Variance estimation using Randomized Response Technique

UNCG Author/Contributor (non-UNCG co-authors, if there are any, appear on document)
Badr Aloraini (Creator)
The University of North Carolina at Greensboro (UNCG )
Web Site:
Sat Gupta

Abstract: The aim of this dissertation is to study the problem of variance estimation for sensitive variables. The problem of variance estimation in the non-sensitive variable case has been well-explored by researchers. However, very little attention has been devoted to the case of sensitive variables. We use Randomized Response Technique (RRT) and Optional Randomized Response Technique (ORRT) models to propose several variance estimators in Simple Random Sampling under measurement errors. We also extend our investigation to the case of the more complex sampling design, namely the Stratified Sampling Design. Bias and the mean square error (MSE) expressions for these proposed estimators are also derived. Performance of the proposed estimators is evaluated using a unified measure of respondent privacy and estimator efficiency through a simulation study. Our results show that the proposed generalized variance estimator is more efficient than the basic variance estimator and the ratio variance estimator. This holds true both in the presence of measurement errors and in the absence of measurement errors. Additionally, it is shown that the ORRT model performs better than the non-Optional RRT model.

Additional Information

Language: English
Date: 2021
Measurement Errors, Optional Randomized Response Technique, Randomized Response Technique, Respondent Privacy, Stratified Sampling, Variance Estimation
Estimation theory
Analysis of variance
Sampling (Statistics)

Email this document to