Thursday, 29 July 2004

Quantify everything

Tyler Cowen calls on Dan Drezner to self-assign p values to his fence-sitting. My gut feeling is that this approach would be ineffective; based on the cognitive psychology literature, I’d have to conclude that Dan is probably not the best judge of his own objective probabilities. Instead, I recommend employing content analysis of Dan’s posts to arrive at estimated p values at given points in time, or using a panel of raters, or some other more accurate technique.

Incidentally, the only p value I have a good handle on for myself is that p=1.00 that, on election night, I will be sitting with my undergraduate methods class at some venue with available libations making fun of Brokaw, Rather, and/or Jennings on the big-screen as they call (and uncall) states. Assuming the dean doesn’t put the kibosh on the short field trip, that is…