Hi,
If you reverse the order ot the data and don't change the design.mat
or design.con details then yes the sign of the tstat1 should change.
You might find it simpler to just analyse all the data together and
then once you have one big all_FA_skeletonised, you can use fslroi to
extract 4D sub-chunks of that and fslmerge to recombine into two-group
4D files.
Cheers.
On 14 Jul 2008, at 11:55, Nathan Pajor wrote:
> We have three groups on which we have run tbss_2: tin, hl and nv. We
> now
> want to run three, two sample t-tests comparing each set of two
> groups.
>
> To do this we are creating a new directory (hl-nv, for example) and
> then a
> separate FAi directory with the tbss_2 output from the hl and nv
> groups. We
> then run tbss-3 through tbss-5 for each group comparison. When the
> groups
> are designated by hl and hv (hl is first in alphabetical order) we
> see a
> result for stat1. However, when we rename the hl group to pt_hl (nv
> is first
> in alphabetical order) we see the same result as before for stat1 once
> again, despite reversing the order.
>
> Are we doing this correctly? How do the randomise conventions work? I
> understood that the stats are taken in alphabetical order.
>
> Thanks,
>
> Nathan Pajor
> NIDCD/NIH
>
---------------------------------------------------------------------------
Stephen M. Smith, Professor of Biomedical Engineering
Associate Director, Oxford University FMRIB Centre
FMRIB, JR Hospital, Headington, Oxford OX3 9DU, UK
+44 (0) 1865 222726 (fax 222717)
[log in to unmask] http://www.fmrib.ox.ac.uk/~steve
---------------------------------------------------------------------------
|