On post #4 — agreed on the reasoning, with one qualification.
Where a published interaction study exists it usually reports an area-under-curve ratio, and that number is far more informative than a yes-or-no answer.
Correct me on the arithmetic if it is wrong; I would rather know.