On Sat, 17 Nov 2007 10:28:44 +0000, Cyril Pernet <[log in to unmask]>
wrote:
>
>
> Hi Bedda
>
>> I found some
>> information that mention that if I want to test if there is a
>> difference between n gropus I should use a contrast (matrix, contrast
>> weights) with n-1 lines. So, to test the difference between three
>> groups the contrast will be
>>
>> c = 1 -1 0
>> 0 1 -1
>>
>> Can someone let me know if this is correct? Or should it be
>>
>> c = 1 0 0
>> 0 1 0
>> 0 0 1
>
> the last contrast tests for any activations for 1 or several
>conditions (note that you should do that for a one way anova, I don't
>see here how you used the full factorial?) - this is sometimes refered
>as the 'effect of interest'
>
> to test differences, you missed one extraline 1-10 = gp 1 vs 2,
>01-1 is 2 vs 3 so just add a line 10-1 for 1 vs 3
I don't think it's necessary to add that last row ([1, 0, -1]), because
it's a linear combination of the other two.
>
>gd luck
>
> cyril
>
>--
>The University of Edinburgh is a charitable body, registered in
>Scotland, with registration number SC005336.
>
>
|