The initial youth cohort sample included 12,686 men and women between 14 and 21 years of age obtained from two sampling frames, a cross-section of youth in the civilian and military population and an oversampling of black, Hispanic, and economically disadvantaged whites. Interviews with this cohort were conducted each year starting in 1979. Following the 1984 survey round, the military component (1,280) was deleted from the sample. By 1988, the original female cohort ranged in age from 23 to 30 years, and 61 percent had children. A final sample of 681 older youth completed a self-administered questionnaire. Multivariate analysis revealed that, although females generally displayed more vulnerability to specific dimensions of family life than males, responses to these risk factors were not constrained to gender-stereotypic outcomes. Marital discord, recent marital disruption, and to a lesser extent living in a single-parent home appeared to amplify female delinquency. Family risk factors did not exert a greater influence on alcohol use by boys versus girls, but living in a stable, single-parent household was a gender-specific risk for girls. Despite the fact that boys reported drinking more often than girls and engaging in more delinquency, their involvement in deviant activities appeared to be relatively independent of their families. In contrast, recent divorce, marital discord, and prolonged single parenthood placed girls at considerable risk for antisocial behavior. 76 references, 2 tables, and 1 figure