Teacher salary incentives are associated with higher levels of student performance.
Despite the growing interest in merit pay for teachers in American schools, the first U.S. evidence of a positive correlation between student test scores and financial incentive systems that reward individual teachers appears in Individual Teacher Incentives and Student Performance (NBER Working Paper No. 12627) by David Figlio and Lawrence Kenny. Up until now, the large school-survey datasets that underlie so much research in education typically have provided little or no information about teacher incentive pay. To overcome this problem, Figlio and Kenny conducted their own survey of school personnel practices in 2000. They surveyed schools represented in the National Education Longitudinal Survey (NELS); they also matched the NELS to the less detailed information on the use of merit pay in the 1993 Schools and Staffing Survey (SASS). Their final dataset contained information on 4,515 students from 502 schools.
Student achievement was measured as the sum of scores on the four twelfth-grade tests in reading, mathematics, science, and history. Eighth-grade test scores were used to control for unmeasured ability. The number of mathematics courses taken in high school and the number of days absent were used as proxies for motivation. Other control variables included family and school characteristics, including whether a student's school was unionized. Schools were classified as Catholic, non-Catholic private, or public.
To account for the possibility that merit pay systems that automatically reward large fractions of teachers probably do not improve performance, the authors constructed three measures of "merit pay exclusivity." The top fifth of merit pay systems had at least a 20 percent salary range and limited merit pay to no more than 5 percent of teachers. Among the bonus systems, bonuses were limited to no more than 7 percent of teachers. As it turned out, the non-unionized schools were more than twice as likely to offer teacher incentive programs. The schools in the sample also were classified by whether they either fired or encouraged the resignation of at least one or more experienced teachers in the last three academic years.
The authors find that the Catholic schools were "more than twice as likely" as the public schools to dismiss novice teachers and were more than three times as likely to fire experienced teachers. Otherwise, they were indistinguishable from the public schools. The non-Catholic private schools were different from the public schools in nearly "every measured dimension of teacher incentives."
Figlio and Kenny find that teacher salary incentives are associated with higher levels of student performance. They cannot be certain whether the test score improvement is driven by teacher incentives or whether the incentives are proxy variables for unobserved school quality. In general, they find, teacher salary incentives are associated with a 1.3 to 2.1 point rise in test scores, about the same increase associated with increasing maternal education by three years. The correlation exists in schools with predominantly low- and middle-income students.
Incentive programs that awarded bonuses to very large fractions of teachers were not correlated with higher student achievement. The authors point out that principals rated four out of five teachers as "good" or "excellent" in the 1993-4 SASS, and that even higher fractions of Florida teachers were worthy of state-funded merit pay in 2000. They caution readers that the benefits of merit pay will likely prove illusory as long as most teachers receive it.
-- Linda Gorman