THE DEVELOPMENT OF SEX ESTIMATION METHODS IN FOUR PREHISTORIC NATIVE AMERICAN SKELETAL SAMPLES

Author

Kyle Douglas Lubsen, Southern Illinois University CarbondaleFollow

Date of Award

5-1-2012

Degree Name

Doctor of Philosophy

Department

Anthropology

First Advisor

Corruccini, Robert

Abstract

Sex estimation in bioarchaeological research is paramount for comparative analyses of skeletal remains and developing estimates of other demographic variables. Frequently, sex is estimated utilizing the morphology of pelvic and cranial bones in archaeological skeletal samples. Regrettably, these bones are often damaged, destroyed, or lost and cannot be employed for the estimation of sex. Fortunately, a variety of metric and visual sex estimation methods have been developed on modern skeletal samples with known demography. Disappointingly, due to the population specific nature of many of these metric methods, they cannot be accurately applied directly to alternate skeletal samples. However, these methods can be redeveloped and retested for archaeological skeletal samples if the proper protocol is utilized. This research utilizes the protocol for developing methods of sex estimation on samples with unknown sex developed by Murail et al. (1999). Utilizing select hand and foot bones from four prehistoric Native American skeletal samples from Alabama and Illinois, multiple discriminant functions were developed and tested on both Archaic and Mississippian Period skeletal samples. Furthermore, the four individual site samples were combined into two geographical and two temporal samples, as well as an all pooled sample in order to test the broader applicability of these methods. The results indicate that sex estimation methods can be developed on samples where sex is not known. Additionally, the discriminant functions developed produce high levels of classification for the sites, individually, as well as for the geographic, temporal, and all pooled samples. The latter suggests these functions have broader applicability for these regions and temporal periods. Moreover, the concordance rates for these functions are similar to the accuracy rates when these bones were applied to samples with known sex.