Generalized Ratio-Product cum Regression Variance Estimator in Two-Phase Sampling

Abstract

This study develops a flexible and efficient generalized ratio-product cum regression-type estimator of population variance utilizing auxiliary variable in two-phase sampling that incorporates the properties of ratio-type and product-type estimators. The properties of the estimator were derived using first order approximation. The theoretical conditions under which the precision and the flexibility of the estimator is better than some classical estimators are also provided. Empirical evidence from five real datasets suggests that the proposed estimator outperforms the classical variance, ratio variance, product, and exponential ratio type estimators in terms of precision and efficiency. The estimator can be utilized to provide better variance estimates for various phenomena such as inflation variation, exchange rate variation and standard of living variation for better policymaking.