A Comparison of Three New Randomized Response Models for Simultaneous Estimation of Three Sensitive Dependent Characteristics and Their Overlaps

• Activity Number: 684
• Type: Topic Contributed
• Date/Time: Thursday, August 4, 2016 : 10:30 AM to 12:20 PM
• Sponsor: Survey Research Methods Section
• Author(s): Oluwaseun Olanipekun* and Stephen Sedory and Sarjinder Singh
• Companies: and Texas A&M University - Kingsville and Texas A&M University - Kingsville
• Keywords: Randomized response model ; privacy protection ; dependent sensitive characteristics ; simultaneous estimation ; simple random sampling ; relative risks and correlation coefficients

Abstract

Randomized response models have been considered the best method for sensitive surveys which provide the maximum privacy protection to the respondents and procure honest responses. In this paper, we concentrate on the issue of estimation of proportions of the population possessing sensitive characteristics. However, there is amazing deficiency of articles that have addressed the simultaneous estimation of three dependent sensitive attributes. In filling this hole, we present three new models to estimate the proportions of three dependent sensitive characteristics simultaneously. Following Lee, Sedory and Singh (2013), one simple model can be used to develop proportion estimators. This simple model could consist of three decks of cards used in series, and in using these procedures, we can derive the expressions for estimators of some of the marginal and conditional proportions and their variances. These estimators will allow us estimate all the conditional and marginal proportions, relative risks, correlation coefficients, multiple correlation coefficient and partial correlation coefficients. The estimators are developed and investigated from the efficiency perspective.