Authors: Michael Ratcliffe*, U.S. Bureau Of the Census
Topics: Population Geography, Applied Geography
Keywords: Census tract, census, statistical geography
Session Type: Paper
Start / End Time: 3:20 PM / 5:00 PM
Room: Astor Ballroom III, Astor, 2nd Floor
Presentation File: No File Uploaded
Census tracts have existed for over 100 years as small geographic units for tabulating, disseminating, and analyzing data. For much of that time, the twin pillars of the census tract concept were homogeneity in terms of demographic, social, and economic characteristics and comparability from one decade to another, typically viewed as spatial comparability. These conceptual pillars supported the view that census tracts were statistical equivalents of neighborhoods. Given the demographic shifts that have occurred over the past 50-60 years, particularly in central cities and older suburbs, it has been difficult, if not impossible, to continue to meet both concepts; that is, to maintain both socio-economic homogeneity and spatial comparability (allowing for splits or mergers to adhere to size criteria). For this reason, the Census Bureau dropped all references homogeneity for the 2010 Census, focusing only on comparability. Concerns about reliability of sample data affected whether tracts in some counties were split to adhere to optimum population thresholds. Concerns about the modifiable areal unit problem have led some analysts to question the continued utility of census tracts. In this presentation, I discuss the changing nature of census tracts, factors that continue to underpin their relevance as well as those that have led to questions about their utility, and in general, provide the conceptual backdrop for other papers addressing the spatial and statistical issues related to data at the tract level.