King Seth, Wang Lanqi, Datchuk Shawn M, Rodgers Derek B
University of Iowa, Iowa City, USA.
University of Nebraska-Lincoln, USA.
J Learn Disabil. 2023 May-Jun;56(3):210-224. doi: 10.1177/00222194221077688. Epub 2022 Feb 14.
Learning disabilities (LD) may affect a range of academic skills but are most often observed in reading. Researchers and policymakers increasingly recommend addressing reading difficulties encountered by students with LD using evidence-based practices, or interventions validated through multiple, high-quality research studies. A valuable tool in identifying evidence-based practices is the meta-analysis, which entails statistically aggregating the results obtained through primary studies. Specific methods used in meta-analyses have the potential to influence their findings, with ramifications for research and practice. This review assessed the methodological features of the systematic reviews and analytic procedures featured in meta-analyses of reading intervention studies that included students with LD written between 2000 and 2020. Identified articles ( = 23) suggest that meta-analyses have become more prevalent and transparent over time, notwithstanding issues related to publication bias and the opacity of coding procedures. A discussion of implications follows a description of results.
学习障碍（LD）可能会影响一系列学术技能，但最常出现在阅读方面。研究人员和政策制定者越来越多地建议采用循证实践，即通过多项高质量研究验证的干预措施，来解决有学习障碍的学生所遇到的阅读困难。荟萃分析是识别循证实践的一个有价值的工具，它需要对通过初步研究获得的结果进行统计汇总。荟萃分析中使用的具体方法有可能影响其研究结果，并对研究和实践产生影响。本综述评估了2000年至2020年间撰写的、包含有学习障碍学生的阅读干预研究的荟萃分析中系统评价的方法学特征和分析程序。所识别的文章（n = 23）表明，尽管存在与发表偏倚和编码程序不透明相关的问题，但随着时间的推移，荟萃分析已变得更加普遍和透明。在对结果进行描述之后，紧接着是对其影响的讨论。
