沃纳模型中的矩估计量及其方差  

Moment Estimators and Their Variances on Warner's Model

作　　者：董俊超[1] DONG Jun-chao(School of Mathematics and Information Science, Yantai University, Yantai 264005, Chin)

机构地区：[1]烟台大学数学与信息科学学院,山东烟台264005

出　　处：《烟台大学学报（自然科学与工程版）》2017年第2期89-92,105,共5页Journal of Yantai University(Natural Science and Engineering Edition)

摘　　要：随机化回答技术由沃纳(Warner)在1965年提出,用于处理在社会调查中的敏感性问题.其抽样过程可分为4类:Ⅰ.人总体以及问题总体都采用放回抽样;Ⅱ.人总体采用不放回抽样而问题总体采用放回抽样;Ⅲ.总体采用放回抽样而问题总体采用不放回抽样;Ⅳ.人总体以及问题总体都采用不放回抽样.本文首先给出了沃纳模型中4种情形下参数的矩估计,然后给出了求估计量方差的一个新方法,该方法有别于沃纳以及以前研究者的方法,该方法更容易理解及更具一般性.In the Warner＇ s randomized response technique,dealing with the sensitive questions in social survey,the sampling process can be divided into four cases： 1. the respondent and question populations are both sampled with replacement; 2. the respondent population is sampled without replacement and question population is sampled with replacement; 3. the respondent population is sampled with replacement and question population is sampled without replacement; 4. both the respondent and question populations are sampled without replacement. We firstly give the moment estimators of the parameters in the 4 cases,and then present a new method to calculate the variance of the parameters. It is shown that our method is much easier and more general than others.

关 键 词：敏感属性 随机化回答技术 放回抽样 不放回抽样 

分 类 号：O212.7[理学—概率论与数理统计]