Last week, state schools Superintendent James Guthrie told a gathering of Nevada school boards that he’s in favor of a slate of reforms, including merit pay for teachers.
Aren’t there studies that contradict the idea that merit pay results in increased student achievement?
The weight of evidence, both from the U.S. and overseas, is that performance pay is related to elevated student achievement. We do have one study done at Vanderbilt three years that found no visible influence on student achievement. I happen to have been the director of that experiment, so I know it very well, and I know how naively it was conducted. But in the subsequent time there have been other studies, particularly from overseas, which display the power of performance pay.
How does it work?
Teachers are just motivated by the idea of gaining more money or not losing a bonus they’ve already been given? Well, that’s a good question. The actual dynamics — that is, what goes on inside the black box that is a teacher’s brain — we don’t know too much. What we do know is that it has a greater effect not so much on those who are teaching now, but on framing the attractiveness of teaching as an occupation. In economists’ terms, it affects the selection influence. The opening up of added remuneration is enticing to those who otherwise wouldn’t consider the occupation.
How do you prevent a teacher from cooking the books?
There shouldn’t be just one measure that determines a teacher’s evaluation. There should be multiple measures. One portion of which, I would say the largest portion, should be the achievements of students in their classes. But there should be other components, too, like the principal’s evaluation and peer evaluations. Some of it ought to be on process, not just student outcomes. So if we had a multiple composite index, it would be harder to cook the books, to use your term. But teachers are not in charge of testing, that’s done independent of them. Certainly there’s cheating that goes on, I’m well aware of it, and we’re going to discourage it in every reasonable way. Cheating is not right and teachers shouldn’t do it, and we will bring the full force of law against any teacher or principal caught doing it. I’m very favorably inclined toward doing this, but I don’t want to do it fast or recklessly. There’s no state that has succeeded in doing this. I want Nevada to be the first, but I want to do it through careful engagement with teachers, with principals, and ensuring that the technical underpinnings are solid.
Do you think there’d be a lot of instances of teachers teaching to the test?
Teachers should teach to the test. That’s exactly what they should be doing. What’s bad is if they teach the test. That they shouldn’t do. But if the test is appropriate, if it captures what we want students to learn, know and be able to do, then they should teach to the test.
Is it possible to design a merit-pay system that fairly judges teachers who work very different neighborhoods and circumstances?
That’s a classic concern. But our capacity for controlling for the characteristics of students is getting better and better. So mathematically we can hold constant the characteristics of the students with whom the teacher deals and then find out what value she adds to what they know. The methods we’re working on in Nevada will be state-of-the-art in this regard. All similarly situated teachers will be treated the same, and they’ll be treated fairly.