Researchers
Abstract:
This project will offer a scientific way of understanding the level and direction of change in residential segregation.The project will undertake three major strands of quantitative research:
Analyse marginal changes on three major dimensions of segregation - Index of Dissimilarity, Index of Isolation and Index of Diversity - due to changes in population definition, data quality and changing boundaries over time.
Two sources of data will be employed: census output as published, available to academic researchers through the CASWEB on-line interface; and 1991 and 2001 consistent mid-year population estimates as one of the statistical outputs of my doctoral thesis. The results are likely to reveal significant differences in measuring segregation and diversity of ethnic groups over time between the two data sources, and separately for 2001 local authority districts, and constituent Super Output Areas (SOA) and wards in England and Wales.
Compare segregation and diversity over time using a consistent population time series for a specific cohort. Although these indices of segregation have straightforward interpretations, they are affected by population growth. This is expected to be particularly significant with the Index of Isolation.
The proposed method will focus on the population by ethnic groups aged 10 and over in 2001. The measure of segregation of a single cohort will reduce the effect on the indices of population growth. Although it would assume that there are no mortality differentials between ethnic groups between 1991 and 2001, the benefits for this comparison are expected to outweigh such disadvantage.
Test the sensitivity of indices of segregation and diversity to regional and locality effects. A method to analyse spatially varying relationships (Geographically Weighted Regression) will be used to predict the values of measures of segregation and diversity (dependent variable) from a set of different regions and localities and their characteristics (predictor variables). The proposed procedure will provide a solution to control two aspects of the Modifiable Area Unit Problem (MAUP).
First, the zone effect, which will be analysed by demonstrating that results of measures of segregation and diversity, may differ between different ways of aggregating exactly the same data to the same scale. For example, the inclusion of rural areas is likely to increase the indices of segregation and reduce the measures of diversity.
Second, the scale effect, which will be highlighted by examining the changing behaviour of such indices as a result of using geographical areas where the size of locality population is not maintained over time. For example, the size of the smallest census areas in the UK reduced between 1991 and 2001, similarly, the number of wards decreased between the same period as a result of boundary changes.
Department:
Cathie Marsh Centre for Census and Survey Research, University of Manchester
Duration:
October 2007- October 2008
Grant Type:
Post Doctoral Fellowship
Publications
Research findings
Presentations
Sabater, A. (2009) Estimating
segregation and diversity of ethnic groups over time, presentation
made at UPTAP workshop, University of Leeds, 24th March
A
Sabater on Residential Ethnic Segregation for age cohorts in England
and Wales, 1991 to 2001 at BURISA on 29th September 2008
