As described elsewhere, (Bucholz et al 2017; McCutcheon et al 2017) parents were considered positive for AUD if they met criteria for lifetime DSM-5 AUD during their interview (a direct assessment, which occurred once). When interview data were not available, affected status was determined by positive family history reports of DSM-IV lifetime alcohol dependence (Rice et al., 1995) or, where family history reports were unavailable, by offspring report that the parent was a heavy drinker or recovering alcoholic in the Important People and Activities interview (Longabaugh et al., 2010). Where information on parental AUD was lacking for all of these sources, observations were coded as missing. Dummy variables indicating mother-only AUD (father negative), father-only AUD (mother negative), and parents who were both positive for AUD were created; a dummy variable representing missing parent AUD was included in regressions to account for missing data and to ensure that the reference category included only parents known to have no lifetime AUD.
