Comparing More Than Two Groups: One-Way ANOVA
摘要
We introduce the analysis of variance framework for decomposing the variation into a part explained by treatments and an unexplained residual part. We introduce several effect size measures and derive the F-test from basic principles. Further, we develop power analysis for the analysis of variance and provide “portable power” methods suitable for quick approximate sample size determination. Finally, we briefly remark on the one-way analysis of variance with unbalanced data.