多个样本率两两比较χ~2检验界值的研究  

Study on χ~2—test Critical Values for the Multiple Comparison among Several Sample Rates

作　　者：罗文海[1] 张世增 高永[1] 郝恩柱[1] 唐军[1] 袁玉华[1] 徐天和[1] 万巧云[1] 曲巍 

机构地区：[1]滨州医学院,山东省滨州市256603 [2]山东省滨州地区卫生防疫站

出　　处：《中国公共卫生》2001年第4期372-374,共3页Chinese Journal of Public Health

摘　　要：目的　为探索更加完善的多个样本率两两比较的统计分析方法。方法　应用蒙特卡洛法在计算机上产生二项分类的样本数据。在反复分析运算之后 ,确定正式实验条件为 :总体率为 0 2 5 ,样本含量为 15 0 ,样本个数为 3、4、5、6、7、8、9、10 ,抽样次数为 10 0 0 0 0 0 ,概率为 0 90、0 75、0 5 0、0 40、0 30、0 2 0、0 10、0 0 5、0 0 2 5、0 0 1、0 0 0 5。结果　在大量正式运算之后 ,研制出多个样本率两两比较 χ2 检验界值。结论　合理的分析表明 ,本界值准确性高。Objective The study designed to approach a more complete statistical method of the multiple comparison among several sample rates obtained from binomial distribution data with Mote Carlo method on a computer.Methods After calculations and analyses,the formal experimental factors were determined.The population rate:0 25;the sample numbers:3,4,5,6,7,8,9,10;the sampling times:1000000;the probabilities:0 90,0 75,0 50,0 40,0 30,0 20,0 10,0 05,0 025,0 01,0 005.Result After a large number of calculations the χ 2—test critical values for the multiple comparison among several sample rates were brought out.Conclusion Reasonable analyses show that the critical values have high accuracy.

关 键 词：样本率 两两比较 X^2检验 界值 卫生统计 

分 类 号：R195.1[医药卫生—卫生统计学]