One-Way ANOVAs and What's REALLY Important

A one-way ANOVA (like any other statistical analysis) cannot magically transform a flawed study into a sound one. And where can a study be flawed the most? The answer to this question is unrelated to F-tests, equal variance assumptions, effect size indices, or Bonferroni adjustments. That's because the potential worth of any study is connected, first and foremost, to the research question that sets the data-collection-and-analysis wheels in motion. If the research question is silly or irrelevant, a one-way ANOVA cannot make the study worthwhile. Hence, don't be impressed by researchers who use one-way ANOVAs until you have first considered the merits of the research questions being addressed.

(From Chapter 12, p. 351)

Copyright © 2003 Schuyler W. Huck
All rights reserved.