NSF Award
2214747
The goal of forensic anthropology is to assist in medicolegal investigations, typically involving decomposed or skeletal remains. When the remains belong to an unidentified decedent, the first step is to construct a “biological profile,” in which the estimation of sex is a crucial component. Accuracy in sex estimation increases when both metric (e.g., the length of a bone) and morphological (e.g., the shape of a bone) data are used and when multiple skeletal elements are assessed. There are currently no methods for sex estimation that meet both of these criteria. Forensic anthropologists are often called upon to testify in legal proceedings and provide expert evidence; however, the field of forensic anthropology lacks an assessment of how sex estimation methods have been evaluated when the admissibility of such evidence is challenged, resulting in little guidance on which methods forensic anthropologists should use and how to develop improved methods. There are currently 13,885 individuals listed as “unidentified” in the National Missing and Unidentified Persons System, including 796 individuals for whom sex is listed as “unsure.” This research will produce a new method for sex estimation that meets legal admissibility standards and additionally can accommodate the wide range of remains that are recovered (e.g., only a skull versus a complete skeleton), with the ultimate goal of reducing the number of unidentified bodies in the United States.<br/><br/>This project will combine metric and morphological data from multiple skeletal elements into a single prediction model that will satisfy key legal considerations during evidentiary challenges, while improving accuracy and reliability. A framework for method development will be created that incorporates (1) legal perspectives on the admissibility of expert witness evidence and (2) forensic anthropology perspectives on methods of sex estimation to reduce barriers for adopting a new method. Using this framework, skeletal data will be collected and integrated into a prediction model, using the novel Mixed Cumulative Probit algorithm, and will be disseminated as a free, open-access web-based graphical user interface available to all forensic anthropologists. This innovative approach will result in a single, statistically valid sex estimation method that can accommodate both continuous and ordinal data, conditional dependence, heteroskedasticity, and missing data and therefore be used in the vast majority of forensic anthropology casework. By using legal admissibility criteria as guidelines, the resulting model will also be robust to evidentiary challenges.<br/><br/>This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.
