2*2 ANOVA with no cell

[Josh]

Hi, I have a small issue with running an ANOVA. I have 2*2 cell experiment with one cell removed due to irrelevance. There are still 2 independent variables. Can the ANOVA still be run without some issues?

[Rich Ulrich]

On Mon, 10 Apr 2023 02:52:42 -0700 (PDT), Josh

Post by Josh
Hi, I have a small issue with running an ANOVA. I have 2*2 cell experiment with one cell removed due to irrelevance. There are still 2 independent variables. Can the ANOVA still be run without some issues?

Mechanically: There can be two tests of main effects but no d.f.
between groups left over for testing an interaction. Tell ANOVA
to ignore interactions.

The more informative approach is apt to be: Do oneway ANOVA on
three groups; ask for the 3 contrasts between pairs as separate tests.
Be especially wary of drawing conclusions based on 'signficance' if
the Ns are not equal.