Panel Paper: The Effects of Early Math Coursework on College Readiness: Evidence from Targeted Middle School Math Acceleration

Friday, November 4, 2016 : 10:15 AM
Columbia 4 (Washington Hilton)

*Names in bold indicate Presenter

Joshua Goodman, Harvard University, Shaun M Dougherty, University of Connecticut, Darryl Hill, Wake County Public School System, Erica Litke, University of Delaware and Lindsay C. Page, University of Pittsburgh


Advanced math coursework can affect college and labor market outcomes but universal policies expanding access to such courses for all students have generally failed. A large literature, mostly relying on correlational estimates, have highlighted the potential importance of math, and algebra in particular, as an important predictor of success later in life. Through a partnership with the Wake County Public School System (WCPSS) in North Carolina, we present the first evidence on a targeted approach, where prior test scores were used to select qualified students into a college preparatory coursework track beginning in middle school. As one of the largest public school districts in the country and with a racially and economically diverse student enrollment, the implementation of this targeted policy allows for the estimation of the causal effects of being induced into advanced coursework in middle school. Rich administrative data, including student transcripts, and the implementation of the policy based on a cutoff on a composite measure of prior performance created the condition for a natural experiment and allowed for the estimation of the causal effect of the policy on subsequent coursetaking and performance. Regression discontinuity estimates show that at the margin of selection only one- fifth of those initially accelerated in seventh grade remain in advanced coursework by eleventh grade, with higher-income and female students persisting at noticeably higher rates. Importantly, acceleration substantially increases college readiness scores and students’ intentions to enroll in four-year colleges. The large sample size and robustness of the results to various specification checks suggest that we have high levels of power to detect small and precisely estimated effects. Our estimates suggest that targeted math acceleration is a promising approach to improving college readiness but alone is not sufficient to keep even the majority of initially qualified students on a college-preparatory track in mathematics through high school. These findings are relevant to local, and state-level policy makers in particular as they search for ways to improve access to and performance in coursework that will prepare students for college. In particular, these findings could inform policies targeting college going and the STEM pipeline.

Full Paper: