Objectives This analysis delivers a composite picture of population structure, admixture, ancestry variation, and personal identity in the United States, as observed through the lens of forensic anthropological casework and modern skeletal collections. It tests the applicability of the probabilistic clustering methods commonly used in human population genetics for the analysis of continuous, cranial measurement data, to improve population inference for admixed individuals without prior knowledge of sample origins. Materials and Methods The unsupervised model‐based clustering methods of finite mixture analysis are used here to reveal latent population structure and generate admixture proportions for craniofacial measurements from the Forensic Anthropology Data Bank (FDB). Craniometric estimates of ancestry are also generated under a three contributor model, sourcing parental reference populations from the Howells Craniometric Dataset. Tests of association are made among the coefficients of cluster memberships and the demographic information documented for each individual in the FDB. Clustering results are contextualized within the framework of conventional approaches to population structure analysis and individual ancestry estimation to discuss method compatibility. Results The findings reported here for contemporary American craniometrics are in agreement with the expected patterns of intergroup relationships, geographic origins and results from published genetic analyses. Discussion Population inference methods that allow for the model‐bound estimation of admixture and ancestry proportions from craniometric data not only enable parallel—skeletal and genetic—analyses but they are also shown to be more informative than those methods that perform hard classifications using externally‐imposed categories or seek to explain gross variation by low‐dimensional projections. Am J Phys Anthropol 160:604–624, 2016. © 2016 Wiley Periodicals, Inc.