For the purpose of the present study, we focused on COGA participants of European ancestry (as determined by their genetic information) because the sample for the discovery GWAS we used to calculate genome-wide polygenic scores were primarily of European ancestry (Liu et al., 2019), and genome-wide polygenic scores derived from GWAS of European-centric samples may be biased in non-European samples (Martin et al., 2017). We included participants who 1) participated at both COGA Phase I (T1) and Phase II (T2), providing longitudinal data, 2) aged 18 of older at T1 and aged 65 or younger at T2, 3) had genomic data available for calculation of genome-wide polygenic scores, 4) reported on sensation seeking and social support, and 5) indicated that they have ever used alcohol. The analytic sample included 2,836 adults, 46% male, age ranged from 18 to 61 (Mean = 35.65, SD = 10.78) at T1, and 22 to 65 (Mean = 41.44, SD = 14.30) at T2.
