OBJECTIVE To determine whether religious affiliation reduces HIV risk among young women in Zambia, and to examine the effects of religious affiliation on sexual initiation and on condom use during first sexual experience. METHODS Data from a representative probability sample of 5534 women aged 13-20 years was analyzed. The instrument included questions on sexual initiation, condom use during first sex, religious affiliation, and sociodemographic characteristics of respondents. Statistical tests were performed at the bivariate and multivariate levels. Cox proportional hazards and logistic regression were used at the multivariate levels. Standard errors were adjusted for the "clustering" effect found in data from multistage cluster samples. RESULTS Affiliation with religious groups that excommunicate members for engaging in premarital sex, and that oppose condom use has both positive and negative effects on behaviors that carry the risk of HIV infection; young women affiliated with conservative groups are more likely to delay sexual initiation but less likely to use condoms during first sex. CONCLUSIONS Denominations that are not only strongly opposed to premarital sex and condom use, but are able to exercise control over adolescents through socialization or the threat of social exclusion, are likely to create conflicting behaviors among adolescents that cancel each other in terms of HIV risk. Overall, these findings suggest that affiliation with conservative religious groups is unlikely to reduce the risk of HIV infection. Additional studies are recommended.