Alright, now set up a study with controls and all that. If you're onto something in general (and not just anecdotal), it'll show in the results :)