Schoolwide bonus pay for teachers in New York City’s now-suspended program didn’t seem to boost student achievement as a general rule—and in schools with many teachers, it may have diluted individual incentives for teachers to boost achievement growth. But in those schools with high levels of teacher collaboration and a small staff, such programs may have had some slight benefits, concludes a new study by two Columbia University economists.
In a school with lots of teachers, authors Sarena Goodman and Lesley Turner wrote, “the diffusion of responsibility for test-score gains across many teachers may erode the incentive that any individual teacher has to increase effort in the classroom.”
Some important context for you before we delve into these findings: For a long time, policy types have disagreed about whether performance incentives should be individual or group-based. Proponents of schoolwide programs have argued that individual pay programs would cause competition among teachers, rather than collaboration. (American Federation of Teachers President Randi Weingarten once famously said she abhorred the individual type.)
On the other hand, others have argued that schoolwide plans would allow less-dedicated teachers to “free-ride” off of the contributions of their peers. (Before you jump all over me for writing that, the term “free-rider” comes from the research literature.)
And, of course, everyone has wondered about whether merit pay of any sort works to raise achievement. The naysayers seem to have the upper hand at the moment: Two recent plans with individual pay elements have not shown many effects in randomized experiments.
For this study, the economists looked at New York City’s schoolwide performance bonus program, created jointly by the city teachers’ union and the mayor in 2007. They compared results for 181 schools implementing the bonus program with those of 128 in a control group with similar characteristics that didn’t implement it, and they also used school survey information collected as part of the city’s accountability system. They also tested to see whether treatment effects on math and reading scores varied by the number of such teachers, since it was those teachers’ contributions that most directly affected the test scores.
The details of the bonus program are complicated, but in essence, schools that reached pre-set schoolwide achievement goals received a lump-sum payment equal to $1,500 or $3,000 per teacher. Teams of two administrators and two teachers in each school had freedom to allot the awards as they liked as long as every teacher got at least some bonus payout.
Here are some of the findings:
• Schoolwide bonus pay didn’t seem to affect student achievement, teacher instructional technique or absenteeism, or the quality of the teaching pool for the majority of schools. In year two, eligibility for the program actually seems to have slightly depressed math achievement in general.
• The authors found evidence of “freeloading.” Participating schools with the fewest math teachers showed some slight but significant achievement growth in that subject. In plain English, these teachers appear to have stepped up to the plate, causing achievement to rise. But in schools with more math and reading teachers, where the pressure to exert an effort was weakened and where there were more obstacles to collaboration, did not see such gains and may even have seen some declines.
• An interesting caveat: Schools with high degrees of teacher collaboration (as measured by a “cohesion index” based on teacher survey reports) also exerted a slight upward pressure on math scores.