Wikipedia may be phenomenally popular, a testament to the Power of Many, the ultimate manifestation of the online hive mind, yadda yadda. But just how credible is it? The Guardian UK asked a handful of experts to review and rate the Wikipedia entries on their specialty topics. The results are not exactly glowing, with most entries scoring a 6 or 7 out of 10 for accuracy and completeness. It seems that topics of broad popular interest (Bob Dylan) make out with higher marks, while more obscure topics (Samuel Pepys) score lower. Which seems to validate the idea that the ability of a Wiki to extract collective intelligence from the masses is best leveraged when the number of writer/editors is high.
The trouble with this rating system is that each judge judges just one entry and has different personal criteria for credibility. I’d like to see a test like this extended to a thousand or so entries/judges to get a better sample size, then see a correspondence map between traffic to an entry and its judged rating.
Thanks Paul
Better yet, let’s set up a wiki where anyone can evaluate Wikipedia entries and revise the evaluations!