Estimation of population mean of sensitive variable using RRT in RSS under the situation of missing data

Here, we propose an efficient exponential type estimator of population mean of a quantitative sensitive study variable using scrambled response model in Ranked Set Sampling (RSS) in case of missing data. It is assumed that information on non-sensitive auxiliary variable is known. The expressions of biases and Mean Square Errors (MSE) of the proposed and other estimators are derived upto first order approximation. The conditions are derived under which proposed estimator will become more efficient as compared to other estimators. To validate the theoretical results, we perform a thorough simulation study. The effects of variance of scrambling variable, responding probabilities and varying correlation coefficient between study and auxiliary variable on MSE s of the proposed and other estimators have been studied in the simulation process. The analysis clearly signifies the superior performance of our proposed estimator as compared to other estimators in the context of Ranked Set Sampling (RSS).