Dear Jeanette,
Thanks a lot for your explanation. Yeah it’s 2x2x2 not 3x2. Sorry for wrong typing.
By the way, it seems some mistake in the number?!
Just like you said, presumably at the first level the order of the 8 EVs is:
A1B1C1
A1B1C2
A1B2C1
A1B2C2
A2B1C1
A2B1C2
A2B2C1
A2B2C2
For AB interaction, if the contrast (A1B1-A1B2)-(A2B1-A2B2) (collapse over both levels of C) is correct at the first level (I think it is correct), the AB interaction within level C1 and C2, respectively, should be
AB w/in C1= [1 0 -1 0 -1 0 1 0]
AB w/in C2= [0 1 0 -1 0 -1 0 1]
(They are previously represented as:
AB w/in C1= [1 0 -1 0 1 0 -1 0]
AB w/in C2= [0 1 0 -1 0 1 0 -1] )
The 3-way interaction:
ABC =[1 -1 -1 1 -1 1 1 -1]
(They are previously represented as:
ABC =[1 -1 -1 1 1 -1 -1 1] )
For the overall AB interaction:
AB =[1 1 -1 -1 -1 -1 1 1]
(They are previously represented as:
AB =[1 1 -1 -1 1 1 -1 -1] )
Am I correct?
Moreover, can we say that this is a better (or correct) way to run a 2x2x2 repeated measures anova rather than my original higher-level setup in feat? I did that because the feat website's introduction of group statistics (though it's 2x2).
Best,
Mark
|