Interpretational Limitations of Global and Local Indices in the Measurement of Residential Isolation

The P* index has been commonly referred to as the “isolation” index in segregation studies because it isolates (more generally, separates) the distributional pattern of a particular group from other mutually exclusive groups. Since the P* index has been conceived to possess useful mathematical properties for capturing the exposure dimension of residential segregation, its spatial versions have been developed by different researchers. However, these isolation indices may be susceptible to differences in population size and their upper-bound has been known to vary across mutually exclusive groups. To shed light on such interpretational limitations, which have not been clearly demonstrated in the literature, population estimates by race and ethnicity in the State of California at three time periods were assessed in this study. Whether aspatial or spatial in nature, both global and local isolation indices showed that racial and ethnic majority groups were more isolated (or, more generally, “segregated”) than their minority counterparts. For comparative and trend analyses, therefore, further considerations are needed to address the interpretational limitations of aspatial and spatial isolation indices by maintaining the essential meaning of isolation in segregation studies (i.e., the extent of group members’ exposure to the same group).