The debate over linking teacher pay to student test scores that ignited on Capitol Hill last month underscores the growing momentum鈥攁nd continued controversy鈥攂ehind tying what teachers earn to what students learn.
Both the National Education Association and the American Federation of Teachers came out swinging against language in a draft bill for reauthorizing the No Child Left Behind Act that would encourage districts to experiment with performance-based pay for teachers.
But across the nation, experimentation with such efforts is mounting, as states and districts start to crack open the traditional salary schedule by providing teacher bonuses based at least in part on student test-score gains.
At least half a dozen states have statewide or pilot programs that provide financial incentives to teachers based on achievement growth at the classroom or school level. And hundreds of districts are experimenting with such programs, including Denver, Houston, and Nashville, Tenn. The U.S. Department of Education has spent nearly $100 million to promote the idea through the federal Teacher Incentive Fund, which supports the development and implementation of performance-based pay in high-need districts. With the exception of Denver, though, few districts have entirely eliminated pay increases based solely on years of experience and course credits.
鈥淓ach year, it seems like there鈥檚 more going on than the previous year,鈥� said Allan R. Odden, a professor of educational administration at the University of Wisconsin-Madison who studies teacher compensation, 鈥渟o the momentum is building.鈥�
Failures of 鈥�80s
The past is littered, however, with largely short-lived and unsuccessful efforts to redesign teacher-pay systems.
During the 1980s, states experimented with merit-pay plans that tied teachers鈥� salaries, in part, to evaluations of their performance, and with career ladders that paid some teachers more for taking on extra roles and responsibilities. But teachers often complained that the evaluations were too subjective, and that the limited pots of money available for such programs encouraged unhealthy competition between colleagues. And when tight budget times came, such initiatives were often the first to go.
鈥淚 think we鈥檙e really at a very critical juncture,鈥� said William J. Slotnik, who has provided technical help to a number of districts working on pay-for-performance plans, 鈥渂ecause we鈥檙e now 25 years beyond the failed merit-pay experiments of the early 1980s. And if we replicate the same mistakes that burdened that movement, we鈥檙e going to lose a generation of compensation reform.鈥�
Research about the effectiveness of performance-based pay is scarce. A 2007 research synthesis by the federally financed Consortium for Policy Research in Education concluded that such plans 鈥渉ave met with some, but limited, initial success. Evidence of a substantial positive impact on either student achievement or teacher performance is lacking, and teachers report a wide variety of both positive and negative reactions to local plans.鈥�
More recent studies have found a generally positive relationship between financial incentives for teachers and improved student achievement.
鈥淲e don鈥檛 think the literature is sophisticated enough to say, 鈥楾his is the right way to do it,鈥� 鈥� said Michael J. Podgursky, a professor of economics at the University of Missouri-Columbia, who conducted a recent synthesis of such studies with Matthew G. Springer, the director of the National Center on Performance Incentives at Vanderbilt University. 鈥淏ut we think it鈥檚 strong enough, at this point, to say districts and states ought to go out there and seriously think about experimenting and attempting to innovate in this area.鈥�
The surge of interest in such experimentation is bolstered by research findings that some teachers are far more effective than others in raising student achievement. Compensation changes that reward talented teachers so that they stay in the profession, and encourage ineffective ones to improve or leave, potentially could have a large impact on student learning.
鈥楴o Clear Consensus鈥�
But despite a growing consensus that compensation systems for teachers need to be altered, said Brad Jupp, a senior academic-policy adviser to the superintendent of the Denver public schools and one of the creators of the district鈥檚 performance-based pay plan, 鈥渢here鈥檚 no clear consensus about what the best way to pay people, other than the single-salary schedule, is. We鈥檙e in that period of time after the old paradigm has gone and before the new paradigm has finally evolved.鈥�
Under the single-salary schedule, the long-dominant approach, teachers are paid based on uniform pay steps that reward years of experience and education coursework completed.
For states and districts willing to wade into an overhaul of how they pay teachers, experts offer a number of pointers based on experience. For starters, they suggest, the basic salary and benefits package needs to be competitive before policymakers embark on a performance-pay plan. Experts also suggest the rewards should be substantial鈥攕uch as 5 percent of base pay鈥攊f they鈥檙e going to motivate teachers.
Stable and adequate funding is also a necessity鈥攁nd one of the hardest issues for school systems, given the sizable costs of such programs. In 2005, the Denver school board persuaded voters in the 73,400-student district to raise their taxes by $25 million a year to support the ProComp Plan, which pays teachers more if they improve student achievement, acquire and demonstrate new knowledge and skills, choose to work in hard-to-staff schools and positions, and receive satisfactory job evaluations.
In contrast, many of the programs launched under the Teacher Incentive Fund are running almost entirely on federal money, with little guarantee they can be sustained without that aid.
Pay increments also need to be open to all teachers, not just those in state-tested subjects or an arbitrary percentage of the teaching force, advised a blue-ribbon panel of teachers, the TeachersSolution group, that called last year for overhauling how teachers are paid. (鈥淭eacher Panel Calls for Overhaul of Pay,鈥� April 11, 2007.)
In addition, compensation changes need the early engagement and broad public buy-in of teachers and their unions, people who have studied the issue say.
鈥淭oo many times in the past, these programs have been imposed, and you can鈥檛 force people to do it,鈥� said Lewis C. Solmon, the president of the Santa Monica, Calif.-based National Institute for Excellence in Teaching, which last year issued recommendations on successful performance-compensation system for educators.
Union concerns about performance-pay policies have been evident in the response to the recent NCLB-reauthorization draft, released by leaders of the House Education and Labor Committee. (鈥淯nions Assail Teacher Ideas in NCLB Draft,鈥� Sept. 19, 2007.)
In a Sept. 20 letter to U.S. Rep. George Miller, the California Democrat who chairs the committee, AFT President Edward J. McElroy reiterated his union鈥檚 opposition to having the federal government 鈥渕andate鈥� the use of student test scores to evaluate teachers.
鈥淭he AFT does believe that the decision to use or not use student test scores as part of a teacher-evaluation system should be made at the local level by the district officials and teachers who are directly affected by and most knowledgeable about the differential compensation plans that will work in their schools,鈥� he wrote. 鈥淭he AFT believes that compensation is a mandatory matter of collective bargaining subject to state and local, not federal, law.鈥�
The Denver plan was crafted in collaboration with the Denver Classroom Teachers Association, an NEA affiliate. Philadelphia, in contrast, had to scale back its pilot compensation program this school year to a handful of charter schools because it could not reach an agreement with its local AFT affiliate about how to implement the program in district-operated public schools.
Denver also permitted veteran teachers to opt into the program, rather than forcing them to participate.
鈥淚 never felt pressured to join it,鈥� said Elizabeth G. Douma, a humanities facilitator at the 650-student Hill Campus of Arts and Sciences, a Denver middle school. 鈥淚 can鈥檛 say enough about the fact that the district has let veteran teachers enter at their own pace, and really given us a lot of advice about where to enter.鈥�
As of this fall, almost half of Denver鈥檚 teachers, including Ms. Douma, are taking part in ProComp.
Experts say there鈥檚 no such thing as 鈥渙vercommunication鈥� when it comes to explaining performance-pay plans to teachers and making their provisions clear. Denver鈥檚 ProComp has its own Web site and a salary calculator so teachers can determine the payout they鈥檒l receive for various accomplishments.
Equally important, such plans need to rest on a strong measurement system for evaluating teachers, particularly if they hope to gain teachers鈥� trust and be seen as fair. Both national teachers鈥� unions remain opposed to financial incentives tied solely to student test scores, such as Houston鈥檚 ASPIRE Awards program.
鈥業t Has Pushed Staff鈥�
In Denver, teachers can earn annual salary increases for student growth on both state tests and on classroom objectives teachers themselves set in collaboration with their principals. An evaluation of Denver鈥檚 pilot program found students whose teachers set higher-quality objectives achieved higher scores on state tests than did pupils whose teachers set lower-quality learning goals.
鈥淚 feel it has pushed staff to look at their own practice and how it impacts student growth,鈥� Marcia W. Cornejo, a school psychologist and social worker at Swansea Elementary School, said of ProComp. The 17-year veteran, who had been at the top of the salary ladder under the traditional pay schedule, received an extra $10,000 last year through the program.
The good news, said Mr. Podgursky of the University of Missouri, is that states and districts are rapidly developing data systems that permit them to estimate teachers鈥� effectiveness based on their students鈥� test-score gains over time. Much debate remains about the accuracy of those measures, particularly when it comes to rewarding individual teachers. But Mr. Podgursky also argued that states and districts don鈥檛 have to rely on a single measure of teacher performance. Evaluations by principals can also be a reliable guide for identifying high- and low-performing teachers.
鈥淎ny single measure has faults,鈥� Mr. Podgursky said, 鈥渂ut by combining a set of independent measures, you get a better fix on true effectiveness.鈥�
One of the most widely used innovations in teacher compensation鈥攖he Teacher Advancement Program, or TAP, started in 1999 by the Milken Family Foundation鈥攑rovides bonuses to teachers who increase students鈥� academic growth and who demonstrate their skills through classroom evaluations that are conducted four to six times a year by multiple evaluators whom TAP trains and certifies.
The first broad evaluation of the program, released this year by the National Institute for Excellence in Teaching, which operates TAP, found that teachers in schools that participate in the program are more likely to significantly raise student achievement than similar teachers in other public schools.
Not Pay Alone
But Mr. Solmon, the president of the institute, emphasized that TAP is a comprehensive program that includes more than performance-based pay. It also gives qualified teachers opportunities to take on more responsibility and get paid for doing so. And it provides teachers with school-based professional development during the day to help them develop their knowledge and skills. Critics argue that performance-based pay plans are unlikely to raise student achievement without a clear path for teachers to actually improve their skills.
Indeed, one of the key lessons coming out of the research is that financial incentives alone are probably not enough to dramatically change the teacher-talent pool or improve student learning. Studies have found that teachers鈥� working conditions鈥攊ncluding school principals鈥� leadership, teachers鈥� opportunities for advancement, the availability of materials and resources, student behavior and discipline, and the chance for teachers to influence decisions and work with others鈥攁re strongly linked to satisfaction and turnover.
One of the biggest mistakes of the past, said Mr. Slotnik, the executive director of the Boston-based Community Training and Assistance Center, which provides help to a number of performance-pay efforts around the country, was that changes in compensation policy were viewed as a 鈥渢ack-on,鈥� separate from the other factors that lead schools to be effective.
