Year of Award


Document Type


Degree Type

Doctor of Philosophy (PhD)

Degree Name


Department or School/College

Department of Anthropology

Committee Chair

Ashley McKeown

Commitee Members

Randall Skelton, Meradeth Snow, Lyle Konigsberg, Brian Steele


Forensic anthropology, Age estimation, Pelvis, Parity


University of Montana


Estimating age from skeletal remains provides a critical component of the biological profile of an individual. To date, there are different methods used on select parts of the skeleton to assess the age of the individual, and currently, the pelvis is relied upon heavily to obtain accurate and reliable age ranges (Berg 2008; Brooks and Suchey 1990; Buckberry and Chamberlain 2002). Many have stated that age related changes follow different trends in males and females, with parity presented as one of the possible causes for such differences (Berg 2008; Meindl and Lovejoy 1989; Resnick and Niwayama 1998; Suchey and Katz 1998). There is reason to believe that parturition may increase the rate at which the areas of interest of the pelvis degenerate. However, this hypothesis has yet to be formally tested on a recent skeletal collection. The purpose of this study, therefore, is to assess the effects of pregnancy and parturition on the pubic symphysis and auricular surface and determine whether it influences the morphology of the pelvis enough to effect the physiological age of the individual.

Data were collected from the William M. Bass Skeletal Collection located at The University of Tennessee, Knoxville. The study contained 434 individuals (males: 234/females: 200/parous: 157/nulliparous: 43). This is a collection of modern forensic skeletons with known age at death, ancestry, sex, and medical background. The features of the pubic symphysis were noted and matched with the best fitting phase in both the Suchey-Brooks and Todd pubic symphysis scoring systems (Brooks and Suchey 1990; Buikstra and Ubelaker 1994; Todd 1921). Next, the features of the auricular surface were noted and matched with the best fitting phase in the system presented by Lovejoy and colleagues and were individually scored resulting in a composite score following the method proposed by Buckberry and Chamberlain (Buckberry and Chamberlain 2002; Meindl and Lovejoy 1989; Lovejoy et al. 1985). Time was designated at the beginning of the second and third days to employ the test-retest method to calculate the intraobserver error rate and ensure reliability of these assessments.

In this study, a statistical comparison was made between females who have given birth and those who have not to determine whether this process affects the rate of degeneration of the areas of interest of the pelvis. The male sample was used as a control for these age assessment techniques. A transition analysis, also known as a cumulative probit analysis, was conducted on the data in order to establish the age-at-transition distributions between the stages of each age estimation method. The results were then compared between the males and the nulliparous and parous female groups using likelihood ratio tests. The purpose for this comparison is to observe whether the age-at-transition distributions differs between sexes and/or the two groups, with the focus being on whether the parous group illustrates an difference in rate of degeneration, or transition to subsequent observable stages, when compared to the nulliparous and male groups.

The data was entered into the statistical software program R version 3.0.2. The transition analysis produced significantly different results between parous and nulliparous females using the pubic symphysis but not the auricular surface when the likelihood ratio test were taken into account. The male group and the nulliparous female group transition around the same age, while the parous females transition at an earlier age. An investigation was made into the implications these findings have for age estimation. The current research suggests that parturition affects the ! "#! pubic symphysis more so than the auricular surface when determining age at death. While the error rate was found to be within the normal range, a suggestion may be made to use the auricular surface age indicators when assessing age of females presumed to be parous.

The present results are significant in that they represent a noteworthy find within the realm of forensic anthropology and similar professional fields that may utilize this information on modern cases from the United States today. The current project is addresses several issues by employing transition analysis to 1) utilize an appropriate age-at-death distribution for the reference sample, 2) combat the issues inherent in linear regression analyses including biased age estimations and lumping older individuals into terminal categories such as 50 years plus, and 3) generate appropriate age-at-transitions for females who have had children, given their rate of degeneration of pelvic age indicators increase due to such a naturally traumatic event. A large part of forensic anthropology is providing a positive identification of the individual from the recovered remains. Inaccurate age determination of a skeleton can result in the remains being misidentified or unidentified. This project indicates that parturition affects the pubic symphysis and not the auricular surface when estimating age at death. This may lead to future estimations or correction factors that allow age estimation to be done more reliably.



© Copyright 2014 Rosanne Bongiovanni