In each cohort (except FTC; see below), we examined measures of typical frequency of alcohol consumption (CON) and heavy episodic drinking (HED), measured in days per month. Each instance of a repeated measure was recorded along with participants’ age at the time of assessment, rounded to the nearest year. As ALSPAC, FinnTwin12, and S4S surveys were administered at structured times, some variability was observed in participants’ age at assessment for each wave. To prevent small cell counts and unreliable model estimates, we generally collapsed these and assigned all individuals to the most common age group within wave (e.g., all S4S participants were coded as age 18 at the first survey wave, while 6% of participants were actually aged 19 and 0.5% older than 19). When sample sizes per age were sufficient, we included multiple age bins (e.g., age ranges of 34–40 in the last wave of FinnTwin12 were recoded as 34–36=35 and 37–40=38). In AddHealth and COGA, we truncated the age range at 32, given the small number of observations outside this range. Based on the low prevalence of drinking
