Tag Archives: statistics

Lies, Damn Lies, and Math

While stormducken Sandy churned offshore, a good portion of the national elections punditocracy proved once and for all that the main reason most of them didn’t go to medical school is that they couldn’t handle the math. Elections statistician Nate Silver’s FiveThirtyEight blog weathered some bizarre attacks, based mostly in utter innumeracy.

As Paul Krugman and others point out, it is one thing to have a reasoned discussion about Silver’s elections model and its embedded assumptions. It’s entirely another to completely misunderstand the nature of statistics and confidence levels. A third baseman hitting .250 is useful information in predicting a likelihood of his getting a hit at his next at-bat. It is not particularly useful information in describing the kinematics of the ball coming off of the bat.  The probability that President Obama will be re-elected is not automatically 50% simply because the only other possibility is that he won’t.

What Silver is doing isn’t really rocket science, but it does have a mathematical foundation. However, as Krugman notes, “On the right, apparently, there is no such thing as an objective calculation. Everything must have a political motive.”



So, I look forward to the conservative alternative to The Weather Channel.