Abstract
BACKGROUND: This paper presents two replications of a heuristic model for measuring environment in studies of gene-environment interplay in the etiology of young adult problem behaviors. METHODS: Data were drawn from two longitudinal, U.S. studies of the etiology of substance use and related behaviors: the Raising Healthy Children study (RHC; N=1040, 47% female) and the Minnesota Twin Family Study (MTFS; N=1512, 50% female). RHC included a Pacific Northwest, school-based, community sample. MTFS included twins identified from state birth records in Minnesota. Both studies included commensurate measures of general family environment and family substance-specific environments in adolescence (RHC ages 10-18; MTFS age 18), as well as young adult nicotine dependence, alcohol and illicit drug use disorders, HIV sexual risk behavior, and antisocial behavior (RHC ages 24, 25; MTFS age 25). RESULTS: Results from the two samples were highly consistent and largely supported the heuristic model proposed by Bailey et al. (2011). Adolescent general family environment, family smoking environment, and family drinking environment predicted shared variance in problem behaviors in young adulthood. Family smoking environment predicted unique variance in young adult nicotine dependence. Family drinking environment did not appear to predict unique variance in young adult alcohol use disorder. CONCLUSIONS: Organizing environmental predictors and outcomes into general and substance-specific measures provides a useful way forward in modeling complex environments and phenotypes. Results suggest that programs aimed at preventing young adult problem behaviors should target general family environment and family smoking and drinking environments in adolescence.