Zooarchaeology in the era of big data: contending with interanalyst variation and best practices for contextualizing data for informed reuse

Abstract

New digital publication technologies facilitate the publication of primary data and increase the ease with which archaeologists are able to share, combine, and synthesize large datasets. The research prospects that these technologies make possible are exciting, but they raise the issue of how comparable the original datasets really are. In this study we demonstrate an issue associated with many archaeological datasets: interanalyst variation. We conduct two independent analyses of one zooarchaeological assemblage and compare data. We consider the implications of the challenge interanalyst variation poses within projects and across projects. We then make recommendations for zooarchaeologists specifically, and for archaeologists more broadly, who are interested in publishing primary datasets in order to improve future understanding of these data and facilitate their reuse. These recommendations include specific guidance of what information needs to be published along with primary datasets to facilitate their responsible reuse in other projects, recommendations for incorporating interanalyst variation studies into research programs, and suggestions about what to do should analysts discover systematic biases in their analyses stemming from interanalyst variation.