This has been a big topic here at Plymouth State University since I was a student, and after a good discussion with Donald I have written up how I evaluate the credibility of web resources.
Consider whether the information makes sense. Chances are you’re reading about a topic you know a bit about. Does this information fit with what you have experienced? Of course what you know might be wrong, but at least you know it from your own experience. If you don’t know anything about the topic you’re researching, stop now. Go find someone who does and have them point you at some good resources.
If you’re looking for factual information you can probably find it on several sites. Do most sites seem to agree on this topic? If so that tends to lend to the credibility. If not you may be on a very opinion-driven topic which has no right or wrong answer.
What is the reputation of the site in question from other, possibly more reputable sites? For instance, Cliff wrote some instructions on Solaris partitioning which were linked by Sun support shortly after that. A link like that gives a high level of confidence in the content presented.
Anonymous and virtually anonymous resources like message boards, wikis, and newsgroups should not be assumed as correct. Misinformation, whether intentional or not, is common in these resources and, while Wikipedia is a great place to look up episodes of your favorite TV show it should not be used as a primary resource. These sites should be considered supporting resources at best.
These rules have not steered me wrong yet. Everyone has their own methods and needs to figure out their own level of risk, but this has suited me well.