Dice Help Scientists Get Honest Answers to Touchy Questions

Dice Help Scientists Get Honest Answers to Touchy Questions

Read the article, and it is a curio.
Farmers in Africa were killing endangered species and were doing this to eliminate predators on their own livestock, but were lying about it to authorities on this activity.

When rolling one die, and being asked by police to honestly answer if a ONE or a SIX was rolled and answering honestly, while being interrogated about their killing of any endangered species in the protection of their own live stock, farmers more frequently admitted their transgressions.

Wow, as a Las Vegas Dice Dealer - this sure as hell doesn't seem to work with crap players when they are rolling two dice with their own money on the line, when they want to cheat or take shots. Hmm.

Obviously, these farmers are not gamblers, but are innocent workers, where this truth-telling activity promotes further truth telling during interrogation.

Maybe the LVPD Metro can try this technique on their busts here in the valley.

They were asked to answer "NO" if they rolled a 1, and yes if they rolled a 6, and the truth on 2-5. The roll was kept secret, so the farmer didn't have to implicate themselves. Obviously if no law breaking occurred, then 1 in 6 farmers would say yes. They got 19% saying yes. I'm not sure whether in 100 interviews if that would be significant or not. Obviously you'd have to look at the odds of this result coming out by chance.

Quote: Paigowdan

When rolling one die, and being asked by police to honestly answer if a ONE or a SIX was rolled and answering honestly, while being interrogated about their killing of any endangered species in the protection of their own live stock, farmers more frequently admitted their transgressions.

Police aren't doing the questioning. It's being done by researchers.

The randomized response technique is considered somewhat valid simply because it allows the possibility of the answer being random. Because the person being questioned has some assurance that the researcher doing the questioning will not know if the answer is a truthful one or a random one, the person is more likely to be truthful on a roll of 2,3,4,5.

I've only heard about this technique being used in questioning of social type things. Like finding out if people do drugs, or have unprotected sex, and various things like that. It's a good way to get more people to be honest in their answers.

Such a methodology could not work in, for example, trying to identify the culprit of a crime. The methodology requires that you are asking a large number of people a question. I suppose you could bring in 100 suspects for some crime and ask them all if they did it. But even so, you could expect to get at least 16.67 yes answers.

Then again, if you got 100 yes answers, then you'd know something was definitely amiss.

Quote: konceptum

Such a methodology could not work in, for example, trying to identify the culprit of a crime. The methodology requires that you are asking a large number of people a question. I suppose you could bring in 100 suspects for some crime and ask them all if they did it. But even so, you could expect to get at least 16.67 yes answers.

How so ? The majority would say "no" even when the rules are to say "yes", since noone is seeing the dice.

Quote: MangoJ

How so ? The majority would say "no" even when the rules are to say "yes", since noone is seeing the dice.

Your assuming that people in the experiment would not follow the rules, which is valid. Getting people to follow the rules of an experiment is never guaranteed. The same problem occurs with questionnaires. You cannot assume that people will give honest answers in a questionnaire, even if they are assured of anonymity.

But then again if you lie and pass the detector test, one would probably not live long enough to exit the room.