The 82nd Annual Meeting of the American Association of Physical Anthropologists (2013)


Estimating age at death in bioarchaeology: A Rostock approach

SAMANTHA M. HENS1 and KANYA GODDE2,3.

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 Evening, Park Concourse Add to calendar

The “Rostock Manifesto” developed from a paleodemography workshop in Germany in 1999. Participants concluded that critical components to a sound theoretical approach to age estimation required Bayes’ Theorem and assessing the probability distribution of the age at death in the target population. Transition analysis, coupled with a Bayesian approach, has been suggested to better reflect the senescence changes and increase the accuracy of skeletal age estimation (Boldsen et al., 2002). However, due to the limited availability of appropriate data to use as informative priors, too few studies have been carried out in bioarchaeology.

In this study, Suchey-Brooks pubic symphysis age phases were scored on a male target sample of documented historic Italians (n=202). The resulting age estimates from this traditional approach were compared for accuracy to the estimates derived from transition analysis combined with a Bayesian approach using both uniform and informative priors. A Gompertz hazard model was applied to a comparable historic Italian sample to generate an informative prior age at death distribution. The hazard parameters were coupled with published ages of transition from the Terry Collection and Balkan genocide victims to calculate the highest posterior density regions for each pubic symphysis phase. The subsequent cumulative binomial tests at 50%, 75%, 90% and 95% indicated that the Bayesian approaches outperformed the traditional Suchey-Brooks method, demonstrating a higher probability of success at each level.

This paper strongly supports the use of transition analysis with Bayesian statistics to estimate age in archaeological and forensic contexts.

The data collection was facilitated by a grant for Samantha M. Hens from the Office of Research and Sponsored Projects, California State University, Sacramento.

Tweet
comments powered by Disqus