A Bayesian Approach to Investigating Age-at-Death of Subadults in a Forensic Context

Abstract

Estimating age at death is among the first steps in the identification of an unknown individual. For subadults, dental formation stage remains the most accurate aging indicator due to minimal environmental impact. Even the most accurate method, however, is affected by "mimicry bias," where the age profile of the target population "mimics" the age profile of the reference population used to develop the age estimation method. Bayesian statistics and transition analysis can control for this bias in archaeological and forensic samples through calculating the average age of transition from one development phase to another, followed by estimating the probability that someone of a certain age has a given phase of development based on a sample of individuals of known age. Here, Bayesian-derived age ranges related to the dental formation phases of Moorrees et al. (1963) were generated using a sample of 201 children of known age (Orthodontics Case File System, Maxwell Museum of Anthropology). In this study, I present age ranges at an exact 50% probability for each tooth at various stages of development. These ranges can be used in forensic cases wishing to control for "mimicry bias" in assessments that rely on sage estimation via dental formation.