There was recent comment that it is troubling McCain is an admitted computer ignoramus, while Obama uses the internet to look up sports scores. I imagine very few 72 year-old executives could even find boobies on the internet if you didn't let them click on their spam (he called Google "The Google").
But if I had a judge who either knew nothing about statistics, versus one who took one course in college and has used Excel's "linest()" function, I'm not sure this limited knowledge would be better than none. Often, a little education is worse than none at all, because, like Taleb thinking Gaussian distributions are presumed perfect by people who use them, many people learn just enough to make them dangerous. The most common statistical vice is judging models primarily by their R2's, and those are people who know a little, but only a little, about econometrics. Or people who apply a train-test-validate approach, repeatedly, find the highest 'validation' result and think it was all 'out of sample.' MBAs learn that beta is a measure of risk, and returns are a linear function of beta. They are sheep to be shorn. I think I didn't really understand statistics until I TA'd for the course three times.
Someone totally ignorant will have no overconfidence in his ability judge matters. Ideally, a decider knows a lot. But choosing between a decider who know's nothing and a little on a subject? I think I'd take the complete ignoramus, he would be more cautious.
No comments:
Post a Comment