> We have one value for each subject at each level.

Then you would have to go with the following: there is no way to differentiate MSAB and MSE, and you would simply consider MSAB as the within-subject MSE as I originally suggested with option -fa.

>> If yes, you use option -fb or -fab. With -fab, you will get two
>> sub-bricks with the first being MSB and the second
>> MSE. Then you obtain MSE using 3dcalc.
>
> If the above does work:
>
> The second MSE is the error with respect to between subjects, correct?
> At that point, Would I have to square the second MSE since the value
> is the square root of the MSE? Or does this require a different step?

Sorry I misstated in my previous message. When you have multiple value per subject under each condition (it seems this is not your case), use use option -fb or -fab. For example, with -fab you will get two sub-bricks with the first being MSB and the second being F-stat = MSAB/MSE. Then you obtain MSE using 3dcalc.

Gang