EdMedia Commons Archive

Five Questions For …. Dale Ballou

Dale Ballou, an associate professor of public policy and education at Vanderbilt University, specializes in the role of regulations and incentives in teacher training and retention. He spoke with EWA about the current debate over teacher evaluation models as a means of public school reform. The interview was conducted and edited by Emily Richmond.

1. Are people putting too much stock in teacher evaluation models as a means of improving the quality of instruction?

What we started with in No Child Left Behind was just to hold every school and district to one strict standard without any regard for the incoming level of achievement of the students. That obviously presented a lot of problems. The student growth models being proposed as an alternative are certainly better than what we were previously doing. On the other end of the spectrum are evaluation models with possible personnel actions based on some score a teacher gets. You’ve got to recognize these are not flawless instruments. Anytime you try to evaluate an individual based on this kind of data, there’s a possibility you’re going to make a mistake. People are overselling these methods and portraying them as if they represent the answer to all kinds of problems, and that they give you the truth. What do they do give us is an estimate, which is subject to error.  

2. What are some of the hallmarks of a well-designed teacher evaluation model?

In a good model, the intent isn’t to be punitive. There is so much you can do with the information. You could make better classroom assignments, or have teachers assigned to the subjects where they are most successful. You can provide better feedback on performance. This can be an eye-opening experience and really change the way schools do business.

3. In a publication outlining concerns about evaluation models, the National Education Association argues that hiring practices and preparation–the steps taken before a teacher enters a classroom–are the most productive ways of ensuring effectiveness. Do you agree?

I disagree. There just aren’t enough good signals people send prior to the actual experience of becoming a teacher that reliably predict their performance. That goes for teacher education programs, as well. There’s only so much you can do in advance to get people ready for teaching. Much of becoming an effective teacher is what you start learning when you actually do the job. You’re talking about a career where people don’t have much advance information as to how well they’re going to do.  It’s also hard for any third party to identify who is going to be effective. When you have two or three years of classroom performance data, you’re really getting information on who is–and isn’t—having success.

4. Some politicians argue that education reform is being held back by union contracts that make it too difficult to fire ineffective teachers. On the flipside, teachers say their job protection is threatened. Who’s right?

I don’t know of any collective bargaining agreement that doesn’t include a provision for dismissing teachers. That being said, some districts obviously have more latitude than others. There can be a big payoff for weeding people out who are not doing well early in their teaching careers, but there’s been a real reluctance on the part of people in education to take advantage of that window. A lot of states are providing tenure in as little as three or even two years. Tennessee recently passed a law that makes teachers wait five years for tenure. That’s a smart move, and other states would be wise to take similar steps. This is not an earth-shaking change. We’re not talking about pushing out 20-year veterans. This is about waiting until there’s appropriate evidence to support the decision.

5. How can education writers do a better job covering these issues?

There are a lot of self-interested voices in this debate and that’s a difficult environment for a reporter to figure out what to believe. Education writers could play an important role in getting the information out there but they need to educate themselves first. The single biggest improvement we could make is to start giving people a much clearer sense that many of these teacher evaluation models are estimates that are often subject to a wide range of error. This is an area where a little bit of knowledge about statistics can really pay off. There are plenty of us in academia willing to help writers’ understanding of the underlying science so you can figure out what the next question should really be. The resources are out there if you tap them. 


This post originally appeared on EWA’s now-defunct online community, EdMedia Commons. Old content from EMC will appear in the Ed Beat archives.