不同复杂抽样调查的设计效率比较分析

王建生; 冯国双; 于石成; 马林茂; 周脉耕; 刘诗瑶

Abstract

王建生,冯国双,于石成,马林茂,周脉耕,刘诗瑶.不同复杂抽样调查的设计效率比较分析[J].Chinese journal of Epidemiology,2012,33(10):1079-1082

不同复杂抽样调查的设计效率比较分析

Comparison of the Designing Effects(DE)among different designs related to complex samplingmethods

Received:May 23, 2012

DOI：

KeyWord: 复杂抽样调查设计效率

English Key Word: Complex sampling design Design effect

FundProject:

Author Name	Affiliation	E-mail
Wang Jiansheng	Policy Research Center for Environment and Economy, Ministry of Environment, P. R. China, Beijing 1 00029, China
Feng Guoshuang	National Center for Public Health Surveillance and Information Services, Chinese Centerfor Disease Control and Prevention
Yu Shicheng	National Center for Public Health Surveillance and Information Services, Chinese Centerfor Disease Control and Prevention
Ma Linmao	National Center for Public Health Surveillance and Information Services, Chinese Centerfor Disease Control and Prevention
Zhou Maigeng	National Center for Public Health Surveillance and Information Services, Chinese Centerfor Disease Control and Prevention	maigengzhou@126.corn
Liu Shiyao	National Center for Public Health Surveillance and Information Services, Chinese Centerfor Disease Control and Prevention	liushiya0641@163.com

Hits: 3630

Download times: 2091

Abstract:

为比较常用的几种复杂抽样调查的设计效率，以“2002年中国居民营养与健康状况调查”的数据为依据构造抽样总体，利用统计模拟的方法，估计出不同复杂抽样调查的设计效率值。结果显示，不同复杂抽样的设计效率值差别较大，样本量越大、抽样阶段数越多、分层数越少，复杂抽样的设计效率值越大。实际中采用复杂抽样时，可尽量减少抽样阶段数，细化分层类别，以降低设计效率值，提高设计效率。

English Abstract:

To compare the designing effects(DE)among different complex samplingdesigning programs.Data from the‘2002 Chinese Nutrition and Health Survey’was used as an example to generate the sampling population，and statistical simulation method was used to estimatethe values of DEs from six complex sampling designing programs.It was found that the values of DE svaried among the six complex sampling designing programs.The values of the DEs were associated with the sample sizes in a positive way.with more sample stages and leSS stratified categories. Reduction of the numbers of sample stages and detailing stratified categories could decrease the DE values SO as to improve the DE.

View Fulltext Html FullText View/Add Comment Download reader