Sex estimation by discriminant function analysis of long bones in prehistoric Southeast Asian populations

Abstract

Biological sex estimation is an integral part of reconstructing the biological profile of an individual in forensic anthropological and bioarchaeological contexts. Formulating population specific discriminant function equations for metric variables is vital for reconstructing biological sex of fragmentary skeletal remains. This study aimed to develop multivariable and univariable sectioning point sex estimation equations from long bones of prehistoric Thailand and Cambodia people dated from 4700 to 1450 BP. A total of 481 individuals (236 females and 245 males) with 997 long bone measurements were analyzed. Discriminant function analysis was used to analyze sexually dimorphic measurements from long bones of humeri (177 females and 183 males), femora (169 females and 178 males), and tibiae (139 females and 151 males). Stepwise and direct multivariable functions offered the highest accuracies of 97.3% for humeri, 97% for femora, and 96.7% for tibiae. Univariable functions indicated that the recommended measurements for use in sex estimations with high cross-validation accuracies are the humeral epicondylar breadth (89.1%), femoral maximum head diameter (87.1%), and tibial midshaft circumference (88.3%). These equations are applicable for use in sex estimation for the specific prehistoric Southeast Asian populations to improve our understanding of the prehistoric demography. Further evaluation and validation of the equations are required to test whether these equations can also be applied to estimate biological sex of contemporary Southeast Asian populations.