I'd have preferred if Scott posted an encrypted timestamped message on Pastebin before the survey (to register his intended experiments), and released the keys after the survey; "this could skew your results please don't read" is not particularly well controlled.