文章基本信息

标题：On Eliminating The Scrambling Variance In Scrambled Response Models
本地全文：下载
作者：Zawar Hussain
期刊名称：International Journal of Academic Research in Business and Social Sciences
电子版ISSN：2222-6990
出版年度：2012
卷号：2
期号：6
页码：39-45
出版社：Human Resource Management Academic Research Society
摘要：To circumvent the response bias in sensitive surveys randomized response models are being used. To add into it we propose an improved response model utilizing both the additive and multiplicative scrambling method. The proposed model provides greater flexibility in terms of fixing the constant Kdepending upon the guessed distribution of sensitive variable and nature of the population. The proposed model yields an unbiased estimator and is anticipated as more protective against the privacy of the respondents. The relative efficiency comparison of the proposed estimator is made relative to Hussain and Shabbir (2007) RRM. Furthermore, the proposed model itself is improved by taking the two responses from each respondent and suggesting a weighted estimator yielding an unbiased estimator having the minimum possible sampling variance. The suggested weighted estimator is unconditionally more efficient than all of the suggested estimators until now. Future research may be focused on privacy protection provided by the scrambling models. More scrambling models may be identified and improved by taking the two responses from each respondent in such a way that the scrambling effect is balanced out
关键词：Randomized response models; sensitive surveys; privacy protection; simple random ;sampling and weighted estimation