Hacker News Reader

Comment by greggarious | Hacker News Reader

>tabooness

Oh, ok. I thought you said something else for a second.

This is interesting... but please be aware publishing data on these things changes it, since the fetish community is full of weird little edgelords who poop themselves in fear (rather than as part of some mutually agreed upon sex game) when they realize you weren't kidding when you told that cutie in Bloomington that Armin Meiwes did nothing wrong and that next time you don't want to just dry fire their wonder nine.

And if you really want to have fun, do as Kinsey did and have them flip a coin and only be truthful if it's heads, then double your sample size to compensate. That's how you gather truly useful data that will keep you entertained long after your IRB certification expires ;-)

View on HN

Do you have any references to that Kinsey thing? I would just love to read a methodology section including that

The technique is called randomised response. Differential privacy formalises it as well

Yeah, I don’t understand why doubling the same size would compensate for instructing half the group to lie.

What I think GP is describing is this. Say you want to survey people about something that they may feel social pressure to deny (e.g., cannibalism fantasy). You have people come in, you give them a fair coin, tell them to go into an unmonitored closet and flip the coin (only the respondent sees the result), then come back out. If they flipped heads, the question they should answer is, "Have you ever drank water?" If they flipped tails, the question they should answer is, "Have you ever fantasized about cannibalism?"

If you get 55% "yes" responses, you know that about 10% of people answered yes to the cannibalism question.

You know that 50% were answering "yes" no matter what, so if you double the percentage above 50%, you get the people answering "yes" to the cannibalism question.

The Kinsey trick is brilliant.

Then why was I downvoted? Kink shaming is not cool :/

It wasn't me! I only downvote with a counterargument comment buttressed by a link to evidence, and I wish that was a thing.

It’s been a persistent issue for me, on other sites too - trying to be better about paying attention to the comms rather than those kinda things but I’ve had a few times here lately I circle back to a convo to downvotes after I said something elsewhere.

People love to abuse their access lol :/

Reply on news.ycombinator.com