Association for Public Policy Analysis & Management

Teacher evaluation has become a central education policy initiative at the federal, state, and district level. While evaluation is partially used to facilitate transfer and dismissal of teachers as well as provide incentives designed to improve teacher retention, little prior research has examined how teacher performance ratings impact whether teachers move within or exit from the public schooling system. This study estimates the impact of receiving a higher rating under the recently implemented Tennessee teacher evaluation program on teacher transfer and exit decisions. The evaluation system uses classroom and school-level student growth scores and achievement data as well as multiple classroom observations to evaluate teacher effectiveness on a numerical rating score ranging from 100 to 500, which is then used to assign teachers to discrete performance categories from 1 (significantly below expectation) to 5 (significantly above expectation). Evaluation reports provided to teachers include the discrete rating but not the underlying score on the 100-500 scale, which makes it difficult for teachers to determine their closeness to the threshold. The structure of the teacher evaluation system in Tennessee is therefore highly suited for empirical identification of an effect through a sharp regression discontinuity design. We also present descriptive findings showing changes in mobility patterns of high and low value-added teachers before and after the implementation of the statewide evaluation system.

Using rich statewide administrative data, we find that though lower rated teachers tend to exit their school or the profession at substantially higher rates than those who are rated more effective, teachers who cross a rating threshold to become rated “above expectation” (the rating associated with tenure decisions) are slightly but significantly more likely to leave their school, particularly if they are situated in a school with a high percentage of economically disadvantaged students. We also find important differences in transfer and exit patterns based on whether teachers teach tested subject areas and their level of teaching experience.

Notably, teachers of untested subjects, for whom the school level achievement measure accounts for a large portion of their overall rating, exhibit the largest effect of being rated above or significantly above expectation. Untested teachers who cross the threshold for higher ratings are roughly 10 percent more likely to leave a high poverty school for another school within their district. Results are generally robust to alternative bandwidths. 

The results of this study suggest that it should not be presumed that the introduction of more rigorous evaluation systems that result in labeling teachers as effective or ineffective will improve the overall quality or equity of the distribution of the teacher labor force. Policymakers should pay particular attention to the potential for labels and the consequences to create incentives for more effective teachers to exit more challenging, high need schools.