In recent years, sexual harassment among adolescents has received significant national attention in the media and from behavioral science researchers. Initial research focused predominantly on describing the prevalence of peer sexual harassment (PSH) among adolescents and understanding the associated outcomes. More recently, researchers have begun to examine individual and contextual risk factors for victimization. Using the problem behavior hypothesis, opportunity theory and developmental theory, a theoretical framework to identify risk factors for PSH was developed, namely risky behaviors (i.e., self-reported delinquency, aggression and friend support for counter conventional behaviors) and sexual saliency variables (i.e., flirting, dating and opposite sex nominations). Moreover, this is the first study that has used classroom demographic factors (i.e., percent male and class size) to predict PSH. Thus, the current cross-sectional study adds to the literature by examining the associations of individual variables with PSH using multilevel modeling techniques, taking into account the nested design of students within classrooms, with theoretically based correlates (i.e., classroom variables, risky behaviors, sexual saliency factors) to examine vulnerability factors for PSH. A cohort of 8th grade students from an urban, culturally diverse and low-income school district was used. Self-report and peer-reported data were obtained from 744 students in the fall. Self-report measures include PSH (AAUW, 1993; 2001); self-reported delinquency (Elliot, Huizinga, and Ageton, 1995), a modified friends' support for counter conventional behaviors scale (Schierer and Botvin, 1998); and dating frequency. Peer ratings of aggressive behavior, flirtatious behavior, and opposite sex peer nominations came from the Revised Class Play (Matesen, Morrison, and Pellegrini, 1985). Hierarchical linear regression analyses indicated that PSH scores did not significantly differ between classrooms. As such, all variables were disaggregated to the individual level. Regression analyses found that classroom variables of class size and percent male were not related to the experience of PSH. Risky behaviors (i.e., self-reported delinquency, friend support for counter conventional behaviors, peer rated aggression) was a significant factor, with self reported delinquency being significant for girls and boys and peer rated aggression was a significant predictor for boys. Sexual salience variables (i.e., dating, flirtatiousness, opposite sex nominations) was also a significant factor, with flirtatiousness being a significant predictor for girls and boys and dating being significant uniquely for girls. Findings indicated that risky behaviors account for a greater variance in PSH victimization in boys, whereas sexual saliency account for a greater proportion of variance in girls. Findings also indicate that 6th grade pubertal status was predictive of 8th grade PSH. These results were discussed within the context of clinical implications for schools and future research directions.