Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Knowledge from small number of debates outperforms wisdom of large crowds (2017)

arxiv.org

Post reply on HN
Print view View original

51–60 of 123 posts

Re: Knowledge from small number of debates outperforms wisdom of large crowds (2017)

#51

Post by hkt »

Deliberative democracy is great. Put people in enriched decisionmaking environments for better outcomes, and then get greater democratic legitimacy from doing so than even some representatives? Win win.

This is the opposite of populism too: when we voted Brexit in the UK, it was an uncontrolled, information-scarce (and falsehood-rich) environment. A room where people were organised and where they asked for their own experts and talked to one another would never have delivered that result.

Re: Knowledge from small number of debates outperforms wisdom of large crowds (2017)

#52

Post by brianpgordon »

amelius wrote: post #36

Perhaps we can save democracy by replacing the voting mechanism by placing people in groups of 10, and letting them reach consensus before making a vote.

That would not be the same situation at all. In the case of general knowledge questions, obviously the small groups will do better because people who are uncertain and would have voted randomly will instead listen to whomever is confident. And everyone works together to pool knowledge.

In political voting the problem is not a lack of information- everyone has a strong opinion about everything already and putting them in small groups would just result in unproductive bickering.

Re: Knowledge from small number of debates outperforms wisdom of large crowds (2017)

#53

Post by AnthonyMouse »

amelius wrote: post #36

Perhaps we can save democracy by replacing the voting mechanism by placing people in groups of 10, and letting them reach consensus before making a vote.

That was how the US Senate originally operated. You would elect your state legislature (where your vote is much less diluted than at the federal level) and they would elect your Senator.

Early 20th century populism took that away. It also meant that the states now have no representation in the federal legislature, which led to an almost immediate federal takeover of basically everything.

Re: Knowledge from small number of debates outperforms wisdom of large crowds (2017)

#54

Post by sbhn »

hkt wrote: post #51

Deliberative democracy is great. Put people in enriched decisionmaking environments for better outcomes, and then get greater democratic legitimacy from doing so than even some representatives? Win win. This is the opposite of populism too: when we voted Brexit in the UK, it was an uncontrolled, information-scarce (and falsehood-rich) environment. A room where people were organised and where they asked for their own…

People in the uk voted brexit, because it was just after the uk us and france were dropping bombs in lybia and syria, arab spring they called it, the destruction of infrastructure created millions of refugees, the bbc amplyied the importance and strength of isis, and flashed images of thousands of the refugees in calais tripping over each other trying to illegally cross the channel into the uk. The good souls of england thought they were under attack, bless em.

Re: Knowledge from small number of debates outperforms wisdom of large crowds (2017)

#56

Post by TrevorAustin »

amelius wrote: post #36

Perhaps we can save democracy by replacing the voting mechanism by placing people in groups of 10, and letting them reach consensus before making a vote.

Policy juries are an under-used decision procedure: https://en.wikipedia.org/wiki/Policy_jury

You could also have one body of the legislature (the US Senate? The House of Lords?) selected by Sortition: https://en.wikipedia.org/wiki/Sortition

Re: Knowledge from small number of debates outperforms wisdom of large crowds (2017)

#57

Post by thenaturalist »

ajuc wrote: post #28
thenaturalist wrote: post #25

Earlier quoted context omitted.

I can’t follow your reasoning. How does asking you the exact same question twice improve your accuracy? If you don’t know how high the Eiffel Tower is, why would you know it the second time? Conversely, you likely created a mental anchor when giving an estimate the first time and would be hard pressed to provide an answer contradicting your first estimate - even if it might be more accurate. There were no answers pro…

Given more time I can remember something, or I can notice that the last question gives some hints or even unrelated associations that help to answer the first question. I don't think "if you don't know the answer in 20 seconds you won't ever know it" is true.

As far as I understand it, these were not simple right/ wrong questions, but questions about estimating a continuous measure (height, age, percentages). I have no reason to believe that questions were related or would give hints. Such a correlation would destroy the power of the experiment.

As far as I understand it, this paper is not about "Do you know what is true or false in 20 seconds" but "what is a value you confidently estimate within 20 seconds". This is a field much studied in psychology and when you look into Kahnemann and associated research I would be surprised to find any scientific evidence that time improves your estimate. I'm not saying it's impossible, I am confident that - on average - it simply does not happen.

Kahnemann showed we're full of biases and this research shows that calibrating ourselves with others is a much higher predictor of improving accuracy of estimations than time.

Re: Knowledge from small number of debates outperforms wisdom of large crowds (2017)

#58

Post by sova »

mchannon wrote: post #41
deytempo wrote: post #37

Earlier quoted context omitted.

You mean like trial by jury. I like it.

Trial by jury gets one set of 10 (12, whatever), whereas the consensus model described would have n sets of 10. I seriously doubt the wisdom of a crowd goes up when the total crowd size shrinks to 10. I suspect it plummets. This does call into question the notion of a multiple-jury trial (maybe videoconferenced in, where the individual juries do not interact with each other). Based on this study's findings, it would…

To summarize the findings of this research article for you:

  Wisdom of the Crowd vs one person = good  

  Wisdom of the crowd vs tiny groups 4+ = bad

Re: Knowledge from small number of debates outperforms wisdom of large crowds (2017)

#59

Post by axiak »

AnthonyMouse wrote: post #53
amelius wrote: post #36

Perhaps we can save democracy by replacing the voting mechanism by placing people in groups of 10, and letting them reach consensus before making a vote.

That was how the US Senate originally operated. You would elect your state legislature (where your vote is much less diluted than at the federal level) and they would elect your Senator. Early 20th century populism took that away. It also meant that the states now have no representation in the federal legislature, which led to an almost immediate federal takeover of basically everything.

That amendment really screwed democracy in the US. Not only does it remove room for debate, but it has IMHO been a major factoring in lessoning voter interest in the state legislature. This has far reaching consequences, such as making it harder to make new amendments and checks against gerrymandering.

Re: Knowledge from small number of debates outperforms wisdom of large crowds (2017)

#60

Post by prewett »

baud147258 wrote: post #39

It reminds me of that assertion that the IQ of a crowd is the lowest IQ divided by the number of people in the crowd. Seems the study validate this with that smaller groups performs better

Except that the previous study refuted that idea: the average answer was better than individual answers, which is definitely not IQ_crowd = IQ_min / n.
Post reply on HN