Rock star economists Roland Fryer and Steven Levitt (along with John List and Sally Sadoff) recently published this NBER working paper on an intriguing experiment in teacher merit pay in Chicago. Rather than using the traditional methods of paying teachers increased amounts for higher performance at the end of the year, they decided to give teachers a bonus at the beginning of the year and then take some of it away if their performance was below average. Yes, you did read that right.

The results of traditional, end-of-the-year teacher merit pay programs have been less than stellar in the past. However, when money was taken away from a $4,000 bonus given at the beginning of the year ($80 per percentage point the class mean fell below the average for all students), the results were huge. Results on math tests for the treatment group teachers ranged from .2 to .4 standard deviations, the equivalent of increasing teacher quality by an entire standard deviation.

As anyone who follows the education debate could probably guess, Diane Ravitch had a conniption fit at the results. She said that this research was “Loathsome. Inhumane. Unethical. Antithetical to the values of a democratic society. Antithetical to decency,” and asked “Have these economists no shame?”

I’m not going to wade into the ocean of irony of an educator being against research farther than to say that this was a program that was agreed upon between the researchers and teachers (as well as the union that represents them) and was entirely voluntary; it’s not like these folks were experimenting on prisoners. And if she thinks the idea of loss aversion, one of the most documented phenomena in the behavioral sciences, is “beyond disgusting,” she has more of a beef with millions of years of evolutionary biology than with the guy who wrote Freakonomics.

Now look, I’m not going to say that this is the type of program that we want to take to scale. It’s a bit crude for my taste, and God help the principal that would have to track down money from underperforming teachers at the end of the year, but it is a really interesting study and should spark some interesting debate.

Far from “loathsome,” this is exactly the type of research that we should be conducting, taking widely understood phenomena from other disciplines and applying them to education to see how we can increase student achievement. They won’t all be winners, but should do more good than harm over time.