存在测量误差时分层抽样层均值方差的估计  

On the Variance Estimation of the Stratum Mean in Stratified Sampling When Measurement Error Exists

作　　者：王克林[1] 

机构地区：[1]广东商学院经济贸易与统计学院,广东广州510320

出　　处：《统计与信息论坛》2011年第3期16-20,共5页Journal of Statistics and Information

基　　金：全国统计科学研究计划项目<中国抽样调查中的非抽样误差问题研究>(2010LC43)

摘　　要：利用模型的方法研究出现测量误差时多变量间的关系是目前国际上的流行方法,但这不利于对单指标的估计。因此,通过在估计量的设计中纳入测量误差信息,推导测量误差方差的定量测度方法,实现了存在测量误差时分层抽样各层均值方差的估计。采用2007年广东省三个市(县)城镇住户调查中的人均消费性支出数据进行实证分析,定量测度了测量误差在层均值方差估计中的大小及其影响,并对不考虑测量误差的估计结果进行了修正。The methods of using models to study the relationship among multi-variable are popular in the world,but these do not do well when it comes to only one variable.Bringing into measurement error information,the paper designs the variance estimation of the stratum mean in stratified sampling and deduces the formula of measuring measurement error quantitatively.By this way,the paper can estimate the variance of the stratum mean in stratified sampling when measurement error exists.The paper explores the households＇ consumer data,including 3 cities or town in Guangdong province,and measures the measurement error and the degree of the influence quantitatively,and amends the estimation taking no account of measurement error.

关 键 词：测量误差 调查条件 真实值 

分 类 号：O212.4[理学—概率论与数理统计]