Yup. Your final paragraph is kind of where I was heading. To perform an experiment you need a prior agreement to what the hypothesis, parameters and analysis are. If the experiment has been done previously anyway then it literally is no-win, and a bit like the offers to participate in a "debate", there is only one likely outcome. That is unless there is an experiment that can on their terms, within their parameters disprove their hypothesis. Which there isn't. Education is the answer, but retrospective eduction is a tough one and requires time and care, such as that which this place provides.