Please register at here to attend!

This talk synthesizes recent historical research demonstrating the close relationship between developments in statistical methods and debates over racial categories and racial admixture in human genetics. I focus on three temporal snapshots, beginning with the early 20th century peak in the collection of quantitative data by physical anthropologists for the purpose of classifying racial groups and detecting histories of human migrations and admixture. In this period, Indian statistician P. C. Mahalanobis formulated his famous distance function—still in use today for applications ranging from finance to machine learning—in order to answer questions about racial admixture in colonial India. By the postwar period, human geneticists collected different kinds of data and developed new statistical approaches, but still largely aimed to address similar questions about “populations” or “ethnic groups” shaped by concepts of racial difference. For example, major scientific disputes emerged between the 1950s and 1970s about the validity of statistical methods used to estimate racial admixture in communities of African Americans and Ashkenazi Jews. I conclude with a reflection on recent computational methods used to analyze “biogeographical ancestry” and the enduring problem of categorizing populations in human genetics.