It can work—but only if performance is broadly defined and all parties agree to the plan.
Donald B. Gratz | Educational Leadership
November 2009 | Volume 67 | Number 3
Multiple Measures Pages 76-79
A new round of interest in performance pay has been growing for the past decade, as more states and districts have introduced mandates related to the push for ever-higher standards. This year, President Obama and U.S. Secretary of Education Arne Duncan have included performance pay among their goals for education. At the 2009 National Education Association (NEA) convention, Secretary Duncan urged teachers to support performance pay, noting that although "test scores alone should never drive evaluation, compensation, or tenure decisions," not including student achievement in teacher evaluation is "illogical and indefensible" (ASCD Educator Advocates, e-mail communication, July 18, 2009).
Definitions of teacher performance pay take different forms. For example, many districts pay experienced teachers to mentor new teachers, serve as curriculum specialists or in similar posts, or teach in inner-city schools. The most common and controversial proposal is to pay teachers on the basis of their students' standardized test scores. It turns out, however, that test-based pay is more useful politically than it is effective educationally. Although few contemporary plans built on student test scores have lasted, this lack of success has not slowed proposals for more such plans. Given the growing prevalence of performance pay, it is worth exploring its history and assumptions.
A Not-So-Stellar Story
Education performance pay stretches back hundreds of years. In the mid-1800s, British schools and teachers were paid on the basis of the results of student examinations, for reasons much like today's. After more than 30 years, however, the testing bureaucracy had burgeoned, cheating and cramming flourished, and public opposition had grown dramatically. The practice was abandoned as a failure.
In 1907, Edmond Holmes, Great Britain's chief education inspector, described schooling in the era of test-based performance pay as the teacher engaged "in laying thin films of information on the surface of the child's mind, and then, after a brief interval, in skimming these off in order to satisfy himself that they have been duly laid" (Nelson, 2001, p. 386). Holmes referred to this kind of recall as being "the equivalent of food which its recipient has not been allowed to digest" (p. 386).
In 1918, 48 percent of U.S. public school districts described their payment systems as "merit based." But "merit" was subjective: White men were paid more than minorities and women, a disparity that eventually fueled a movement toward a uniform pay scale. Two years after women won the vote, the first uniform pay plans appeared in Denver, Colorado, and Des Moines, Iowa. By the 1950s, only 4 percent of U.S. school districts described themselves as merit based (Murnane & Cohen, 1986; Protsik, 1996).
There were brief attempts to implement performance-based pay in the early 1960s after Sputnik, and again when President Nixon launched an experiment with "performance contracting," which ended in cheating scandals and failure. In the early 1980s, when A Nation at Risk alarmed citizens with the prospect of "a rising tide of mediocrity" threatening to engulf U.S. schools, President Reagan reintroduced experiments with merit pay, with similarly negative results. Some school districts experimented through the 1980s with incentive programs based on merit, management by objectives, and career-ladder or differentiated staffing approaches. Few such experiments had any staying power. A new wave of experiments developed in the 1990s, most of which were also based on career ladders, teacher skills and knowledge, or differentiated staffing.
A New Approach Emerges
In 1999, the Denver, Colorado, school board and teachers association jointly sponsored a pay for performance pilot based largely on student achievement. As head of the outside research team for the first half of the pilot, I can attest to the energy and commitment with which the joint labor-management design team approached the task.
Although the pilot was successful and teachers in pilot schools supported it, designers saw that measures of student performance were still inadequate, that connections to teacher performance were hard to establish, and that standard measures of student learning were not applicable to more than half of the teachers—including gym, art, and music teachers; media specialists; special educators; and so forth. The model didn't address incentives for teachers to work in difficult situations and didn't include the contributions that many teachers make in support of their schools, younger colleagues, students, and their students' parents.
A much broader assessment of teacher performance was needed to capture the breadth of the teacher's role (Gratz, 2005). After four years and substantial effort, teachers and administrators collaborated to produce a new plan that the board, teachers, and voters ultimately approved. In the process, Denver expanded its definition of performance.
Denver's groundbreaking professional compensation plan replaces the traditional "steps and lanes" approach to compensation, in which teachers receive annual "step" increases as well as "lane" increases if they earn additional degrees. Only one of the new plan's four components directly addresses academic achievement goals—and that one is based significantly on teacher-set objectives, not just standardized test scores. In addition to student academic growth, the plan addresses teacher skill and knowledge, professional evaluation, and market incentives—compensating teachers who work in hard-to-serve schools or in hard-to-staff positions.
The Flawed Logic of Most Plans
Although today's performance pay plans take many forms, the most commonly proposed version—in which teachers are rewarded on the basis of their students' standardized test scores—flows from flawed logic and several troublesome assumptions.
Assumption 1: Teachers Lack Motivation
If we believe that additional pay will motivate teachers to work harder, we must also believe that teachers know what to do to improve student achievement— and that they aren't doing it because they aren't sufficiently motivated. The assumption is that they must value financial rewards more than student success.
Does anyone really think that large numbers of teachers know what their students need but are willfully withholding it? That they would help students learn more, if only someone offered them a bonus to do so? This is a highly cynical view of teachers, one that teachers understandably find demeaning, not motivational.
Most teachers care about their students and want them to succeed. Why else enter the profession? But although presenting information may be simple, successful teaching is more complex. Some teachers could certainly do a better job, but they mostly need mentoring, support, supervision, and training in new techniques—plus opportunities to learn, grow, and take on additional responsibilities—just like the rest of the workforce.
Assumption 2: Schools Are Failing
The broad call by state and national leaders for performance pay and other "reforms" is based on the widespread presumption that U.S. schools are failing. Schools have been labeled "in crisis" since the 1800s, but this designation has usually been more political than real. Schools were blamed for letting the USSR's Sputnik "beat us into space" in the 1950s, for economic collapse in the 1980s, and for economic inequality in the early 2000s. U.S. students are accused of lagging behind their peers in other countries, and a wide range of reports over the past two decades has predicted economic disaster in the future because, the reports claim, today's students are unprepared for work and will not be productive.
In fact, the United States had a satellite nearly ready to launch in the 1950s, but kept it under wraps because of its anticipated use in spying. Explorer I was launched just four months after Sputnik. The downturn of the 1980s, for which schools were often blamed, was followed in the 1990s by the longest period of sustained growth in history, for which schools received little credit.
As for the failure of U.S. students to measure up to their foreign counterparts, this is largely not the case. Rather, as many researchers have shown, the test score gap often results from comparing older or more select students in other countries with a broader range of U.S. students and from confusing test scores with achievement (Bracey, 2005; Mathews, 2008; Rotberg, 2008).
In fact, although poor results on specific tests make headlines, U.S. students compare well with their international peers on many tests, and U.S. workers excel in measures of economic success, such as creativity and innovation. Further, test scores don't correlate with economic success. The countries whose students outscore their U.S. peers do not have stronger economies or more productive citizens. Worker productivity in the United States soared in the 1990s and has remained high.
Schools make an easy target, but school change moves too slowly to affect short-term economic cycles. It takes at least 12 years for a restructured K–12 curriculum to produce its first newly trained students. So, although an educated workforce is important, schools have little effect on economic cycles. Fortunately, schools have not yet been blamed for the current economic debacle.
It's true, of course, that we have some very troubled schools in the United States—mostly in large, bureaucratic districts and mostly serving poor children and children of color. By one estimate, the majority of failing schools in the United States are found in only 29 districts (T. W. Slotnik, personal communication, July 10, 2008), suggesting that these districts need improvement at the school and district leadership levels, not just among teachers. Despite the existence of troubled schools and districts, however, most students achieve more academically now than in past decades, and most parents give their schools high marks and support them (Bradshaw & Gallup, 2008).
Assumption 3: Measuring Academic Achievement Is All That Counts
The third assumption—the most perilous for the United States—is that standardized test scores accurately measure student academic achievement and that academic achievement constitutes the full range of goals we have for students. However, beyond basic academic skills, corporate leaders have consistently cited the need for critical thinking, problem solving, teamwork and collaboration, communication skills, and a good work ethic as the keys to worker success. And because there is more to life than work, most citizens want children to learn about art, music, and other aspects of civilization; to explore and develop their own skills and talents; and to become good neighbors and active, productive citizens.
Look for the megacompanies that control standardized testing to produce new tests that claim to measure all these attributes—but don't believe them. If we want students to develop as well-rounded human beings who are empathetic, thoughtful, and creative, we will have to include these characteristics among our goals for schools and seek ways to gauge our success. A system that rewards schools, students, and teachers only for test scores will get mostly test scores. This is not what most of us want for our children.
The most promising aspect of the current discussion is the surprising extent to which district leaders, corporate leaders, and teachers unions are recognizing that they have common interests—interests that include accountability, expanded professional responsibility, and improvements in teaching conditions. Many parties are coming to see the value of higher and differentiated pay, and although we must carefully consider the specifics, the potential for change that benefits both teachers and students is real.
Denver's pilot has helped to demonstrate some of the possibilities. In an increasing number of districts, teachers who teach in hard-to-serve schools, such as those in inner cities, or in hard-to-fill positions, such as advanced physics, may earn additional pay. Teachers who mentor younger teachers, develop components of the curriculum, or take on other specialized duties may also earn more. Such pay doesn't insult teachers. Instead, it provides experienced teachers with the opportunity to learn, grow, and support their colleagues—critical opportunities in all professional fields to keep people refreshed and engaged.
Denver's plan also involves teachers in setting objectives for their own students, an approach that engages both their professional judgment and interest. Such opportunities have often been missing from teaching in the past.
Beyond Denver, it is also promising that some larger districts, often with outside technical assistance, are using performance pay as a catalyst for fundamentally changing how they do business— reorganizing their processes around their goals for their students and how best to reach them.
Finally, it is crucial that the discussion of performance pay—which requires districts to develop a new definition of performance—leads states and districts, including Denver, to a new consideration of their true goals for their students. An exclusive focus on academic achievement is a relatively new idea in U.S. education. Thinkers like Thomas Jefferson and Ralph Waldo Emerson did not expect schools to teach all children the same facts over an extended 12-year period. Rather, they believed that schools should provide individual students with basic skills and tools for learning so these students could pursue their particular goals and find their own place in society.
This requires a national discussion. Do we want a world of critical thinkers or a world of test takers? Do we care about citizenship, civic engagement, and the ability to work with others? Do we value the arts and humanities? Do we want each child to develop his or her individual talents and abilities? If so, how do these goals fit into our definition of student and teacher performance? How do they align with No Child Left Behind?
Until we can answer these kinds of questions—until we determine the breadth, depth, and individual scope of student achievement that is worth pursuing—paying for performance will not produce the results we want for our children or our society.
Bracey, G. W. (2005, October). 15th Bracey Report on the condition of public education. Phi Delta Kappan, 87(28), 138–153.
Bradshaw, W. J., & Gallup, A. M. (2008, September). Americans speak out: Are educators and policy makers listening? Phi Delta Kappan, 90(10), 7–31.
Gratz, D. B. (2005). Lessons from Denver: The pay for performance pilot. Phi Delta Kappan, 86(8), 569–581.
Mathews, J. (2008, Spring). Bad rap on the schools. Wilson Quarterly, 32(2), 15–20.
Murnane, R. J., & Cohen, D. K. (1986, February). Merit pay and the evaluation problem: Why most merit pay plans fail and a few survive. Harvard Education Review, 56(1), 2.
Nelson, W. (2001, January). Timequake alert: Why payment by results is the worst "new" reform to shake the educational world, again and again. Phi Delta Kappan, 82(5), 384–389.
Protsik, J. (1996, May). History of teacher pay and incentive reforms. Journal of School Leadership, 6, 265–289.
Rotberg, I. C. (2008, June 11). Quick fixes, test scores, and the global economy. Education Week, 27(41), 27, 32.
Donald B. Gratz is Director of Graduate Programs in Education and Chair of the Education Department, Curry College, Milton, Massachusetts, and author of The Peril and Promise of Performance Pay (Rowman and Littlefield Education, 2009); firstname.lastname@example.org.