DNA, Self-Reported Ancestry, and Social Scientific Inquiry

Today, one in five Americans has taken a genetic ancestry test (GAT), which estimates the proportion of a person’s DNA that fall into various geographic categories. While much research has focused on [i] conceptually distinguishing race from ancestry and [ii] understanding how people interpret GAT information, few social scientific studies construct and utilize these quantities – which we call genetic similarity proportions (GSPs). Consequently, the empirical distributions and measurement properties of GSPs remain underexplored. After conceptually distinguishing between genealogical and genetic ancestry, we leverage computational genomic methods to estimate GSPs of genotyped respondents from the nationally-representative Add Health study. Utilizing a survey question on respondents’ self-reported ancestral countries, we observe that – though a person’s self-reported ancestry is associated with their GSPs – there exists considerable variation in GSPs among individuals with the same self-reported ancestry. Then, we explore how GSPs can inform our understanding of racial construction and racialization in the U.S. today by examining their relationship with skin tone and racial classification by others. We find that, among respondents who self-report ancestry from the Americas and/or from Sub-Saharan Africa, GSPs are robustly associated with both skin tone and racial classification – suggesting they may be useful for studying within-group heterogeneity and stratification.