Interesting question! And coincidentally, one I was just thinking about, while working with a group of Canadian business leaders looking to stimulate innovation...
FWIW, when I hear "crowd source," I usually think of the exchange-type games like the Iowa Electronic Markets (
www.biz.uiowa.edu/iem/index.cfm ) and the Hollywood Exchange (
www.hsx.com/). If you spend some time Googling the guy who's done much of the compelling research in this area, David Pennock(probably an unfortunate turn of phrase, since he appears to be working for Yahoo now), you'll find a list of publications that includes many different references to these sorts of exchange-based simulations:
dpennock.com/publications.html
As theoretically sound as Pennock's research appears to be (I have to stick to "appears," since much of the math is beyond me), the basic problem I've found with the whole collective wisdom idea is that prediction games don't really attract masses of participants. Sure, you can get groups of experts to debate an issue on a specialized site - you see prediction-type threads on the Red Sox fan site Sons of Sam Horn pretty often - but that isn't really what the "wisdom of the crowds" is supposed to represent. And, you can get crowds excited about things like a goofy YouTube video, but that's more about shared experience than wisdom...
One more nit, before I get down from this soapbox - it's hard to find widely-used, well-tested, low cost tools that can be used to build "wisdom of the crowds" apps. In one of Pennock's papers, he mentions NewsFutures (
www.newsfutures.com/www/home.html), but it wasn't obvious to me what their pricing/packaging model is (and my confidence in software always takes a hit when the vendor's
site is buggy!). If you know of any good options along these lines, though, please let me know - I'm a believer in wisdom whatever the source, and I'd like to get a chance to try to apply the Crowds theory!