I would assemble a team to try and replicate several scientific experiments that have came out with inexplicable results. While a large number of those will turn out fraudulent, it is very possible that a few claims were true, but sounded so outlandish that they were considered fake right away.