An Examination of the Relationship Between Data Recording Strategies and Intrasite Spatial Analysis: San Xavier Archaeological Project

During the course of the San Xavier Archaeological Project over 18 square miles in the southern Tucson Basin were intensively surveyed. This work resulted in the recording of 150 sites, of which 147 contained components dated to either the prehistoric or protohistoric periods. In a previous study (Altschul and Rose, Statistical Research Technical Series 3) block cluster analysis was used to derive a site classification. Each site with a prehistoric and/or protohistoric component was classified on the basis of a set of categorical artifact and feature variables into one of six site classes. Although the analysis reduced the data into a manageable set of analytic categories, it sidestepped a fundamental issue: exactly what is a site?

At first glance this question may seem trivial. After all, sites are locations where artifacts and other evidence of past behaviors are present. Such a "common sense" definition, however, is not really very useful, for a site can be a single artifact or an entire ancient city. Moreover, a site can represent a single occupation of an area by one social group, repeated occupations by the same social group, or overlapping occupations by different social groups.