People could be using actual random number sources to do this, although I think that’s probably not the explanation.
I want to point out that getting these polls right is a much different task than generating a good pseudorandom sequence. For example, people could come with a few pseudorandom numbers stamped onto their soul (EG, a birthdate), which they use to answer polls like this. This would not help with generating long pseudorandom sequences.
(I don’t actually think everyone uses their birthdate as a pseudorandom number, but maybe they use similar things such as address etc. But mainly, there could be other explanations which similarly don’t imply people would be good at generating larger sequences.)
Hm, could we tell apart yours and Zack’s theories by asking a fixed group of people for a sequence of random numbers over a long period of time, with enough delay between each query for them to forget?
This occurred to me, but I didn’t see how it could work with different ratios. I guess if you have a sample from a variable with a big support (> 100 events) that’s uniformly distributed, that would work (e.g. if x is your birth date in days, then x/365 < 20 would work).
It would be interesting to test this with a very large sample where you know a lot of information about the respondents and then trying to predict their choice.
People could be using actual random number sources to do this, although I think that’s probably not the explanation.
I want to point out that getting these polls right is a much different task than generating a good pseudorandom sequence. For example, people could come with a few pseudorandom numbers stamped onto their soul (EG, a birthdate), which they use to answer polls like this. This would not help with generating long pseudorandom sequences.
(I don’t actually think everyone uses their birthdate as a pseudorandom number, but maybe they use similar things such as address etc. But mainly, there could be other explanations which similarly don’t imply people would be good at generating larger sequences.)
Hm, could we tell apart yours and Zack’s theories by asking a fixed group of people for a sequence of random numbers over a long period of time, with enough delay between each query for them to forget?
This occurred to me, but I didn’t see how it could work with different ratios. I guess if you have a sample from a variable with a big support (> 100 events) that’s uniformly distributed, that would work (e.g. if x is your birth date in days, then x/365 < 20 would work).
It would be interesting to test this with a very large sample where you know a lot of information about the respondents and then trying to predict their choice.