Our subjects were participants in the adolescent to young adult Prospective sample of COGA (N = 3286). The COGA study started in 1989 and families were recruited between 1989 and 1995. Each family was recruited through a proband with an alcohol use disorder (at that time [25] DSMIII-R and [26] Feighner), targeting successive admissions to treatment facilities. There was a family size requirement (at least two living first-degree relatives) with the idea of prioritizing larger families. All first-degree relatives were interviewed and families were extended through affected subjects (i.e., the identification of an affected uncle of the proband would then lead to invitations for interview of that uncle’s family members). The subjects in the present study were offspring of the proband (or in some cases second-degree relatives). The response rate for recruitment was about 70% or more (with some inter-site variation). More information about the COGA study may be found in [23] Bucholz et al., 2017 and [27] Reich et al., 1998. All offspring in the age range (12–21) at the start of follow-up (2006–2007) were included. Offspring reaching the
