Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Cognitive Dissonance in Monkeys - The Monty Hall Problem (nytimes.com)
27 points by nickb on April 8, 2008 | hide | past | favorite | 9 comments


The actual paper is linked at http://www.som.yale.edu/faculty/keith.chen/papers.htm

The article doesn't give what seems to be a pretty important detail. Children and monkeys chose the new item 63 and 60% of the time, respectively. That suggests buyer's remorse rather than choice rationalisation, but only small amounts. Possibly not even significant, depending on the sample size.


There are 2 kinds of people in the world: Those behind Door #1 who don't understand the Monty Hall problem, and those behind Doors #2 and 3 who do.

It's amazing how many smart people don't get the problem at all.


Paul Erdos got it wrong (!).


As someone explained earlier on News.yc (can't find the link), Paul Erdos (and possibly other mathematicians) got it "wrong" because they weren't told the correct problem. If you don't mention that: (1) the game host knows where the car is beforehand (2) he tells beforehand that he'll reveal a goat,

then there's no paradox.


Do you mean this thread?

http://news.ycombinator.com/item?id=78322

I haven't read the book mentioned so I can't comment, perhaps someone else can.


Summary: one shouldn't assume uniform prior distributions in psychological experiments. Bayes-rule is your friend.


To quote the article: "He says the psychologists wrongly assumed that the monkey began by valuing all three colors equally."

So, basically psychologists are dumb; why assume that monkeys have no color preference? Even if you don't think they've got any intelligence, you could probably assume that they have some knowledge of colors associated with looking for food.


But also from the article: "After identifying three colors preferred about equally by a monkey -"

They didn't assume there were no preferences. Their mistake was to assume they could measure them precisely enough.


This monkey thinks that perceived novelty/rarity could also explain the green-preference. When at the second forced-choice, blue has appeared in 2-of-4 positions, while both red and green have only appeared in 1-of-4 positions each. Further, blue has always been at least one choice. That makes it look common, so less interesting.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: