According to this well-written Business Week article, Nature found that, out of a random sample of 42 entries of a scientific bent, Wikipedia had, on average, just 1 more error per entry than the much-vaunted Encyclopædia Britannica.
What's more, they also polled more than 1,000 scientists who had contributed to Nature, and found that 17% consulted Wikipedia on a weekly basis. And we should all know, by now, that the more smart people that use Wikipedia, the better — because they can FIX MISTAKES that they find there.1
By the way, you can check out Nature's own report of their findings here.
I'll repeat myself one more time, just for good measure: viva Wikipedia!
1 Yes, in fact, I did read the Nature article, and I saw that part where they said that while 17% of those scientists use Wikipedia on a weekly basis, only 10% of them contribute to Wikipedia while they're there. What do I have to say about that? Two things. First, it's better than nothing. Second, shame on the rest of them.