Magazine article Policy Brief Series (Hamilton Project)

Identifying Effective Teachers Using Performance on the Job

Magazine article Policy Brief Series (Hamilton Project)

Identifying Effective Teachers Using Performance on the Job

Article excerpt

Public education ultimately succeeds or fails based on the talent and skills of America's 3.1 million teachers in elementary and secondary schools. Everything else-educational standards, testing, school buildings, and school and district leadership-is background, intended to support the crucial interactions between teachers and their students.

Traditionally, policymakers have tried to improve the effectiveness of the teacher workforce by raising certification requirements. Research shows, however, that these credentials have little to do with teaching excellence, as measured by student performance. Once teachers are hired, school districts do very little additional screening and commonly award tenure after two or three years, regardless of teachers' performance. Moreover, the most effective teachers generally receive no incentives to work in the poorest districts. These policies are particularly problematic because there is a large gap between the most effective and least effective teachers, and the most effective teachers are underrepresented in schools serving low-income youth.


In a paper for The Hamilton Project, Robert Gordon, Thomas J. Kane, and Douglas O. Staiger propose a new five-point reform that would address these difficulties. It would increase the pool of potential teachers, make it tougher to award tenure to those who perform least well, and reward effective teachers who are willing to work in schools serving large numbers of low-income, disadvantaged children.

A Five-Point Plan to Identify Effective Teachers

What Makes a Good Teacher?

Typical teacher credentialing systems rely heavily on specific coursework and test scores. The federal No Child Left Behind Act, for example, requires all teachers of core academic subjects to be "highly qualified." They must have a minimum of a bachelor's degree, full state licensure and certification (which generally requires a degree in education), and competence in the subject areas they will teach, demonstrated by completing academic coursework or passing standardized tests.

Yet a growing body of research suggests that many such paper credentials have little to do with whether a teacher is effective, as measured by student achievement. For example, Gordon, Kane, and Staiger studied some 150,000 Los Angeles students in grades three through five from 2000 to 2003 and found no statistically significant achievement differences between students assigned to certified teachers and students assigned to uncertified teachers.

Other recent studies similarly have found that differences in teaching quality between certified and uncertified teachers are small compared with the differences in teaching quality within each group. In other words, there are good teachers and poor teachers, regardless of their certification. Much more relevant to predicting long-term performance is performance in the first few years of teaching.

Gordon, Kane, and Staiger conclude that good and bad teachers can be identified after only a year or two in the classroom. In particular, they find that teachers' performance during their first two years on the job provides a lot of information about their likely effectiveness in year three. On average, students assigned to third-year teachers who performed poorly during their first two years (in the bottom quarter of all teachers) lose ground relative to other students, whereas students of third-year teachers who performed well (in the top quarter) gain ground. In fact, students assigned to the best quarter of teachers ended up about 10 percentile points ahead of students assigned to the worst quarter of teachers.

In short, a considerable body of evidence shows the following:

* The effectiveness of teachers varies widely, even after adjusting for student's baseline test performance and other characteristics.

* Even with only two years of student performance data, a district can learn a lot about which teachers are likely to generate large student learning gains and which are not. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.