Estimation of Sensitive Attributes Using Randomized Response Techniques in Sample Surveys

Abstract

The survey is a research technique that involves asking inquiries of a sample of components. Sample studies may draw attention to respondents' private or sensitive information. The tendency of respondents to respond on the basis of what is socially acceptable may harm the response rate of surveys. This may be because a large number of respondents decline to participate in the survey or give erroneous or conditional replies, which have a significant impact on the estimates' accuracy and dependability. To safeguard respondents' privacy and lessen the chance that they may avoid answering difficult questions, the Randomized Response (RR) approach was devised. Warner's work has spawned a large body of literature and been applied in numerous sectors, although these methods have challenges and drawbacks. The Horvitz Randomized Response method is employed in this article to address the sensitive properties. In order to obtain findings that are more accurate than those obtained by using previous RRT algorithms, we estimate two sensitive attributes: lying and bullying. To compute our findings, we employ the R codes.