A Practical Implementation and Bayesian Estimation in Franklin's Randomized Response Procedure

Abstract

A Bayesian estimator based on Franklin's randomized response procedure is proposed for proportion estimation in surveys dealing with a sensitive character. The method is simple to implement and avoids the usual drawbacks of Franklin's estimator, i.e., the occurrence of negative estimates when the population proportion is small. A simulation study is considered in order to assess the performance of the proposed estimator as well as the corresponding credible interval.

Keywords:

• Bayesian inference
• Computer algebra system
• Proportion estimation
• Randomized response method

Mathematics Subject Classification:

• 62D05
• 62F15
• 62P25