High School Teachers and College-Relevant Skill Production: Evidence from ACT Score Value Added

• In Event: Teacher Value Added and Long-Run Student Outcomes

Abstract

The transition from high school to college is crucial for children’s later life outcomes, yet little is known about the impacts of teachers on children’s skill development during the final years of high school. I provide the first estimates of teacher value added on college admissions test scores, using administrative data from North Carolina to estimate the impacts of high school teachers on students’ ACT performance. My identification strategy leverages quasi-random assignment of teachers within high schools, conditional on course choice, and I employ standard empirical Bayes methods from the value added literature (Chetty et al. 2014). I find that assignment to an English teacher with value added one standard deviation above the mean increases English and reading ACT scores by 0.06 standard deviations. Assignment to a comparably skilled math teacher increases math ACT scores by 0.08 standard deviations, approximately 0.5 points on the 36-point ACT scale.

I next estimate the long-run impacts of ACT score value added on on-time college enrollment using a nested logit model in which students choose between 2-year colleges, 4-year colleges, and foregoing college attendance. Both math ACT score value added and English ACT score value added increase 4-year college enrollment and enrollment in selective colleges, conditional on 4-year enrollment. Assignment to a math teacher with ACT score value added one standard deviation above the mean increases on-time enrollment in 4-year colleges by approximately 1.88 percentage points, 7% relative to the full sample on-time enrollment rate, while assignment to a similarly skilled English teacher increases 4-year college enrollment by 0.08 percentage points. Increases in 4-year college enrollment are offset by declines in 2-year college enrollment, relative to the outside option of foregoing college. Assignment to math and English teachers with high ACT score value added increases enrollment in selective colleges with high expected completion rates, conditional on 4-year college enrollment.

To understand whether the impacts of ACT score value added on college enrollment translate to improved college performance, I estimate the impacts of ACT score value added on persistence (measured by dropout during or after freshman year), freshman year GPA and accumulated credits, and graduation within 5 years of initial enrollment. I employ a selection correction procedure from Lee (1983) to disentangle the direct effects of ACT score value added on academic performance from indirect effects through selection into different college enrollment choices. I find that students exposed to math and English teachers with high ACT score value added earn higher freshman college GPAs and credits, are less likely to drop out during or after freshman year, and are more likely to graduate within five years of initial college enrollment. Effects of math and English teacher value added are similar in magnitude. In particular, both math and English teacher value added reduce freshman dropout rates by roughly 0.7 percentage points or 8% of the full sample mean. My results suggest that high school teachers have significant scope to influence the accumulation of college-relevant cognitive skills.

Author

• Hema Shah, Duke University
  Presenting Author