Dear Maria,
Firstly I'd decide where you wish to publish the work. If similar papers
in this journal use a fixed effects analysis then you could go with that.
Otherwise RFX.
Assuming you go for RFX you should note that Random Field Theory (RFT)
is inaccurate with low degrees of freedom (ie. small number of
subjects). I therefore suggest you do a nonparametric random effects
analysis. See eg.
http://www.fil.ion.ucl.ac.uk/spm/doc/biblio/Keyword/NONPARAMETRIC.html
Best wishes,
Will.
Maria Densmore wrote:
> Dear SPMers,
>
> I have a study where there are two groups of Patients , Z(n=8) and P(n=7), Each
> subject had 2 scans (initial then 2 months later).
> Zpre Ppre Zpost Ppost
>
> With such small group sizes at this point in the study, is it still acceptable
> to present fixed effects model results or must I run Random Effects as well?
>
> I would like to test for the following contrasts:
>
> Zpre vs Zpost
> Ppre vs Ppost
> Zpre vs Ppre
> Zpost vs Ppost
>
>
> Please advise me as to what would be the most acceptable way to analyze this
> data set using spm99.
>
> thank you for your time,
> Maria
>
>
--
William D. Penny
Wellcome Department of Imaging Neuroscience
University College London
12 Queen Square
London WC1N 3BG
Tel: 020 7833 7475
FAX: 020 7813 1420
Email: [log in to unmask]
URL: http://www.fil.ion.ucl.ac.uk/~wpenny/
|