Results of logistic regression models to test hypothesis 2 and 3.

Notes: exp(B) = exponentiated estimate, CI = confidence interval, RERI = relative excess risk due to interaction. Attraction is coded 1 if same-sex attracted, 0 if other-sex attracted. Victimization is coded 1 if victimized, 0 if not victimized. 95% CI exp(B) is statistically significant when the CI doesn't include 1. 95% CI RERI is statistically significant when the CI doesn't include 0. Results of logistic regression models to test hypothesis 2 and 3.