Treating the problem of non-response in stratified random sampling under calibration approach

In the present study, we suggested a chi-square-type distance function that works well for non-response situations and we have derived a new calibration-based estimator using such a kind of distance function, which is used to estimate the mean of a stratified population. The proposed estimator is derived through the implementation of a reweighting technique referred to as the calibration approach, aimed at improving upon Hansen and Hurwitz's (Citation1946) estimator. Typically, in situations where some units do not respond, researchers adjust the stratum weights to obtain an estimator for the mean of a stratified population using the calibration approach. In this study, the weights given to the responding units in the sample and the units in the sub-sample taken from the non-responding sample were calibrated to derive the calibration estimator. We assume that the unit non-response problem impacts both study and auxiliary variables. To examine the performance of the proposed calibration estimator in scenarios involving non-response, we have carried out empirical studies.