RESEARCH NOTE—MEASURING ETHNIC RESIDENTIAL SEGREGATION: PUTTING SOME MORE GEOGRAPHY IN91

Abstract: Most studies of urban ethnic residential patterns rely on various single-number indices to demonstrate the degree of spatial segregation. These have been criticized on a variety of grounds, and various other approaches have been proposed, including the use of measures of statistical autocorrelation and typologies of areas based on their population composition. These alternatives provide a greater geographical appreciation of segregation than the indices. It is argued here—using Auckland, New Zealand as a case study—that their integration could substantially increase our evaluation of segregation levels. [Key words: ethnic segregation, measurement, local statistics, Auckland.] There is a massive literature on ethnic residential segregation in urban areas across many countries (see recent reviews in Kaplan and Woodhouse, 2004, 2005; Reardon, 2006). Much of it provides estimates of the degree of segregation of individual groups, almost all using single-index measures—some of ever-increasing complexity and unin-terpretability. Our contention is that, useful though they are, such indices conceal as much as they reveal, if not more. A fuller, more informative discussion of segregation levels needs more geography. Some have started doing just that in recent years, but we conten