Structural Equation Modeling to Assess Sexual Violence and Associations With Drinking Among College Students

Structural equation modeling (SEM) allows for the examination of complex relationships between violence and associated behaviors. We sought to understand the bidirectional longitudinal associations between sexual violence (SV) victimization and alcohol use on college campuses. We used a subsample (n = 1,545) of 18-24-year-old students, seeking care at health centers across 28 campuses in Pennsylvania and West Virginia, who completed surveys at baseline, 4 months, and 12 months as part of a larger trial (N = 2,291). We used panel models and SEM; controlling for covariates, SV predicted future alcohol use and binge drinking (β's range = .101-.111; p < .001). Alcohol use and binge drinking did not predict future SV. Research-informed interventions that address prior exposure to violence victimization including SV are integral to prevention strategies addressing SV and harmful alcohol use.