I really wish Keith Ng would write more often. I like his forthwright, easy to read musings on what passes for economics or at least economic spin in this country. He's usually quite good at pointing out how badly people are using statistics in the pursuit of supporting a pre-determined point of view. The idea that you can use statistics to prove anything is a load of bollocks. The notion that you can use statistics badly to support any idea you want has a lot more going for it.
In a similar vein, best post title I've seen this month: Bill English and metaphysico-theologo-cosmolonigology
Classes in statistics and logic forMr English methinks. And everyone who takes what he says at face value.