Re: Simple question

Enrique Cruiz wrote:

I have a question concerning the two-way anova analysis though. The
obtained p-values are very small, which is a strong indication that the
average varies from one row/column to another. If I understand well,
that is not an absolute proof that they are different. Is there a way
to prove that they are different, i.e. to prove that the surface is not
flat?

No, there is no absolute proof.


A last question for my own curiosity: why the interaction (XY) does not
have its own p-value?

The interaction has no F because there is no suitable denominator
("error term"), and no p because there is no F.

.