心理研究元分析报告标准：现状与建议.doc

• 作者： 刘宇 ¹ 陈树铨 ² 樊富珉 ¹ 邸新 ³ 范会勇 ⁴ 封春亮 ⁵ 甘怡群 ⁶ 李会杰 ⁷ 吕小康 ⁸ 徐鹏飞 ⁹ 袁博 ¹⁰ 左西年 ⁷ 胡传鹏 ^11,12
• 作者单位：
  1.  清华大学心理学系

  2.  Department of Clinical and Counseling Psychology, Teachers College, Columbia University

  3.  Department of Biomedical Engineering, New Jersey Institute of Technology

  4.  渤海大学教育与体育学院

  5.  北京师范大学信息科学与技术学院

  6.  北京大学心理与认知科学学院

  7.  中国科学院行为科学重点实验室

  8.  南开大学周恩来政府管理学院社会心理学系

  9.  深圳大学心理与社会学院

  10.  宁波大学心理学系暨研究所

  11.  Deutsches Resilienz Zentrum (DRZ), University Medical Center of the Johannes Gutenberg University

  12.  Neuroimaging Center (NIC), Focus Program Translational Neuroscience (FTN), Johannes Gutenberg University
• 提交时间：2018-09-23

摘要: 元分析通过综合实证研究结果，积累证据、去伪存真，在研究与实践中有着非常重要的作用。元分析所用方法种类繁多，其选择对研究的质量影响较大，当前元分析报告迅猛增长使得其质量更应多加关注。为规范元分析流程、提升报告质量，生物医学（如由医疗卫生领域专家国际性小组制定的PRISMA）及心理学（如美国心理学会推荐的MARS）领域中均制定了元分析报告标准。然而，这类标准开放度不够而且没有中国数据支撑。结合心理学的特点以及最近开放科学的趋势，本文拟提出更具有开放性与透明性的元分析报告标准，共包括7部分30个条目。基于该标准的方法与结果报告部分的条目，本文对2017年国内心理学期刊上发表的元分析报告进行评估。结果发现：已有元分析报告存在方案预注册信息缺乏、研究方法与结果部分内容的混淆、数据来源与分析过程开放性不够等问题。本文提出的元分析报告标准，既可以帮助研究者实施元分析，又有助于提升国内元分析研究的质量。

Abstract: Meta-analysis is a cluster of statistical techniques to quantitatively synthesize empirical findings. By doing meta-analysis, the science community can accumulate existing evidence and distinguish true effects from false positives. Therefore, meta-analytic results were regarded as the high-level evidence in both basic research and clinical settings. Despite its importance, the qualities of published meta-analysis studies, however, is not always warranted. With the rapid increase in the number of published meta-analytic studies, the science community increasingly pays attention to the quality of meta-analytic studies, and several standards have been developed in different fields of science, for example, the Preferred Reporting Items for Systematic Reports and Meta-Analyses (PRISMA) were developed in medical science and the Meta-Analysis Reporting Standards (MARS) were developed in psychology. However, there is no standard for meta-analysis in Chinese, although many meta-analytic studies were published in Chinese every year. To fill this gap, we developed Preferred Reporting Items for Open and Reproducible Meta-analysis (PRIOR-MA). PRIOR-MA integrates items from PRISMA and MARS, as well as the Transparency and Openness Promotion (TOP) guidelines. Therefore, PRIOR-MA provides three advantages: (1) it is based on the most widely used reporting standard for meta-analysis – PRISMA, therefore have recommendation value; (2) it integrates items from MARS so that it has higher practical value than PRISMA for researchers in psychology and related field, and (3) it embraces the latest open science initiative and encourages opening data and analytic code. PRIOR-MA consists of seven sections, and 30 items in total. Based on PRIOR-MA, we evaluated meta-analytic studies published in Chinese psychology journals in 2017, and the results showed that many of these articles did not fully report important information that is necessary for evaluating their qualities, for example, there is lack of protocol and registration information, confusion between method and results, and unavailability of data and analytical code. These results suggested that Chinese researchers need a reporting standard like PRIOR-MA. We recommend Chinese psychology journals adopt PRIOR-MA as part of their policy. We believe applying PRIOR-MA in the journal policy will benefit not only researchers who conduct meta-analysis but also the science community as a whole.

• 来自： 刘宇
• 分类： 心理学 >> 心理统计
• 引用： ChinaXiv:201809.00177 (或此版本 ChinaXiv:201809.00177V1)
  doi:10.12074/201809.00177
  CSTR:32003.36.ChinaXiv.201809.00177.V1
• 推荐引用方式： 刘宇,陈树铨,樊富珉,邸新,范会勇,封春亮,甘怡群,李会杰,吕小康,徐鹏飞,袁博,左西年,胡传鹏.(2018).心理研究元分析报告标准：现状与建议.doc.中国科学院科技论文预发布平台.[ChinaXiv:201809.00177] (点此复制)