Unfortunately I was not able to test it as I encountered some error messages beforehand. As I had similar problems with the example designs I started a new thread (see https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=spm;9406dd3d.1202 ) about this topic.
---
Concerning contrasts for 2x3x3, I think I have understood the principle more or less. See below. Still, I'm not sure about the t-Tests. If I wanted to compare e.g. the first level of factor 1 with the second level of factor 1, would it be enough to simply write something like [+1 -1] corresponding to the two rows and filling up the rest with zeros, or would I have to take into account some of the other factors/error terms as well?
2x3x3-Within-Subject-ANOVA / F-contrasts
%%%Main factors
I.Cons(1).Groups = {1 2};
I.Cons(1).Levs = [2];
I.Cons(1).ET = 2;
I.Cons(1).mean = 0;
I.Cons(2).Groups = {3 4 5};
I.Cons(2).Levs = [3];
I.Cons(2).ET = 3;
I.Cons(2).mean = 0;
I.Cons(3).Groups = {6 7 8};
I.Cons(3).Levs = [3];
I.Cons(3).ET = 4;
I.Cons(3).mean = 0;
%%Two-way interactions
I.Cons(4).Groups = {1 2 6 7 8};
I.Cons(4).Levs = [2 3];
I.Cons(4).ET = 5;
I.Cons(4).mean = 0;
I.Cons(5).Groups = {1 2 3 4 5};
I.Cons(5).Levs = [2 3];
I.Cons(5).ET = 6;
I.Cons(5).mean = 0;
I.Cons(6).Groups = {3 4 5 6 7 8};
I.Cons(6).Levs = [3 3];
I.Cons(6).ET = 7;
I.Cons(6).mean = 0;
%%%Three-way interaction
I.Cons(7).Groups = {1 2 3 4 5 6 7 8};
I.Cons(7).Levs = [2 3 3];
I.Cons(7).ET = 8;
I.Cons(7).mean = 0;
