Explaining Couple Cohesion in Different Types of Gay Families

Abstract

This Internet-based study used data from a convenience sample of 176 gay men in current partnerships to examine differences in outness, cohesion, and relationship quality between three types of gay male couples: first cohabiting partnerships, repartnerships, and gay stepfamilies. Also, we tested whether relationship quality mediated the link between outness and cohesion and the moderating role of type of relationship. Results showed that those in first cohabiting partnerships had the lowest levels of relationship quality and cohesion, whereas those in gay stepfamilies reported having the highest levels of relationship quality and those in repartnerships reported the highest levels of cohesion. For all couples, the link between outness and cohesion was partially mediated by relationship quality, and this was moderated for those in repartnerships.

Keywords

cohesion gay couples relationship quality stepfamilies

There is a vast history of research that examines the role of family structure in influencing and differentiating various types of heterosexual families and their relationship experiences (e.g., Bradbury, Fincham, & Beach, 2000; Brown & Booth, 1996; Coleman, Ganong, & Fine, 2000). Generally, this research indicates that certain types of heterosexual families (cohabiting, intact, remarried, stepfamilies) differ in terms of relationship processes and outcomes, thereby suggesting the importance of structure. There is also strong and consistent research suggesting that gay couples are comparable to various types of heterosexual couples (e.g., cohabiting, married) in terms of relationship processes and outcomes (Kurdek, 2004). However, one potentially important factor unique to gay couples is that of outness, or the degree to which others know one’s sexual orientation, and research suggests this influences relational outcomes (e.g., Elizur & Mintzer, 2003; Mohr & Fassinger, 2006).

In spite of calls for more studies that address within-group variation (e.g., Demo, Aquilino, & Fine, 2005), studies such as those mentioned above (e.g., Kurdek, 2004) often have treated gay couples as monolithic or simply controlled for the influence of certain variables (e.g., length of relationship), potentially masking differences in structure (see, Lewis, Derlega, Berndt, Morris, & Rose, 2001). It is likely that the lack of social recognition, the changing nature of legal status, and the resulting difficulty of conceptually differentiating types of gay couples has played a role in their monolithic presentation. However, recent changes in legal status (e.g., establishment of civil unions, domestic partner registries, same-sex marriage) led some scholars to examine the diversity within gay families (Solomon, Rothblum, & Balsam, 2004). Such research suggests that type of relationship or family structure might be an important determinant of relationship processes and outcomes.

Here, we add to this literature by focusing on the relational differences among three types of gay relationships: those in first cohabiting relationships (first partnerships), in repartnerships (second cohabiting relationships), and in gay stepfamilies (gay cohabiting couples in which one or both were in a previous relationship and had at least one child in that relationship). Given previous research that finds some gender differences among same-sex couples (e.g., Kurdek, 2004; Lynch, 2000), especially when children are present (e.g., Goldberg, 2009) and to maximize our exploration of within-group variation among gay couples, only gay male couples are examined here. Additionally, we hypothesized that the more gay men in relationships are out (i.e., the extent to which others are aware of one’s sexual orientation), the more cohesive their relationships will be (i.e., perceived couple unity), and this link will be mediated by relationship quality (i.e., perceived match between current and ideal relationship standards around companionship, trust, and conflict) and moderated by type of relationship.

Conceptual Framework

We draw on the work of Demo et al. (2005) and use a family composition and transitions framework. This framework suggests that common transitions affecting family composition include transitions from singlehood to cohabitation, marriage to divorce, and divorce to remarriage. These transitions are associated with myriad stressors (e.g., changes in family relationships among extended family members, changes in parenting and child custody, changes in daily routines and rituals, changes in residence). Such stressors are linked to changes in individual, couple, and family well-being and other outcomes (Demo et al., 2005). Unique to gay couples, outness would be expected to affect couple/family functioning and outcomes. Theoretically, if one or both partners are less out to others, they likely will hide their relationships and this can create stress (Elizur & Mintzer, 2003). Furthermore, these stressors also might differ by family composition. For example, married couples who divorce necessarily experience different stress based on the presence or absence of children (Coleman et al., 2000). Couples without children will experience stress from dividing marital assets, whereas couples with children will experience the added stress of both dividing assets and determining custody and visitation arrangements.

For gay couples, transitions that affect family composition can take several forms, such as entering a first cohabiting partnership or subsequent cohabiting partnerships, which we term repartnerships. Another transition affecting family composition occurs when a heterosexually married man comes out as gay (at least to himself), divorces his wife, and eventually enters a cohabiting relationship with another man. When a child is present from his previous marriage, the resulting family is a gay stepfamily (Berger, 2000; Crosbie-Burnett & Helmbrecht, 1993). We recognize that some gay stepfamilies are not formed following a heterosexual marriage, but this is the most common pathway currently (Berger, 2000). Our interest here is how these three family structures (first partnerships, repartnerships, and gay stepfamilies) differ. This framework also is consistent with Cherlin’s (1978) incomplete institutionalization hypothesis, which asserts those in remarriages and stepfamilies experience more difficulty and decreased relational stability because their relationships are not supported socially. Applied to gay couples, none of their relationships are institutionalized in most respects or in most states; for those in repartnerships and stepfamilies, there is a double deinstitutionalization. Furthermore, within the gay community children who are part of a relationship are embraced by some and said to be a symbol of heteronormative conformity by others (Lannutti, 2005). As such, it may well be that gay stepfamilies are particularly vulnerable to ambiguity and effects of incomplete institutionalism.

Differences in Outness, Relationship Quality, and Cohesion

Outness

For gay men, coming out is a significant life event that partially defines one’s sexual identity and is associated with various individual and couple outcomes (e.g., Elizur & Mintzer, 2003; Haas & Stafford, 1998; Savin-Williams, 2001). However, research has focused more explicitly on how outness and other gay identity variables influence individual outcomes. What we do know suggests that gay men who are more out report less depression and higher relationship quality than those who are not out (Elizur & Mintzer, 2003). Alternatively, those who are less out also report lower self-acceptance and lower relationship quality.

Coming out can occur at various ages over one’s life and continues as men interact with new others (Savin-Williams, 2001). In this way, it is a dynamic and continuous developmental process. Because transitions that affect family composition include meeting and interacting with new individuals (Demo et al., 2005), we hypothesized that outness differs by type of couple. Based on where individuals likely are developmentally, including their age (i.e., younger), we expect those making a transition to a first cohabiting relationship are least out (Rutter & Schwartz, 1996). Conversely, we assumed that those in a repartnership will be more out on average than those in a first partnership. Gay stepfamilies present a unique group (Lynch, 2000). Previous research (Berger, 2000) suggests that some gay individuals in stepfamilies are likely to hide their sexual orientation, most often because of the added difficulty over child custody issues. Additionally, many of these individuals come out as part of the divorce process nonetheless, and this is more common among gay men exiting a previous heterosexual marriage than it is among lesbians (Lynch, 2000). Thus, we predicted that gay men in stepfamilies will be more out than those in first partnerships but similar to those in repartnerships.

Relationship quality

The second variable of interest here is relationship quality. Generally, research (e.g., Brown & Booth, 1996) suggests that relationship quality among heterosexual couples differs by family structure and is comparable between heterosexual and gay relationships (e.g., Kurdek, 2004). However, findings that compare relationship quality between those heterosexual couples in first marriages and remarriages are mixed. For example, some studies found higher relationship quality among first marriages than remarriages, whereas others reported no differences between groups (Coleman et al., 2000). Taken together, we do not expect to find differences in relationship quality between those in first cohabiting relationships and repartnerships. However, again we believe that gay stepfamilies represent a unique group, and they will report higher levels of relationship quality than those in first partnerships (Crosbie-Burnett & Helmbrecht, 1993; Lynch, 2000). We posited this for two reasons. First, gay men who were in a heterosexual relationship should experience a heightened level of relationship quality when they enter into a partnership that is consistent with their sexual identity (Stets, 1993) and their relationship ideals (Fletcher, Simpson, & Thomas, 2000). Second, lowered marital quality in stepfamilies often is attributed partially to issues of stepparenting (e.g., Braithwaite & Baxter, 2006). Because the focus here is on gay men, and men are less likely to be custodial parents, we believe that the negative effects of stepfamily living on relationship quality will not be present or will be lessened. In this way, gay stepfamilies should differ from first partnership and repartnership structures.

Cohesion

Previous studies found differences in cohesion associated with family structure and composition (Brines & Joyner, 1999; Lavee & Olson, 1991). However, to date no research has examined cohesion in gay relationships explicitly or from a quantitative perspective, as is done here. Thus, we offer a tentative prediction on how cohesion differs among the groups we studied. We speculated that the lack of legal status and difficulty finding social support to build and maintain cohesion will result in lower levels of cohesion among those in first partnerships than either of the other two groups (LaSala, 2000). For those in repartnerships, we expected that they are more adept at navigating these issues because of prior relationship experiences. Thus, prior partnerships might help couples develop certain relationship maintenance skills of which cohesion is one such skill. In fact, cohesion includes the ability to establish and maintain boundaries with the outside world, as well as maintain emotional bonds within the family system (Olson & Gorall, 2003). Given the discrimination experienced by gay men and couples, the need to navigate and establish boundaries within a stigmatizing social world is paramount to couple development (Mohr & Fassinger, 2006). As such, prior relationship experience may be crucial to the development of cohesion. Gay stepfamilies are formed after prior relationship experiences that aid in general relationship skill development specific to heterosexual relationships; however, they might lack skills needed to combat social stigma that accompanies gay stepfamilies (Crosbie-Burnett & Helmbrecht, 1993; Lynch, 2000), both as part of being gay and a stepfamily. Thus, we speculated that cohesion among gay stepfamilies will fall between those in first partnerships and those in repartnerships.

Our Model Explaining Cohesion in Gay Relationships

Our second goal was to examine how level of outness, relationship quality, and cohesion are linked and to explore variations by relationship type. Specifically, we hypothesized that the more out gay men in relationships are, the more cohesive their relationships will be, and that this link will be mediated partially by the quality of their relationship. Furthermore, we posited that the mediated model will differ by type of relationship.

As stated earlier, previous research indicates that level of outness influences various relationship processes and outcomes (Elizur & Mintzer, 2003). Specific to outness and couple outcomes, Mohr and Fassinger (2006) found that each dimension of LGB identity influenced relationship quality and that perceived identity similarity between partners also affected relationship quality. However, actor rather than partner effects matter most. Additionally, LaSala (2000) interviewed 20 gay couples and found that being out played an important role in allowing them to feel more connected as a couple, a feeling that reflects cohesion. It is likely that being out reduces internal stressors such as homonegativity and stigma consciousness (Pinel, 1999) and that fewer internal stressors reduce depression and anxiety among individuals, thereby influencing the relationship directly and indirectly. Additionally, other research found that being out also is linked to increased relationship quality (e.g., Lewis, Kozac, Milardo, & Grosnick, 1992) and that relationship quality is associated with cohesion (e.g., Fisiloglu & Lorenzetti, 1994). Qualitative scholars (e.g., McQueeney, 2003; Oswald, 2000, 2002) found that being out as a couple and subsequently being recognized and treated as a couple provided a sense of cohesion and positively influenced relationship quality. However, none of these studies examined or measured cohesion explicitly. Similar findings related to the link between relationship quality and cohesion also are reported for heterosexual couples (e.g., Craddock, 1991; Fisiloglu & Lorenzetti, 1994) and grandparent–grandchild dyads (Ruiz & Silverstein, 2007). Thus, we argue that the link between outness and cohesion is at least partially mediated by relationship quality. Because we outlined previous research earlier suggesting that each of the variables in our model differs by type of relationship, we assumed that the links in the proposed model are moderated by the type of relationship.

Method

Procedures and Sample

An Internet-based survey was used to collect data because previous findings demonstrate increased access to marginalized populations (e.g., LGBT; Davis, Bolding, Hart, Sherr, & Elford, 2004), efficiency, and the cost-effectiveness of using the Internet in this way (e.g., Murray & Fisher, 2002; Rhodes, DiClemente, Cecil, Hergenrather, & Yee, 2002; van Eeden-Moorefield, Proulx, & Pasley, 2008). Although we acknowledge that this method limits the ability to reach certain populations (e.g., those who live in rural areas where Internet access is not widely available, those who have fewer financial resources), Internet users are becoming more diverse in ethnicity and socioeconomic status (Pew Internet Life Project, 2007; U.S. Department of Commerce, 2004). Most research on LGBT families has relied on more affluent and homogeneous samples (e.g., Kurdek, 2004). Although we hoped to obtain a more diverse sample through this method, we also recognize that samples obtained by Internet methods might be similar to those garnered by more traditional sampling methods. The Internet provides an added benefit particularly appropriate for research on gay men—that of greater confidentiality. As such, we hoped that marginalized individuals, who might otherwise not participate for fear of being outed or experiencing further social stigmatization (Davis et al., 2004; van Eeden-Moorefield et al., 2008), would participate in this study.

To date, most online research relies on creating survey websites and registering them with search engines or establishing links from other sites (Rhodes et al., 2002). In many cases, the surveys are open to anyone (as opposed to screening and use of passwords) for participation (including multiple completions by the same individual), thereby producing an opportunity for increased bias (i.e., lack of independence; Copeland & White, 1991). Measures taken in the current study to reduce this potential are outlined below.

A convenience sample of 43 nationally and locally based (southeastern) organizations that cater to gay men and or gay couples was contacted. Local organizations were targeted for oversampling due to needs related to meeting the goals of a larger study. Eighteen e-mails were returned as invalid, and 14 responded and agreed to assist in recruitment by sending study announcements via listservs or newsletters. Those who responded and declined to assist in recruitment (11) did so because of organizational policies or because their next newsletter would not be sent until after the scheduled study completion date. Although organizations also were asked for available member demographic information (e.g., the number of gay men that potentially would receive study notification), they reported that they did not maintain such information.

Potential participants contacted the first author directly to obtain additional information and be screened for participation (i.e., participants must self-identify as gay, currently be involved in a gay relationship of at least 3 months, and only one partner was asked to participate). Those interested were provided with the secure survey link, a username, and a password. Adding screening and e-mail communication steps to the participation process decreases the likelihood of problems with independence (Murray & Fisher, 2002; Rhodes et al., 2002).

These methods resulted in a total of 249 e-mail inquiries, of which 228 met study criteria and received passwords to complete the survey. Because online studies that offer incentives demonstrate high completion rates (Reips, 2002), participants were given the opportunity to chose from three national charities (e.g., Human Rights Campaign), and an anonymous donation was made on their behalf. The resulting participation rate was 83%, which is consistent with previous studies and represents a high rate of participation.

The sample used in this analysis was drawn from a larger study (van Eeden-Moorefield, 2005) and includes only those men (N = 176) who self-identified as gay, were in a cohabiting relationship, and did not adopt or otherwise bring children into the current relationship (e.g., use of a surrogate; see Table 1). Overall, the sample was mostly White, well educated (79% had at least a bachelor’s degree), and employed, and almost half (48%) earned at least $50,000 or more. They represented 26 states and the District of Columbia. Their mean age was 40.7 years (SD = 11.66 years), the average length of their relationship was 8.48 years (SD = 8.44 years), and most (72.9%) were in closed (i.e., monogamous) relationships. Taken together, the sample was fairly homogeneous.

Table 1.

Demographic Characteristics of the Sample (N = 176) in Frequency and Percentage

Variable	n	Percentage
Race/ethnicity
Asian	3	1.7
Black	2	1.1
Hispanic/Latino	8	4.5
White	158	89.8
Other	5	2.8
Education
Less than or equal to High school	5	2.7
Some college	34	18.2
Four-year degree	44	25.0
Some graduate school	17	9.7
Graduate degree	68	38.6
Other (e.g., specialist’s degree)	10	5.3
Employment
Unemployed	6	3.4
Full-time	127	72.2
Part-time	24	13.6
Retired	12	6.8
Other	7	4.0
Gross annual income (2003)
<$10,000	15	8.5
$10,000-29,999	31	17.7
$30,000-49,999	49	27.8
$50,000-69,999	19	10.8
$70,000-89,999	24	13.6
≥$90,000	38	21.6
Current sexual relationship
Open	48	25.5
Closed	137	72.9
Other	3	1.6

Measurement

Family structure

Family structure was measured using responses to several relationship history questions (e.g., whether they were previously cohabiting with another man, whether they were previously married, whether they had a previous relationship) and questions about children from previous heterosexual marriages. Based on responses to these questions, respondents were grouped into three structures: first partnerships (n = 85), repartnerships (n = 63), and gay stepfamilies (n = 28). Unfortunately, no questions assessed residential status, age of child, or whether there was more than one prior cohabiting relationship with another man.

Outness

The Outness Inventory (Mohr & Fassinger, 2003) is a 15-item summed measure that assesses the extent to which the respondent’s sexual orientation is known and openly discussed with others (e.g., mother, best friend). Although it is likely that the importance of any individual who knows one’s sexual identity (e.g., mother vs. coworker) might differentially affect certain outcomes, scholars suggest a more linear process in who is told and when (Savin-Williams, 2001; close friends, followed by siblings, mothers, etc.). As such, the more people to whom one is out represents a global level of outness, and that is our focus here. Individual item responses vary from does not know (1) to definitely knows, openly discussed (7). Higher scores reflect higher levels of being out with a possible range from 15 to 105, and the α for this sample was .66. Previous research reports αs ranging from .72 to .97 (Balsam & Szymanski, 2005).

Relationship quality

The Marital Comparison Level Index (Sabatelli, 1984) was used to measure relationship quality; it is a 31-item index assessing perceptions of how current aspects of the relationship match their ideals (sample item: “amount of companionship”). Responses range from worse than I expected (−3) to better than I expected (+3). All words related to marriage were replaced with relationship or partnership. Possible scores range from −93 to +93; the α for this sample was .94.

Cohesion

The Cohesion Subscale of the Family Adaptability and Cohesion Evaluation Scale (Olson, 1986) was used to measure cohesion. The subscale contains 10 items that ask about the jointness, or cohesion, between the partners (sample item: “togetherness is important”) and should not be interpreted as a measure of enmeshment. The summed items are scored on a Likert-type scale, ranging from almost never (1) to almost always (5). Possible scores range from 10 to 50, and the α for this sample was .92.

Results

Preliminary Analyses

Findings showed that the full sample was fairly out, believed that their current relationships matched their ideals slightly more than expected, reflecting moderate relationship quality, and were moderate to high on cohesiveness (see Table 2). Results of bivariate correlations showed that all variables were correlated in the expected direction, which provides evidence that supports testing for mediation (Baron & Kenny, 1986).

Table 2.

Means, SDs, Alphas, and Correlations Among Variables

	OI	RQ	CO
OI	—	.35*	.41*
RQ	—	—	.62*
CO	—	—	—
Mean	70.93	21.96	42.84
SD	11.47	21.65	6.41
α	.66	.94	.92

Note: OI = outness; RQ = relationship quality; CO = cohesion.

p < .05. **p < .01.

Differences by Family Structure

As hypothesized, results from three one-way ANOVAs suggest that there are differences in mean scores by family structure for each of the variables (see Table 3). Post hoc analyses showed that generally our hypotheses were supported with one exception. Specifically, we hypothesized that mean outness scores would be lowest for those in first partnerships and highest for those in repartnerships and that these would be significantly different. We also suggested that those in gay stepfamilies would have scores between the other two groups but differ significantly from those in first partnerships only. Results supported these assertions.

Table 3.

Comparisons of Mean Scores by Family Structure (N = 176)

	Family Structure Means (SD)
Variables	First Partnerships (n = 85)	Repartnerships (n = 63)	Stepfamilies (n = 28)	F
OI	67.87 (12.04)_a	74.33 (8.85)_b	72.54 (12.77)_ab	6.45*
RQ	17.13 (20.20)_a	25.60 (22.26)_b	28.43 (21.95)_c	4.43*
CO	41.51 (6.52)_a	44.33 (5.13)_b	43.50 (7.88)_ab	3.82*

Note: OI = outness; RQ = relationship quality; CO = cohesion. Means in the same row that do not share subscripts differ at p < .05 based on Bonferroni post hoc analyses.

p < .05. **p < .01.

We also hypothesized that those in first partnerships would have the lowest relationship quality scores, but they would not differ significantly from those in repartnerships. In our analyses this was not the case. In fact, those in repartnerships reported higher relationship quality than those in first partnerships. Furthermore, we hypothesized that those in gay stepfamilies would report scores between the other two groups. Our results indicated that gay stepfamilies differed significantly from both those in first partnerships and repartnerships with the highest levels of relationship quality. Finally, we hypothesized that those in first partnerships would report the lowest levels of cohesion followed by gay stepfamilies and then those in repartnerships. Although this was the case in terms of mean scores, the difference was between first partnerships and repartnerships only.

Generally, our results suggest that those in their first partnerships are the group most at risk for poor relationship outcomes, given their lower levels of outness, relationship quality, and cohesion compared with those in repartnerships and stepfamilies. However, overall the differences are not large.

Testing the Overall Model Explaining Cohesion

We first tested the mediating model with the full sample, in which we hypothesized that the effect of outness on cohesion would be partially mediated by relationship quality. Relationship duration, education level, and income were included as control variables given the correlation between these variables and both outness and relationship outcomes (e.g., Kurdek, 2004; Lewis et al., 1992). Education level and income were not significant and, thus, were dropped from the model. Relationship duration was retained. Results of the hierarchical regression suggest that relationship quality partially mediated the relationship between outness and cohesion, accounting for 44% of the variance in cohesion (see results of Model 3 in Table 4). Specifically, those who reported being more out to others also reported higher relationship quality, which then was associated with higher levels of cohesion. This finding is consistent with previous qualitative studies (e.g., LaSala, 2000; Oswald, 2000, 2002). Note that the second step suggested by Baron and Kenny (1986), in which outness predicts relationship quality, is not reported in the model. Results were significant for this step, β = .35, t = 4.87, p = .00. Additionally, results of Sobel’s (1982) test suggest that partial mediation was significant, t(173) = 4.28, p = .00.

Table 4.

Summary of Hierarchical Regression Analysis Predicting Cohesion as Mediated by Relationship Quality (N = 176)

	Model 1			Model 2			Model 3
Variable	B	SE B	β	B	SE B	β	B	SE B	β
RD	.02	.01	.23**	.01	.01	.15*	.01	.01	.14*
OI				.21	.04	.38**	.11	.03	.19**
RQ							.16	.02	.54**
R ²		.05			.19			.45
ΔR²		.05			.18			.44
F for change in R²		9.72*			29.13**			79.95**

Note: RD = relationship duration; OI = outness; RQ = relationship quality; CO = cohesion.

p < .05. **p < .01.

Next, we examined the same mediating relationship with the inclusion of family structure as a moderator (see Table 5). Results showed that family structure did moderate the model once the mediator was included. Compared with those in first partnerships, there was a significant interaction between being out and family structure for those in repartnerships. Thus, it appears that being out may play a more significant role in predicting relationship quality and cohesion for those in repartnerships beyond that of first cohabiting couples and stepfamilies. However, before the mediator was entered the link between outness and cohesion was also moderated such that outness played a significant role for those in first partnerships only. Consistent with results for the full model, the final model was partially mediated, again accounting for 44% of the variance in cohesion.

Table 5.

Summary of Hierarchical Regression Analysis Predicting Cohesion as Mediated by Relationship Quality and Moderated by Family Structure (N = 176)

	Model 1			Model 2			Model 3			Model 4
Variable	B	SE B	β	B	SE B	β	B	SE B	β	B	SE B	β
RD	.02	.01	.23*	.01	.00	.16*	.01	.00	.15*	.01	.00	.14*
OI				.19	.04	.34**	.18	.04	.32**	.10	.04	.18**
FS
1 versus 0				−.55	.33	−.22	−.49	.33	−.11	−.32	.28	−.07
2 versus 0				−.44	−.43	−.08	−.48	.43	−.08	.18	.37	.03
FS × OI
1 versus 0 × OI							−.02	.02	−.08	.03	.02	.17*
2 versus 0 × OI							−.04	.02	−.18*	.02	.02	.08
RQ										.17	.02	.58**
R ²		.05			.20			.22			.46
ΔR²		.05			.18			.19			.44
F for change in R²		9.72*			10.74**			2.03			76.51**

Note: OI = outness; RQ = relationship quality; CO = cohesion; FS = family structure. FS was represented as two dummy variables with first partnerships (0) serving as the reference group: 1 = repartnerships; 2 = stepfamilies.

p < .05. **p < .01.

Discussion

The pathways that some gay couples and families take to recognize and legitimize their relationships have transformed with the development and passage of domestic partnership policies and civil union and marriage laws in several U.S. states (Solomon et al., 2004) although it can be argued that these families remain incomplete institutions (Cherlin, 1978). Whereas some gay couples continue to live in “invisible” relationships (e.g., LaSala, 2000; Oswald, 2000), others now enjoy some of the benefits of cohabiting, civil unions, and legal marriage, or at least the opportunity to choose these and other relational options that recognize and make visible their partnerships (Solomon et al., 2004). Added to this is the increasing likelihood and opportunity for gay couples to bring a child into their relationships either through adoption, surrogacy, or from a previous heterosexual marriage (Crosbie-Burnett & Helmbrecht, 1993; Lynch, 2000). As such, the individual and relational experiences of transitions that affect family composition differ, leading to an increased need to examine within-group variation (Demo et al., 2005).

The results of our study contribute to the literature on within-group variation in gay couples in three ways. First, the results respond to the call by scholars (e.g., Demo et al., 2005) to add to a scant literature examining within-group variation among couples in general and gay couples in particular (Elizur & Mintzer, 2003). However, the current study did not produce the desired diverse sample especially in terms of ethnicity, which likely is another key moderating variable to consider in future research. There is research showing that responses to homosexuality and experiences of such couples is ethnically unique (e.g., McQueeney, 2003). With few exceptions, our findings related to differences in outness, relationship quality, and cohesion between those in first partnerships, repartnerships, and gay stepfamilies were as expected. In some cases our findings were somewhat unlike those of others (e.g., Kurdek, 2004) where, on average, gay couples were similar to their heterosexual counterparts. We assert that this is because our findings demonstrate variation that potentially was masked in previous research where all gay couples were treated as similar rather than distinct. Instead, our findings are more consistent with those of Lewis et al. (2001), who suggested some level of influence of structure on relational experiences.

A second and noteworthy contribution is the finding that gay couples in repartnerships reported higher relationship quality and cohesion than those in first partnerships—a finding that adds to the mixed findings from other studies about the nature of marital quality among heterosexual couples in first marriages compared with remarriages (Coleman et al., 2000). We also found support for our prediction that relationship quality would be higher among gay stepfamilies compared with other gay couples, consistent with previous research (Crosbie-Burnett & Helmbrecht, 1993; Lynch, 2000). Earlier we suggested that gay stepfamilies were unique in that they present an opportunity for a man in a previous heterosexual marriage to reconcile his sexual identity (Stets, 1993) and his relationship ideals (Fletcher et al., 2000). We also suggested that lowered marital quality in stepfamilies is attributed partially to issues of stepparenting or selection effects (e.g., Braithwaite & Baxter, 2006) and that because gay stepfamilies were less likely to have custodial children present, they would have higher relationship quality. The potential for selection effects to operate in our sample is possible. From Cherlin’s (1978) hypothesis, we argue that none of these couples, regardless of type, enter institutionalized relationships. Thus, these hypothesized effects are equally experienced by all gay couples, but the degree to which they are experienced likely differs by state of residence and the individual marriage laws within the state. Taken together, the fact that we did not ask specific questions about child custody, stepparenting, or perceived social deinstitutionalization are limitations of the current study and should be addressed in future research.

The findings from testing our model of relationship processes predicting cohesion showed that the model worked equally well for all couples. However, when family structure was added as a moderator, and before relationship quality was entered, there was a moderating effect of family structure on cohesion. That is, those in stepfamilies who were more out had lower cohesion. Once relationship quality was entered as a mediator, this effect was reduced to nonsignificance and a significant effect emerged for those in repartnerships (e.g., higher outness was related to higher cohesion as mediated by relationship quality). It may be that this finding can be explained in future studies that also incorporate adaptability (Olson & Gorall, 2003) into the model. For example, those in repartnerships were more out, and this may mean that less change and adaptability are needed in daily life and relationships. As such, cohesion might be more important for these couples. Alternatively, those in first cohabiting relationships and stepfamilies were less out, representing more potential changes, in addition to the added transitions related to forming a stepfamily. For these two groups of couples, adaptability may be more salient in conjunction with relationship quality. Certainly, these are speculations and should be tested in future studies. Furthermore, the lack of data on custodial arrangements and other stepfamily life variables are limitations affecting interpretation of our results. Importantly, this result differs from the established literature on heterosexuals that generally suggests relationship processes do not differ among those in remarriages and those in stepfamilies (e.g., Coleman et al., 2000). Our findings also slightly differ from literature suggesting differences between those in first marriages compared with remarriages and stepfamilies (e.g., Bradbury et al., 2000; Brown & Booth, 1996). Clearly, the inclusion of family structure and level of outness are important in understanding the diversity within gay relationships, and future research should include these. We make this recommendation realizing that the reliability of the outness measure was less than desirable and any recommendations must be made with caution.

Overall and consistent with previous research, the results presented here suggest that although family structure is important, family processes appear to be more important in explaining family life and family outcomes among gay couples in our study. Clearly our results support those of Kurdek (2004), suggesting that gay and heterosexual families are more similar than different. Our hypotheses were derived from the broader literature on remarriage and stepfamilies and generally were consistent. For gay men, however, we believe that the role of outness is both unique and important when examining the relationship experiences of gay couples.

Footnotes

This article was presented at the 70th National Council on Family Relations Annual Conference, November 5-8, 2008.

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Balsam

Szymanski

(2005). Relationship quality and domestic violence in women’s same-gender relationships: The role of minority stress. Psychology of Women Quarterly, 29, 258-269.

Baron

Kenny

(1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173-1182.

Berger

(2000). Gay stepfamilies: A triple-stigmatized group. Families in Society, 5, 504-516.

Bradbury

Fincham

F. F.

Beach

(2000). Research on the nature and determinants of marital satisfaction: A decade in review. Journal of Marriage and the Family, 62, 964-981.

Braithwaite

Baxter

L. A.

(2006). “You’re my parent but you’re not”: Dialectical tensions in stepchildren’s perceptions about communicating with the nonresidential parent. Journal of Applied Communication Research, 34, 30-48.

Brines

Joyner

(1999). The ties that bind: Principles of cohesion in cohabitation and marriage. American Sociological Review, 64, 333-355.

Brown

Booth

(1996). Cohabitation versus marriage: A comparison of relationship quality. Journal of Marriage and the Family, 58, 668-678.

Cherlin

(1978). Remarriage as an incomplete institution. American Journal of Sociology, 84, 634-650.

Coleman

Ganong

Fine

(2000). Reinvestigating remarriage: Another decade of progress. Journal of Marriage and the Family, 62, 1288-1307.

10.

Copeland

White

(1991). Studying families. Newbury Park, CA: Sage.

11.

Craddock

A. E.

(1991). Relationships between attitude similarity, couple structure, and couple satisfaction within married and de facto couples. Australian Journal of Psychology, 43, 11-16.

12.

Crosbie-Burnett

Helmbrecht

(1993). A descriptive empirical study of gay male stepfamilies. Family Relations, 42, 256-263.

13.

Davis

Bolding

Hart

Sherr

Elford

(2004). Reflecting on the experiences of interviewing online: Perspectives from the Internet and HIV study in London. AIDS Care, 16, 944-952.

14.

Demo

Aquilino

Fine

(2005). Family composition and family transitions. In Bengtson

V. L.

Allen

Acock

Dilworth-Anderson

Klein

(Eds.), Sourcebook on family theory and research (pp. 119-142). Thousand Oaks, CA: Sage.

15.

Elizur

Mintzer

(2003). Gay males’ intimate relationship quality: The roles of attachment security, gay identity, social support, and income. Personal Relationships, 10, 411-435.

16.

Fisiloglu

Lorenzetti

A. F.

(1994). The relation of family cohesion to marital adjustment. Contemporary Family Therapy, 16, 539-552.

17.

Fletcher

Simpson

Thomas

(2000). Ideals, perceptions, and evaluations in early relationship development. Journal of Personality and Social Psychology, 79, 933-940.

18.

Goldberg

(2009). Lesbian parents and their families: Complexity and intersectionality from a feminist perspective. In Lloyd

Few

Allen

(Eds.), Handbook of feminist family studies (pp. 108-120). Thousand Oaks, CA: Sage.

19.

Haas

S. M.

Stafford

(1998). An initial examination of maintenance behaviors in gay and lesbian relationships. Journal of Social and Personal Relationships, 15, 846-856.

20.

Kurdek

L. A.

(2004). Are gay and lesbian cohabiting couples really different from heterosexual married couples? Journal of Marriage and Family, 66, 880-900.

21.

LaSala

M. C.

(2000). Gay male couples: The importance to coming out and being out to parents. Journal of Homosexuality, 39, 47-71.

22.

Lewis

Derlega

Berndt

Morris

Rose

(2001). An empirical analysis of stressors for gay men and lesbians. Journal of Homosexuality, 42, 63-88.

23.

Lewis

Kozac

Milardo

Grosnick

(1992). Commitment in same-sex love relationships. In Dynes

Donaldson

(Eds.), Homosexuality and psychology, psychiatry, and counseling: Studies in homosexuality (pp. 174-194). New York, NY: Garland Press.

24.

Lannutti

(2005). For better for worse: Exploring the meanings of same-sex marriage within the lesbian, gay, bisexual, and transgendered community. Journal of Social and Personal Relationships, 22, 5-18.

25.

Lavee

Olson

D. H.

(1991). Family types and response to stress. Journal of Marriage and the Family, 53, 786-798.

26.

Lynch

J. M.

(2000). Considerations of family structure and gender composition: The lesbian and gay stepfamily. Journal of Homosexuality, 40, 81-95.

27.

McQueeney

(2003) The new religious rite: A symbolic interactionist case study of lesbian commitment rituals. Journal of Lesbian Studies, 7, 49-70.

28.

Mohr

Fassinger

(2003). Measuring dimensions of lesbian and gay male experience. Measurement and Evaluation in Counseling and Development, 33, 66-90.

29.

Mohr

Fassinger

(2006). Sexual orientation identity and romantic relationship quality in same-sex couples. Personality and Social Psychology Bulletin, 32, 1085-1099.

30.

Murray

Fisher

(2002). The Internet: A virtually untapped tool for research. Journal of Technology in Human Services, 19(2/3), 5-18.

31.

Olson

(1986). Circumplex model VII: Validation studies and FACES-III. Family Process, 25, 337-351.

32.

Olson

Gorall

(2003). Circumplex model of marital and family systems. In Walsh

(Ed.), Normal family processes (3rd ed., pp. 514-547). New York, NY: Guilford Press.

33.

Oswald

(2000). A member of the wedding? Heterosexism and family ritual. Journal of Social and Personal Relationships, 17, 107-131.

34.

Oswald

(2002). Inclusion and belonging in the family rituals of gay and lesbian people. Journal of Family Psychology, 16, 428-436.

35.

Pew Internet Life Project. (2007). Demographics of Internet users. Retrieved from http://www.pewInternet.org/trends/User_Demo_6.15.07.htm

36.

Pinel

(1999). Stigma consciousness: The psychological legacy of social stereotypes. Journal of Personality and Social Psychology, 76, 114-128.

37.

Reips

(2002). Internet-based psychological experimenting: Five dos and don’ts. Social Science Computer Review, 20, 241-249.

38.

Rhodes

DiClemente

Cecil

Hergenrather

Yee

(2002). Risk among men who have sex with men in the United States: A comparison of an Internet and a conventional outreach sample. AIDS Education and Prevention, 14, 41-50.

39.

Ruiz

S. A.

Silverstein

(2007). Relationships with grandparents and the emotional well-being of late adolescent and young adult grandchildren. Journal of Social Issues, 63, 793-808.

40.

Rutter

Schwartz

(1996). Same-sex couples: Courtship, commitment, context. In Auhagen

von Salisch

(Eds.), The diversity of human relationship (pp. 197-226). Cambridge, England: Press Syndicate of the University of Cambridge.

41.

Sabatelli

(1984). The marital comparison level index: A measure for assessing outcomes relative to expectations. Journal of Marriage and the Family, 46, 651-662.

42.

Savin-Williams

R. C.

(2001). “Mom, Dad. I’m gay.” How families negotiate coming out. Washington, DC: American Psychological Association.

43.

Sobel

(1982). Asymptotic confidence intervals for indirect effects in structural equation models. In Leinhart

(Ed.), Sociological methodology 1982 (pp. 290-312). San Francisco, CA: Jossey-Bass.

44.

Solomon

Rothblum

Balsam

(2004). Pioneers in partnership: Lesbian and gay male couples in civil unions compared with those not in civil unions and married heterosexual siblings. Journal of Family Psychology, 18, 275-286.

45.

Stets

J. E.

(1993). The link between past and present intimate relationships. Journal of Family Issues, 14, 236-260.

46.

U.S. Department of Commerce. (2004, September). A nation online: Entering the broadband age. Retrieved from http://www.ntia.doc.gov/reports/anol/NationOnlineBroadband04.htm

47.

van Eeden-Moorefield

(2005). Links between the statuses of gay men in relationships and couple identity: A theoretical extension and examination (Unpublished doctoral dissertation). University of North Carolina at Greensboro, Greensboro, NC.

48.

van Eeden-Moorefield

Proulx

Pasley

(2008). A comparison of Internet and face-to-face qualitative methods in studying gay couples. Journal of GLBT Family Studies, 2, 181-204.