Evaluation of small area population estimates produced by housing unit, ratiocorrelation, and component method II compared to 2000 census counts

Abstract

This paper evaluates small area population estimates produced by the Housing Unit Method, Ratio-correlation Method, Component Method II, and the average of all three methods compared to the 2000 Census counts for the 254 counties and 1,279 places in Texas. This evaluation of three estimation methods shows the expected patterns of error by population size and population change. Of the three methods tested, no single method produced more accurate estimates than the average of two or three methods. The assessment of the accuracy of the place-level estimates show substantially higher levels of errors than those found for counties.