Nici Wenderoth wrote:
> Hi SPMers,
>
> if I want to run a random effect analysis and compare two different
> experimental groups, does statistical power suffer if groups differ with
> respect to
> 1. subject size (e.g. one group contains 12 the control group just 10
> subjects) and/or
>
> 2. between-subject variance?
>
The two-sample t-test used in RFX analysis assumes that the between
subject variances are the *same* for each group.
The test is robust to violation of this assumption *if* the number of subjects
in each group is equal or nearly equal [Wackerley et al., p. 443].
Therefore, you should try to match the number of subjects in each group.
In practice, however, I would say that 10 subjects from one group and 12 from
another looks fine.
Best wishes,
Will.
Wackerley et al. (1996) Mathematical Statistics with Applications, Duxbury
Press (A great book !).
>
> So generally speaking, is it appropriate if my control group contains less
> subjects than my treatment group?
>
> Thank you in advance.
>
> With best regards,
> Nici
>
> Motor Control Laboratory
> Dep. of Kinesiology
> K.U.Leuven
> Tervuurse Vest 101
> 3001 Heverlee
> Belgium
>
> Tel: 32 16 32 90 72
> Fax: 32 16 32 91 97
--
William D. Penny
Wellcome Department of Cognitive Neurology
University College London
12 Queen Square
London WC1N 3BG
Tel: 020 7833 7478
FAX: 020 7813 1420
Email: [log in to unmask]
URL: http://www.fil.ion.ucl.ac.uk/~wpenny/
|