A Quantitative Analysis of the Relative Effects of the Components of Change Used in Postcensal County Population Estimates

Sam T. Davis, U.S. Census Bureau

The question often arises as to the relative effect of the five components of change used to produce postcensal county population estimates. Unlike previous data displays and discussions, this analysis looks at the effects of the components over all counties, and looks at the characteristics of counties having strong influences by each component. The first approach in examining this involves looking at the components for the post Census 2000 estimates and tabulating their relative effects on the estimates. The other approach looks at the components from an evaluation perspective. We take sets of 1990-based postcensal estimates for April 1, 2000 omitting one component for each set. The differences in error from the Census 2000 results between the routine set of estimates and the sets omitting a component are obtained. We compare these differences among the components to determine the quantitative effect on the estimates in relation to the Decennial Census.