Using measures of spatial autocorrelation to describe socio-economic and racial residential patterns in US urban areas
Residential patterns are investigated using an exploratory approach with GIS-based thematic maps and spatial statistics. The research aims to contribute to the debate whether geopolitical fragmentation enhances spatial segregation of households based on racial and socio-economic characteristics. Using 1990 US census data for three key variables (income per capita, percent black and percent hispanics), measures of spatial association serve as quantitative index to describe the level of homogeneity or heterogeneity of residential neighborhoods. Statistical results are mapped and the correlation of significant homogeneous clusters with political entities is assessed. The results show considerable variation of household cluster sizes depending on the geopolitical structure as well as the population composition of the urban area. Also, cluster sizes and locations are correlated with municipal boundaries in some of the areas; this supports the proposition that the location of segregated communities is influenced by political and policy frameworks.