ABSTRACT
The application of local spatial statistics (Anselin 1995; Getis and Ord 1992; Ord and Getis 1995), as discussed in Chapter 4, is rapidly becoming more widespread in the field of spatial analysis. Often there is interest in searching for significant spatial associations with no or little prior information about their location. The third category of tests in Besag and Newell’s classification, namely, tests for the detection of clustering, serve this purpose and essentially carry out multiple local tests of the null hypothesis of no raised incidence across the entire study region. These tests “scan” the study region to find subareas that constitute spatial clustering over and above what could be expected by chance alone. Often, such scanning is done not only across all locations within the study area, but also at different spatial scales.
