mmoayedi@uhnres.utoronto.ca wrote:
Hi,
I am studying two groups and have found that age is a confound (or is a covariate with an interaction between the two groups). Simply, when I correlate age between the two groups, there is an interaction (i.e., the slopes are different).
My questions are:
- is it appropriate to proceed with a map-wise t-test (with or without age as a nuisance covariate)?
If there is an age effect in the data, then you should include it in your model.
- is there a way to pull out age effects on each group specifically, i.e., use age as a covariate of interest?
I'm not sure what you mean. It sounds like that is what you've already done. You can create a contrasts to test the age effect in each group separately.
If neither of these methods is statistically sound, is there a method that can be used to describe age effects in each group, or characterize the findings?
Thanks,
Massih Moayedi
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer