Almost Unbiased Variance Estimators Under the Simultaneous Influence of Non-response and Measurement Errors

Abstract

This paper addresses the problem of biases in variance estimators under the simultaneous influence of both measurement and non-response errors. Three classes of estimators are suggested and their properties are studied up to first order of approximation. Also, their efficiency conditions over some existing estimators are established. Resultant almost unbiased estimators of the proposed estimators are obtained and their properties are discussed. Theoretical results reveal that the suggested class of estimators under the measurement and non-response errors are more efficient. Results of simulation study also indicate superiority of suggested estimators over other estimators considered in the study. © 2022, Grace Scientific Publishing.