Eleven rape prevention items developed as the result of a telephone survey pretest of 21 rape prevention items were included in a subsequent telephone survey about crime that was administered to 1,600 adults in 3 cities: San Francisco, Philadelphia and Chicago. Factor analyses of the responses of random samples of men and women from the three cities indicate the presence of two relatively independent dimensions of rape prevention attitudes: (1) beliefs about measures calling for restrictions in women's behavior, and (2) beliefs about measures involving changes in the environment, or assertive actions by women. Two-way analyses of variance of restrictive measures and of assertive measures each showed significant main effects for race and sex, but no significant interactions. In both instances, women and blacks rated restrictive and assertive measures as more effective than did men and whites. When the effectiveness ratings of the two types of measures were compared within race-sex groups, three of the groups endoresed assertive measures as more effective, while one group, black females, rated restrictive strategies as more effective. It is concluded that rape prevention indices are useful in summarizing and organizing people's beliefs about preventing rape. The results support a multidimensional concept of rape-prevention beliefs, in accordance with the multidimensionality of rape attitudes reported by Feild (1978). The two dimensions of rape prevention apppear to operate differently in subpopulations that vary in the likelihood of victimization. Black females, the group with the highest risk of rape, rate restrictive measures as more helpful than assertive ones, whereas groups with lower risks of victimization rate assertive strategies as more effective. Thus, there may be differences in the meaning and dynamics of rape beliefs for various race and sex groups. Eleven references, two footnotes, and two tables are provided.