ALTERNATIVE TWO-PHASE VARIANCE ESTIMATORS IN THE PRESENCE OF RANDOM NONRESPONSE

Abstract

Variance estimators are utilized to estimate the variability of population under study, and this variation estimates can aid in devising better policies. In this study, a two-phase population variance estimator under two-phase sampling is suggested. The properties of the estimator such as bias and mean square error were derived, and they were compared theoretically with some existing estimators. The efficiency conditions of the modified estimators under two realistic situations of random non-response were derived theoretically. The performances of the estimators were assessed using the criterion of mean square error and percentage relative efficiency. The empirical results using real data sets revealed that the proposed estimators performed better than the existing variance estimators considered. Thus, the proposed estimators in this study can be used to estimate variations that exist in real-world problems when there is random non-response.