Doctor of Philosophy
Dr C.J. Ellis / Dr E. Molto
This study explores a field of spatial statistics known as Point Pattern Analysis (PPA) and its application in archaeology. The overall goal is to provide a resource which will guide and assist the reader in the proper application of PPA. Past archaeological applications are combined with more recent geographical and statistical mathematics to create a more inter-disciplinary, synthesized approach. Included are a discussion of analytical methods and two detailed case studies/applications. The study begins with an overview of PPA approaches in archaeology, starting with a general introduction and several commonly understood concepts such as first and second order effects and simple and labeled point patterns. It also describes options for calculating statistical significance and their appropriate uses which depend on the analysis being performed --something which is not well articulated in the literature. It goes on to describe appropriate techniques for analysis introducing another new concept called scale of analysis, which facilitates comparison of various statistics in the analysis of first order effects. Finally, it provides logical structured approaches to conducting a PPA and selecting appropriate statistics for various kinds of analysis including some refined and new routines. A series of PPA statistics developed in R are provided. The first case study analyzes the distribution of surface material in a 1.9 ha Davidson Archaic site in Ontario. An analysis of first and second order effects of the distribution of lithic debitage using multiple statistics leads to the conclusion that the Broad Point occupation represents an aggregation site with a series of similar clusters representing socially distinct groups of people. A second order analysis of the distribution of more formal artifacts shows a more complex deposition than the flake clusters. The second case study examines the distribution of discrete genetic traits in the Kellis-2 cemetery in Egypt evaluating the hypothesis that the cemetery was organized on a kinship basis and that male kin ties governed grave placement. In addition, it is shown that a lower than expected number of males in the cemetery is not spatially random but tends to occur more frequently in some of the kin-based groupings.
Keron, James R., "The Use of Point Pattern Analysis in Archaeology: Some Methods and Applications" (2015). Electronic Thesis and Dissertation Repository. 3137.