促進中文閱讀理解教學成效量化研究統合：調節變項影響與評估

Abstract

以實徵研究證據為基礎，探討促進學生中文閱讀理解教學議題逐漸受到重視，而使用量化統合技術推動本議題的研究，亦愈趨興盛，然而，受限研究數量，鮮少文獻針對調節變項進行深入分析，據此，本研究目的除了分析促進中文閱讀理解教學之量化統合效果外，更聚焦於探討教學介入、參與者背景特徵、個別研究品質評估等因素，對於教學效果量可能產生的調節影響。本研究經統合分析與刪除極端值，發現國內促進中文閱讀理解教學平均效果量，分別具有+0.43立即效果（168筆）與+0.39 延宕效果（46筆），為近中等程度教學成效，續經調節變項分析，發現文本重複閱讀具有顯著偏低平均教學效果、使用預測策略則具有較高教學成效，同時，參與者為特殊生或國中階段學生、使用研究者自編評量工具、實驗設計：其他與納入個別研究的品質較差者，皆可產生顯著偏高教學平均效果量。最後針對分析結果提出幾點未來實務與研究建議。

Effective instructions for reading comprehension have been studied over the past few decades,and synthesis of quantitative research has also been conducted on such instructions to analyze theireffect considering cumulative findings across relevant empirical studies. However, due to the limitednumber of included studies, more research into moderating effects is required. This study conducteda meta-analysis of the moderating effects of characteristics of instructional intervention, participants,and methodologies. A systematic literature review was undertaken to identify and collect relevantstudies. General keywords were searched for in a Taiwanese literature database. A total of 880Chinese reading-related articles were retrieved, and 173 articles were formally included on the basisof five inclusion and exclusion criteria. After the meta-analysis, an overall average effect size of+0.43 and +0.39 was obtained separately for immediate and follow-up effects. In the analysis ofmoderating effects, the findings indicated a significantly low average effect size for repeated readinginstruction but a large average effect size for the use of a prediction strategy or research-developedmeasurement tool, struggling or middle-school-aged participants, and studies with poor experimentaldesign or failure to meet relevant standards. Finally, some suggestions and implications are given forfuture research.