This brief was developed using Microsoft Copilot and edited by Charlotte Sutcliffe, Duke undergraduate research assistant; for full text and references see
Rose, E. K., Schellenberg, J. T., & Shem-Tov, Y. (2022). The effects of teacher quality on adult criminal justice contact. Retrieved from https://www.nber.org/papers/w30274
Background:
This study investigates whether and how the quality of teachers in elementary and middle school affects students’ future interactions with the criminal justice system (CJC), including arrests, convictions, and incarceration. Using a dataset that links North Carolina public school records (1996-2013) with statewide administrative arrest records (2006-2020), the authors analyze approximately two million students and 40,000 teachers to estimate the long-term social impact of teacher quality.
Findings:
Teacher quality significantly affects students’ future criminal justice outcomes. One standard deviation difference in teacher overall effectiveness on students corresponds to an 11% change in the likelihood of a student being arrested and a 24% change in the likelihood of incarceration. Specifically, teachers who improve students’ behavioral outcomes - such as reducing suspensions, improving attendance, and preventing grade repetition - reduce the likelihood of future arrest by 2–4%.
The study shows that teacher effects on behavior and test scores are largely uncorrelated. This means that a teacher who excels at improving test scores may not necessarily help reduce disciplinary issues or improve attendance. Consequently, policies that rely solely on test score-based evaluations may overlook teachers who are instrumental in developing students’ non-cognitive skills - traits like perseverance, self-control, and social cooperation - that are important for avoiding criminal behavior. Teacher effects on behavior are more persistent over time than effects on test scores, which tend to fade. This suggests that non-cognitive skill development may be a key mechanism through which education reduces crime.
The authors also explore how these effects vary across student demographics and school environments. They find that teacher impacts on criminal justice outcomes are heterogeneous - meaning they differ across student groups (e.g., by race, gender, socioeconomic status) - but that teacher impacts on short-term behaviors are more consistent across these groups. This suggests that while some teachers may be particularly effective with certain student populations, the ability to improve behavioral outcomes is broadly transferable.
To assess policy implications, the study simulates teacher retention strategies. It finds that removing the bottom 5% of teachers based on their long-run impact on criminal justice and academic outcomes could significantly improve student outcomes. However, policies based only on short-run metrics like test scores would achieve only a fraction of these gains. The authors argue that incorporating behavioral outcomes into teacher evaluations would yield more socially beneficial results.
Takeaways:
This research highlights the critical role teachers play in shaping not just academic success but also long-term social outcomes. It calls for a reevaluation of teacher quality metrics to include behavioral impacts, which are more predictive of future criminal justice contact than test scores. This shift could inform more effective education and criminal justice policies aimed at reducing incarceration and promoting social mobility.