Modelling residential segregation as unevenness and clustering: A multilevel modelling approach incorporating spatial dependence and tackling the MAUP

Abstract

Traditional studies of residential segregation use a descriptive index approach with predefined spatial units to report the degree of neighbourhood differentiation. We develop a model-based approach which explicitly includes spatial effects at multiple scales, recognising the complexity of the urban environment while simultaneously distinguishing segregation at each scale net of all other scales. Moreover, this model distinguishes segregation as unevenness and as spatial clustering in the presence of stochastic variation. The modelling approach, unlike traditional index approaches, allows hypothesis evaluation concerning alternative scales and zonation through an accompanying badness-of-fit measure. Ultimately, this permits the identification of the scale and zonation regime where the spatial patterns come into focus thereby directly tackling the modifiable areal unit problem. The model is applied to Indian ethnicity in Leicester, UK, finding segregation as unevenness and as spatial clustering at multiple scales.