Using data from the 1991 and 1992 National Household Survey on Drug Abuse, multiple logistic regression models were used to estimate the relative risk of initiating non-medical drug use by 12-17-year-old adolescents living in families without both biological parents present compared to those living with both biological parents. To hold constant unmeasured socio-environmental confounding factors, we employed an epidemiological strategy that involves post-stratification of adolescents into neighborhood risk sets. The results indicated that adolescents living in families without both parents were more likely to initiate non-medical drug use, but these associations were varied by sex. These results help measure the impact of divorce/separation/death and/or remarriage of parents on adolescents and identify subgroups at a higher risk for initiating non-medical drug use.