Racial Frontier: Machine Learning, Biracials, and the Future Racial Hierarchy

• In Event: Intragroup Differences and Identity among Latinos and Black Americans

Abstract

Researchers often examine whether Biracials are more similar to either their lower or higher status racial groups to make inferences about the racial hierarchy. The former circumstance implies stable racial group boundaries and minority exclusion, while the latter augurs assimilation for racial minorities and waning intergroup prejudice. This study develops machine learning algorithms using data on individuals' political characteristics to predict single-race ancestry. The algorithms are then applied to Biracials to determine how likely they are to be classified with either of their single-race groups in order to empirically pinpoint their theoretical placement between them. Leveraging symbolic politics and impressionable-years hypotheses positing stable racial affect beyond early-adulthood, the algorithms are applied separately to generational cohorts to examine whether racial boundaries are shifting overtime. Although Black ancestry continues to possess the most gravity, overtime all Biracial subgroups are becoming increasingly similar to their higher status racial group, indicating decreasing adherence to hypodescent.

Author

• Gregory John Leslie, The Ohio State University