Evaluation of Osteometric Measurements in Forensic Anthropology

As submitted by the proposer:
Many methods and techniques in forensic anthropology employ osteometric measurements as their basis, but little work has been done to investigate the error rates associated with these measurements. This research will provide the forensic anthropology community with error rates for standard osteometric measurements (intraobserver repeatability and interobserver error), use these error rates to evaluate the efficacy of these measurements, and evaluate alternatives for problem measurements. Particular focus will be given to the measurements that interface with the Fordisc computer program used by forensic anthropologists to estimate stature, sex, and ancestry. These error rates will provide a foundation for forensic case analyses and research. As such, this research addresses the recommendations put forth in the National Academy of Sciences Report Strengthening Forensic Science in the United States: A Path Forward (2009), in which the authors stressed the need for "rigorous systematic research to validate the discipline's basic premises and techniques".

Data collection will be done on the William M. Bass Donated Collection housed at the University of Tennessee. Four observers will take 78 standard measurements (34 cranial and 44 postcranial) and 20 additional measurements on a sample of 50 skeletons using standard osteometric equipment (sliding calipers, spreading calipers, an osteometric board, and a mandibulometer). These skeletons will be measured five times by each observer to establish interobserver and intraobserver error rates. The following elements will be measured: cranium, mandible, clavicle, scapula, humerus, radius, ulna, femur, tibia, fibula, os coxa, sacrum, and calcaneus. A number of statistical methods have been employed in the literature for examining and quantifying measurement accuracy, precision, and bias, including mean absolute difference, absolute and relative technical error of measurement (TEM), coefficient of reliability (R), coefficient of variation, percent error, scaled error index (SEI), two-way ANOVA without replication, t-tests for paired comparisons, Fisher's sign test, and nested ANOVA (Adams and Byrd 2002; Barnes and Wescott 2007; Mueller and Martorell 1988; Muñoz-Muñoz and Perpiñan 2010; Utermohle and Zegura 1982). Each of these statistics will be calculated as appropriate in order to compare the results of this comprehensive study with some of the results in the published literature.

The results will be disseminated to the forensic community in the form of a new edition of a widely-used laboratory manual designed to interface with the Fordisc program (Data Collection Procedures for Forensic Skeletal Material), continuing education workshops on osteometric techniques, and a free instructional video. The updates to the Data Collection Procedures (DCP) manual will include tables with measurement error information, photographic images demonstrating how each measurement is to be taken, and clarifications of some measurement definitions; DCP formatting will be done by a graphic artist. This information will also be made available to the authors of the Fordisc program for inclusion in the "Help file" of the program. The video demonstrating measurement techniques will be made available to the scientific community via YouTube in order to maximize the dissemination of the grant results and offer free training to practitioners who are unable to attend continuing education workshops. Each year semi-annual progress reports will be submitted to the program officer at the end of months 6 and 12. In addition, a final cumulative technical report will be submitted for review 90 days prior to the end of Year 2.

ca/ncf