Well, I don't they did this, but it's an interesting technique used for this sort of scenario. Unfortunately, I can't find a wiki article to explain it (I thin

Well, I don't they did this, but it's an interesting technique used for this sort of scenario. Unfortunately, I can't find a wiki article to explain it (I think it's something like the hugh-jones coin or something). Anyway, I think the idea was someone needed statistics on how many people were stealing things after some sort of event, but obviously they couldn't just ask people because who would admit to breaking the law. So what you do is you ask the question: "Are you stealing" then you have the person flip a coin out of your view. If the coin is heads then they answer honestly, but if the coin is tails, then they always answer yes. This way the person in question has plausible deniability (I didn't steal, I said that I did because the coin came up tails). Then once you have the data you just need to divide by half and maybe tweak some other parameters, but you end up with realistic data (or so the statisticians say).