DESCRIPTION (provided by applicant): Illicit anabolic-androgenic steroid (AAS) use represents an important public health problem; NIDA has recently funded a major public education campaign to counter AAS use, with television advertising in English and Spanish. Surprisingly, however, we are not aware of any NIDA-funded studies of illicit AAS users (aside from a team-based substance abuse prevention program) in the last four fiscal years. Because of the limited available research, risk factors for illicit AAS use are poorly understood. In a recent pilot study, we found that AAS use may be associated with two clusters of risk factors, an antisocial cluster and a body image cluster. To test these hypotheses, we propose a cross-sectional cohort study- an advantageous and underutilized design, which we show is equal in validity to the more traditional case control design, but particularly suited for the study of risk factors in our setting. We will assess 300 men, drawn from a population of regular weightlifters, and shown by our pilot data to contain approximately 50% of AAS users. To minimize observer bias, one interviewer will assess subjects for the outcome variable (history of AAS use), and a separate blinded interviewer will assess subjects for all exposure variables (childhood and adolescent attributes that may be risk factors for AAS use). These assessments will use a battery of structured personal interviews and self-report measures, developed from our pilot study and extensive previous AAS research. We hypothesize that AAS use will be significantly associated with 1) childhood and adolescent antisocial traits and associated attributes, including parental substance use, lack of parental supervision, risk-taking, and drug use among peers; and 2) childhood and adolescent body image concerns and associated attributes, including eating disorders, body dysmorphic disorder, and other disorders in the obsessive-compulsive spectrum. We will also test the secondary hypothesis that the antisocial cluster predicts drug use in general, whereas the body image cluster predicts AAS use in particular. Findings from this study will lead to more focused and efficient strategies for 1) identifying boys and young men most likely to use AAS; 2) targeting populations at greatest risk, to ensure optimum deployment of educational resources for reducing AAS use; 3) treating psychiatric disorders (such as body image disorders) in individuals at high risk for AAS use; and 4) designing subsequent prospective intervention studies of strategies to prevent AAS use.