Theoretically, merit pay sounds great. You take the industrial model of quality improvement — measuring output and rewarding those producers who perform best — and you simply apply it to education. It’s such a simple concept: What could go wrong?
Then again, I also love the idea of rocket ships that travel at the speed of light. Unfortunately, we haven’t quite figured out a way to make that work, and the same is true of merit pay. It’s a complicated idea with a lot of moving pieces — more complicated than rocket science, in fact, because of human variability — and unless you get them all right, you can get it all wrong.
Nonetheless, Gov. Nathan Deal — echoing a campaign by President Obama and the U.S. Department of Education — has announced that he will ask the Georgia Legislature to take what he calls a “significant step” toward implementing a statewide merit-pay program.
“We’re not going to go to a fully merit-based pay system, but I do think there is a portion of the teachers’ pay that should go to how good a teacher they are,” Deal said. “Now, getting the education community to support that is sometimes difficult.”
Yes, and that’s just the beginning of the difficulties.
For example, if you’re among the growing number who fear that we already place too much emphasis on “high stakes” standardized testing, resulting in schools that relentlessly “teach to the test,” then merit pay ought to trouble you immensely. Standardized test results are the primary means of assessing teacher proficiency, and if you want to see someone “teach to the test,” give him a $5,000 bonus if his student’s test scores rise.
It’s easier in other industries, where raw material comes in at the loading dock and a standardized, easily measurable final product comes out the other end. That does not describe a schoolhouse. Students do not come in standardized, easily measurable units and neither do classrooms or schools. The variation is immense, and statisticians warn that a test sample of 25 or 30 students — the size of the standard classroom — is much too small and random from which to draw robust, actionable conclusions.
That’s in part why the American Statistical Association — a professional group whose members believe deeply in the power of statistics — cautions against the approach. According to the ASA, research shows that teacher quality accounts for somewhere between 1 percent and 14 percent of the variation in student test scores; the rest is “attributable to factors outside of the teacher’s control such as student and family background, poverty, curriculum, and unmeasured influences.”
In addition to measurement issues, we face the danger of financial incentives producing behavior that runs contrary to your goal, as occurred in the test-driven Atlanta Public Schools scandal.
For example, if a low-performing student in your classroom is going to endanger your $5,000 merit-pay bonus, as a good teacher you might target that student for extra attention and support. But you might also label the child as a discipline problem or as learning disabled to get him removed from the classroom. Merit pay opens the door to all kinds of ways to game the system, by administrators as well as teachers, and none of them is any good for kids.