Douglas N. Arnold and Kristine K. Fowler have self-archived "Nefarious Numbers" in arXiv.org.
Here's an excerpt:
Despite numerous flaws, the impact factor has been widely used as a measure of quality for journals, and even for papers and authors. This has created a strong incentive to manipulate it. As we have demonstrated, it is possible to vastly increase impact factor without increasing journal quality at all. The actions of a few interested individuals can make a huge difference, yet require considerable digging to reveal. The cumulative result is that impact factor gives a very inaccurate view of journal quality. We primarily discussed one extreme example, but there is little reason to doubt that such techniques are being used to a lesser degree by many journals.
