identified in the data as a specific union status distinct from marriage and consensual union. As in the Netherlands, registered partnerships are open to both same- and different-sex couples. In Belgium, two coresident individuals can obtain legal cohabitation status whatever their gender or kin relationship. This status provides a way for homosexual couples to formalize their union. In these three countries, all couples can still opt for a civil partnership even now that same-sex marriage is legal.

Table 1

Legal status of same-sex couples in the survey countries

Table 1
Country Designation Period of implementation GGS year (Wave 1) Information available in GGS Legal cohabitation 2000– No Belgium 2008–2010 Marriage 2003– Yes PACS 1999– Yes France 2005 Marriage 2013– No Registered partnership 1998– No Netherlands 2002–2004 Marriage 2001– Yes Registered partnership 2001–2017 No Germany 2005 Marriage 2017– No Registered partnership 1993–2009 Yes Norway 2007–2008 Marriage 2009– No Registered partnership 1995–2009 Yes Sweden 2012–2013 Marriage 2009– Yes

Legal status of same-sex couples in the survey countries

20The signification of ‘registered partnership’ is different in Sweden and Norway. Before 2009, it was open only to same-sex couples. It granted rights similar to those of marriage for different-sex couples and was presented as equivalent (Andersson et al., 2006), so in the data used here, it is assimilated to marriage. From 2009, registered partnership in these two countries was replaced by marriage for new same-sex couples, thereby harmonizing the designation and legal aspects of their status with that of different-sex couples (Banens, 2010; Wiik et al., 2014).

21In Germany, registered partnership is open only to same-sex couples and grants fewer rights than the Nordic model (Festy, 2006).

22At the time of the first GGS wave, we can consider that Belgium, Norway, the Netherlands, and Sweden had opened marriage (or its close equivalent) to same-sex couples. In France, the PACS was the only available option for formalizing a union. In Germany, registered partnership was legal at the time of the survey, but no information is given in the GGS.

IV – Method

1 – Definitions of union and partnership history

23In the GGS, just one member of each sample household is interviewed, but information on other members is also collected. The regular presence of a partner in the dwelling indicates a cohabiting union. Besides their current union at the time of the survey, respondents are asked to describe all previous cohabiting unions. The survey data contain nothing on previous non-cohabiting relationships nor the date when the couple first met before moving in together.

24Consensual unions—those not recognized by a civil partnership or marriage—are included in the partnership history. All types of unions, formalized or otherwise, are thus recorded in this part of the questionnaire. This is of particular importance for same-sex couples who have only recently been able to register their partnership.

2 – Variables

25The dependent variable in our statistical analyses is union dissolution. In the longitudinal models measuring separation risk by union duration, each couple is followed from the start of cohabitation, if it occurred after 1980, [3] up to the date of separation (if any). For married couples, the date they report having stopped living together is used as the separation indicator, not the legally recognized divorce date. To lower the risk of under-reporting unions in the distant past, the study’s scope is limited to individuals aged 60 and under.

26The main independent variable is the gender composition of the union (or type of couple), i.e. two people of the same sex or of different sexes. Unfortunately, coding errors on the sex of the respondents or their partners may lead to an overestimation of the number of same-sex unions recorded in the database (Festy, 2007). The proposed checks (and corrections) are described in the Box. However, they do not guarantee that all errors are eliminated (Régnier-Loilier, 2018a).

27Independent and control variables are also included in the models:

  • Union formalization. This variable provides a means to verify the second hypothesis on the difference in separation risk for the two types of couples. Because some live together before marrying, this variable is considered as time-varying in all models. Registered partnerships, when identifiable in the survey (in France) or when equivalent to marriage for same-sex couples (in Sweden and Norway), are counted as a form of legally recognized union.
  • Presence and number of children (in the household or not). This variable serves to verify the third hypothesis on increased separation risk for childless couples of both types. In the model, we distinguish between two measures of parenthood: a birth within the union or the presence of a child at the start of the union. [4] The first measure is time dependent, and the second is fixed. The couple’s shared children, living in the household or not at the time of the survey, include biological children (and adopted children in Sweden). For previous unions reported in the partnership history, it is assumed children born after the start of cohabitation were born within that union. Children from previous unions include those of the respondent or his or her partner’s.
  • Availability of a legal framework for same-sex couples. Depending on the period and country, union formalization may not always have been possible for same-sex couples, so when the mode of formalization is indicated in the survey, it only introduces heterogeneity for different-sex couples. This time-dependent variable thus gives an indication of the period and the possibility of marriage (or civil partnership) for same-sex partners.
  • Age difference between partners [5]
  • Respondent’s age at entry into union
  • Intergenerational transmission of separation measured by the separation (if any) of the respondent’s parents before age 18
  • The respondent’s union order, distinguishing first cohabiting unions and higher-order unions
  • Respondent’s age at the time of the survey
  • Respondent’s educational level
  • Country of residence

Box. Identifying same-sex unions in the GGS

Same-sex couples cannot be identified in the GGS simply by matching the variables corresponding to the gender of the respondent and his or her partner. Sex-coding errors exist, and even if the error rate is low, the estimated proportion of same-sex unions in the population may be severely distorted by these data capture biases (Festy, 2007).
Checks can be made to identify and limit bias linked to ‘false’ same-sex unions. First, for married couples, the marriage date must be later than the date when same-sex marriage became legal in the country concerned. For supposedly same-sex married couples, if the marriage took place before this legalization date, then the union is recoded as heterosexual. Second, having a biological child with one’s current partner in the household is a highly unlikely situation for same-sex couples. The last check concerns the answer to a question of opinion, i.e. agreeing with the statement that ‘same-sex couples should have the same rights as different-sex couples.’ Individuals who (totally) disagree with this statement are recoded as being in different-sex unions. These checks eliminated 125 false same-sex unions out of 530 among unions formed after 1980 by respondents aged 60 and under at the time of the survey (Table 2).
Table 2

Number of same-sex unions and percentage of sex-coding errors by country

Table 2
Country n same-sex unions % % Biological % n same-sex unions Marriage children Opinion after correction % Total error Belgium 81 16.1 2.5 2.5 64 20.1 Germany 73 46.6 4.1 1.4 35 52.1 Netherlands 123 11.4 4.9 — 103 16.3 Norway 87 0.0 3.5 6.9 78 10.4 Sweden 60 1.7 5.0 0.0 56 6.7 France 106 23.6 9.4 8.5 62 41.5 France (first name) 106 — — — 49 53.8 Total 530 16.4 5.1 3.4 385 24.9

Number of same-sex unions and percentage of sex-coding errors by country

Interpretation: Of same-sex couples in Belgium, 16.1% are identified as false after checking the marriage date.
Note: The types of errors are considered sequentially. A single couple may be concerned by both a marriage-date error and a biological-children error.
Coverage: Cohabiting unions formed since 1980 by individuals aged 60 and under.
Sources: GGS Wave 1, GGS Wave 1 France (database of first names of respondents and their partners).
For Norway and Sweden, the population registers used to collect basic sociodemographic information on the sampled individuals contain fewer errors on the gender of individuals or their partners. Error rates are higher for France and Germany. Data collection in these two countries and in Belgium is by face-to-face interview only, while in the Netherlands self-administered questionnaires are also used (Fokkema et al., 2016). This may explain the higher rates of coding errors in these countries than in those with administrative databases that provide more reliable information on gender.
In France, access to a database giving the first names of respondents and their partners provided a more precise alternative method for detecting false same-sex couples. This is the preferred correction method used in our analyses. The higher percentage of errors detected in France than elsewhere using the first-name database suggests that the previous corrections do not cover all false same-sex couples, notably in Belgium or the Netherlands, where the rate appears lower than in France and Germany, which used very similar data collection protocols. The Dutch questionnaire, for example, does not include the opinion question, so false same-sex couples cannot be detected based on this item.

3 – Statistical and event history models

Checking the consistency of partnership history

28Most surveys of this type include small proportions of dating errors and missing values in individual partnership histories. The resulting biases in the models need to be corrected. Such errors concern 4.2% of individuals under age 61 in the sample of 37,443 people. The 1,503 individuals presenting inconsistent dates in their partnership history were excluded. The analysed sample thus comprises 35,940 individuals for 46,861 unions. After excluding unions formed before 1980, the sample is reduced to 28,347 individuals for 36,475 unions.

29The analyses were conducted using a monthly time unit. The dates of union formation and dissolution are given to the nearest month when the individual can remember them. Certain months were imputed, notably when the respondent gave a season rather than a specific month. Missing months with no indication of an approximate date were imputed randomly with a uniform distribution, taking care to maintain the consistency of the partnership history. When a union was formed and dissolved in the same year, it was generally imputed to start at the beginning of the year and to finish at the end of the year.

Descriptive analysis of event histories

30After correcting the sex and dating errors, a union survival model was built. The basic survival function corresponds to a non-parametric Kaplan– Meier estimator. It is possible to differentiate the survival functions for same-sex and different-sex unions and to determine the number of months before which a defined percentage of couples has separated. A separation probability is then calculated as the number of cohabiting union dissolutions in a given period, divided by the total number of unions at risk at the start of the period. These probabilities are calculated by 6-month union periods.

31The descriptive analysis of event histories thus gives the separation risks by union duration and type of couple. Because the focus here is on cohabiting unions, these unions begin when the partners move in together, no matter how long they have already been in a non-cohabiting relationship.

Cox proportional hazards model

32Longitudinal analysis models with explanatory factors are then developed. A Cox model is built for each of the survey respondents’ unions. The risk function of experiencing separation is conditioned by fixed and time-varying independent variables. The standard errors are grouped by individual, since an individual may have a history of several unions.

33The first two models include only the independent variable of interest, i.e. the type of couple (same-sex, different-sex), while the following ones include the independent variables and control variables. An interaction term between union formalization and type of couple serves to assess the specific effect of marriage by gender composition of the union. The same process is used to determine the effect of the presence of children born within the union or from a previous union, by type of couple.

V – Results

1 – Characteristics of unions and respondents

34Table 3 summarizes the characteristics of the sample and of same-sex (n = 385) and different-sex (n = 36,090) cohabiting couples. Most respondents in a union were still in their first union at the time of the survey.

35Among all unions formed since 1980, 1.05% are same-sex unions, whatever the period of union formation. For same-sex partners, the year of entry into union is associated with the likelihood of cohabiting, which is higher in recent periods. It is difficult to determine whether these figures are consistent with those of other surveys because few studies of individuals in same-sex unions have been conducted in Europe. Most official figures available to compare these results concern legally registered unions only. In Norway, for example, 0.7% of registered partnerships between 1993 and 2001 were between same-sex partners, as were 0.5% of partnerships in Sweden between 1995 and 2002 (Andersson et al., 2006). In their study of assortative mating of same-sex couples in the Netherlands between 2001 and 2007, Verbakel and Kalmijn (2014) found a proportion of 0.9% (both married and unmarried couples). In France, according to the Family and Housing Survey associated with the 2011 census, same-sex couples (cohabiting or not) represent 0.6% of all couples (Buisson and Lapinte, 2013). Our sample includes both couples who have formalized their relationship and those who have not. The proportion of same-sex couples may actually be slightly higher in the GGS, given that they formalize their relationship less often and are thus less visible in certain official statistics, unlike the GGS where they are reported (Banens, 2010). Identified same-sex couples also include some remaining false couples not detected by the various correction methods. Last, reluctance to report same-sex unions for fear of social stigma may also lead to an underestimation of their number (Régnier-Loilier, 2018a).

Table 3

Characteristics of same-sex and different-sex unions (%)

Table 3
Variable Overall Same-sex unions Gay couples Lesbian couples Different-sex unions Union status Marriage/partnership 15.9 19.9 11.8 53.6 Non-formalized 77.1 74.3 80.0 43.3 Unknown 7.0 5.8 8.2 3.1 Children within union Yes 6.5 3.6 9.5 56.7 No 93.5 96.4 90.5 43.3 Children from previous union Yes 17.2 12.5 22.0 20.5 No 82.8 87.5 78.0 79.5 Age difference between partners (years) 0–4 55.2 53.8 56.6 67.4 5–8 21.7 20.7 22.7 19.2 9+ 18.9 21.2 16.7 10.6 Unknown 4.2 4.4 4.0 2.8 Previous partner Yes 39.9 32.3 47.7 27.6 Different sex 9.6 10.0 9.3 27.6 Same sex 30.3 22.3 38.4 0.0 No 60.1 67.7 52.3 72.4 Birth cohort Before 1960 18.4 17.6 19.2 18.6 1960–1969 39.9 46.2 33.5 39.0 1970–1979 29.0 27.8 30.3 30.8 After 1979 12.7 8.4 17.0 11.6 Educational level Primary 2.6 2.7 2.5 2.2 Lower secondary 10.9 9.8 12.0 15.1 Upper secondary 35.0 32.8 37.3 42.2 Higher education 49.7 54.3 45.1 37.5 Unknown 1.8 0.4 3.1 3.0 Parental separation before age 18 Yes 18.1 10.7 25.6 13.0 No 76.8 84.5 69.1 81.0 Unknown 5.1 4.8 5.3 6.0 N cases 385 170 215 36,090

Characteristics of same-sex and different-sex unions (%)

Note: Weighted percentage by country.
Coverage: Cohabiting unions formed since 1980 on by individuals aged 60 and under.
Source: GGS Wave 1.

36Taking account of these limitations, we find that these two types of couples differ in several ways. Same-sex partners less frequently formalize their union (15.9%) than different-sex partners (53.6%) (Table 3). That 7.0% of same-sex couples have unknown union status is doubtless linked to the legal ambiguity sometimes associated with the process of union recognition and to the legislative changes over the last 2 decades.

37Not surprisingly, a much larger proportion of different-sex couples have children. Alongside those with children from a previous union, 56.7% have children born within the union. Among lesbian couples, the proportion is 9.5%. In 20.5% of different-sex couples, at least one partner has a child from a previous union; among same-sex couples, the proportion is almost equivalent at 17.2%. The proportion of childless couples, on the other hand, is nearly 2.5 times higher among same-sex couples (78.2% vs. 32.1%). For same-sex partners who are parents, the children are most often from a previous heterosexual union (Moore and Stambolis-Ruhstorfer, 2013; Gross et al., 2014).

38The partnership history of same-sex partners more frequently includes previous unions (39.9%), sometimes with a partner of the opposite sex (9.6%). The mean age difference between partners is greater for homosexual couples. Experience of parental separation during childhood is more frequent among women in lesbian couples, at 25.6%, compared with 10.7% among men in gay couples and 13.0% among heterosexual partners. With 1 in 2 partners having post-secondary qualifications, the mean educational level of respondents in a same-sex union is higher than that of heterosexual partners, who have an upper secondary qualification in 42.2% of cases.

2 – Duration models, separation risks

39A survival function represents the probabilities for a cohabiting union of ending after a given number of months (Figure 1). Same-sex unions are clearly more unstable than different-sex unions. After 10 months, 10% of same-sex couples are no longer living together, while for different-sex couples, this percentage is reached after 21 months. After 5 years, almost 50% of same-sex couples have separated versus just 25% of heterosexual couples.

40The separation probabilities (Figure 2) of heterosexual couples increase slightly during the first 12 months and then decrease over the long term. In other words, the longer the partners stay together, the less likely they are to separate. The separation risks of same-sex couples, on the other hand, do not decrease with time. After 3 years of cohabitation (36 months), the probabilities even tend to start increasing again. The separation probability for all of the first 5 years of union is 4.6 per 1,000 different-sex couples [CI: 4.5, 4.7] and 10.2 per 1,000 same-sex couples [CI: 8.6, 12.2]. These results reveal the relative stability of heterosexual unions compared to homosexual unions, although more analysis is needed.

Figure 1

Union survival probability by gender composition

Figure 1

Union survival probability by gender composition

Interpretation: Of cohabiting heterosexual couples, 10% have separated after 21 months.
Note: The percentages indicate the proportion of unions dissolved after a given duration.
Coverage: Cohabiting unions formed since 1980 by individuals aged 60 and under.
Source: GGS Wave 1.
Figure 2

Separation probabilities by duration and gender composition of unions

Figure 2

Separation probabilities by duration and gender composition of unions

Note: Union durations are given for 6-month periods to limit variance. For clarity, the narrower confidence intervals for different-sex couples are not presented on the graph.
Coverage: Cohabiting unions formed since 1980 by individuals aged 60 and under.
Source: GGS Wave 1.

3 – Factors behind separation

Influence of legislative context and marital status

41The differences in separation risks for same-sex and different-sex couples are linked to group composition effects. Table 4 presents six models that include the different determinants of separation risk. The unadjusted separation risk for same-sex couples is 2.37 times higher than for different-sex couples, thus confirming the first hypothesis. Lesbian unions have the highest separation risk with respect to different-sex unions, with a risk that is 25% higher than that of gay unions (Model 2). The difference in separation risk between same- and different-sex unions becomes smaller (relative risk [RR] = 1.64) when the model includes the union status (marriage/partnership or not) and the possibility of marriage for same-sex couples (Model 3).

42An interaction term between the type of couple and marriage/partnership (Model 4) enables us to determine whether the effect of union formalization on separation is identical for both same-sex and different-sex couples. In this model, the type of couple parameter represents the separation risks of same-sex couples among all married couples. The interaction term specifies whether the effect of marriage differs between different-sex and same-sex couples. For heterosexual couples, marriage substantially reduces separation risk (RR = 0.26). The strength of the effect of marriage on separation differs between same-sex and different-sex couples, however. Union formalization protects against separation for both types of couples but less strongly for same-sex couples (RR = 0.26 × 1.74 = 0.45). This appears to support the hypothesis of a differential impact of union formalization on separation risk for homosexual couples even if those who marry generally remain together for longer than those who do not. However, the number of same-sex couples in the sample who are married or in a civil union is too small (60) to draw robust conclusions. The interaction term is set aside in the other models with control variables.

43Because marriage only recently became possible for same-sex couples, separation risks cannot strictly be compared with those of couples formed before same-sex marriage was legalized. Complementary analyses (Appendix Table A.1) limited to couples formed after same-sex marriage was introduced (Belgium, Norway, Netherlands, and Sweden) show results similar to the presented models. This confirms the lesser effect of marriage on the longevity of same-sex unions compared to different-sex unions.

44All the other independent variables are included in Models 5 and 6. The separation risk remains higher for same-sex couples, but the difference is less pronounced (RR = 1.49). Lesbian unions are less stable than heterosexual unions (RR = 1.72 with respect to heterosexual unions).

Table 4

Relative risk of dissolution of cohabiting unions (Cox model)

Table 4
Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Type of couple (Ref.: Different-sex) Same-sex 2.37*** 1.64*** 1.60*** 1.49*** Gay couple 2.12*** 1.30* Lesbian couple 2.65*** 1.72*** Union status dep (Ref.: Not formalized) Marriage/partnership 0.26*** 0.26*** 0.34*** 0.34*** Same-sex × marriage/ partnershipdep 1.74* Same-sex marriage/partnership dep (Ref.: Not legalized) Legalized 1.24*** 1.24*** 0.99 0.99 Children in the union dep (Ref.: No) Yes 0.46*** 0.46*** Children from previous union (Ref.: No) Yes 1.12*** 1.12*** Age difference (years) (Ref.: 0–4) 5–8 1.00 1.00 9+ 1.23*** 1.23*** Unknown 2.97*** 2.97*** Age at entry into union (Ref.: Over 35) Under 25 1.38*** 1.38*** 25–35 1.06 1.06 Parental separation before age 18 (Ref.: No) Yes 1.27*** 1.27*** Unknown 1.60*** 1.60*** Previous union (Ref.: No) Yes 1.05 1.05 Birth cohort (Ref.: Before 1960) 1960–1969 1.31*** 1.31*** 1970–1979 1.66*** 1.66*** After 1979 2.94*** 2.94*** Educational level (Ref.: Upper secondary) Primary 1.15 1.15 Lower secondary 1.08* 1.08* University 1.01 1.01 Unknown 1.08 1.08 N unions 36,475 N separations 13,287 Wald chi-square 782.50 809.37 4,076.51 4,068.24 6,718.11 6,749.32

Relative risk of dissolution of cohabiting unions (Cox model)

Notes: dep signifies a time-dependent variable. Survey country is controlled for in all models. Couples with unknown union status are considered unmarried in the models.
Coverage: Cohabiting unions formed since 1980 by individuals aged 60 and under.
Statistical significance: *** p < .001, ** p < .01, * p < .05.
Source: GGS Wave 1.

The influence of children

45Among the independent variables included in Models 5 and 6, the presence of children born within the union reduces separation risks for all unions. Children from previous unions, on the other hand, slightly increase the separation risk (RR = 1.12); they do not have the same effect on partnership dynamics as children within the union.

46To assess the effects of parenthood by type of couple in more detail, new interaction terms can be used to measure the specific impact of children born within the union or from a previous union on the separation risks of same-sex and different-sex couples (Table 5).

47Models 1 and 2, which only consider the unadjusted effect of parenthood on separation risks (with no control variable), indicate a much lower separation risk in the presence of children born within the union and no significant effect of children from a previous union. Model 3 shows that among all heterosexual couples, separation risk is halved by the presence of a child born within the union (RR = 0.46). For same-sex couples, however, the presence of a common child (a rare occurrence) has a destabilizing effect (RR = 0.46 × 3.55 = 1.63). The presence of a child from a previous union (Model 4) increases the separation risk for heterosexual couples, all other things being equal (RR = 1.13). However, the interaction term shows the opposite is again true for same-sex couples. Having one or more children from a previous union almost halves the separation risk for same-sex couples (RR = 1.13 × 0.51 = 0.58). These results reveal that the effect on union longevity of children born within the union, and likewise of children from a previous union, is specific to the gender composition of the couple.

Table 5

Relative risk of dissolution of cohabiting unions, interactions with presence of children born within the union or from a previous union (Cox model)

Table 5
Model 1 Model 2 Model 3 Model 4 Type of couple (Ref.: Different-sex) Same-sex 1.75*** 2.37*** 1.38*** 1.66*** Children within union dep (Ref.: No) Yes 0.34*** 0.46*** 0.46*** Children from previous union (Ref.: No) Yes 1.04 1.12*** 1.13*** Same-sex × children within union 3.55*** — Same-sex × children from previous union — 0.51* Union status dep (Ref.: Not formalized) Marriage/partnership 0.34*** 0.34*** Homosexual marriage/partnership dep (Ref.: Non-legalized) Legalized 0.99 0.99 Control variables No No Yes Yes N unions 36,475 N separations 13,287 Wald chi-square 2,673.47 790.58 6,734.85 6,802.12

Relative risk of dissolution of cohabiting unions, interactions with presence of children born within the union or from a previous union (Cox model)

Notes: dep signifies a time-dependent variable. Survey country is controlled for in all models. Control variables: age difference, parental separation, age at entry into union, birth cohort, educational level, previous union. Couples with unknown union status are considered as unmarried in the models.
Coverage: Cohabiting unions formed since 1980 by individuals aged 60 or less.
Statistical significance: *** p < .001, ** p < .01, * p < .05.
Source: GGS Wave 1.

48Our third hypothesis is thus refuted: parenthood alone does not explain the residual differences in separation risk between the two types of couples. It even produces opposing effects on the separation risk of same-sex and different-sex couples, depending on the presence of children born within the union or from a different union.

49While these models offer a way to isolate the effect of parenthood on separation risk, they have a major limitation. For heterosexual couples, not having a child within the union may signify that the partners do not share the same desire for a child or that one of them is sterile. These factors may increase the separation risk if plans to have a child are not realized. This question of the desire for a child within the union is less acute for same-sex couples, many of whom may be discouraged by the difficulty of achieving this goal. In our sample, this is reflected by the small number (25) of same-sex couples with a child within the union, so caution is needed when drawing conclusions about this effect.


50In this study, combined data from six countries in the GGS were used to measure the stability of cohabiting unions by gender composition. The aim was to determine whether the separation risk for same-sex couples is greater than for different-sex couples and to measure the determinants of this risk. Marriage and children, previously denied to same-sex couples, were then studied as potential sources of the difference in separation risk between the two types of partnership.

51Our results show that same-sex unions are dissolved more frequently than different-sex unions. Five years after entry into union, almost 50% of same-sex couples have separated versus just 25% of heterosexual couples. Group composition effects and selection effects are considerable, however. For example, being married and having children with the current partner—characteristics found more frequently among different-sex couples—reduce the risk of separation. However, risk is not affected in the same way for same-sex couples with these two characteristics.

52For all couples, unions are dissolved more quickly if they have not been formalized, but formalization influences union stability less strongly for same-sex couples than different-sex couples. Marriage (or civil partnership) is insufficient to close the gap in separation risk between the two types of couples. The relationship norms for same-sex unions are less clearly defined and have a much shorter history than those for different-sex unions, suggesting commitment through marriage or registered partnership may have a lesser influence on union stability for homosexual couples (Lerch, 2007). Differences in union stability between married and unmarried same-sex couples are therefore smaller.

53Even if they are able to marry, some same-sex couples might not see marriage as an important symbol of their commitment. Other means can be used to formalize their union; for example, through a private ceremony (Reczek et al., 2009). Others may reject the norms associated with marriage, preferring to remain in a consensual union or to enter a registered partnership, if distinct from marriage (Jörgens, 2008; Courduriès, 2011; Rault, 2018). Given the larger proportion of homosexual partners who do not live together (primarily gay men), moving into a shared home may in itself represent a strong partnership commitment (Courduriès, 2011; Rault, 2018). For same-sex couples, the meaning of marriage, civil union, and consensual union is variable. The recent legalization of same-sex marriage has introduced, however, an additional factor into the relationship trajectory of same-sex partners. In the countries where the option is open to them, same-sex couples must now decide where they stand with respect to the formalization of their union (Courduriès, 2011). With greater hindsight, future studies will allow us to determine whether same-sex marriage and civil union are growing in popularity, whether norms are changing, and if so, whether the evolving context will affect future separation risk differently.

54Same-sex parenting is another area where norms have only recently begun to emerge. It has a specific impact on separation risk. While the presence of children born within the union is associated with greater stability for heterosexual couples, the reverse is true for homosexual couples. The presence of children from a previous union, on the other hand, seems to increase the longevity of same-sex unions. To have a child, same-sex couples must go through various legal procedures that differ from one country to another, such as adoption, surrogacy, or medically assisted reproduction for lesbian couples (Gross et al., 2014). This process is long and sometimes unsuccessful, so becoming a parent takes more time than for different-sex couples. If a child is born or adopted, same-sex couples may find it more difficult to achieve a work–life balance. Without predefined gender roles, same-sex couples wishing to share their roles equally may face greater constraints in their family and parental lives, and this may threaten union stability (Wiik et al., 2014). A child from a previous union, however, does not raise the same child-rearing and task-sharing issues, as its arrival was not a joint parental project of the current partners. For at least one partner, this type of parenthood forms part of a partnership trajectory in which a fertile heterosexual union preceded the current same-sex union. The traditional family values adopted with the previous different-sex partner may also be applied with the same-sex partner, bringing stability to the new union.

55Even after controlling for the composition effects of the two types of couples, above all the effects of marriage and the presence of children, separation risk remains higher for same-sex couples. Between lesbian and gay couples, it is the former who have the higher separation risk. This may be because women are more sensitive to the quality of the relationship than men; in different-sex couples, it is they who more frequently initiate divorce (Kalmijn and Poortman, 2006; Kurdek, 2006). The interval between first meeting and entry into union may be another factor, as it is shorter for lesbians than for gays, which may reflect a more rapid commitment to a cohabiting relationship, but also a greater risk of instability (Lau, 2012; Rault, 2018).

56Other factors not directly observable in the GGS data may explain the residual difference in union longevity by gender composition of the couple. For example, heteronormativity and the resulting stigmatization of homosexual couples may affect union stability. While social acceptance of same-sex relationships is increasing over time, it may still be difficult to challenge the heterosexual norm, especially without social support (Jörgens, 2008; Van Eeden-Moorefield et al., 2011). Marriage may be a strategy adopted by homosexual couples to gain social recognition.

57This research has several limitations, the main ones being the residual sex-coding errors in the survey, the small number of same-sex couples in our statistical analyses, and the observation period that limits the comparison between type of couple and all forms of union formalization. The use of longitudinal and administrative data sources (registers and annual tax data) provides a valuable future means to limit sex-coding errors. Such errors will be less frequent than in surveys and population censuses as the sex of individuals can be verified via the administrative bodies that use these data. It is more difficult to identify unmarried same-sex couples in these sources, however, as little information is available on the relationship between two unrelated cohabiting individuals (Festy, 2007). In all quantitative research aiming to identify trends specific to same-sex couples, it is crucial to take account of the methodological challenges involved. These data sources will also make it possible to incorporate new explanatory dimensions into the study of partnership trajectories, such as residential trajectories via national registers or career trajectories via tax data. They will enable us to broaden the analysis of factors influencing union stability and thereby to encompass the full range of partnership and family situations characterizing Europe in the 21st century.


I would like to thank Ester Rizzi for kindly agreeing to oversee the progress of this research project. My thanks also to Philippe Bocquier for his expert assistance with the longitudinal analysis models. Likewise. this study would not have been of the same quality without the help of Arnaud Régnier-Loilier. who helped me to grasp the challenges of measuring same-sex couples.
The data were obtained from the Generations and Gender Programme Data Archive in full compliance with the norms of confidentiality.
Table A.1

Relative separation risk of cohabiting unions, among unions formed after legalization of same-sex marriage (Cox model)

Table A.1
Model 1b Model 3b Model 4b Model 5b Type of couple (Ref.: Different-sex) Same-sex 1.74*** 1.54** 1.45** 1.48** Union status dep (Ref.: Unmarried) Marriage 0.25*** 0.24*** 0.33*** Same-sex × marriage dep 2.66* Children born within union dep (Ref.: No) Yes 0.36*** Children from previous union (Ref.: No) Yes 1.14* Age difference (years) (Ref.: 0–4) 5–8 0.91 9+ 1.12 Unknown 1.92*** Age at entry into union (Ref.: Over 35) Under 25 1.33** 25–35 0.97 Parental separation before age 18 (Ref.: No) Yes 1.11* Unknown 2.29*** Previous union (Ref.: No) Yes 0.92 Birth cohort (Ref.: Before 1960) 1960–1969 1.93*** 1970–1979 2.09*** After 1979 3.01*** Educational level (Ref.: Upper secondary) Primary 0.97 Lower secondary 1.42*** University 1.19*** Unknown 1.20 N unions 11,205 N separations 3,597 Wald chi-square 177.94 694.49 688.19 1,649.97

Relative separation risk of cohabiting unions, among unions formed after legalization of same-sex marriage (Cox model)

Notes: dep signifies a time-dependent variable. All models control for the survey country (Belgium. Norway. the Netherlands. and Sweden). Models 2 and 6 of Table 4 are not considered.
Coverage: Cohabiting unions formed since legalization of same-sex marriage in the countries concerned. individuals aged 60 and under.
Statistical significance: *** p < .001. ** p < .01. * p < .05.
Source: GGS Wave 1.


  • [1]
    The terms ‘heterosexual union’ and ‘homosexual union’ are not used in this article with reference to sexual orientation, but as synonyms, respectively, of ‘different-sex union’ and ‘same-sex union’.
  • [2]
    Each country chooses its own sampling method, giving rise to variability when the data of several countries are assembled. This article uses the national survey weightings. For more detail, see the article on data collection by Fokkema et al. (2016).
  • [3]
    Restricting the period to unions formed after 1980 makes it easier to compare same-sex and different-sex unions because same-sex unions formed before the 1980s were rare.
  • [4]
    To identify these children, we compare their year of birth with that of entry into union, as well as the links with household members or any children not living in the household. For previous unions, the respondent is asked if his or her former partner already had children when the union was formed.
  • [5]
    Because the age of former cohabiting partners is not asked in Norway and the Netherlands, it is imputed randomly to approximate the distribution obtained in the other countries.

Same-sex couples are one component of the growing diversity of family and partnership situations in Europe. Data from six countries that took part in the Generations and Gender Survey (Belgium, France, Germany, the Netherlands, Norway, and Sweden) were used to analyse the respondents’ partnership histories and the sex of current and previous cohabiting partners. After correcting for the coding errors inherent in this type of survey, a Cox model was used to compare the separation risks of different-sex and same-sex couples. The results show that the unadjusted separation risk for same-sex couples is 2.4 times higher than that for different-sex couples, but that this risk decreases after the independent and control variables are included in the model. Marriage and the presence of children, less frequent among same-sex couples, are two key factors that increase the stability of different-sex unions.

  • union
  • same-sex couples
  • cohabitation
  • separation
  • Generations and Gender Survey

La séparation chez les couples corésidents de même sexe et de sexe différent

Les couples de même sexe s’inscrivent dans un schéma de diversification des situations conjugales et familiales en Europe. L’enquête Générations et genre (Generations and Gender Survey), utilisée ici pour six pays européens (Allemagne, Belgique, France, Norvège, Pays-Bas, Suède), permet d’analyser l’histoire conjugale des individus et le sexe des partenaires avec qui ils ont cohabité au cours de leur vie. Après avoir tenu compte des erreurs d’encodage du sexe inhérentes à ce type d’enquêtes, nous comparons les risques de rupture des couples formés d’un homme et d’une femme aux couples de personnes de même sexe, avec un modèle de Cox. Ces résultats montrent que les couples de même sexe présentent un risque brut plus élevé de connaître une séparation (2,4 ois) que les couples de sexe différent, mais qui se réduit lorsque les variables explicatives et de contrôle sont introduites dans les modèles. Le mariage et la présence d’enfants constituent deux éléments majeurs augmentant la longévité des couples de sexe différent, situations que les couples de même sexe connaissent moins fréquemment.


La separación de las parejas co-residente del mismo sexo y de sexo diferente

Las parejas homosexuales se inscriben en un esquema de diversificación de las situaciones conyugales y familiares en Europa. La encuesta Generations and Gender Survey, utilizada aquí para seis países europeos (Alemania, Bélgica, Francia, Noruega, Países-Bajos, Irlanda), permite analizar la historia conyugal de los individuos así como el sexo de los conyugues con quienes han cohabitado en el curso sus vidas. Una vez considerados los errores de codificación del sexo inherentes a este tipo de encuestas, comparamos, con un modelo de Cox, el riesgo de ruptura de las parejas formadas por un hombre y una mujer con el de las parejas del mismo sexo. Los resultados muestran que las parejas del mismo sexo presentan un riesgo bruto más elevado de separación (2,4 veces más) que las parejas de sexo diferente, pero esta diferencia se reduce cuando las variables explicativas y de control se introducen en los modelos. El matrimonio y la presencia de hijos constituyen dos elementos importantes que aumentan la longevidad de las parejas de distinto sexo, situaciones que las parejas del mismo sexo conocen con menos frecuencia.


  • OnlineAndersson G., Noack T., Seierstad A., Weedon-Fekjær H., 2006, The demographics of same-sex marriages in Norway and Sweden, Demography, 43(1), 79–98.
  • OnlineAndress H.-J., Borgloh B., Bröckel M., Giesselmann M., Hummelsheim D., 2006, The economic consequences of partnership dissolution: A comparative analysis of panel studies from Belgium, Germany, Great Britain, Italy, and Sweden, European Sociological Review, 22(5), 533–560.
  • OnlineBanens M., 2010, Mariage et partenariat de même sexe en Europe [Vingt ans d’expérience], Politiques sociales et familiales, 99, 73–84.
  • Becker G., 1981, A treatise on the family, Cambridge, Harvard University Press.
  • OnlineBlackwell D., Lichter D., 2004, Homogamy among dating, cohabiting, and married couples, The Sociological Quarterly, 45(4), 719–737.
  • Blumstein P., Schwartz P., 1983, American couples: Money, work, sex, New York.
  • OnlineBozon M., Rault W., 2012, From sexual debut to first union. Where do young people in France meet their first partners? Population, English Edition, 67(3), 453–490.
  • OnlineBrines J., Joyner K., 1999, The ties that bind: Principles of cohesion in cohabitation and marriage, American Sociological Review, 64(3), 333–355.
  • Buisson G., Lapinte A., 2013, Le couple dans tous ses états: non-cohabitation, conjoints de même sexe, Pacs…, Insee Première, 1435, February.
  • OnlineCourduriès J., 2011, Être en couple (gay). Conjugalité et homosexualité masculine en France, Lyon, Presses universitaires de Lyon.
  • OnlineFesty P., 2006, Legalizing same-sex unions in Europe: Innovations and paradoxes, Population & Societies, 424, June.
  • OnlineFesty P., 2007, Numbering same-sex couples in censuses and population registers, Demographic Research, 17(12), 339–368.
  • OnlineFokkema T., Kveder A., Hiekel N., Emery T., Liefbroer A., 2016, Generations and Gender Programme Wave 1 data collection: An overview and assessment of sampling and fieldwork methods, weighting procedures, and cross-sectional representativeness, Demographic Research, 34(18), 499–524.
  • OnlineGross M., Courduriès J., De Federico A., 2014, Le recours à l’AMP dans les familles homoparentales: état des lieux. Résultats d’une enquête menée en 2012, Socio-logos, 2870(9), 1–17.
  • OnlineJörgens F., 2008, Projets de couple, projets de famille. Discours homosexuels sur le mariage et la parentalité, in Descoutures V., Digoix M., Fassin E., Rault W. (eds.), Mariages et homosexualités dans le monde, Paris, Autrement, 127–137.
  • OnlineKalmijn M., Loeve A., Manting D., 2007, Income dynamics in couples and the dissolution of marriage and cohabitation, Demography, 44(1), 159–179.
  • OnlineKalmijn M., Poortman A.-R., 2006, His or her divorce? The gendered nature of divorce and its determinants, European Sociological Review, 22(2), 201–214.
  • OnlineKurdek L., 1998, Relationship outcomes and their predictors: Longitudinal evidence from heterosexual married, gay cohabiting, and lesbian cohabiting couples, Journal of Marriage and Family, 60(3), 553–568.
  • OnlineKurdek L., 2006, Differences between partners from heterosexual, gay, and lesbian cohabiting couples, Journal of Marriage and Family, 68(2), 509–528.
  • OnlineLau C., 2012, The stability of same-sex cohabitation, different-sex cohabitation, and marriage, Journal of Marriage and Family, 74(5), 973–988.
  • OnlineLerch A., 2007, Normes amoureuses et pratiques relationnelles dans les couples gays, Informations sociales, 144(8), 108–117.
  • OnlineLesthaeghe R., 2010, The unfolding story of the second demographic transition, Population and Development Review, 36(2), 211–251.
  • OnlineLevinger G., 1965, Marital cohesiveness and dissolution: An integrative review, Journal of Marriage and Family, 27(1), 19–28.
  • OnlineLyngstad T., Jalovaara M., 2010, A review of the antecedents of union dissolution, Demographic Research, 23(10), 257–292.
  • OnlineManning W., Brown S., Stykes B., 2016, Same-sex and different-sex cohabiting couple relationship stability, Demography, 53(4), 937–953.
  • OnlineMoore M., Stambolis-Ruhstorfer M., 2013, LGBT sexuality and families at the start of the twenty-first century, Annual Review of Sociology, 39(1), 491–507.
  • OnlineOppenheimer V., 1988, A theory of marriage timing, American Journal of Sociology, 94(3), 563–591.
  • OnlineRault W., 2018, La distance dans les relations conjugales et familiales des gays et des lesbiennes, in Imbert C., Lelièvre E., Lessault D. (eds.), La famille à distance, Paris, INED, 257–275.
  • OnlineReczek C., Elliott S., Umberson D., 2009, Commitment without marriage: Union formation among long-term same-sex couples, Journal of Family Issues, 30(6), 738–756.
  • OnlineRégnier-Loilier A., 2018a, Are the Generations and Gender Surveys well suited for studying same-sex couples? European Journal of Population, 34(4), 567–578.
  • OnlineRégnier-Loilier A., 2018b, Faire couple à distance. Prévalence et caractéristiques, in Imbert C., Lelièvre E., Lessault D. (eds.), La famille à distance, Paris, INED, 165–193.
  • OnlineRosenfeld M., 2014, Couple longevity in the era of same-sex marriage in the United States, Journal of Marriage and Family, 76(5), 905–918.
  • Van De Kaa D., 1987, Europe’s second demographic transition, Population Bulletin, 42(1), 1–59.
  • OnlineVan Eeden-Moorefield B., Martell C., Williams M., Preston M., 2011, Same-sex relationships and dissolution: The connection between heteronormativity and homonormativity, Family Relations, 60(5), 562–571.
  • OnlineVerbakel E., Kalmijn M., 2014, Assortative mating among Dutch married and cohabiting same-sex and different-sex couples, Journal of Marriage and Family, 76(1), 1–12.
  • OnlineWeisshaar K., 2014, Earnings equality and relationship stability for same-sex and heterosexual couples, Social Forces, 1–31.
  • OnlineWiik K., Seierstad A., Noack T., 2014, Divorce in Norwegian same-sex marriages and registered partnerships: The role of children, Journal of Marriage and Family, 76(5), 919–929.
Benjamin Marteau
Institut national d’études démographiques (INED); Center for Demographic Research, Université Catholique de Louvain (UCL).
Institut national d’études démographiques, Campus Condorcet, 9 cours des Humanités - CS 50004, 93322 Aubervilliers Cedex, France.
Translated by
Catriona Dutreuilh
This is the latest publication of the author on cairn.
Latest publication on cairn or another partner portal
Uploaded on on 26/02/2020
Distribution électronique pour I.N.E.D © I.N.E.D. Tous droits réservés pour tous pays. Il est interdit, sauf accord préalable et écrit de l’éditeur, de reproduire (notamment par photocopie) partiellement ou totalement le présent article, de le stocker dans une banque de données ou de le communiquer au public sous quelque forme et de quelque manière que ce soit.
Loading... Please wait