Using posterior probability informed thresholds to develop best practice recommendations for MorphoPASSE using the innominate, cranial, and combined traits

MorphoPASSE is a free program that estimates sex based on morphological traits of the innominate, cranium, or a combined set of traits; however, MorphoPASSE does not provide recommendations on which set of traits performs best or recommend a posterior probability (PP) threshold for sex classification in modern forensic casework. The goals of this study were to compare accuracy rates when using different sets of traits and when imposing posterior probability informed thresholds (PPITs). Innominate and cranial trait score data were collected from four modern US documented skeletal collections (n = 285). Accuracy rates for five mutually exclusive PP intervals were calculated, and PP intervals that were significantly different from chance were condensed into PPITs. Using a PPIT of 0.85-1.00 produced high accuracy rates of 97.3% for the innominate traits and 90.2% for the cranial traits. Using the combined set of innominate and cranial traits resulted in significantly higher accuracy (99.6%) with a lower PPIT of 0.75-1.00. Additionally, the combined trait model corrected all previous misclassifications by either producing a correct sex classification or leaving the individual unclassified for failing to reach the required PPIT. Therefore, when both elements are available, the combined set of traits is recommended with a 0.75-1.00 PPIT. Individuals with a PP falling between 0.75 and 1.00 should be reported as consistent with the male or female reference samples, while those falling below should be reported as "could not be estimated." Use of these recommendations will help standardize the use of MorphoPASSE and reporting sex estimation results.