A better way to grade teachers: Grading on how teachers promote student learning rather than test scores

SALT LAKE CITY — Nikhil Goyal says the learning stopped in third grade when his class began to prepare for state-mandated math and English tests. Worksheets replaced science projects and "all the fun and joy in it just evaporated." Goyal, now 17, says it was his own love of learning that drove him to write and publish his first book, "One Size Does Not Fit All: A Student's Assessment of School."

Goyal aims to revolutionize education in the U.S., a goal that includes developing holistic systems of evaluation for students as well as teachers. He says the current system overemphasizes student test score gains and fails to help teachers and students grow. For example, he says that instead of improving their own work and that of their students, some of his teachers instruct kids to strive for low scores each September, to create the appearance of academic growth when they take the test again in June.

Policymakers are beginning to recognize what Goyal and 50 million public schoolkids nationwide already know: Good teaching goes far beyond getting children to fill in answer bubbles. Over the next few years, school systems in cities across the U.S., including Memphis, Chicago, New York, Pittsburgh and Denver, will phase in teacher evaluations that combine several measures to give a better picture of teacher success.

This approach is supported by the work of Harvard professor Thomas Kane, director of the Measure of Effective Teaching project. Since 2009, Kane and his research team have tested various methods of teacher evaluation to determine how best to fairly and accurately measure a teacher's effects on student success in general, regardless of quirks of any one student's performance on any one test.

Some results have been surprising. For instance, student achievement is not highly correlated to a teacher's level of education or scores on content knowledge tests. Rather, student success depends on whether a teacher demonstrates and promotes classroom behaviors that support student learning. According to preliminary findings released in January, the best way to predict whether a teacher's future students will make academic gains is to combine test data from last year's students with two different sources of information on teacher and student behavior in the classroom: rigorous classroom observations and student surveys. Teachers who scored well on all three measures were not simply "teaching to the test." Rather, this combination accurately predicted student growth on outside assessments of writing ability and conceptual mathematics, as well as on state tests. The MET project final report will be released in January.

Rigorous observation

According to preliminary findings of the MET project, in order for classroom observation to be rigorous and accurate, teachers must have clear, specific expectations for teacher and student behavior in the classroom. To improve accuracy, teachers must be observed multiple times each year by different trained observers, including people with no personal relationship to the teacher.

In Memphis, rigorous classroom observation is based on a list of expected teacher behaviors called the Memphis City Schools Teaching and Learning Framework. The framework, 17 pages in all, describes behaviors for five levels of teaching. For example, to achieve the highest ratings, teachers must demonstrate, that they "use classroom space to enrich learning of current or recent content" and that they provide opportunities for students at all levels to "explain concepts to each other with the support of teacher facilitation." Memphis teachers can expect to be observed at least four times per school year. The results count for 40 percent of each teacher's overall evaluation.

Student surveys

Memphis City Schools will also use student surveys to evaluate teachers, though the surveys only count for 5 percent of the total evaluation.