Genetic ancestry principal components were computed from GWAS data using Eigenstrat (Price et al., 2006) and the 1000 Genomes, Phase III reference panel. COGA participants were assigned an ancestry classification (European, African, or Other) based on the first two principal components. In the present study, we focused on COGA participants of European ancestry (~75%) because the sample of the discovery genome-wide association study (GWAS) we used to calculate education polygenic scores (see further details below) was primarily of European ancestry (Lee et al., 2018). Polygenic scores derived from discovery GWAS results of largely European ancestry perform poorly in samples of non-European ancestry (Martin et al., 2017). The primary analytic sample included 4,089 individuals of European ancestry who completed at least one SSAGA interview and provided information on educational attainment, had genomic data analysis for calculating education polygenic score, aged 18 or older, and had genomic data available for at least one biological parent. These individuals came from 840 extended families and 1,893 nuclear families. Offspring participants’ age at the last interview ranged between 18–79 (M = 30.07, SD = 9.22;
