The 81st Annual Meeting of the American Association of Physical Anthropologists (2012)

Age-at-death estimation in bioarchaeology: does a uniform prior perform better than an informative prior in transition analysis?


1Anthropology, California State University Sacramento, 2School of Social Sciences, Humanities and Arts, School of Natural Sciences, University of California Merced, 3Anthropology, University of Tennessee Knoxville

Thursday All day, Plaza Level Add to calendar

Transition analysis offers a Bayesian statistical method for adapting skeletal aging techniques to the unique age-at-death structure of the population under study. An appropriate informative prior is always considered the best option, however, this may not be feasible with bioarchaeological samples. This context often requires the use of a uniform prior, which is usually considered undesirable. This paper compares transition analyses using a uniform prior and an informative prior on a documented historic Italian sample from Sardinia. We seek to 1) demonstrate the utility of transition analysis in bioarchaeological contexts, and 2) compare the performance of the informative prior, uniform prior, and the traditional Suchey-Brooks method.

Pubic symphyses were scored (n=390) on the Sardinian sample using the Suchey-Brooks method and subject to transition analysis. The Terry Collection approximated a uniform prior to which a Gompertz hazard model was fit. The informative prior consisted of documented individuals from Torino, Italy. The highest posterior density regions (HPDR) were generated from the hazard model parameters and ages-at-transition. In order to test the accuracy of the uniform and informative priors, cumulative binomial tests were run on the HPDR. The results indicate the informative and uniform priors worked similarly well, both out-performing the corresponding Suchey-Brooks age ranges. These findings suggest a uniform sample may be a suitable prior when an informative prior is unavailable. A uniform prior eliminates limitations bioarchaeologists face in locating an informative prior. Adoption of the transition analysis method will strengthen age-at-death estimations and, in turn, paleodemographic investigations.

The research was funded in part by a grant to Samantha M. Hens from the Office of Research and Sponsored Projects, California State University, Sacramento.

comments powered by Disqus