Front Page - Archives - Search 
Class Notes
How do you judge who merits pay?
(Published April 3, 2006)


In February, Florida joined Denver, Houston, North Carolina and many others on the list of school districts and states to implement pay-for-performance in teaching. The current U.S. Congress has shown tremendous willingness to go around our local government and experiment with the latest free-market education fad in D.C., e.g. charter school funding. Is pay-for-performance next? Is it a good idea? Should pay-for-performance go from Bush's mouth (in this case, Jeb) to D.C. schools?

Pay-for-performance, also known as merit pay, is the idea that better teachers should get better pay. In other words, the teachers who do the most to increase student achievement should be rewarded with more money. Just like the most productive worker on an assembly line would get a bonus for putting together more toasters, teachers in the school line would get a bonus for producing more proficient students. Any teacher with confidence in their teaching ability should be arguing for this. Instead, most teacher unions argue against merit pay. Why?

Merit pay seems like a good idea. It's a seductive premise: rewarding good teaching will encourage teachers to work harder and give teachers incentive to do more of the things that are proven to work in the classroom. How can you argue against a system that gives bonuses to people who are lifting student achievement? The current D.C. pay system, like Florida's, doesn't recognize differences in quality. In the current pay system, good and bad teachers with the same experience and educational level receive the same pay, regardless of their evaluation or classroom outcomes. To quote the Florida Department of Education: "We know there to be no significant correlation between either years of teaching experience, or level of degree obtained with actual growth in student performance. For example, the best and worst teachers in a district, as measured by higher student achievement levels, may have both served the district for 15 years and have the same education level attained, so currently they are paid the same." (

Critics of merit pay often point to tricky implementation issues. School districts don't have good measures of teacher quality and teacher performance. Florida's "as measured by higher achievement levels" system, like many others, uses standardized test scores as the yardstick of good teaching. What's wrong with that? In the March 20 issue of this paper, I argued that the coming DCCAS standardized test will give valuable feedback on what was taught well, and what needs to be taught. Some might say that if I had the courage of my convictions, I should be supporting bonuses (and penalties, since any bonus money would reduce money available for other programs) for standardized test score gains.

Using standardized test scores to quantify good teaching has three glaring flaws. It fails to accurately measure many aspects of good teaching. It aims too low. It gives the wrong incentives to teachers.

Much of good teaching is more than getting the right answer. Inspiring a child to research independently and to read for pleasure has only secondary effects on test scores. More importantly, teachers outside of English and math get short shrift: how do I reward my outstanding music or art teacher? Do I punish the social studies department because test scores declined in math? Also, which teacher is better (i.e. deserves more reward): the teacher whose students improve from 20 percent to 30 percent, 60 percent to 66 percent, or 85 percent to 90 percent? Is it harder to lift the kids at the bottom or the top? Should we give more credit to a teacher with many kids who just scrape over passing (65 percent?) or the teacher who has a few kids go from lumps to super-stars? Do we only reward teachers in the grades tested? (Currently, D.C. tests fourth, eighth and 10th grade students.)

Using standardized test scores to determine bonuses also aims too low; much of the 10th grade math test is really seventh grade math. A full, challenging and relevant curriculum is more than just what will be on the test. Standardized tests are not performance assessments where kids demonstrate the application of their skills in a real context.

Pay-for-performance based on tests also gives the wrong incentives to teachers. Students at my school currently need to improve their reading and writing. Math teachers have been asked to incorporate literacy in every lesson, including having students read articles and write essays describing their mathematical problem solving. If I'm rewarded based on my students' math scores, why would I take time during math instruction to work on reading? More callously, why should I tutor another teacher's student after school? That child's improvement might not fatten my wallet. And at the extreme, I should ask the parents of lower-performing kids to transfer them to other schools. In fact, one of the reasons why No Child Left Behind has a percentage-tested requirement is to ensure that schools don't quietly encourage lower-achieving students to stay home during testing week to make their scores look good.

Maybe all we need is a good way to evaluate teachers, perhaps one that uses other measures in addition to test scores, that somehow avoids the pitfalls of subjective processes and that motivates teachers to achieve success for all. I don't think so. The problem is not in the way we evaluate teachers.

Merit pay is an intrinsically bad idea because teaching is not a factory job. In fact, economists might call the entire teaching profession a "discontinuity;" people don't teach because of the free market. People might leave teaching because of market forces, but every teacher I know who's worth anything isn't staying in teaching for the money. We know when we start teaching, especially in D.C. Public Schools, that the only reward we can expect is to see our children become successful in their lives; if you're the kind of person who needs a lot of external motivation, or recognition, or even safe, modern and professional working conditions, there are plenty of other jobs out there. During the school year, I work six days a week, usually from 8 a.m. to around 7 p.m. There are many people who work harder I know the custodians at my school work also work six days a week from 11 a.m. to midnight yet the notion that my colleagues and I need to be encouraged to work harder to help kids shows a profound ignorance of why teachers teach.

Above and beyond whether we have a good way to evaluate teachers, incentive systems encourage only the behaviors we need to pay for. Is that what the community wants in its teachers?


Wennersten teaches mathematics at Bell Multicultural High School in Columbia Heights and a graduate of the D.C. Teaching Fellows program ( Please send stories, comments or questions to

Copyright 2006 The Common Denominator