摘要
2×2列联表的统计推断在现代医学,各种临床试验以及现代生物学的研究中通常是感兴趣的问题,其中归因风险的统计推断经常出现在流行病学和健康服务研究中.本文我们针对2×2列联表中归因风险的统计推断给出了一种新的方法,利用2×2列联表中边际和条件概率提出了Score检验,给出了建立在置信区间之上的Score检验的描述.并对Score检验和似然比检验的效果进行评价,依据覆盖概率、左尾错误率及右尾错误率得出结论:Score检验与似然比检验相比更有效.
Statistical inference for correlated 2×2 table is frequently of interest in modern medicine study, various clinical trails and modern biological studies, in which attributable risk is often used in epidemiology and health services reasearch. In this article, we develop a new approach to statistical inference for attributable risk in a 2×2 table, we propose a score test of hypotheses pertaining to the marginal and conditional probabilities in a 2 × 2 table. Scoretest based confidence interval will alo be described. In addition, the effects about the score test and likehood ratio test are also evaluated. According to the coverage pribability, the left tail Error Rate and Right-tail Error rate, the performance of the score test is better than that of the existing likehood ratio test.
出处
《数学的实践与认识》
CSCD
北大核心
2009年第8期188-196,共9页
Mathematics in Practice and Theory
基金
河南省科学发展计划基础与前沿技术研究项目(082300460190)
