A few days ago, I was asked by a young lady of my acquaintance, who is a high school student, if I had any books that might be relevant to a paper she was assigned to write, about aspects of the Holocaust. Unfortunately, I could not think of any that were related to the specific area in which she was interested, and some other people present tried to come up with some suggestions. One asked the question, “Have you looked at Wikipedia?” She replied that her school did not allow students to use Wikipedia as a reference, apparently on the grounds that “it can edited by anyone.” And I was given to wonder a bit about whether this rule was a good idea.
Now it is true that Wikipedia is a prime example of an open project. It does not impose any requirements, in terms of background or credentials, on those who wish to contribute. (Incidentally, there is an “About Wikipedia” page that gives a fairly thorough introduction to the process.) It is certainly possible for someone to introduce incorrect or malicious information into a Wikipedia article, but it is also possible for anyone to challenge an entry that (s)he believes to be wrong. Those of you who have read this blog from time to time probably have noticed that I often provide links to Wikipedia articles in order to provide background information on specific topics. I do read those articles before linking to them, and on the whole find that they are reasonably accurate and complete. Admittedly, these tend to be articles on relatively technical subjects, ones in which Wikipedia’s technically-oriented contributors are likely to be knowledgeable. For what it’s worth, I have also looked at articles in science and economics, and have found them to be of good quality as well.
A few years ago, a comparison was made, by the staff of Nature, of the relative quality of a set of science articles from Wikipedia and from the Encyclopedia Britannica. There was a certain amount of argument about the interpretation of the results (some of which is summarized in an article at Wired), but I think the main important message was that, contrary to what some people expected, there was no huge gulf in quality between the two sources. Prof. Ed Felten of Princeton (who I have mentioned here recently in another context) also conducted a small experiment of his own, which he reported in two posts at the Freedom to Tinker blog. His conclusion was similar: that the quality of both sources was good, but that there was more variance in the quality of Wikipedia’s entries. My own suspicion is that Wikipedia is most likely to be suspect on matters where there is current controversy (although errors or biases tend to be challenged quickly).
However, I think there is enough evidence to indicate that Wikipedia can be a credible information source; certainly there are sources, both on-line and in print, that are demonstrably less reliable. In thinking about the blanket ban on Wikipedia as a source, though, my main conclusion is that a really valuable educational opportunity is being thrown away: the opportunity to teach students how to evaluate the credibility of a source. This is not a new issue; we had to learn this back in the days when the Internet didn’t exist. It is an important skill that is of enormous value in real life, and students will not learn it just because some sources are arbitrarily marked “unsatisfactory”.
Update Wednesday, 29 December 2010, 10:23 EST
According to an article in the Washington Post this morning, there are a number of history textbooks used in Virginia that qualify as printed sources that are demonstrably less reliable than Wikipedia. Apparently, the main focus of the state’s textbook review process is ensuring that the books cover specific topics from the state “Standards of Learning”:
The department is required to find texts that meet the state’s stringent Standards of Learning, which includes lists of themes that each textbook must cover. That disqualifies many books produced for the national textbook market.
In other words, the books have to teach the material that will be covered in a particular set of standardized tests. It seems that mere factual accuracy is not such a concern.