Partner: Muskingum Valley Educational Service Center
Team: Michael Feffer, Alice Lai, Kit Rodolfa

Background

Across the country, school districts have been creating support programs to help students in need of additional help to meet their academic goals. These goals range from on-time graduation to college admissions to college persistence and graduation. While many of these programs exist, they are limited in their capacity and can only enroll a small number of students. School districts have been using various heuristics to identify students who may be in need to such programs such as using information about GPA, attendance, or test scores. While there have been signs of some success, these simple indicators often don’t use all the information available about the students, don’t allow schools to prioritize who are most in need of support, and miss many students. In our collaboration with Muskingum Valley ESC, the largest educational service center in Ohio, we are building on previous work to develop early warning systems to identify students who are at risk of not meeting their academic goals. Much previous work has focused on high school graduation as the definition of student success, but we are interested in other longer-term outcomes like college enrollment and college graduation. In addition, we want to characterize the types of students who are missed by current early warning systems, and build models to better identify those students.

This work is an extension of a related project from the 2016 Data Science for Social Good Fellowship.