Zooarchaeological Survivorship Models using Ordered Logistic Regression

Summary

Archaeologists investigate past hunting and herding strategies using models of animal survivorship derived from long bone fusion and/or mandibular tooth wear patterns. As biological and behavioral variation makes estimating precise biological ages problematic, researchers typically assign "age stages" that describe ranked age groups. Ordered logistic regression models take advantage of the information in these rankings to estimate and analyze patterns in ranked/ordered data based on other variables. This poster describes fitting ordered logistic regression models to zooarchaeological survivorship data using Bayesian inference (via Stan) to (a) estimate uncertainty in survivorship estimates due to sample size and (b) compare survivorship between different sites and sub-assemblages within sites. The model is also able to incorporate uncertain stage assignments using aoristic analysis; as most zooarchaeological assemblages contain specimens that cannot be assigned to a single "age stage". The ability to use uncertain stages allows aoristic survivorship analyses to combine patterns of long bone fusion and mandibular tooth wear data, despite their variable specificity. The models shown here can be adapted to any archaeological situation that uses ordered or ranked variables.

This Resource is Part of the Following Collections

• Society for American Archaeology 83rd Annual Meeting, Washington, DC (2018)

Keywords

Spatial Coverage

Record Identifiers

Abstract Id(s): 20670