An improved randomized response model for cheater detection using unrelated question
Year: 2026
Type: article
Abstract: {Statement: Developed a novel unrelated-question cheater detection model that simultaneously estimates the proportion of the sensitive attribute and the proportion of cheaters in the responses, improved efficiency and privacy protection.} The randomized response technique (RRT) is an effective method for collecting sensitive information from respondents while ensuring their privacy. Randomized devices incorporating unrelated questions have gained significant popularity due to their psychological acceptability among participants and their ability to protect privacy. Traditional RRT models estimate only the proportion of the sensitive attribute π, under the assumption that respondents answer truthfully. However, these models often overlook the presence of respondents who provide dishonest answers. In this article, we propose a novel unrelated question cheater detection model that simultaneously estimates the proportion of the sensitive attribute π and the proportion of cheaters in the responses γ. A simulation study is conducted to validate the theoretical properties of the proposed estimators. Furthermore, a numerical comparison of efficiency and privacy protection is carried out between the proposed model and the model developed by Reiber et al. [Cheater detection using the unrelated question model, Sociological Methods & Research, 2023;52(1):389–411] under simple random sampling (SRS).
Source: Figshare
Authors: Ghulam Narjis, Farah Abbasi, Javid Shabbir
Language: English
Subfield: Statistics and Probability
Field: Mathematics
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7234344204
