I recently read an article in the current edition of Education Week about the debate regarding language in the reauthorization of the No Child Left Behind Act that would encourage experimenting with performance-based pay for teachers. While both the American Federation of Teachers (AFT) and the National Education Association (NEA) ardently oppose the proposed language, experimentation with performance-based pay is mounting across the nation as states and districts try to reform teacher compensation.

A main issue in designing performance-based pay in teacher compensation is how to evaluate teachers. Currently, performance-based pay schedules are at least partly based on test scores because they are a quantitative measure for student achievement. This is in contrast with the traditional teacher pay schedule, where teachers are paid based on uniform pay steps that reward years of experience and education coursework completed.

The article states that during the 1980s, the main concerns about performance-based pay schedules relate to evaluating teachers based on criteria that were too subjective and that the limited scope of these compensation programs promote unhealthy competition between teachers. Education policymakers and administrators are trying to avoid the same problems in designing current performance-based pay schedules.

Research on the success of performance-based pay schedules is scarce. The article states that a 2007 research synthesis by the federally financed Consortium for Policy Research in Education found the impact of performance-based pay on student achievement is limited. While recent studies have generally found a positive relationship between financial incentives for teachers and student achievement, there is no consensus on how to design performance-based pay schedules.

In my experience researching education policy, when it comes to evaluating teacher effectiveness, current evaluation methods and the collection of quantitative data can provide a higher quality estimate of teacher effectiveness than the data and evaluation methods used when many performance-based pay schedules failed in the 1980s. For example, the article maintains that states and districts have developed data systems that estimate teachers’ effectiveness based on their students’ test-score gains over time. Such use of value-added measures does have its shortcomings, but combining these measures with qualitative measures such as improvement in student behavior and student satisfaction can possibly provide a more holistic assessment of teacher performance.