This study examined the effect of two interventions (didactic and perspective-taking) on five dimensions of heterosexism (hostile, aversive, amnestic, paternalistic, and positive stereotypic). Participants were 117 undergraduate students, randomly assigned to one of three groups in this posttest-only control group experiment. After the intervention, the participants completed measures of sexual prejudice, right-wing authoritarianism (RWA), gender role beliefs (GRB), and a demographic questionnaire. Results from the first step of hierarchical regression models showed that RWA and GRB were significantly predictive of all dimensions of heterosexism except for positive stereotypic, accounting for 13.8% to 56.7% of the variance. This suggests that interventions targeting sexual prejudice should examine both variables. When between-group comparisons were added during the second step, the didactic intervention in comparison with the control group was significantly predictive of lower paternalistic heterosexism; the perspective-taking intervention in comparison with the control group was significantly predictive of lower positive stereotypic heterosexism.