中文体育类核心期刊元分析的选题和方法学问题

Topics Selection and Methodological Issues of Meta-analysis in Chinese Sports Core Journals

目的:探讨中文体育类核心期刊中元分析的选题和方法学问题,进行评价和举例分析,提出相应建议。方法:以中国知网数据库中16本中文体育类核心期刊发表的元分析研究为对象,依据Cochrane手册和经典元分析文献探讨体育科学领域元分析存在的选题问题,并评估所选元分析研究的方法学问题。方法学问题的主要评估指标为:1)是否存在异常值(效应量>3);2)是否将标准误换算成标准差;3)是否忽略了研究内的相关性;4)是否忽略了研究内的方差;5)是否关注研究内而非研究间;6)报告多个效应量的研究是否仅提取某一个效应量。结果:在选题方面,71.1%的中文元分析的研究主题与既有英文元分析相似,其中27.3%的中文元分析题目名称与英文元分析几乎相同。在方法学方面,82.2%的元分析至少存在1个方法学问题,这些问题可能导致43.8%的元分析结果发生变化。结论:元分析应以解决具有实际意义的研究问题为出发点。在进行元分析时,研究者应仔细检查原始数据的报告形式,并采用多种效应模型、分析方法、合并策略、元分析软件及展示方式,通过团队合作确保元分析结果的稳定性。更重要的是,需在方法部分详细介绍和解释分析过程,以提高研究的透明度和可重复性。

Objective:To explore the topic selection and methodological issues in meta-analyses which published in Chinese sports core journals,and to provide recommendations for the improvement of research quality.Methods:Meta-analyses published in 16 Chinese sports core journals were retrieved from the CNKI database.Based on the Cochrane Handbook and classic meta-analysis literature,this study discussed the topic selection issues and evaluated the methodological issues in the selected meta-analyses,the items including:1)The presence of outliers(effect size>3);2)conversion of standard errors to standard deviations;3)consideration of within-study correlations;4)oversight of within-study variances;5)focus on within-study rather than between-study differences;6)extraction of only one effect size from multiple effect sizes in one study.Results:In terms of topic selection,71.1%of Chinese meta-analysis research topics were similar to existing English meta-analysis,with 27.3%of them were nearly identical to their English counterparts.In terms of methodological problem,82.2%of the studies had at least one methodological issues,which could potentially alter 43.8%of the meta-analyses results.Conclusions:The aim of meta-analyses should focus on the resolution of major research problems with practical significance.When performing a meta-analysis,researchers should carefully check the reporting form of original data and employ various effect models,analysis methods,combining strategies,meta-analysis software,and presentation methods,and collaborating with other researchers to ensure the stability of meta-analysis results.More importantly,the author need to detail and explain the analysis process in the methods section to enhance the transparency and reproducibility of the study.