Dear Richard,
Richard Perry wrote:
> Dear Ian,
>
> Apologies: Rik Henson has pointed out an error in what I wrote yesterday.
>
> >Dear Ian...
> >...It's not so much that you are assuming that the observations from
> >all subjects are equally reliable (this is clearly not the case!)...
>
> Apparently you DO in fact assume equal 'efficiency' of estimates of
> the betas in each individual in a RFX analysis. I'm not quite clear
> why (and it worries me a bit, because I know that the variance often
> varies considerably from one individual to another!).
I think what Rik is simply saying is that in the (one sample) second-level
analysis (as currently implemented), the
betas are assumed to be i.i.d.
Sincerely,
Eric
> I thought I
> should at least warn you of this immediately; perhaps someone will be
> able to write an e mail explaining why this is so later!
>
> Best wishes,
>
> Richard.
> --
> from: Dr Richard Perry,
> Clinical Lecturer, Wellcome Department of Cognitive Neurology,
> Institute of Neurology, Darwin Building, University College London,
> Gower Street, London WC1E 6BT.
> Tel: 0207 679 2187; e mail: [log in to unmask]
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|