Hi all,
Hi Stefan,
I have a question related to that of Boris (see below) concerning the
flexible factorial design.
For the paired t-test, the number of subjects corresponds to the number of
conditions in the design description (while the number of actual conditions
isn't mentioned at all...). Further, the subject-specific effects (mean of
each subject) appear in the design matrix without explicitly specifying them
(see attached file: paired_t-test.ps).
In contrast, in the flexible factorial design specification the number
of conditions corresponds to the number of columns in the design matrix
(number of conditions plus number of subjects). Further, the
subject-specific effects (mean of each subject) appear in
the design matrix only when they are explicitly specified as an
additional condition (see attached file: flexible_factorial.ps).
Q1: Am I doing the correct thing by specifying the subjects as an extra
CONDITION?
Q2: If so, in the paired t-test description number of subjects=number of
conditions, while in the flexible factorial design description number of
subjects+number of conditions=number of conditions.
Is this really how things are supposed to be?
Q3: In SPM2 the subject-specific effects appeared in the design matrix
without explicitly specifying them. Has this changed in SPM5 or am I doing
something wrong?
Many thanks for your help,
Mira
Hi Boris,
I had a look at the paired t-test in SPM5. What you get is correct.
The subject-specific effects (mean of each subject) appears as the 1st
block of the design matrix. The 2nd block are the (interesting)
condition-specific effects. So you were right in your intuition: in SPM2
it was the other way around.
If you have 7 pairs, your contrast for a difference between condition 1
and 2 should be (SPM5...):
[zeros(1, 14) -1 1]
Stefan
Hi all,
I tried to setup a paired t-test which shouldn't be that hard, but if I
define 7 pairs, there are first the subjects the design matrix and the
last two columns are the conditions.
The parameters say that I have 7 conditions...
Is this correct? When setting up the contrast should I only use the last
columns with the conditions? In SPM2 the design matrix was the other way
around, first the conditions and than the subjects...
Many thanks
boris
|