BACKGROUND
Sexual violence (SV) is prevalent among US college athletes, but formal reports are rare. Little is known about adaptations to institution-level reporting policies and procedures that could facilitate reporting.
METHODS
We conducted a discrete choice experiment (DCE) survey with 1004 student-athletes at ten Division I NCAA member institutions to examine how attributes of the reporting system influence the decision to formally report SV to their institution. Changes in utility values were estimated using multinomial logistic regression and mixed multinomial logistic regression. Importance scores were compared to understand student-athlete preferences.
RESULTS
In order of relative importance, the two attributes most preferred by student-athletes were higher probabilities of students perpetrating SV being found in violation of code of conduct policies (relative importance score = 33), and the availability of substance use amnesty policies (relative importance score = 24). Student-athletes with prior SV experiences were more likely to opt out of formally reporting in the DCE paired choice, had lower estimated utility values for all attributes, and had less between-person heterogeneity. While anonymous reporting and survivor-initiated investigations were preferred by student-athletes on average, there was considerable valuation heterogeneity between student-athletes (sizeable deviations from mean estimated utilities). These two attributes also varied in relative importance; anonymous reporting had higher relative importance after interacting levels with prior SV experiences and competitive status, but lower relative importance after interacting levels with whether a student-athlete played on men's or women's sports teams.
CONCLUSIONS
Changes to reporting policies and procedures (e.g., transparency about SV reporting outcomes, implementing substance use amnesty policies) may be promising institution-level interventions to increase formal reporting of SV among student-athletes. More research is needed to understand preference heterogeneity between students and generalize these findings to broader student populations.