Abstract
PURPOSE: Sexual assault is a widespread issue on college campuses, wherein men disproportionately perpetrate this form of violence against women. While cross-sectional research has helped identify salient risk factors related to sexual perpetration, it remains critical for researchers to understand how these risk factors vary over time; however, numerous challenges exist to conducting longitudinal research on sexual perpetration. The current study was conducted to address one of the logistical barriers to longitudinal research - linking data anonymously - by examining two methods for participant-generated study IDs. Further, the current study examines if specific predictors of sexual assault (i.e., trait aggression, sexual assault victimization, and perpetration history) are related to study retention and successful re-creation of study IDs over time. METHODS: Participants included 271 college students (95.1% men) who completed the baseline survey, with 56.1% (n = 152) of the total sample whose data was retained for analyses were randomly assigned to create type A IDs and 43.9% (n = 119) randomly assigned to create type B IDs. RESULTS: A total of 35.1% of the study sample (n = 95) completed the one-week follow-up survey and successfully matched their IDs across time-points. There were no differences between ID types in match rates. Further, results indicated that no differences were detected related to sexual assault-specific predictors on either successful matching of IDs across time, or in study retention. CONCLUSIONS: While none of the individual difference variables significantly predicted study retention or successful ID matching across time, numerous practical implications for improving longitudinal research related to sexual assault are discussed.