Key points are not available for this paper at this time.
The relationship between the index of residential differentiation (RDI), a recently introduced measure of residential segregation which is designed to summarize the distribution of c social groups in a single index, and the indices of dissimilarity (D) and segregation (SI) is examined. It is demonstrated that RDI is a natural extension of D, and that when c = 2, RDI = D; when c greater than 2, it is possible to stipulate RDI in terms of SI, but it is not possible to stipulate RDI in a straightforward manner in terms of D. An adjusted RDI is then suggested in which random segregation rather than complete desegregation is used as the baseline.
Morgan et al. (Fri,) studied this question.