Incorporating direct responses into optional randomized response models without compromising respondents’ privacy

Accepted - March 2024

Abstract

Survey researchers use randomized response models for collecting data in sensitive surveys. The existing optional scrambling models lack the option of direct response. In practice, some of the respondents may be willing to provide the direct response. An efficient randomization technique is proposed where each respondent is free to provide either direct or protected response. The findings reveal that the option of direct response increases the efficiency of the models without compromising on data privacy. A real-life application of the proposed technique is also presented and the improvement in efficiency is shown for various choices of parameters.