First of all, you have to understand that this is a very conservative number. It's based on large numbers of treatment programs, some of which are effective and some of which are not, dealing with some people who are very dangerous and high-risk and some of whom are a lower risk. So as for what you're looking at, the way you measure it is that you look at a matched sample of people who didn't get treatment, and you take a look at the recidivism rate for them over 5, 10, or 15 years, and then you compare that with the recidivism rate for people who went through treatment.
The large study I mentioned that was done as part of the collaborative database found that generally you got a reduction from I think 17% to 9.9%. That's a significant decrease.