Minority Stress and Relationship Well-Being in Sexual Minorities: The Varying Role of Outness on Relationship and Sexual Satisfaction

Abstract

Objectives: Empirical inquiry has targeted how same-sex couples grapple with unique stressors in their romantic relationships. Meta-analyses demonstrate that the adverse link between minority stress and sexual and relationship well-being is contingent on the type of stressor and relationship component. Methods: We conducted a cross sectional study examining a sample of 238 individuals currently in same-sex relationships. Results: Outness predicted relationship satisfaction, sexual satisfaction, and sexual dissatisfaction relative to other stressors, and depressive symptomatology mediated the minority stressors and relationship well-being associations. Conclusions: The current findings offer practical applications for practitioners working with sexual minority clients and researchers who continue to reveal the boundaries in the minority stress process.

Given the recent legalization of marriage in the United States, the opportunities for sexual minorities, or individuals in same-sex romantic relationships have dramatically evolved and captured the interest of researchers and professionals. Specifically, researchers have explored how different stressors particular to this group predict numerous aspects of romantic relationships (Cao et al., 2017). Minority stressors are germane to marginalized social status groups and negatively correlate with mental and physical health outcomes (Hatzenbuehler, 2009; Lick et al., 2013; Meyer, 2003). Indeed, sexual minorities report lower levels of good health and greater levels of depressive symptomatology compared to straight people and are designated as a health disparity group (Cochran & Mays, 2009; Meyer, 2003; National Institute of Health, 2016; Yarns et al., 2016). Support from romantic partners may buffer health discrepancies, as relationship satisfaction is positively linked to optimal health outcomes (Bookwala & Gaugler, 2018; Rosenthal et al., 2019). Though empirical attention has been directed toward stress in same-sex couples, inconsistencies and gaps in the literature persist. In particular, the research questions of the present paper are to examine how visibility of relationship status predicts relationship satisfaction and sexual evaluations, as well as identify whether depressive symptomatology mediates these exchanges.

Minority stress theory

According to Minority Stress Theory, members of marginalized groups experience a surplus of additive stress due to societal systems of inequality (Meyer, 2003). Minority stressors are shared antecedents of stress and health that target only those belonging to a marginalized group. The salience among the stressors rests on the distance between the stressor and the self. Proximal stressors come from within an individual, such as internalized homonegativity (IH) and outness, whereas distal stressors come from external sources, such as experienced discrimination. Being out implies that a person actively discloses and makes her/his/their sexual minority status and/or same-sex relationship known in public, whereas passing refers to the ability and/or choice to hide one’s sexual minority status and be mistook as straight, a phenomenon that is stressful for both an individual and a relationship (Pachankis et al., 2020). Internalized homonegativity (i.e., internalized homophobia, internalized heterosexism) is an internal negative attitude about non-heteronormative individuals that also has been associated with several negative health outcomes (Berg et al., 2016; Szymanski et al., 2008). Lastly, experienced discrimination is unfair treatment, adverse attitudes or judgements, and exposure to negative events due to being a sexual minority (Herek, 2009).

Minority stress and relationship satisfaction

Over the last 20 years, researchers have continually outlined whether outness, IH, and/or experienced discrimination predict various aspects of romantic evaluations. Two meta-analyses (Cao et al., 2017; Doyle & Molix, 2015) and two reviews (Lavner, 2017; Rostosky & Riggle, 2017) summarize the inconsistencies in the stigma and relationship quality association. Generally, the results of the two meta-analyses conclude that relationship satisfaction, commitment, and intimacy are hampered by the differential relationship to minority stressors, with the condition that not all stressors are equally sound predictors. More specifically, the authors found evidence that IH and discrimination reliably predict different features of relationship quality; however, they diverge in their consensus on outness. Similarly, Cao et al. (2017) reviewed 13 research reports and 51 effect sizes on outness and concluded a trivial association with relationship quality. Trivial findings possibly are due to having fewer research reports in addition to methodological limitations; in fact, other commentaries highlight how findings regarding outness are inconsistent, understudied, and challenging to measure (Knoble & Linville, 2012; Lavner, 2017; Pachankis et al., 2020; Rostosky & Riggle, 2017).

Passing may provide protection from becoming a target of discrimination and/or stigmatization and is not always perceived as harmful. Yet, qualitative accounts detail that being out and having an equally out partner are positive for relationships because both nurture a sense of affirmation and validation as a couple (Knoble & Linville, 2012). Quantitative examinations are ambiguous as there are a host of research reports (Balsam et al., 2008; Berger, 1990; Caron & Ulin, 1997; Guschlbauer et al., 2019; Jordan & Deluty, 2000; Reeves & Horne, 2009) supporting that outness predicts relationship satisfaction, and a litany of reports (Balsam & Szymanski, 2005; Beals & Peplau, 2001; Dispenza, 2015; Mohr & Daly, 2008; Oetjen & Rothblum, 2000; Šević et al., 2016; Todosijevic et al., 2005; Tornello et al., 2013) that fail to establish this link. Adding to the complexity of this issue is that in some cases, outness predicted only particular aspects of relationship well-being (Ackbar & Senn, 2010). Similarly, outness subscales predict satisfaction differently between self and partner ratings in dyadic studies (Clausell & Roisman, 2009; Totenhagen et al., 2018). In a recent meta-analysis, a small positive correlation was found between outness and internalizing mental health issues, such as depression (Pachankis et al., 2020). This correlation was conditional on the type of measurement being used, year of study and age of sample. Exhaustively, the relationship between outness and relationship satisfaction indicators is mixed at best.

One way to interpret the inconsistency of the predictive utility of outness is in the way it is measured. When researchers examine outness, they tend to use inventory methods that calculate specific instances of disclosure to a predetermined list of individuals (Mohr & Fassinger, 2000); however, these measures are static, ignore other concealing behaviors outside of verbal disclosure, and assume that each target is weighted the same. The process of coming out of the closet is lifelong and requires continual balance between disclosure and inhibition; therefore, it is important to examine it by using non-inventory methods that assess more than the number of people with whom one is out and/or the degree of that outness to this limited list. One study that included both inventory and non-inventory approaches found that the best predictor of relationship quality was a non-inventory question aimed at disclosure of one’s relationship (Caron & Ulin, 1997). Non-inventory methods allow for a more nuanced understanding of the complexities of both disclosure and concealment. For example, one may be out to their parents, but still avoid showing affection to a partner in public. Recent evidence using non-inventory approaches found strong support that outness predicted relationship satisfaction (Guschlbauer et al., 2019).

Minority stress and sexual satisfaction and dissatisfaction

One area of relationship fulfillment that has been neglected in the literature and rarely studied in conjunction with minority stressors, is sexual satisfaction and dissatisfaction (Peplau & Fingerhut, 2007). There is a cultural expectation that romantic couples have desirable sexual relations with sex being a vital feature of relationship well-being (Maxwell & McNulty, 2019). Although general daily stressors and depression predict lower sexual satisfaction in all couples, less is known about the relationships between sexual satisfaction and minority stressors in same-sex couples (Peplau & Fingerhut, 2007; Peleg-Sagy & Shahar, 2013; Syme et al., 2013; Totenhagen et al., 2012). This lack of inquiry is surprising considering stigmas about sexual minorities are inherently sexual. A few findings suggest that minority stressors and sexual functioning are related, but these conclusions are tentative at best (Henderson et al., 2009; Kuyper & Vanwesenbeeck, 2011; Meyer & Dean, 1998). More specifically, although some findings demonstrate that higher levels of IH predict higher sexual dysfunction, worse sexual health behaviors, and lower sexual satisfaction (Henderson et al., 2009; Kuyper & Vanwesenbeeck, 2011; Meyer & Dean, 1998), other findings suggest no relationship between IH and sexual well-being (Cohen & Byers, 2015; Šević et al., 2016). Similarly, experienced discrimination predicted sexual problems in one report, but the link was not supported elsewhere (Zamboni & Crawford, 2007). Of particular interest to the present study is whether outness relates to sexual outcomes; however, this construct has received less attention, with only one attempt failing to connect it with sexual satisfaction and dysfunction (Kuyper & Vanwesenbeeck, 2011). We contend that this lack of substantiated relationship between outcomes and sexual outcomes is related to the use of inventory methods, rather than its lack of importance. Across these findings, researchers have conceptualized sexual well-being differently and often include either a positive (e.g., sexual satisfaction) or negative (e.g., sexual dysfunction, sexual problems) measure of sexual behavior. Therefore, it is necessary to examine the non-mutually exclusive aspects of sexual satisfaction and dissatisfaction independently to shed more light on these novel associations (Shaw & Rogge, 2016).

Mediating role depressive symptomology

Researchers in this area postulate a need to identify the mechanisms of the minority stressor and relationship quality exchange (Lavner, 2017; Rostosky & Riggle, 2017). One such mechanism that has been examined in sexual minorities is depressive symptomatology. Although there is some evidence that depressive symptomatology partially mediates the relationship between IH and relationship quality, this mechanism has yet to be examined with either outness and experienced discrimination (Frost & Meyer, 2009). Depressive symptomatology demonstrates a robust link between minority stressors and lower mental health outcomes (Meyer, 2003; Newcomb & Mustanski, 2010; Pachankis et al., 2020; Szymanski et al., 2008). Additionally, there is a well-established association between depressive symptomatology and poorer relationship quality, including sexual satisfaction, in mixed-sex couples (Mackinnon et al., 2012; Whitton, & Whisman, 2010). The robust findings between depressive symptoms and relationship satisfaction in couples do not consider other antecedents of depression, such as minority stressors (Mackinnon et al., 2012; Whitton & Whisman, 2010). According to minority stress theory, internalized mental health symptoms are direct consequences of minority stressors and the links between the minority stressors and relationship quality should be explained in part by these mental health symptoms.

Current study

The purpose of this study is to address three current gaps in how minority stressors predict same-sex relationship quality. First, exploring how a non-inventory approach to outness uniquely predicts relationship quality separate from IH and experienced discrimination. Second, examining how sexual satisfaction and dissatisfaction are connected to minority stressors. Finally, assessing the mediating role of depressive symptomatology between minority stressors and relationship and sexual satisfaction. Using a sample of individuals in same-sex relationships, we specifically hypothesize the following:

Hypothesis 1: Outness will positively predict relationship and sexual satisfaction and negatively predict sexual dissatisfaction.

Hypothesis 2: Internalized homonegativity and experienced discrimination will negatively predict relationship and sexual satisfaction and positively predict sexual dissatisfaction.

Hypothesis 3: Depressive symptomatology will mediate the relationship between each independent minority stressor and relationship outcomes.

Method

Participants

Data were collected on 238 individuals who had been in a same-sex relationship¹ for at least one year. The exclusionary criterion for participation was that only one partner of a dyad was permitted to participate to avoid multicollinearity of couples within the sample. Demographic information about the sample can be found in Table 1. Generally, the individuals represented the lifespan, were primarily white, about equal in gender representation, primarily identified as being gay or lesbian, highly educated, varied in relationship duration, lived with their partner/spouse, and did not have children.

Table 1.

Demographics of individuals in same-sex relationships.

Variable	Total
Age, range (M)	18–80 (40.93)
Race, %
Black	3.40
Non-Latinx White	87.60
Latinx	2.60
Asian	1.30
Mixed Race	5.10
Gender, %
Woman	50.80
Man	43.70
Gender Diverse	5.50
Sexual Orientation, %
Lesbian/Gay	80.10
Bisexual	10.60
Another Orientation	9.30
Highest Level of Education, %
High/Vocational School	4.80
Some College	18.20
College	29.90
Post College Degree	47.20
Length of Relationship, range (M)	1–56 (10.41)
Married, %	47.00
Monogamous, %	84.90
Live with Partner, %	75.90
Have Children, %	30.00

Note. The non white racial Black, Data was missing randomly for the following variables age (5), race (4), sexual orientation (2), highest level of education (7), length of relationship (3), marital status (2), living arrangement (1), and parental status (8). The data above do not reflect the missing data.

Materials

Demographics were collected using a self-report questionnaire. Participants were asked to disclose their age, race, gender, sexual orientation, highest level of education, length of relationship, marital status, monogamy status, living arrangement, and parental status.

Outness

To assess outness, or disclosure of relationship status, participants completed a slightly modified version of the Lesbian, Gay, Bisexual Visibility Management Scale (LGB-VMS; Lasser et al., 2010). The responses of the 28-item non-inventory LGB-VMS questionnaire ranged from 1 (Strongly Disagree) to 6 (Strongly Agree) with higher scores representing greater outness. Examples of the phrasing of items were minimally changed such that orientation labels were replaced to be more inclusionary of an individual’s relationship (e.g., “I do my best to let most people know that I’m gay/lesbian/bisexual” was altered to “I do my best to let most people know that I’m in a same-sex relationship”). The LGB-VMS held good reliability with a Cronbach’s alpha of .95.

Internalized homonegativity

The Internalized Homophobia Scale (IHP; Herek et al., 1998) was used to measure IH. The scale is comprised of 9 items on a 6-point scale, ranging from 1 (Strongly Disagree) to 6 (Strongly Agree), with higher scores indicative of higher IH. Questions were minimally modified from their original content so that they did not specify a particular sex/gender/orientation. For example, a question such as, “I feel alienated from myself because of being lesbian/bisexual” was altered to “I feel alienated from myself because of being attracted to a person of the same sex.” The IHP demonstrated sound reliability with a Cronbach’s alpha of .86.

Experienced discrimination

The Everyday Discrimination Scale (EDS; Williams et al., 1997) was adapted to measure experienced discrimination at the couple level. The EDS historically is used to measure racial discrimination, so the phrasing of targets was altered to apply to an individual’s relationship(s). The 9 questions were changed so that the phrase “because of your race” was changed to “because of being in a same-sex relationship.” Similar adaptations of measures have been used in sexual minority research (Rosenthal et al., 2019). The EDS was measured from 1 (Never) to 6 (Always) and demonstrated excellent reliability (α = .91). Higher scores indicated more experienced discrimination.

Relationship satisfaction

To assess quality in their relationship, participants completed the Gay and Lesbian Relationship Satisfaction Scale (GLRSS; Belous & Wampler, 2016). The GLRSS is a validated measure of relationship satisfaction intentionally created for same-sex couples. It is comprised of 16 Likert-type scale items on a scale of 1 (Strongly Disagree) to 6 (Strongly Agree), with higher scores suggesting more satisfied relationships. An example of an item is, “My mate has the qualities I want in a partner.” The GLRSS demonstrated sound reliability with a Cronbach’s alpha of .81.

Sexual satisfaction & dissatisfaction

To assess both sexual satisfaction and dissatisfaction we used the Quality of Sex Inventory (QSI; Shaw & Rogge, 2016). The QSI is a general measure of sexual satisfaction (QSI-S) and dissatisfaction (QSI-D) on a 6-point response scale from 1 (Not at All True) to 6 (Completely True). Each subscale is comprised of 12 items, and higher scores represent higher respective evaluations. An example item of sexual satisfaction is “My sex life is fulfilling,” which contrasts in directionality of the dissatisfaction subscale, “Sexual activity with my partner is not fun.” The QSI subscales have been used with people in same-sex relationships (Shaw & Rogge, 2016) and the satisfaction subscale held better reliability compared to its counterpart (QSI-S, $\alpha$ = .97, QSI-D, $\alpha$ = .72).

Depressive symptomatology

Depressive symptomatology was measured using The Center for Epidemiologic Studies Depression Scale (CES-D; Radloff, 1977). To better account for missing items and aid interpretation, the 20 items were scaled from 1 (Rarely or None of the Time, Less than One Day) to 4 (Most or All of the Time, 5–7 days) and assessed the mean across all items. An example of an item is, “I felt everything I did was an effort.” The CES-D recorded great reliability (α = .93).

Procedure

After receiving Institutional Review Board approval at the University of Akron, data were collected online via computer or smartphone using Qualtrics, a survey generator. Participants were recruited primarily through the use of social media, in addition to online groups or forums specific to LGB persons, word of mouth, and snowball sampling. Several previous researchers have been successful in attaining large sexual minority samples using online assessments (Riggle et al., 2005). After giving their informed consent, participants were required to answer the following question that identified whether they were eligible for the study: “I have been with a current same-sex partner for longer than a year who has not also participated in this research project”. If they did not meet the requirements of the inclusion criteria, they were filtered out of the survey. After giving their consent and completing the inclusion criterion question, they completed the demographic questionnaire and the battery of self-report assessments, which took about 20–25 min to complete. Participants were provided the chance to win a $50 gift card upon completion of the study.

Results

Analytic plan

Prior to testing hypotheses, the validity of the modified LGB-VMS, missing data, normality assumptions, outliers, and descriptive statistics were addressed. The first set of hypotheses used three hierarchical multiple regressions to test whether outness predicts each relationship evaluation on its own and relative to the other stressors. Each regression was comprised of three steps and predictors were entered progressively to identify their unique variance. Step 1 controlled for covariates, step 2 examined the unique predictability of outness (Hypothesis 1), and in step 3 we entered IH and experienced discrimination to examine both the predictive power of outness relative to the other stressors, as well as how the other stressors uniquely predict relationship well-being (Hypothesis 2). The third hypothesis was tested with Hayes’s (2013) PROCESS macro to test whether depressive symptomatology mediated the relations between each stressor and each outcome with a 10,000 bootstrapped distribution. The presence of both a significant indirect and direct effect confirms partial mediation. Significance of these effects are determined when confidence intervals (CI) do not include zero. Although directional claims inferred by mediation cannot be supported due to the cross-sectional nature of the study, the current use of mediation still provides insight into the theoretically derived pathways proposed by Minority Stress Theory (Hayes, 2013; Meyer, 2003).

Validity of the modified LGB-VMS

An exploratory factor analysis (EFA) tested whether the items in the modified version of the LGB-VMS loaded in the same three-factor solution as the original measure. The EFA was tested with oblique direct oblimin rotation and explained 57.64% of the variance in the LGB-VMS. Bartlett’s test of sphericity, χ² (378) = 2,219.89, p < .001, suggested that the factors were intercorrelated and the Kaiser–Meyer–Olkin metric of sampling adequacy was .91, indicating adequate sample size. The factor structure was identical to the original with one item cross loading and one item loading on a different factor (Lasser et al., 2010; factor loadings available in the online supplementary materials). To test for convergent validity, the LGB-VMS was correlated with the single-item measure of outness (i.e., “Do you consider yourself to be ‘out of the closet?’” with response options from 1 (Never) to 5 (Always)) and both were correlated with a known correlate of outness, IH. The single item outness measure and the LGB-VMS strongly correlated (r = .66, p < .05) and the latter held a medium association with the IHP (r = −.44, p < .01), which is more consistent with prior research compared to the single item measure (r = −.32, p < .01; Jackson & Mohr, 2016).

Missing data and preliminary screening

Initially, ineligible participants were screened out of the survey, then we examined for missing data rates and attrition. Rates of missingingness for the demographics were as high as 3% (See Table 1). Rates of missing data for the predictors and criterion variables progressively increased with the order of our surveys indicating survey attrition (See Table 2). The length of the study, social media environment, and method of taking the survey (e.g., smartphone, tablet) are likely suspects that explain the attrition rate. There were no significant differences between participants who finished the survey and those who stopped at any point across any demographic or construct variables. The available data were missing at random, noted by a non-significant Little’s test; therefore, we used Parent’s (2013) suggestion of available item analysis (AIA) to handle missing data by making a mean substitution to compute a mean score for every variable relative to every subject’s scores on other items of the measure.

Table 2.

Correlations and descriptive statistics.

Measure	1	2	3	4	5	6	7	8	9
1. LGB-VMS	—	−.44**	−.19*	.39**	.22**	−.26**	−.29**	.20**	.14*
2. IHP		—	.26**	−.28**	−.18*	.25**	.37**	−.17	.05
3. EDS			—	−.07	.10	−.09	.37**	−.08	−.02
4. GLRSS				—	.71**	−.68**	−.33**	.12	−.01
5. QSI-S					—	−.82**	−.20**	−.12	−.16*
6. QSI-D						—	.26**	.03	.12
7. CES-D							—	−.17*	−.07
8. Age								—	.73**
9. LOR									—
M (SD)	4.31 (.75)	1.50 (.69)	2.19 (.86)	4.38 (.65)	4.28 (1.28)	1.44 (.79)	1.61 (.54)	40.93 (15.0)	10.41 (10.1)
N	216	201	186	191	200	198	172	233	235

Note. Transformed data were used on the IHS and QSI-D correlation coefficients; however, we reported the raw means and standard deviations for ease of interpretation. Higher scores designate increased responding in that construct across all variables. LGB-VMS: Lesbian, Gay, Bisexual Visibility Management Scale; IHP: Internalized Homophobia Scale; EDS: Everyday Discrimination Scale; GLRSS: Gay Lesbian Relationship Satisfaction Scale; QSI-S: Quality of Sexual Satisfaction subscale; QSI-D: Quality of Sexual Dissatisfaction subscale; CES-D: Center for Epidemiologic Studies Depression Scale; LOR: Length of Relationship.

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

We screened the data for skewness, kurtosis, and multicollinearity. Skewness and kurtosis were explored using histograms and p-plots. The IHP and QSI-D were severely positively skewed and leptokurtic, which were addressed with an inverse transformation.² Violations of multicollinearity were not suspected across predictors as the only correlation above .80 was between sexual satisfaction and dissatisfaction, which were analyzed independently. Using a z-score distribution for each score, we found one univariate outlier (z > 3.29) on the LGB-VMS, one on the GLRSS, and one on the CES-D. All scores were within range and were vetted further, using Mahalanobis distance and no multivariate outliers were found.

Descriptive statistics

Table 2 provides the breakdown of the means, standard deviations, sample sizes, and intercorrelations among the variables. The sample had low minority stress as indicated by high outness, low IH, and low experienced discrimination. The sample also was content with their sexual and relationship satisfaction. Both outness and IH were significant predictors of the three relationship well-being variables and depressive symptomatology correlated with all three minority stressors and with the three relationship variables.

Control variables

Given that gender has moderated the minority stressor and relationship well-being exchange, it was held as a control variable (Guschlbauer et al., 2019). All other demographics were inspected and kept as controls if they differed across any of the criterion variables. Age was correlated with all three relationship variables (See Table 2). Married individuals had higher relationship satisfaction (t(189) = −2.57, p < .05, d = .37) compared to those who were non-married (e.g., dating). Those who were non-monogamous had lower relationship and sexual satisfaction (t(189) = 2.70, p < .01, d = .51; t(198) = 2.42, p < .05, d = .46, respectively). Lastly, parents had higher relationship satisfaction and sexual satisfaction and lower sexual dissatisfaction (t(189) = 2.58, p < .05, d = .44; t(198) = 2.52, p < .05, d =.40; t(139.76)³ = 3.18, p < .01, d = .47, respectively). Therefore, age, gender, marital status, monogamy status, and living arrangement were considered covariates. Gender was dummy coded, such that women were a reference group compared independently to both men and gender diverse people.

Analyses

Hypotheses 1 and 2: Minority Stressors Predicting Relationship Well-being. Across the three different hierarchical regressions run on our three outcomes, specific coefficients for the minority stressors in steps two and three are presented in Table 3 and coefficients for the control variables are in the supplementary materials.

Table 3.

Hierarchical regressions predicting minority stressors on relationship satisfaction, sexual satisfaction, and sexual dissatisfaction.

	GLRSS						QSI-S						QSI-D
	Step 2			Step 3			Step 2			Step 3			Step 2			Step 3
Predictor	B	SE B	β	B	SE B	β	B	SE B	β	B	SE B	β	B	SEB	β	B	SE B	β
Constant	3.41*	.34		3.41*	.38		4.60*	.67		3.76*	.76		.71*	.12		.56*	.14
LGB-VMS	.32*	.06	.36	.27*	.07	.30	.47*	.12	.27	.42*	.13	.24	−.09*	.02	−28	−.07*	.02	−.24
IHP				−.39†	.23	.13				−.78†	.44	−14				.17*	.08	.18
EDS				<−.01	.06	<−.01				.24*	.11	.16				−.04*	.02	−.17
ΔR2	.12*			.01			.07*			.03*			.08*			.04*

Note. Transformed data were used on the IHS and QSI-D coefficients, and the signs of the variables were altered to reflect their conceptual relationships. Higher scores designate increased responding in that construct across all variables. LGB-VMS: Lesbian, Gay, Bisexual Visibility Management Scale; IHP: Internalized Homophobia Scale; EDS: Everyday Discrimination Scale; GLRSS: Gay Lesbian Relationship Satisfaction Scale; QSI-S: Quality of Sexual Satisfaction subscale; QSI-D: Quality of Sexual Dissatisfaction subscale.

*p < .05. †p < .10.

Relationship satisfaction

The controls in the first regression predicted 13% of the variance in relationship satisfaction. The addition of outness improved the model significantly and accounted for an additional 12% of variance in relationship satisfaction (ΔR² = .12, F(1, 175) = 28.39, p < .01). Outness maintained its significance with the addition of IH and experienced discrimination, yet this inclusion did not significantly improve the model (ΔR² = .01, F(2, 173) = 1.57, p = .21), nor was either variable a significant predictor.

Sexual satisfaction

Using sexual satisfaction as an outcome, 14% of variance was predicted by the control variables. Step 2 accounted for an additional 7% of the variance, which significantly improved the model (ΔR² = .07, F(1, 175) = 15.41, p < .01). The final step improved the model and accounted for an additional 3% of sexual satisfaction’s variance (ΔR² = .03, F(2, 173) = 3.78, p < .05). Divergent from the pattern between minority stressors and relationship satisfaction, both outness and experienced discrimination, but not IH, were significant predictors of sexual satisfaction.

Sexual dissatisfaction

In the final regression model predicting sexual dissatisfaction, the first step predicted 8% of the variance. The following step, using outness as a predictor, accounted for an additional 8% of the variance of sexual dissatisfaction (ΔR² = .08, F(1, 175) = 15.58, p < .01). In the final model, an added 4% of the variance of sexual dissatisfaction, was explained with the addition of IH and experienced discrimination, which were both significant predictors (ΔR² = .04, F (2, 173) = 3.97, p < .05).

Hypothesis 3: mediating role of depressive symptomatology

The mediating role of depressive symptomatology in the stressor and relationship well-being association was tested with three different mediation models for each relationship outcome totaling nine mediations.⁴ Figure 1 outlines the findings for the stressors individually; however, in each mediation one minority stressor is the independent predictor and the other two stressors, the male dummy code, and monogamy status are controlled.⁵ The a path, in which the stressors and controls predict depressive symptomatology, are statistically identical across all models. The b paths, in which depressive symptomatology predicts each relationship factor above each stressor and control, are the same within each outcome. However, the calculation of the direct and indirect effects is unique to each predictor and is presented for each of the nine models.

Figure 1.

Meditational Models between Minority Stressors and Relationship Well-Being through Depressive Symptomatology. Note. All models controlled for gender, monogamy status, and other respective stressors. The following panels outline the nine different mediation models. All models were run in the PROCESS (Hayes, 2013) macro. Panels A-C represent the models examining relationship satisfaction with outness, IH, and experienced discrimination as the antecedents, respectively. Panels D-F represent the models examining sexual satisfaction for each stressor, respectively. Lastly, panels G-I represent the models examining sexual dissatisfaction for each stressor, respectively. The direct effect is denoted by C and the indirect effect is denoted by C’. The sample size for all mediations is 172. *p = significant effect indicating that 0 was not in the 95% CI.

The a paths, predicting each stressor on depressive symptomatology, were significant for only IH (B = −.54, SE = .18, 95% CI [−.91, −.18]) and experienced discrimination (B = .18, se = .04, 95% CI [.10, .27]). Outness approached significance (B = −.10, SE = .06, 95% CI [−.21, .004]). The b paths of depressive symptomatology predicting relationship satisfaction were significant (B = −.26, SE = .09, 95% CI [−.44, −.08]). Additionally, outness also predicted relationship satisfaction (B = .27, SE = .07, 95% CI [.14, .40]). Hypotheses were confirmed that depressive symptomatology partially mediated the association of outness on relationship satisfaction (See Panel A in Figure1) due to a significant indirect effect (B = .03, SE = .02, 95% CI [.001, .08]) and direct effect (B = .27, SE = .07, 95% CI [.14, .40]). The link of IH on relationship satisfaction supported full mediation (See Panel B in Figure 1) indicated by a significant indirect effect (B = .14, SE = .08, 95% CI [.03, .34]), and non-significant direct effect (B = .24, SE = .22, 95% CI [−.20, .67]). The same pattern was found that depressive symptomatology fully mediated the link between experienced discrimination and relationship satisfaction (See Panel C in Figure 1) because the indirect effect was significant (B = −.05, SE = .03, 95% CI [−.11, −.01]) and the direct effect was non-significant (B = .04, se = .05, 95% CI [−.07, .15]).

The b paths for the second outcome, depressive symptomatology predicting sexual satisfaction, were significant (B = −.38, SE = .19, 95% CI [−.76, −.01]). Outness and experienced discrimination also significantly predicted sexual satisfaction (B = .35, SE = .14, 95% CI [.08, .62]; B = .24, SE = .11, 95% CI [.01, .46], respectively). Depressive symptomatology did not mediate the links between either outness or experienced discrimination (See Panels D and F in Figure 1) as both indirect effects were non-significant (B = .04, SE = .03, 95% CI [−.002, .14], B = −.07, SE = .05, 95% CI [−.19, .001], respectively). However, there was support for full mediation between IH and sexual satisfaction (See Panel E in Figure 1), such that the indirect effect, was significant (B = .21, SE = .14, 95% CI [.001, .57]) and the direct effect (B = .34, SE = .47, 95% CI [−.58, 1.25]) was non-significant.

The last groups of b paths for sexual dissatisfaction were significant (B = −.09, SE = .01, 95% CI [−.15, −.02]). Experienced discrimination and outness again were significant predictors of sexual dissatisfaction (B = .05, SE = .01, 95% CI [.01, .09]; B = .07, SE = .02, 95% CI [.03, .12], respectively). Depressive symptomatology did partially mediate the relationship between outness and sexual dissatisfaction (See Panel G in Figure 1) because both the indirect effect (B = .01, SE = .01, 95% CI [.001, .03]), and the direct effect (B = .07, SE = .02, 95% CI [.03, .12]) were significant. Full mediation was evident for the relationship between IH and sexual dissatisfaction (See Panel H in Figure 1), which held a significant indirect effect (B = .05, SE = .03, 95% CI [.01, .11]) and non-significant direct effect (B = .09, SE = .08, 95% CI [−.06, .24]). Finally, both the indirect and direct paths were significant indicating partial mediation (See Panel I in Figure 1) for the relationship between experienced discrimination and sexual dissatisfaction (B = −.02, SE = .01, 95% CI [−.04, −.003]; B = .05, SE = .02, 95% CI [.01, .07], respectively).

Discussion

These findings provide three new insights to the minority stress-relationship literature in same-sex couples: (1) Outness is a predictor of relationship satisfaction, (2) minority stressors, including outness, also predict sexual evaluations, and (3) depressive symptomatology mediates these relationships. At the bivariate level, outness and IH, but not experienced discrimination, were significant predictors of relationship satisfaction supporting claims from Doyle and Molix (2015) that proximal stressors have more resonance compared to distal stressors. When the stressors were examined concurrently, the role of IH was minimal as it solely related to sexual dissatisfaction. Interestingly, experienced discrimination yielded the opposite pattern with the sexual evaluations, as it gained significance. This change in pattern highlights the necessity of examining the stressors simultaneously (Frost & Meyer, 2009). Measuring a single stressor endangers accuracy if variance among multiple stressors overlap. Outness retained its significance across all three outcomes, even when analyzed with established minority stress predictors. Premature claims about the relation between outness and relationship satisfaction made in meta-analyses minimize the importance of the construct without properly disentangling its nuances. Specifically, given the inconsistencies in the predictive value of outness, the current study demonstrates that the operationalization of the construct may help to delineate its importance in multiple relationship satisfaction outcomes.

The finding of outness driving the relationship beyond more historically supported stressors may be a function of the use of non-inventory over inventory measures. Recent evidence using more contextual measures of outness found that it predicted relationship quality; however, this examination was independent of other stressors and focused on individual-level disclosure (Guschlbauer et al., 2019). These more contextual methods allow for the proliferation of outness by capturing the hazards of decision-making regarding relationship status visibility as opposed to the stress of interpersonal disclosure. Nevertheless, there is still utility for inventory measures when needing to specify outness with specific relationships. For example, disclosure to family vs. friends has been found to distinctly predict well-being (de Miguel et al., 2018). Integrating different methodologies to assess outness would provide the best insight into the stresses and processes of being out. Outness of a relationship may act as a source of validation and/or an act of investment for a couple. The present data support that this sense of validation is important and that it overcomes insidious self-stigma.

The results regarding the oft-overlooked links with sexual evaluations can be integrated into models such as The Dual-Process Sexual-Evaluation Model, where sexual evaluation is proposed as a mechanism and/or moderator of relationship quality (Maxwell & McNulty, 2019). The findings are novel in that they emphasize sexual evaluation rather than performance, frequency, or dysfunction thus providing unique implications. The individual effect sizes of each stressor were stronger in their association for dissatisfaction; however, roughly the same amount of variance was predicted across both types of evaluation. Internalized homonegativity was found to be a stronger predictor of sexual dissatisfaction than satisfaction. The next step in this intriguing pattern will be to replicate with relationship dissatisfaction. The focus on relationship satisfaction alone within the present study may have been shortsighted, as low scores could represent ambivalence rather than true dissatisfaction. Different relationship outcomes should also be explored, specifically with how they relate to downstream health outcomes in the minority stress process.

Depressive symptomatology and relationship and sexual satisfaction were inversely related, suggesting that having more satisfaction with your relationship predicts less depressive symptomatology. The role of depressive symptomatology helped explain the associations between minority stressors and relationship well-being. Full mediation models were supported between IH and all outcomes, which replicate the model found by Frost and Meyer (2009). This previously established model was broadened also to partially explain outness and experienced discrimination. Depressive symptomatology had a stronger association with IH and experienced discrimination compared to outness and with sexual dissatisfaction compared to sexual satisfaction, as evidenced by only partially mediating the relations between outness and sexual dissatisfaction and not mediating the links with sexual satisfaction. Additionally, full mediation was found between experienced discrimination and relationship satisfaction, partial mediation was found between experienced discrimination and sexual dissatisfaction, and no mediation was found between experienced discrimination and sexual satisfaction. Although sexual satisfaction appears to be less sensitive to depressive symptomatology, mediation was significant for IH. Future work is needed to provide consistency in the interceding role of depressive symptomatology and actively theorize and test whether this role is dependent on the type of stressor and relationship evaluation.

Although this study has made innovative advances, it has several methodological limitations. The data were self-reported and correlational, thus directional links across variables and causal mediation pathways cannot be inferred. The link between outness and relationship measures could be bi-directional, such that happier couples are more visible about their relationship. Further, our sample scored low on IH and experienced discrimination and high on outness, indicating low total minority stress, which is consistent with other studies (Frost & Meyer, 2009; Totenhagen et al., 2018). In fact, given the results of multiple regressions, we had to be less demanding in the level of significance for some of the stressors, in order to proceed with examining their mediational models. Simply agreeing to participate in a study on same-sex relationships is not representative of a full sexual minority population. Beyond selection bias, this issue is compounded by a lack of sensitivity in how both IH and outness are measured, as the lived experiences of participants in qualitative research support that these are relevant stressors in their lives (Frost, 2011; Knoble & Linville, 2012). Lastly, data were collected from only one partner. Future studies should follow the steps of previous researchers (Clausell & Roisman, 2009; Mohr & Fassinger, 2006; Otis et al., 2006; Totenhagen et al., 2018) and should consider using dyadic approaches to study overlooked minority stressors and aspects of relationship well-being.

Future directions should propel research on same-sex couples by incorporating consistency in their approaches. Specifically, researchers should continue to use innovative or multimethod measures of outness, evaluate multiple stressors concomitantly, identify the boundaries of relationship outcomes, elaborate the relevance of sexual well-being, corroborate depressive symptomatology as a mediator, and explore other possible mediating factors. Future researchers should also focus on longitudinal designs and cohort effects, and more specifically, how minority stressors differ for people of different ages and how this leads to differing levels of relationship satisfaction across time. There are some short-term longitudinal studies (Mohr & Daly, 2008; Totenhagen et al., 2018) exploring these circumstances, but they do not examine long-term impact. One important question is whether outness dimensions become less salient the longer you are together. Given the way our population is aging, understanding the way older sexual minorities appraise their stress across time is imperative. A broader scope can identify what factors and circumstances are influential to specific cohorts and sociopolitical contexts (Vale & Bisconti, 2020). Additionally, intersectional approaches across multiple identities will provide more insight into how varying minority stressors interact and compound relationships. Another avenue for future researchers is to examine more practical outcomes and consequences, such as the likelihood to break up or relationship conflict. Problem interaction tasks provide an innovative lens into measuring discernible relationship quality (Clausell, & Roisman, 2009). Practitioners working with sexual minority couples will benefit from understanding the minority stress process and be able to guide and apply interventions respectively. These approaches may intervene with relationship well-being specifically or address the actual minority stress process.

There is a need to embrace contextual insight on the deleterious link of minority stressors on relationships and sexual well-being. Although current reviews and meta-analyses provide an overview of the state of the current literature and make suggestions on how to drive the field forward, they are drawing conclusions prematurely. The present findings highlight how discrepant results should be disentangled and tested as methodological and conceptual limitations may underlie inconsistencies. Understanding how and why minority stressors have downstream associations with relationships and overall well-being will provide instrumental implications for practitioners caring for this recently deemed health disparity group.

Disclosure statement

The authors report no conflict of interest.