Get stories like these delivered straight to your inbox. Sign up for The 74 Newsletter
“Merit pay,” according to education historian Diane Ravitch, “is the idea that never works and never dies.” New York City teachers union president Michael Mulgrew has argued, “In study after study, experiments with merit pay have failed to improve student performance.”
Journalists have echoed these claims. NY1 political anchor Errol Louis wrote that there’s “a mountain of evidence suggesting that [merit pay], which sounds fine in theory, simply doesn’t work.”
Statements like these have become received wisdom, often repeated without citation in news coverage.
In fact, the research on merit pay (also called performance pay, usually by supporters) is mixed, with some studies finding no effect but many others finding positive impacts. The emergence of the conventional narrative — that it has been proven not to work — shows how an education myth is quickly built, repeated, and self perpetuated: even when it’s wrong or missing important nuance.
A handful of high-profile studies make the case against merit pay — and a myth is made
In 2010, just before the release of a major study on performance pay, American Enterprise Institute scholar Rick Hess worried that the study wasn’t designed for “understanding what we care about when it comes to performance pay.”
He predicted that “the study will confuse the issue, obscure the actual question of interest, and (depending on the results) lend either simple-minded advocates or performance-pay skeptics a cudgel that they will henceforth freely misuse in the name of ‘evidence.’”
The next day, researchers at Vanderbilt University announced that their investigation, in which teachers were awarded $15,000 if they produced gains in student test scores, found that students with teachers eligible for bonuses did not perform better than those whose teachers were not eligible. “Teacher bonuses don’t raise student test scores,” concluded USA Today.
Soon to follow was a 2011 New York City study examining schools that received the opportunity for building-wide bonuses — distributed equally to each teacher — for strong performance. Again there was no evidence of a positive effect for merit pay.
Yet another study, which produced a number of headlines, found that performance pay only worked when it was given to teachers at the beginning of the year, but taken away if the teacher didn’t hit performance goals. Traditional incentive pay, though, didn’t produce results — yet another nail in the merit pay coffin, it seemed.
The limits of these studies — usually pointed out by the researchers but often overlooked in news accounts — is that they examined just one of multiple ways performance pay might have a beneficial effect. As Hess put it, “Whether the merit pay experiment shows big test jumps or none at all, it won't tell us a damn thing about the ability of performance pay to attract new talent to teaching, undergird efforts to promote professionalism, retain talent, or boost regard for the profession.”
Recent studies find more promising results
Matthew Springer, a professor at Vanderbilt who was coauthor of both the Tennessee and New York City reports, explained that such studies only measure one part of the potential benefits of merit pay. He calls this the “motivational effect”: whether teachers perform better or work harder due to the opportunity of extra pay.
(Or, as this idea is often caricatured, here by journalist Linda Perlstein: “Presuming that merit pay alone would elevate student achievement makes sense if you assume teachers have a hidden trove of skills and effort they are not unloosing on their students only because they lack the proper incentives to do so.”) This seems to be the most common theory of action for merit pay, at least as conceived by skeptics.
What these studies don’t look at though, according to Springer, is a “compositional effect,” meaning the possibility that performance pay would help attract and retain effective teachers. Springer says there is good evidence that merit pay can have a positive impact in this area.
For instance, Springer recently coauthored a study finding that a different performance pay system in Tennessee helped retain teachers in tested grades and subjects. Another study of Denver’s system showed that pay incentives improved retention in high-poverty schools.
And recent studies — in D.C., Minnesota, Austin, and several districts across the country — have suggested that performance pay can in fact directly improve student achievement, though it’s not always clear if the gains resulted from compositional, motivational effects, or some combination of the two.
Springer says these favorable recent studies have attracted less publicity than some of his earlier papers showing no impact. A quick search of news clips shows that Springer’s 2010 research was cited in coverage of recent performance-pay debates in Kansas, Georgia, and Pittsburgh. His newer findings, by contrast, have been largely ignored outside of local outlets in Tennessee, where Vanderbilt is located .
Springer remains optimistic that performance-based compensation can be an improvement over traditional salary schedules: “In large part the way in which we currently compensate teachers — using experience and degrees held — is highly inefficient. What the better practice may be we’re still trying to figure out. With careful design and implementation, we could be moving to a much better way.”
Towards a nuanced understanding of the research
Most summations of research findings do not fit on a bumper sticker, and the performance-pay literature is no exception.
Here’s what we can say: exclusively test-based bonuses are unlikely to change teachers’ behavior in a way that improves student achievement. However, performance pay linked to a comprehensive evaluation and support system is more promising on that front. Moreover, there’s encouraging — if still far from conclusive — evidence that merit pay can help attract and keep effective teachers.
Yet implementing performance pay systems, particularly determining how to measure performance, can be challenging. More research is needed to determine the long-term impact, potential unintended consequences, cost effectiveness, and best designs of merit-based pay.
These are questions worth discussing carefully and without overly broad declarations that performance pay “doesn’t work.” In education research, it’s usually a lot more complicated than that.