We propose a framework for value-added models that flexibly characterizes heterogeneous teacher productivity based on multidimensional student characteristics. We show that teacher effectiveness heavily depends on the specific attributes of their students. For example, the difference in value-added between well-matched and poorlymatched students for the average teacher is approximately 0.1 standard deviations in test scores. Notably, these matching effects are particularly pronounced among lowachieving students. In language arts, the standard deviation in teacher value-added is one-third larger for low-achieving students compared to high-achieving students.

---

27 October 2025

---

The Review of Economics and Statistics 2025

---

DOI: 10.1162/REST.a.1629

https://direct.mit.edu/rest/article/doi/10.1162/REST.a.1629/133692/The-Importance-of-Student-Teacher-Matching-A

This Journal article is published under the centre's Education programme.