Efficient Imputation Methods to Handle Missing Data in Sample Surveys

This paper develops a class of modified difference-cum-exponential type imputation methods and corresponding point estimators to estimate the finite population mean in case of missing data problem. We have shown that the proposed strategies use the available auxiliary information effectively and give better results in comparison to some of the existing estimators. The performance of the proposed procedures over others is supplemented by the Monte Carlo simulation technique and presented through tables and graphs. Suitable recommendations are made to survey researchers.