It Depends on How You Look at It: The Role of Alcohol in Men’s Sexual Aggression Perpetration

We thank the commentators for their thoughtful responses to our article. We agree that analyses provide important insights into the role of alcohol in college men’s sexual perpetration while raising additional questions. Alcohol’s role in sexual assault is complex and multifaceted, because its effects may be distal and indirect as well as acute, with effects on perpetrators and victims. Given the impossibility of including all facets within a single study, we (Testa & Cleveland, 2017) considered the distal effects of heavy episodic drinking (HED) and party and bar attendance, as well as individual difference variables, on sexual aggression perpetration over five semesters. HED did not have an independent effect on risk of perpetration; however, that does not mean that alcohol plays no role in sexual assault perpetration. We found robust between- and within-person effects of both party and bar attendance. If we assume that men also drank in these settings, then we are observing significant effects of drinking in social settings on perpetration.

As noted by Abbey (2017) and George and Davis (2017), laboratory analog studies reveal robust effects of administered alcohol on responses suggestive of sexual aggression. We do not see Testa and Cleveland as contradicting these findings. Acute intoxication may have facilitated some sexually aggressive events; however, our wide-angle lens does not permit detection. As part of the larger project of which these analyses are a part, we examined alcohol’s acute, temporal effects in a 56-day daily report study. Drinking episodes facilitated the occurrence of sex with new partners, both with and without aggression, in the next 4 hours (Testa et al., 2015). Moreover, higher intoxication at the time of sex predicted greater use of verbal sexual aggression and intentional intoxication (but not physical force; Brown et al., 2016).

As Abbey (2017) points out, sexual assaults are heterogeneous, and it would be a mistake to imagine that all college sexual assaults involve intoxicated hookups at parties. Although not reported in this article, men who reported perpetration were asked follow-up questions about the most recent incident. Most assaults involved girlfriends or dating partners, fewer than half involved drinking at the time, and most occurred at home (although some encounters may have begun at a party or bar). Although increasing the specificity of the outcome might increase precision of prediction (e.g., alcohol use predicts alcohol-involved events; Kingree & Thompson, 2015), we opted to consider any perpetration as the outcome for both practical and theoretical reasons. Examining predictors of specific types of sexual assaults at multiple time points would complicate the analysis substantially and reduce power to detect effects. More importantly, we believe that aggression-prone men will take advantage of opportunities to perpetrate with or without alcohol and regardless of setting or relationship to the victim. Perpetrators often re-offend (Zinzow & Thompson, 2015), and alcohol-using and sober perpetrators are similar in most respects (e.g., Zawacki et al., 2003).

We believe that the robust effects of party and bar frequency on perpetration represent not simply direct effects of behavioral choices on sexual assault (e.g., increased exposure to intoxicated women or deliberate seeking out of opportunities by predatory men). Rather, given that most assaults did not involve intoxication or new partners, we view party and bar frequency as representing an orientation toward somewhat risky, sensation-seeking, and even deviant behavior (e.g., men were not of legal age to drink or attend bars), an orientation that ebbs and flows over semesters. As George and Davis (2017) aptly state, “Such men experience alcohol as a stage-setting accompaniment for sexual assault proclivities they are already motivated to enact” (p. TKTK).

Finally, George and Davis (2017) noted our failure to report interactions of HED with other covariates. We did test these interactions (both Level 2 and cross-level interactions), as well as the interactions between party and bar attendance and personality variables. Most were not significant, and none support the notion that HED and personality characteristics act synergistically to increase risk of assault. Results are available on request.

Our findings do not point to a simple, magic bullet that can be targeted to prevent sexual assault (e.g., reducing men’s HED). Although men’s alcohol use is not necessarily causal, policies that reduce alcohol availability on campus appear to be a promising strategy to reduce opportunities for perpetration or to alter a culture in which sexual assault is prevalent (Lippy & DeGue, 2016). In a recent study involving 524 college campuses, policies permitting alcohol on campus for students over 21 emerged as a significant predictor of campus rape rate (Stotzer & MacCartney, 2016). Our findings and those of Thompson et al. (2015) point toward the importance of individual difference factors, which are established before college. This suggests a need for earlier intervention (e.g., Foshee et al., 2004) and societal commitment to changing the culture.