Another problem with piece-rate evaluation had been noted in similar evaluation schemes in mid-nineteenth-century England: when judged by test scores, teachers were likely to ignore the subjects that were not tested, meaning there was a narrowing of the curriculum, and important skills were ignored because they were hard to test.
The most significant obstacle to performance pay based on test scores, the authors argued, is the nature of teachers’ work. Very few private schools engage in the practice, even though it would be easy for them to do so. They don’t do it because schools have multiple goals, not just raising test scores. Even if you could isolate the teachers’ contribution to test scores, you would simultaneously destroy the teamwork that is necessary to accomplish other important goals, such as encouraging good behavior or reducing drug use. School principals recognize the multiplicity of goals and the importance of teamwork; they also understand the perverse incentives introduced by piece-rate compensation. That may explain, said Murnane and Cohen, why “paying teachers on the basis of their students’ test scores is extraordinarily rare in American education.” Even if the quality of standardized tests were far better, even if they perfectly matched what students were supposed to learn, they argued, the negative consequences of piece-rate compensation would still make it objectionable.
The authors also found fault with the “old style” of merit pay, where the principal decides who gets a bonus. Most of these experiments failed too. The problem is fairness and the perception of fairness. The supervisor must be able to answer two questions: One, why did worker X get merit pay and I didn’t? Two, what can I do to get merit pay? The task of unloading boxes from a truck is an activity for which it is easy to answer these questions. Worker X unloads more boxes than you; if you unload more boxes, you too can get merit pay. That makes sense to the worker. It is defensible. But when a principal awards merit pay to teacher X and not to teacher Y, he needs an explanation better than “I know good teaching when I see it.” Did the teacher get merit pay because she was outstanding or because she was a favorite of the principal’s? Will teachers know what to do to improve their chances of getting merit pay? And will they hide their problems from the principal instead of asking for help? Thus, it is the nature of the teaching activity itself that causes “old style” merit pay to fail repeatedly.
When Murnane and Cohen interviewed teachers and administrators, they learned that when teachers got poor ratings, they didn’t work harder; they “responded by working less hard.” If teachers felt their evaluation was unfair, they reacted angrily. Some of the principals who were interviewed said that before the introduction of merit pay, they often gave higher ratings than the teachers deserved because teachers would try to live up to their praise. With a higher rating (even if undeserved), teachers not only worked harder but were more open about seeking help for their problems.
Once merit pay was introduced, the principals no longer had the freedom to give so many high ratings because it would cost too much in bonuses. They came under pressure from budget-conscious superintendents to limit the number of high ratings, so they had to make choices. As one principal said to the authors, “Merit pay turns my job from being a coach into being a referee.” As the principal became a judge rather than a coach, teachers viewed him differently, with less trust. Merit pay changed the role of the principal in some respects by removing one of his motivational tools. He has to rely on his teachers to do their best and be highly motivated, but every time the principal gives a rating that is less than the teacher expected, he undermines the teacher’s morale. The primary reasons that school districts decided to eliminate merit pay were the demoralization of teachers and the problems of administering the program fairly.
The authors looked for districts where merit pay had survived. They found only a few. Not one of them was an urban district. Not one was a once-troubled district that successfully used merit pay to improve its performance. Most were small homogeneous districts that offered very small amounts of money. The few districts where merit pay had survived were notable for high salaries and good working conditions. But of greater importance was that the merit pay was extra pay for extra work, not bonuses for higher test scores or for something indefinable that the principal liked. The districts with merit pay also pursued a strategy of “make everyone feel special,” because almost every teacher received some monetary award, and the differences among them were not large. That eased tensions among teachers.
In other districts with long-lasting merit pay plans, the conditions for winning merit pay were so onerous and the rewards so small that no one cared much who got it. In all of the districts with merit pay, it was understood that no one would find out who won the extra money or how much they won. Another reason for the longevity of the merit pay plans is that all of them were designed by teachers to make them acceptable, non-divisive, and relatively unobtrusive.
I have delved into the Murnane and Cohen paper in detail because I think it sheds light on the dilemmas of merit pay today. When they were writing, in the mid-1980s, there was very little interest in the idea of pay for performance and very few districts where any form of merit pay existed.
What they learned then remains pertinent.
Although many districts had tried test-based performance pay in the past, not a single school district in the United States at that time awarded bonuses based on student test scores.
Merit pay turned out not to motivate teachers. In fact, it typically caused resentment and dissension among teachers who did not get merit pay.
Merit pay was not associated with improved student performance.
Where merit pay survived, it was connected to the extra work that teachers do, it was inconspicuous, and the awards were kept secret to avoid causing bad feelings.
Certainly, there are lessons for us today in this not-so-distant history. Only a generation ago, educators, school board members, and the general public abandoned the idea that bonuses might motivate teachers, yet today we have a reform movement that demands what has never worked in the past. Not only has the U.S. Department of Education committed to spend $1 billion of taxpayer money to support test-based bonuses (“piece-rate compensation,” to use Murnane and Cohen’s term), but many states have passed programs committing hundreds of millions of dollars for state-based merit pay programs at a time of budget cuts.
The goal of current merit pay plans is simple: the lure of a bonus is supposed to motivate teachers to work harder and therefore to get students to produce higher test scores. The goal is also to get the “effective” teachers to remain in the classroom and to attract new teachers who are motivated by the chance to make more money than a standard teacher’s salary.
It is not clear what the connection is between teachers working harder and students getting higher test scores. I recall hearing the late Albert Shanker say at a public meeting many years ago, “Let me get this clear. If you offer teachers more money, the students will work harder and get better grades?”
But that is in fact the main theory of action today behind merit pay. Offer bonuses to teachers, they will be motivated to get the money, they will teach more and better, and students will get higher test scores. Of course teachers want to be paid more, but they don’t want to compete with one another for bonuses tied to test scores. Some reformers believe that merit pay will lead to a restructuring of the teaching profession, making it more attractive to people who were in the top third of their graduating class at university; they will be attracted to teaching because of the opportunity to make more money. But frankly, anyone who becomes a teacher with the expectation of making big money is not making a wise career choice. He or she would be better off in another field.
The National Center on Performance Incentives at Vanderbilt University carried out the most definitive study of merit pay. Some economists speculated that the reason merit pay had never gotten good results in the past was that the reward was so small. So the Vanderbilt study offered a bonus of $15,000 for
teachers whose students got higher math scores. Over three years, the researchers closely observed the performance of teachers in a control group and an experimental group in the Nashville public schools. In 2010, the results were released: the study found no significant difference in the test scores of the students taught by teachers in the two groups. Presumably, both groups of teachers were doing the best they could.2
Within days of the release of the Vanderbilt study, the U.S. Department of Education released the first $500 million for the Teacher Incentive Fund to encourage more districts to try merit pay.
At the same time that the Nashville experiment was initiated, New York City embarked on its version of merit pay. After negotiation with the teachers’ union, the city’s Department of Education announced that two hundred schools would receive school-wide bonuses if test scores went up. A committee at each school would decide how to divide the bonus and whether to split it up among teachers only or among the entire school staff. The city paid out $56 million in bonuses but at the end of three years abandoned the experiment. A study by the Rand Corporation documented the failure of the New York City plan. It did not raise test scores in reading or mathematics. It did not increase teacher satisfaction. It was a waste of public money.3
Undaunted by the failure of the school-wide bonus plan, Mayor Michael Bloomberg announced in 2011 that he wanted to initiate a new merit pay plan, offering $20,000 to individual teachers who were able to raise test scores.4 This plan closely paralleled the one that failed in Nashville.
Meanwhile, a four-year study of the Teacher Advancement Program in Chicago concluded that the schools offering merit pay showed a somewhat higher retention rate of teachers but no significant impact on student test scores.5 Merit pay didn’t work in Texas, either. There the program got poor results in a pilot test, but the state legislature responded to this failure by expanding the program to $200 million a year (the funding was slashed in 2011 when the state legislature cut $5.4 billion from the public schools’ budget). In sum, all contemporary evidence on merit pay plans in American schools shows that they have had no effect on student test scores.6
The business expert W. Edwards Deming strongly opposed performance pay. Speaking of corporations, not schools, Deming warned against the corrosive effect of performance pay in the workplace. He said that it promotes rivalry and destroys teamwork. It ruins morale. It gets workers focused on themselves, not on what is good for the organization. It promotes short-term thinking and undermines long-term goals. He saw no value in using money to encourage employees to compete with one another. Deming contended that if management is doing its job, hiring the right employees, supporting them in their work, and keeping the system stable, most employees will perform well. As one biographer, Andrea Gabor, put it, “Deming believes that ranking employees is a cop-out for inadequate leadership.”7
The paradox of merit pay in education is that even if it did work, it would still fail. The more that teachers and schools are compelled to focus on raising test scores, the more they are likely to narrow the curriculum; the more likely that districts and schools will game the system to inflate scores; the more likely that there will be cheating; the more likely that teachers will seek to avoid low-scoring students. So, to the extent that schools are promised rewards for raising test scores (and punishments for not raising test scores), the quality of education will suffer. Schools will reduce the time for the arts, history, science, foreign languages, physical education, and anything else that is not tested. Or test everything, which would reduce instructional time even more than testing only basic skills. That way lies bad education, whether it is a narrow curriculum or a school that expends disproportionate amounts of time testing and preparing for testing.
Economists discount the importance of experience and education on grounds that they have no effect on test scores. They should also discount merit pay—for the same reason.
Merit pay is the idea that never works and never dies. Merit pay is faith-based policy. No matter how many times it fails, its advocates never give up. They are true believers. Next time it will work, they say. Next time we will get it right. No matter how many hundreds of millions of dollars are squandered with no benefit to anyone, there is always next time. Their belief in the magical power of money is unbounded. Their belief in the importance of evidence is not.
CHAPTER 13
Do Teachers Need Tenure and Seniority?
CLAIM Schools will improve if tenure and seniority are abolished.
REALITY There is no evidence for this claim.
One of the cardinal rules of the corporate reformist dogma is that teachers should have no job protections. Reformers believe that schools will be more successful if teachers serve at the will of their supervisors, who may dismiss them at any time for any reason. Reformers say that the business world works this way, and it is wrong for educators to have tenure or seniority. They say that these job protections make it impossible to fire bad teachers. Since they believe that firing bad teachers is the key to improving education, it follows that teachers should have no job protections. Reformers assume that if teachers can be easily fired, schools can quickly get rid of the low performers, and those who remain will work harder and produce higher test scores.
Reformers want to retain only effective teachers. The effective teacher is the one whose students get higher scores every year. If a first-year teacher can get students to produce higher test scores, then she is a better teacher than the teacher with fifteen years of experience. The reformers are convinced that experience, credentials, and degrees don’t matter. Bill Gates told the nation’s governors in 2010 that teachers should not be paid more for master’s degrees or years of experience because they do not increase test scores. Only “effectiveness,” that is, the ability to raise test scores, matters. The most effective teachers, he proposed, should be paid more to teach larger class sizes and work in the neediest schools.1
Thus, the traditional way of paying teachers is completely wrong, in the eyes of the reformers, because it rewards additional education and experience, not effectiveness.
Not only are ineffective teachers retained and promoted, the reformers say, but it is impossible to fire them because of tenure and seniority.
The reformers’ mission has therefore been to tie compensation to test scores (which they refer to as either “effectiveness” or “performance”) and to abolish tenure and seniority, thus making it easier to get rid of the deadbeat teachers, the incompetents who hang on forever, the ones who (as some governors charged) are paid “just for breathing” or “just for showing up.”
Michelle Rhee, Michael Bloomberg, and Joel Klein, among others, declared war on tenure and seniority in articles, public testimony before legislative bodies, and political advertisements. Michelle Rhee campaigned in state after state against tenure and seniority, and her organization StudentsFirst funded politicians who agreed. Their arguments received a warm reception in conservative states in the South, the West, and the Midwest. The Gates Foundation helped establish new organizations of young teachers with names like Teach Plus and Educators 4 Excellence, who publicly opposed tenure and seniority in legislative hearings. Other reform groups like Democrats for Education Reform, Education Reform Now, and Stand for Children have lobbied against job protections for teachers and against the seniority principle that they call LIFO (or “last in, first out”).
The reformers believe that scores will go up if it is easy to fire teachers and if unions are weakened.
But is this true?
No. The only test scores that can be used comparatively are those of the National Assessment of Educational Progress, because it is a no-stakes test. No one knows who will take it, no one knows what will be on the test, no student takes the full test, and the results are not reported for individuals or for schools. There is no way to prepare for NAEP, so there is no test prep. There are no rewards or punishments attached to it, so there is no reason to cheat, to teach to the test, or to game the
system.
So, let’s examine the issues at hand using NAEP scores as a measure. The states that consistently have the highest test scores are Massachusetts, New Jersey, and Connecticut. Consistently ranking at the bottom are states in the South and the District of Columbia. The highest-ranking states have strong teachers’ unions and until recently had strong tenure protections for teachers. The lowest-ranking states do not have strong teachers’ unions, and their teachers have few or no job protections. There seems to be no correlation between having a strong union and having low test scores; if anything, it appears that the states with the strongest unions have the highest test scores. The lowest-performing states have one thing in common, and that is high poverty. The District of Columbia has a strong union and high poverty; it also has intense racial isolation in its schools. It has very low test scores. Most of the cities that rank at the very bottom on NAEP have teachers’ unions, and they have two things in common: high poverty and racial isolation.
Having unionized teachers does not get in the way of high academic achievement, nor, in the case of the District of Columbia (and other high-poverty urban districts), does it guarantee high achievement. Conversely, the states where unions are weak or nonexistent have low achievement. Eliminating unions does not produce higher achievement, better teachers, or even higher test scores. Eliminating unions silences the most powerful advocate for public education in every state. It assures that there will be no one at the table to object when the governor or legislature wants to cut the budget for public schools. The union’s main role is to advocate for better working conditions and better compensation for its members. Better working conditions translate into better learning conditions for students, such as reduced class size and more resources for the schools. Better compensation attracts and retains teachers, which reduces teacher attrition. That too benefits students.2
Reign of Error: The Hoax of the Privatization Movement and the Danger to America's Public Schools Page 15