Dear Donald,
many thanks for your help.
Best regards, Elisabeth
> A couple points:
> (1) The MEG is not a valid contrast in a within-subject/repeated-measures
> design. The reasoning is explained in my early SPM posts on the topic.
> (2) The contrast for MEC needs to be an F-test with (N-1) rows where N is
> the number of conditions.
> (3) The interaction contrast needs to be an F-test with (N-1)*(M-1) rows
> where N is the number of conditions and M is the number of groups.
>
> For more complex design, please read my previous posts on the topic.
>
> For your 2x2 design, the contrasts are correct.
>
> Best Regards,
> Donald McLaren, PhD
>
>
> On Tue, Jan 12, 2016 at 6:30 AM, Elisabeth Kasper <
> [log in to unmask]> wrote:
>
>> Dear All,
>> I have a longitudinal study design (ANOVA with repeated measures and two
>> factors: factor group and factor condition with two levels each).
>> According
>> to the SPM8 manual, I generated a design matrix without problems.
>> Then I formed contrasts following the manual by Gläscher and Gitelman (
>> http://www.sbirc.ed.ac.uk/cyril/download/Contrast_Weighting_Glascher_Gitelman_2008.pdf)
>> in the following way:
>>
>> 1. I set the following variables in matlab:
>> ?nc? for number of conditions, ?ng? for number of groups,
>> ?MEc? for main effect of condition >>MEc = [1:nc] ?mean(1:nc),
>> ?Meg? for main effect of group>>MEg = [1 -1]
>>
>> 2. The contrast template for main effect of group in the manual is
>> defined
>> as follows:
>> ?MEg zeros(1,nc) ones (1, nc) /nc ?ones (1,nc) /nc?
>> >> So for my design it would be: ? 1 -1 0 0 0.5 0.5 -0.5 -0.5?
>>
>> 3. The contrast template for main effect of condition in the manual is
>> defined as follows:
>> ?zeros(1,ng) MEc MEc*[n1/(n)] MEc*[n2/(n)]?
>> >> So for my design it would be: ?0 0 0.5 -0.5 0.25 -0.25 0.25
>> -0.25?
>> Now I`m wondering about the 0.5 for condition. I would assume it
>> should
>> be 1 -1 ?
>>
>> 4. The contrast template for group/condition interaction in the manual
>> is
>> defined as follows: ?0 0 0 0 MEc ?Mec?
>> >> So for my design it would be:?0 0 0 0 0.5 -0.5 -0.5 0.5?
>>
>> Could you tell me whether my contrasts are correctly specified?
>> Many thanks, Elisabeth
>>
>
Dipl.Psych. Elisabeth Kasper
Neuropsychologin
Universitätsmedizin Rostock
Zentrum für Nervenheilkunde
Klinik für Psychosomatik und Psychotherapeutische Medizin
Gehlsheimer Str. 20
18147 Rostock
Tel.: +49 381 494 – 9618
Fax: +49 381 494 – 9472
E-mail: [log in to unmask]
Deutsches Zentrum für Neurodegenerative Erkrankungen e.V. (DZNE)
Standort Rostock/ Greifswald
Sitz am Zentrum für Nervenheilkunde
Gehlsheimer Straße 20
D- 18147 Rostock
|