Do the Most Effective Math and Science Teachers Leave?: Evidence on Turnover among Middle School Teachers in a Large Urban School District

Ruth Curran Neild, Johns Hopkins University
Vaughan Byrnes, Johns Hopkins University
Elizabeth Farley-Ripple, Johns Hopkins University

Using a merged teacher-student data set from a large urban district, we examine whether the most effective teachers are more likely to leave their school district or transfer within the district. We employ a multi-level model to see whether middle school math and science teachers who added the most value (in terms of fall to spring academic growth, measured by standardized tests) were more likely to leave or transfer. We find that teachers who remained in the district value were more effective with their students than those who left. Among teachers who remained in the district, value added did not predict whether a teacher would transfer schools.