Assigned sex estimation with the clavicle and scapula: A study in a Portuguese reference sample

Area of Science:

• Forensic Anthropology
• Bioarchaeology
• Skeletal Biology

Background:

• Estimating biological sex is crucial for skeletal profile assessment.
• Pelvic bone fragmentation or absence necessitates alternative methods for sex estimation.
• Clavicle and scapula offer potential for sex determination when primary indicators are unavailable.

Purpose of the Study:

• To develop new predictive models for biological sex estimation using clavicular and scapular measurements.
• To assess the accuracy and reliability of these models in a Portuguese skeletal sample.

Main Methods:

• Utilized clavicular and scapular measurements from the Lisbon Identified Skeletal Collection.
• Employed decision tree classifier (C4.5) and logistic regression (LR) for model development.
• Generated univariable and multivariable models for sex prediction.

Main Results:

• Achieved high classification accuracy (up to 93.8%) under cross-validation.
• Demonstrated minimal bias (<5%) in the developed models, especially multivariable ones.
• Logistic regression models provide probabilistic sex estimation, accommodating overlapping sexual dimorphism.

Conclusions:

• Clavicular and scapular measurements provide a viable alternative for biological sex estimation in forensic anthropology.
• The developed models offer accurate and reliable tools for sex determination, particularly in cases with incomplete skeletal remains.
• Probabilistic models enhance the utility of these bones for forensic casework.