Email Debate::Wales Discusses Political Bias on Wikipedia

Conservative blogger Robert Cox, who writes the National Debate blog, told me he was amazed at the quality of Wikipedia and thought it was a great resource. But there was something about the free online community-generated encyclopedia that was getting under his skin — what Cox believed was a liberal bias in many hot-button topic entries, despite Wikipedia’s principle of giving a neutral point of view (NPOV).

Cox felt there was a liberal tilt to the entry on George W. Bush, Bill Clinton, and the partial-birth abortion entry, to name a few. Plus, at one point, Wikipedia founder Jimmy Wales invoked the dreaded WP:OFFICE command — basically a unilateral edit done only by Wales — to tone down a scathing liberal point-of-view entry on the conservative site NewsMax.com.

So I thought it might be a good idea to have a three-way email discussion between Cox (pictured at left) and Wikipedia founder Wales (pictured above) to find out what the project’s effervescent leader thought about political bias and how Wikipedia deals with it.

The following is an edited version of that email discussion, which touches on political bias, transparency issues, and how articles get to be “semi-protected” and what that means.

MARK GLASER: This is a question for both of you. How would you characterize the political leanings of the Wikipedia community — both the existing one and the earlier, smaller community? Is there any data or surveys to back that?

JIMMY WALES: The Wikipedia community is very diverse, from liberal to conservative to libertarian and beyond. If averages mattered, and due to the nature of the wiki software (no voting) they almost certainly don’t, I would say that the Wikipedia community is slightly more liberal than the U.S. population on average, because we are global and the international community of English speakers is slightly more liberal than the U.S. population.

There are no data or surveys to back that.

MARK GLASER: So you feel that Wikipedia having a “slightly more liberal” slant than the U.S. is OK? How does it affect the goal of neutral point of view and should you do something to counteract it in some way?

JIMMY WALES: I do not think it affects the goal at all. The question totally misapprehends the process. The idea that neutrality can only be achieved if we have some exact demographic matchup to United States of America is preposterous, as I am sure you will agree.

ROBERT COX: If there is any data out there I assume Jimmy would know about it so I believe that all these [come from] anecdotal evidence on this question.

They may out there but I do not recall ever seeing an entry that was non-NPOV on the “conservative” side but I have read many entries that are non-NPOV on the “liberal” side.

Wikipedia says this: “Earlier on, we had a systemic bias toward liberal issues. However, as Wikipedia has grown, and become more mainstream, the liberal contingent has declined as a proportion of Wikipedia in general. Perhaps our other biases will be partially neutralised in the same way.”

This suggests to me that the Wikipedia community was early on biased
“liberal.” That they say bias might now me “partially neutralised” suggests that it has not been “fully neutralised.” That the proportion has declined means the issue has not gone away. So, I would say that Wikipedia itself sees this as an ongoing issue.

I have had my own direct experience with editors of the Keith Olbermann page which suggests this is the case. I edit a blog called Olbermann Watch. Not that it was ever my goal in life but I am now the leading blog critic of Keith Olbermann and a recognized authority on Keith Olbermann (citation: quoted in Washington Post, New York Observer, Hartford Courant, Online Journalism Review, etc.).

Not only do I know a great deal about Keith Olbermann, I also have a good deal of familiarity with some of the Wikipedia editors who have watch-listed his entry — liberal fans of Keith Olbermann. Some of these fan/editors have declared online that the Keith Olbermann page is their “pet project” and, not surprisingly, the entry reads more like a “fan site” than an encyclopedia entry. Some of these editors have openly sought to use that page to market their own fan sites and forums. Not surprisingly, the Keith Olbermann entry is massively non-NPOV.

From time to time, I have attempted to correction misinformation or edit a section to make it NPOV. Those edits are typically “reverted” within the hour without explanation or discussion. Over the past month, I signed up for an account with my name in the User ID; many of the editors know who I am and are openly hostile to my editing the site. These editors aggressively revert any edits I make to the entry. When I attempt to discuss my recommended edits they ignore me. When I make the edits they criticize me for not discussing them. If I continue to make edits they complain to the “Wikipedia cops.” Even after posting a detailed exposition on why the page is massively NPOV they have ignored the substance of my post and instead attacked the messenger.

It’s a neat trick — they demand that I propose changes on the discussion page, ignore me, then when I go ahead and make those changes they revert them, all the while complaining to an admin that I should be banned from editing because I won’t “discuss” changes. The real issues is that these people WANT the page to be massively non-NPOV and resent any efforts to alter their “pet project.”

MARK GLASER: What do you think, Jimmy, about this case? Is there a way that Wikipedia could be more transparent about how the editorial process works, who has the power to make final edits and ban people, etc.?

JIMMY WALES: I can’t imagine how we could possibly be any more transparent. All of that information is discussed openly on the website. There is no such thing as a “final edit.” The banning process is clearly specified.

Some of these fan/editors have declared online that the Keith Olbermann page is their “pet project” and, not surprisingly, the entry reads more like a “fan site” than an encyclopedia entry.

I dispute this characterization. It sounds like Bob is just unhappy that (a) we don’t follow his lead in calling fans of Olbermann “Olby Loons” and (b) he was banned from editing Wikipedia for his misbehavior on this page.

ROBERT COX: The edits I made to the Wikipedia entry are available on the site. The edits I attempted to make were made in good faith and done solely to make the site more NPOV.

I never proposed or made an edit “calling fans of Olbermann ‘Olby Loons’” and have never used that term or any other such disparaging term on Wikipedia. I HAVE stated a fact — that most of the active editors on the site are left-wing, Keith Olbermann supporters.

I was banned from editing for 24 hours because I was accused of violating a rule (3RR — you can’t revert an article three times in 24 hours) that I did not — to my understanding — violate. I did not know about this rule and I did not know, initially, that the person who “warned” me about violating the 3RR rule was an admin but once I understood that to be the case I complied with his warning and have not made any major edits to the page since. I still do not believe I violated this rule but have made it my own policy to make only one or two edits to the page per day.

There is something close to a “final edit” on Wikipedia; that is when a handful of individuals take “ownership” of a Wikipedia entry and aggressively revert any attempts to edit “their” page. I believe that is the case here with the Keith Olbermann entry.

JIMMY WALES: Just make some good faith edits, and write in a non-hostile manner on the talk page that you have an interest in trying to make the article high quality and neutral. Reach out with love and kindness to your opponents and see what happens. I will watch and not interfere.

MARK GLASER: How about the George W. Bush entry? How and why did you decide to lock up the entry? You talk about Wikipedia being very transparent, so explain in detail what has happened with that entry and who decided to change the editing status of it.

JIMMY WALES: This is a question like “do you still beat your wife?” in the
sense that the question contains a false premise. The false premise is that the George W. Bush entry is “locked up.” It is not. Any fully registered user of the site can edit the entry right now. Robert Cox can edit the entry right now. If you have a user account which is more than 4 days old, you can edit the entry right now.

The entry is edited all the time, hundreds and hundreds of edits can be
seen simply by clicking on “history.” The question seems to imply that there is some mysterious and possibly non-transparent process by which some entries get locked from further editing, quite possibly for political reasons. It’s a daft question, though, because there is no such process, no articles are locked for political reasons, and you can easily and quickly confirm this for yourself.

In the old days, we had to “protect” articles temporarily to deal with episodes of vandalism or when a particularly ugly flame war broke out. While an article is protected, no one is allowed to edit it, not even admins. (Admins have the technical ability to edit, and rarely do edit to correct obvious vandalism which was inadvertantly protected, but routine editing is never done while articles are protected.)

We didn’t like this, because it is the sort of “hard security” we prefer to avoid. I am unaware of ANY cases where protection was done for political reasons, and in fact I am unaware of any serious allegations of it. If you examine the protection logs, you will see what I mean.

There are a number of social rules surrounding protection to guard against protection as a mechanism for introducing bias. First, admins are not allowed to protect articles (or use any admin powers) in any editing situation in which they are personally involved. Second, any other admin can undo any admin action, so that there are checks and balances in the system against abuse.

Even with all that, we still were not happy. Protection to deal with vandalism was overkill. So we invented what is called “semi-protection.” Semi-protection is a state in which articles can still be edited by any user of the site, but not by anonymous IP numbers. (We also exclude newly created accounts, defined as the most recent 1% of all accounts, I believe, which is about 4 days currently.)

In this way, we preserve the ability of editors to edit the article in a routine fashion, while eliminating virtually all the vandalism. This is a much softer solution than the old solution.

You can view the protection log here. this shows who has protected/semi-protected the article and their reasons for doing so.

Also, the discussion here frequently includes discussions of the current semi-protection status. (Take special note of the archive filing cabinet icon over to the right, which contains older discussions.)

You will see that the entire protection discussion, through and through, makes absolutely no mention of content issues, OTHER THAN SIMPLE VANDALISM. This is as it should be, and is the routine process in the community.

To sum up, the attempt to spin that this article is “locked down” for what might be political reasons is just confused. The article is not locked down, and any and all protections that it has had at various points in time are clearly justified by the amount of vandalism that the article sees, not by any political bias.

MARK GLASER: Thanks for explaining what “semi-protection” means and how it works. I appreciate it. I actually never said anything about locking it down for political reasons. I was much more interested in understanding the process around an entry that obviously has a lot of vandalism and is highly controversial — and in my opinion, is pretty critical of Bush.

I keep hearing from my readers (many of whom I’m guessing are Wikipedians or ex-Wikipedians) that attaining NPOV is impossible, that everyone has bias and introduces it in some way. However, they say, why not just accept that there is bias and be transparent about how people are biased? So if I was a new user to Wikipedia, I would expect it to be this free online resource that had NPOV I would be surprised at this bias, no?

JIMMY WALES: You asked about the article being locked down in the context of a discussion on political bias in Wikipedia, and cited this as a possible example. I don’t know how else to interpret the question except as a question about whether the protection policy on Wikipedia is used for political purposes.

I was much more interested in understanding the process around an entry that obviously has a lot of vandalism and is highly controversial — and in my opinion, is pretty critical of Bush.

It is open to editing, if you have a problem with it.

So if I was a new user to Wikipedia, I would expect it to be this free online resource that had NPOV I would be surprised at this bias, no?

I would say that the thing that surprises and astonishes new users is that for the first time, they are confronted with a very large public resource which is well written and astoundingly free of bias. This is commonly reported upon by many new users, and indeed, by many people on all sides of the political spectrum.

MARK GLASER: You said, “The Wikipedia community is slightly more
liberal than the U.S. population on average.” Is it possible that a “slightly more liberal” community can write from an NPOV? Can anyone write from an NPOV? Doesn’t everyone have inherent biases? How does the community as a whole enforce NPOV?

I agree that I too have been amazed by the quality of most Wikipedia articles, and I think Bob would concur in his experience. But then, upon digging deeper, we find errors and bias in some politically sensitive entries. It’s easy to say “go fix it!” to any problem I bring up, but the reality is that anyone can revert what I do and it’s a more complicated problem than that…The community rules in an ad hoc manner which newbies have no power to deal with.

JIMMY WALES: Of course, any sincere person can write in a solidly neutral manner, and this is enhanced when one has good faith assistance and help from others who may not agree in every respect. There is no magic here, of course. Neutrality is a goal, and NPOV is a process.

As I often say, I am not the next Aristotle. I have discovered no new solutions to the age old problems of good solid expository writing. But I think we are onto a technique that does a great job.

It is somewhat silly and U.S.-centric in a rather frightening way to imagine that “slightly more liberal than the U.S. population on average” is an impediment to neutrality. Is there an inherent assumption that somehow, the U.S. popluation on average is exactly the right balance of people to do intellectual work?

I agree that I too have been amazed by the quality of most Wikipedia articles, and I think Bob would concur in his experience. But then, upon digging deeper, we find errors and bias in some politically sensitive entries. It’s easy to say “go fix it!” to any problem I bring up, but the reality is that anyone can revert what I do and it’s a more complicated problem than that…The community rules in an ad hoc manner which newbies have no power to deal with.

There are thousands of counter-examples to this every day. Yes, if someone comes in and makes blind edits to insert POV, then yes, those edits are likely to be reverted. But solid edits, which are properly sourced, and which acknowledge multiple perspectives, survive quite easily all the time.

I daresay that I could log out and edit as an anonymous user, making 25 straight edits in the course of an evening, and not be reverted once. And I also daresay that I could do the same thing and get reverted repeatedly.

I am hopeful that Bob will take me up on the idea of him trying again to edit the Olbermann page, with a sincere attempt to edit in a neutral fashion. If he goes to the talk page and writes a note of explanation, and (even better) an apology for his prior behavior, I think he will be warmly welcomed, and if his edits are good, they will stand.

ROBERT COX: For the sake of discussion, I’d like to grant Mark his premise that the George W. Bush entry is biased to the liberal side in some sections. I’m interested in how that bias might transmitted to a user, or is it? I am thinking of my son who is a 9th grader. He is doing a term paper and comes across Wikipedia. He reads what it says, takes some notes and writes his paper. Misinformation or biased information gets into his paper. Later, through the process Jimmy has been describing, that entry is edited, in fact it is done so well it is perfectly NPOV and approved as a featured article.

My son now has some bad information in his term paper and is marked down accordingly by his teacher. He shows the teacher that he got the information from the Wikipedia page on George W. Bush but the information he referenced in his page is no longer contained in the entry. My son is puzzled — and he has a bad grade.

Now, if my son really wanted to delve into it he could sit there with the teacher and review the discussion page and look at the history page to view past edits to see what happened to that information, how it got in there, why, who took it out and why. To that extent Wikipedia is transparent, something I understand to be an important value of Wikipedia proponents.

My question is what good does the transparency of the editing process do for my son who is highly unlikely to do anything other look at the entry itself on the day he is working on his paper. Also, aren’t users like my son far more prevalent than editors and admins?

JIMMY WALES: Wikipedia is bad now because we correct errors when we find them? I’m very sorry but I find this argument to be entirely perplexing. You could say the same of Britannica, you know. They update articles on their site, too.

ROBERT COX: I was hoping we could discuss “transparency” as it relates to editing an entry in the context of someone (in this case my son) who takes what he finds in an entry at face value, unaware that there is a long discussion/edit history behind how that section reads at the moment he arrives on the site, and that he might be arriving soon after a controversial edit has been made, and leave shortly before that change is edited again.

The edit history for the entry Mark mentioned lists somewhere around 400 edits this month alone so the entry is changing constantly — about twice an hour on average. Given this, it seems quite conceivable that several different readers could get entirely different information from the identical source.

What you suggest I tell my son in order that he be a “smart consumer” of information from Wikipedia?

JIMMY WALES: OK, let’s set a framework for that discussion. What counts as transparency? How do we compare to any other source? I frankly find questions about transparency a bit odd simply because there has never been a more transparent edit process than Wikipedia. Blogs are not as transparent, Britannica is totally opaque, newspapers are totally opaque, etc.

What you suggest I tell my son in order that he be a “smart consumer” of information from Wikipedia?

Explain how it works, and how it is better than every other source which has ever existed in terms of transparency, and not as good as almost every other source which has ever existed in terms of stability. Point out that the open editing process has strengths and weaknesses.

Show him the Nature study, so that he understands that Britannica had an average of 3 errors per article, and Wikipedia an average of 4. The important lesson there is that traditional resources are pretty good…but that despite the leather binding and the feel of permanence, there’s a lot of problems there too.

Teach him that multiple sources should be used, and talk about the kinds of facts that are more likely to be suspect.

ROBERT COX: I did ask my son whether he was aware of the editing feature of Wikipedia and, to my surprise, he said yes. He said that his friend kept editing the Adolf Hitler entry to include the observation “Hitler did not like flowers” and for some reason my son and his friends thought that to be hilarious.