Statistics in archaeology

Abstract

The basic statistical tools described above provide for fundamental quantitative description and comparison, for establishing the confidence with which the samples available permit characterizations of the populations from which they come, and for assessing the strength and significance of the relationships between pairs of variables. More complicated statistical tools for multivariate analysis, which cannot be summarized here, are also often used in archaeology: cluster analysis, factor analysis (or principal components analysis), discriminant analysis, multidimensional scaling, multiple regression, and others. As in many fields where statistical analysis is applied, examples of valid and insightful use of statistical tools can be found in archaeology, along with serious misunderstanding of statistical results. It is clear that archaeologists have not always fully grasped the meaning of statistical significance (and its mirror image, confidence) or entirely understood how this notion differs from strength. In concluding summary, then, strength starts with what can be observed in samples: Is there much relationship between two variables in the samples that we have? How much difference is there between two categories in the samples that we have? Confidence and significance concern the inferences we make about the populations our samples come from. They tell us how confident we should be that our samples are large enough to sustain the conclusions we would like to make (based on the strength of what we observe in them) about the populations from which they come. © 2008 Copyright © 2008 Elsevier Inc. All rights reserved.