Kanellopoulou Afroditi, Dwan Kerry, Richardson Rachel
Methods Support Unit, Evidence Production and Methods Directorate Cochrane London UK.
Department of Hygiene and Epidemiology Faculty of Medicine University of Ioannina Ioannina Greece.
Cochrane Evid Synth Methods. 2025 Jan 29;3(2):e70013. doi: 10.1002/cesm.70013. eCollection 2025 Mar.
The aim of this article is to present the most common statistical errors in meta-analyses included in systematic reviews; these are confusing standard deviation and standard error, using heterogeneity estimators for choosing between a common-effect and random-effects model, improper handling of multiarm trials, and unnecessary and misinterpreted subgroup analyses. We introduce some useful terminology and explain what authors can do to avoid these errors and how peer reviewers can spot them. We have also developed a micro-learning module to provide practical hands-on tutorial.
本文旨在介绍系统评价中纳入的Meta分析中最常见的统计错误；这些错误包括混淆标准差和标准误、使用异质性估计量来在固定效应模型和随机效应模型之间进行选择、对多臂试验处理不当，以及进行不必要且解读错误的亚组分析。我们引入了一些有用的术语，并解释作者可以采取哪些措施来避免这些错误，以及同行评审人员如何发现这些错误。我们还开发了一个微学习模块，以提供实际操作教程。
