Skip to main content

NSF funds $2 million project to help college students struggling in math and science courses

A project led by researchers at the University of North Carolina at Chapel Hill will receive $2 million in funding from the National Science Foundation to pioneer new ways to help undergraduate college students who are struggling in mathematics and science courses.

The project, which involves researchers from Carolina’s School of Education and College of Arts and Sciences, and two higher education institutions in Nevada, will use existing campus academic data systems to identify struggling students then provide them with supports that have been shown to help students succeed in rigorous courses.

Leaders of the project, School of Education faculty members Jeff Greene and Matt Bernacki, will work with a team that includes College of Arts and Sciences researchers Abigail Panter, senior associate dean for undergraduate education, and Kathleen Gates, professor of psychology and neuroscience, and researchers from the University of Nevada-Las Vegas.

“We’re excited about this project as it promises to refine and extend work at Carolina and UNLV to help students, especially those from underrepresented minorities and first-generation students, groups that have often struggled in college mathematics and sciences courses,” said Greene, an associate professor of educational psychology and learning sciences who also serves as the School of Education’s interim associate dean for academic affairs.

The grant will build on work already being done at UNC-Chapel Hill and at UNLV to identify students struggling in STEM – science, technology, engineering and mathematics – courses, then test interventions designed to help them succeed.

The project will rely upon learning analytics and machine learning methods applied to data collected from students’ everyday interactions with existing learning management systems such as Sakai to identify students likely to struggle based on their early learning activities in large gateway science courses.

The team’s intervention efforts include academic coaching methods developed at Carolina as part of the Finish Line Project, which has been led by Panter and has received $3 million in U.S. Department of Education funding to help first-generation students succeed in college. It will also incorporate digital skill training approaches developed by Bernacki at UNLV, where Bernacki worked before joining the School of Education this year.

Researchers will work in partnership with campus data systems offices and academic departments in the College of Arts and Sciences, to ensure that machine learning algorithms identify struggling students accurately and as early in the semester as possible to maximize the benefits early interventions can achieve.

The College of Southern Nevada, a two-year community college, will also take part in the project and explore an intervention approach called Supplemental Instruction.

Thousands of students in North Carolina and Nevada will be involved in the five-year project.

Bernacki, an assistant professor of learning sciences and psychological studies, said: “We anticipate that mining students’ learning behavior will help us design and test models that identify students we can support, and that these powerful interventions can increase achievement, especially for underrepresented students.

“By replicating a learning analytics solution across school types and testing different intervention methods across schools, we’re hopeful a model will emerge that can be adopted across a wide range of higher education institutions,” Bernacki said. “The learning analytics approach can help schools target their existing student resources and make them more efficient.”

Combining early identification, academic coaching

The project will use data-collection systems already being used at Carolina and at UNLV that trace student use of digital materials their instructors provide, and mine the data to identify struggling learners before they begin to fail.

The project will incorporate earlier NSF-supported work in which Bernacki developed a system called Learning Theory and Analytics as Guides to Improve Undergraduate STEM Education Learning, or “LearningTAGS.” LearningTAGS used students’ online learning behaviors in the earliest weeks of STEM courses to successfully identify students who need additional support to be successful in those large lecture settings.

Using that information, instructors and researchers reached out to students identified as at risk for struggling in the courses and provided advice from successful course completers and a series of online instructional materials that have been shown to improve students’ learning behaviors and exam performance after completing the skill training modules.

Those students identified as needing help will also receive coaching by counselors using methods developed in Carolina’s Finish Line Project, which has demonstrated that first-generation college students benefit from early intervention, academic coaching, online instructional courses in self-regulated learning, and the use of active-learning methods in STEM classrooms.

“Too many students struggle with STEM courses, delaying their progress to graduation or in some cases forcing them into other academic pathways,” Greene said. “We expect this work will expand understandings of how higher education institutions can better prepare students for challenging STEM courses and support them when they encounter challenges.”

Other researchers taking part in the project include Kelly Hogan, associate dean of instructional innovation at Carolina’s College of Arts and Sciences, and co-principal investigators Erin Windsor, a faculty member at the College of Southern Nevada, and Christy Strong, a faculty member at UNLV.