A Monte Carlo Approach to Estimating the Significance of Segregation

Abstract

Simulation techniques are not new to geography, but they have focused traditionally on attempts to replicate observed patterns rather than on seeking to assist in the formulation of a null hypothesis. The benefits and problems of a Monte Carlo simulation methodology are examined with reference to the latter objective. The flexibility of the approach in the evaluation of spatial statistics is shown to be considerable. Specific case studies are pursued which exemplify its potential in the investigation of social segregation using familiar measures such as the dissimilarity index and the P* index.