A Bayesian approach to 2000 Census evaluation using ACE survey data and demographic analysis

Demographic analysis of data on births, deaths, and migration, together with coverage measurement surveys that use capture-recapture methods, have established that U.S. Census counts are flawed for certain subpopulations. Previous work using 1990 Census data in African-Americans age 30-49 proposed a hierarchical Bayesian model that assembled Census, follow-up survey, and demographic data, providing a principled solution to the problem of negative estimated counts in some subpopulations, smoothing highly variable estimates across subpopulations, and providing estimates of precision that incorporate uncertainty in the demographic analysis estimates. This article extends that effort by refining the hierarchical model design, expanding the set of models considered, considering the presence of bias in the Census or follow-up survey counts, obtaining Bayes factors for use in model selection, and applying the methods to the entire 2000 U.S. Census. Comparisons with the 1990 U.S. Census results are included as well. KEY WORDS: Bayes factors: Capture-recapture: Gibbs sampling; Postenumeration survey.