Sometimes you have to conduct an experiment again, again, — and again — before you come to a conclusion.