Options for Studying Teacher Pay Reform Using Natural Experiments

How public school teachers are paid in the U.S. has been a controversial issue for many years. Critics of the traditional system, in which teachers are paid on the basis of years of experience and educational attainment, claim that it does not reward or promote good teaching as fairly as systems that tie pay to performance: having certain skills, or being willing to teach in high-need areas. Proponents of the traditional system argue that experience and education are important predictors of teacher performance, and that the simplicity, transparency and fairness of the system make it the logical choice. In an attempt to achieve the best of both worlds, educators and policymakers have devised numerous approaches to reforming the teacher payment system. Choosing between the various approaches is difficult because scientific evidence on their effectiveness is limited. This report presents findings and recommendations from a review of how secondary data could be used to study a variety of teacher incentive programs to address the question of whether teacher incentive programs improve student learning, either by making teachers more productive (productivity effort) or by attracting and retaining good teachers at higher rates (composition effect.) Following background information on attempts to study these programs, the preliminary steps of identifying and categorizing teacher incentive programs and the process of narrowing down the list to the most promising are discussed. Detailed profiles of the seven candidates selected for further study are provided. Cross-cutting issues of sample size adequacy and cost are identified, and a final section summarizes the findings and offers recommendations for next steps. (Contains 8 tables.) [This report was submitted to the US Department of Education, Institute of Education Sciences/National Commission on Excellence in Education by Mathematica Policy Research, Inc. It was prepared by Mathematica Policy Research, Inc. and Chesapeake Research Associates.]