Robust small domain estimation using random effects modeling - ARCHIVED

This paper develops a design consistent small domain estimator using a random effects model. The mean squared error of this estimator is then evaluated without assuming the random effect component of the model is correct. Data from a complex sample survey shows how this approach to mean squared error estimation, while perhaps too instable to be used directly, can be employed to determine whether the design consistent small domain estimator proposed here is better than the conventional design-based estimator.