I have been teaching the course Topics on the border of CS and economics with Michal Feldman in the fall semester. One of the course assignments was to write a Wikipedia entry on a topic of the students’ choice that is related to the course. This was the first time that any of us tried this, so we left this assignment pretty open and only made sure that the class’s choice was injective and more or less in range. We were not aware at the time that there are suggested ways on how to organize such a thing, e.g. this entry in English or this one in Hebrew (we allowed both English and Hebrew entries, and most students chose the latter.)
We saw various reasons why this assignment is a good idea: First, we figured that having the “world” reading your work is an added incentive to write well. (This is especially so given the fact that, due to the terrible budget cuts in Israeli universities in the last few years, we had no TA.) Second, we wanted to ensure that the effort that students put into their assignment is not wasted but rather put to a good use. Third, there was the idea of doing a public service. Fourth, there was a somewhat vague notion of immersion in the main motivator of the course: what goes on the Internet.
As the course has ended, we compiled a list of the entries written for the course. I would say that the experiment has ended with mixed results: the quality of entries varies. Some of them are really good, others are OK — a reasonable start and hopefully will be improved, but some are quite badly written (in various senses: format, writing style, and even technical content.) It seems that adding incentives for what is usually done by volunteering (i.e. giving course credit for writing an entry that is usually done voluntarily) introduces problems. While normally Wikipedia writers only do so when they want to and are able to do a god job, here we incentivized people to do so even if they could not do a good job or did not want to put enough effort into doing so. (This is somewhat similar to the observation that paying for blood donation reduces the willingness to do so.) I can’t really tell if the students that did a really bad job are incapable of doing significantly better or just didn’t put enough effort into it.
Michal and me now feel somewhat responsible for doing some harm to the (mostly Hebrew language) Wikipedia. We are not really able to go over all the entries ourselves and “fix” them (again, no TA, and many dozens of entries.) We are making small changes here and there as well as adding a few comments in discussion pages, but this is small in magnitude, and doesn’t help much for the worst entries, so we have decided to hire a TA/RA for a while to “clean after this course”.
One of the interesting new entries was the Algorithmic Game Theory entry in the English Wikipedia. It turns out that there was no such entry previously, despite a call to write one by the “committee to improve Wikipedia’s arguably-somewhat-sketchy coverage of theoretical computer science” over a year ago, which mentioned AGT together with a list of other desired entries, many of which still remain unwritten. (Looking at the history page of AGT, it seems that an entry for AGT was previously written, but the entry was not appropriate and focused on Algorithmic Mechanism Design and so was “moved” there about two years ago.)
My (preliminary) conclusion: I would do it again, but next time, only with a smaller class size and with a devoted TA.
Turing's Invisible Hand 
It looks like a good idea. Thanks for sharing with us. May be you should have 2 phase assignment. First what you did, and the second, collect everybody’s entry and ask the class to collaborate and improve each other’s enties. Sharing of knowledge is precisely the idea behind Wikipedia, and to a good extent behind the web too.
May be another class do the second phase?
I had mentioned a similar course on Expanders (Fall 2005) by Manoj Prabhakaran, see http://globofthoughts.blogspot.com/2005/12/wikinotes.html
I have been doing something like this in my Information Retrieval class, except that students are asked to improve an existing entry (sometimes in a small way). And then to report on what they did.
The only downside so far is that one student (in 3 years) has complained that it is free labor and refused to do it. Good for him!
I’m pretty sure it has improved the French wikipedia.
I share the sentiment that this is a good idea. Perhaps a better approach would be to have a local installation of MediaWiki where students could write their articles, and only if/when they were approved would they be placed on Wikipedia. Students could get extra credit for improving entries that did not pass muster by the end of the class.
Have one student as the writer and another as the editor for a particular page. So the editor gets the text from the writer and then attempts to clean it up and improve it before the final posting.
I have seen this done before with lecture notes and it seemed effective.
Each student can also serve as a writer and editor on different pages.
Michal and me now feel somewhat responsible for doing some harm to the (mostly Hebrew language) Wikipedia.
My impression is that Wikipedia (at least the English one) prefers a incomplete entry on a notable topic to no entry. So I doubt that anyone else considers your class to have harmed Wikipedia. After all if some of the articles are so bad as to be useless someone can always delete and start over.
I did something similar in my coding theory course last year (and am continuing the practice this year). However, I had 7 students in the class so I could give them comments on their preliminary entries. (As Adam suggested above, I had a local wiki for students to play with.) Also the students had to get my explicit permission to upload their final entries on Wikipedia.
I agree with your preliminary conclusion that doing it in a smaller class (with some quality control) is the way to go.
[…] may look at my recent blog post about my evaluation of the “wikipedia chore”. In particular, we are looking to hire a […]
[…] Wikipedia course assignments « Algorithmic Game Theory […]
[…] Wikipedia course assignments « Algorithmic Game Theory Share and Enjoy: […]
[…] Wikipedia course assignments An interesting post by Noam Nisan where he talks about his experiment in his course. Basically, he […]
For a stupid student like me … wikipedia is the 1st place to look for … I mean to say I start for understanding wikipedia then go to book… This is really a nice move Sir. Hope other prof around the world should do the same…. I now realize I could have done the same while doing my masters .
I have been doing something like this in my Information Retrieval class, except that students are asked to improve an existing entry (sometimes in a small way). And then to report on what they did.
The only downside so far is that one student (in 3 years) has complained that it is free labor and refused to do it. Good for him!
I’m pretty sure it has improved the French wikipedia.
Basically I llike the idea but perhaps a better solution would be to use MediaWiki locally they could write their articles, and only if/when they were approved would they be placed on Wikipedia.