The Sensitivity of Teacher Performance Ratings

Recent policy reforms have dramatically changed how educators are evaluated. To better differentiate teacher effectiveness, newly implemented teacher evaluation systems incorporate multiple teacher performance measures and performance rating categories. Yet, little evidence exists on the consequences of these system design features for the distribution of teacher effectiveness. Using data from the Measures of Effective Teaching (MET) study, we conduct simulation-based analyses that demonstrate the critical role that performance measure weights and ratings thresholds play in determining teacher proficiency rates. Further, our findings reveal that teacher proficiency rates depend crucially on classroom observations, which external MET raters scored substantially lower than do school administrators. This work aims to inform policymakers' decisions as they refine and possibly redesign teacher evaluation systems.