The often ingenious Jay Greene has been ingenious again. Greene chairs the Department of Education Reform at the University of Arkansas, and with his co‐​author Josh McGee has come up with a way of ranking every school district in the United States against the international average. The idea is simple: 1) find out how each district performs in mathematics compared to the other districts in its state, 2) find out how that state compares to the U.S. national average, 3) find out how the U.S. compares to the international average, and 4) do a few straightforward statistical manipulations to make each of those findings comparable to the others, and then add them together. Rinse and repeat with reading scores.


The results, of course, vary from one district to another. The single most elite district I could think of in Washington state, Mercer Island, manages an international percentile ranking in the low 80’s—well above the international average (by definition = 50) but still a good ways away from the top of the international heap. Many other elite districts do dramatically worse. How about yours?


The one caveat I’d offer for this ranking is that it uses the PISA test to compare the performance of nations, and there are good reasons to prefer the TIMSS test as a measure of academic performance. Results on the two tests are highly correlated overall—as is the case for virtually all academic tests—but their results can differ quite substantially when a nation’s curriculum is aligned with one and not the other (Finland is a case in point, ranking #1 on TIMSS, but having ranked between 10th and 14th the last time it decided to participate in TIMSS, a decade ago).


I’d love to see Jay and Josh plug the TIMSS results into their model and provide the alternate rankings it would generate. Any chance of that, guys?