Some of my regular readers will probably note that I frequently link over to Wikipedia entries on topics I’m discussing. I do this not so much because I believe Wikipedia to be a fantastic reference source of unimpeachable truth, but rather that it is convenient and thorough. But, in this week’s journal Nature, Jim Giles unveils findings of a randomized survey that suggests that for science topics, Wikipedia is only slightly less accurate than the encyclopedia Brittanica (link may not be free access).
The author and the journal’s editors point out that participation in creating the information on Wikipedia by scientists will only lead to improvements in the quality of the entries. This is a fantastic point, and again I applaud Nature for being on top of the changes occurring in information dissemination because of the Web.
Wikipedia is remarkable in that it nearly perfectly satisfies its function. It is a central resource that is so totally complete that its visitors become sucked in by their constantly successful searches. Those with expertise find no problems adding content, though there are certainly turf battles, and controversial issues are not its strong point. That said, by allowing controversy to exist in Wikipedia entries, disagreements become more understandable to those outside, thus Wikipedia may be informative in that sense as well.
I don’t know if Wikipedia is going to change the world, but it is an amazingly effective tool that is a great first-stop resource for information. Its entries should be taken with a grain of salt, though perhaps on scientific and computing issues, those grains may not be as large as we all once thought.
Just keep in mind that Nature was generous to Wikipedia since it only compared scientific articles. Wikipedia is good for things that are generally unquestionable and provable, namely math, science, and especially computer technology since a hugely disproportionate amount of users and supporters of wikipedia fall into that category. Thus the entries on Klingon, C++, Ruby on Rails, and Vector Spaces are likely to be very accurate.
However, in anything subjective, especially history, news, politics, literature, religion, etc, you are likely to find something awful. That’s where my beef usually lies.
Also, one more thing to consider, that the Nature article did not consider quality of writing, readibility, or extraneous information. I would assume that professionals working for Britannica with a single supervisory editor would be more likely to produce well-written articles than the participatory method.
There’s a site called WikiWatch, one of many I am sure, that has some interesting notes and articles about various problems with wikipedia. According to his header, the author doesn’t go out of his way to find bad articles, he just has to deal with them because google too frequently spits out a wikipedia link as one of the highest references. The author concludes that wikipedia’s problems aren’t math and science, but subjective information, etc.
http://users.erols.com/mwhite28/wikiwoo.htm
Tom,
Thanks for the the link. Yeah, you and I understand that the subjective topics in Wikipedia are probably not to be trusted, but I’m not sure that the average Googler does. That message should be made more clear in the publicity of Wikipedia, and in discussions of its value. Where there is no objectivity to determine who is correct in a dispute over an enntry, there can be no single entry.
Wikipedia could harness this fact and “teach the controversy” as I mentioned in the original post. By putting this debate up on the front page in some cases, it may be more enlightening to the average viewer.