Especially when results are borderline significant. For example, if I get a values of F(1,23)=4.11448 and p=0.054 for one main effect and F(1,23)=3.07060 and p=0.093 for the interaction effect, is performing a post hoc test valid?

And, perhaps more importantly, if I get a significant difference in a post hoc test, should I trust it?

More Fernanda Nogueira Lotz Alves's questions See All
Similar questions and discussions