Variance Estimation in Stratified Random Sampling in the Presence Of Two Auxiliary Variables

Abstract

The objective of this thesis is to develop improved population variance estimators in the presence of two auxiliary variables under stratified random sampling. New estimators of population variance of the study variable were suggested using prior information on two auxiliary variables. The mean square errors of the proposed estimators have been obtained using first order approximation of Taylor series method. Efficiency comparisons of proposed estimators have been discussed and achieved improvement under certain conditions. Results are also supported by numerical analysis. Based on result obtained, the proposed ratio- type variance estimators may be preferred over traditional ratio- type 𝑠2𝑡 and sample estimator of population variance 𝑠2𝑠𝑡,𝑦 for the use in practical applications.