Predicting College Persistence among High School Students

Fellows: Benedict Kuester, Michael Stepner, Masha Westerlund
Data Science Mentor(s): Anushka Anand
Project Manager: Alan Fritzler
Project Partner: Noble Network of Charter Schools, KIPP: New Jersey, KIPP: Chicago, Perspectives Charter Schools

Many high-performing charter schools graduate students from low-income backgrounds at much higher rates than their traditional school counterparts, and these graduates are several times more likely to finish a four-year college degree than the national average. To build on this success and further improve college completion rates, these charter networks provide additional support for alumni after graduation and during their college years. These organizations would like better methods to predict which students will most benefit from these support initiatives.

DSSG will build a useful, data-driven model that can predict which graduating students will struggle in college. Fellows will interface with multiple charter networks, balancing their various interests and needs, while also working towards a system that can be used by additional schools around the country in their own college persistence programs. These tools will help more schools achieve their core mission of preparing students for successful post-secondary education and careers.