Journal of Official Statistics, Vol.16, No.1, 2000. pp. 1–13
Correcting the Bias in the Range of a Statistic Across Small Areas
David R. Judkins and Jun Liu
Abstract:In small area estimation problems, there is usually a large number of areas for which simultaneous estimates are required for comparison purposes. Such a collection of statistics can be called an ensemble estimator.
The properties of several ensemble estimators are compared. Design-based ensembles are found to have too wide a range. Bayes and empirical Bayes ensembles are found to have too narrow a range. Constrained empirical Bayes ensemble estimators appear to be best for some purposes.
Keywords:Range estimator; small area estimation; design-based; empirical Bayes; Bayes.
Copyright © Statistics Sweden 1996-2018. Open AccessISSN 0282-423XCreated and Maintained by OKS Group