Intimate partner aggression (IPA) is a serious public health problem that affects more than 1 in 3 women in the United States [1]. Although the definition of IPA continues to be debated [2], we consider IPA to include psychologically, physically, or sexually aggressive or coercive behaviors by a romantic or sexual partner [3]. Research on IPA among sexual minority women (SMW; eg, lesbian, bisexual) is relatively new; this research has primarily focused on heterosexual couples and largely, although not exclusively, on male-to-female aggression [4-6]. However, a growing body of literature suggests that SMW may be at an even greater risk of experiencing IPA than heterosexual women [7-12]. Among women surveyed in the Centers for Disease Control and Prevention’s 2010 National Intimate Partner and Sexual Violence Survey, 32% of heterosexual women, 42% of lesbian women, and 57% of bisexual women reported lifetime physical violence by an intimate partner [8]. Rates of psychological violence or aggression were also higher among lesbian (67%) and bisexual (73%) women than among heterosexual women (47%).

Ample research demonstrates strong linkages between hazardous drinking and IPA among heterosexual couples [5,6]. Hazardous drinking, defined by the World Health Organization as a pattern of alcohol use that increases the risk of harmful consequences (operationalized in this protocol using several indicators such as heavy drinking, heavy episodic drinking, and intoxication), is among the most prominent health-related disparities in comparisons of heterosexual women and SMW. Research examining the associations between IPA and hazardous drinking in SMW’s intimate relationships is limited [13]. Yet, in our research [14], and others’ [15-17], SMW report substantially higher rates of drinking alcohol, heavy drinking, and drinking-related problems than heterosexual women, which may increase the risk of IPA.

Drawing on the sample of SMW enrolled in the 21-year longitudinal Chicago Health and Life Experiences of Women (CHLEW) study, we will recruit the partners of our current participants to examine factors associated with hazardous drinking and IPA in this population. Specifically, we will take an innovative dyadic approach to test the influences of factors known to be associated with IPA perpetration and victimization among women generally (eg, hazardous drinking, relationship conflict, depression, and childhood abuse), as well as sexual-minority-specific factors (eg, sexual identity concealment and internalized stigma), along with new couple-level minority stress constructs that stem from society’s stigmatization of same-sex relationships.

Research in the general population has typically found a positive association between an individual’s drinking pattern and relationship dissatisfaction or IPA [18,19]. For example, a meta-analysis by Cafferky et al [18] indicated a statistically significant association between alcohol use, particularly problematic use, and both perpetration and victimization, replicating the previous 8 meta-analyses that examined alcohol use and IPA. Moreover, because of the expanded number of studies, this meta-analysis demonstrated that the association between alcohol use and perpetration was stronger for men than women, a finding that is consistent with earlier substantive reviews [19]. Although too few to permit meta-analysis, studies of couples’ drinking patterns from a dyadic framework have also demonstrated a relationship between alcohol use and relationship satisfaction or intimate partner violence. A number of large-sample longitudinal studies have demonstrated that the lowest satisfaction and the highest risk for divorce are found in couples in which one member is a heavy drinker, and the other is not [20,21]. The association between more nuanced measures of couples’ drinking patterns and intimate partner violence is somewhat more complicated, although it appears that heavy drinking by either member of the couple or by both is associated with an increased risk of IPA [22-24].

In contrast to studies among heterosexual couples, the few studies on alcohol use and IPA among SMW have largely been descriptive and have had multiple methodological limitations (eg, samples that overrepresent younger White participants or are too small or homogenous for subgroup analyses, lack of guiding theoretical frameworks or perspectives, and lack of dyadic research) [20,23,24]. Thus, the links between SMW's drinking and IPA are poorly understood. We know that hazardous drinking may contribute to or be an indicator of relationship stress and conflict in SMW’s relationships similar to heterosexual relationships. For example, Kurdek [25] asked same-sex couples to rate the issues about which they fought the most, such as finances, sex, and household tasks. Among same-sex female couples, the most frequent disagreements were about drinking or smoking [25]. Drabble and Trocki [15] found that SMW were almost 11 times as likely as heterosexual women to report relationship or social problems (eg, fighting and partner being angry) related to their drinking. Kelley et al [26] found that, controlling for psychological and physical aggression, lesbian women who reported discrepant alcohol use between themselves and their partners also reported poorer relationship quality.

A further limitation of research on the links between alcohol use and IPA among SMW is that it has typically focused on experiencing but not perpetrating IPA. For example, using data from the California Health Interview Survey, Goldberg and Meyer [27] found that both SMW and heterosexual women who binge drank on a daily or weekly basis had significantly higher odds of having experienced IPA. Research on the perpetration of IPA among SMW has tended to focus on discrepancies in drinking between partners as a potential causal factor. For example, in a longitudinal study of lesbian women, discrepant drinking was prospectively associated with being psychologically but not physically aggressive at 6- and 12-month follow-ups [28]. Being physically or psychologically aggressive at baseline was additionally associated with discrepant drinking. Thus, theoretically grounded research that includes both general and sexual minority–specific risk factors and that examines both IPA perpetration and victimization from the perspective of each partner is needed.

Research on IPA has historically taken a gendered perspective (women as victims or men as perpetrators), and for the most part this research has concentrated on either victims or perpetrators, not both simultaneously. Moreover, research on the links between hazardous drinking and IPA has rarely focused on the effects of both partners’ alcohol use. However, findings from existing studies (all with heterosexual couples) suggest that each partner’s alcohol use can independently predict both partners’ physical IPA perpetration. For example, Cunradi et al [46] found that among White and Black couples in which the female partner had alcohol-related problems, rates of female-to-male IPA were five to six times higher than in couples in which the female partner did not have alcohol problems. Among couples in which the male partner had alcohol problems, the risk of female-to-male IPA was 3 to 4 times higher than among couples in in which the male partner had no alcohol problems. These researchers found that among Hispanic couples, women’s alcohol problems were not associated with IPA perpetration. However, among Hispanic couples in whom the male partners had alcohol problems, the odds of IPA perpetration were more than 2 times higher than in couples in whom the male partner had no alcohol problems. Other research among heterosexual couples suggests that one partner’s heavy episodic drinking predicts the other partner’s anger (partner effects), and both actor and partner alcohol use predict physical and psychological IPA [47-49].

There is almost no dyadic research on hazardous drinking among sexual minority couples. In the only published study of which we are aware, LeBlanc et al [50] reported that in a sample of same-sex male and female couples, participants who felt that their relationships were not recognized to the same extent as the relationships of heterosexual couples (a couple-level minority stressor) had higher rates of problematic drinking. Moreover, results showed both actor and partner effects: one partner’s perceived unequal recognition was positively associated with the other’s problematic drinking [44].

Less well-understood are mediators and moderators of the hazardous drinking–IPA link, which is important for identifying modifiable mechanisms that may be used in prevention and intervention efforts. Although there has been research on mediators or moderators of the associations between alcohol use and violence, as noted above, much of it has focused on violence between men. Far less research is focused on couples. Research in this area tends to focus almost exclusively on factors that influence alcohol use and IPA in heterosexual couples [51]. For example, relationship dissatisfaction seems to play a mediating role in this link among heterosexual couples [51,52]. Among women in heterosexual couples reporting bidirectional violence, wanting to appear tough or wanting to intimidate one’s partner mediated the link between alcohol misuse and IPA. However, self-defense and the need to express negative emotions did not [53]. Whether the mediators and moderators of the associations between hazardous drinking and IPA are similar in SMW relationships is not yet known.

Although all SMW are exposed to minority stress, most do not perpetrate or experience IPA—even in the context of hazardous drinking—so there are clearly other factors that influence the risk of IPA in this population. Some evidence suggests that among women aggressors in heterosexual relationships, stress interacts with coping styles to increase aggressive behavior, and stress and coping may moderate the association between alcohol use and IPA [54]. Among lesbian women, general stressors have been found to be positively associated with hazardous drinking, and hazardous drinking is associated with IPA. Those who reported higher levels of emotional distress were more likely to drink to cope, consume more alcohol, and experience more drinking-related problems [55]. In a sample of lesbian women, Mason et al [56] found links between minority stress and physical IPA perpetration and that negative affect and interpersonal intrusiveness (eg, possessiveness) mediated this association.

Together, these findings suggest that hazardous drinking may be an important contributor to IPA perpetration in SMW’s relationships, particularly among couples who experience high levels of minority stress. They also highlight the importance of examining alcohol-related IPA (inclusive of both partners’ reporting of perpetration and victimization) within a dyadic framework and potential mediators and moderators of these associations. For example, as illustrated in Figure 2, both individual- and couple-level minority stressors may influence the experiences of hazardous drinking and IPA perpetration among SMW and their partners (pathway A). The B pathways illustrate the hypothesized association between hazardous drinking and IPA perpetration for individual couple members (actor effects). The C pathways illustrate partner effects (the effects of each member’s hazardous drinking on their partner’s IPA perpetration).

The specific aims of the study are discussed in the following sections.

In the baseline CHLEW study (wave 1; K01AA00266), we collected comprehensive data from a large and diverse sample of SMW who resided in the greater metropolitan Chicago (United States) area. This mostly descriptive study replicated and extended the National Study of Health and Life Experiences of Women (R01AA004610, SCW, PI), a 20-year longitudinal study of drinking patterns, problems, risk factors, and consequences among adult women in the general US population. Since wave 1, four follow-up waves of the CHLEW have been funded (R01AA013328-14, TLH, PI). Wave 2 extended the CHLEW to examine changes in alcohol use patterns and risk and protective factors for hazardous drinking, wave 3 examined the impact of accumulated childhood and adult stressors on drinking outcomes among SMW, and waves 4 and 5 (currently underway) focus on the impact of legalization of same-sex marriage in the United States on hazardous drinking and health. The procedures for each wave have been reviewed and approved by the institutional review board of the university where the principal investigator (TLH) held her primary appointment (waves 1-5 by the University of Illinois at Chicago; waves 4 and 5 and this study by Columbia University). Detailed information about the CHLEW study methods can be found elsewhere [58,59].

We will draw on the large cohort of SMW currently enrolled in the longitudinal CHLEW study for the recruitment of participants. Approximately half of the CHLEW cohort has been followed since 2000 and a half since 2010. We recruited the wave 1 baseline sample using social network or snowball sampling strategies with additional efforts to maximize sample representativeness. Women were eligible if they were aged ≥18 years, lived in the greater Chicagoland area, and self-identified as lesbian. Unlike most previous studies, participants (N=447) represented a wide age range (18-83 years), and more than half were women of color. In wave 2, we successfully located and reinterviewed 86% of the participants. In wave 3, we reinterviewed nearly 80% of all original participants (we were able to locate 85% of the women). In wave 3 (2010-2012), we also recruited a supplemental sample (N=373) of younger (aged 18-25 years), Black, Latina, and bisexual women.

As noted above, the sample was quite diverse in terms of race or ethnicity (35.9% Black, 23.1% Latinx, and 37.4% White). We will invite all CHLEW participants who currently have partners to participate in the CHLEW Couples Study. To determine eligibility, each CHLEW participant will be asked if they are dating or in a committed relationship of at least 3 months with an English-speaking partner aged ≥18 years. CHLEW participants will be asked to invite their eligible partners to participate with them in the CHLEW Couples Study. During the course of the CHLEW study, many participants have changed their sexual or gender identity; we will include otherwise eligible participants regardless of sexual or gender identity.

To aid retention in the longitudinal CHLEW parent study, participants have provided their social security numbers, cell phone numbers, and email addresses and listed the names, addresses, and phone numbers of 4 people who would always know their whereabouts. Letters with return postcards requesting address updates are sent to all participants at 6-month intervals. Other retention strategies included birthday and holiday cards, reminder calls before interview appointments, and graduated monetary incentives—from US $35 in wave 1 to US $80 in this study (US $40 for the telephone interview, US $20 for the web-based self-administered survey, and US $20 if the participant and their partner can be interviewed within 7 days of each other). Of the 820 participants enrolled in CHLEW (as of the writing of the grant application for this study), 49 (6.0%) were deceased, dropped out of the study, or were unable to participate for health reasons, leaving a sample size of 771 (94.0%) [59].

In previous CHLEW surveys, 61%-69% of participants reported that they were in a committed relationship; partial data from wave 4 (underway when the grant application was submitted) indicated that at least an additional 10% were in dating relationships. Although it is possible that the proportion of participants in relationships will be lower in the proposed study, we expect that it will be the same or possibly higher (given marriage equality and improved societal attitudes about sexual minority people and same-sex partnerships). Therefore, we estimate having a recruitment pool of at least 405 (71% of 771) CHLEW participants. In waves 4 and 5 of the longitudinal study, we asked all participants whether their partner (if they had one) was interested in participating in the CHLEW Couples Study. Again, using partially collected data from wave 4, we found that <10% said no. On the basis of a conservative estimate of 12% refusals, our sample would be 357 couples. However, CHLEW includes approximately 50 couples (100 women reported that their partner was also in the study), which reduces the estimated sample size to 307 couples. Of these, we expect that a few partners will not meet the eligibility criteria (age ≥18 years and able to speak English). Therefore, we budgeted for a total of 302 couples (604 individual interviews).

All SMW enrolled in the CHLEW parent study will receive an invitation from the principal investigator, describing the CHLEW Couples Study. Those in relationships of at least 3 months, whose partner is aged at least 18 years, and who can complete an interview in English will be invited to participate. The invitation letter will include a description of the study procedures, information about incentives (US $60 for each partner, plus an additional US $20 each if the couple agrees to be interviewed in the same week), and will emphasize confidentiality (eg, each member of the couple will have a different interviewer; no information from the interview will be shared with the participant’s partner or anyone outside the study team). In the invitation letters, CHLEW participants will be asked to call, text, or send an email message to the research office to indicate their interest. They will also be asked to have their partner contact the research office so that they can be screened and enrolled if eligible. If a partner does not contact the research office within 1 week, a member of the research staff will contact the index CHLEW participant to reassess the couple’s interest in participation. Participation will include a one-time, 60- to 90-minute telephone or videoconference interview conducted by a trained interviewer and a 25-minute web-based survey to be completed within 1 week of the telephone interview. We will ask each member of the couple to refrain from sharing information about the interview or survey until both partners complete the study.

We have collected data using interviews (face-to-face in waves 1 and 2 and telephone or Zoom in waves 3-5; we also completed approximately 100 wave 3 interviews by phone because many study participants had moved outside the Chicago Metropolitan area) as the primary mode of data collection. We feel that these interactions are key to our ability to retain such a high number of participants for 21 years. To assess potential mode effects in wave 3, we compared self-reports of alcohol and drug use among participants interviewed in person with those interviewed by telephone. Although women interviewed by telephone were less likely to report the use of cocaine, we found no differences in any of the hazardous drinking measures. These findings were consistent with the assessments of the 1990 and 2000 National Alcohol Surveys [60,61]. Our finding of limited mode effects in wave 3 provided confidence that this would not significantly influence the self-reports of key variables. Further, in wave 4, we asked participants how they preferred to be surveyed in the future (ie, phone, video, or on the web), and most people had no preference. We moved scales that are of interest but not central to the study’s aims to a web-based module to reduce possible participant fatigue associated with a long telephone or Zoom interview (Tables 1 and 2).

The demographic measures for this study are included in Table 1. The baseline CHLEW instrument was adapted from the National Study of Health and Life Experiences of Women, which used measures with established reliability and validity. In addition to retaining hazardous drinking, sociodemographic, and other key measures from CHLEW’s previous surveys, we added measures that address new research questions and hypotheses relevant to the study aims, including measures of couple-level minority stressors. The CHLEW Couples Study survey instrument retained measures of major drinking variables, relationship variables (eg, relationship satisfaction, commitment, conflict, and IPA), as well as risk and protective factors for hazardous drinking (eg, depression, anxiety, resilience, and social support) included in the longitudinal CHLEW study. We also included several new measures. For example, we added questions about the frequency of physical, sexual, psychological, and verbal IPA—both victimization and perpetration—from the Conflict Tactics Scale [65] and the Psychological Maltreatment of Women Inventory [66]. Each of these measures has been used in prior studies that included SMW [20,107,108]. We also added new couple-level variables (eg, perceptions that the relationship is devalued) developed by LeBlanc et al [50,109]. To account for the potential effects of the COVID-19 pandemic, we included a few questions about the impact of the pandemic on participants, and we will include the date of the interview as a covariate to account for the timing of the interview and at what stage in the pandemic the interview occurred. Table 2 summarizes the measures of the major variables in this study.

Most measures in the CHLEW Couples Study have established reliability and validity. Nevertheless, we will examine dimensional consistency and internal reliability of all scales and functioning across subgroups. We will correct outliers, data entry errors, or other logical inconsistencies. Given that we used a modified version of respondent-driven sampling to recruit the supplemental sample in CHLEW wave 3 [58], we will adjust for potential interdependence by including a sample cluster corresponding to the seed or referral chain (n=75 chains exist in our data) through which each supplemental sample participant was recruited. We will control for individual-level variables such as age, education, sexual identity, gender, race or ethnicity, education, date of interview, and other variables as appropriate in all APIMs. We may also control for couple-level variables such as relationship length, relationship status (eg, legally married), whether there are children aged <18 years living at home, and income. The choice of covariates will depend on the specific hypotheses being tested and will be guided by the current literature.

The changing social landscape, including the legalization of same-sex marriage, has led to heightened interest in research on same-sex relationships; however, this topic is substantially underrepresented in the literature. Furthermore, although research on alcohol use and IPA among sexual minority people has grown in the past 2 to 3 decades, important gaps remain, which the CHLEW Couples Study will address. First, most research on SMW has focused on the prevalence of IPA rather than on individual, relational, and contextual factors. Second, existing studies have focused predominantly on the main effect of associations between hazardous drinking and IPA. Less examined and understood are potential mediators and moderators of these associations, which are particularly important in informing interventions. Third, there are very few published studies on the association between hazardous drinking and IPA among SMW that have included both partners, and even fewer have examined both victimization and perpetration. Addressing these gaps represents an important shift from our (and others’) predominant focus on individual risk factors to also considering couple-level factors. Finally, sample sizes of sexual minority people in population-based studies are typically small, and subsamples of those who are both sexual and racial or ethnic minorities are even smaller; consequently, research on subgroups (eg, racial or ethnic minority SMW) is very rare.

This study is guided by minority stress and IPA theoretical perspectives to better understand both individual and dyadic factors that promote or deter hazardous drinking and IPA among SMW and their intimate partners. Building on our previous work that examined the impact of hazardous drinking on general and sexual minority–specific stressors that accumulate throughout the life span, we will capitalize on an opportunity to examine the links among minority stress, hazardous drinking, and IPA. This study also has a more balanced perspective that considers both risk and resilience from the perspectives of both members of the couple. We hope to advance the fields of sexual minority health and women’s health. Specifically, we will add significantly to the currently sparse knowledge about individual, interpersonal, relational, and contextual factors that contribute to elevated rates of physical and mental health problems among SMW.

In addition to the strengths outlined above, we recognize the limitations of the study. First, it is possible that some study participants may be reluctant to share information about IPA in their relationship because of feared loss of confidentiality or social desirability. These fears may be complicated by the COVID-19 pandemic, given that couples who live together may have less privacy because of shelter-at-home guidelines. For purposes of confidentiality, partners will be interviewed by separate interviewers and assured that no information they provide will be shared with their partners and that all data will be treated as confidential. We will provide all participants with earbuds (with microphones) to help protect privacy (ie, so that others in the household cannot hear the questions being asked of the participant). We will employ interviewers experienced in conducting interviews that ask questions about sensitive topics and will provide them with extensive training on how to ensure participant privacy and maintain confidentiality. To reduce social desirability bias, we will preface IPA questions with a brief introduction that contains additional confidentiality assurances [121-123]. Also, to address the possibility of underreporting of IPA by one or both members of a couple, we will compare their reports of IPA perpetration and victimization. We will consider IPA to have occurred if either partner reports having perpetrated or experienced a particular form of IPA. We will conduct additional analyses of cases in which partners provide discrepant reports of perpetration or experience of IPA to gain additional insights about each partner’s experiences. We will also examine variables such as hazardous drinking, that may be associated with discrepancies in partner perceptions of IPA.

Second, we expect to have a small number of partners who are men or report transgender or nonbinary gender identity. We will conduct exploratory analyses to glean information about potential similarities and differences in findings based on the sex or gender of partners. As we will not know how many mixed-gender dyads we will have in our final sample until the recruitment is over, it is unclear whether we will have large enough subsample sizes of, for example, transgender partners or cisgender men partners to examine couple-level gender differences. There is very little research on couples among whom at least one member is transgender; however, there is some suggestion in the literature that alcohol use may be influenced by unique aspects of those relationships [124-126]. There is also little research with SMW in relationships with cisgender men; however, available data suggest that partner gender plays a key role in drinking-related behaviors in these couples [127-129].

Third, our past experiences of scheduling interviews with individual CHLEW participants suggest that it will take more time and effort to schedule both the CHLEW participant and their partner’s interview within close proximity (the same week). As it is important that the same periods be reflected in reports from each member of the couple, we will provide a US $20 incentive to each member of the couple who is willing to be interviewed in the same week (in separate, private locations). As we will train at least 5 interviewers and because research staff will also conduct interviews, we can be flexible and accommodate most schedule requests. Finally, the CHLEW Couples Study uses a cross-sectional design. For this reason, inferences about causal relationships will be considered cautiously, with careful attention to assumptions about causal ordering that cannot be directly tested.

We expect that the findings of this study will provide the basis for future research aimed at clarifying the causal pathways linking hazardous drinking and IPA among SMW. This research can support the development of individual and dyadic prevention and intervention strategies for SMW and their partners. Research with heterosexual couples suggests that relationship factors are especially important to consider in intervention development [130-132]. Given the existing empirical evidence, there is reason to believe that dyadic interventions are particularly well-suited for treating IPA, especially among couples who do not wish to separate. We are particularly interested in gaining information about modifiable targets for intervention—both general (eg, couple communication, alcohol use behaviors, and emotion regulation) and sexual minority–specific (eg, conflicts about differing levels of sexual identity disclosure)—that can be used in working with SMW. For example, if we find that rejection from families of origin is a particularly strong modifier of the link between minority stress and IPA among couples in the study, dyadic interventions that aim to help couples cope with familial rejection and find other sources of support could be tested.