Abstract

In this paper, a ratio-cum-product estimator of the mean of a finite population is proposed using a number of auxiliary variables, some of which are positive and some negatively correlated with the study variable. Expressions for the bias and the mean squared error (MSE) of the estimator are obtained, and the conditions determining the sign of the bias are derived, as well as those under which the estimator is more efficient than the sample mean. A simple random sample drawn from a finite multivariate population is used to illustrate the determination of the optimum weights and the gain in efficiency obtained over the situations where only positively correlated auxiliary variables or only negatively correlated variables are used in the estimator.

Author: G. Chattapadhyay and K. B. Panda

Received on: January, 2023

Accepted on: June, 2024