Dear John et al
Hi - we're doing a multi-gp multi-subject multi-condition SPECT analysis and
we have a couple of questions that we really hope you can answer...:
Are we correct in thinking that with this model, contrasts of
(condition x vs condition y) within a group
does a paired t-test?
Is there a way of also doing between group (unpaired t-test) comparisons
((gp 1 condition1) vs (gp 2 condition 2))
using this model? (We've tried setting up the contrast in the regular
way (eg -1 0 1 where there are 2 conditions
for the 1st group) but the contrast manager gives "invalid contrast" )
Also, we have some scans belonging to an extra "group" of mixed
subjects, that is, each of these subjects should actually
belong to one of the other defined groups (controls, responders,
non-responders) except that their response
to treatment is not determined. Would it be sensible to add these as a
separate group to the model or should we
leave them out?
With many thanks for any answers or suggestions
Hava :)
--
************************************************************************
Hava Lester, PhD [log in to unmask] +972 (0)2 677 7148
home: [log in to unmask] +972 (0)2 567 0778
Dept Nuclear Medicine, Hadassah Hospital, Ein Kerem, Jerusalem, Israel
.
--
************************************************************************
Hava Lester, PhD [log in to unmask] +972 (0)2 677 7148
home: [log in to unmask] +972 (0)2 567 0778
Dept Nuclear Medicine, Hadassah Hospital, Ein Kerem, Jerusalem, Israel
|