Paying Students To Do Well in School: What Economists Are Learning about Pay-4-Performance

Economists believe in incentives and what better incentives are there than writing checks to students who perform well on tests. The idea had been around for a long time but few policymakers or researchers could say with any confidence whether paying students worked.

Certainly, the policy climate in the U.S. since the early -1980s of increasing parent choice in schools, more accountability, and higher academic standards has welcomed economic incentives. States have paid cash for higher test scores and used dollars to get high school graduates into the labor market thereby strengthening the entire economy in an ever-changing and competitive global market.

Now comes Roland Fryer’s recent study “Financial Incentives and Student Achievement from Randomized Trials.”(See student incentives)

Nearly 40,000 students in 261 schools in four districts (Washington, D.C., Chicago, New York, and Dallas) participated. Researchers gave $6.3 million in rewards to schools.

In New York City, the researchers paid fourth and seventh grade students for their performance on ten tests given by the NYC Department of Education to all students. In Chicago, they paid ninth graders every five weeks for grades in five core courses. In Dallas, they paid second graders $2 per book to read and pass a short quiz to confirm they read it. In the District of Columbia, they paid sixth, seventh, and eighth grade students for improved attendance, behavior, and three other criteria chosen by individual schools. Thus, researchers ran controlled experiments by varying what was rewarded, how often students were given incentives, students’ grade levels, and how much students received.

In two cities, money was paid to students who read books, improved their daily attendance and behavior–in short, the “inputs” to achievement, while in two cities students received checks for performance on tests and grades, the “outputs.” To avoid selection bias, researchers used school-based randomized choices for students who took the tests. Finally, except in Dallas where students were paid three times a year, students received personal checks within three days of their test results being confirmed.

What did the researchers find out when it came to these incentives? “Remarkably,” Fryer says, “incentives for output did not increase achievement (p.5).” This occurred across all grades, across cities. However, “paying students to read books (Dallas) yields a large and statistically significant increase in reading comprehension (p. 5).” Ditto for students in Washington, D.C. who improved their test scores when they were focused on improving the “inputs” (attendance and behavior) to achievement.

Surely, it is dicey to lean too much on any one study even if the choice of schools were randomized and the design was experimental. I do not. What flashed through my mind, however, was the connection between students’ pay-4-performance touted so often by policymakers who lust after randomized experiments and econometric thinking and, yes, teacher pay-4-performance. The connection?

For those reformers who believe that teacher accountability, i.e., personal responsibility for student performance on tests, has to be incorporated into teacher evaluations, tenure decisions, and salaries, “now it’s time to tackle teacher tenure and evaluations head-on,” according to Mike Petrilli. “That’s what we’re seeing in Florida, with the far-reaching bill just vetoed by Governor Crist. That’s what we’re seeing in Colorado, with a bold proposal just released by state senator and uber-reformer Mike Johnston. And that’s what we could see nationwide if states were willing to step up to the Race to the Top’s challenge for meaningful teacher accountability, ” Perhaps.

Here is where I return to Roland Fryer’s study. He showed that paying students directly to get higher test scores failed to yield results–some of the students did worse on the test–while paying students to read books, come to school daily, and sit in class without causing a ruckus, the so-called “inputs” showed far higher yields on test scores. I do wonder whether there is an application to the current crescendo of support for teacher pay-4-performance.

Were one to translate this study to the growing policy support that teachers should be evaluated, given tenure, and paid according to how their students do on tests, one might pause for a moment and think twice about “inputs” to improved teaching. Maybe considering instructional coaches, more time for professional development, teachers-helping-teachers, and dozens of other ways that district and school administrators and teachers have used historically to gain knowledge and skills might be considered worthwhile.
Again, Fryer’s study is, thus far, singular. What he has done is use a close-to-gold-standard design and tested the taken-for-granted assumption by those policymakers and their cohorts endorsing market-based incentives that paying students to do well on tests would work. Not in these four urban districts. But the “inputs” associated with improved performance did yield positive results. Applied to teacher accountability hullabaloo now, folks should , at the least, pause and consider the rush for students’ test scores to evaluate, tenure, and pay teachers.