Co-funded with the Ford Foundation
Substantial evidence (largely from twin studies) suggests that nearly every phenotype of interest in the social or biomedical realm is, to some degree, heritable in the sense that phenotypic differences are partly accounted for by genetic differences. But questions about genomic influence have frustrated social scientific efforts to understand processes of educational, occupational and economic attainment because of the methodological difficulties associated with carrying out genetic analyses. Recently, however, research regarding the association between genetic variants and educational attainment has emerged based on genome-wide association study (GWAS) methods. GWAS results are based on very large samples and are used to generate a polygenic score (PGS) that captures the cumulative predictive power of the genome for some outcome. The PGS can then be applied to other samples that have both social science data and genetic data. Recent evidence indicates that state-of-the-art PGS for educational attainment predicts educational attainment out-of-sample and does so largely by indexing genetic differences that play some causal role.
Social scientists are just beginning the process of understanding the role of the PGS in (1) occupational and educational attainment processes, (2) the mechanisms by which it predicts these attainments, and (3) the extent to which early life environments modify its influences. Professors Benjamin Domingue, Jeremy Freese and Pamela Herd will use the Wisconsin Longitudinal Study (WLS) to analyze newly available genomic data on a sample of older adults. They will address three questions: To what extent are education PGS differences associated with differences in educational, occupational and economic attainments? To what extent do cognitive and non-cognitive characteristics mediate the relationship between the PGS and socioeconomic outcomes? To what extent do early-life environments, including schools, moderate the effects of the polygenic scores?