Hi Thomas,
Thanks for replying.
On 22 Feb 2008, at 11:33, Thomas Nichols wrote:
> Dear Colm,
>
> You very nearly have it, just one tweak and you should be there...
Thanks, that's a relief!
>
> I have 3 groups (ctrl, long, and short) of 9 subjects each. I'm
>> attempting to ascertain whether there are grey matter differences
>> between these three groups.
>>
>> I've set up the design matrix by following the FEAT example fro F
>> tests with 3 groups of subjects. So I have one input for each
>> subject, the first 9 are assigned to group 1, second 9 to group 2 and
>> final 9 to group 3.
>>
>
> I've added 3 EVs to the matrix with EV1 being 1 only for control
>> subjects, EV2 is 1 only for long subjects and EV3 is 1 only for short
>> subjects.
>
>
> Good so far....
>
> I'v created 3 contrasts: ctrl - short, ctrl - long and long
>> - short and asked for 3 F-tests: one for each contrast in the Glm
>> window.
>>
>
> The t contrasts are correct, but as you describe it, the F
> contrasts aren't
> right. However, the design.con and the design.fts you sent *are*
> correct.
I realised after I sent the mail that I'd chosen the files from a
different directory and that they didn't match up with the story in
my email.
> The issue is that you are interested in *any* differences between the
> groups, and so 3 separate F-tests, one based on each t-test won't
> give you
> this. You want what is what you have actually have, *1* F-test
> based on the
> 2 contrasts (just 2 t-contrasts is sufficient, because there are
> only 2 DF
> in the differential effect of interest).
So given that I've got 3 groups I can only do 2 planned contrasts? So
I could perform the contrasts as in the design.fts and design.con
files but couldn't, for example, perform long - short on the same
test? Aside from running a separate invocation ofo randomise with
different design.{con,fts} is there any way way around this problem?
The problem is that the three contrasts that I mention are of interest.
Thanks,
--
Dr Colm G. Connolly
School of Psychology and Institute of Neuroscience
The Lloyd Building
University of Dublin
Trinity College, Dublin 2, Éire
Fax: +353-1-671-3183
|