Hi Arno,
this has been answered before on the fsl list:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind1001&L=FSL&P=R21782...
For more scenarios, have a look at the archives with the keywords randomise + missing + value
Cheers,
Gwenaelle
--- En date de : Sam 28.8.10, Arnaud Messé <[log in to unmask]> a écrit :
> De: Arnaud Messé <[log in to unmask]>
> Objet: [FSL] Missing values in Randomise
> À: [log in to unmask]
> Date: Samedi 28 août 2010, 17h45
> Dear all,
> I have a similar question, regarding missing data within
> randomise.
> Briefly, I have initially 2 groups (for example 3 controls
> (C) and 4 patients (P)), with the following design:
>
> C P
> 1 0
> 1 0
> 1 0
> 0 1
> 0 1
> 0 1
> 0 1
>
> Next, I split my patients group into subgroup by prognosis,
> for example: good or bad outcome.
> So, I have finally 3 groups: controls, good patients (G),
> and bad patients (B).
> However, some patients have no prognosis (due to
> missing clinical variable), and my design looks like that:
>
> C G B
> 1 0 0
> 1 0 0
> 1 0 0
> 0 0 1
> 0 1 0
> 0 0 0
> 0 0 1
>
> The patient of the penultimate line does not belong to a
> group... this causes problems? in statistical efficiency
> (df)?
> Do i need to add one EV for each subject ouside the groups?
> Or one EV for all the subjects ouside?
> Many thanks in advance,
> Arno
--------------------------------------------------------------------
Gwenaëlle Douaud, PhD
FMRIB Centre, University of Oxford
John Radcliffe Hospital, Headington OX3 9DU Oxford UK
Tel: +44 (0) 1865 222 523 Fax: +44 (0) 1865 222 717
www.fmrib.ox.ac.uk/~douaud
--------------------------------------------------------------------
|