If we didn’t already know not to trust figures quoted by Paul Krugman, this would be an object lesson. Iowahawk breaks from his usual format and explains how Krugman’s latest attack on Texas education is full of crap. Alas, my beloved Economist is guilty as well.
Iowahawk debunks figures showing that Wisconsin education is superior to Texas education (Krugman uses dropout rates, the Economist uses standardized test scores), showing that they fail to control for a very important variable:
As a son of Iowa, I’m no stranger to bragging about my home state’s ranking on various standardized test. Like Wisconsin we Iowans usually rank near the top of the heap on average ACT/SAT scores. We are usually joined there by Minnesota, Nebraska, and the various Dakotas; Vermont, Maine, New Hampshire…
… beginning to see a pattern? Perhaps because a state’s “average ACT/SAT” is, for all intents and purposes, a proxy for the percent of white people who live there. In fact, the lion’s share of state-to-state variance in test scores is accounted for by differences in ethnic composition.
If you control for ethnicity in standardized tests, Iowahawk shows that Texas beats Wisconsin in 17 out of 18 categories. (Wisconsin won the last category by a statistically insignificant margin.)
If you control for ethnicity in dropout rates, Wisconsin does slightly better than Texas for white and hispanic students, but Wisconsin does much worse than Texas for black students. Texas beat the national average in all three categories; Wisconsin beat the national average in two, but did much worse for black students.
Paul Krugman is an economist. The Economist is presumably staffed by economists. As any economist knows, you need to control for hidden variables when analyzing data. Neither of them did that. Is that merely malpractice? Or dishonesty?
UPDATE: Iowahawk takes a few questions from people who don’t understand statistics.