Meta-analyses were performed on 9 studies of subjects' (N = 434) penile responsivity to rape stimuli. The mean effect size for raw score data suggested that most sexual aggressors against women exhibit slightly more rape arousal than control or comparison subjects, whereas the mean effect size for rape index (rape arousal: consenting sexual arousal) data suggested a moderate between-groups difference. Rape index effect sizes in individual studies, however, were heterogeneous. The rape index appears sensitive to differences between sexually aggressive men and those who are not but is not sensitive to differences between men who sexually aggress against women versus other types of sexual aggressors. Also discussed are the needs for standardization and methodological and statistical improvements.