Relationship Characteristics and Motivations behind Agreements among Gay Male Couples: Differences by Agreement Type and Couple Serostatus

Abstract

Gay men in relationships are often overlooked in HIV prevention efforts, yet many engage in sexual behaviors that increase their HIV risk and some seroconvert as a result. While different aspects of gay male relationships have been studied, such as sexual agreements, relationship characteristics, and couple serostatus, little research combines these elements to examine HIV risk for this population. The present study recruited 566 gay male couples from the San Francisco Bay Area to study their sexual agreements, motivations behind making agreements, and other relationship characteristics, such as agreement investment, relationship satisfaction, intimacy, and communication. Participants rated their level of concurrence with a set of reasons for making their agreements. They were also measured on relationship characteristics using standard instruments. Analyses were conducted by agreement type (monogamous, open, discrepant) and couple serostatus (concordant negative, concordant positive, discordant). A majority reported explicitly discussing their agreements and nearly equal numbers reported being in monogamous and open relationships. A small number (8%) reported discrepant agreements. Across all agreement type and serostatus groups, HIV prevention as a motivator for agreements fell behind every motivator oriented toward relationship-based factors. Only concordant negative couples endorsed HIV and STD prevention among their top motivators for making an agreement. Mean scores on several relationship characteristics varied significantly. Couples with monogamous agreements had higher scores on most relationship characteristics, although there was no difference in relationship satisfaction between couples with monogamous and open agreements. Scores for concordant positive couples were distinctly lower compared to concordant negative and discordant couples. Agreements, the motivations behind them, and the relationship characteristics associated with them are an important part of gay male relationships. When examined by agreement type and couple serostatus, important differences emerge that must be taken into account to improve the effectiveness of future HIV prevention efforts with gay couples.

Introduction

Epidemiological studies point to unprotected anal intercourse (UAI) with primary partners as a significant source of many HIV infections (Davidovich et al., 2001; Moreau-Gruet, Jeannin, Dubois-Arber, & Spencer, 2001; Sullivan, Salazar, Buchbinder, & Sanchez, 2009). Behavioral studies report that gay men in committed relationships engage in higher rates of UAI with their primary partners than single gay men with their casual partners, which for some could increase risk (Ekstrand, Stall, Paul, Osmond, & Coates, 1999; Elford, Bolding, Maguire, & Sherr, 1999; Hoff et al., 1997) and for others not change their level of risk at all (Jin et al. 2009). Accordingly, there is a recent and increasing recognition for the need to examine other factors operating within gay couples to help explain the high rates of HIV infection for this population (Harawa et al., 2004). Despite this, the pace of research involving gay couples and the relationship characteristics that may affect risk has fallen out of step with the pace of the epidemic. Therefore, we sought to investigate aspects of gay male relationships – sexual agreements, couple serostatus and relationship characteristics – which may contribute to understanding the dyadic context wherein HIV risk takes place.

One aspect of gay couples that may play an important role in determining HIV risk is sexual agreements (hereafter referred to as ‘agreements’). Widespread among gay couples, agreements are the decisions couples make about whether they allow sex with outside partners and the sexual behaviors they engage in together. In the 1970’s and 1980’s, studies of non-monogamous gay couples showed that they were motivated by sexual adventurism and variation (Bell & Weinberg, 1978; Blumstein & Schwartz, 1983; Kurdek, 1989). In a more recent study, Hoff and Beougher (in press) found that having agreements allowing sex with outside partners (hereafter referred to as ‘open agreements’) has several other benefits, including supporting efforts to actualize a non-heteronormative identity and providing structure to the relationship by establishing boundaries that foster a sense of primacy between partners. HIV prevention, however, was not a primary motivating factor for having agreements.

Research into agreements has focused on their use as an HIV prevention tool. Negotiated safety uses agreements to reduce infection among concordant negative couples by encouraging partners to be monogamous or to have safe sex with outside partners so they can have sex without condoms together. Initial studies of negotiated safety agreements found that they lowered HIV risk among those who had them compared to those who did not (Kippax et al., 1997; Crawford, Rodden, Kippax, & Van de Ven, 2001). Subsequent studies, however, found that the two factors required for them to be effective – knowledge that one’s partner is HIV-negative and having safe sex with outside partners – were not always present, potentially corroding their effectiveness as an HIV prevention strategy (Diaz, Morales, Bein, Dilan, & Rodriguez, 1999; Elford, Bolding, Maguire, & Sherr, 1999). In light of the inconclusive findings regarding the effectiveness of negotiated safety and recent research suggesting that HIV prevention is not a primary motivating factor for having agreements, it is imperative to delve further into agreements and their association with other characteristics of gay male relationships.

Agreements have rarely been examined with relationship characteristics – the exception being relationship satisfaction, where findings are mixed. While some report that couples with open agreements are less satisfied compared to monogamous couples (Bell & Weinberg, 1978; Saghir & Robins, 1973), others report that couples with open agreements have more favorable attitudes than their monogamous counterparts (Kurdek & Schmitt, 1986). The most consistent finding, however, was that relationship satisfaction did not differ significantly by agreement type (Blasband & Peplau, 1985; Kurdek, 1988; LaSala, 2004).

Previous research has investigated other characteristics among gay couples, such as commitment, attachment, autonomy, and equality (Kurdek, 1995; Kurdek, 1996; Kurdek, 2000), but has not often linked them to agreements nor studied them in the context of HIV. Communication skills are relevant to making agreements, but few studies have investigated a direct link between agreements such as negotiated safety and partners’ abilities to discuss complex relationship dynamics with each other (Prestage et al., 2006). It is likely that these relationship characteristics (e.g., commitment, communication) influence couples’ decisions to make and maintain agreements. Therefore, it is necessary to include these factors into investigations of agreements.

Many of the above studies were conducted prior to or outside the context of the AIDS epidemic. A critical component to take into account in current research with gay couples is the HIV status of each partner in the relationship. For instance, Hoff and colleagues (2009) found that concordant negative couples were more invested in their agreements and less likely to engage in risky sexual behaviors with outside partners than were concordant positive and discordant couples. Prestage (2008) found that over time an increasing proportion of concordant negative couples required monogamy as their agreement and an increasing proportion of discordant couples permitted UAI within their relationship. Clearly, differences emerge in sexual behaviors, relationship characteristics, and agreement type when examined by couple serostatus, yet much remains unknown about how the interplay among these factors may be associated with HIV risk.

The present study builds on the existing literature by factoring in agreement type and couple serostatus to explore couples’ motivations behind making agreements as well as their relationship characteristics. Our aim is to begin to understand the associations between agreement type, couple serostatus, motivations for making agreements, and relationship characteristics to increase the effectiveness of HIV prevention efforts for gay couples.

Methods

Recruitment

Gay couples (n = 566) were recruited from the San Francisco Bay Area between June 2005 and February 2007 using active and passive recruitment strategies. These recruitment procedures have been detailed elsewhere (Neilands et al., in press).

Measures

Agreement Type

Respondents were first asked to describe their current agreement via an open-ended response. Next, they were asked to categorize this response in the question: “Which one of the following scenarios best describes the current agreement that you and your primary partner have? (1) Both of us cannot have any sex with an outside partner, (2) We can have sex with outside partners but with some restrictions, (3) We can have sex with outside partners without any restrictions, (4) We do not have an agreement.” Respondents who chose option (4) were not asked any further questions about their agreement. Most who chose option (4) described their agreement in a way that could be clearly categorized into one of the other three options which was done by the investigators. Also, each participant’s response was compared to his partner’s to create three couple-level agreement categories: monogamous agreements where both partners chose option (1), open agreements where partners chose either option (2) or (3), discrepant agreements where one partner chose option (1) and the other partner chose either option (2) or (3). A number of other measures including standardized ones were also recorded and are documented in Table 1.

Table 1.

Measures recorded

Demographic characteristics	Age, Race/Ethnicity, Employment, Income
HIV serostatus Relationship	Both self and partner Length Status	Married/Registered as domestic partners/Had a commitment ceremony/Boyfriends or lovers/Other	Frequencies of couples with legal recognition of their relationship status (i.e., married, registered domestic partners) are reported
Use of couples therapy Agreement	Live together? Ever? In the previous six months? Whether discussed explicitly Reasons for making agreement	Items shown in Tables 3 and 4; responses recorded on 4 point scale: ‘Strongly Disagree’ to ‘Strongly Agree’	Responses were dichotomized into two broader categories – ‘Agree’ (for responses “Agree” or “Strongly Agree”) and ‘Disagree’ (for responses “Disagree” or “Strongly Disagree”).

Standardized Measure	Reference	No. of items	Response Scale	Sample Item
Sexual Agreement Investment Scale	Neilands et al., in press	13	5-point: ‘Not at All’ to ‘Extremely’	“How much do you appreciate having your current agreement?”
Dyadic Satisfaction (from Dyadic Adjustment Scale)	Spanier, 1976	10	6-point: ‘All of the Time’ to ‘Never’	“How often do you and your partner quarrel?”
Miller Social Intimacy Scale	Miller & Lefcourt, 1982	17	10-point: ‘Very Rarely’ to ‘Almost Always’ OR ‘Not Much’ to ‘A Great Deal’, depending on the question	“When you have leisure time, how often do you choose to spend it alone with your partner?”
Commitment	Sternberg, 1988	8	9-point: ‘Not at All True’ to ‘Extremely True’	“Because of my commitment to my partner, I would not let other people come between us.”
Trust	Rempel, Holmes & Zanna, 1985	17	7-point: ‘Strongly Disagree’ to ‘Strongly Agree’	“Though times may change and the future is uncertain, I know my partner will always be ready and willing to offer me strength and support.”
Equality	Kurdek, 1995	8	9-point: ‘Not at All True’ to ‘Extremely True’	“My partner shows as much affection to me as I think I show to him” and “My partner and I have equal power in the relationship.”
Autonomy	Kurdek, 1995	8	9-point: ‘Not at All True’ to ‘Extremely True’	“I make most decisions on my own.”
Attachment	Kurdek, 1995	8	9-point: ‘Not at All True’ to ‘Extremely True’	“I can never get too close to my partner.”
Communication Patterns Questionnaire:	Christensen & Shenk, 1991
Mutual Constructive Communication Subscale		7	9-point: ‘Very Unlikely’ to ‘Very Likely’	“When some problem in the relationship arises, both of us try to discuss the problem.”
Mutual Avoidance and Withholding Subscale		3	9-point: ‘Very Unlikely’ to ‘Very Likely’	“After the discussion of a relationship problem, both of us withdraw from each other.”

Note: For all scales above, higher scores represent higher levels of the characteristic under consideration. To achieve this, appropriate items within each scale were reverse scored prior to computing the composite score.

Data Analysis

Responses to several questions were appropriately categorized and descriptive statistics such as means and frequencies were calculated. Next, the percentages of respondents who agreed with each of a set of specified reasons for making their agreements were calculated. This was done by the three categories of agreement type (monogamous, open, and discrepant) and the three categories of couple serostatus (concordant negative, concordant positive, and discordant). The differences in the proportions of endorsement of each reason across the groups were tested using the Rao-Scott chi-square test of association.

The couple-level scores on the relationship measures were derived by averaging both partners’ scores on the measures. These scores were compared across the agreement types and serostatus groups using the least square means from PROC GLIMMIX. The resulting p-values from the multiple pairwise comparisons were adjusted via the simulation option and in a stepdown fashion (Westfall & Young, 1993). All models controlled for relationship length and the level of significance was set at 0.05. The five couples who reported having no agreement were omitted from the analyses because their small group size precluded obtaining reliable, generalizable inferences. All analyses were done using SAS 9.2.

Results

Ninety nine percent of couples reported having an agreement (see Table 2). Specifically, 45% had monogamous agreements, 47% had open agreements, and 8% reported discrepant agreements. In sixty four percent of couples, both partners reported discussing their agreements explicitly with their partner. Concordant negative couples comprised 55% of the sample, 22% were concordant positive, and 23% were discordant. The sample was racially and ethnically diverse, with the largest proportions being either interracial (47%) or white (45%). Individual incomes were less than $60,000 per year for a majority of participants and in most cases both partners were employed. Over 60% of couples reported being in the relationship for more than two years and 77% lived together. The average age of individuals was 41 years and the majority of couples (57%) reported age differences between partners to be six years or less. Almost 20% of couples reported being married and close to 30% were registered as domestic partners. One quarter of couples reported they had been in couples therapy at some point in their relationship and 9% said they had been in therapy in the past six months.

Table 2.

Characteristics of the Sample

	% (N)
Agreement type
Monogamous	45 (255)
Open	47 (262)
Discrepant	8 (44)
No agreement	<1 (5)
Agreement explicitly discussed by both partners	64 (364)
HIV status of couple
Concordant negative	55 (310)
Concordant positive	22 (124)
Discordant	23 (132)
Race of couple
Interracial	47 (268)
Caucasian	45 (254)
African-American	5 (26)
Hispanic/Latino	2 (11)
Asian-American/Pacific Islander	1 (5)
Native American	<1 (2)
Employment
Both partners employed	49 (279)
One partner employed	32 (180)
Both partners unemployed	19 (107)
Individual Income
Less than $30,000	45 (507)
$30,000 to $59,999	30 (343)
$60,000 or $99,999	16 (177)
$100,000 or higher	9 (105)
Length of relationship
6 months or less	9 (51)
More than 6 months and up to 2 years	28 (156)
More than 2 years and up to 5 years	26 (147)
More than 5 years and up to 10 years	17 (98)
More than 10 years	20 (114)
Partners live together	77 (468)
Age difference between partners
No difference	5 (30)
Between 1 and 3 yrs	32 (184)
Between 4 and 6 yrs	20 (114)
Between 7 and 9 yrs	16 (92)
Between 10 and 15yrs	16 (89)
Over 20 yrs	10 (57)
Relationship status
Married	19 (108)
Registered as domestic partners	29 (164)
Been in couples therapy
Ever	24 (136)
In the previous six months	9 (52)

Note: Not all applicable percentages sum to 100% due to rounding errors.

Motivations for making agreements were similar across all three agreement types, with slight differences when ranked by frequency of endorsement (see Table 3). Among couples with monogamous agreements, the most frequently endorsed motivations for making one were: to build trust in the relationship (86%), to be honest in the relationship (84%), and to protect the relationship (83%). Couples with open agreements most frequently endorsed the following: to be honest in the relationship (83%), to build trust in the relationship (73%), and to protect the relationship (66%). The reasons most frequently endorsed by couples with discrepant agreements were: to be honest in the relationship (76%), to build trust in the relationship (75%), and to strengthen the relationship (74%). Although protecting oneself or one’s partner from HIV was endorsed by 74% of the couples with monogamous agreements, 53% of the couples with open agreements, and 53% of couples with discrepant agreements, it was not among the top reasons chosen by couples with any of the three agreement types.

Table 3.

Endorsement of specific reasons for making the agreement, categorized by agreement type

	Agreement Type
	Monogamous	Open	Discrepant
Reasons for Making the Agreement	n = 510	n = 524	n = 88
To build trust in the relationship with primary partner	86	73	75*
To be honest in the relationship with primary partner	84	83	76
To protect the relationship	83	66	69*
To strengthen the relationship with primary partner	82	66	74*
To protect partner/self from STDs	81	56	68*
To please primary partner	75	61	60*
To protect partner/self from HIV	74	53	53*
To have satisfying sex	64	63	53
To promote gay identity of self	35	24	24*
To be more sexually adventurous	31	65	41*
To keep self from getting bored in relationship with primary partner	28	61	35*

Note: The values above denote the percentage of men within each agreement type who endorsed the specified reason.

^*Denotes a significant difference (at the 0.05 level) in proportions across agreement types. The denominator for the proportions reported in the table above consists of all respondents who could be classified into one of the three agreement types described in the Measures section.

Motivations for making agreements were also similar across all three couple serostatus groups, however, only concordant negative couples endorsed HIV or STD prevention among their top three motivators (see Table 4). Among these men, the most frequently endorsed motivations for having an agreement were: to be honest in the relationship (85%), to build trust in the relationship (81%), and to protect partner or self from HIV (79%). For concordant positive couples the primary motivators were the desire to be honest in the relationship (77%), to build trust in the relationship (74%), and to strengthen the relationship (68%). Men in discordant relationships most frequently endorsed the desire to be honest in the relationship (79%), to build trust in the relationship (76%), and to protect the relationship (73%).

Table 4.

Endorsement of specific reasons for making the agreement, categorized by HIV status of couple

	HIV status of couple
Reasons for Making the Agreement	Concordant Negative	Serodiscordant	Concordant Positive
	n = 620	n = 264	n = 248
To be honest in the relationship with primary partner	85	79	77*
To build trust in the relationship with primary partner	81	76	74
To protect partner/self from HIV	79	50	34*
To protect partner/self from STDs	77	58	52*
To strengthen the relationship with primary partner	77	71	68*
To protect the relationship	76	73	67*
To please primary partner	67	66	64*
To have satisfying sex	64	61	61
To be more sexually adventurous	47	45	51*
To keep self from getting bored in relationship with primary partner	42	43	47*
To promote gay identity of self	31	27	26*

Note: The values above denote the percentage of men within each couple serostatus group who endorsed the specified reason.

^*Denotes a significant difference (at the 0.05 level) in proportions across couple serostatus groups.

On comparing scores on relationship characteristics by agreement type, couples with monogamous agreements were distinctive from couples with open and discrepant agreements (see Table 5, (A)) in that they reported significantly higher investment in their agreements, greater intimacy with their partner, more trust toward their partner, more commitment to their partner, more attachment to their partner, and greater equality in the relationship. For each of the above relationship characteristics, couples with discrepant agreements scored the lowest. Couples with discrepant agreements also scored significantly lower in the measure of constructive communication and significantly higher in mutual avoidance and withholding communication than couples with monogamous and open agreements. The level of relationship satisfaction and the level of autonomy in the relationship did not differ significantly among the three agreement types.

Table 5.

Mean scores^* on Relationship Characteristics in Couples

Relationship Characteristic	(A) By Couple’s Agreement Type			(B) By Couple’s HIV status
	Monogamous	Open	Discrepant	Concordant Negative	Serodiscordant	Concordant Positive
Investment in Sexual Agreement **	43.39ab	38.27a	37.89b	41.78a	40.52b	37.59ab
Dyadic Satisfaction	39.39	38.14	37.35	39.49a	38.46b	36.79ab
Intimacy	147.31ab	142.48a	140.31b	146.06a	144.09	141.1a
Trust	26.10ab	22.53a	20.61b	26.21a	23.41b	19.19ab
Commitment	64.37ab	60.86a	60.25b	63.28a	62.32	60.38a
Attachment	53.41ab	48.46a	50.53b	51.45	49.89	50.47
Autonomy	48.67	49.76	47.81	49.23	49.19	48.74
Equality	59.13ab	55.84a	53.42b	57.94	57.26	55.05
Mutual Constructive Communication	9.24a	8.56b	5.17ab	10.11a	8.67b	4.8ab
Mutual Avoidance and Withholding	9.11a	9.19b	10.83ab	8.84a	9.35	10.29a

Note: The superscripts beside the scores denote a statistically significant difference (p<.05) in the adjusted means between groups.

^*All means have been adjusted for length of relationship.

^**The SAIS measure had roughly 20% missing values that were imputed using the multiple imputation procedure MI and analyzed using PROC GLIMMIX. The results were combined using PROC MIANALYZE and the p-values from the multiple pairwise comparisons were adjusted by the step-down Bonferroni method using PROC MULTTEST.

On comparing relationship characteristics by couple serostatus, concordant positive couples were found to be significantly different from concordant negative and discordant couples on many measures (see Table 5, (B)). Concordant positive couples were least satisfied, least invested in their agreements, least trusting, and least able to engage in mutually constructive communication. They were also significantly less intimate, less committed, and more likely to communicate in an avoidant manner in their relationships compared to concordant negative couples. Levels of attachment, autonomy, and equality did not differ significantly by couple serostatus.

Discussion

Our results demonstrate the complex and competing factors behind the making of sexual agreements and the significant differences in relationship characteristics that emerge on examining couples by serostatus and agreement type.

Nearly equal numbers reported monogamous and open agreements, while 8% reported discrepancies. Motivations for making agreements varied by agreement type and couple serostatus. The majority of couples across the agreement type and serostatus groups were motivated to make their agreements by the desire to strengthen or improve their relationship. ‘Preventing HIV infection’ as a motivator for making an agreement fell behind every motivator oriented toward relationship-based factors, such as building trust and honesty. Only concordant negative couples endorsed HIV and STD prevention among their top motivators. The scores on several relationship characteristics varied significantly by agreement type and couple serostatus. Couples with monogamous agreements reported greater agreement investment, intimacy, trust, commitment, attachment, and equality than couples with open or discrepant agreements. Concordant positive couples were found to be distinctly different from concordant negative and discordant couples.

These findings raise several noteworthy issues that have implications for HIV prevention efforts with gay couples. First, the motivations behind agreements are important in a manner separate from the importance of the agreement itself. Our data suggest that couples seek strong, healthy relationships and that partners support each other by making agreements that fit their needs for intimacy, sound health, and sexual fulfillment. Only concordant negative couples were likely to endorse HIV and STD prevention among their top motivators for making an agreement, which suggests their heightened awareness of infection and that they are taking precautions to avoid HIV and STDs. Most other couples were motivated by relationship-based factors, indicating that they associate agreements more with the building, protecting, and strengthening of their relationships than with disease prevention. This is critical to understand in light of the efforts many researchers invest in tying HIV prevention directly to agreements. If indeed most couples conceptualize their agreements primarily as a relationship dynamic and not a risk-reduction strategy, these researchers may have placed the cart before the horse by emphasizing disease prevention as a focal point of the agreement. Researchers and providers working with gay couples may find that knowing what motivates couples to make agreements yields a more effective, albeit indirect, path to HIV prevention.

Second, as a result of surveying both partners we found that some couples give discrepant reports about their agreement. Possible explanations for this are: the agreement is in transition; there is no clear agreement; the agreement is out of date and the understandings and sexual behaviors of one or both partners have drifted away from its original intent; or there is difference of opinion over the agreement. Couples with discrepant agreements are worrisome from an HIV prevention perspective as they may be at substantially higher risk. For example, for a couple with no consensus about which sexual behaviors are acceptable, any risky behaviors with outside partners that might occur may go unreported within the relationship, effectively severing a valuable, risk-reducing feedback loop. An additional point of concern is that previous research has traditionally targeted only one partner in HIV risk reduction efforts. By not involving both partners researchers may not know the full scope of the couple’s agreements and whether there are discrepancies. Prevention providers must encourage couples to discuss their agreements explicitly and regularly so that both partners can make informed choices about acceptable levels of risk.

A final consideration is agreements and their association with relationship characteristics. Couples with monogamous and open agreements did not differ significantly in their relationship satisfaction. This is in line with previous findings (Blasband & Peplau, 1985; Kurdek, 1988; LaSala, 2004) and is encouraging because it shows that gay couples continue to display a great deal of resiliency maintaining satisfying relationships. Other relationship characteristics, however, showed differences by agreement type. Mutual constructive communication, a fulcrum for agreements, was most difficult for couples reporting discrepant agreements. These couples were also most likely to have avoidant and withholding communication patterns with their partners. This is unsurprising given the considerable disconnect implied by the discrepancies in their reported agreements. Couples with open and discrepant agreements scored significantly lower than couples with monogamous agreements on most other relationship characteristics. Monogamous couples may report higher levels of agreement investment, trust, and commitment because they have more at stake if their agreement breaks down. Individuals in concordant negative, monogamous relationships who engage in UAI must trust that their partners remain true to their agreements because their health and sense of fidelity depends on it (Worth, Reid, & McMillan, 2002).

The main strength of the present study is its inclusion of both partners. Other strengths are a racially/ethnically diverse sample, the inclusion of couples with all types of agreements, and large sample sizes of different couple serostatus groups. The generalizability of the findings is restricted due to the non-probability nature of the sample, while the cross-sectional nature of the data precludes causal inferences. Further, couples were eligible only if they reported one primary partner and knowledge of their own and their partners HIV status, thus limiting our understanding of the process of more complex relationship and sexual negotiations.

Agreements play an important role for gay couples and examining the motivations behind them as well as the relationship characteristics associated with them should be an integral part of HIV prevention efforts. Agreement type and couple serostatus further refine these associations and may affect HIV risk. The present study identified several relationship characteristics that, when nurtured, could support couples in their efforts to establish and maintain healthy, satisfying relationships, but, when ignored, could make couples vulnerable to HIV. In-depth examinations of how agreement type, couple serostatus, and relationship characteristics motivate agreements result in more relevant and efficacious strategies to fight the progress of the epidemic in this population.