On P-value Combination Procedures
作者机构:University of Chinese Academy of Sciences School of Statistics Capital University of Economics and Business Academy of Mathematics and Systems Science Chinese Academy of Sciences
出 版 物:《Acta Mathematica Sinica,English Series》 (数学学报(英文版))
年 卷 期:2024年
核心收录:
学科分类:02[经济学] 0202[经济学-应用经济学] 020208[经济学-统计学] 07[理学] 0714[理学-统计学(可授理学、经济学学位)] 070103[理学-概率论与数理统计] 0701[理学-数学]
基 金:Supported by the CAS Project for Young Scientists in Basic Research (Grant No. YSBR-034) National Nature Science Foundation of China (Grant No. 12201432)
摘 要:Combining p-values is a well-known issue in statistical inference. When faced with a study involving m p-values, determining how to effectively combine them to arrive at a comprehensive and reliable conclusion becomes a significant concern in various fields, including genetics, genomics, and economics, among others. The literature offers a range of combination strategies tailored to different research objectives and data characteristics. In this work, we aim to provide users with a systematic exploration of the p-value combination problem. We present theoretical results for combining p-values using a logarithmic transformation, which highlights the benefits of this approach. Additionally, we propose a combination strategy together with its statistical properties utilizing the gold section method, showcasing its performance through extensive computer simulations. To further illustrate its effectiveness, we apply this approach to a real-world scenario.