Computing the effect of measurement errors under additive scramble response of the sensitive variable

Abstract

Abstract The use of scramble variable is considered herein randomized response technique to estimate the parameters of the sensitive variable. We propose a generalized class of regression-cum ratio estimator for the mean estimation of the sensitive variable. Variables under study are assumed to be recorded with measurement error. Mean square error (MSE) of the proposed estimator and members of the estimators are computed in the presence of measurement error. The efficiency of the estimators and the proposed estimator is shown by the graph for different values of the correlation coefficient. A simulation study is also conducted to examine the influence of measurement error on the MSE of the proposed estimator and other estimators at the different levels of scramble variable response.