Development and Validation of a Subadult Sex Estimation Method Using Pelvic Metrics

Abstract

The goal of this research is to explore the utility of pelvic interlandmark distances (ILDs) collected from partially and fully fused subadult innominates in sex estimation. The individuals were from the contemporary US subset of the Subadult Virtual Anthropology Database (SVAD) and between the ages of eight and 20 years (n = 364). Thirty-four pelvic landmarks were placed on 364 left innominates from which 11 standard ILDs were calculated using the X, Y, and Z coordinates. Linear discriminant function analyses using training (75%) and testing (25%) subsets were performed using the ILDs. An additional variable, the ILIS (triradiate cartilage) fusion score, was also incorporated as a predictor variable to capture biological maturity. Total correct sex classifications ranged from 90.88% to 97.78%; correct sex classification of females ranged from 94.29% to 100%, and correct classification of males ranged from 87.95% to 98.18%. Testing and training subsets were comparable, and there was minimal sex bias. Accuracies are remarkably similar to adult sex estimation methods and are stronger than pelvic morphological subadult sex methods currently published. Large samples enabled exploration of metric data to reveal the utility of pelvic ILDs for subadult sex estimation. These results are transformative to the field, as we can now confidently estimate sex, with a high level of accuracy, on individuals with actively fusing innominate or a completely fused innominate, which is younger than previously assumed.