Friday, February 24, 2006

Why I read Brad DeLong

Yes, he's a talented and smart economist from whom I learn stuff. But he smacks stupidity down very nicely:
The excellent Jonathan Chait makes the mistake of actually *believing* Newt Gingrich:
The Plank: Gingrich expounds at length upon his ideas in a recent interview with Brian Carney of The Wall Street Journal editorial page. At one point Gingrich offers up what Carney describes as a "telling anecdote":
We just got a report yesterday that was in the newspaper that children on Medicaid are six times as likely to be obese as children who are not on Medicaid. Now even for a liberal that should give them some pause as to how bad Medicaid is as a system.
Actually, this is a statistic, not an anecdote...
Actually, Jonathan, it's neither a statistic nor an anecdote. 24% of kids are obese. 34% of kids on Medicaid are obese. 21% of kids not on Medicaid are obese. It's not 6 to 1--it's 3 to 2.
And correlation isn't causation, so it doesn't necessarily mean that Medicaid is at fault. Perhaps obese kids are more likely to need Medicaid's help?

Or perhaps, as DeLong notes, since poorer children are more obese on average, and poorer children are the ones one Medicaid, we ought to ponder how bad poverty is, and what we can do to alleviate its effects.

This is one of those cases where simple math education isn't enough, but it's a necessary part of the intuition that Gingrich's statement is bogus (regardless of whether the number are right). The Journal should have known better, as should Chait, I guess. Understanding numbers is tricky, and it's easy to lie with them if you can't do the manipulations for yourself. It's nice to have Brad DeLong to catch you, but it's even nicer to know what he's going to say before he says it.