Effects of different calibration constraints on calibration estimators under the randomized response technique

Abstract

In survey sampling, to handle the situation where human bias arises due to some personal or illegal questions, the randomized response technique developed by Warner is used [1]. The calibration technique is used to reduce the bias between the original and estimated weights in stratified random sampling. We estimate the population total, using the calibration estimation technique in which the study variable corresponds to a scrambled variable due to the randomized response technique. We have discussed the estimation procedure of population total and suggested some new estimators, such as ratio and exponential estimators, using different distance measures and calibration constraints. The simulation study has been applied tn newly developed estimators in sensitive survey sampling using a scrambled randomized response (SRR). We see that exponential estimators perform better than ratio and simple estimators of the total stratified random samplings when the variable of interest is sensitive.