This is why I wish the sample size was bigger, closer to that magic 2000 person populace that gives a relatively random cross section. I just don't think they're going to get the value they need out of this limited experiment. 2000*24000 = 48million... I'm sure they have that kind of scratch laying around.