Figure 1.
In 2008 there were 4.1 new highly similar pairs of manuscripts per 1,000 published papers in Medline and deposited in the Déjà vu database. This is a major decline that has taken place in the last 2 years. One could speculate on a number of reasons, including fear of detection by would-be perpetrators, but whatever the cause, the problem is getting better but it is still significant in size.