本研究之目的有三:(1)對國內歷年閱讀理解教學研究成果,做有系統的整理;(2)使用統合分析法瞭解國內閱讀理解教學研究對閱讀理解、相關閱讀能力,以及閱讀情意表現的成效;(3)並探討影響閱讀理解教學效果的中介變項。 本研究使用電腦資料庫檢索、溯洄法,以及相關期刊回顧等方式,蒐集到26篇閱讀理解教學的實徵性研究。依Hedges與Olkin (1985)統合分析的技術來計算加權平均效果量d+值,並以其「類別模式」尋找影響閱讀理解效果的中介變項。根據研究結果,可獲得以下結論: 1.國內歷年閱讀理解教學研究方案,對學生閱讀理解能力的提昇,具有中等的助益效果;在立即效果和持續效果上,亦有中等以上的效果。在測驗編製者方面,閱讀理解教學研究的效果量並沒有差異存在。 2.國內歷年閱讀理解教學研究方案,對學生閱讀相關能力及閱讀情意表現亦有中等效果。 3.在中介變項的尋找上,閱讀教學法,閱讀教學者、學生學習階段、學生學習能力是影響閱讀解表現的中介變項。 最後依據研究結果加以討論,提出閱讀理解教學以及統合分析研究的建議。
The purposes of this study were: (1) to integrate the outcomes of experimental studies in reading comprehension instruction; (2) to investigate the overall weighted mean effect sizes of reading comprehension abilities, related reading abilities, and affective performance in reading with meta-analysis method; (3) to find out moderator variables that may influence the effects of reading comprehension instruction. It began with a systematic search of the literature from electronic database, the ancestry approach and related Journals to locate 26 reading comprehension instruction studies. The effect sizes were calculated by Hedges and Olkin’s (1985) procedure. Further, this studies tested all coded features by ‘Categorical Model ‘to find moderator variables. The results of this study were as follows: 1.The overall weighted mean effect size of reading comprehension instruction studies was middle magnitude. The immediate and retentive effect sizes were also middle magnitude. The weighted mean effect sizes of researcher-developed tests and non-researcher-developed tests were not different. 2.The overall weighted mean effect sizes in reading related abilities and reading affective performance were middle magnitude. 3.The moderator variables of reading comprehension instruction studies were reading instruction methods, reading instructors, learning stages of the students, and the competence of the students. Based on the results of the study, the researcher also provided a discussion and made some recommendations.