Estimating sex among South African groups using the dentition

Abstract

Biological profiles are used to assist in the identification of an unknown person. Sex estimation is important in this process as other aspects of the biological profile such as age-at-death, population affinity and stature depend on accurate sex estimates. While the pelvis and long bones, respectively, take preference over the dentition and cranium for sex estimation, dentition remains a good proxy for this parameter due to their post-mortem longevity. The purpose of this study is to examine dental size variation in incisors, canines, premolars and molars of black, white and coloured South Africans and to use discriminant function statistics (LDA) to develop population-specific formulae for the estimation of sex. A total of 906 adult crania were analysed. Measurements included four permanent tooth crown dimensions: maximum mesiodistal, maximum buccolingual and molar diagonal diameters (mesio-buccal – disto-lingual and mesio-lingual – disto-buccal). Statistical analyses included TEM, Student’s t-test, ANOVA, and discriminant function analysis (DFA). Dental dimensions are repeatable with low intra and inter-observer errors ranging from 0.09% to 4.17% and 0.18–6.17%, respectively. Of the 36 dental variables, 26 were statistically significant for biological sex and 17 for population affinity, and included all tooth types. Stepwise discriminant functions with a LOOCV provided correct classification rates of up to 86% for sex. The raw data for the dentition of black, white and coloured South Africans is provided on https://doi.org/10.5281/zenodo.5226935.

Keywords

• Forensic anthropology population data
• Sex estimation
• Odontometrics
• Linear discriminant function analysis