A Bayesian Approach to Population Estimation with Administrative Data

Abstract

Abstract The article describes a Bayesian approach to deriving population estimates from multiple administrative data sources. Coverage rates play an important role in the approach: identifying anomalies in coverage rates is a key step in the model-building process, and data sources receive more weight within the model if their coverage rates are more consistent. Random variation in population processes and measurement processes is dealt with naturally within the model, and all outputs come with measures of uncertainty. The model is applied to the problem of estimating regional populations in New Zealand. The New Zealand example illustrates the continuing importance of coverage surveys.