Schoolhouse Rock

The Bonus Lottery

It’s one thing to talk hypothetically about merit pay for teachers. It’s another to actually put a merit-pay system into practice.

Under Gov. Jeb Bush, Florida became one of the first states to experiment with merit pay. In 2006, after years of negotiation, the state created a new compensation system to reward successful teachers. Each district was offered additional state funding if they opted into the plan. Districts were given some flexibility in how to measure student success, but standardized-test scores had to make up at least 60 percent of the formula. In February, a reporter for the St. Petersburg Times found that in one urban district that had opted to take the deal—Hillsborough, which includes Tampa and its suburbs—”merit” checks were going in overwhelming numbers to teachers at affluent schools.

[O]nly three percent of the educators deemed worthy of the $2,100 bonuses worked in the low-income schools that struggle most, where at least nine in 10 students qualify for free or reduced-price lunch. And almost two-thirds taught in A-rated schools, where they arguably were least needed.

That wasn’t how it was supposed to work.


But just because merit pay didn’t work in Florida doesn’t mean it can’t work, period. A more effective merit-pay program wouldn’t reward every successful teacher, it would reward teachers who succeed with those students that are hardest to teach. And it also might define “success” in a more creative way than Florida does. One thing most teachers seem to agree on: they hate being judged on their students’ test scores alone. If you ask current teachers about a deal like the one in Hillsborough, most of them say no thanks. In the teacher-compensation study I mentioned yesterday , researchers surveyed teachers in Washington state and found that 83 percent of them opposed merit pay, including almost 60 percent who “strongly” opposed the idea. But the survey defined merit pay very narrowly, as a system that gave bonuses only for standardized-test gains. If we could come up with a more flexible system, the number of teachers opposing change might start to fall.

I got an e-mail yesterday from a reader, a former New York City teacher, making the case against judging teachers by standardized tests alone. (She asked to remain anonymous, “in case my old principal actually learned how to use the Internet.”)

In the school where I used to work, I would be alternately canonized and demonized each year, as we compared not how each student had progressed each year, but how one eighth grade compared to the grade in front of or behind it. The year our eighth grade went from 400 motivated kids to over 500 “challenging” students, test scores fell, and I was told not to be creative in my classroom, “at least until after the test.”

The next year, the eighth grade shrank to 250 students, and a good group of them, at that. When test scores for that class were higher than the previous eighth grade, I was praised to the heavens as if I had actually had anything to do with the increase. If my performance review and job security were tied to that kind of a measure, I’d be terrified, and would demand a salary equal to two or three years what I actually deserved just to build in security for my inevitable firing.

Accountability nuts get nervous when teachers say they want to be judged by more flexible standards than test scores alone. And for good reason—without some kind of objective measure, it would be way too easy to design an accountability system that could be gamed, where teachers were rewarded based on “parental satisfaction” or some equally vague measure, or where sub-par teachers received bonuses along with high-quality teachers, out of a sense of “fairness.”

Still, the Florida system—bonuses based only on test scores, and no extra compensation for working in high-needs schools—clearly isn’t working.