Almost three years ago, as many pay-for-performance initiatives for educators were in their formative stages and advocates were intensifying their campaign for expansion, we warned against moving forward faddishly, or mindlessly mimicking the mistakes associated with 1980s-style merit pay. (鈥淭eacher Pay for Performance: Another Fad or a Sound and Lasting Policy?鈥� Commentary, April 5, 2006.) We endorsed a gradual, incremental development of such compensation plans鈥� design, and urged rigorous independent experimentation and objective evaluation.
In the intervening years, growth in this area of education policy has been anything but incremental. Performance pay has become a wildly popular option, and appears to be poised for even more dramatic future expansion.
There are now widespread federal, state, and local initiatives, along with foundation endorsements, technical support, and financial inducements (more than $500 million allocated for the 2008-09 school year) to expand incentive pay. States as diverse as Florida, Minnesota, South Dakota, Tennessee, and Texas have enacted statewide policies promoting performance pay for educators. National performance-pay models, such as the Milken Family Foundation鈥檚 Teacher Advancement Program, or TAP, continue to expand, while creative district models have further evolved, including Denver鈥檚 widely publicized ProComp plan. New models also are emerging, such as New York City鈥檚 鈥渟choolwide performance bonus program鈥� supported by the United Federation of Teachers and the New York City Department of Education.
In some form or fashion, performance pay exists in approximately 10 percent of the nation鈥檚 school districts and affects at least 20 percent of K-12 teachers and students. Its expansion continues despite the sour aftertaste left from the failed incentive-pay experiences of the 1980s鈥攁nd perhaps in some part because of steadily increasing media attention.
Those who already are engaged in the operation of such plans, or are considering their adoption, should understand that there is still only the slenderest research base undergirding the effort. As yet, there are no rigorous empirical validations to show that U.S. performance-pay programs in education are linked to substantial and sustained successes, either in elevating student achievement or in accelerating the occupational attractiveness of education for a wider pool of able teacher candidates. In effect, while policy-system enthusiasm for the idea is building, the research-and-evaluation jury is still out on educator performance pay.
Even so, lessons can be learned from recent research results. These early findings and preliminary practical observations may provide little in the way of specifying a perfect performance-pay model, but they do suggest several important considerations in the design and implementation of such plans鈥攁s well as pitfalls to be avoided. Here is a brief summary of the knowns and unknowns about teacher performance pay, and what this knowledge may mean for future development of the concept:
What we know with a high degree of certainty:
鈥� An effective teacher can contribute substantially to student achievement, regardless of students鈥� innate abilities and home and neighborhood socioeconomic circumstances.
鈥� Sustained, multiyear contact with an effective teacher can materially mitigate students鈥� accumulated achievement deficits.
鈥� Currently, the distribution of identified effective teachers favors students from higher socioeconomic circumstances.
鈥� While current considerations in the determination of public school teacher pay鈥攕eniority and added academic credits鈥攈ave ameliorated past injustices and provided predictability and objectivity, they display only minimal relationships with elevated student academic achievement.
What we suspect is true from experience and observation, but do not now know with certainty:
鈥� Rewards hold the prospect of strongly shaping the effort of employees. But there is a possibility of dysfunctional goal-displacement鈥攚orking hard on the highly rewarded goals but slighting others equally important. Consequently, architects of performance-pay plans must make sure that the conditions being measured and rewarded (accurately and comprehensively) reflect the organization鈥檚 desired outcomes.
鈥� Incentive-program designs almost always necessitate trade-offs between desirable qualities, such as transparency on the one hand and accuracy of performance measurement on the other. The increased complexity associated with accurately measuring teacher and school effectiveness, for example, often is associated with diminished transparency.
鈥� Hastily conceived and overly simple plans often lead to mistakes and can engender low employee acceptance as a consequence. Hence, given the complexity of performance-award programs, it is important to initiate evaluative procedures and anticipate midcourse corrections from the outset.
鈥� Three obstacles on which contemporary performance-pay programs appear most frequently to founder are (1) inaccurate, incomplete, or unfair measurements of student achievement and other outcomes; (2) unclear or insufficient efforts to explain the program to important stakeholders and gain their commitment; and (3) inaccurate projections of possible financial exposure and a consequent inability to pay as promised or sustain funding for the program.
What we do not yet know:
鈥� The power of financial awards in promoting more-effective teaching and elevating student performance.
鈥� The effects of group awards relative to individual performance awards.
鈥� The preferable mix of financial and nonpecuniary awards.
鈥� The long-term effect of performance awards on the supply of effective teachers.
鈥� The consequences of offering higher pay for teachers in subject shortage areas and hard-to-staff schools.
鈥� The cost-effectiveness of performance incentives relative to alternative strategies for elevating academic achievement, such as class-size reduction, enhanced reliance on educational specialists, or intensified deployment of technology.
In light of these conditions, our message to educators and policymakers remains much the same as it was three years ago: a cautious endorsement of careful performance-pay-plan design and a strong plea for rigorous independent experimentation and objective evaluation. What may be needed most is the kind of pragmatic experimentation advocated by Jack V. Matson, an environmental engineer, in his 1996 book Innovate or Die. Matson recommends, as one element of innovation, an analytical process for eliminating unsuitable design options that he terms 鈥渋ntelligent fast failure.鈥�
Intelligent fast failure is not a goal, but an outcome from risking effort. Each experiment undertaken is carefully considered, with the goal being to determine the conditions necessary for success. Experiments are crafted to minimize downside risks and the time and resources needed to learn quickly from the outcomes. While ordinarily one would test ideas sequentially, starting with the best, with intelligent fast failure an institution experiments with multiple ideas simultaneously. As a result, it is possible to accelerate the learning process, compress failure time, and progress more rapidly toward resolution.
With today鈥檚 proliferation of performance-pay plans, it is important that educators systematically collect, document, and share widely what is being learned from such programs. Lessons passed on by practitioners and researchers represent assets of tremendous value to states and districts interested in designing and implementing performance-pay plans of their own. Similarly, empirically based assessments of the short- and long-term consequences of performance pay will allow both the policy and practitioner communities to more forward with greater certainty.
By harvesting contextually rich insights based on research, and by harnessing the potential of processes such as intelligent fast failure, we can move the performance-pay phenomenon further from being a passing fad and closer to being a sound and lasting policy.
