Abstract

In the era of survey sampling, there is an increasing interests in sensitive features of the population that people generally prefer to hide from people. A novel technique called the Randomized Response Technique (RRT)is used to gather reliable data, protectthe confidentiality of respondents, and estimate the population proportion, bearing the sensitive attribute. But the perception of sensitivity is notthe same for every person. It has been observed that a group of people are willing to disclose their true nature rather than the compulsory Randomized Response (RR). Considering this fact, the concept of optional RRT was developed.In the present work, we reformulate an estimation method of sensitive population proportion addressed by Singh, Singh, Mangat and Tracy s’two-stage RRT. The procedure is more generalized here when respondents are chosen exclusively byan unequal probabilitysampling scheme and an unbiased estimator is derived along with its unbiased variance estimator. Further, we develop an Optional Randomized Response (ORR) method based on our proposed work. A simulation study has been carried out to find out the efficacy of the proposed RR and ORR procedures.

Author: Dipika Patra and Sanghamitra Pal

Received on: October, 2019

Accepted on: March, 2021