An Evaluation of the Consistency of Subcounty Estimates of Population Based on the American Community Survey and the Population Estimates Program

Gregory Robinson, U.S. Census Bureau
David Dixon, Migration Policy Institute

The American Community Survey (ACS) data are controlled so that the population estimates defined by age, sex, race and Hispanic origin agree with the Census Bureau's official population estimates. Such adjustments are important to correct the survey data for non-sampling and sampling errors. The controls are applied at the county level; the estimates for subcounty areas are not directly controlled. This paper compares the ACS estimates for subcounty places to published population estimates. The differences in the two sets of estimates are systematically analyzed, in terms of percent difference by size of place, the pattern of differences (whether the ACS estimate is higher or lower than the population estimate and the characteristics of the places), and the variability in the implied change based on an estimates time series. The pattern of differences identified informs additional research to address the use of population estimates as ACS controls for subcounty areas.