Domain Mean Estimators Assisted by Nested Error Regression Models

Abstract

This paper introduces estimators of domain means assisted by nested error regression models. The new estimators are modifications of empirical best linear unbiased predictors that takes into account the sampling weights. They are obtained by summing up the model-based predicted values adjusted by a weighted sum residuals. The paper studies the sampling-design properties of the introduced estimators by means of simulation experiments. The simulation results show that the new estimators present a good balance between sampling bias and mean squared error.