Modelling residential segregation: an empirical and simulation-based study

This research study aims to take advantage of the potential of computer simulation as a new methodological approach to modelling residential segregation and measuring the variables of interest. Simulation applied to census data will be used as a novel approach to test the proposed model.

Census data provides rich demographic information and it is the intention of the proposed research to exploit this data resource to validate simulation model-based estimates of segregation patterns over time. A series of simulated longitudinal investigations will be made, following through each Census cohort, to test the simulated prediction of the models, and hypotheses about the processes and choices behind residential segregation.

The causes, patterns, and consequences of urban residential segregation by ethnicity and socio-economic status are important topics in contemporary urban geography and sociology. Segregation is important because of its potential impacts on issues such as access to education and healthcare opportunities. However, there are significant methodological challenges faced by any study of changing patterns of residential location: for example, measuring static spatial patterns of segregation is difficult, and neighbourhood change is in any case a process not merely a succession of static outcomes. Ponsonby, in Auckland, for instance, was, until even relatively recently, a residential area with a strong presence of Pacific Peoples. Today, it is a totally transformed social landscape.

Studies that can describe and explain the details of such processes of transformation, at the level of both individual household decision making, and of aggregate demographic shifts over time, are of great value for urban geographers and sociologists in understanding neighbourhood formation and social stratification.

1. To explore linkages between the social micro (household level) and macro levels (social structure expressed through demographic patterns in geographical space), and to operationalise these linkages in a simulation model.
2. To demonstrate the usefulness of simulation modelling as an approach to understanding the dynamics of urban neighbourhood change.
3. To develop an approach to the transparent verification of simulation models of residential segregation using detailed census data.
4. To describe and measure past, present, and possible future patterns of residential location by ethnicity, focusing particularly on change in New Zealand urban (especially Auckland) settings.
5. To establish census datasets as potential ‘test-beds’ for future modelling-based research which can address theoretical issues in urban geography and sociology.