> If I run a group-level analysis (e.g., 3dANOVA) on a subject-wise statistic
> (e.g., a pair of betas), which I assume are independent identically distributed
> normal random variables, should I ever expect to see fewer than 5% of the
> voxels exceeding the p < .05 threshold?

I don't see any reason for such an expectation. Theoretically the proportion of voxels that exceed a specific threshold can be anything within [0, 1] depending on how much activity the brain has.

> If 2% of my voxels exceed a given threshold, can I use that value as the
> p-threshold in AlphaSim to determine the cluster size threshold? Intuitively, if
> 2% of my voxels exceed a particular threshold, isn't 2% the upper limit my
> "true" p-value?

No, the p-threshold in AlphaSim is supposed to be one at individual voxel level. You should be able to check the corresponding q-value right underneath the threshold slider bar on AFNI viewer.

Gang