Exploring Mechanisms of Racial Disparities in Intimate Partner Violence among Sexual and Gender Minorities Assigned Female at Birth

Abstract

Objectives.

Sexual and gender minority people of color (SGM-POC) report higher rates of intimate partner violence (IPV) than White SGM, adding to growing evidence that people holding multiple stigmatized social identities are at particular risk for adverse experiences. We aimed to identify mechanisms underlying the racial/ethnic disparities in IPV among SGM, focusing on childhood experiences of violence, structural inequalities, and sexual minority stress.

Method.

308 SGM assigned female-at-birth (AFAB; 82 White, 133 Black, 93 Latinx; age 16-31) self-reported on minor psychological, severe psychological, physical, and sexual IPV victimization and perpetration, and three proposed mechanisms: childhood violence (child abuse, witnessing interparental violence), structural inequalities (economic stress, racial discrimination), and sexual minority stressors (internalized heterosexism, anti-SGM victimization, low social support). Indirect effects of race on IPV victimization via hypothesized mechanisms were estimated using logistic regression with 5000 bootstrapped samples.

Results:

Compared to White participants, Black participants were 2.5-7.03 times more likely to report all eight IPV types; Latinx participants were 2.5-4.8 times more likely to experience four IPV types. Univariate indirect effects analyses indicated that these racial/ethnic disparities were partially explained by higher economic stress, racial/ethnic discrimination, and childhood violence experiences (for Black and Latinx participants) and lower social support (Black participants). In multivariate models, the most robust indirect effects were through racial/ethnic discrimination and childhood violence.

Conclusions:

Findings underscore the need for policy and interventions aimed at preventing IPV among SGM-POC by targeting factors that contribute to IPV disparities in this group, particularly racial/ethnic discrimination and family violence.

Intimate partner violence (IPV), which includes physical aggression, emotional abuse, and sexual violence between dating partners, is prevalent among all U.S. adolescents and young adults (e.g., Breiding et al., 2014; Kann et al., 2016). There is growing evidence, however, that risk for IPV is highest among young people who hold multiple stigmatized or marginalized social identities. Individuals who identify as sexual minorities (i.e., as lesbian, gay, bisexual, queer or any other non-heterosexual identity) and/or gender minorities (i.e., as a gender that does not match their sex assigned at birth) consistently report more IPV than heterosexual and cisgender individuals (e.g., Kann et al., 2011; Porter & Williams, 2011). Further, among sexual and gender minorities (SGM), those whose sex or race represents an additional minority identity are at even higher risk for IPV. Sexual minority adolescents girls report higher rates of physical (Martin-Storey, 2015) and sexual IPV victimization (Olsen et al., 2017) than sexual minority boys, and in samples of SGM young people, those assigned female at birth (FAB) were more likely to experience verbal and physical IPV than those assigned male at birth (MAB; Reuter et al., 2017; Whitton, Newcomb, et al., 2019). Regarding race, sexual minority women of color report significantly more IPV than White sexual minority women (e.g., Balsam & Szymanski, 2005; Steele et al.,, 2020) and Black SGM youth are 3-4 times more likely than White SGM youth to experience verbal and physical IPV victimization (Reuter et al., 2017).

Recent research suggests that SGM who are both AFAB and racial minorities (i.e., have three minority identities) may be at particularly high IPV risk. In a sample of young SGM-AFAB (ages 16-32) who reported on IPV during the last 6 months, Black and Latinx participants were 2.5 to 4.7 times more likely to report psychological IPV victimization and perpetration of psychological and physical IPV than White participants (Whitton, Dyar, et al., 2019). Black, but not Latinx, SGM-AFAB were also more than twice as likely to report physical and sexual IPV victimization, and over six times more likely to have perpetrated sexual IPV, than White participants. These findings are consistent with concerns that SGM-AFAB people of color (POC) are particularly vulnerable to a variety of negative experiences and health outcomes due to their multiple minority identities (e.g., Greene, 1995; Bowleg et al., 2003). Despite these concerns, SGM-AFAB-POC remain understudied, limiting our ability to develop policies and interventions to reduce these IPV disparities.

Grounded in Intersectionality theory (Crenshaw, 1990), which emphasizes how individuals’ experiences are shaped by the unique convergence of their multiple identities, scholars are increasingly calling for research on how co-occurring social identities- including those related to sex, race, and sexual identity- may intersect to affect health in minority populations (Graham et al., 2011). As part of these efforts, research is needed to better understand the factors that place SGM-AFAB-POC at greater risk for adverse events, even compared to White SGM-AFAB. Toward that aim, the current study sought to identify mechanisms underlying the racial disparities in IPV among SGM-AFAB. Our investigation was framed within the dynamic developmental systems perspective (DDS; Capaldi et al., 2012), which highlights how IPV risk is influenced by both individual background/developmental factors and contextual factors. Accordingly, we focused on three potential mechanisms of risk: childhood experiences of violence, a key background/developmental risk factor, and two broader contextual factors highly relevant to this population-- race-based structural inequalities, and sexual minority stress.

First, we explored whether childhood experiences of violence may partially explain the racial differences in IPV among SGM-AFAB. Compared to White children, Black and Latinx children are more likely to experience physical, sexual, and emotional abuse (Oscea Hawkins et al., 2010; Wildeman et al., 2014) and to witness interparental violence (Covey et al., 2013), likely because families of color are more likely to live in disadvantaged neighborhoods and face economic hardship, both robust correlates of family violence (e.g., Coulton et al., 2007). According to social learning theory (Bandura & Walters, 1977), witnessing and experiencing violence as a child leads to use of violence towards others, including romantic partners, later in life through socialization and learning processes. Much research supports this intergenerational transmission of violence: witnessing aggression between parental figures and experiencing abuse during childhood each increase the risk for IPV during adolescence and adulthood (reviewed by Capaldi et al., 2012; Smith-Marek et al., 2015). This suggests that the higher rates of IPV among SGM of color compared to White SGM may be, in part, attributable to their higher risk of experiencing violence within the family of origin.

Second, social-structural theories (Gelles, 1985) propose that IPV reflects not the cultural characteristics of a given demographic group, but the social structural conditions they face (e.g., poverty, discrimination). In other words, racial differences in IPV may be largely driven by structural inequities, which place greater burdens upon and offer fewer opportunities to racial minorities than White people. According to the US Census, 20.8% of Black Americans and 17.6% of Latinx Americans live in poverty, versus only 8.1% of White Americans (Semega et al., 2019). Black and Latinx individuals also report greater financial hardship than White individuals (Mimura, 2008; Short, 2005). Because economic hardship is a risk factor for IPV (Lucero et al., 2016; Copp et al., 2016), the heightened risk for IPV among people of color may be driven by these income inequalities. Indeed, racial differences in IPV have disappeared when controlling for household income in heterosexual samples (Cazenave & Straus, 1979; Rennison & Planty, 2003). Interestingly, however, socioeconomic status did not account for racial differences in IPV in a sample of sexual minority women (Steele et al., 2020). More research is needed to explore economic hardship as a potential mechanism of racial disparities in IPV among SGM people.

People of color in the U.S. also face racism and discrimination, which can be disempowering and limit access to educational, occupational, and social resources that reduce IPV risk. Theorists have emphasized the role of such cultural oppression in violence between romantic partners, which mirrors broader public violence toward oppressed groups (Almeida et al., 1994). Accordingly, experiences of ethnic/racial discrimination have been associated with greater risk for IPV victimization among heterosexual Asian American (Cho, 2012) and Black women (Waltermaurer et al., 2006). Further, among Black and Latina young adult women, racial discrimination was positively associated with victimization and perpetration of both psychological and physical IPV, controlling for other risk factors (Stueve & O’Donnell, 2008). We could only locate two studies examining these issues among SGM; both used samples of adult cisgender male sexual minorities and found that racial discrimination was associated with higher likelihood of physical and sexual IPV victimization, but not perpetration (Finneran & Stephenson, 2014; Stephenson & Finneran, 2017) . It is particularly important to examine whether similar associations are present for SGM assigned female at birth (AFAB), who experience higher rates of IPV than sexual minorities assigned male at birth (Reuter et al., 2017).

Third, we evaluated whether differences in sexual minority stress, defined as the social stress that people face as a result of stigma against their minority sexual identity (Meyer, 2003), contribute to racial differences in IPV among SGM-AFAB. In largely White samples of SGM, IPV risk has been associated with various sexual minority stress experiences, including anti-SGM discrimination (Balsam & Szymanski, 2005), victimization (Edwards & Sylaska, 2013), internalized heterosexism (e.g., Balsam & Szymanki, 2005; Edwards & Sylaska, 2013; Finneran et al., 2012), and low social support from family and peers (Katz-Wise & Hyde, 2012; Whitton, Newcomb, et al., 2019). Although SGM of all races experience these sexual minority stressors, stigma against SGM may be higher in racial/ethnic minority than White communities. Compared to White Americans, Black and Latinx Americans report higher religiosity (The U.S. Religious Landscape Study, 2014), which is associated with more negative views about non-heterosexuality (e.g., Ellison et al., 2011; Sherkat et al., 2010). In qualitative research, sexual minorities of color describe heterosexism and social isolation within their own racial/ethnic group (e.g., Ghabrial, 2017); for example, engaging in same-sex behavior may violate gender roles that are particularly important within the Latinx culture (Pachankis & Goldfried, 2013). Studies examining racial differences in sexual minority stress, however, have yielded conflicting findings. Some indicate that, compared to White SGM, SGM-POC experience more anti-SGM victimization (Comstock, 1989), heterosexist discrimination (Swank et al., 2013; Whitfield et al., 2014), and internalized heterosexism (Barnes & Meyer, 2012; Finneran & Stephenson, 2014; Glick & Golden, 2010). However, other studies found no racial differences in these sexual minority stressors (Balsam et al., 2013; Huebner et al., 2004; Moradi et al., 2010; Meyer et al., 2008). Further research is needed to explore whether SGM-POC experience higher levels of sexual minority stressors than do White SGM, and if so, whether these differences explain racial disparities in IPV.

In this paper, we aimed to better understand the high risk for IPV among SGM POC by exploring potential mechanisms through which holding a racial minority identity may increase risk for IPV. The current analyses build upon a previous study using the same sample of young SGM-AFAB, which found that Black and Latinx participants experienced higher rates of several types of IPV victimization and perpetration than did White participants (Whitton, Dyar, et al., 2019). In the current study, we focused on explaining the racial differences in four major domains of IPV—minor and severe psychological abuse, physical violence, and sexual IPV--between the three most highly represented racial categories in the sample: Black, Latinx, and White (other racial groups were too small to include with adequate power). We hypothesized that identifying as Black or Latinx, compared to White, would be indirectly associated with higher rates of IPV through childhood experiences of violence (child abuse and witnessing violence between parents), structural inequalities (economic stress, ethnic/racial discrimination), and sexual minority stressors (internalized stigma, SGM victimization, and low social support).

Methods

Participants and Procedure

Data were drawn from FAB400, an ongoing cohort study of 488 young SGM AFAB (Whitton, Dyar, et al., 2019). It includes two cohorts: (1) a late adolescent cohort recruited in 2016-2017 (N = 400; 16-20 years old at baseline), and (2) a young adult cohort comprised of AFAB participants from FAB400, a study of SGM youth that began in 2007 (N = 88; 23-32 years old at baseline in 2016-17). Inclusion criteria required participants to be female-assigned at birth, 16-20 years old at original enrollment, speak English, and either identify with a sexual or gender minority label, report same-gender attractions, or report same-gender sexual behavior. Each cohort was recruited using an incentivized snowball sampling approach, in which participants were recruited directly from various venues (i.e., SGM community organizations, health fairs, high school/college groups) and online social media advertisements, and then could refer up to 5 peers to the study.

In 2016-2017, all 488 participants completed the baseline assessment. Participants were paid $50 for completing a battery of self-report measures using computer-assisted self-interview. The study protocol was approved by the Institutional Review Board (IRB) at Northwestern University with a waiver of parental permission for participants under 18 years of age under 45 CFR 46, 408(c). Participants provided written informed consent, and we used mechanisms to safeguard confidentiality (i.e., federal certificate of confidentiality).

The current analytic sample excluded 136 participants who reported no romantic relationship in the last 6 months. We then selected the 308 participants who identified as Black, Hispanic/Latinx, or White (excluding 44 participants of other races); by focusing on these three groups, fairly equally represented in the sample (ns = 82-133), we maximized power to detect racial differences. This analytic sample was diverse in terms of sexual and gender identity (see Table 1 for sample demographics).

Table 1.

Demographic characteristics of analytic sample (N = 308).

Variable	N (%)
Race/ethnicity
Black/African-American	133 (43.2)
Hispanic/Latinx	93 (30.2)
White	82 (26.6)
Sexual Orientation
Gay/Lesbian	85 (27.6)
Bisexual/Pansexual	171 (55.5)
Other Sexual Identity	52 (16.9)
Queer	31
Unsure/Questioning	8
Straight/Heterosexual	5
Asexual	4
Not Listed	4
Gender Identity
Cisgender Female	240 (77.9)
Gender Minority	68 (22.1)
Male	8
Transgender	10
Gender Non-Conforming	18
Gender Queer	18
Non-binary	9
Not listed	5
Age (M, SD)	20.47, 3.90

Measures

Demographics

Age.

Age was calculated by subtracting self-reported birthdate from the date of assessment.

Race.

Participants were asked to select the option(s) that best described their race from the following: American Indian or Alaskan Native, Asian, Black or African American, Native Hawaiian or Other Pacific Islander, White, or other (please specify). Participants also indicated if they identified as Hispanic or Latino/Latina/Latinx, regardless of race. As recommended by the National Institutes of Health (2001), all who selected a Latinx ethnicity were classified as Latinx regardless of race.

Gender identity.

Participants responded to the question, “What is your current gender identity?” with the following options: Male, female, transgender, gender non-conforming, genderqueer, non-binary, and not listed (please specify). We coded participants as either cisgender women (self-identified as female) or gender minorities (identified with any other gender identity).

Sexual identity.

Participants were asked, “Which of the following commonly used terms best describes your sexual orientation?” with the options: Gay, lesbian, bisexual, queer, unsure/questioning, straight/heterosexual, pansexual, asexual, and not listed (please specify). For analyses, sexual identity was recoded into three categories: monosexual (gay or lesbian), non-monosexual (bisexual/pansexual), and other.

Potential Mechanisms of Racial Disparities

Childhood Experiences of Violence.

Child Abuse Experiences.

Participants reported (yes = 1; no= 0) whether during childhood they experienced four forms of verbal abuse (e.g., “Swore at you, called you names, said insulting things like you're "fat", "ugly", "stupid", etc. more than a few times a year;” α = .78) and six forms of physical abuse (e.g., “Hit you so hard that it left marks for more than a few minutes;” α = .78) on the Maltreatment and Abuse Chronology of Exposure (MACE)(Teicher & Parigger, 2015), and six sexual abuse experiences on the ENGAGE Child Sexual Abuse scale (α = .85; Leserman, 2005). Scores for verbal, physical, and sexual abuse were summed to create an overall index of child abuse experiences; this scale has shown test-retest reliability and convergent validity with other child abuse measures (Teicher & Parigger, 2015).

Witnessing violence between parents.

On the witnessing physical abuse between parents subscale (α = .77) of the MACE (Teicher & Parigger, 2015), participants responded to 5 items (e.g., “saw adults living in the household push, grab, slap or throw something at your mother, stepmother or grandmother;” yes = 1; no= 0)). Item responses were summed.

Structural Inequalities.

Economic stress.

On the “Can’t make ends meet” subscale of the Economic Pressure Scale (Conger et al., 1999), participants responded to three items (e.g., “I had enough money to meet my expenses”) on 4- and 5-point Likert-type scales. Responses were standardized and summed (α = .77); higher scores indicate greater economic stress and have shown strong correlations with other economic strain indices (Conger et al., 1999).

Ethnic/racial discrimination.

Participants completed a 12-item version of the Brief Perceived Ethnic Discrimination Questionnaire, Community Version, which has shown evidence of validity in Black and Latinx individuals (PEDQ-CV; Brondolo et al., 2005). Participants reported how often certain experiences of discrimination happened to them in the past 6 months because of their ethnicity (e.g., “Have policemen or security officers been unfair to you?” 1 =Never happened to 5 = Happened very often). Responses to items were averaged (α = .89).

Sexual Minority Stressors.

Internalized Sexual Minority Stigma.

On the 8-item Desire to be Heterosexual subscale of the Internalized Homophobia Measure, which has demonstrated convergent and predictive validity in SGM (Puckett et al., 2017), participants rated their agreement with statements (e.g., “Sometimes I feel ashamed of my sexual orientation”; 1 =Strongly disagree to 4 =Strongly agree). Items were averaged (α = .87); higher scores indicate more internalized stigma. Five heterosexual participants did not complete this scale (all were gender minorities).

SGM Victimization.

On a 10-item scale developed by Pilkington and D'Augelli (1995), participants rated the frequency (0 =Never to 5 =More than ten times) of SGM-based victimization experiences (e.g., “In the past six months, how many times have you been threatened with a knife, gun, or another weapon because you are or were thought to be gay, lesbian, bisexual or trans*?”). Items were averaged (α = .79) and multiplied by ten to facilitate interpretation.

Social support.

The Multidimensional Scale of Perceived Social Support, which has shown test-retest reliability and discriminant validity (Zimet et al., 1988), includes 12 items tapping support from family, friends, and a significant other (e.g., “I can count on my friends when things go wrong;” 1=Very strongly disagree to 7 =Very strongly agree). Scores represent the mean of all 12 items (α = .90); higher scores indicate higher levels of social support.

Intimate Partner Violence

The Sexual and Gender Minority Conflict Tactics Scale (SGM-CTS2; Dyar, Messinger, et al., 2019) is an adaptation of the CTS2 (Straus et al., 1996), modified to be culturally appropriate for SGM individuals. The SGM-CTS2 has demonstrated the same factor structure as the CTS2, along with evidence of reliability and validity (Whitton, Dyar, et al., 2019). Each item, which describes a specific IPV behavior or event, is asked first for perpetration (e.g., “I slapped [partner name]”) and then for victimization (e.g., “[Partner name] slapped me”). Participants indicated how frequently the given event occurred on a scale of 0 (never), 1 (once), 2 (twice), 3 (3-5 times), 4 (6-10 times), 5 (11-20 times), 6 (more than 20 times), and 7 (not in the past 6 months, but it did happen before). Because we were interested in the past 6 months only, responses of 7 were recoded as 0.

For this study, we used four SGM-CTS2 subscales: minor psychological, severe psychological, physical, and sexual IPV. Four pairs of questions measured minor psychological IPV (e.g., "I swore at my partner;" α = .74 for perpetration and .75 for victimization), and four others assessed severe psychological IPV (e.g.: “I destroyed something that belong to [partner name]”; α = .67 for perpetration and .67 for victimization). Twelve pairs of items assessed physical violence (e.g., “I pushed or shoved [partner name];” “I used a knife or gun on [partner name]” (perpetration subscale: α = .85; victimization subscale: α = .82). Five pairs of items captured sexual IPV (e.g.: “I used force -like hitting, holding down, or using a weapon- to make [partner name] have sex”) (perpetration subscale: α = .54; victimization subscale: α = .49). The lower than typical alphas, likely due to many 0 responses (Sijtsma, 2009), are not concerning because scores were dichotomized to capture presence of each IPV type.

Results

Rates of each type of IPV are presented by race in Table 2. Consistent with growing evidence that IPV is often bidirectional (e.g., O’Leary & Slep, 2012), participant reports of victimization and perpetration of each IPV type were associated (rs = .78 for minor psychological IPV, .67 for severe psychological IPV, .74 for physical IPV, and .48 for sexual IPV). Correlations among different IPV types ranged from 0.16-0.49, suggesting that 2-24% of their variance is shared.

Table 2.

Racial Differences in Intimate Partner Violence

	Rates of IPV (%)			Racial Differences (Odds Ratios)
IPV Type	Black	Latinx	White	Black	Latinx
Minor Psychological Victimization	78.2	62.4	48.8	2.81**	1.51
Minor Psychological Perpetration	80.5	69.9	59.8	3.41**	1.71
Severe Psychological Victimization	36.8	30.1	14.6	3.90**	2.49*
Severe Psychological Perpetration	43.6	35.5	14.6	4.76**	3.11**
Physical Victimization	30.8	18.3	14.6	2.52*	1.19
Physical Perpetration	33.8	25.8	9.8	4.69**	3.08*
Sexual Victimization	20.3	15.1	8.5	2.78*	1.92
Sexual Perpetration	17.3	10.8	2.4	7.03*	4.86*

Note.

^*p < .05

^**p < .01; Reference group: race (White). Odds ratios were estimated controlling for age, participant gender identity, and sexual orientation identity.

Racial Differences in IPV

We first tested for racial differences in IPV using logistic regression analyses. Separate models were run for each IPV type, including dummy variables for Black and Latinx as predictors (reference group = White), and controlling for age, gender identity (coded as 0 = cisgender; 1 = gender minority), and sexual orientation identity (dummy variables for bi/pansexual and other identities; gay/lesbian = reference group). The resulting Odds Ratios, shown in Table 2, indicated that Black participants were more than 2.5 times as likely to experience each type of IPV than White participants, with the most striking difference seen for sexual IPV perpetration (OR = 7.03). Latinx participants did not differ from White participants in rates of minor psychological perpetration or minor psychological, physical, or sexual IPV victimization, but were 2.5-4.8 times more likely to experience the other IPV types. Note that some ORs differ slightly from those presented in Whitton, Dyar, et al (2019) due to a slightly different sample (excluding other race/ethnicities) and control variables.

Racial Differences in Risk Factors for IPV

Second, we tested whether each IPV risk factor differed by race using ANCOVAs with race as a 3-level factor (Black, Latinx, White) and age, gender identity, and sexual orientation identity as covariates. Statistically significant F tests were followed up with pairwise comparisons using Fisher’s Least Significant Difference (LSD) test. As shown in Table 3, Black and Latinx participants reported more child abuse experiences, interparental violence, economic stress, and racial discrimination than White participants. There were no racial differences in sexual minority stressors other than that social support was lower in Black than White participants.

Table 3.

Means (Standard Errors) of Risk Factors for Intimate Partner Violence, by Race/Ethnicity.

Risk Factor	Black (n = 133	Latinx (n = 93)	White (n = 82)	Test of Group Differences
				F (2, 301)	p-value
Childhood Experiences of Violence
Child Abuse Experiences	5.94 (.34)a	5.25 (.39)a	3.29 (.42)b	11.89	< .001
Witnessing Violence Between Parents	0.80 (.11)a	.92 (.13)a	.35 (.14)b	5.15	.006
Structural Inequalities
Economic Stress	0.18 (.07)a	.09 (.08)a	−.41 (.09)b	14.81	< .001
Racial Discrimination	1.79 (.05)a	1.64 (.06)a	1.06 (.07)b	38.84	< .001
Sexual Minority Stressors
Internalized SM Stigma	1.64 (.05)	1.69 (.06)	1.61 (.06)	0.44	0.64
SGM Victimization	3.42 (.40)	2.31 (.47)	2.18 (.50)	2.26	0.11
Social Support	5.03 (.10)a	5.27 (.12)a,b	5.52 (.13)b	4.45	0.01

Note. All means are adjusted for age, gender identity, and sexual orientation identity. For tests comparing the racial/ethnic groups on internalized stigma, n = 129 for Black, n = 92 for Latinx, n = 82 for White participants; degrees of freedom are 2 and 296.

Note. Potential range for racial discrimination is 1-5, for child abuse experiences is 0-16, for witnessing violence between parents is 0-5, for internalized SM stigma is 1-4, for SGM victimization is 0-5, for social support is 1-7. Actual range for economic stress is −1.43-1.74.

^a,bGroup means with different superscripts differ significantly from each other at p < .05 according to Fisher’s LSD tests.

Associations Between Risk Factors and IPV

Next, we assessed whether the potential mechanisms of racial disparities were associated with IPV using separate logistic regression models predicting each IPV type with each potential mechanism, controlling for demographic variables. Results are presented in Table 4. Childhood experiences of violence showed expected associations with IPV. Odds ratios indicate that, for each additional child abuse experience reported, participants were 11-19% more likely to report various forms of IPV in last 6 months. Similarly, likelihood of all IPV types except minor psychological victimization increased by 25-44% for each type of interparental violence witnessed. Both structural inequality variables, economic stress and ethnic/racial discrimination, were associated with greater likelihood of experiencing all IPV types except minor psychological perpetration. Contrary to hypotheses, internalized stigma was not associated with any IPV outcome. SGM victimization was only associated with higher risk for minor psychological victimization and perpetration and sexual IPV victimization. Social support was negatively associated with risk for each type of IPV except minor psychological.

Table 4.

Associations between Risk Factors and Intimate Partner Violence experiences (Odds Ratios).

Risk Factor	Minor Psych Vict	Minor Psych Perp	Severe Psych Vict	Severe Psych Perp	Physical Vict	Physical Perp	Sexual Vict	Sexual Perp
Childhood Experiences of Violence
Child Abuse Experiences	1.14**	1.18**	1.19**	1.12**	1.17**	1.13**	1.11*	1.12*
Witnessing Violence Between Parents	1.25	1.31*	1.33**	1.43**	1.25*	1.36**	1.43**	1.44**
Structural Inequalities
Economic Stress	1.43*	1.34	1.66**	1.36*	1.74**	1.69**	1.49*	1.68*
Racial Discrimination	2.41**	2.48**	3.11**	2.67**	2.22**	2.06**	2.35**	2.58**
Sexual Minority Stressors
Internalized SM Stigma	1.23	1.39	1.49	1.49	1.11	1.33	1.35	1.16
SGM Victimization	1.17**	1.22**	1.05	1.04	1.06	1.04	1.07*	1.04
Social Support	.84	.80	0.75*	0.73**	0.63**	0.78*	0.67**	0.59**

Note.

^*p < .05

^**p < .01. Odds ratios were estimated controlling for age, participant gender identity, and sexual orientation identity.

Indirect Effects

Finally, we assessed for indirect effects of race on IPV via each of the proposed mechanisms that differed by race (i.e., excluding internalized stigma and SGM victimization). Using PROCESS for SPSS (Hayes, 2013), Model 4, we estimated the indirect effects of identifying as Black and as Latinx (versus White, the reference group) on each IPV type via each proposed mechanism with 5000 bootstrapped samples. We first ran separate models for each risk factor, and then multivariate models including all risk factors simultaneously to assess for unique indirect effects. All models controlled for age, gender identity, and sexual orientation identity. Results are shown in Table 5. When risk factors were considered separately, there were indirect effects of race (for both Black and Latinx races) on all forms of IPV via childhood experiences of violence; interestingly, for minor psychological IPV this was only true for child abuse and not witnessing parental IPV, whereas for sexual IPV, this was only true for witnessing parental IPV and not child abuse. There were also indirect effects of Black and Latinx race via racial/ethnic discrimination on all eight IPV types, and via economic stress only on psychological IPV victimization and physical IPV victimization and perpetration. For Black participants, there were indirect effects of race on IPV via lower social support on four types of IPV (severe psychological perpetration, physical victimization, and sexual victimization and perpetration). There were no indirect effects via social support for Latinx participants, who did not differ from Whites in social support.

Table 5.

Indirect Effects of Race/Ethnicity on IPV through Putative Mechanisms of Effect.

		Univariate Models				Multivariate Model
		Indirect Effect: Black		Indirect Effect: Latinx		Indirect Effect: Black		Indirect Effect: Latinx
Outcome	Risk Factor	Estimate	95% CI	Estimate	95% CI	Estimate	95% CI	Estimate	95% CI
Minor Psychological Victimization	Child Abuse Experiences	.28	.07 - .61	.21	.04 - .49	.23	−.02 - .55	.17	−.01 - .45
	Witnessing Parental IPV	.07	−.02 - .25	.09	−.03 - .27	.02	−.09 - .17	.02	−.13 - .18
	Economic Stress	.13	−.07 - .36	.11	−.07 - .32	.06	−.16 - .30	.05	−.14 - .26
	Racial Discrimination	.48	.14 - .95	.38	.11 - .76	.36	−.01 - .86	.28	−.01 - .69
	Social Support	.06	−.05 - .23	.03	−.03 - .13	−.02	−.17 - .13	−.01	−.11 - .07
Minor Psychological Perpetration	Child Abuse Experiences	.39	.15 - .76	.29	.09 - .61	.32	.06 - .71	.24	.03 - .57
	Witnessing Parental IPV	.11	−.004 - .30	.13	−.002 - .34	.02	−.09 - .19	.03	−.12 - .21
	Economic Stress	.10	−.11 - .34	.09	−.09 - .29	.01	−.24 - .26	.01	−.20 - .23
	Racial Discrimination	.57	.19 – 1.11	.44	.15 - .89	.40	−.02 - .97	.32	−.01 - .77
	Social Support	.08	−.04 - .25	.04	−.03 - .15	−.01	−.17 - .15	−.004	−.10 - .08
Severe Psychological Victimization	Child Abuse Experiences	.41	.19 - .72	.30	.11 - .58	.30	.07 - .62	.22	.05 - .51
	Witnessing Parental IPV	.11	.003 - .26	.13	.01 - .29	.03	−.09 - .16	.03	−.12 - .18
	Economic Stress	.22	.01 - .51	.19	.01 - .46	.09	−.14 - .36	.08	−.12 - .31
	Racial Discrimination	.73	.39 – 1.15	.58	.29 - .96	.55	.18 – 1.00	.43	.13 - .84
	Social Support	.11	−.003 - .28	.06	−.02 - .19	−.01	−.16 - .15	−.003	−.09 - .10
Severe Psychological Perpetration	Child Abuse Experiences	.23	.05 - .47	.17	.03 - .37	.07	−.16 - .30	.05	−.12 - .24
	Witnessing Parental IPV	.13	.03 - .30	.17	.05 - .34	.10	−.01 - .27	.12	−.01 - .29
	Economic Stress	.08	−.13 - .31	.06	−.11 - .28	−.05	−.30 - .17	−.04	−.25 - .15
	Racial Discrimination	.55	.22 - .93	.43	.17 - .76	.44	.09 - .86	.34	.07 - .70
	Social Support	.12	.01 - .29	.06	−.02 - .19	.06	−.07 - .22	.03	−.04 - .13
Physical Victimization	Child Abuse Experiences	.40	.19 - .71	.30	.12 - .55	.27	.03 - .59	.20	.02 - .46
	Witnessing Parental IPV	.09	−.01 - .23	.12	−.01 - .26	.01	−.12 - .14	.01	−.15 - .15
	Economic Stress	.30	.07 - .60	.25	.06 - .55	.17	−.08 - .47	.14	−.06 - .41
	Racial Discrimination	.55	.19 - .96	.44	.14 - .80	.28	.11 - .71	.22	−.08 - .60
	Social Support	.21	.06 - .42	.10	−.02 - .27	.11	−.04 - .32	.06	−.03 - .19
Physical Perpetration	Child Abuse Experiences	.25	.05- .51	.18	.04 - .39	.14	−.10 - .42	.10	−.07 - .32
	Witnessing Parental IPV	.12	.01 - .28	.15	.02 - .30	.08	−.03 - .22	.09	−.04 - .25
	Economic Stress	.23	.005 - .52	.19	.004 - .46	.15	−.09 - .44	.13	−.07 - .38
	Racial Discrimination	.36	.01 - .70	.28	.01 - .59	.18	−.20 - .56	.14	−.15 - .47
	Social Support	.09	−.03 - .26	.05	−.02 - .16	.01	−.13 - .16	.01	−.08 - .10
Sexual Victimization	Child Abuse Experiences	.23	−.01 - .51	.17	−.01 - .39	−.002	−.36 - .32	−.001	−.28 - .23
	Witnessing Parental IPV	.15	.02 - .32	.19	.05 - .36	.11	−.01 - .32	.14	−.02 - .34
	Economic Stress	.18	−.08 - .50	.16	−.06 - .43	.03	−.27 - .35	.02	−.24 - .31
	Racial Discrimination	.57	.18 – 1.00	.45	.14 - .80	.41	.01 - .94	.33	.01 - .76
	Social Support	.18	.02 - .42	.09	−.02 - .27	.11	−.07 - .36	.05	−.04 - .22
Sexual Perpetration	Child Abuse Experiences	.22	−.04 - .55	.16	−.02 - .41	−.03	−.40 - .30	−.02	−.30 - .22
	Witnessing Parental IPV	.15	.01 - .33	.18	.02 - .39	.10	−.06 - .31	.13	−.07 - .36
	Economic Stress	.23	−.04 - .56	.19	−.03 - .47	.07	−.24 - .41	.06	−.20 - .34
	Racial Discrimination	.56	.15 – 1.01	.44	.11 - .82	.37	−.10 - .95	.29	−.07 - .76
	Social Support	.24	.04 - .55	.12	−.03 - .33	.18	−.03 - .49	.09	−.04 - .30

In the multivariate models, once accounting for the other risk factors, there were no longer indirect effects of race through many of the IPV risk factors. In fact, there were no statistically significant estimates of unique indirect effects through any risk factor on minor psychological victimization, physical IPV perpetration, or sexual IPV victimization or perpetration. The most robust findings were for racial discrimination; identifying as Black or Latinx (vs. White) had an indirect effect on severe psychological perpetration via racial discrimination, identifying as Black (vs. White) had an indirect effect via racial discrimination on three forms of IPV victimization (all but minor psychological), and identifying as Latinx (vs. White) had an indirect effect via racial discrimination on psychological and sexual IPV victimization. Child abuse was the only other risk factor through which race indirectly affected IPV in multivariate models; identifying as Black or Latinx (vs. White) was indirectly associated with minor psychological perpetration and with severe psychological and physical IPV victimization through child abuse. Direct effects of race were nonsignificant in these models, with the exception of Black identity in the prediction of severe psychological perpetration, B = .99 (SE= .42), p =.02 and physical perpetration, B = 1.01 (SE= .47), p =.03.

Discussion

The current findings emphasize racial disparities in the IPV experiences of young SGM-AFAB, and help shed light on the mechanisms behind these disparities. Replicating previous research with other SGM (e.g., Edwards et al., 2015; Reuter et al., 2017), Black and Latinx participants in this SGM-AFAB sample reported significantly higher rates of a wide range of IPV experiences than did White participants. These findings suggest that young people who hold multiple stigmatized identities (i.e., related to race in addition to sex and sexual orientation) are particularly vulnerable to IPV. As such, they are consistent with theories that SGM-AFAB-POC are at ‘triple jeopardy’ for adverse outcomes due to the compounding negative effects of minority stressors associated with each of their social identities (e.g., Greene, 1995; Bowleg et al., 2003) and highlight the importance of using an intersectional approach that attends to individuals’ multiple social identities in health-related research (Bauer, 2014; IOM, 2011). Study results also build understanding of the developmental/background and contextual factors contributing to these racial disparities in IPV, within the DDS persepctive (Capaldi et al., 2012).

Consistent with theories proposing that social-structural conditions are primary drivers of racial disparities in violence (Gelles, 1985), racial discrimination emerged as the most consistent mechanism through which Black and Latinx identities were associated with IPV risk. Extending previous evidence that racial discrimination is associated with IPV victimization among heterosexual women of color (Cho, 2012; Waltermaurer et al., 2006) and male sexual minorities (Finneran & Stephenson, 2014; Stephenson & Finneran, 2017), we found that racial discrimination was associated with victimization and perpetration of minor and severe psychological, physical, and sexual IPV in this sample of SGM-AFAB. Further, racial discrimination mediated racial differences in every type of IPV assessed when considered in isolation (i.e., without other putative mechanisms in the model), and fully explained race effects on six of the eight types of IPV (all except psychological and physical IPV perpetration). Together, these findings suggest that the elevated rates of some types of IPV among SGM-POC may be largely accounted for by their experiences of race-based discrimination. As such, they add to a large literature identifying discrimination as a key driver of racial disparities in health and wellbeing (Williams & Mohammed, 2013) in the U.S., and suggest that efforts to reduce IPV among SGM-POC should include macrolevel strategies to decrease racial discrimination as well as individual interventions to help racial minorities cope with discriminatory experiences.

However, when all other putative mechanisms were also considered, there were no longer statistically significant indirect effects of race via racial discrimination on half of the IPV outcomes. It may be that the models simultaneously testing five indirect effects for two racial groups lacked power to detect small unique indirect effects, especially given that the outcomes were dichotomous and in some cases had low variance due to very high or low base rates (i.e., minor psychological and sexual IPV, respectively). The fact that no indirect effects remained significant in multivariate models predicting three IPV types supports this possibility. Alternately, although racial discrimination explains higher rates of severe psychological IPV and of physical and sexual victimization among SGM-AFAB of color, it may not be as influential in understanding their heightened risk for other forms of IPV. Future research is needed to explore this possibility and identify alternate mechanisms that do explain racial differences in those other IPV forms, particularly physical and sexual perpetration.

The background factor, childhood experiences of violence, also emerged as a fairly strong explanatory mechanism for racial disparities in IPV among SGM-AFAB. Echoing racial differences observed in large national samples (e.g., Wildeman et al., 2014), Black and Latinx participants experienced more child abuse and witnessed more interparental violence than White participants. Consistent with an intergenerational transmission of violence, these experiences were each associated, in turn, with greater risk for victimization and perpetration of psychological, physical, and sexual IPV. Further, indirect effects analyses indicated that identifying as Black or Latinx, versus White, was indirectly associated with all six types of IPV via childhood experiences of violence (though results differed slightly for indirect effects via child abuse vs. via witnessing interparental violence). Even in multivariate models when all other risk factors were controlled, race was indirectly associated with severe psychological and physical IPV victimization and with minor psychological perpetration through child abuse, suggesting that experiencing abuse as a child uniquely accounts for racial differences in these IPV types. Together, these findings highlight how efforts to reduce IPV among SGM-POC must attend to racial disparities in early exposure to violence, as well as the structural inequities (neighborhood disadvantage; poverty) that contribute to them. Future longitudinal research might explore whether racial disparities in economic hardship, like those we observed in this sample and are a well-established risk factor for family violence (e.g., Coulton et al., 2007), may lead to racial differences in childhood experiences of violence that, in turn, raise risk for IPV.

Findings were weaker for economic stress. Mirroring the income inequalities between racial/ethnic groups in the U.S. (Semega et al., 2019), Black and Latinx SGM-AFAB reported greater economic stress than White SGM-AFAB in this sample. In univariate models, there were indirect effects of race via economic stress on three of the eight types of IPV. However, there were no indirect effects via economic stress in univariate models predicting the five other IPV types or in any multivariate model. Together with previous evidence that socioeconomic status did not account for racial differences in IPV among sexual minority women (Steele et al., 2020), these findings suggest that economic inequities may not be as powerful in explaining racial disparities in IPV among SGM as among heterosexual individuals (Rennison & Planty, 2003). Further research is need to explore why this might be the case.

Similarly, findings failed to provide strong support for social support as a mechanism of racial differences in SGM-AFAB IPV. Although low social support was associated with all types of IPV other than minor psychological, it did not differ between Latinx and White participants. Further, though Black participants reported less social support than White participants, and having a Black racial identity was indirectly associated with several types of IPV via lower social support when other proposed mechanisms were not included in the models, these indirect effects disappeared in all multivariate models. This pattern of findings suggests that Black SGM-AFAB may be particularly vulnerable to social isolation, perhaps due to heterosexism in their own racial community (e.g., Ghabrial, 2017), racism from peers and dating partners (Stonewall, 2018; Wilson et al., 2009), or both. However, this low social support does not uniquely explain Black SGM-AFAB’s elevated risk for IPV in the context of other, potentially more potent, risk factors such as racial discrimination and childhood violence.

Notably, findings did not support the sexual minority stressors of internalized stigma or victimization as mechanisms through which SGM-POC are at heightened risk for IPV. We did not observe any racial differences in either variable, which is consistent with some past studies (e.g., Balsam et al., 2013; Moradi et al., 2010) but not others (e.g., Swank et al., 2013; Whitfield et al., 2014). Further, contrary to previous findings that both of these minority stressors are risk factors for IPV (e.g., Balsam & Szymanki, 2005; Edwards & Sylaska, 2013; Finneran et al., 2012), internalized stigma was not associated with any IPV outcome and SGM victimization only predicted minor psychological IPV and sexual victimization. It is possible that the low levels of internalized stigma (1.64 on a 1-4 scale) and SGM victimization (0.28 on a 0-5 scale, representing 1-3 victimization events in 6 months) in this sample may have limited our ability to detect associations with other variables. Future research should explore whether other more commonly experienced sexual minority stressors may differ by race and help explain racial disparities in SGM IPV. It is also possible that, as suggested by Intersectionality theory (e.g., Bowleg, 2008), our examination of SGM stressors as separate from racial minority stressors failed to capture unique experiences of SGM of color that are particularly impactful for IPV risk. Future studies would benefit from using intersectional measures that capture minority stressors related to both identities (e.g., “how often have you been called a name because you are a sexual minority person of color?”; Balsam et al., 2011).

Study limitations include the cross-sectional data, which cannot speak to direction of effects or mediational processes. Our models predicted each type of IPV separately, though the various types of IPV were intercorrelated, and because we did not assess both partners, individual effects were not considered in the context of partner effects. The non-probability sample limits generalizability of findings: Recruitment via SGM-focused events and social media connections may have selected for SGM-AFAB highly integrated in the LGBT community, perhaps leading to the low internalized stigma and victimization observed, and findings may not reflect SGM-AFAB’s experiences in regions less accepting of racial and sexual minorities than Chicago. Also, we excluded SGM-AMAB and racial groups other than White, Black, and Latinx; other studies are needed to explore racial disparities in the IPV experiences of these groups. In multivariate models, none of the assessed mechanisms accounted for racial differences in minor psychological IPV victimization or physical and sexual IPV perpetration, suggesting that we may have failed to consider other, more influential mechanisms. Identifying these mechanisms is critical, as physical and sexual IPV perpetration show the largest racial disparities among the IPV types and may be particularly harmful to victims (Bonomi et al., 2007).

Despite these limitations, this study yielded novel information about racial disparities in IPV among young SGM-AFAB, with important clinical and policy implications. High rates of IPV among participants of color highlight the need for IPV prevention and intervention efforts targeting this high-risk group specifically. At a minimum, existing IPV prevention programs and victim services must be inclusive, culturally sensitive, and accessible to individuals who are both racial and sexual/gender minorities. Further, efforts to reduce the racial disparities in SGM IPV should target racial discrimination and childhood violence, which emerged as the most consistent mechanisms behind those disparities. Specialized services for SGM-POC might include strategies based in Racial Encounter Coping Appraisal and Socialization Theory, which show promise in reducing the negative psychological effects of racial discrimination (Anderson et al., 2019). Services that support effective parenting, especially among racial minority parents facing significant economic stress that increases the likelihood of harsh physical discipline, may also help break the intergenerational transmission of violence in communities of color. More importantly, these findings speak to the importance of broad societal change to reduce structural inequities in the U.S., as discrimination appears to be a key driver of racial disparities in IPV among SGM-AFAB.

Public Significance Statement:

Among sexual and gender minorities assigned female at birth, those who also identify as a racial minority are at heightened risk for intimate partner violence (IPV). An exploration of the mechanisms behind these racial disparities revealed that they are largely accounted for by higher levels of racial discrimination and childhood violence experienced by racial minority, compared to White, individuals.