Improved Randomized Response Technique for Two Sensitive Attributes

TL;DR

This paper introduces enhanced estimators for survey sampling that more accurately estimate the proportion of individuals with two sensitive attributes, outperforming previous methods especially as the prevalence of these attributes increases.

Contribution

It extends Mangat's (1994) work to develop more efficient estimators for two sensitive attributes, improving upon Lee et al.'s (2013) models.

Findings

Proposed estimators are more efficient than existing models.

Efficiency increases with higher population proportions of sensitive attributes.

The method improves accuracy in estimating sensitive attribute prevalence.

Abstract

We proposed new and more efficient estimators for estimating population proportion of respondents belonging to two related sensitive attributes in survey sampling by extending the work of Mangat (1994). Our proposed estimators are more efficient than Lee et al (2013) simple and crossed model estimators as the population proportion of possessing the sensitive attribute increases.