Person-Centered Analysis of Assailant Coercion Tactics Used Prior to Sexual Assault Incidents

Background: Bystander-based programs are an innovative approach to sexual assault prevention because they seek to engage university students as allies in changing community norms and preventing sexual violence when feasible. Although promising, empirical work is needed to better understand assailant behaviors that signal the likely occurrence of sexual assault (Banyard, 2008). Using a person-centered approach, this study sought to uncover common constellations of sexual assault coercion tactics to inform bystander-based prevention curricula. Person-centered analysis represents a novel and useful technique for examining assailant coercion tactics because it investigates relationships among various tactics at a personal level. Such an approach is warranted given that assailants rely on a repertoire of coercive behaviors as opposed to a single tactic (White, Kadlec, Sechrist, 2006).

Methods: Data were collected from 415 female college students who experienced sexual assault. Women were recruited from two colleges in the same urban northwestern region. Those who consented were asked to recall the most severe sexual assault and complete a self-report questionnaire consisting of questions about their experiences prior to, during, and after the assault, including information regarding assailant coercion tactics.

Mplus6 was used to perform latent class analysis with 19 dichotomous indicator variables assessing assailants’ use of coercive tactics prior to the assault. These 19 variables asked about assailants’ use of controlling, manipulative, and malicious coercive behaviors, as well as coercion involving the use of substances. To determine the number of classes that best fit the data, models were systematically and iteratively compared using substantive meaningfulness, Bayesian Information Criteria, Lo-Mendell-Rubin, entropy, and posterior probability statistics. The best fitting model was validated by examining group differences on substantively and empirically-related variables using chi-square and one-way ANOVAs with Bonferroni post hoc comparisons in SPSS18.0.

Results: A three-class model best fit the data. Class one (33%) was distinguished by women with a high likelihood of experiencing coercion tactics of control and a moderate probability of their assailant using substances prior to the assault. Class two (53%) consisted of women with moderate to high probabilities of experiencing multiple controlling, manipulative, and malicious coercion tactics. Class three (14%) was characterized by low to moderate probabilities of experiencing any of the 19 coercion tactics. Group difference analyses revealed three robust and substantively meaningful classes. For instance, participants in class two were most likely to report having a disability, looking for a long-term relationship at time of assault, appraising self as accountable for assault, feeling emotionally hurt at time of assault, and responding to the assault by trying to negotiate with assailant.

Implications: This study found three distinct constellations of coercion tactics used by sexual assailants. Content based on the three classes of assailant coercion tactics can be used to help college students identify common clusters of behaviors that signal the likely occurrence of a sexual assault incident. Further, the results can be used to guide the development of bystander-based sexual assault prevention curricula as well as other secondary prevention programs that help women to appropriately recognize and respond to assailants’ sexual assault coercive tactics.