One-Sample Tests in Regional Archaeological Analysis: New Possibilities through Computer Technology

Archaeologists commonly employ two-sample statistical tests in regional locational analyses that compare environmental measurements obtained at site locations against measurements taken at random locations from the background environment. One-sample tests that compare a site sample against a background standard are conceptually and statistically superior, but have been difficult to implement for continuous data types. This situation now is changed owing to a relatively new computer technology known as Geographic Information Systems (GIS). GIS can provide a complete description of the nature of the background environment of entire regions for categorical and continuous data types, thereby allowing the ready application of one-sample testing strategies. Examples of several GIS-based one-sample tests are given using data from east-central Arizona. Such conventional tests only should be applied, however, when the observations can be shown to be statistically independent through tests for spatial autocorrelation.