Developing Better Teacher Evaluation Systems

With many states passing merit-pay laws, finding a good method to evaluate educators is imperative. A new report from The Brookings Institution a Washington, D.C.-based think tank, says a way to accurately judge student progress, or a teacher's "value added," needs to be developed, and these systems need to be uniform among school districts. Many school districts that evaluate teachers currently put a 50 percent weight on principal or administrator evaluations of teachers, and the remaining 50 percent on teacher "value added," typically measured through student test score improvement.

I talked with Grover Whitehurst, director of Brookings' Brown Center on Education Policy, about states' decisions to revamp teacher pay and the future of teacher evaluation systems.

With many states moving toward merit-pay systems, is there a sense of urgency to develop better teacher evaluation systems?

You've got numerous states moving to require that school districts evaluate teachers. A significant portion of that evaluation is value added. States want to tie that to pay-for-performance, but we have 1,000 different rulers to measure what the value is. Some of those rulers are really good, others are really lousy. The problem is no one is paying attention to the quality of these systems themselves.

Do you think there needs to be nationwide teacher evaluation standards, or will statewide systems be enough? 

I feel strongly that there should be a lot of latitude for local experimentation. We don't know enough yet for the state of Ohio or Texas or the federal government to be saying 'This is how we should do teacher evaluation.'

What are some of the problems with the ways we're doing teacher evaluations now? 

We don't know enough yet about the best way to deal with 'value added.' Fifty percent of teacher evaluation is a number people are plucking out of midair. Why shouldn't it be 30, 40, 72 percent?

Because teacher evaluation systems are still in their infancy, is it fair for states to implement merit-pay systems? 

Evaluating complex human behaviors—whether it's teachers or reporters—every system you try to apply to [evaluating these behaviors] is seriously flawed. They will generate mistakes. That doesn't mean they shouldn't be used. Teachers are getting differential pay now, and they're getting it for things we know don't matter, like pure seniority and the number of graduate credits they've earned.

We've seen that after the first three to five years, experience doesn't add anything. [Basing pay on teacher performance] is better than basing it on something irrelevant. Even though these will generate the wrong conclusions, almost every evaluation system will. If there's rough justice in the system, and that's all we can expect, it will have positive effects on teachers.

One of the real values of putting a meaningful teacher evaluation system in place is it helps a teacher understand where they are when teaching similar kids. Right now, if I were a teacher, I wouldn't know how I was doing, if my failures were due to a student's background or due to things I have no control over. This is not all about performance pay or management—it's also about giving teachers real information about the jobs that they're doing.

In your report, you reference a study that found in four states using a pass/fail teacher rating system, 99 percent of teachers were judged as "satisfactory." Is that because teachers are being evaluated by people within their schools?

It's because there are no consequences to the evaluations—the principal is not put in the situation of having bonus money to give out. It's a meaningless exercise. Once there are consequences, principals will make more discriminating judgments because they have to. With no consequences for the teacher or principal, why place yourself in an unpleasant interaction with one of the teachers who report to you?

Are there any districts that are on the right track with teacher evaluations? 

Hillsborough County, Fla. is way out in front. They're a leader in trying to create a teacher evaluation system. They are trying to produce value-added data. They pre-test and post-test for every subject and every grade.

Cincinnati also has an evaluation system that's been frequently studied, because it's been around for 10 years. It's about the only system people can look at that has a long track record. And, in years teachers are being evaluated, they do a better job. Student scores go up compared to off years.

See how your school stacks up in our rankings of Best High Schools. Have something of interest to share? Send your news to us at highschoolnotes@usnews.com.