Metascientists commonly distinguish between tests of replicability, reproducibility, robustness, and generalizability. Whereas these distinctions are sensible, they are not often the subject of in-depth definitional exploration. This presentation examines the relation between two of these—replication and generalizability—seeking to clarify their differences in the context of sample diversity. Arguments in favor of increasing the generalizability of our tests focus on the values of inclusion and representation, with little attention to generalizability as an inferential practice (e.g., Roberts et al., 2020). The most detailed accounts of generalizability as an inferential matter have focused primarily on study designs, stimuli, and settings, but not sample diversity (e.g., Nosek and Errington, 2020; Yarkoni, 2021). This presentation will focus on the lack of portability of polygenic scores, derived from genetic-wide association studies, across genetic ancestry groups to argue that existing frameworks for understanding the distinction between replication and generalizability may require revision.