What percentage of New York City’s teachers are performing at an unsatisfactory level? Did anyone guess 2.3 percent? That’s how many were rated unsatisfactory by the school system in 2009-201 0— and it actually represents an enormous uptick (up from 0.89 percent) in "unsatisfactory" ratings because of the city’s emphasis on improving the system.

How do we square such low rates of teacher failure with the fact that, despite real improvements to the system, students in New York City's public schools perform poorly in large numbers? Simple. The current evaluation system depends very little on answering the one question we care about most: Are students learning in a teacher’s classroom? Incorporating analysis of student test scores helps focus evaluations on answering that essential question.

Test scores are important because they’re objective measures of the schooling outcome. It’s appropriate to emphasize student achievement on math and reading tests because these are the building blocks for success, and far too few students attending public schools today adequately possess these basic skills. Developing new tests and the right methods for analyzing them can be costly. But their potential contribution to improving teacher quality — the single most important school-based factor for fostering student learning — far outweighs the upfront cost.

Of course, test-score analysis can’t tell us everything we want to know about a teacher’s performance. Using it in isolation to evaluate teachers creates bad incentives and can miss a great deal of what makes a teacher effective. But research shows that evaluations of a teacher’s contribution to her student’s test scores this year is a far better predictor of how much her future students will learn than are the factors prioritized by the current system: years of experience and possession of advanced degrees. Failing to utilize such important and accessible information about a teacher’s effectiveness is scandalous.

Standardized tests are imperfect measures of student achievement, and the statistical analyses that utilize such tests are imperfect tools for evaluating teachers. But despite their limitations, standardized tests provide important information about teacher quality that we should use to improve our terribly flawed system for evaluating teachers. New York City’s movement toward increased use of test scores to evaluate teachers is a step in the right direction.