ABSTRACT
The Modifiable Area Unit Problem (MAUP) is recognized as one of the major issues in the spatial analysis of demographic patterns and processes. Using aggregated data that refer to regular geographies (i.e., grids), is a suggested way to avoid MAUP, although these kinds of data are frequently limited at national level and, when available, they typically cover a restricted number of socioeconomic variables. Moreover, it is relatively unclear how relevant is the overall effect of the MAUP, e.g. with reference to certain phenomena of demographic interest, such as the residential segregation of foreigners in urban contexts. Real empirical testing of this issue is relatively difficult because of lack of relevant geo-referenced databases on regular lattices. The present study tries to address this issue with an empirical analysis of residential segregation of foreigners in the three largest Italian cities (Rome, Milan, and Naples) by comparing the outcome of segregation indexes calculated using regular (grid) and non-regular (enumeration area) spatial domains. The empirical results of this study indicate that the effect of MAUP on segregation estimates is relatively minor, thus raising important questions about the (real) limits of using irregular geographies when analyzing spatial segregation of foreigners.
