Take as an example the study predicting diabetes. Nobody does a study that just looks at one particular census; you're always tracking the changes over five years. So you'd be looking at the determinants of health outcomes not just in 2006 but also in 1991, 1996, 2001, and 2006, and looking at the changes in lifestyles and then predicting from that.
Even if the voluntary survey worked out fairly well, we'll never be able to establish a bridge from 2006 through 2011. As I said, in my view it will take three or four cycles of a voluntary survey before we're able to make use of it. So we're going to have this continuous, pretty solid data through 2006, and then I think we're going to have a fog for a number of periods.
It's kind of a shame for the people who did take the trouble to fill out surveys in 2001, 2006, and other years. We won't get the full benefit of theirs, because it's going to hit a dead end.