Getting Serious about Education: Why Can We Measure Students but Not Teachers?

Last week, Michelle Rhee, chancellor of D.C. public schools, made national news by firing 241 — six percent — of the District’s teachers deemed underperformers. Rhee’s move came after negotiations in June with the Washington Teachers’ Union that created a merit-based bonus system that permits well-performing teachers to earn up to a 21 percent pay increase. The agreement also allows the District to fire those who did not meet minimum benchmarks. Teacher assessment scores will be based half on student improvement and half on in-class teacher evaluations.

While performance-related pay has been around since the 1700s and affects the pay scale of over 85 percent of private sector employees, the debate over merit pay for teachers is still highly contentious. On one hand, proponents argue merit pay will help cash-strapped schools retain good teachers and shed bad ones. They also argue that this will create a salary scale that is fairer than the system of seniority pay that currently exists in most school systems. On the other hand, opponents contend that merit pay may work for seamstresses, but teaching is too complicated to base quality on student performance on a standardized test.

The argument goes, evaluating teachers based solely on a set of student-achievement benchmarks will incentivize teachers to neglect the essential but non-tested responsibilities of educators. As George Parker, current president of the Washington Teachers’ Union put it, “It [merit pay] takes the art of teaching and turns it into bean counting.” Yet numerous other professions that require complex skills and responsibilities have adopted merit pay with positive results. For example, the department of Homeland Security has recently implemented performance-related pay for security analysts, and few would equate scrutinizing terrorist threats with “bean counting”.

The real question for education policy makers is to what degree can metrics assess the added value of different teachers? Part of the answer to this question relates to the availability of good data. Teacher performance may vary significantly depending on a number of variables such as student household income or the percentage of students with English as a second language. Without significant aggregate databases recognizing and accounting for such variables when developing performance pay systems may be difficult or even impossible. Yet technological advancements in the longitudinal data systems being put in place in states and districts are increasingly allowing for a more granular understanding of where educators do and do not add value to the learning process. Although it’s probably true that the current level of data may not be enough to predict exactly what makes a good teacher, what’s important is to use the data, along with the ways of assessing teacher performance, we have to make a better incentive system for the nation’s educators.

Yet that hasn’t been the case. In 1950, for example, 97 percent of public school teachers were paid based on seniority and education attainment (because data did not exist to fairly reward teachers based on any other benchmarks). But by 2007, 96 percent were still being paid based on these payment schedules; regardless of the numerous studies that have actually found experience (after the first two years) and teaching certifications are two of the worst indicators of teacher performance.

The blatant disregard of the evidence is not accidental. Several players in education policy—particularly teachers’ unions—have described evidence-based pay as some sort of pedagogical chimera, sucking the lifeblood out of what it takes to be a good teacher. For example, Doug McAvory, general secretary to the National Association of Head Teachers, a teachers’ union in the UK, argues, “The extension of performance-related pay based on pupil progress will further demoralize and demotivate teachers.” Yet given that educators readily embrace handing out grades to seven year olds, the argument that performance metrics might “demoralize” underperforming adult professionals seems unpersuasive. Such arguments do little more than distract educators from the importance of using technology and advanced metrics to create better schools.

ITIF and others have emphasized the importance of an educated workforce to improving the innovative and global competitiveness of the United States, but U.S. students are falling behind when measured against their counterparts in other countries. For example, in 2007, only ten percent of U.S. fourth-graders and six percent of eighth-graders scored at or above the international benchmarks in mathematics. Yet as other nations, such as Finland, South Korea and Sweden, have embraced data-based pay in public schools, United States has resisted.

Educators and policy makers should keep in mind the simple syllogism that there is nothing better than a good teacher and nothing worse than a bad one. As a society, we should do what is necessary to get more of the former and fewer of the latter, whether that requires more money, monitoring or better metrics.