New ratio estimators for population mean in simple random sampling using robust regressionNew ratio estimators for population mean in simple random sampling using robust regression

Abstract

Using the traditional least square regression estimator we can validate the assumption of normality of the residual error when outliers occur in the data. In this paper, the alternative ratio estimators for estimating population mean using robust regression are  proposed for use in the case where data is contaminated with outliers, under simple random sampling.  The bias and mean square error of the proposed estimators have been investigated.  A simulation study has been conducted to compare the efficiency of the proposed estimators with traditional estimators.