Quaap home
Adventures of a stay-at-home, skeptical, homeschooling dad (etc.)


Main

The Wikipedia experiment

Though I use it all the time, I often make fun of Wikipedia and its supporters for several obvious reasons, including terrible writing and a "patchwork" feel to the articles that arises from too many editors. But the main problem is that, as a source of reliable information, it, well, blows. When confronted with these negative opinions, the defenders of Wikipedia always tell me that incorrect information, lies, and vandalism are "corrected within minutes" and that it's just as accurate as other encyclopedias.

For a select number of high traffic articles or articles on controversial topics, that is most likely true. If I went into the article on, say, abortion and added the phrase "studies show that abortion doctors love to eat the babies" or "anti-abortion protesters enjoy beating women", then yes, that will quickly be reverted as obvious vandalism. What's more, someone reading the article would probably recognize it as vandalism as discount it even if they don't feel like going to the trouble of editing it themselves.

But what if I where to add a real-sounding but incorrect statistic such as "75% of abortions are covered by Medicare"? Someone would probably remove it quickly, probably within an hour, probably a lot less. Which is good, I suppose, but a lot of people can see it in that hour. Possibly thousands of people. Do they know it's false? Do they see that it's not cited? Do they know if they are viewing the article right before or right after some Wikipedian has cleaned up the incorrect information? Now, I realize that the abortion article is locked and only logged in users can edit, it probably gets a lot of scholarly attention, etc, but the problem still exists.

However, the vast majority of Wikipedia is open for editing by anonymous users, and the vast majority of articles do not receive much scrutiny, and approximately no scholarly love. So, if someone were to go to a lower-traffic article, and add something to it that was incorrect, but would be hard to verify or even merely sounded plausible, would anyone ever notice? Would it be corrected "within minutes"? Would it get one of those "citation needed" tags quickly?

Approximately two years ago, I edited several articles on Wikipedia and added plausible but almost certainly incorrect statements to each. I say "almost certainly incorrect" because they may very well be true, entirely by accident. I specifically made them resemble things someone might have heard somewhere and has believed ever since, but has no basis in fact.

The result? No reverts. No "Citation needed" stamps. No corrections. Nothing. Well, there were a few edits: someone changed the wording of one sentence and someone else changed the format of a link it contained. Someone even added some text which expounded on my incorrect statement and attempted to give reasons why it might be that way.

I won't tell you which articles I edited, because I really want to see how long they last. But suppose I edited the article on lentils and added something wrong but plausible, like the following:

In several communities of the Indus river valley, a paste made from lentils, coriander, and cumin is used as a folk remedy for infertility.

If you saw this in an article on lentils, would you know it was a false statement? Would anyone? Does it substantially differ from 90% of the content of Wikipedia? Someone added it, and no one else has the ability or the inclination to refute it, so it stays there. Basically, it's just the same as any other statement in Wikipedia without a reference stamp: possibly true, or possibly just pulled out of someone's ass which no one has reviewed. By the way, I think the links to the other topics helps out its believability, don't you?

Are many people doing this? Adding incorrect information to articles? Almost certainly yes, and probably in a massive way, but they think they are adding correct information: something they heard once, something they read in a book, something that is slanted or just plain wrong. I'm willing to bet there are a huge number of these "facts" scattered all over Wikipedia, just waiting to be read and passed on. And again, even if newly added incorrect information is corrected quickly, how many people read it in the meantime? Much worse is that Wikipedia is often the first or second result returned by Google for almost any topic search, making it the de facto source for research on anything.

My experiment was a "success", in that it proved what I thought it would, namely that there's not nearly as much verification going on as some proponents would have you believe, and you can trust basically nothing that you read there. If you use Wikipedia, only use it as a starting point, not a destination for research, and keep a whole lot of pinches of salt handy.

By the way, to ease my feelings of guilt for adding the information, I chose several other related articles and corrected incorrect data, reworded poorly written sections, and just generally made them better. Why? Because I actually care about real facts, and, despite all its faults, Wikipedia is what people are using.

2010-02-06
.