Designing databases and using generalized estimating equations for entheseal datasets: An example from the Tiwanaku culture (AD 500–1100)

Abstract

AbstractStudying entheseal changes (EC) in human skeletal remains involves questions surrounding how researchers should collect, process, and evaluate these data as there are no set standards. Osteoarcheological research should also be able to answer population‐level queries using entheses, such as if group A was moving their body in different ways from group B? Or are there age‐related tasks or gendered labors? However, not all entheseal data can be easily evaluated in this fashion. If researchers select one area of the body, they may not be able to discuss population‐level differences. In addition, if data are evaluated by each muscle attachment area throughout a body, the results can be overwhelming. Further, grouping EC may produce problems with statistical assumptions of independence. To address these design and scalar issues, we discuss proper database construction, including the importance for consistent data collection strategies and in anchoring individuals under a specimen number. We also show how generalized estimating equations (GEE) can address how individual‐level scores can be collected and population‐level research questions can be answered with entheseal marker data processed in SAS 9.4 or the free alternative, R. We utilize a sample of over 1200 adults from the Tiwanaku culture (AD 500–1100) of Bolivia and Peru. We demonstrate evaluating significant overall differences while also pinpointing specific EC by sex or by age at death to discuss various ways in which bodies moved in the past.