Two-group interlandmark distance analysis for skeletal sex estimation using resampling and posterior probability contour plots.

The attribution of taxonomic and/or demographic parameters based on the analysis of interlandmark linear measurements remains a widely used approach in both anthropology and zoology, despite the availability of more sophisticated techniques. Especially in forensic contexts, linear measurements are commonly employed for the estimation of sex and/or ancestry to facilitate identification. In the present study we evaluate the potential of applying a resampling-based approach that uses a newly developed R function to efficiently compute interlandmark distances from their Cartesian coordinates and allows identification of which distances are robustly associated with a biological factor; specifically sex. Following the identification of significant dimorphic measurements, we evaluate their associated sex classification accuracy. Furthermore, we demonstrate how results from a training dataset can be applied to novel cases using a rarely implemented, yet straightforward and effective, graphical method. When limited to two variables, this technique involves generating a scatterplot overlaid with interpolated contour lines representing posterior probabilities. By plotting new observations within this data space, users can visually classify sex and its associated probability, without requirement for further computation, analogous to how elevation is estimated on a topographic map using contour lines. We discuss the advantages and limitations of this novel approach and its statistical reproducibility. Broader application of this method could enhance understanding of population specificity in sexually dimorphic cranial measurements and support the development of contour-based tools for the rapid and accurate estimation of skeletal sex in unidentified human remains.