Technical note: Prediction of sex based on five skull traits using decision analysis (CHAID)

Abstract

Osteologists commonly assess the sex of skeletal remains found in forensic and archaeological contexts based on ordinal scores of subjectively assessed sexually dimorphic traits. Using known-sex samples, logistic regression (LR) discriminant functions have been recently developed, which allow sex probabilities to be determined. A limitation of LR is that it emphasizes main effects and not interactions. Chi-square automatic interaction detection (CHAID) is an alternative classification strategy that emphasizes the information in variable interactions and uses decision trees to maximize the probability of correct sex determinations. We used CHAID to analyze the predictive value of the 31 possible combinations of five sexually dimorphic skull traits that Walker used previously to develop logistic regression sex determination equations. The samples consisted of 304 individuals of known sex of English, African American, and European American origin. Based on practical considerations, selection criteria for the best sex predictive trait combinations (SPTCs) were set at accuracies for both sexes of 75% or greater and sex biases lower than 5%. Although several of the trees meeting these criteria were produced for the English and European American samples, none met them for the African American sample. In the series of out-of-sample tests we performed, the trees from the English and combined sample of all groups predicted best.

A Validation Study of the Langley et al. (2017) Decision Tree Model for Sex Estimation

“…The Walker equation correctly sexed 227 of the 281 individuals (80.8%), although there was a strong sex bias toward classifying individuals as males (males = 93.0% correct; females = 68.3% correct; Table ). Applying Stevenson et al's European American CHAID decision tree to the data, a correct classification rate of 82.6% was obtained for the pooled sample (Table ). There is a strong sex bias when males and females are pooled (females = 96.4% correct; males = 69.0% correct), but the low male accuracy is a consequence of including the Black male samples.…”

“…In comparison with the DT1 results, Walker's American/English logistic regression discriminant equation utilizing glabella and mastoid scores ( Y = −1.568 * glabella x − 1.459 * mastoid + 7.434) and Stevenson et al's CHAID European American decision tree utilizing glabella and mastoid scores were also applied to the data. Percent correct classifications were calculated for the total, pooled sexes, and ancestry‐ and collection‐specific samples.…”

“…Intraobserver agreement was not analyzed. Other studies have reported similar levels of interobserver agreement for the mental eminence , although both Walker and Stevenson et al use the mental eminence in their sex estimation methods. Langley et al report that during their analyses, the mental eminence was not selected by any of the decision tree models.…”

“…Stevenson et al used Walker's trait score data and a chi‐square automatic interaction detection (CHAID) procedure to create a decision tree model for sex estimation. They explain that decision trees are better suited for ordinal data than regression models and are easy to implement.…”

“…Before this new method can be confidently applied to forensic casework, tests of validity and reliability by independent observers are necessary. This study seeks to test the Langley et al decision tree on samples of U.S. White and Black individuals, compare the results to those obtained using the Walker and Stevenson et al methods, and provide an independent evaluation of zygomatic extension scoring reliability.…”