Surprisingly popular

#1

Post by tosh » Sun, Jul 28, 2019, 12:44 PM UTC

Re: Surprisingly popular

#2

Post by Noumenon72 » Sun, Jul 28, 2019, 2:34 PM UTC

The capital is not Philadelphia, it's Harrisburg. So the idea must be that most people think they are right and others will agree with them, but there are people who get it right and know that others get it wrong, and they are responsible for the difference in scores.

Re: Surprisingly popular

#3

Post by carbocation » Sun, Jul 28, 2019, 2:35 PM UTC

It's an interesting heuristic, but without an exposition it's hard to know if there is any reason to believe it. Has this technique been assessed across various domains? Does it work better for political facts than for scientific facts? Where does it fail? It's a provocative article and a cool thought, but it seems more like a hypothesis.

(In particular, I'm referring to the assertion at the end of the article: "Because of the relatively high margin of 10%, there can be high confidence that the correct answer is No.")

Re: Surprisingly popular

#4

Post by mathgeek » Sun, Jul 28, 2019, 2:38 PM UTC

I’m not sure what to think of a Wikipedia article without any sections on Uses, and the only Example being unreferenced.

Re: Surprisingly popular

#5

Post by bengarvey » Sun, Jul 28, 2019, 2:40 PM UTC

Philadelphia should be the capital of Pennsylvania

Re: Surprisingly popular

#6

Post by nacc » Sun, Jul 28, 2019, 2:42 PM UTC

Digging down a bit this method is published in Nature! https://www.nature.com/articles/nature21054

Re: Surprisingly popular

#7

Post by ncmncm » Sun, Jul 28, 2019, 2:42 PM UTC

bengarvey wrote: post #5

Philadelphia should be the capital of Pennsylvania

Pittsburg/Philadelphia should be the capital.

Re: Surprisingly popular

#8

Post by busfahrer » Sun, Jul 28, 2019, 2:45 PM UTC

I wonder if this could be viewed as an application of https://en.wikipedia.org/wiki/Dunning%E2%80%93Kruger_effect

Re: Surprisingly popular

#9

Post by tofof » Sun, Jul 28, 2019, 2:46 PM UTC

carbocation wrote: post #3

It's an interesting heuristic, but without an exposition it's hard to know if there is any reason to believe it. Has this technique been assessed across various domains? Does it work better for political facts than for scientific facts? Where does it fail? It's a provocative article and a cool thought, but it seems more like a hypothesis. (In particular, I'm referring to the assertion at the end of the article: " Bec…

One of the wiki's sources might have been a better linked article.

The MIT summary [1] notes "The researchers first derived their result mathematically, then assessed how it works in practice, through surveys spanning a range of subjects, including U.S. state capitols, general knowledge, medical diagnoses by dermatologists, and art auction estimates." Across all those areas, this technique had error rates about 20% lower than other competing techniques. Those techniques included simple majority vote to two different kinds of confidence-weighted scoring.

The paper earned a prestigious publication in Nature.

1: http://news.mit.edu/2017/algorithm-better-wisdom-crowds-0125

Re: Surprisingly popular

#10

Post by tofof » Sun, Jul 28, 2019, 2:51 PM UTC

Noumenon72 wrote: post #2

The capital is not Philadelphia, it's Harrisburg. So the idea must be that most people think they are right and others will agree with them, but there are people who get it right and know that others get it wrong, and they are responsible for the difference in scores.

That is indeed exactly the underpinning of the technique, not that you can tell from the wiki article.