方差分析效果大小报告的新指标

心理学期刊论文中可重复性不高的现象,原因之一是研究结果的效果普遍较小。并且在报告效果大小的文章中,效果大小指标使用不当。在方差分析中最经常报告的是η2和η2_p,但是在不同的研究设计中,这些效果大小是无法直接进行比较的。广义eta方(η²_G)是近年来新出现的一种新的效果大小指标,可克服的η2和η2_p不足,灵活处理重复测量等多种研究设计下个体差异的计算问题,实现跨研究设计效果大小的可比性。论文结合实例介绍了η²_G的原理和计算方法,并对其优缺点、使用和报告等问题进行了讨论。研究人员在报告效果大小时要考虑到不同的研究设计和研究假设,并选择恰当的指标防止过高估计效果大小。

A New Effect Size Indicator of Analysis of Variance

One of the reasons for the low repeatability in psychological research is that the effect size are generally small.And in the article reporting the effect size,the effect size indicator is not used properly.The most frequently reported indicators for ANOVA are eta squared and partial eta squared.But these two indicators are not comparable across different research designs.Generalized eta squared that has emerged in recent years can overcome traditional indicators' shortcomings,and achieve comparability by considering individual differences in many kinds of research designs.The present paper introduces what is generalized eta squared and how to calculate it with examples.Advantages,disadvantages and reporting of it are also discussed.Researchers should consider different research designs and research hypotheses and choose appropriate size measures to prevent overestimation.