Print

Print
Hi Sam,

Please, see below:


On 15 December 2015 at 12:38, Sam Rogers <[log in to unmask]> wrote:
Hi Anderson.

I am sorry. I have an F-test that is more correct here. For the attached factor model, the F-test is non-significant in region X. This F-test should be the within-subject effect of time, so this result suggests that there is no change in grey matter between sessions.

No no. This F-tests suggest that there is no difference between groups on what concerns differences between sessions. This test is about differences on differences, as opposed to about simply changes between sessions. There may be vast changes between sessions, and still the test may give no significant results, simply because the changes are the same across all three groups.
 
The results from the correlate model (which takes the subtraction image) shows that changes in region X are associated with score. Should there be agreement here?

Nope... The hypotheses are different. An eventual agreement would be a coincidence or perhaps something interesting to be investigated, but it isn't something expected in principle.

 
I am confused why the factor shows no change but the correlate shows a change associated with the score.

The first is about differences between groups on changes between sessions, not just on changes (in fact, even if there were a sole group, it still would be a different hypothesis, so no similar results, unless due to coincidence).

Hope this helps.

All the best,

Anderson

 

Thanks.
Sam.