2by2 ANOVA post-hoc correction?


New Member
I would like a second opinion on the claim that if you have a 2 by 2 ANOVA (2 factors, 2 levels each) then for pairwise comparisons (simple effects) you do not need to introduce any correction like LSD (or Bonferroni or Šidák) ?



TS Contributor
With a 2x2, the interaction includes consideration of 4 cell means, so there are 6 possible 2-way comparisons of these 4 means. If you plan to do all 6, then it would be a good idea to apply a correction to the alpha level to protect against increasing Type I errors.