Dear Expert
I 'm doing a fMRI FEAT analysis with 2 groups.
I think if we suspect that the variances are different between groups, we can set up different group variables in FEAT.
* Statistically, I think this is equal to saying that we suspect that the "assumption of homogeneity of variance" is broken?
(I meant the assumption which we always use in ANOVA, like the Levene's test)
* Also, what if the variances are actually "NOT DIFFERENT" between groups, and I still set up "DIFFERENT" group variables in FEAT?
Will the result be the same when I set up "same" group variables in FEAT?
Thanks super much!
Jie-Yu
|