> I would like to run an ANOVA with 2 between subjects factors and one within
> subjects factor. It does not seem like this is possible using 3dANOVA3 or the
> matlab GroupAna package if you want to include subject as a random factor.
> Is this true or have I missed something?
Yes that is true.
> I know this 3-way ANOVA could be done in 3dRegAna, but there again I don't
> think that you can code for random effects, right?
Yes you can, but it is just excruciately painful to code the design and obtain results when dealing with 3dRegAna.
A better choice is to try the new baby program 3dLME:
http://afni.nimh.nih.gov/sscc/gangc/lme.html
Gang