Univariate problems: the modifiable areal unit problem

Abstract

We have already discussed in Chapters 1–3 the problem of the modifiable unit. The difficulties created by modifiable units are noted by Yule and Kendall (1957; p.310) and for spatial problems have been termed, following Openshaw (1977) the “modifiable areal unit problem”. It is worth recalling here that the spatial problem consists of two aspects: the first concerns the different results obtained with the same set of data grouped at different scale levels (e.g. individuals, counties, regions). This is referred to as scale problem. The second considers the variability of results not due to variations in the size of the areas, but rather to their shape and it is therefore referred to as aggregation problem. In Chapter 4 we put forward a methodological framework that enables us to discuss the modifiable areal unit problem in the context of the theory of stochastic processes. In the present chapter we introduce the detailed analysis of the problem.