Abstract
Background: Gene-environment correlations ( r GEs) are often observed for psychiatric traits, including externalizing problems. Some researchers believe r GEs reflect the effects of offspring genotype on psychosocial environments. Others consider r GEs to be a nuisance, reflecting effects of confounders, such as population stratification, assortative mating, or parental genotypes. Our study aims to examine whether r GEs persist when confounding effects are controlled for in a family-based design. Methods: We examined r GEs between a polygenic score (PGS) for externalizing problems and adolescent psychosocial environments (parent- and child-reported parental relationships, delinquent peer affiliation, and school connectedness) in the National Longitudinal Study of Adolescent to Adult Health. We estimated r GEs within 777 families (1,589 twins/full-siblings) and also at the population-level using 7,635 unrelated individuals. Results: r GEs were observed with parental relationships, delinquent peer affiliation, and school connectedness in the population sample. In family-based analyses, r GEs with parental relationships persisted at the within-family level. However, r GE with delinquent peer affiliation was attenuated and non-significant at the within-family level, suggesting the presence of confounding. Analyses regarding school connectedness were inconclusive. Conclusions: While PGSs are often associated with psychosocial environments, some of these associations may be subject to substantial confounding. Our findings highlight the importance of testing r GE while controlling for potential confounds in large population-based studies.
Funding Information
  • National Institute of Mental Health (R01 MH134039)
  • National Institute of Mental Health (R01 MH128371)
  • Eunice Kennedy Shriver National Institute of Child Health and Human Development (P50 HD105353)

This publication has 59 references indexed in Scilit: