OBJECTIVE: To examine the evolution of risk behaviors over 2 years among a community-based cohort of low-income African American preadolescents and young adolescents enrolled in a randomized trial of an acquired immunodeficiency syndrome risk reduction intervention. DESIGN: Longitudinal, community-based cohort. SETTING: Nine recreation centers serving 3 public housing developments. SUBJECTS: Three hundred eighty-three African American youths aged 9 through 15 years at baseline. INTERVENTIONS: Frequency distributions, chi 2 analyses, and regression analyses regarding 10 risk behaviors were conducted. To assess whether a specific risk behavior or its protective nonrisk) behavioral analogue, composing a risk-nonrisk behavioral complex (eg, was sexually active and was sexually abstinent or used drugs and refrained from drugs), was stable over time, kappa values were determined for the 10 risk-nonrisk behavioral complexes. MAIN OUTCOME MEASURES: Instrument assessing risk/ behaviors administered at baseline and every 6 months aurally and visually via talking computer. RESULTS: The prevalence of sexual intercourse, cigarette smoking, alcohol consumption, and drug use increased notably over time. Drug use increased from a 6-month cumulative prevalence of 7% at baseline to 27% at the 24-month follow-up (P < .001). Cumulatively over the 2-year study interval, 81% of youths had engaged in fighting, 58% had engaged in sexual intercourse, and from 33% to 40% had engaged in truancy, knife or bat carrying or both, alcohol consumption, drug use, and cigarette smoking. All of the risk-nonrisk behavioral complexes except weapon carrying were stable during the semiannual assessment intervals. Fighting kappa = 0.22, P < .01), sexual intercourse (kappa = 0.33, P < 001), alcohol consumption (kappa = 0.21, P < .001), and unprotected sexual intercourse (kappa = 0.34, P < .05) were stable for 2 years. Six risk-nonrisk behavioral complexes were stable for the 2-year interval among youths aged 13 through 15 years at baseline, while only 2 risk-nonrisk behavioral complexes were stable among younger youths. The intervention seemed to affect the stability of 4 risk behaviors: truancy, drug use, unprotected sexual intercourse, and, possibly, fighting. For unprotected sexual intercourse, this intervention effect seemed to be due to stabilization of nonparticipation in risky behavior. Intervention youths were less likely to adopt a risk behavior ie, engage in it for > or = 2 risk assessment periods) than control youths, but they were not less likely to experiment with a risk behavior. CONCLUSIONS: There is evidence that although the prevalence of risk behaviors does change with age, most risk-nonrisk behavioral complexes seem to be relatively stable over time and stability may increase with time. Risk reduction interventions seem to decrease risk adoption, stabilize nonrisk behaviors, and possibly destabilize risk behavior.

*Adolescent Behavior *Blacks *Risk-Taking

www.aegis.org