Based on a life course developmental perspective, new and secondary longitudinal data from 3 ongoing nationally representative birth cohort studies (born 2001, 1970, 1958; each >10,000 cases) will examine prevalence, predictors, and consequences of alcohol use in early adolescence through midlife. Specific aims of this developmental epidemiological application are to: (1) Examine variation in the prevalence of early alcohol initiation by conducting the first prospective national study of alcohol use and attitudes via the age 11 survey of the 2001 Millennium Cohort Study (MCS), and testing theory-based hypotheses about child temperamental, cognitive and family influences on early drinking in the MCS and 1970 British Cohort Study (BCS) with structural equation models; (2) In the BCS and 1958 National Child Development Study (NCDS), examine how changes in work, family, and civic social roles across adulthood predict concurrent fluctuations in alcohol use and problem drinking (socialization hypothesis), using fixed effects models to control for stable unobserved effects; (3) Document long-term consequences of mid 20s alcohol use for health and social roles in midlife (age 38/50 in BCS/NCDS) using propensity score models. J-shaped associations are hypothesized in which light/ moderate alcohol use predicts greater midlife health and role success (consequences hypothesis). All aims will use multi-source?parent, teacher, medical, cognitive, self-report?and multi-wave longitudinal data to control for child and adolescent social and personal capital (selection hypothesis) and test whether observed effects are buffered for those with greater capital (selection x socialization, selection x consequences hypotheses). Long-term, the project will follow the MCS cohort through adolescence and beyond and the BCS and NCDS cohorts into older adulthood to document population and sub-population patterns, effects, and variation in vulnerability to heavy and problem drinking. Prospective national data on drinking patterns across the life course, with appropriate controls for potentially spurious risk and protective factors and for conditional effects, is vital for locating vulnerabilities, understanding mechanisms, and informing and targeting developmentally-appropriate programs to decrease harms of alcohol abuse and dependence including disease, injury, and social role problems.