Article Spotlight: Older Collections and Meta Analyses

Our blog post series highlighting recent articles is back,  starting with the first of our summer publications, “The Promise and Peril of Older Collections: Meta-analyses and the zooarchaeology of Late Prehistoric/Early Historic Mexico” by Emily L. Jones and Caroline Gabe. Emily Jones is an Assistant Professor at the University of New Mexico, New Mexico, US.

In the past 5 years analyses rooted in “big data” have become ubiquitous in archaeology as elsewhere. Frameworks for successful conservation projects, the origins and spread of animal husbandry, the growth of London’s northern trade – all these topics and more have been the subject of recent meta-analyses. Such results are clearly breaking new scientific ground, but the plethora of such big studies has also highlighted concerns. Critiques of “big data” (for example, this 2014 opinion piece in the New York Times) have pointed out that when data included in analysis are collected with varying methods, spurious findings can result. A Harvard Business Review article title puts it succinctly: “Google Flu Trends’ Failure Shows Good Data > Big Data”.

My student and colleague Caroline Gabe and I both work with archaeological collections – that is, material recovered from excavations that took place sometime in the past. These collections are sometimes the result of relatively recent excavations, but sometimes they were recovered much longer ago, as far back as the early 1900s. The archaeologists excavating in the first half of the twentieth century used very different methods than we do now. In addition, older methods of curation can lead to the loss of some data, further biasing results. In short, archaeological meta-analyses involving both older and newer collections have the potential to be problematic for many of the same reasons that Google Flu Trends failed.

 

This specimen from the Comanche Springs collection was originally identified as a “skull” – even though all we found during reanalysis was bone fragments, roots, and dirt (photo by C. Gabe).

So when the opportunity arose for us to participate in a symposium on zooarchaeological meta-analyses, we leapt at the chance to explore how the simultaneous use of older and newer zooarchaeological data might impact analyses. Our Open Quaternary article, The Promise and Peril of Older Collections: Meta-Analyses and the Zooarchaeology of Late Prehistoric/Early Historic New Mexico, is the result. While (as our title suggests) we did find many potential problems in archaeological meta-analyses that involve both older and more modern collections, we also explored some ways to address these problems. The “promise” we identify begins with understanding the perils!