On A Method of Bias Reduction in the Product Method of Estimation

Abstract

In this paper, we focused our attention on the creation of an almost unbiased predictive product estimator after estimating and correcting bias of the classical product estimator under predictive approach. Considering mean square error as the performance measure, superiority of the proposed estimator has been analyzed compared to the classical product estimator and Robson’s [1] unbiased product estimator under (i) a finite population set-up, (ii) an infinite population set-up assuming bivariate normal distribution between the variables, and (iii) the assumption of a super-population model.