A total of 125 girls with a history of sexual abuse confirmed by youth protection services or a competent professional were administered standard questionnaires and structured interviews. The girls ranged in age from 12 to 17. Information on consensual sexual activity was obtained through a 19-item self-report questionnaire; and the Sexual Abuse History Questionnaire was used to obtain a description of the sexual abuse experienced by the girls, with attention to severity, length, frequency, perpetrators, and disclosure. The Inventory of Family Problems was administered primarily for descriptive purposes. Topics addressed included family adversity, verbal or physical violence from parents, and spousal violence between parents. The study focused on the identification of factors that might increase the probability that sexually abused adolescent girls engaged in at-risk sexual behaviors. The study identified three sexual abuse characteristics that were associated with sexual at-risk behaviors. The severity of sexual abuse was generally found to be associated with the presence of sexual at-risk behaviors on three levels. First, when the abuse included penetration, girls were more likely to be sexually active and at considerable risk for reporting a pregnancy. Second, sexual abuse by more than one person (in one or more incidents) increased the likelihood of all sexual at-risk behaviors except precocity. Third, physical coercion during abuse was associated with victims being sexually active, having multiple consensual sexual partners in the preceding year, and having been pregnant. Of these three characteristics, sexual abuse by more than one person was apparently most consistently associated with at-risk sexual behaviors. Regarding family characteristics, financial difficulties in the family, verbal or physical violence from parents, and witnessing verbal or physical domestic violence were, as a whole, related to being sexually active among abuse victims; however, although family adversity proved to be significant, it did not predict sexual activity when severity of abuse was controlled in the hierarchical regression. Implications of these findings are discussed for both prevention and intervention. 3 tables and 39 references