U-M data science projects explore connection between student achievement, well-being
ANN ARBOR—Can Big Data analyses improve classroom achievement and well-being for college students or teaching effectiveness for professors?
Researchers at the University of Michigan will explore these questions through two new projects focused on innovative and interdisciplinary applications of learning analytics.
Supported by the first round of the Michigan Institute for Data Science (MIDAS) Challenge Initiatives program, the projects bring together teams of researchers from across campus to leverage educational and behavioral data collected from U-M students, and translate it into practical insights on how to improve student outcomes and teaching quality.
The two projects are each receiving $1.25 million in MIDAS Challenge Initiatives grants.
One of the projects, led by Rada Mihalcea, professor of electrical engineering and computer science, seeks to uncover connections between students’ personal attributes such as values, beliefs, interests, behaviors and backgrounds and their success in school or overall sense of well-being. The other, led by Stephanie Teasley, a research professor in the School of Information, will build a holistic model of student achievement—one that uses multiple statistical and Big Data methods to analyze students’ written work, behavioral data, and institutional data.
“These interdisciplinary projects will result in new ways of helping students achieve their full potential,” said MIDAS co-director Brian Athey, professor and chair of computational medicine and bioinformatics. “The projects will also advance the state of the art in data science, helping to address privacy issues and other important methodological concerns.”
The funding is part of U-M’s Data Science Initiative, which was announced in September 2015.
Mihalcea said the goal of her research is to expand the traditional analysis of educational success to include student life, personality and background outside of the classroom. Much of the previous research in learning analytics focuses only on academic behavior, to the exclusion of factors that may play important roles in student success or failure.
“Our project envisions using new ways of analyzing data to consider all aspects of a student’s experience, lifestyle and background when developing tools to aid in educational success—of course, with due concern and assurances regarding the protection of students’ privacy,” Mihalcea said.
Mihalcea’s collaborators on the project include researchers from the College of Engineering, School of Information, School of Public Health, School of Education, and Department of Statistics.
One of the aims of Teasley’s project is to integrate the learning analytics work being done by U-M research in several different fields—from visualizing educational data to studying active learning technologies to analysis of in-class behavior—into a single, data-driven model that promises to yield new insights into the learning process.
“This research has the potential to yield new understandings of how people learn,” Teasley said. “Our goal is to demonstrate how data-driven inquiry can improve teaching and learning in higher education.”
Researchers from the School of Information, College of Engineering, School of Education and the departments of Physics and Astronomy are also working on Teasley’s project.
Michigan has been a leader in learning analytics since 2012 when the provost’s Learning Analytics Task Force was launched to help the university community develop new ways to use existing student record data to improve teaching and learning. The task force helped sponsor the Symposium on Learning Analytics at Michigan (SLAM) seminar series, create the Practical Learning Analytics MOOC, and launch of the Digital Innovation Greenhouse, where new education technologies like ECoach, Student Explorer and GradeCraft are being grown to scale every day. To take learning analytics to the next level, researchers are starting to take advantage of new, much more complex kinds of data.
The goal of the multiyear MIDAS Challenge Initiatives program is to foster data science projects that have the potential to prompt new partnerships between U-M, federal research agencies and industry. The challenges are focused on four areas: transportation, learning analytics, social science and health science. For more information, visit midas.umich.edu/challenges.
“U-M is in a great position to advance the field of learning analytics because we have a great trove of conventional academic data, other related databases, and an outstanding team of researchers across many disciplines to analyze them,” said Alfred Hero, co-director of MIDAS and professor of electrical engineering and computer science.