PAKISTAN JOURNAL OF STATISTICS, cilt.26, sa.3, ss.461-477, 2010 (SCI-Expanded)
Estimation of the mean of a univariate normal population with, unknown variance is a well-known problem in the presence of an uncertain prior information. In this study, we propose a new estimator, shrinkage pre-test estimator, which is a combination of pre-test and shrinkage estimators. The mean squared error (MSE) for the shrinkage pre-test estimator is also derived and theoretical comparisons based on MSE criterion of this new estimator with the restricted, the pre-test, and the shrinkage estimators are given. We show that the shrinkage pre-test estimator performs better than the other existing estimators. In addition, the performance of this newly proposed estimator is being assessed by conducting a simulation study.