蒙特卡洛模拟在多分类敏感问题两阶段整群抽样信度与效度评价中的应用  

Reliability and Validity Evaluation Based on Monte Carlo Simulations in Two-stage Cluster Sampling on Multinomial Sensitive Question Survey

作　　者：杜俏俏[1] 高歌[1] 靳宗达[1] 朱宏儒[2] 李伟[1] 陈向宇[1] 

机构地区：[1]苏州大学公共卫生学院流行病与卫生统计学教研室,215123 [2]江苏省血吸虫病防治研究所

出　　处：《中国卫生统计》2013年第2期217-220,共4页Chinese Journal of Health Statistics

基　　金：国家自然科学基金项目(项目编号:30972548;81273188)

摘　　要：目的采用蒙特卡洛模拟方法,评价多分类敏感问题RRT模型下两阶段整群抽样调查方法及其统计公式的信度和效度,为多分类敏感性问题提供准确可靠的复杂抽样调查方法及其统计量计算公式。方法采用两阶段整群抽样,随机抽取北京市6个区28个男同性恋活动场所1523名男同性恋者;采用多分类敏感问题RRT模型调查;根据本项目已推导出的相关公式,计算出敏感问题各属性的样本比例及其方差。以此样本为模拟总体,应用Matlab语言编程进行多分类敏感问题两阶段整群抽样下的蒙特卡洛模拟调查30次,分别作模拟调查的样本比例与模拟总体比例比较的Z检验。结果 30次Z检验P值的范围:0.05<P<0.99,差异均无统计学意义。结论蒙特卡洛计算机模拟抽样调查结果表明多分类敏感问题RRT模型下二阶段整群抽样调查方法与统计量的计算公式信度、效度较高。Objective To evaluate the reliability and validity of the RRT method to investigate qualitative sensitive questions with the sample selected by two-stage cluster sampling and the statistical formula. Methods 28 chambers of MSM from 6 districts of Beijing city were ran- domly selected by two-stage cluster sampling, and all the 1523 MSMs from these chambers were surveyed by RRT,The estimator of the population pro- portion an~ its variance were calculated based on corresponding formulae, this sample was used as simulate overall. Monte Carlo simulation investiga- tion was did on two-stage cluster sampling of qualitative sensitive ques-tions, Sample proportion compared with the overall proportion by Z test. Results Sample survey and the actual results were compared by the Z- test （0.05 〈 P 〈 0.99 ）, the difference was not statistically significant. Con- dusion Monte Carlo computer simulation results show that the two-stage cluster sampling method and corresponding formula are feasible, It provides a scientific and reliable survey methods and statistical calculation formula for the investigation of multinomial sensitive questions under complex sam- piing.

关 键 词：蒙特卡洛模拟 分类敏感问题 随机应答技术(RRT) 二阶段整群抽样信度评价 

分 类 号：R195[医药卫生—卫生统计学]