Thanks. Someone (not on through this list) also suggested that I might
be able to code it [1 -0.05 -0.95] and have it be close enough. Could
that also work? Thanks much!
Thomas Nichols wrote:
> Dear Vina,
>
> Typically an additive model (i.e. a contrast [1 0 -1], or just
> including a 0/1/2 covariate) will give you the most sensitive
> inferences. But if you want to test your 3-level (genotypic) model
> for a monotonic effect, the only way to do it is by a conjunction of
> two contrasts
> [1 -1 0]
> and
> [0 1 -1]
>
> You can do this with contrast masking in Feat.
>
> -Tom
>
>
> On Tue, Jul 29, 2008 at 2:59 AM, Vina Goghari <[log in to unmask]
> <mailto:[log in to unmask]>> wrote:
>
> Hi everyone,
> I searched the archives for this topic and I did not see it, so I
> apologize if it is a repeat. My question is what is the best way
> to do an analysis in three groups, where you want an analysis
> where Group 1 > Group 2 > Group 3, as my question has to do with
> genetic load, which I think will affect brain activity.
>
> Thanks,
> Vina
>
>
>
>
> --
> ____________________________________________
> Thomas Nichols, PhD
> Director, Modelling & Genetics
> GlaxoSmithKline Clinical Imaging Centre
>
> Senior Research Fellow
> Oxford University FMRIB Centre
|