Sex Differentiation Using Cephalic and Somatic Measurements: A Binary Logistic Regression Approach

One of the key applications of anthropometry lies in sex estimation, particularly in forensic anthropology, where investigation of missing individuals and unidentified cadavers aids in the identification of unknown individuals. Binary logistic regression was deemed useful for sexual differentiation by developing a predictive model that estimates an individual's sex based on various variables. This model assigns predicted probabilities, and a cut-off value (typically 0.5) is used to classify individuals as male or female. The model's predictive power is often assessed using the Receiver Operating Characteristic (ROC) curve, with the area under the curve (AUC) indicating the strength of the model. Thus, it was performed to assess the ability of cephalic and somatic measurements to predict sex. This exploratory study was conducted on 131 adult population, consisting of both males and females, employing snowball sampling and purposive sampling method, using. Martin’s Anthropometer, Spreading, sliding calipers and a flexible steel tape were utilised as the instrument for this study. The model was statistically significant (p < 0.001), and predictors including Max Head Length (p < 0.001), Max Head Breadth (p < 0.001), and Stature (p = 0.006) contributed significantly. Head Circumference was not statistically significant (p = 0.456). The odds of being female decreased with increasing values of Max Head Length, Breadth, and Stature. Key words: Anthropometry, Morphometry, Cephalometry, Multivariate Statistics.