ABSTRACT
The triple-source estimator avoids the correlation bias inherent in the double-source estimator, which has been a popular estimator of population size for assessing the quality of a census. The triple-source estimator relies on the census list, the quality-assessment survey list, and the administrative record list. It also provides an estimate of the net census coverage error. It is established in population strata of equal probability of being drawn and based on a sample. The triple-source estimator provides an unbiased estimate of the net error.
Acknowledgments
The authors thank the anonymous reviewers for their useful comments.