The 81st Annual Meeting of the American Association of Physical Anthropologists (2012)


Let's get real about MSMs: reliability in scoring techniques

C. BRADY. DAVIS1, KRISTRINA SHULER2, MARIE E. DANFORTH1 and KELSEY HENDRON2.

1Department of Anthropology and Sociology, The University of Southern Mississippi, 2Department of Sociology, Anthropology, and Social Work, Auburn University, 3Department of Anthropology and Sociology, The University of Southern Mississippi, 4Department of Sociology, Anthropology, and Social Work, Auburn University

Thursday All day, Plaza Level Add to calendar

Studies of musculoskeletal stress markers (MSM) generally note relatively low rates of interobserver error. However, levels of error could vary depending upon the selection of attachment site and/or experience of observers. All could potentially affect the quality of MSM data and have notably negative effects on interpretation of activity patterns, especially given the small size of many taphonomically compromised bioarchaeological samples. This study assesses both intra and interobserver reliability in MSM scoring using 58 individuals from two Late Woodland/Mississippian skeletal series from west central Alabama. Using two primary scoring methods [(Hawkey and Merbs (1995) and Villotte et al. (2010)], seven evaluators representing a range of experience levels (undergraduate to PhD) recorded MSM expression at 17 muscle attachment sites on the long bones. Results strongly suggest that variability in scoring patterns were specific to MSM sites using both methods. Attachment points such as supinator and popliteus that generally exhibited little morphological remodeling had overall high agreement among all scorers. Other sites, such as deltoideus and quadriceps femoris, that displayed greater morphological development showed a broad range of scores, with up to 20% of sites having scores assigned to non-adjacent categories on Hawkey-Merbs' three-point scale. No pattern by level of experience was evident. In contrast to interobserver results, intraobserver error levels were generally low. These findings suggest that MSM data must be carefully interpreted, especially concerning site selection, and data comparison across researchers, since this could adversely affect interpretation of specific activity patterns.

Tweet
comments powered by Disqus