A reappraisal of the role of spatial science and statistical inference in geography in Britain

Abstract

It is argued that recent criticisms of the tradition of spatial science in Anglo-American geography have been broadly correct. To overcome these criticisms it is suggested that spatial science requires much better theory and techniques to differentiate spatial process (which is argued to be the prime centre of interest) from structures of data arising from map pattern and spatial structure. This in turn leads to the conclusion that the preoccupation with statistical inference must be replaced by a wider concern with confirmatory (e.g. Bayesian) and exploratory modes of analysis. This, it is argued, can be achieved only by a recasting of geographical methodology to give a better understanding of spatial process. This in turn requires a reestablishment of the core of geographical enquiry as a social view of spatial science.