Hi Mark,
Thanks! How exactly should I set that up? Just the positive effect plus F-test for each EV, or both positive and negative plus one F-test for each (and mask the t-test results with the results of the F-test)?
Best,
Renske
-----Original Message-----
From: FSL - FMRIB's Software Library [mailto:[log in to unmask]] On Behalf Of Mark Jenkinson
Sent: vrijdag 3 juli 2015 19:58
To: [log in to unmask]
Subject: Re: [FSL] FEAT 2-sided tests
Hi,
Using an F-test is the easiest way to get a two-sided test.
All the best,
Mark
> On 3 Jul 2015, at 17:38, Renske Huffmeijer <[log in to unmask]> wrote:
>
> Hi,
>
> This is probably a very silly question, but I can't seem to find the answer: We've run lower level analyses for each subject that resulted in several copes, and now want to run a higher level analysis to test the group effect (EV1) as well as effects of a continuous variable (EV2). We want to test everything 2-sided and I was wondering how to set up the contrasts. If we'd just include both the positive and negative effect of each EV, we'd be testing 1-sided twice, which isn't quite correct. Can we use F-tests to remedie this or should we half a p-value somewhere?
>
> Thanks!
> Renske
|