I don't think it makes sense to correct using cohen's d. I've never heard of it at least doug
On 03/27/2014 06:31 AM, L. Schweren wrote:
Dear FreeSurfer experts,
For my group-level analyses, I created Cohen’s D effect size maps. Now I looking the different methods to correct for multiple comparisons (FDR and MC-simulations), it seems to me they are based on p-values, not on effect sizes. I tried applying FDR-correction in tksurfer on the effect size maps. This indeed results in a higher threshold for clusters to appear, but I am not sure whether this is a valid tool to use in statistical maps other than .sig ?
As a p-value reflects the “chances” of finding this effect by chance and the MC-correction is about reducing the “chance of false positives”, and Cohen’s D not reflecting such chances, is correction for multiple comparisons even sound when reporting effect sizes? If so, what would be the best approach to determine an MC-corrected threshold for effect sizes?
Thanks again for your help, you have been very helpful so far!
Best wishes,
Lizanne
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer