AIMS: Our study has two goals: to evaluate variation in symptoms of substance abuse/dependence by family structure and to examine several potential explanations for this association, including differences in socio-economic status, social support, social stress and perceived approval and use of substances by family and friends. DESIGN: Ordinary least squares (OLS) regression is used to examine the association between family type and problematic substance use and to assess the hypothesized mediators. SETTING: Data were collected between 1998 and 2000 as part of a study of the prevalence and social distributions of psychiatric and substance use disorders. The study involved face-to-face interviews with a representative sample of young adults in a South Florida community. PARTICIPANTS: Respondents (n = 1760) were between 18 and 23 years of age. Approximately 25% were of Cuban origin, 25% other Caribbean basin Hispanic, 25% African American and 25% non-Hispanic white. MEASUREMENTS: Four family types are examined: mother-father families, single-parent families, single-parent families that include other adult relative(s) and stepfamilies. Problematic substance use is measured by a set of 22 substance abuse/dependence symptoms. FINDINGS: Controlling for race-ethnicity and gender, respondents from single-parent families report a significantly higher level of problematic substance use than those from mother-father families. Although nearly all explanations receive support, we find the strongest evidence for differential association with deviant peers and exposure to stress. CONCLUSIONS: Our findings suggest that--rather than representing a unique and independent predictor of substance use problems--family structure can be viewed as a marker of the unequal distribution of factors influencing the risk of problematic substance use.