In this paper, we have proposed an improved estimator of the population mean in simple random sampling without replacement by applying Searls (1964) technique on the estimator given by Roy (2003). We have also proposed its extension to stratified random sampling which is a generalization of an estimator given by Grover (2006). We have also used robust regression method to these proposed traditional estimators and developed new robust estimators in simple random sampling as well as in stratified random sampling. The biases and mean square errors of these proposed traditional estimators have been obtained, up to first order of approximation. After comparison of mean square errors of various traditional estimators, it has been found that the proposed traditional estimator is more efficient than its competitor estimators considered in this paper. The proposed estimator based upon robust regression method is more efficient than proposed traditional estimators in simple random sampling and in stratified random sampling under some conditions. The theoretical results so obtained have been verified with the help of some empirical examples and simulation studies.

在本文中，我们对Roy（2003）给出的估计量应用Searls（1964）技术，提出了一种简单的随机抽样中无需替换的总体均值的改进估计量。我们还建议将其扩展到分层随机抽样，这是Grover（2006）给出的估计量的概括。我们还对这些建议的传统估计量使用了鲁棒回归方法，并在简单随机抽样以及分层随机抽样中开发了新的鲁棒估计量。这些拟议的传统估计量的偏差和均方误差已经获得，直到近似一阶。通过比较各种传统估计量的均方误差，我们发现所提出的传统估计量比本文中考虑的竞争者估计量更有效。在某些条件下，基于鲁棒回归方法的估计器在简单随机抽样和分层随机抽样中比传统估计器更有效。如此获得的理论结果已通过一些经验实例和仿真研究得到了验证。