Journal of Statistics Applications & Probability


In this article, authors have tried to implement a modified version of [1] estimator in a two-stage unequally clustered population where the second stage units can be either of responding or non-responding type. We have utilized imputation technique to tackle the non-response persisting even after second call in the situation of two-staged unequal cluster population which is completely a new attempt. The purpose is to strengthen the new estimator by improving the ‘second call’ part of the estimator. Instead of taking a simple weighted average of the responding part after ‘second call’ a difference type imputation (to deal with the non-responding part yet after ‘second call’ in a different way) is deployed to extrapolate the average and thus, examine the performance of the proposed strategy. Empirical studies carried over the data set of natural population. Suitable recommendations to the survey statistician are made.

Digital Object Identifier (DOI)