Dear Xingbao,
We are analyzing a fMRI study
using SPM in which we have two groups. If the two groups have different
smaple sizes (eg 14 vs 30), how do I can report second level result with
a same p value and a same extend cluster size?. Of course, we can do two
samples t-test with SPM for comparison. However, we need to report each
group result respectively. If we set p value at 0.001 for both groups,
you get a high t value reqired for group with less sample size ( t =3.4
vs t=3.8). How do I can interprate the results of two groups with
unbalance sample sizes?
To report the activation within each group separately use a single sample
t-test to show that the
contrast is greater (or less than) zero in the usual way.
Then report the group differences in a third SPM analysis using an
unpaired t-test. Again report
the results in the usual way.
I hope this helps - Karl
