Keywords

archaeology, computer-intensive methods, quantitative methods

Abstract

Among Keith Kintigh’s many contributions to archaeology was his emphasis on understanding the connections among quantitative methods, archaeological problems, and what archaeologists can reasonably infer from their data. In both publications and in the classroom, he demonstrated the value of simple computer simulations to understand quantitative measures and how they behave when applied to actual archaeological data. Archaeological research increasingly incorporates analysis of large databases and quantitative methods appropriate to “big data”, but simple statistics are still important to archaeological research. This paper uses computer-intensive methods to demonstrate that archaeologists (and others) frequently misunderstand and misapply some of the simplest statistics used in archaeology: chi-square tests and related methods for examining cross-tabulated data (e.g., artifact counts by provenience). These tests rely on assumptions about sampling that are usually not appropriate for archaeology, and, when applied to archaeological data in textbook fashion, they often give wildly misleading results. Computer simulations provide a better understanding of the issues involved, and of how to use these (not so) simple statistics to provide valid insights about archaeological questions.

Original Publication Citation

James R. Allison 2019 Simple Statistics and Archaeological Problems. Paper presented at the 84th Annual Meeting of the Society for American Archaeology, Albuquerque, New Mexico.