During the past few decades, we have seen the persistence of failed policies in education. Most of them were codified by No Child Left Behind and Race to the Top: give standardized tests; punish teachers and schools where scores are low or do not rise; reward teachers and schools where scores go up. Pay bonuses to teachers if their students’ scores go up. Tie teacher pay to student scores. Close schools with low scores. Turn low-scoring schools over to private management. Give vouchers to parents to send their children wherever they want.

All of these remedies failed. They encouraged cheating and gaming the system. They encouraged educators to avoid schools that enrolled the neediest students. They demoralized teachers who were idealistic and wanted to teach the joy of learning. Test prep became far more important than intellectual curiosity.

All of these are zombie policies. No matter how consistently they fail, policymakers won’t let go of them.

Merit pay is a policy that has been tried since the 1920s. It has never accomplished anything. I summarized the research on merit pay in my last three books: The Death and Life of the great American School System; Reign of Error; and Slaying Goliath. The research is overwhelming: merit pay doesn’t improve education and doesn’t even raise test scores. Yet in true zombie style, it never dies. It should.

John Thompson writes here about the revival of the merit pay zombie in Oklahoma:

As the “mass exodus” of teachers from Oklahoma schools continues, the legislature has rejected an across-the-board pay raise for teachers. Instead, several legislators are searching for a fix for the state’s “flawed” bonus system. If that doesn’t work, maybe Walters’ use of public money to spread his attacks on “on the radical left” will bring educators back to Oklahoma …

Seriously, Walters’ push for his vision of incentive pay prompted some education advocates to ask me to research performance pay. I sure appreciated the oportunity to reread new and older research on the subject.

Twenty-five years ago, I opposed performance pay because there were better ways to improve teacher quality. But I didn’t have major concerns; although its likely benefits would be small, I thought its downsides shouldn’t be a big deal. However, starting with No Child Left Behind and taking off with Race to the Top, test scores were weaponized, and the dangers of performance pay grew dramatically. Output-driven teachers’ salaries, joined at the hip with unreliable and invalid accountability metrics, promoted educational malpractice that undermined meaningful teaching and learning, increasing in-one-ear-out-the-other, worksheet-driven instruction. Teamwork was damaged, trust was compromised, the flight of educators from classroom increased, and the joy of student learning declined significantly.

During that time, I communicated frequently with data-driven analysts working for think tanks, who almost never had experience in urban schools. Their job was to provide evidence that performance pay, and other incentives and punishments, can work. They ignored educators and social scientists who tackled the real policy question – how will those experiments work? 

Sometimes, merit pay produced modest test score gains, but there was no way of determining whether those test scores revealed an increase or a drop in meaningful learning. Neither did they address the overall learning losses due to teachers being pressured to focus on metrics, as opposed to children. In 2012, a Rand study concluded, “most studies have found no effects on student outcomes.” By 2015, the U.S. Department of Education found that large incentives, such as $15,000 per teacher, may attract talent, but:

In addition to creating an environment that lends itself to narrowed pedagogical approaches and teaching to tests (and even cheating on them), this article suggests that merit pay schemes that require teachers to compete with one another may likely undermine positive collaboration.

Around the time of the 2018 Oklahoma teacher walk-out for higher pay, Denver threatened a strike to get rid of performance pay. Chalkbeat explained the complexity of balancing for larger or smaller payments to teachers in diverse classrooms. It went into depth answering the question, “How did a pay system that once seemed to hold so much promise bring teachers to their breaking point?”  The concise conclusion was, “lack of trust.”

Education Week studied the minimal effects of performance pay in Tennessee and Texas, which implemented expensive reward-and-punish, and often short-lived programs. The negative effects of the Houston plan, which State Superintendent Ryan Walters seems to support, are especially relevant for Oklahoma. The Houston teachers’ union president explained, “Performance pay demeans students and undermines teachers, so if the focus is on pay for performance, you’re incentivizing the test-and-punishment model.” Similarly, Education Week cited comprehensive studies that concluded that the relatively more effective programs “avoided an overemphasis on test scores.” But even many or most of the more successful programs were unlikely to survive.

Finally Education Week reported how the $200 million Bill and Melinda Gates Foundation merit pay experiments “did little to boost retention of high-performing teachers,” and it “had little effect on student achievement.”

At the peak of merit pay mandates, and now, Bixby Superintendent Rob Miller explained, “Teacher merit pay is one of the more persistent and seemingly indestructible zombie ideas related to education.” Miller said, “Merit pay for teachers has been tried again and again since the 1920s.” He cited cognitive and social science that explained why performance pay experiments were doomed to fail, as well as numerous evaluations of how it failed in the 21^st century.  Miller now asks, “Is it fair to place the primary responsibility on teachers and schools for outcomes strongly affected by factors outside their control?” and answers, “Doing so damages school culture and teacher morale and obstructs meaningful dialogue about school improvement.”

At a time when Ryan Walters is threatening to put the worst of the failed policies of the last twenty years on steroids, I was struck by a recent column by Thomas Dee, a fervent believer in output-driven accountability. Even though he seems to think that teachers were to blame, Dee also seems to acknowledge that performance pay had disappointing results. Now he recommends:

It may be possible to achieve durable political support for a teacher evaluation system if that system focuses narrowly on identifying master teachers and providing them with training and extra pay to coach their peers but takes a more incremental approach toward dismissing underperforming teachers.

Dee’s latest almost brings me back to 25 years ago, before NCLB, when the schools I knew were improving, and a win-win approach to performance pay didn’t seem so problematic. At the urging of the union, the Oklahoma City Public School System briefly implemented the Toledo peer review plan, which included a fair and efficient plan for removing ineffective teachers. The best evidence is that the plan was a reliable method for improving classroom instruction. But, it and so many other promising programs were undercut by corporate school reform.

Maybe I’ll once again be open to a compromise involving constructively built, non-punitive merit pay incentives, once the destructive school cultures advanced by corporate school “reform” have disappeared. But, I won’t hold my breath.