A recent study from Brown University’s Matthew A. Kraft and John P. Papay and Harvard’s Olivia L. Chi uses nine years of administrative data from Charlotte-Mecklenburg Schools in North Carolina to examine teacher improvement through the lens of principal evaluations.
Although teachers were evaluated on eight domains (e.g., Management of Instructional Time, Management of Student Behavior), it’s not clear that these domains actually capture different dimensions of teacher quality. Nor is there any evidence that teachers improve more quickly in some domains than others. So for simplicity’s sake, the bulk of the paper relies on a single measure of teacher quality, which is based on teachers’ ratings across all eight domains.
Overall, the authors find that “new teachers make large and rapid improvements in their instructional practices throughout their first ten years on the job,” rather than plateauing after three to five years, as some early research suggested. For example, between the first and tenth years on the job, the average teacher moves from the 31st to the 62nd percentile of subjective (i.e., principal-assessed) performance (though it’s worth noting that most teachers still get favorable evaluations).
Yet the rate and duration of teachers’ improvement also varies significantly across and within settings. For example, middle school teachers seem to stop improving after three years on the job, while both elementary and high school teachers continue to improve. And somewhat more encouragingly, there is a significant negative correlation between teachers’ initial performance and their rate of improvement (meaning weaker teachers improve more quickly and should perhaps be given a few years to improve rather than hastily dismissed).
Notably, although the authors find “little evidence of differences in the improvement profiles among teachers in tested vs. non-tested grades,” principals do rate teachers in high-stakes classrooms slightly higher than their peers in low-stakes classrooms—a difference the authors attribute in part to the fact that receiving a lower performance rating modestly increases a teacher’s odds of being reassigned to a non-tested grade, though there’s no evidence that high-performing teachers in low-stakes classrooms are reassigned to high-stakes classrooms.
According to the authors, the findings generally “underscore the potential of human capital investments in the teacher labor force, and the perils of relying on a revolving door of inexperienced teachers to staff schools.” And yet, some high-performing charter schools seem to have just such a door (though not necessarily because they want one). So one way or another, there must be a bit more to the story.
SOURCE: Matthew A. Kraft, John P. Papay, and Olivia L. Chi, “Teacher Skill Development: Evidence from Performance Ratings by Principals” (Annenberg Institute at Brown University, 2019).
