Hacker News new | past | comments | ask | show | jobs | submit login
Wikipedia and the Wisdom of Polarized Crowds (nautil.us)
84 points by dnetesn 7 days ago | hide | past | web | favorite | 31 comments





One of the reasons I still frequent Hacker News is because there is a very similar dynamic at play here on topics where informed dissent can bring about greater awareness and lead to chinks in the hive mind. I think the common thread is the appeal to authority, and the willingness to admit you've been "out-sourced" in your stance, whether you're with or against the majority. Of course there are always those threads that won't die or those individuals that just keep rehashing the same point, oblivious to the lack of scientific reasoning or rigor to the arguments, but you learn to spot those pretty quickly.

What kind of bad science rehashed points are you talking about specifically, which appear on HN?

There are plenty of items that pop up on HN that get the almonds activated. Basically anything that has non-conclusive scientific proof and also some component of ideology gets drama churned up. Here is a short list of examples:

- IQ, and how much of it is biological

- Actual differences between biological sexes

- Anything related to government fiscal or monetary policy

- The extent (or even existence) of privilege for sexes or races

- If software should be free

- If software can be "stolen"

- Whether it is right or wrong for people to give up so much of their privacy to FANG

- ... and so on.


I tend to stay away from any and all science/physics posts for these same reasons.

I wonder how these would be interpreted with the author's point that:

"Liberal readers preferred basic science (physics, astronomy, zoology), while conservatives went for applied and commercial science (criminology, medicine, geophysics). 'It seems like conservatives are happy to draw on science associated with economic growth—that’s what they want from science,' Evans said. “Science is more like Star Trek for liberals: traveling through worlds, searching for new meanings, searching for yourself.'"

I'm curious how these examples would be categorized: when going down the list, how would a "liberal" and "conservative" group, in general, think about each example; as basic science or as applied/commercial science?


That's interesting, I consider myself more or less conservative, but really enjoy basic sciences/fields like physics, math, philosophy etc

Yeah, I'm curious in the way that the author defines these terms. Did they take into account fluidity of situations as you relate to? Because I'm sure there are many groups out there just like this. And what are the boundaries to these definitions?

A lot of soft "social science" (pseudoscience in my opinion) is unquestioningly defended by many because of appeal to authority, ideological, and political reasons.

I don't want to put words in you mouth, but social science is not necessarily pseudoscience. A lot of it is, or rather, a lot of what is popularized is, because it's a lot easier to make bullshit interesting.

> a lot of what is popularized is

I think that's what they're saying; ie: (a lot of "social science") (namely the pseudoscience) is unquestioningly defended, while the minority of social (non-pseudo-) science is usually ignored because it doesn't appeal to authority, ideology, or politics.


In my experience, this also happens a lot with "hard science" too, and a charge against various investigations in soft science and philosophy are led by certain "appeals to authority" themselves - including and not limited to unquestioned positivism, the unsubstantiated belief in objectivity in science etc. And these are as ideological as any other "soft science" defender - only more pernicious, since such views are rarely admitted to be ideological in themselves.

Science is as much predicated in faith as any religion. (Faith that people correctly report the "truth", etc.). The congregations of, let's say, Newtonian physics may be pretty skeptical, as they can examine claims by throwing a ball around, but most modern "hard" science takes place on scales too small or large for any person to observe. That takes a lot of faith in instrumentation and models.

The fact that your computer works is proof that our theories about very small scales are extremely accurate.

You only have to depend on faith if you don't understand the math.


The article title is misleading. It is about the wisdom of diversity in well refined arguments. People tend to magnify their stupidity and ignorance in groups to achieve confirmation bias as the beginning of the article eludes.

I don't really see how the article title is misleading when you view it as a play on the classic wisdom of the crowds model. Moderate polarisation creates more diverse opinions which makes the Wikipedia pages better because they present a wider array of arguments.

Maybe you mistakenly interpreted "polarized crowds" in the title to mean that a crowd that is extreme in one direction? In fact, it actually does mean "polarized" in the sense that the crowd contains both poles.

I focused more on the word crowd than polarized and I stand by my previous comment.

> If you have these different ideologies, it’s associated with different filters on the world, different intakes of information, and so when it comes to constructing reference knowledge on an encyclopedic web page that’s supposed to thoroughly characterize an area, you do a much better job because you have a lot more information that’s attended to by this ideologically diverse group.

It's worth noting that this doesn't just apply to groups. There is such a thing as an ideologically diverse individual. See for example the Ideological Turing Test [1]. If an individual can successfully apply the filters of diverse ideologies, they at least have the potential to swap out those filters and apply them to their scientific judgments.

Perhaps we should should systematically apply and score such tests to scientists and use the results as additional data with which to evaluate their work. True, ideologically homogeneous people could learn just enough about their opponents' views to score well on such a test. But even cynically pursued, learning to do that would train them to have those diverse lenses in their toolkit, making them easier to apply and consider.

[1] https://www.econlib.org/archives/2011/06/the_ideological.htm...


Link in parent is dead; looks like correct is: https://www.econlib.org/archives/2011/06/the_ideological.htm...

Your premise is appealing, but how would it fare knowing that conservatives are better at understanding liberals than liberals are at understanding conservatives (the former are able to empathise with liberals’ values, all the while holding their own as superior, whereas liberals do not consider conservative sensibilities to be ethical values at all)[1]? Don’t you think that an ability to empathise, which already exists at least in the case of conservatives, doesn’t seem to lead to more nuanced views? I’m not questioning the obvious benefits of education or injection of diverse points of view into one’s environment, only your premise.

Wouldn’t a simpler premise, making one’s bias apparent by stating a binary preference (conservative or liberal, but maybe less grossly-defined) somewhere in one’s piece of work, achieve the positive outcome of allowing readers to better analyse content, without betting on the potentially positive side-effects of trying to game a test?

[1]: https://theindependentwhig.com/haidt-passages/haidt/conserva...


This article inspired me to go and compare the Barack Obama and Donald Trump Wikipedia pages. It is layering subjectivity on top of subjectivity, but the treatment of Trump does seem pretty fair and the talk page goes into excruciating detail on any number of topics. The result being an excellent job done of finding references and characterising the situation with a careful and mature perspective.

I hadn't thought about it in that light before, but the highly partisan nature of the Trump Presidency does seem to have resulted in a level of quality that neither group of partisans would be able to achieve alone. Very encouraging what good editorial policy can do.


That's not a representative example - those are extremely high profile pages (maybe the highest). Wiki pages for less-known polarizing figures are far below that standard.

A sample size of 1 is not overwhelming evidence, it is true, but the page being high profile is a point in favour not against the argument. But if polerization did reduce quality, I'd expect to see evidence of it on Trump's wiki page. That doesn't seem to be the case.

Wiki pages for less-known figures being to a lower standard is consistent with the idea that popular interest + little consensus => high quality, rather than an alternative such as general consensus on topic => high quality.


The edits to most popular pages will be most scrutinized. Obama's and Trump's talk pages have the most revisions out of all Wikipedia articles [1] and are second and third most viewed [2], so they are the hardest articles to edit in a way that won't have hundreds of eyes on them right away. The Wikimedia Foundation gets $100 mil in revenue per year and having these two pages look non-neutral would be awful for their fundraising. I think it's just that extremely high interest => usually high quality and polarizing topics have lower quality than average. The consensus was never established on some quite popular controversial articles e.g. [3] and this is heavily criticized by one side of the argument [4].

[1]https://en.wikipedia.org/wiki/Wikipedia:Database_reports/Pag...

[2]https://en.wikipedia.org/wiki/Wikipedia:Multiyear_ranking_of...

[3]https://en.wikipedia.org/wiki/Gamergate_controversy

[4]https://www.reddit.com/r/KotakuInAction/comments/9l30zj/how_...


That is hardly neutral.

Right in the second full paragraph, Trump's page turns nasty. The fourth full paragraph is nasty too. OK, seen enough...

Over on Obama's page, every one of the intro paragraphs is nice. I also note the large section on religion that does not even acknowledge that famous interview in which he slipped up and said "my Muslim faith".


In the book study mentioned at the beginning of the article, how is causality proven such that politics guides scientific interest and not the other way around? Isn't it a more natural conclusion that your politics is formed by the media you consume?

Wikipedia toxicity and edit wars?

Sadly enough.

TFA states:

> On the contrary, they showed politically diverse editor teams on Wikipedia put out better entries—articles with higher accuracy or completeness—than uniformly liberal or conservative or moderate teams.

The problem is that many of Wikipedia's editor teams are not politically diverse. And some, based on my limited experience, seem totally nonfunctional and dominated by trolls.

Maybe polarization does generate high-quality articles about popular topics. But for topics that aren't so broadly polular, toxicity and edit wars seem more likely.


Wikipedia is the version of past events that people have decided to agree upon.

Where people, version, and agreed upon are, or can be, dynamic.

To wit, Wikipedia allows us to view history, view / participate in talk, and even edit an entry.

Indeed, for any sufficiently complex topic, it would seem to me, there is ongoing and lively debate, or at least discussion. Perhaps even new evidence from time to time.


Yes, but not just any people. There is a bias toward people with free time and the ability to write. This heavily favors unemployed people (such as those with humanities and liberal arts degrees) and those without families. This clearly introduces a political bias.



Applications are open for YC Summer 2019

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

Search: