Many would-be educational innovators treat technology as an end-all and be-all, making no effort to figure out how to integrate it into the classroom. "Computers, in and of themselves, do very little to aid learning," Gavriel Salomon of the University of Haifa and David Perkins of Harvard observed in 1996. Placing them in the classroom "does not automatically inspire teachers to rethink their teaching or students to adopt new modes of learning."

At last week's dog-and-pony show, Duncan bemoaned how the U.S. is being outpaced in educational technology by countries such as South Korea and even Uruguay. ("We have to move from being a laggard to a leader" was his sound bite.)

Does Duncan ever read his own agency's material? In 2009, the Education Department released a study of whether math and reading software helped student achievement in first, fourth, and sixth grades, based on testing in hundreds of classrooms. The study found that the difference in test scores between the software-using classes and the control group was "not statistically different from zero." In sixth-grade math, students who used software got lower test scores — and the effect got significantly worse in the second year of use.