I'm pretty sure that Clark was comparing the trial size of 9902A to the trial size of 9901 in his analysis, not the small trial size of 9902A per se.