Abstract:
ABSTRACT
UNCERTAINTY IN SKELETAL AGING: A RETROSPECTIVE STUDY
AND TEST OF SKELETAL AGING METHODS AT THE JOINT
POW/MIA ACCOUNTING COMMAND CENTRAL
IDENTIFICATION LABORATORY
by
Carrie Ann Brown
Master of Arts in Anthropology
California State University, Chico
Spring 2009
Adult skeletal age estimation is an important facet of forensic anthropology,
paleodemography, and bioarchaeology. Estimating the age-at-death of adults is problematic because of human variability in the aging process. Analysis of the error associated with skeletal age estimation methods is necessary so that the performance of these methods is not overestimated and so that the uncertainty in these skeletal techniques can be quantified and better understood.
The purpose of this thesis is to analyze and describe the error associated with skeletal age estimation methods used at the Joint POW/MIA Accounting Command Central Identification Laboratory (JPAC/CIL) from 1972 to 31 July 2008. There
Were six general categories of age estimation methods used: epiphyseal fusion, suture closure, dental formation and eruption, and morphological changes in the pubic symphysis, auricular surface, and sternal rib end. The total identified known age-at-death sample was 979 individuals, although method sub-samples were much smaller. Additional interobserver error research was conducted with three methods that were problematic for the JPAC/CIL sample.
Results indicate that adult age estimation methods perform well for the JPAC/CIL identified known age-at-death sample, most likely because of the young age composition of this sample. Bias, inaccuracy, and scaled error index (SEI) values are low for most methods and phases or stages of methods. Correlation between estimated and known age-at-death is statistically significant for maxillary suture closure, pubic symphysis, auricular surface, and sternal rib end methods. The auricular surface is the poorest age indicator of those examined in the JPAC/CIL sample. It is also recommended that fusion of the sacral segments no longer be used for age estimation since this method had a correct classification rate of only 32.1%. Future research in adult skeletal age estimation and refinement of existing techniques should include estimation of measurement uncertainty.
