American Association of Physical Anthropologists 80th Annual Meeting
JAMIE M. GOMEZ and BRIDGET F. B. ALGEE-HEWITT
The estimation of age-at-death with some degree of statistical certainty is a key component of any osteological analysis that seeks to extract biological information from an individual or reveal the demographic structure of a skeletal population. Transition Analysis (TA) has emerged as a valuable statistical tool for obtaining robust age estimates that circumvents many of the statistical problems that plague other, traditional, methods. Moreover, the multiple-trait, component scoring approach associated with TA (Boldsen et al. 2002) is known to be a promising alternative to the single indicator methods when dealing with unknown, incomplete, and fragmentary cases. For these reasons, we believe that the TA approach merits more refined testing. This project evaluates the effect of education and experience on the age estimates obtained. We are concerned, specifically, with the ease of learning and repeatability when TA is applied by a practitioner who is familiar with age estimation theory but has minimal TA experience. We pose the following questions: (1) how interpretable are the criteria? (2) are some components more difficult to score than others? (3) how repeatable, and, so, reliable are the age estimates produced? (4) do TA scores, and, in turn, age estimates improve with practice? (5) are the patterns of intraobserver error and repeatability consistent between male and female specimens? Data for this study was collected on the Suchey-Brooks pubic symphysis casts: specimens were assigned unbiased identifiers, selected at random and repeatedly "blind"-scored over a two month period. Special attention was paid to the date and conditions of the analysis in order to detect patterned results and tests of intraobserver variation were conducted. Pedagogical and practical-use recommendations are provided.