Here is an article about his on New Scientist:
http://technology.newscientist.com/article/mg19626335.500
And here is Tom Griffith’s page, the guy that came up with this conclusion. His paper on it can be found below:
http://cocosci.berkeley.edu/tom/papers/retrieval.pdf
The paper also explains in detail how the PageRank algrithm is “supposed” to work (no one knows except Google).
