Wikipedia: Caveat Lector

I received an anonymous email from someone who was concerned that the birthdate of author Howard Zinn was presented inaccurately in this anonymous comment that was left on this old thread on this blog.

The emailer explained that the birthdate misinformation from the anonymous comment had propagated via Google, and that someone had used it as a source of biographical information for the Wikipedia entry on Howard Zinn. The email quoted a purported transcript of a message from Zinn himself, attesting to his real birthdate, which was not the same as the one cited by our anonymous commenter or, apparently, Wikipedia. (I have no idea which of these birthdates, if any, is the real one, but that doesn’t matter for this discussion.)

The emailer was eager for any Zinn birthdate records on this blog to be accurate, and I reopened comments on the old post so that he could leave a note. Let no one think that this blog would deny Howard Zinn his rightful birth anniversary.

All of the above seemed kind of wacky to me, but the serious point is what it reveals (or confirms) about Wikipedia. Here was a biographical entry about someone whose life is no mystery, who is probably available by email and who probably has a publisher or agent who can provide accurate biographical data. Yet one of the Wikipedia authors relied on uncorroborated assertions from an anonymous comment in an obscure online discussion, apparently found via a casual Google search, as a source of factual information. And now someone is making the Internet rounds in an effort to clean up the sources of the misinformation that was being spread by Google, lest other Wikipedia authors repeat the first guy’s error.

Whatever else I learned from this curious experience, it is obvious that Wikipedia cannot be trusted at all, not even as a source of routine biographical information that has no political or ideological significance. Reader beware.

8 thoughts on “Wikipedia: Caveat Lector”

  1. The problems with Wikipedia are simple:

    1) people make mistakes, and without snopes-like resolve to research, mistakes tend to persist

    2) when mistakes persist, they tend to become urban legends, which are notoriously difficult to debunk

    3) people have biases, including those at the top of the Wiki dispute-handling chain, which means some contentious issues will be misrepresented (see, for example, the entry on Little Green Footballs.)

  2. You have discovered what everyone knows, that Wikipedia contains errors. What is interesting is whether those errors get fixed. As we can see, they do.

  3. Constant,

    What is interesting is whether those errors get fixed,

    Yes, but there is a time interval during which fallacious information has the wikipedia imprimatur. That information gets copied and will persist long after the wikipedia entry has been corrected.

    The internet has dramatically lowered the logistical cost of disseminating information. The central informational problem now is sorting bad information from the good.

  4. Give this “anonymous” fact-checker the Pulitzer Prize!

    If this guy has a blog, maybe he could turn a buck by posting Wikipedia’s misinformation next to his researched corrections – submit to search engines, then court advertisers. Or he could sell his research to Wiki’s editors. He must be rewarded!

    Still, I don’t think we can expect one internet encyclopedia to be the font of our modern knowledge. I believe we will always need at least three competing search engines to counterbalance the flaws each has. I think the development of an information oligopoly is a better long-term goal, than the creation of an information monolith.

  5. The cure is stylebooks requiring versioning of wikipedia content. If you see that a wiki reference is version 150 of a page and you look it up and see that this page is on version 2000, it’s as reasonable to doubt the reference as it is to doubt a Britannica cite from 1922 on a scientific topic.

  6. Wikipedia or Wankerpedia? Wikipedia is an attempt to use evolutionary-style approaches to winnowing of information. Evolution requires variation and selection according to some survival-related criteria. Wikipedia has variation, but its selection mechanisms are perverse — generally, the versions that survive are those done by the most fanatic wankers who have nothing better to do. This is probably inversely proportional to the motivation to actually arrive at correct facts. I suggest it should be referred to as “Wankerpedia”.

    Linus worrks as a production model becaause its products are tested against a real-world criterion that everybody can agree upon — either the code works, or it crashes. Wankerpedia does not have this readily-available correction mechanism, which is whyu it doe not deliver the enviable results of Linux despite attempting to use its model.

  7. Caveat lector applies to chicagoboyz as much as anywhere else. For example, this thread assumes I am male. I am not. I am the person who brought the error in Zinn’s birthdate to your attention.
    Do you believe everything you read in the newspaper?
    A dictionary? (Which one? Dictionaries are governed by style du jour on some “facts.” )
    Do you believe everything you read in encyclopedias on printed pages in bound books? You ought not. Books, including encyclopedias, sometimes err, and are governed by the viewpoint, bias, and world view of their writers and editors.
    Everything written turns on a point of view, including the selection of facts, which to include and which to exclude.

  8. Anon Girl,

    I think that you are missing the point. Newspapers, books, dictionaries, encyclopedias, TV news shows, and blogs for that matter are not necessarily more accurate than Wikipedia. However, unlike Wikipedia the other media have clear chains of accountability for what they publish. For example, the identity of a book’s author or a blog poster is usually known and can be factored into readers’ calculations about the accuracy of the material. Even dictionaries and encyclopedias, which are usually written by committees, typically publish the names of the committee members; and their brands are valuable, so publishers have a strong incentive to avoid publishing inaccuracies under their names. This is not the case for Wikipedia, where nobody stands to lose money or be fired because of bias or factual inaccuracy in articles, and where contributors are volunteers with sometimes unclear motives (and often anonymous as well).

    The open-source model works for software because software is easily evaluated by users: it either works or it doesn’t, and if it works its features are readily compared to the features of competing products. An open-source encyclopedia is much more difficult to evaluate by users, who cannot readily determine if, for example, a reported birth date is accurate. The open-source encyclopedia won’t crash if someone inserts a grossly inaccurate entry, so the whole body of work has to be continually checked by volunteers, of whom there are never enough and who may have their own agendas. With a dictionary or a TV news show or a blog I can eventually figure out the biases, and if necessary read between the lines or ignore reports on particular topics. But what am I supposed to do if I don’t know the authors or their biases — if entries are composed, and may be revised at any time, by an anonymous (or even partly anonymous) swarm of authors? I have to distrust the entire body of work even though some of it is no doubt of good quality.

    BTW, I did not assume that you are male, nor do I now assume that you are female. You probably are female, because why else would you quibble about such an issue, but all I have to go on is your assertion. I don’t really know anything about you except your IP address. Your defense of Wikipedia would be more convincing if we knew more about you and the other Wikipedia contributors. I’m not saying that you should reveal your identity, merely that there appear to be inherent flaws in the Wikipedia model.

Comments are closed.