[VoxBo] Orthogonal contrasts including conditions labeled as zero

[Ranjani Prabhakaran]

Thanks, Charan and Dan.  Your responses were really helpful!!

Ranjani

Quoting Daniel Y Kimberg <kimberg at mail.med.upenn.edu>:

> Ranjani Prabhakaran wrote:
> > What would be the appropriate weights for this contrast if I want to
> > get a beta value?
>
> Charan gave a detailed answer, here's the one-liner: if you weight
> each covariate 0.333, you get the average of the three effects
> (betas), while if you use 1, you get the sum.  It may or may not make
> a difference which you use, dpending on what you plan to do with the
> values.
>
> dan
>