Dear Patricia,
This is not what we would recommend for this analysis because
the age values are not demeaned. As there is some difference
between the mean age of each group (16.39 16.59 19.07 17.89)
then this will be driven by some combination of age *and* group
difference, which is not what you are likely to want. It will also
affect the group difference contrasts. So I strongly recommend
that you demean each group.
Also, splitting age between the different groups allows you to fit
different slopes (amounts of variation of quantity with age) for
each group. Is that what you want, or do you just want to account
for age as a covariate of no interest across all groups? Given
that you want a different mean age in each group then doing
the latter is difficult, but you must be aware of what you are assuming
when you are doing the former.
Jeanette Mumford has a very good discussion of the various
demeaning issues at:
http://mumford.fmripower.org/mean_centering/
As for your contrast - C1 asks the question of whether the
average quantity (fMRI effect size, or FA, or whatever you are
testing) is different from zero for the first group only. If you
want to know whether groups differ from each other then you
need to have differential contrasts such as:
-1 1 0 0 0 0 0 0
-1 0 1 0 0 0 0 0
-1 0 0 1 0 0 0 0
0 -1 1 0 0 0 0 0
0 -1 0 1 0 0 0 0
0 0 -1 1 0 0 0 0
which ask the questions:
Is the average quantity in group 2 greater than the average quantity in group 1?
Is the average quantity in group 3 greater than the average quantity in group 1?
Is the average quantity in group 4 greater than the average quantity in group 1?
Is the average quantity in group 3 greater than the average quantity in group 2?
Is the average quantity in group 4 greater than the average quantity in group 2?
Is the average quantity in group 4 greater than the average quantity in group 3?
Note that I've given the 8 column version, which is not to meant to mean that
I agree with splitting the age covariate, I'm just trying to make it more similar
to the ones you provided. This depends on the points I discussed earlier,
but the contrasts can be converted to the 5 column version by dropping the last
3 zeros.
I hope this helps.
All the best,
Mark
On 1 Aug 2011, at 16:05, SUBSCRIBE FSL Patricia Pires wrote:
> Dear experts,
>
> i want to analize in Glm 4 different groups (Controls, Mild, Medium and Severe disease).
> Also, i have included age covariate in EV's because this was significant.
> This is the matrix:
>
> /Matrix
> 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 6.300000e+01 0.000000e+00 0.000000e+00 0.000000e+00
> 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 5.900000e+01 0.000000e+00 0.000000e+00 0.000000e+00
> 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 5.800000e+01 0.000000e+00 0.000000e+00 0.000000e+00
> 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 7.400000e+01 0.000000e+00 0.000000e+00 0.000000e+00
> 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 6.700000e+01 0.000000e+00 0.000000e+00 0.000000e+00
> 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 6.000000e+01 0.000000e+00 0.000000e+00 0.000000e+00
> 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 7.400000e+01 0.000000e+00 0.000000e+00 0.000000e+00
> 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 6.900000e+01 0.000000e+00 0.000000e+00 0.000000e+00
> 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 6.400000e+01 0.000000e+00 0.000000e+00 0.000000e+00
> 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 6.100000e+01 0.000000e+00 0.000000e+00 0.000000e+00
> 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 6.000000e+01 0.000000e+00 0.000000e+00 0.000000e+00
> 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 7.100000e+01 0.000000e+00 0.000000e+00 0.000000e+00
> 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 6.400000e+01 0.000000e+00 0.000000e+00 0.000000e+00
> 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 7.400000e+01 0.000000e+00 0.000000e+00 0.000000e+00
> 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 6.600000e+01 0.000000e+00 0.000000e+00
> 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 6.900000e+01 0.000000e+00 0.000000e+00
> 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 6.800000e+01 0.000000e+00 0.000000e+00
> 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 6.300000e+01 0.000000e+00 0.000000e+00
> 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 6.600000e+01 0.000000e+00 0.000000e+00
> 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 6.300000e+01 0.000000e+00 0.000000e+00
> 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 6.600000e+01 0.000000e+00 0.000000e+00
> 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 7.100000e+01 0.000000e+00 0.000000e+00
> 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 6.400000e+01 0.000000e+00 0.000000e+00
> 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 6.200000e+01 0.000000e+00 0.000000e+00
> 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 7.000000e+01 0.000000e+00 0.000000e+00
> 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 6.200000e+01 0.000000e+00 0.000000e+00
> 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 6.900000e+01 0.000000e+00 0.000000e+00
> 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 7.000000e+01 0.000000e+00 0.000000e+00
> 0.000000e+00 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 8.300000e+01 0.000000e+00
> 0.000000e+00 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 7.800000e+01 0.000000e+00
> 0.000000e+00 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 7.900000e+01 0.000000e+00
> 0.000000e+00 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 7.800000e+01 0.000000e+00
> 0.000000e+00 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 7.200000e+01 0.000000e+00
> 0.000000e+00 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 7.400000e+01 0.000000e+00
> 0.000000e+00 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 7.800000e+01 0.000000e+00
> 0.000000e+00 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 7.200000e+01 0.000000e+00
> 0.000000e+00 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 7.900000e+01 0.000000e+00
> 0.000000e+00 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 7.500000e+01 0.000000e+00
> 0.000000e+00 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 7.900000e+01 0.000000e+00
> 0.000000e+00 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 7.500000e+01 0.000000e+00
> 0.000000e+00 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 6.900000e+01 0.000000e+00
> 0.000000e+00 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 7.700000e+01 0.000000e+00
> 0.000000e+00 0.000000e+00 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 7.700000e+01
> 0.000000e+00 0.000000e+00 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 7.400000e+01
> 0.000000e+00 0.000000e+00 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 7.000000e+01
> 0.000000e+00 0.000000e+00 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 6.400000e+01
> 0.000000e+00 0.000000e+00 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 7.300000e+01
> 0.000000e+00 0.000000e+00 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 5.900000e+01
> 0.000000e+00 0.000000e+00 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 7.800000e+01
> 0.000000e+00 0.000000e+00 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 6.700000e+01
> 0.000000e+00 0.000000e+00 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 7.900000e+01
> 0.000000e+00 0.000000e+00 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 7.100000e+01
> 0.000000e+00 0.000000e+00 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 7.700000e+01
> 0.000000e+00 0.000000e+00 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 7.300000e+01
> 0.000000e+00 0.000000e+00 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 7.600000e+01
> 0.000000e+00 0.000000e+00 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 6.400000e+01
>
> And the contrast:
>
> /Matrix
> 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00
> 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00
> 0.000000e+00 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00
> 0.000000e+00 0.000000e+00 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00
>
> Matrix and contrast are well built?
> If there is any difference in contrasts (e.g. C1) does it means this group differs from all others?
>
> Thanks very much,
>
> Patricia.
>
|