Intimate Partner Violence Perpetration and Victimization in Same-Sex and Mixed-Sex Couples: A Comparative Analysis

Background: Intimate partner violence (IPV) is a pervasive social issue with numerous detrimental impacts. Current estimates of IPV prevalence among people who identify as lesbian, gay, bisexual, and/or queer (LGBQ) are wide-ranging, though estimates are consistently equal to or greater than rates for people who identify as heterosexual. For LGBQ groups, IPV prevalence and impacts may be heightened given risk factors associated with stress from being in a marginalized population. Researchers have begun examining differences in IPV perpetration/victimization prevalence between individuals in same-sex and mixed-sex couples; however, this work has not yet employed robust methods for balancing non-equivalent groups across potential confounders. Thus, we employed propensity score analysis (PSA) techniques and logistic regression to examine IPV victimization and perpetration prevalence in same-sex versus mixed-sex relationships across various violent behavior types.

Methods: We used U.S. data from the International Dating Violence Study (2001-2006). This analytic sample of university students included 3,960 respondents who provided details on their most recent mixed-sex and 121 respondents on their most recent same-sex intimate relationship. Scales from the Revised Conflict Tactics Scale were used to measure IPV perpetration and victimization of physical assault, injury, psychological aggression, and sexual coercion. To compare potential differences in IPV victimization and perpetration prevalence, we analyzed multiple logistic regression models iteratively with the key predictor of couple-type (mixed-sex versus same-sex). For comparison, we conducted pre-adjusted, post-matching, and post-weighting analyses while controlling for five covariates in each analysis (gender, age, family income, relationship status, and relationship length).

Results: Pre-adjusted results indicated that of the five types of perpetration and victimization examined, only injury was significantly different by couple type. Those in same-sex relationships had twice the odds of reporting perpetrating IPV resulting in partner injury (OR = 2.06, p < .01), and almost twice the odds of reporting being injured by a partner (OR = 1.95, p < .01) relative to those in mixed-sex relationships. Using PSA, results indicated that both matching and weighting procedures effectively balanced subgroups across all eight covariates. The significant odds ratios for injury perpetration and victimization were reduced to non-significance in the post-matching analysis; however, the odds ratios remained relatively large and approached significance (OR = 1.64 and 1.30, respectively). Post-weighting analyses re-established the significance of the odds ratios associated with injury and generated even larger odds ratios (OR = 2.56, p < .01 and OR = 2.50, p < .01).

Implications: Our results suggest that U.S. university students in same-sex couples may experience higher levels of IPV injury perpetration and victimization than those in mixed-sex relationships. Pre-adjusted and post-weighting analyses suggest that individuals in same-sex relationships have nearly twice the odds (or higher) compared to those in mixed-sex relationships of reporting injury perpetration or victimization. However, post-matching analyses did not support this finding, which may be an artifact of reduced statistical power. These findings underscore the need for employing more robust analytic methods in future research on this topic and the importance of recognizing the unique IPV prevention and service needs of people in same-sex relationships.