The NYC Teacher Pay-for-Performance Program: Early Evidence from a Randomized Trial

Matthew Springer and Marcus Winters
Center for Civic Innovation at the Manhattan Institute
April 2009

This random assignment study of New York City's School-Wide Performance Bonus Program isn't going to calm the controversy over performance-based pay--at least not yet. It looks at the math scores of 100,000 students in 200 Gotham schools and finds that the bonus program had no effect on achievement or on students', teachers', or parents' perceptions of the school learning environment. (Oddly enough, they found limited evidence that math performance decreased in larger participating schools but those findings are likely "spurious"--i.e., researcher talk for "lacks validity.") But don't despair yet. Although MI's report is an impact study, there were less than three months between the start of the performance-pay program and the administration of the state test. Virtually any educational intervention is doomed to have negligible effects if given 90 days or less to take effect. Seeking to downplay the impact angle of the study, analysts rightly term it a "baseline" endeavor and essentially say to hang tight until the program matures a bit. (Perhaps that's why MI made little effort to promote its findings.) Unfortunately, the program itself is weak. It uses the school rather than the individual as the unit of accountability, making for a feeble incentive (it was UFT-negotiated, after all), and the payout is small potatoes (three-fourths of the schools awarded a maximum individual bonus of $3,000 or less). But at least the name of New York's plan is honest: This is indeed a school wide bonus plan--and a far cry from individual merit pay.

More By Author