Advances in Research Methods for the Study of Urban Segregation

Abstract

The analysis of spatial segregation involves several methodological issues. First, there are multiple dimensions of socioeconomic residential segregation and different ways of measuring each. Second, social processes in space will not necessarily correspond to the geographical definitions used in comprehensive data collection, such as national population censuses, thus creating problems of measurement. But the field of spatial statistical analysis and geocoded census data creates new opportunities for addressing these issues and advancing the state of the art. Finally, of particular relevance to this volume, is that population censuses have yet to be standardized across countries in the Americas, as has been the case with the European Union, and, thus, formal, quantitative cross-country comparisons are not yet feasible. In this chapter the methodological frame used in the case studies that follow is presented. This chapter first defines the multiple dimensions of segregation and their measurement. The types of measurement errors common to this type of analysis, including the problem of areal definition and spatial autocorrelation, are explored. Finally, the types of variable and areal definitions adopted in the case studies are presented.