Replies

I'm sure some people will see this as simply "sour grapes" by someone who just wants to denigrate Bush, but I think the reality is that the grunt work of producing these numbers is being done by career bureaucrats, many of them the same bureaucrats that were there way before Bush. When you tell them "We want to see a 10% reduction in these numbers, or heads are going to roll.", then they'll find some way to produce that 10% reduction. It's not so much an issue of Bush or Walters as it is of the way things work inside the beltway.

Looking at their sampling methodology, I don't think a 10% fluctuation due to sampling variance is out of reason, nor do I think it unreasonable to question wheather the "random selections" were entirely random.

There is always the possibility of random error, and perhaps also sampling bias. They are not specific about how they chose their geographical areas, for example. So I think this is a legitimate criticism.

Assuming this is a geniunely random sample, however, it is a good bet, probability-wise, to conclude that there is a significant difference in this newest cohort, compared to past cohorts. I'd be willing to reject the null hypothesis here, but maybe you are looking for more stringent criteria, which is fine too. I can appreciate that.

I appreciate the skepticism, which is the appropriate attitude toward any study such as this. When skepticism crosses over into cynicism, however, I grow a bit concerned.