The quality of scientific articles on the English Wikipedia

From Devwiki

Bill Wedemeyer

We have been evaluating the quality of Wikipedia's scientific articles in several respects, such as accuracy and completeness of facts, quality of writing and illustration, depth and breadth of coverage. Drawing upon nearly 100 independent peer reviews by tenured faculty at research institutions, we find that the Featured Articles, Good Articles and A-level science articles are reliably good, and generally better than their counterparts in the Encyclopedia Britannica. In contrast, Wikipedia's Start and Stub articles are reliably bad, with B-level articles forming a twilight class. Taken as a whole, Wikipedia seems excellent for scientific outreach and background information, but not yet adequate as a textbook. Some possibilities for improvement will be discussed.

Although Wikipedia is peerless for its quantity, Jimmy Wales and other knowledgeable commentators have noted that its next great challenge is quality: in writing, illustration and factual accuracy. In particular, Wikipedia faces the challenge of assuring quality, giving the reader a sense of how much an article can be trusted. Although Wikipedia has long been critiqued as unreliable and poorly written, there is surprisingly little scientific data to support this conclusion. Here, we have analyzed a representative cross-section of Wikipedia's scientific articles, subjecting them to independent peer-reviews by tenured faculty at research institutions, and found certain classes of them to be excellent.