Urban Transition Project

Even without HGIS maps, the 100% microdata now available for the 1900-1940 decades provide abundant information for spatial analyses of residential segregation. The Urban Transition team has worked extensively with the microdata for more than 190 cities to be able to identify people’s location more accurately at multiple spatial scales: the household, building, adjacent buildings, streets, enumeration districts, and wards. These data are now available for download in the Urban Transition Geographic Reference File.

We have now made use of these data to analyze residential segregation between blacks and whites in 134 major cities that had at least 1000 black residents in any decade in this period.  A report on our findings is now available in the following working paper.  There are three main findings.  First, racial segregation was already high at a local scale in 1900 and increased greatly in nearly all cities.  Second, the scale of segregation in all kinds of cities shifted from individual streets with many black residents toward entire wards that were more highly divided by race.  Third, among the many Northern cities with modest black populations, those with greater black presence were more segregated, consistent with a “group threat” hypothesis.  In Southern cities and major Northern Destination cities, however, there is no support for this explanation. These findings lead toward a conclusion that the main trends creating the black ghetto by the mid-20th Century were national in scope, and future research should focus less on why some cities differed from others than on why such similar trajectories are found everywhere.

John R. Logan, Benjamin Bellman, and Elisabeta Minca. 2020. “From Side Street to Ghetto: Understanding the Rising Levels and Changing Spatial Pattern of Segregation, 1900-1940

The segregation measures for every city, every year, and every spatial scale can be downloaded from HERE. The download is a single csv file. It includes measures of black isolation (P*bb) and Dissimilarity (D) between blacks and whites at every spatial scale mentioned above. Another type of measure is the correlation ratio, eta-squared (originally proposed Bell). It also includes two measures of segregation based simply on the order of enumeration, sometimes referred to as “next-door-neighbor” measures – the SIS (used by Grigoryeva and Ruef) and what we call the NIS (used by Logan and Parman).

Please cite the Logan, Bellman, and Minca working paper and this webpage as the source of measures when you use them in your own research.