Some General Classes of Efficient Estimators in Case of Missing Data

To handle the missing data problem in sample surveys, the imputation technique for missing values may suits well in reducing the negative impact of non-response in estimating the population mean. The socio-economic data yield fruitful results while imputation method employed to missing observations. Keeping this in mind, we have proposed three new general classes of difference-cum-ratio type imputation methods and the corresponding estimators in three different sampling strategies using the dual (rank) of an auxiliary variable in the presence of non-response. The biases and mean square errors of the proposed estimators are obtained up to the first-order approximation. The theoretical comparisons of the proposed estimators with usual mean imputation and the works Lee et al. [12], Kadilar and Cingi [11], Gira [9], Diana and Perri [8], Bhusan and Pandey [4], and Bhusan and Pandey [5] have been made which are also the special cases of the proposed estimators apart from being less efficient. The results are computed under an empirical study where the proposed work shows the efficacious performance over the above discussed works.