Bayesian analysis of a sensitive proportion

다국어 초록 (Multilingual Abstract)

Respondents tend to answer untruthfully when they are asked to response sensitive questions, as they are reluctant to expose their identity. In order to reduce the response bias that can be generated through this, Warner (1965) proposed a randomized design which uses a randomization device that conceals individual response and protects the respondent. Thereafter, various survey designs to reduce the response bias were proposed, and Bayesian estimation and simulation methods have also been studied. This study proposes an analysis method to reduce posterior standard deviations of the sensitive proportions in a survey with sensitive questions and compare the results with the existing analyzes through the simulation. In addition, this study applies the proposed analysis method to the actual survey with the sensitive questionnaires related to the organizational commitment to the experienced employees.

목차 (Table of Contents)

• Abstract
• 1. Introduction
• 2. Bayes estimation of a sensitive proportion
• 3. Computations
• 4. Simulation study
• 5. Real data analysis
• 6. Discussion
• References