Hi
Wrt to a simple mean regressor of interest, de-meaning and orthogonalisation are the same thing. If the variable is already demeaned and you add the 'orthogonalise' option then FEAT will remove 0 from the already demeaned regressor again, i.e. not do anything.
hth
Christian
On 26 Jul 2011, at 10:30, Max Beermann wrote:
> Hi Steve,
>
> I read Jeanette Mumford's page and understand the demeaning issue.
>
> There is only one thing I did not find, maybe you could give me a short answer to this?
>
> What is the relationship between demeaning and orthogonalisation? (If I demean my variable, it is orthogonalised wrt the group mean EV; but what happens if I just tick the "orthogonalise" box in FEAT - is my variable then demeaned?)
>
>
> Many thanks and best regards,
> Max
|