心理与教育元分析中相关系数的合并

Combining Effect Sizes Based on Correlation in Psychology and Education Research

国内对效应量r的合并与分析缺乏系统性的介绍,加上没有简便的统计软件,使该领域的研究受到了很大的限制。对效应量r的元分析一般有5个步骤:指标的获得、效应量的合并、异质性分析、发表偏倚分析和异质性原因探索。前4步可通过EXCEL完成,最后一步可通过HLM和SPSS完成。

It is rare to see meta -analysis based on Correlation r in Chinese due to lack of systematic introduction and statistical soft- ware of this filed. Generally,five steps should be took to do a meta - analysis on r,that is selecting effect size r,combining effect size r, heterogeneity test, analysis of publication bias and exploring factors affecting variance of effect size. One can use EXCEL to finish the former four steps and HLM and SPSS to the last step.