Geostatistics as measures of spatial segregation

Traditional measures of segregation, such as the index of dissimilarity, have been criticized as aspatial in nature. Spatial measures of segregation have been proposed, but they are difficult to use. Based on the idea that segregation implies a spatial separation of ethnic groups, the degree of spatial correlation among groups can reflect the level of segregation. This paper suggests that several geostatistical measures, especially the standard deviational ellipse, are effective tools for capturing the spatial characteristics of a population group. By comparing the ellipses of different groups, measures of segregation can be derived. The paper demonstrates this approach to measuring segregation by way of both a simulation and a case study.