I don't have the stats on hand, but they show that new pilots, around 0-300 hours have less accidents than pilots with 300-1500, and that 1500+ have less accidents than the 300-1500. These numbers are straight out of thin air, but they were something along those lines. I guess the reasoning is that the 0-300 guys are too scared to do anything, the 300-1500 have just enough confidence to start doing something stupid, and the 1500+ guys know better.
To be fair, it's a bit better than that. Anyone can make those subjective observations, but are the observations accurate? And can you fix the problem? The scientists gathered objective data showing it was true and showing that people can be educated out of it.
You are reffering to the stats in "The killing zone" a great book about fatal aircraft accidents.
In it the author shows that 0-50 hour pilots have one of the lowest fatal accident rates. The absolute worst accident rate is for pilots with 50-500 hours. After 500 hours the accident rate trends gradually downwards.
Take note that this is only for fatal accidents, "fender benders" were not counted at all. I would assume that the 0-50 hour pilots have disproportianlly more of these.