Hi James -
You can safely run each subject separately at the second level if
you do fixed effects at that level.
You will then only have 42 subjects for the next level.
T
On 21 Mar 2007, at 15:42, James N. Porter wrote:
> Hello-
>
> I'm following the web guide for Multi-Session & Multi-Subject
> (Repeated Measures - Three Level Analysis). I'm wondering if FLAME
> requires that I create the subject means within the context of the
> entire sample, or if I can split up my sample into more manageable
> bits and still obtain accurate results. That is to say, is the
> program calculating an overall variance statistic for the entire
> sample and then using it to evaluate each individual subject? I ask
> because I am currently running an analysis of 42 subject who each
> have 3 sessions, and the accompanying 126 x 42 matrix to define my
> EVs and the 42 x 42 matrix to define my contrasts are very unwieldy
> and slow when FEAT tries to generate them on-screen. Would
> splitting my analyses into something like 3 groups be a bad move?
>
> --
> Jim Porter
> TRiCAM Lab Coordinator
> Elliott Hall N437
> 612.624.3892
> www.psych.umn.edu/research/tricam
|