本文主要以教育學者對教育學之各領域所做統合分析之結果加以綜合，不少統合分析除了總平均效應量之外，還作細部分析，檢定細分變項(Moderator variables)是否也達顯著。由於有些細分變項的效應量數目太少(例如少於10個)，故不加以列述。本文研究方法是對重要期刊，尤其是Review of Educational Research做地毯式搜尋，並用電腦網路之教育期刊以Meta-analysis為關鍵自查出教育學門統合分析之研究報告。本文除介紹各統合分析之自變項與依變項之操作型定義之外，還將各研究報告所列之效應量求出總平均效應量、標準差、標準誤及t值，看總平均效應量是否顯著的不等於零。如果原來的研究報告已算出t值，則從其值。本文結果擬形成「教育科學導論」，以供將來教育學者重複驗證，並對本文所沈澱之教育學知識體加以增修，使教育學真正成為一門科學。 This study synthesizes the meta-analyses done by educational empirical researchers in various fields of education. A lot of meta-analyses calculated not only average mean of effect sizes, but also average means of moderator variables. But many of moderator variables were not adopted by this study, because of too few number, say less than 10. The method used for synthesizing the results of meta-analyses was using t-test to examine whether the mean effect size was significantly different from zero. The articles of meta-analysis were searched from the important educational journals, such as Review of Educational Research, as well as from computer data bases, such as EBSCOhost. This study has described the operational definitions of each variable and calculated from the available N, mean, and standard deviation, the standard error and t-value of each meta-analysis. If the t-value has been presented originally in article, than it would be adopted directly. The contribution of this study would be the formation of a textbook "introduction to the educational science", which is characterized by the precipitated empirical knowledge body of education.