Teacher evaluation systems that use in-class observations, particularly in high-stakes settings, are frequently understood as accountability systems intended as non-intrusive measures of teacher quality. Presumably, the evaluation system motivates teachers to improve their practice – an accountability mechanism – and provides actionable feedback for improvement – an information mechanism. No evidence exists, however, establishing the causal link between an evaluation program and daily teacher practices. Importantly, it is unknown how teachers may modify their practice in the time leading up to an unannounced in-class observation, or how they integrate feedback into their practice post-evaluation, a question that fundamentally changes the design and philosophy of teacher evaluation programs. We disentangle these two effects with a unique empirical strategy that exploits random variation in the timing of in-class observations in the Washington, D.C. teacher evaluation program IMPACT. Our key finding is that teachers work to improve during periods in which they are more likely to be observed, and they improve with subsequent evaluations. We interpret this as evidence that both mechanisms are at work, and as a result, policymakers should seriously consider both when designing teacher evaluation systems.
Enacting the Rubric: Teacher Improvements in Windows of High-Stakes Observation