A Trust-Enhanced Binary RRT Model with Measurement Errors

Human surveys in a variety of disciplines experience challenges such as social desirability bias, confidentiality distrust by the respondents, and errors in measurement. All these factors may significantly reduce the quality of the data. To handle these challenges, a variant of the Randomized Response Technique (RRT) for binary sensitive variables has been suggested in this study. This builds on recent studies where trust enhancement has been implemented in the binary domain but not in the presence of measurement errors. This study attempts to fill that gap. Extensive simulations are used to validate theoretical results. We also highlight the risks involved in naively assuming that measurement errors do not exist.