View all Articles
Commentary By Marcus A. Winters

Can School Performance Be Measured Fairly?

Education Pre K-12

Major League Baseball teams use quantitative data to improve outcomes. So do local police forces all across the country. It would be foolish to believe that data can provide perfect information, or that data cannot be misused. But countless sectors have learned that there is information in quantitative measurement that can improve decision-making.

Public schools are finally entering the modern age. School systems across the nation are incorporating quantitative measures of a teacher’s effect on her student’s test scores into performance evaluations. These so-called Value Added Measures (V.A.M.) of teacher quality, based on the test scores, use a complicated statistical algorithm to estimate the teacher’s contribution to student performance.

Test-based evaluations provide information that improve a school system’s ability to distinguish between effective and ineffective teachers.

Critics point out that V.A.M. incorporates measurement error. Thus, it is certain that some teachers will receive scores that are either higher or lower than they deserve. The critics are correct.

But perfection is too high a standard. The relevant question is whether there is information in V.A.M. estimates that improve a school system’s ability to distinguish between effective and ineffective teachers. There undoubtedly is.

Improving upon the current evaluation system is a remarkably low bar. Research has discovered wide variation in teacher quality. Yet nearly all teachers receive satisfactory ratings under the current system. Further, research shows that the characteristics used to identify effective teachers — for instance, obtainment of a master’s degree — are unrelated to classroom effectiveness.

In contrast, V.A.M. can be used to predict a teacher’s future performance in the classroom. For instance, research by Dan Goldhaber and Michael Hansen shows that a teacher’s performance as measured by V.A.M. in her first two years in the classroom is a far better predictor of student learning than are conventional measures like a master’s degree.

But reformers shouldn’t turn a blind eye to V.A.M.’s imperfections. V.A.M. should never be used in isolation to make employment decisions. It must be coupled with meaningful, subjective evaluations by administrators.

Classroom observations can identify cases when a teacher’s V.A.M. score is too low. Similarly, V.A.M. analysis can raise red flags about teachers who appear to be doing well but whose students aren’t making academic progress. When both the subjective and quantitative assessments tell us that a teacher isn’t getting the job done, the school system should work to either remediate that teacher or remove him from the classroom.

We will never have a perfect evaluation system. But careful use of student test scores to evaluate teachers can drastically improve the quality of teachers in American public schools.

This piece originally appeared in The New York Times

This piece originally appeared in The New York Times