Age-at-death estimation in an Italian historical sample: a test of the Suchey-Brooks and transition analysis methods

A growing body of research is demonstrating increased accuracy in aging from a relatively new method, transition analysis. Although transition analysis was developed for paleodemographic research, a majority of subsequent studies have been in the forensic arena, with very little work in bioarchaeological contexts. Using the Suchey-Brooks pubic symphysis phases, scored on a target sample of historic Italians from the island of Sardinia, we compare accuracy of aging between transition analysis combined with a Bayesian approach and the standard Suchey-Brooks age ranges. Because of the difficulty in identifying a reasonable informative prior for bioarchaeological samples, we also compared results of both an informative prior and a uniform prior for age estimation. Published ages-of-transition for the Terry Collection and Balkan genocide victims were used in conjunction with parameters generated from Gompertz hazard models derived from the priors. The ages-of-transition and hazard parameters were utilized to calculate the highest posterior density regions, otherwise known as "coverages" or age ranges, for each Suchey-Brooks phase. Each prior, along with the parameters, were input into cumulative binomial tests. The results indicate that the Bayesian approach outperformed the Suchey-Brooks technique alone. The Terry Collection surpassed the Balkans as a reasonable sample from which to derive transition analysis parameters. This discrepancy between populations is due to different within phase age-at-death distributions that reflect differences in aging between the populations. These results indicate bioarchaeologists should strive to apply a Bayesian analysis when aging historic and archaeological populations by employing an informative prior.