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

A comparison of aging methods and hazard analyses using a skeletal sample from the Larsen Village Cemetery (39WW2) in South Dakota


Department of Anthropology, University of Tennessee, Knoxville, Central Identification Laboratory, JPAC

Thursday All day, Park Concourse Add to calendar

Age estimation methods utilized in Paleodemography have long been criticized for their inability to accurately reflect the population being studied. Milner and Boldsen (2012) argue that there are few methods to systematically and quantitatively combine information from independent features within a single skeleton, and that age estimations without terminal ages or that do not reflect the older individuals in a population produce bias age structures. The age-at-death estimation method chosen has the potential to change the hazard analyses reliant on accurate estimations.

This study examines the adult mortality profile for the protohistoric, Postcontact Coalescent, Larsen Village Cemetery (39WW2) from South Dakota (N=197). The purpose of this study is to demonstrate how different aging techniques affect the mortality profiles using traditional aging methods (e.g. Lovejoy et al. 1985), transition analysis, and the subjective/experience-based approach suggested by Milner and Boldsen (2012).

Results from the Gompertz and Gompertz-Makeham hazard models suggest a similar demographic profile for the Larsen Village Cemetery as suggested by Owsley et al. 1979. The risk of death for young adult females and individuals over 50 years of age is high. An increase in older individuals (70-80 years old) is suggested when transition analysis is used. Ironically, there were significant correlations between the different age-at-death methods utilized, which supports Wilson and Steadman’s (2007) findings using the Orendorff skeletal sample. Also, these findings support Milner and Boldsen’s (2012) suggestion that experience based methods may fair no better or worse than maximum likelihood estimations derived from transition analysis.

comments powered by Disqus