部分核实数据下基于比例差的置信区间宽度的样本量确定  

Sample Size Determination Based on the Width of Confidence Intervals for Proportion Difference with Partially Validated Series

作　　者：王黎明 刘青松 张晓良 邱世芳 WANG Liming;LIU Qingsong;ZHANG Xiaoliang;QIU Shifang(College of Science,Chongqing University of Technology,Chongqing 400054,China)

机构地区：[1]重庆理工大学理学院,重庆400054

出　　处：《重庆理工大学学报（自然科学）》2021年第5期280-294,共15页Journal of Chongqing University of Technology：Natural Science

基　　金：国家自然科学基金项目(11871124,11471060);重庆市基金研究与前沿探索项目(cstc2018jcyjAX0241)。

摘　　要：基于无金标准的部分核实数据,从比例差的区间宽度的角度给出了比较两组患病率是否有显著性差异所需要的样本量。在2种不同的二重抽样模型下,分别基于Wald、对数变换、Logit变换、Score和似然比检验统计量构造的置信区间给出了在给定置信水平下区间宽度控制在指定范围内的样本量估计方法。模拟研究了在估计的样本量下区间估计的经验覆盖概率和经验覆盖宽度,并基于疟疾数据利用所提出的方法进行了验证。结果表明:所提出的样本量估计方法是准确有效的。Based on the partially validated series without gold standard test,this article investigates the sample size determination for the significant testing of two disease prevalence rates in terms of interval width for the proportion difference.Based on Wald confidence interval(CI),logtransformation CI,Logit transformation CI,Score CI and likelihood ratio CI under two different double-sampling models,the approximate sample sizes which can control the width of a confidence interval with a pre-specified confidence level are developed.Simulation studies are conducted to investigate the empirical coverage probability(ECP)and empirical coverage width(ECW)of all CIs under the estimated sample size,and the results show that sample size estimation methods proposed in this paper are accurate and effective.Finally,the malaria data is used to illustrate the proposed methods.

关 键 词：置信区间 比例差 无金标准部分核实数据 区间宽度 样本量 

分 类 号：C811[社会学—统计学] O212.2[理学—概率论与数理统计]