More from one of our favorite econ guys: Dan Goldhaber crunched eight years of Denver schools data to find out if that city's pioneer performance pay project known as ProComp has paid off. And unlike some other recent studies on teacher pay, the findings suggest it does.

While ProComp allows teachers to earn their bonuses a multitude of ways, it was the two choices most closely related to improving test scores that did in fact improve test scores. One gave bonuses to teachers whose students outperformed similar students on state tests, and the other provided bonuses to teachers who met their own self-designed goal for student growth.

Other softer or less objective goals, such as rewarding teachers for taking professional development or being rated well on the classroom observation, didn't make a dent.

We still can't conclude the lack of effect from teachers getting bonuses for these goals means that they should be eliminated from the menu of options. The variety of bonuses embedded within ProComp's reward structure may be part of its success. By providing teachers with several avenues to earn rewards, it's possible to compensate teachers for the more intangible and nuanced outcomes of teaching.