Dear all,

I have an experimental design in which there are 3 conditions (control and 2 experimental). I need to estimate sample size based on desired power and FDR but I've been reading only SSPA allows for more than 2 groups. However, the example provided in the original paper only contemplates a microarray experiment with 3 conditions and is pilot-data based, data which I don't have.

Has anyone used SSPA for 3-group sample size?

Does it make sense to perform 2-group sample size calculations for each combination, and keep the largest sample size, as suggested in a previous post ?

Thank you in advance!

You are (I assume) planning to make two different comparisons, right? And you want both to be suitable powered. Assuming that you have evidence that one group will be more variable than the rest, or if you think the average delta between groups will be different, you might need to do the sample size calculations separately. But if you don't know that to be the case and are simply assuming a particular variability and delta (based on some pilot data or whatever), then it doesn't matter that you have two comparisons because you are assuming the same within-group variability regardless.