Hello,
For designs where row permutation is inappropriate randomise will use row inversion instead.
Many Regards
Matthew
> Hi FSLers,
>
> Can you please say what randomise does (i.e. how permutations are performed) when we have a single group and want to do multiple regression (looking at the effect of one EV when controlling for nuisance variables)? I get the point when we want to compare 2 groups, but not for just 1. I remember a nice post from Tom Nichols on this topic, but after extensive search I've been unable to get it back.
>
> Thanks a lot,
> Miguel
>
>
>
|