OUP - Abstract

Using Geo-Demographic Methods for Improving Small-Area Population and Housing Unit Estimates

Dissertation School: University of Wisconsin-Madison

Abstract:
This study introduces, tests and demonstrates the efficiency of a methodology for improving post-census population and housing units estimates for Census Block Groups—neighborhood-sized geographical units of roughly 1,000 persons. For twelve Midwestern states, iterative cluster analysis was performed on their approximately 60,000 Census Block Groups, based on the block group's values for a selection of variables from the 1990 Census of Population and the 1987 Census of Agriculture. This resulted in a typology of roughly 35 different clusters—neighborhood “types” based on such demographic and economic variables as age composition, racial makeup, income, poverty status and educational attainment. This segmentation was then used to test a methodology known as “social area analysis”, which theorizes that neighborhoods with a similar demographic and economic composition will have a similar population dynamic as well. The hypothesis of this study are: (1) within-cluster variation in population growth will be less than between-cluster variation in population growth, and (2) knowledge about population dynamics for selected neighborhoods of a given type can be applied to other neighborhoods of the same type. To test the hypothesis, a sample of block groups is drawn from each cluster and their 1980-1990 percent growth rates in housing stock were calculated. These rates are then applied to the respective clusters to produce estimates of housing units for 1990. A different set of estimates for number of housing units in 1990 is calculated using the constant shares method. These two set of estimates are then compared with the actual 1990 observed housing unit counts at the block group level.

Back to Search Result of DDRG Dissertations