Principal Investigator: Jeff Greene
Co-Principle Investigator: Matt Bernacki
Funding Agency: National Science Foundation


The focus of this Development and Implementation Level II project is to support, retain, and increase the achievement of undergraduates who traditionally do not persist in STEM majors (e.g., underrepresented minority groups, first-generation college students) by (1) applying an existing data-driven solution to identify struggling STEM learners before they begin to fail, (2) developing targeted, effective achievement and retention interventions combining the expertise of two universities who are leading sources of empirically supported approaches for STEM success, and (3) demonstrating the applicability of these solutions to a variety higher education contexts.

College STEM degree completion rates hover just under 50%, and only a quarter of the students least represented in the STEM workforce (i.e., underrepresented minority and first-generation college student groups) graduate within six years. When college students are asked about their reasons for leaving STEM, many point to introductory-level STEM coursework and classroom environments that are unengaging, unwelcoming, and overly-challenging. Active learning classrooms bolster student performance and shrink achievement gaps. However, such initiatives require students who can effectively self-regulate their learning within and outside of class, things many first-year college students struggle to do. This project will marry an existing data-driven, web-based approach for early identification and support of struggling college STEM learners with a robust student success initiative, focusing on fostering all students’ ability to self-regulate their learning.

The University of Nevada Las Vegas’s Learning Theory and Analytics as Guides to Improve Undergraduate STEM Education (LearningTAGs; NSF DRL-1420491) Project has used a data-driven approach to identify and directly intervene with struggling students. Prediction models that use only data generated from student use of online course materials before their first exam can successfully identify >80% of students who will have to retake a critical introductory course like anatomy and physiology. Using these prediction models, targeted web-based interventions have been delivered to struggling students, who subsequently outperformed their peers on exams taken just days after the intervention, and throughout the semester. LearningTAGs methods can be expanded to better serve students underrepresented in STEM by integrating findings from the Finish Line Project at the University of North Carolina at Chapel Hill, a “First in the World” project funded by the United States Department of Education. Finish Line Project researchers have found first-generation college students benefit most from early intervention, accessible academic coaches, and active learning STEM classrooms.

The proposed collaborative project will: (1) develop and test LearningTAGs prediction modeling and digital intervention at another highly-selective, public institution (i.e., UNC), as well as at a community college (i.e., College of Southern Nevada); (2) leverage Finish Line Project findings on academic coaching to test various support interventions (i.e., online self-regulated learning instructional modules, academic coaching, and supplemental instruction); and (3) identify whether support efficacy varies across traditional, underrepresented minority and first-generation college students.