Urban racial residential segregation: A comparison of a-spatial and spatial measures of dissimilarity in a multi-ethnic city

Background. Indices of urban racial residential segregation have been linked to racial disparities in health. First generation of a-spatial segregation indices such as the Dissimilarity index (Duncan and Duncan, 1955) and the Isolation index (Bell 1954), are insensitive to spatial arrangements of population among spatial units. Distance-based spatial indices like the Spatial Proximity measure are an attempt to account for the spatial distribution of race/ethnicity across spatial features (Jakubs 1981, Morgan 1983, Wong 1993, Reardon and O'Sullivan 2004). In 2000 Wong implemented a set of spatial segregation measures in a GIS environment. In this study we further explore both, a-spatial and spatial, dissimilarity indices using a relatively small multiethnic urban sample. Methods. Data for this analysis are from a two-stage random probability sample of Detroit residents (≥25 years old), conducted in 2002 by the Healthy Environments Partnership, a community-based participatory research partnership in Detroit. Association of a-spatial and spatial dissimilarity index will be assessed using correlation coefficients. We will also test their prediction power in regression models to assess association with cardiovascular disease indicators such as Allostatic Load. Results. We will present findings examining the association of a-spatial and spatial measures of segregation using the Dissimilarity index. We will determine the usefulness of developing complex spatial segregation measures for a relatively small sample from a multiethnic urban community of the City of Detroit.