The following web page created by Jeanette Mumford may be helpful to you
to understand the impact of variable centering.
http://mumford.fmripower.org/mean_centering/
cheers,
-MH
On Thu, 2011-04-21 at 15:30 +0200, Mondadori Christian wrote:
>
> dear community
>
> i'm trying to run a simple regression in spm8 with RT as covariate and
> got a bit puzzled about the 'centering' option, since turning it on or
> off gives completely different spm's in the results section. i
> searched the archives and looked in the manual but didn't find a lot
> of help.
> first i would like to know what is the statistical/theoretical
> implication of this option. second i'd like to know wheter i should
> turn this option on or off, when correlating the con.images with RT.
>
> thanks for your answers
>
> chris
> ----------------------------------------
>
>
>
|