Even if they won't admit it, students are using Wikipedia to kick off their research and fill the gaps in their class notes ... right now. It might not show up in the bibliography, but the free, open source online resource has long since become the starting point for settling factual disputes, brainstorming paper ideas and even offering suggestions for further reading.
If that's an open secret, then so is this: For all the hand-wringing over whether Wikipedia is a legitimate source for completing college assignments, some professors are quietly incorporating it into their classrooms and even their research. Others, noting features of the Web site that contribute to inaccuracies and shortchange the value of expertise, are building variations on the model that are more amenable to academics and to peer review.
A good number of college professors "work on Wikipedia pretty regularly, and of course their work is one of the main reasons why Wikipedia is as good as it is," said Larry Sanger, one of the founders of the site who has since become a critic of what he refers to as its "worst elements."
A quick glance at the syllabus for Breno de Medeiros's Advanced Topics in Cryptography and Network Security course at Florida State University, to take one example, reveals reading assignments -- in addition to the usual textbook chapters and published papers -- that direct students to pages on Wikipedia. Introduction to complexity theory? See the page on P and NP classes. Brush up on probability theory? See Wikipedia's entry on the Chernoff bound. Far from the amateurish, typo-ridden entries some have come to expect, the articles are straightforward and include definitions, illustrations and explanations that at least match similar content from comparable textbooks.
"Information on computer science subjects in Wikipedia is likely to be accurate and informative, often using unique resources to illustrate concepts that are not available to print media," wrote de Medeiros in an e-mail. "This probably derives from the fact that computer scientists use the computer as their main form of access to scientific articles and journals, that they take advantage of electronic forms to disseminate their research, including instructional materials in various formats. Researchers and educators of high caliber are probably behind most Wikipedia articles in computer science."
In all likelihood, tech-savvy scholars are among those keeping such isolated corners in the digital stacks of Wikipedia relevant, up to date and accurate. For computer science, especially, many topics on Wikipedia are in a form polished and accessible enough to assign to students as reading, and the subjects aren't controversial in a way that would inspire the sort of back-and-forth citation wars that cause some articles to fluctuate wildly between competing versions. But other topics get assigned from Wikipedia as well -- not least in courses about digital culture itself.
"I use Wikipedia a lot for my own research and for course preparation. Often, to the extent that [Wikipedia articles] appear on my syllabi it’s to give students a quick overview of a subject or concept when I’m looking less for a theoretical or critical perspective and more for this kind of open-source knowledge, or kind of 'crowd-sourced' perspective," said Mark Tribe, a professor of modern culture and media studies at Brown University.
Tribe's course outlines are posted on wiki-like pages, and he uses the format in some assignments for students. Some professors have even experimented with assigning students to create their own Wikipedia pages. Still, some continue to worry that the very structure of Wikipedia encourages editors (who can be anyone) to disregard expertise and undermine the basic mechanics of peer review and academic credibility.
In other words, what happens to articles once they're posted? Will they be watered down or made inaccurate by someone with no relevant credentials? Wikipedians would argue that credentials are besides the point -- that anyone with a computer can police the encyclopedia by judging source material, sifting through edits and using a neutral tone to describe disputes. It's a dynamic that Sorin Matei, a communications professor at Purdue University, describes this way: "He who can sit for the longest in front of the computer is right."
Wikipedia "speaks with a million voices, and these voices kind of jump on each other," he said. In this view, the site's openness -- the ability of everyone to participate, without having to identify themselves by name -- leads to an erosion of accountability and, often, an increasingly shrill cacophony. Matei is academic content coordinator of the Eduzendium initiative, an attempt to bridge the wiki format and academe. Entire classes can join up so that professors can coordinate the production of certain articles with students -- either individually or as a group assignment. It's up to the instructors how exactly to divide the labor, but the idea is that more accurate content will result from putting one's real name (and, possibly, grade) on the line and from the checks and balances set up to minimize the kind of interference seen on Wikipedia pages.
The project is only a few months old, part of the larger Citizendium enterprise initially set up by Sanger as a "fork" of Wikipedia. One of his main criticisms of that site, which he co-founded and helped conceive and develop in its early years, was what he characterized as users' general skepticism of the value of expertise in a given field. "I think that it would be possible for people to, if they were properly motivated, to make a lot of good solid changes to Wikipedia articles. My question is whether they would think that it’s really worth the effort ...," he said, to "do battle" with more energetic editors who perhaps don't have the same expertise.
"A lot of people might come to Wikipedia with the false idea that well of course, my credentials are going to count for something," continued Sanger, who holds a Ph.D. in philosophy and taught at Ohio State University. "For some sensible Wikipedians ... that’s true. For a lot of people, and in particular the people who cause the problems, basically, it doesn’t matter at all what sort of training you might have had. They don’t care, and in fact, I’ve heard from many people who come back and say, well I was told that because I have published a lot in this area I have a conflict of interest."
Citizendium, with over 6,100 articles and counting, tries to attack the problem by enabling experts to essentially "freeze," or approve, articles when they have reached a certain level of completion. Anyone can still participate in creating content -- but their work is "gently guided by experts," as the site puts it. Dubbing itself a "republic of letters," Citizendium is described in terms borrowed from political philosophy. Sanger refers to "good governance," and notes that "it’s a social contract. It’s called the Citizens’ Compendium and not Expertpedia for a reason. What we are emphasizing is reliability and responsibility both in terms of the quality of information and also the way that it’s created."
Contributors use their real names, and experts have their work evaluated and revised by others with similar qualifications, rather than just an "anonymous crowd," an arrangement that Sanger said was paramount in his discussions with academics in attracting them to devote time and resources to the project. On Wikipedia, "the peer review is done by self-selection or a process of selection that is totally opaque, in my view," Matei said.
Building on that specific complaint, another variation of the model is trying to make peer review work in a wiki environment. On Scholarpedia, articles are written by "experts," who are elected by users or selected by other experts. If an article is peer-reviewed -- a process that is anonymous and restricted to other designated experts, per scholarly convention -- its author automatically becomes a "curator" for that article and presumably for related subjects as well.
"Our goal is not to create ... content very fast. We want to get the best people (living legends) to participate in the project, so that the content will survive and will be maintainable by the future generation of experts 100 years from now. This is why we are so selective," said Eugene M. Izhikevich, a senior fellow in theoretical neurobiology at the Neurosciences Institute in San Diego, Calif., and Scholarpedia's editor-in-chief, in an e-mail.
Izhikevich said his site offers the necessary incentives for academics to be "willing to invest their time" in the project, most notably acknowledgment (by name), full editorial control and peer review processes that allow scholars to cite their Scholarpedia work in their CVs, and in applications for tenure or grants. He added that Scholarpedia has its own International Standard Serial Number from the Library of Congress, and he has cited his own works on the site in peer-reviewed, published journal articles. "Scholarpedia is under consideration to be included [in] citation databases, such as MedLine, PubMed and ISI Web of Science," he noted. (The site hosts encyclopedias in several subfields, such as astrophysics and computational intelligence, and there are plans to publish those in print.)
"The only difference between the peer-review process in Scholarpedia and in Science, Nature, [ Proceedings of the National Academy of Sciences] and many other journals, is that the reviews are open to the public and the reviewers may choose to (but do not have to) reveal their name," Izhikevich continued. "I would love to read the review of Einstein's 'Space-time' entry in Britannica. I would love to know what was the state of mind and the prevailing dogma at that time. Unfortunately, it is not available, like most reviews of the other peer-reviewed journals. Scholarpedia maintains the history of its reviews and all article versions for the future generations to read and analyze."
It's unclear whether such sites will gain popularity in the academic mainstream, but there's no doubt that interest is increasing. Sanger said he's noticed a significant uptick in editing activity recently on Citizendium. "I don’t know that we will in the long run have the same number of articles as Wikipedia, but I can say that we are definitely on track to create large numbers of very, very good articles. I see every reason to think that we are going to continue to grow ... throughout this year and the next year and so forth," he said.
A sign that people might become more receptive to a modified form of Wikipedia is that Google is getting into the mix with a tool of its own, dubbed knol, that won't be restricted to experts but will offer name recognition (and presumably, other benefits that come with a higher page rank) to those who have credentials.
In a blog post, the company elaborated: "The key idea behind the knol project is to highlight authors. Books have authors' names right on the cover, news articles have bylines, scientific articles always have authors -- but somehow the web evolved without a strong standard to keep authors names highlighted. We believe that knowing who wrote what will significantly help users make better use of web content."