Looking at the study I brought up as an example, the incredible thing is that the 700 potential subjects are not people off the street, but those who have already self-identified as meeting the basic parameters for the study.

Another intersting point, slightly OT, about the weight loss diet study cited above, is that it included a similar run in period where subjects were all brought to a predetermined baseline. Of course this subject is very closely related to the T2 diabetes problem as more enlightened research begins to prove that insulin dynamics are the key to weight gain. I did not realize, for instance, that while insulin therapy is generally associated with weight gain, medications like Metformin cause weight loss.