Wikipedia's Not as Biased as You Might Think, Say Harvard Researchers (qz.com)

An anonymous reader shares a report on Quartz:In a sea of biased content, Wikipedia is one of the few online outlets that strives for neutrality. After 15 years in operation, it's starting to see results. Researchers at Harvard Business School evaluated almost 4,000 articles in Wikipedia's online database against the same entries in Encyclopedia Brittanica to compare their biases. They focused on English-language articles about US politics, especially controversial topics, that appeared in both outlets in 2012. In its initial years, Wikipedia's crowdsourced articles were tinted very blue, slanting more toward Democratic views and displayed greater bias than Britannica. However, with more revisions and more moderators volunteering on the platform, the bias wore away. In fact, the upper quartile of the Wikipedia's sample had enough revisions that there was no longer any difference in slant and bias from its offline counterpart. More surprisingly, the authors found that the 2.8 million registered volunteer editors who were reviewing the articles also became less biased over time.

Comments

Top quartile, you say?

[LynnwoodRooster]

So they've managed to validate that just 1 out of 4 articles is free of bias. And the other 3?

Re:Top quartile, you say?

[serviscope_minor]

So they've managed to validate that just 1 out of 4 articles is free of bias. And the other 3?

You're right. The article on semigroups, particularly the bit about monoids has a distinctly liberal bias.

Might not be neutral but...

[TWX]

...the fact that the editorial system still follows the king-of-the-hill model, where those that choose to sit on pet-pages win simply by undoing any other changes simply because they don't like them, will leave the entire thing biased in some fashion or another.

I will not contribute to Wikipedia anymore. I've had edits that I could provide support for undone by some self-important busybody whose only credentials were the ones they defined when they signed up for an account on Wikipedia. Forget that.

Re:Might not be neutral but...

[Anonymous Coward]

This is, in fact, a terrible fact about Wikipedia. It's also the only way to actually achieve neutrality: a very strong bias against new people, since they're usually just there to slant an article towards their preferred point of view.

Re:Might not be neutral but...

[Anonymous Coward]

Experienced editors like former Wikipedia admin Ryulong are among the most guilty of introducing bias and incorrect information into articles. The worst part is that even in the face of these actions, it was ridiculously difficult to have an individual like this removed. The old hands can be just as damaging to an article.

Re:Might not be neutral but...

[AthanasiusKircher]

The "king of the hill" model you describe has built an AMAZING encyclopedia... a fact that you ignore and shows you are willing to throw the baby away with the bath water.

Huh? At no point did I see GP saying that Wikipedia wasn't a significant accomplishment. He was pointing out a flaw. Are people not allowed to criticize something without bowing before it first and acknowledging how "great and mighty" it is?

And please cite where GP said we should throw Wikipedia away. He said it has some problems, and he's not willing to waste his time editing because of those problems. I don't see anything where GP suggested that we get rid of Wikipedia -- he's noting something's wrong. The rational response would be to say, "Hey, yeah, sometimes that happens -- how can we improve?"

If you think Wikipedia sucks, then don't use it or you are hypocritical.

What's funny is that you accuse GP of taking a "my way or the highway approach" when GP didn't, but then you turn around a second later do exactly what you accuse GP of doing -- taking a completely irrational stance and making false dichotomies.

What -- if you recognize flaws in Wikipedia, you're hypocritical to ever use it?? Huh? In what universe do people only ever make use of absolutely perfect products or resources? Are they all just supposed to keep their mouths shut when something's wrong (but not necessarily terrible enough to "throw the baby out with the bathwater," which you just said was bad).

Second, your characterization for the most part is inaccurate. Wikipedia does NOT work on a "king of the hill" model. It works based on a set of policies and guidelines that focus on writing a good, free encyclopedia (see https://en.wikipedia.org/wiki/... ). Sometimes you do get people that are too heavy-handed but there are ways of dealing with that. More likely is that your edits were problematic but you couldn't understand why.

TL;DR -- you didn't learn to Wikilawyer, like me.

Plus, your comment above already shows YOU are violating some of the established rules: you are not assuming good faith when you write "undone by some self-important busybody whose only credentials were the ones they defined when they signed up for an account on Wikipedia". It's clear that you didn't even try to discuss the changes but just got upset that your changes were removed and rage quit because you didn't get your way. That's the kind of uncollaborative person that is poison to a group project.

How is it that you know so much about this editor but know GP was being a jerk and didn't know what he wasn't talking about?

Pretty sure you're the one NOT "assuming good faith."

Without seeing the particular edits you are upset about I cannot say for sure

Oh, oops... I guess it's actually CERTAIN that you're the one who doesn't know anything about the situation and is just not "assuming good faith."

but in my experience, basically ALL people who make the complaints you are making were making SHITTY edits that lowered the quality of the articles but through either ignorance of Wikipedia's policies and guidelines or sheer inability and incompetence were unable to realize it.

And in my experience, basically Wikipedia has a major vandalism and troll problem that only gets worse every day. That causes editors to reflexively overreact to any changes on an article they're "sitting on" as "klng of the hill" and revert. That doesn't mean those reversions are always unjustified, or even than these people are acting badly.

What it does mean is that Wikipedia's vandalism and poor edit problem is so bad from fighting bad the idiots and the trolls that good, well-meaning editors don't have time to make reasonable evaluations.

What it does mean is that Wi

Re:Might not be neutral but...

[HornWumpus]

Even the article you link admits that the policy is not working (see discussion of 'tag teams').

Until people that participate in 'tag teams' are banned from all editing this shit will continue.

Re:Might not be neutral but...

[TWX]

This is it in a nutshell.

The article in-question that was the straw breaking the camel's back was about a particular form of ballroom dance. The article did not actually describe how the dance was performed. Didn't mention time signature or beats per minute, didn't describe the basics of the footwork (the "box step" that is common to a dozen dances of European heritage) or anything about the grip or how lead-follow works.

When I remember back to reading the encyclopedia as a child, was that the article would have some practical description of what the topic was and if it was an action, what that action entailed, followed by history and possibly a comparison of competing schools of thought on the subject. The Wiki lacked the practical part; reading the article would not tell someone what the dance actually was; one could not recognize the dance if seeing someone do it after reading the article.

I described the dance. I had paragraphs on box-step, lead expectations for lead and follow, types of "position", and time signature and speed along with the kinds of music commonly danced-to. Within minutes the changes I had made were reverted with a line, "this should be on Wikihow," or something like that. No discussion on the discussion page, no suggestion for doing something like forking-off the box-step, or anything else related to using the content that at that point did not exist.

While I grant that I did not specifically dig through the rules to figure out an appeal process, I have a life, a job, and hobbies. If Wikipedia is not able to figure out how to use contributions from people like me without requiring me to spend an inordinate amount of time making those contributions stick, then they do not deserve contributions from people like me.

Re:The Harvard article doesn't mention the strike

[JoshuaZ]

So? This is recentism in a nutshell. It is a mistake to think just because an event is happening now that it is therefore more important. Harvard has a 300 year history; an event in one specific year needs to be a really big deal to make it into the primary article on its history.

Re:Show me the Wikipedia entry for "Bradley Mannin

[Anonymous Coward]

Bradley Manning. You're welcome.

Filter bubble effect

[QuietLagoon]

... More surprisingly, the authors found that the 2.8 million registered volunteer editors who were reviewing the articles also became less biased over time. ...

Which shows what happens when people get information from outside their comfortable filter bubble ... they tend to take on less extreme views.

.

Which is one of the reasons why some of the more extreme news sites often say something to the effect, ~you don't have to go anywhere else, we tell you all the news you need to know~

Absolute horseshit

[Anonymous Coward]

Anyone who thinks Wikipedia isn't biased is someone who has never tried to contribute to Wikipedia.

BOLD, revert, discuss cycle

[tepples]

Did you ask the editor who reverted that change to explain the reversion? If you ask, wait a week, and try the edit again, the burden falls on the reverting user to explain why the edit should not stick.

Re:BOLD, revert, discuss cycle

[sexconker]

"Not properly sourced."
But you provided all the links to sources.

"No original research."
But it's not my own research. Look at the sources I provided.

"Not noteworthy."
What? This is a hugely significant!

Page has been locked, your IP has been banned from editing.

Thank you for contributing to Wikipedia. Please click on Jimmy Wales's ugly face to donate.

How do you accurately measure bias?

[Wycliffe]

Yes, like their main news channel being self-described as "fair and balanced"... oh, wait!

I have plenty of conservative friends that think fox news is less biased and plenty of liberal friends who disagree. I even have conservative friends that think that fox news is a little too left leaning.

I'm not sure how you even go about measuring bias. Do you find the most conservative and the most liberal person you can find and ask them? I would love to see more shows with opposing views but it would be hard to do the extremes without turning in to the Jerry Springer Show. I've found that it's easy to talk politics to people who are in the rational middle but if you get a diehard democrat or a diehard republican then they get very angry if you disagree with them. Both sides have taken the moral high ground and think anyone who disagrees with them is evil and irrational.

Re:How do you accurately measure bias?

[MightyYar]

I know it's not popular to try and approach things as scientifically as possible, but in this case they describe their methodology in the original paper.

In simple terms, they identify text samples which express either a Republican or a Democratic view, and then tally them up.

This seems like a more disciplined approach than asking one's friends what they think.

Re:How do you accurately measure bias?

[bluefoxlucid]

That approach only indicates an equal amount of bias. It doesn't indicate if the Democrat views are expressed where the bulk of evidence suggests their views aren't loony, and the Republican views are expressed where the bulk of evidence suggests they're actually correct.

What if you wrote an article about vaccine causing autism, where you cited 12 rigorously-evaluated scientific opinions, and 12 fervently-espoused lunatic ravings? The article would tally as "unbiased", but it would actually be biased: a preponderance of evidence would discredit the lunatics, yet the article would be shored-up to give them equal footing. In other words: you'd handicap the visibly-correct view and give additional, unmerited support to the visibly-unsupportable view, slanting the article toward a side that can't stand on its own merits.

Re:How do you accurately measure bias?

[phantomfive]

I'm not sure how you even go about measuring bias.

Statements can fall (roughly) into three categories:
1) Statements that are well supported by facts.
2) Statements that are poorly supported, thus contain quite a bit of uncertainty.
3) Statements that are preferences (ie, I prefer lower taxes and a weaker safety net).

You can measure bias on all of these. The first is easy, because a biased viewpoint is outright wrong:
1 example) "Rich people pay no taxes" or "we could get rid of the national debt if we got rid of welfare" (both wrong).

When there is quite a bit of uncertainty, biased people tend to state things as fact:
2 example) "The treaty will prevent Iran from getting nuclear weapons" "The treaty guarantees Iran will get nuclear weapons."

Often people who utter #3 type statements state their preferences as fact:
3 example) "GMOs hurt the world" or "taxes are too high."

(Most political disputes actually fall into category #3, believe it or not. Things like gay marriage, abortion, allowing immigrants, allowing refugees......these are framed as questions of "right" or "wrong;" but mostly they are preferences. Even a topic like AGW, the political question is, "what should we do about it?" and there are a huge range of options even if we could agree on what the goal is).

Re:How do you accurately measure bias?

[bluefoxlucid]

They provided a definition of "balanced", yes. My point is more that "balanced" is hard to measure and that their definition has procedural flaws.

There's a secondary issue in that people already perceive the concept of "fair" and "balanced" as a positive thing, and so will tend to accept a reasonable ideal of what "balanced" might mean and then accept conclusions. That means a large amount of the population won't scrutinize the semantics in the way I did, and then will accept that a certain set of information is more-accurate because of the emotional sense of a lack of bias, fairness, and other ideals which provide comfort.

At this point, we're getting into complex psychology and persuasive argument. As you say, measuring things against the amount of actual evidence included and excluded and on the emphasis on that evidence would show something akin to reliability. Science can be wrong; it tends to be less-wrong than other approaches, and to improve its correctness over time. At the same time, it's possible to emphasize poorly-supported conclusions and de-emphasize strongly-supported conclusions--for example by talking at length and in detail about something which is plausible given the evidence but unlikely, while talking scantly on something which is almost-certain given large volumes of high-quality evidence. We can abuse science as much as anything else.

That doesn't even get into the Einstein trick--drawing large conclusions from small sets of data. I use that methodology a lot, and frequently to great effect (so much so that I independently re-discovered a large amount of modern macroeconomics in a few days of examining a very small set of samples spanning economic history, including everything Malthus was right about, the advantages of trade, and how technical progress affects an economy, without actually bothering to read up on economics); however, you tend to either be surprisingly-correct or hilariously-wrong for the dumbest reasons. Einstein conjecturing works when you have enough data points to align your trajectory properly; if you're a few degrees off and you reach too far, you miss by enormous margins (often going completely backwards). Usually you want to use it as a guide to figure out what data points you need to improve your understanding, and to verify your facts (i.e. if what you've worked out is true, many things should also be true; are they?).

Stuff like that lets you (cautiously) extend well-established science to further conclusions; compare that against some other fringe-science, and you might find the poorly-supported view is ridiculous--or that it makes a lot of sense, and merits more examination.

Strict analysis is hard. Wikipedia's policy tries to avoid stuff like people drawing long, unsupported bullshit conclusions (Einstein conjecturing ending in stupidity) by banning original research; that doesn't stop people from cherry-picking research, giving you what might look like a strongly-scientific article when really it's a load of horse shit.

Re:How do you accurately measure bias?

[jdavidb]

I'm not sure how you even go about measuring bias

Wikipedia's NPOV policy is actually a good standard, although sometimes the implementation leaves something to be desired. And trying to follow it was a very good exercise for me earlier in life. It taught me how to speak about multiple points of view respectfully even when I disagreed with others.

Under NPOV you don't report "absolute truth." You report what people believe, why they believe it, etc. It can be very insightful as a tool to understand other viewpoints.

Re:How do you accurately measure bias?

[bluefoxlucid]

Asking for a better approach is usually the last refuge of people who can't prove themselves right. You're also opening yourself up for much of the other arguments I made by claiming what you intuitively expect from the word "balanced": most people expect "balanced" to equate to "informative" and, by extension, "correct."

I maintain that giving less-credible positions similar weight to credible positions is not balance. It would not be balanced to write an article on homosexual adoption that gave equal weight to both the rights of homosexual couples and the ethics of leaving children with a group of people known to be mostly child-predators.

Re:Wikipedia no more biased than British ivory tow

[fred6666]

Let's face it, the rest of the English-speaking world is more liberal than the USA.
The problem here is US politics which has a conservative bias, not Wikipedia.

Bias of people versus bias of organization

[sjbe]

Anyone who thinks Wikipedia isn't biased is someone who has never tried to contribute to Wikipedia.

Wikipedia as an organization and as a website generally isn't biased on most topics. Kind of hard to have a bias about some random regurgitation of a technical fact like a chemical or math equation. Some of the people who contribute to Wikipedia very much are biased because most people carry assorted biases with them. But these biases generally don't seem to lean coherently towards one political persuasion or another across Wikipedia but rather are generally confined to specific hot button topics. The hope is that the various biases of the contributors will mostly balance out and the objective facts will remain. This doesn't always happen but it seems to happen often enough that one can say it usually works and not look stupid saying so.

What is amusing/depressing is that some people reflexively claim that any facts that don't match their pre-existing world view must be biased.

Voltaire's Opinion

[avandesande]

"Perfect is the enemy of good". It's not perfect but it's good enough to be useful.
https://en.wikipedia.org/wiki/...

Re:So Britannica is more republican

[GLMDesigns]

You realize that creationism is a Democratic and Independent "thing" as well.

Your position is poor. It's like saying that since Democrats (some democrats) believe in homeopathy and gaia that Democrats are anti-science and anti-facts.

re Wikipedia

[harvey the nerd]

Hahahaha Havard !

Re:So Britannica is more republican

[fred6666]

I don't really care about some democrats and some republicans in the general population. A party can't be accountable for every single supporter. But some high profile republicans, including some elected officials, reject climate science and/or are creationists.

Are there Democratic elected officials who reject evolution? Like it or not, it's much more a republican thing.

Concerning homeopathy, do they vote laws against science-based medicine and favoring homeopathy? If not I don't really care if they waste their personal money on homeopathy. Anyways don't you think republicans are just as likely to believe in homeopathy?

I am not saying Republicans are anti-science and anti-facts on all issue, but they are on these very important two. So as a whole, I expect a pro-science and pro-facts encyclopedia to be "biased" against such a party, if only for this reason, and I consider it to be a good thing.

Wiki wars

[BrookHarty]

Real drama is in the talk pages. Check ones like how dangerous pit bulls and whether its owners vs breeds, gamergate, mens rights, any politician, politics, political view, etc.

Basically anything with a point of view, will slant to the left, that's because more editors are of leftish ideology, and wikipedia staff are mostly leftist.

The article itself says it has a slight leaning left, what is "slight" in numbers, 10%, 20%?

Re:Fox News headline version

[Impy the Impiuos Imp]

Fox news was a stroke of business genius -- have a different slant that doesn't make people feel attacked. It worked and shot to popularity.

The problem with ABC and to a lesser degree CNN is their True Believers refuse to recognize an equal bias on "their side" (to say nothing of NPR.)

Most bias isn't in fact checking, but the choice of stories to harp on, all day, every day, shouldn't the govrnment do something, wink!

Re:Fox News headline version

[Chris Mattern]

Pay $10 to watch what you want to hear

Pay $20 to listen to what you want to see.