It makes sense to take a lot of these numbers with a healthy amount of skepticism. That second 'A' means something. AA has no membership, no census, no graduat

It makes sense to take a lot of these numbers with a healthy amount of skepticism. That second 'A' means something. AA has no membership, no census, no graduation. Additionally, I doubt the people who stay involved are a random sample, so identifiable 12-steppers are going to be a skewed distribution. I'm sure at least some of that can be controlled, but claims about their efficacy have been all over the map, I think, because of this.