Geographical units of analysis and the analysis of crime

Abstract

When spatial analysis of crime is conducted, the analyst should not ignore the spatial units that data are aggregated into and the impact of this choice on the interpretation of findings. Just as several independent variables are considered to determine whether they have statistical significance, a consideration of multiple spatial units of analysis should be made as well, in order to determine whether the choice of aggregation level used in a spatial analysis can result in biased findings. This chapter considers four classes of problems that can arise when data bounded in space are analyzed. These problems, inherent in most studies of space, include: issues associated with politically bounded units of aggregation, edge effects of bounded space, the modifiable aerial unit problem (MAUP), and ways in which the results of statistical analyses can be manipulated by changes in the level of aggregation. Techniques that can be used to alleviate each of the methodological difficulties described in this chapter are then discussed. © 2009 Springer New York.