Admissibility of Continuous Unbiased Estimators in Linear Mixed Models

Document Type : Original Article

Author

Abstract

For a given linear function of the fixed effects in the usual mixed linear model, within the class of estimators, a discontinuous unbiased estimator is introduced. This estimator is shown to be uniformly better (as measured by the variance) than the estimator given by Azzam, Birkes, and Seely (1986).

Keywords