Dear Experts,
I define a full factorial design A(a1,a2)xB(b1,b2)in spm5. I assign each con
image to the corresponding factor level:
a1_b1: 1 0 0 0
a1_b2: 0 1 0 0
a2_b1: 0 0 1 0
a2_b2: 0 0 0 1.
Because the factors are within-subjects I set "no" for Indipendence for both
factors. In the end I get a design matrix with four regressors without a
further regressor which models subject-effect (I'm interest in modeling
subject-effect to increase statistical power, I'm not interest in it
directly).
My questions are: Is this analysis really within-subjects or is it
between-subjects because I'm not modeling subject-effect? Is it sufficient
to specify "no" for Indipendence for both factors to run a within-subject
analysis or shall I specify a further factor (for example subject or
something like)?
Many thanks.
Federico.
--
No virus found in this outgoing message.
Checked by AVG Free Edition.
Version: 7.5.441 / Virus Database: 268.18.3/696 - Release Date: 21/02/2007
15.19
|