Prescription for driving otherwise sane scientists into "atheoretical" self-delusion: Take the most appealing idea imaginable, devise a simple-seeming experiment to test it, and have the experiment be more complicated than it seems, with an easy mistake that everyone is sure to make on their first attempt, guaranteeing the same erroneous conclusion. The experiment then gets replicated over and over until the appeal of the idea overwhelms the scientists' innate skepticism.