> How can Alphasim be used to find cluster sizes for a group ANOVA?
I can't think of anything with AlphaSim that would be different from what is done with individual subject analysis.
> I've been reading some work and it seems that cluster sizes for
> ANOVA analyses are often higher than cluster sizes for single
> subject T-maps, which doesn't make entire sense to me. I am
> guessing it has to do with spatial correlation of the data?
I have never done such a comparison, but I agree with you that the difference in minimum cluster size is most likely due to different spatial correlation since spatial correlation of the residual error might be slightly different between the two analyses.
Gang