May 4, 2010
House Education and Labor Committee HearingSupporting America's Educators:
The Importance of Quality Teachers and Leaders
Testimony by Marcus Winters, Senior Fellow, The Manhattan Institute
Chairman Miller, senior Republican Castle, members of the committee, I appreciate the opportunity to testify before you today on the important issue of teacher quality. This is an issue that I have studied and written about extensively as a senior fellow at the Manhattan Institute for Policy Research. However, I emphasize that the opinions that I express here today are my own.
The last decade and a half of empirical research has dramatically increased our understanding of teacher quality and the factors related to it. The findings of modern research strongly confirm what parents, teachers, and school administrators have always known: The quality of a child's teacher is the most important factor within a school's control that determines a student's learning in a given year. The best estimates indicate that the difference for a student being taught by a good or bad teacher amounts to about an additional grade level's worth of learning at the end of the school year.
Unfortunately, despite the substantial variation in teacher quality, the current system fails to distinguish between our best and worst teachers. Nearly all teachers are rated satisfactory or higher according to their official evaluations.
When the current system does distinguish between teachers it is according to two attributes that research consistently finds have little to no relationship to a teacher's performance in the classroom: the attainment of advanced degrees and years of experience.
Researchers consistently find no discernible relationship between whether or not a teacher has earned a Master's degree and the learning her students acquire in a given year. Further, the positive experience with alternative certification programs such as Teach for America that recruit motivated, bright individuals without education backgrounds to teach in low-performing public schools shows that great teachers need not have ever attended a single course in an education college.
The research evaluating the relationship between classroom experience and effectiveness is only slightly more encouraging. According to the research, the benefits from classroom experience plateau in about the third to fifth year. Of course, some teachers do get better over time. But some teachers don't improve, while others burn out and actually get worse over time.
In addition, whether an individual teacher is better at her job today than she was yesterday is insufficient for determining whether she is more effective than the teacher down the hall. Empirical studies consistently find that experience and other easily observed characteristics explain very little of the difference in teacher effectiveness.
That credentials and experience tell us so little about a teacher's effectiveness is disappointing because most school districts rely on those attributes alone to determine a teacher's salary. Teachers have responded to the incentives of their pay scale by pursuing unproductive advanced degrees. According to the National Center for Education Statistics, the percentage of public school teachers with a Master's degree or higher has increased from about 24 percent in 1961 to about 52 percent today.
Under the current system, whether or not a teacher is allowed to remain in the classroom is nearly entirely a function of how many years she has been employed there already. It is common for school systems to determine layoffs based entirely on seniority within the system. Those "first-in, last-out" layoff rules are now coming into play as states across the country are finding it necessary to reduce their teaching staffs during this time of fiscal strife. The result of basing layoffs on factors unrelated to classroom effectiveness will be that many wonderful young teachers will be let go and several poorly performing but more experienced teachers will remain in the classroom.
Further, in most school systems upwards of 95 percent or more of the teachers who remain in the classroom the three years or so required to become eligible for the job protections of tenure receive it. It's true that tenure only requires that a teacher cannot be fired from her position unless the school system first goes through a due process proceeding. However, in practice that due process is so burdensome and expensive that most administrators don't even bother with it. For instance, just 10 of New York City's 55,000 tenured teachers were fired for any reason in 2007.
A common argument made by tenure's defenders is that school systems have effectively weeded out many of the low performers by the third year. However, even if we were to believe that schools were capable of identifying and removing ineffective teachers so early in their careers, the practice of tenure still essentially assumes that anyone not shown to be incompetent by her third year will be effective in the classroom in her thirtieth.
Given the complexity of a teacher's job, it's not so surprising that basic attributes like experience and credentials explain so little of their effectiveness. Most of the qualities that differentiate a great teacher from a not-so-great teacher can't be collected in an administrative data set. Qualitative attributes, such as a teacher's patience, classroom management skills, and knack for presenting complex information clearly explain most of her influence on her students' learning.
Unfortunately, those are attributes that do not lend themselves to simple salary schedules and layoff policies. If we take the lessons of modern research seriously we have to conclude that today's system has its priorities backwards. A better system would measure a teacher's actual performance in the classroom and then reward the most effective teachers and remove the least effective ones accordingly.
The first step towards creating a better system is to improve teacher evaluations. Not all teachers in today's public schools are succeeding in the classroom, and that the current evaluation tools tell us otherwise makes them essentially useless. School districts should replace the current evaluation system with one based in part on a teacher's measurable influence on her students' standardized test scores. Data analysis is far from perfect, and it should certainly not be used in isolation to make employment decisions. But modern statistical techniques can raise red flags and thus help administrators distinguish between teachers whose students excel and teachers whose students languish or fail.
Once a school system has identified the best and worst teachers, it should act upon that information. States and districts should continue to experiment with different ways to tie some portion of a teacher's compensation to her performance in the classroom. Further, states should streamline the process for administrators to remove ineffective teachers after they have been identified.
The political hurdles to adopting such a reasonable system are daunting. Though we have heard some encouraging words from the American Federation of Teachers at the national level, local union affiliates continue to fight hard against meaningful change.
Consider New York's recent experience. When it appeared that New York City Mayor Michael Bloomberg was prepared to use test scores as part of evaluating teachers for tenure, the city's teachers union went to Albany and pushed through legislation making it illegal for any school system in the state to do so. Further, with an estimated 15,000 teacher layoffs on the horizon in New York due to stressed budgets, it is the state and city teachers unions that have stood strongest against proposed legislation that would replace the state's law requiring layoffs to occur according to seniority with legislation granting discretion to principals so that they can determine which teachers should remain in the classroom.
The unions and other defenders of the current system also frequently argue that it would be more productive to focus efforts on reducing class size rather than removing ineffective teachers. The argument for reducing class size depends on a single study from 79 public schools in Tennessee during the late 1980's. The study followed a high quality random assignment design and found some evidence that student learning was greater in smaller class size environments. However, when taken to scale, the results of class size reduction programs have been disappointing. For instance, a study by the Rand Corporation found that California's class size reduction program has had no influence on student proficiency.
It is essential to America's future to ensure that each of the nation's public school classrooms is staffed with an effective teacher. The current system's failure to accurately measure teacher quality, its emphasis on rewarding teachers for attributes unrelated to their effectiveness, and its powerful protections for even the worst teachers makes it incapable of achieving that goal. It is time for school systems to rethink the way that they evaluate, compensate, and hold accountable the public school teachers who are educating the nation's youth.
I look forward to answering your questions as this important discussion moves forward.