The centralization index: A measure of local spatial segregation

Abstract

Segregated areas may occur around an attractive park or a waste incinerator, but the magnitude and group membership of the people in closest proximity will likely be difierent. We therefore introduce a local segregation measure that can be applied to any location within a metropolitan area, and that can identify the group that is relatively more concentrated around that reference location. We further introduce an inference approach to identify the statistical significance of a particular segregation value. In an exploratory setting the index can be used to generate a map of hot spots, and seed the question: “why is this group significantly concentrated around that location?”

Is Poverty Decentralizing? Quantifying Uncertainty in the Decentralization of Urban Poverty

“…. to identify the centre" (Folch and Rey 2016). Centralization measures remain a useful measure, however, not least because they offer a simple, transparent, and replicable measure well suited to intercity comparison and can be a useful way to monitor changes to city structure.…”

“…Measuring changes to the evenness of poverty across aerial units is typically measured using the dissimilarity index (DI). Folch and Rey (2016) also proposed using the RCI as a local measure of segregation. We apply both methods here and address the respective issues of inference.…”

“…How do we know whether an apparent shift in the suburbanization of poverty is not simply an outcome of random events (e.g., population churn) or measurement error (particularly problematic if suburban boundaries are also changing)? In the context of ethnic residential mix, Folch and Rey (2016) provided a simple procedure for estimating whether a particular RCI value is statistically different to zero. This is essentially a distance from randomness measure, and does not help us address the question of whether observed change over time is statistically significant.…”

“…Folch and Rey (2016) recently proposed the RCI as a local measure, applied to each and every aerial unit, to give a moving window of the locus of relative position of one group relative to another. The RCI does not replace the richness of bespoke suburbanization approaches.…”

“…This is highly relevant not only for the provision of welfare services but for some of the neighbourhood effects mechanisms that assume or rely on spatial concentration (Galster 2007;Manley et al 2013). Folch and Rey (2016) demonstrated how relative centralization measures can be usefully applied to capture local spatial concentration. Again, though, as far as we are aware, robust statistical inference has yet to be developed for measuring changes to the local centralization of poverty.…”