不同實驗設計之標準化平均數差異效果量整合探討

Combining Effect Size Estimates across Different Experimental Designs in Meta-Analysis

在國內後設分析研究中，對於結合來自不同實驗設計所延伸之標準化平均數差異效果量議題，多鮮少著墨，而此舉就形同間接捨棄某些較為適當之效果量，再加諸原始分數、共變項調整後分數及進步分數等相關效果量計算公式來源不僅複雜且過去文獻或多或少存在某些錯誤，因此，本文目的在於系統性評閱及歸納各分數量尺常見統計量與標準化平均數差異效果量之轉換，最後，研究者開發一套免費的「SMD Calculator」軟體，並搭配數項範例說明及建議，以便未來使用者能輕易上手運用。

Combining standardized mean difference (SMD) from different experimental designs were rarely discussed， especially domestic meta-analytic studies. This would ignore some more suitable effect sizes for further meta analysis. Besides， the formula of SMD deriving from raw score， covariate adjusted score and gain score are not only complicated， but also some mistakes made in the past literature. The aim of this study is to conduct a systematic review of methods or formula of effect size dealing with results from different experimental designs. Finally， software， named SMD Calculator was designed and demonstrated with three examples and suggestions for future usage are proposed.