Estimating South African socio-cultural identity using sub-specific discriminate 3D shape matrices

Abstract

In forensic anthropology, the probable identification of an unknown individual is based on the presence of quantifiable phenotypic variations and the relationship of these variations to the individual's socio-cultural identity. This study aims to create sub-specific discriminate shape matrices to estimate socio-cultural identity among a modern black South African sample, with a particular emphasis on developing standards for predicting mid-facial variation within this population.

The sample consists of 191 adult South Africans representing nine modern black South African socio-cultural identity groups obtained from the Pretoria Bone Collection in the Department of Anatomy at the University of Pretoria. Three-dimensional (3D) modelling of the relevant anatomical area was performed using an EinScan H 3D scanner. The 3D anatomical extraction was performed by placing 37 standard craniometric landmarks and 388 sliding landmarks on 3D models.

The analysis of variance associated with the linear model “shape against socio-cultural identity” explained 95.5% of overall shape variation showed that variations in midfacial shape configurations were statistically significant (MANOVA: p= 0.001; 50-50 MANOVA: p <2e-16) for all shape configurations, including sub-specific discriminate shape matrices, separately. Additionally, cross-validated linear discriminant function analysis yielded an accuracy between 73.01% and 91.53% for all shape configurations and sub-specific discriminant shape matrices, reflecting the discriminative power of socio-cultural identity groups in the black South African population.

Our findings support the utilization of geometric morphometric methods (GMM) for socio-cultural identity estimation as they allow us to retain the objects' geometry and statistically analyze subtle structural differences.