Hi,
yes - or yo do an F-test.
Cheers-
Andreas
________________________________________
Von: FSL - FMRIB's Software Library [[log in to unmask]] im Auftrag von David Harper [[log in to unmask]]
Gesendet: Donnerstag, 8. April 2010 18:54
An: [log in to unmask]
Betreff: [FSL] one or two tailed?
Greetings,
I'm using TBSS to look at differences between two groups (one normal
and one with pathology). I noticed in the instructions that output
can consist of two maps (tbss_tfce_corrp_tstat1.nii.gz and
tbss_tfce_corrp_tstat2.nii.gz). The first map tests controls >
pathology and the second tests pathology > controls at the p<0.05
level. Does this mean that the two tests are one tailed? If I wanted
to be agnostic about the direction of the finding would I need to
threshold at p<0.025?
Thanks,
David
|