A Primer on Testing the Minimum Difference among Standardized Effects

在變異數分析中，有兩種衡量多群組標準化效用差異的綜合指標，分別是簡稱為訊號對雜訊比值及標準化效用平方之平均值的開方根。雖然二者的用途簡單明瞭，但相關的最低差異檢定卻鮮少討論。基於研究規範對效果量呈現和最低差異檢測的提倡，本文介紹並評比相關基本差異而非毫無差異之檢定力和樣本數方法。相關資料分析和研究規畫的軟體也一併開發，藉以增加應用便利性。

關鍵字

效果量；檢定力；樣本數；標準化期望值差異

並列摘要

Two different average measures of the sum of squared standardized effects have been proposed to represent the degree of difference among treatment effects in the analysis of variance design. One is the square root of the signal-to-noise ratio and the second is the root mean square standardized effect. Despite the obvious usage of the two measures, the associated test procedures and properties for detecting a minimal important difference among standardized means are not well illustrated. In view of the advocated practice of reporting effect sizes and testing minimal effects in quantitative research, this paper presents and compares the power and sample size procedures for testing the hypothesis that treatments have negligible effects rather than that of no difference. To enhance the practical usefulness, the corresponding computer algorithms for data analysis and design planning are also developed.

並列關鍵字

effect size ； power ； sample size ； standardized mean difference

參考文獻

Alhija, F. N. A.,Levy, A.(2009).Effect size reporting practices in published articles.Educational and Psychological Measurement.69,245-265.

Google Scholar

American Educational Research Association(2006).Standards for reporting on empirical social science research in AERA publications.Educational Researcher.35,33-40.