The definition that was used here was the one the Department of Justice used when it made its briefing to you earlier, which was looking at that point at one year. Recidivism can be one year or ten years or it could be a lifetime. The point is that when you're trying to do these studies, you need a short enough period of time to know whether you're getting better or worse.
One-year recidivism is only good in comparing conditional sentencing to prison. What we find is that with prison, after one year there's a 30% recidivism, and with conditional sentencing there's 17%. So that's a dramatic improvement, and it's indicative that something worthwhile is happening.
It's not to say that after that year there's no recidivism, but we can say that recidivism drops the longer a person is out. So even from federal institutions, after two years in the community the recidivist rate drops quite substantially.
So mechanisms that keep people engaged in the community and allow them to have jobs, work, and so on have greater potential over the longer term, because they are not going through that isolation. The longer you can keep people supervised and productive in the community, the better your chances that they will not recidivate over the longer term.