Abstract
Sexual assault on college campuses remains a pervasive issue, with students who consume alcohol at heightened risk for more frequent and severe victimization. A history of childhood trauma and specific drinking motives may increase this vulnerability, yet little research has examined how these factors interact to inform the impact of sexual assault frequency and severity during college. In this study, we used Bayesian multiple regression and random forest modeling to analyze the impact of childhood trauma and drinking motives on sexual assault among college students (N = 624). Bayesian regression analysis showed that risky alcohol consumption was the strongest predictor of sexual assault, B = 12.22, 95% CI [3.63, 5.35]. Other credible predictors included childhood sexual abuse, B = 4.87, 95% CI [1.27, 8.41]; childhood physical abuse, B = 8.78, 95% CI [4.43, 13.18]; childhood physical neglect, B = 7.28, 95% CI [2.34, 12.34]; and conformity drinking motives, B = 9.19, 95% CI [5.40, 13.02]. The random forest model achieved an R
2 value of .68 and identified risky alcohol consumption as the most important predictor, followed by childhood sexual abuse, childhood physical abuse, and conformity motives. These findings were consistent across both analytic approaches. By integrating hypothesis-driven Bayesian modeling with prediction-oriented machine learning, this study provides a nuanced understanding of how risky alcohol consumption, childhood trauma, and drinking motives shape the severity and frequency of sexual assault victimization in college. Results underscore the need for trauma-informed and substance use–focused prevention strategies to reduce the impact of campus sexual assault.