Reforming Teacher Pay

While some effective practices are emerging, even the best-designed performance pay plans can be undermined by SUSAN FREEMAN BURNS AND CATHERINE D. GARDNER

A recent Google search for information regarding performance pay in education produced 6.1 million results. This number should come as no surprise given the current level of interest in incentives as a popular reform option in public education.

Public school systems in New York City, Denver and Houston all offer compensation packages that reward teacher performance. Several states, most prominently Florida, Minnesota and Texas, have allocated hundreds of millions of dollars to performance pay. In 2009, the Obama administration designated a substantial portion of the American Recovery and Reinvestment Act for the development and implementation of teacher pay-for-performance programs.

Susan Burns and Catherine GarnerSusan Burns (right) is program manager and Catherine Garner is a policy and outreach analyst at the National Center on Performance Incentives in Nashville, Tenn.

The popularity of these programs lies in what some see as a link between well-designed performance pay plans and improved teacher effectiveness. Proponents reason that effective teachers should be rewarded. They also argue that rewards can attract more qualified applicants to teaching, especially in hard-to-staff subjects or schools. Finally, supporters believe pay-for-performance programs encourage less effective teachers to improve and will reduce teacher turnover.

However, clear challenges exist. Most notable is the difficulty of monitoring teacher performance. Even in states and districts using value-added measures, ensuring accurate student-teacher linkages and data systems are developed enough to make high-stakes personnel decisions can be difficult. Critics also contend that individual performance awards may negatively affect the collaboration among professionals that is essential to teaching, particularly if the award structure sets up teachers competing against one another for bonuses.

Another argument against performance pay programs involves the issue of multitasking. Given the multidimensional nature of teaching, how is it possible to implement a program that captures all of the important roles school personnel play? Focusing on too few dimensions could result in a narrowing of the curriculum, better known as teaching to the test.

One important but relatively understudied consideration is the design components of a pay-for-performance program. In a 2009 article published by the Organisation for Economic Co-operation and Development, Matthew Springer and Ryan Balch address the myriad of design options for performance pay plans. They highlight the complexity, outlining considerations such as the structure of incentives, whose performance should determine eligibility for awards (individual teachers, teams, whole schools), what the evaluation criteria and performance standard should be, and how frequently awards should be distributed.

Despite the relatively thin knowledge base around the design, implementation and operation of performance pay in education, a few key ideas are emerging as school districts put in place pay-for-performance plans and researchers evaluate those efforts. It’s possible to distill some lessons that have been learned from recent incentive pay programs in an effort to better inform superintendents and their school boards as they examine potential implementation.

As research progresses around performance pay, six key elements are emerging:

•  the importance of nesting a performance pay program in a district’s larger reform agenda and aligning it to key district priorities; 

•  the need for ongoing, effective communication about all aspects of the initiative; 

•  the need for stakeholder involvement at all levels of the program; 

•  the need for stable funding streams and meaningful awards; 

•  the importance of multiple, well-defined perform-ance measures centered on district goals and priorities; and 

•  the importance of sophisticated, well-managed and well-maintained data systems.

Nesting Incentives
Any pay-for-performance program should be one component of a larger district reform, with program elements well aligned to and complementary of school district priorities. Merely changing the way teachers are compensated will not improve student achievement outcomes in our schools. Instead, teacher compensation must be one element of a systemwide reform that may include supports to improve teacher quality, opportunities for teacher advancement, or recruitment and retention of highly effective teachers.

“Because pay-for-performance programs typically entail substantial investments,” says Harvard Professor Susan Moore Johnson in her book Redesigning Teacher Pay: A System for the Next Generation of Educators, “districts should consider deliberately integrating their approaches to pay reform so they provide clear and complementary incentives, which are aligned with their priorities for improvement.”

Minnesota’s Q Comp provides a fine example. Q Comp is a statewide reform program composed of five core elements: career ladder/advancement options, job-embedded professional development, teacher evaluation, perform-ance pay and an alternative salary schedule. Though the program is voluntary, each component must be included in a participating district’s broader plan focused on improving student achievement.

Similarly, the REACH initiative in the Austin Independent School District in Texas has four main goals: a quality teacher in every classroom, especially in the highest-needs schools; improved student learning at all schools and for all students; professional growth for teachers; and increased retention rates among teachers and principals.

As Johnson says in her book, a pay-for-performance program should “support, not substitute for, a school district’s strategy for improvement. … It is not an isolated bonus for a single success, but rather an integrated strategy for advancing a wide range of goals in the human capital system.”

Communicating Well
Effective communication is critical during development of a program, and even more so once it is implemented.

In a 2007 study of teacher attitudes toward pay for performance in Florida, Brian Jacob and Matthew Springer discovered only 49 percent and 39 percent of teachers had a clear understanding of the two pay-for-performance programs in their state — respectively, Special Teachers Are Rewarded, or STAR, and the Merit Award Program, or MAP. Yet most teachers also had strongly negative views of the programs and did not think the programs were beneficial, which may be a direct consequence of the design components being poorly understood among teachers who were directly affected by it.

In Austin, on the other hand, where communication about REACH was strong and a top priority for district officials, teachers were more supportive of that program than they were of pay for performance in general.

These examples underscore the abundance of misunderstanding surrounding pay for performance. They suggest the need for clear communication with teachers about the elements of the incentive pay program in which they are participating. Beyond design and implementation stages, communication must be ongoing, allowing feedback and revisions to the process when necessary.

Stakeholders’ Role
As the pay program is developed, involvement by key stakeholders will directly affect success. Historically, teacher associations have opposed pay-for-performance plans, but in recent years, both the American Federation of Teachers and the National Education Association have shown a willingness to relax their position when engaged early in the decision-making process.

In 2007, the AFT’s union in New York City negotiated a pay-for-performance program to operate in more than 200 Manhattan schools. At the time, AFT President Randi Weingarten spoke to the importance of the collaborative effort, saying, “Rather than being pilloried as an obstacle … we created a program that may promote the collaboration and respect that are necessary for great schools … We have taken a negative — individual merit pay — and come up with a positive alternative that makes it a plus for educators and kids.”

In addition to teacher associations, early engagement with several other stakeholders is essential. In particular, teachers and other school district officials, local business leaders, parent-teacher-organization representatives, local politicians and school board members should be included in early design and implementation.

While promoting collaboration and stakeholder engagement, consider one important caveat: Not everyone will agree on program goals and elements. Recognize and consider multiple perspectives and options. There should be a balance, however, between having all voices heard, thereby increasing potential buy-in for the program and designing a sound program. An example from Texas is illustrative.

The Texas Educator Excellence Grant program required plans be developed by multiple stakeholders at the school level. Best-practice guidelines were suggested but not required. As a result, most schools designed programs that distributed decidedly small awards equally across all school personnel regardless of their individual performance. It is no surprise the researchers found the program to have no meaningful impact on student achievement. While the goal for stakeholder input should be development of a strong plan, engaging in critical conversations should likewise be valued.

Tenuous Funding
The financial costs associated with operating pay-for-performance programs are substantial and often exceed district and state projections. At present, most perform-ance pay programs are funded by either federal grants or foundation support. These revenue streams are short-term at best. For a pay-for-performance program to be given the chance to be effective, school districts must carefully examine, document and plan ways to develop stable, long-term sources of revenue.

In a 2008 Center for Educator Compensation Reform publication, James Guthrie and Cynthia Prince warned that school systems that either underestimate the financial costs associated with these programs or miscalculate the financial exposure of these programs risk not only substantial financial losses but possible legal action and loss of credibility. Moreover, several empirical studies indicate that if teachers are not confident that funding is secure, their engagement with the program will wane.

Guthrie and Prince offer several strategies for sustaining programs, including redeploying existing resources, redirecting future expenditures, repackaging existing state and federal categorical funds, seeking additional public funds (funding for Denver’s ProComp was secured through a voter-approved $25 million mill levy), and seeking philanthropic or corporate support.

Measuring Teaching
Perhaps one of the most challenging aspects of developing a performance pay program is how the program measures teacher effectiveness. The vast majority of performance pay programs today rely solely on standardized achievement test scores as their primary measures. As has been the case for years, arguments abound as to whether test scores are a valid and reliable measure of performance. These arguments have intensified as greater stakes are placed on scores from standardized assessments.

Critics of pay for performance cite the unfairness of a single test being used as the measure of teacher effectiveness and subsequent bonuses. Moreover, according to a 2008 publication from the Center for Educator Compensation Reform, approximately 69 percent of teachers teach courses that are not covered by a standardized assessment. While sophisticated value-added measures hold promise for measuring teacher effectiveness, all of these criticisms and concerns point to the need to incorporate multiple measures of teacher effectiveness into a pay-for-performance program.

Alternative measures of teacher effectiveness do exist. Teachers participating in Denver’s ProComp and Austin’s REACH work with their principal to develop two individualized, data-based learning goals for their students. In New York City’s School Wide Performance Bonus Program, a school’s standing is determined, in part, on the basis of student attendance and student, parent and teacher perceptions of the school learning environment.

While education research, practice and policy continue to collaborate on the design and development of alternative measures of teacher effectiveness, including classroom observation protocols, there is much work to be done before reliable, established measures are readily available.

Data Quality
Once measures are clearly identified, a robust, well--managed data system is crucial to the success of the program. Most district- and state-operated information systems are simply insufficient to manage the data necessary to administer a teacher pay-for-performance program. According to research gathered by the Data Quality Campaign, a national organization supporting states in their efforts to use high-quality data to improve student achievement, data systems in only 11 states meet all 10 of what they define as essential elements of a data system for school districts.

The data obtained from district and state information systems also are often riddled with inaccuracies and errors that can wreak havoc on the operation of a performance pay program. In an analysis of data system quality, Battelle for Kids documented several common weaknesses, including too few data snapshots (data collected only once or twice a year does not accurately capture what is happening in schools), inaccurate course codes, errors with the unique student identification number (multiple students with the same number) and incorrect student-teacher linkages.

These types of errors have the potential to completely undermine even the most well-designed pay-for-perform-ance program. Both Battelle for Kids and the Data Quality Campaign offer several best-practice guidelines for data systems. Among the guidelines is the recommendation that systems should collect multiple data, including student course enrollment, teacher assignment data, principal assignment data and translation or association of course codes to state tests. In addition, systems should use secure accounts that validate user access. Finally, sophisticated audit systems must be implemented that allow administrators and teachers to verify the accuracy of the data.

Susan Burns is program manager at the National Center on Performance Incentives at Vanderbilt University in Nashville, Tenn. E-mail: susan.f.burns@vanderbilt.edu. Catherine Gardner is a policy and outreach analyst at NCPI.