Many students arrive on college campuses with the intention of pursuing science, technology, engineering or mathematics (STEM) majors and careers, but about half of them never complete a STEM degree (48% in a 2013 US Department of Education report). Of those who leave their STEM major, half eventually graduate with a non-STEM degree, but the rest leave college without earning any degree at all. Preparing undergraduates to join the STEM workforce is a high priority for universities, and this rate of attrition has captured the attention of university administrators and educational researchers.
Research on STEM dropouts indicates that students tend to leave their majors when they perform poorly in a course, perceive they lack the critical skills needed to perform tasks in STEM courses and professions, and lack the motivation to continue onward with their training. Dr. Matthew Bernacki and his project team are working to prevent course attrition and improve achievement in STEM courses, thus increasing the number of graduates who are hired into STEM professions.
Bernacki is Principal Investigator and recipient of a $499,973 award from the National Science Foundation to explore how students use learning management systems - such as UNLV's WebCampus - and whether features designed to help students build their learning skills and maintain their motivation can increase student achievement in STEM courses. The three-year, collaborative project with faculty from Biology (Jenifer Utz), Math (Carryn Warren-Bellomo, Monika Neda), and Engineering (Donald Hayes, Jeffrey Markle) and support from the Office of Information Technology began last fall and will continue through the summer of 2017.
Bernacki and his project team designed a set of three web-delivered modules - known as the Learning to Learn series - to teach students key learning skills known to improve learning and achievement. The three modules review the common challenges students face when learning while at college, then introduce them to learning principles. Each module explains a set of learning principles, describes a case where use of the principles improved student achievement, and provides students with interactive opportunities to practice and receive feedback as they incorporate these principles into their own approach to learning. In total, the Learning to Learn series requires about two hours of training that can be completed on students' own time via WebCampus.
Initial findings show that the Learning to Learn intervention holds promise for improving students' academic performance in their Biology course work - especially for students who struggled in the early weeks of the course. Results of a randomized control study indicate that students who completed the Learning to Learn intervention after their initial exam outperformed a control group on the next two exams, and that the benefits of the intervention were greatest for the students who scored most poorly on the initial exam. In essence, training students to adopt evidence-based learning strategies improved achievement, and effects were larger and more persistent for those students mostly likely to need support.
While the Learning to Learn series can build students' learning skills, other tools are needed to help warn students about impending poor performances that can lead to STEM dropout. Bernacki and his team are also designing a tool to decrease the time it takes for students to be notified that they are in danger of failing a course.
Because typical "early warning systems" rely on students' performance data (like midterm exam scores) to generate alerts, notifications to students who are struggling often arrive too late for them to recover. Bernacki and his team aim to move up this "early warning window" so that alerts arrive before major tests and assignments so that students can adapt before a poor grade is earned. To accomplish this, the team is building a warning system that relies on WebCampus usage data to detect potentially problematic behavior and alert students after just a few weeks of learning.
Preliminary findings show that their warning system accurately identifies approximately 75% of students in Biology courses who would eventually fail to obtain the grade needed to move on in their major (e.g., a B or better for nursing and pre-med majors). This behavior driven system is able to warn students about the strong possibility of obtaining a poor grade approximately six weeks before a midterm generated "early warning" alert would typically arrive. This early warning provides greater opportunity to change learning strategies and seek out help, increasing alerted students' likelihood of success in the course. With this early warning system for Biology nearing completion, the team has turned its attention to building similar detectors for Calculus and Engineering courses and aim to implement them in fall 2016.
Bernacki sees promise for the Learning to Learn series and other tools designed to support student STEM achievement and retention: "It is exciting to find that an investment of just a few hours spent Learning to Learn more effectively enhances students' achievement. With some work to improve the effectiveness of modules and the precision of the warning system here at UNLV, we can continue to improve student outcomes locally and eventually share these innovations to help STEM majors at other universities, too."
Dr. LeAnn Putney, Chair of the Department of Educational Psychology & Higher Education, sees his research as an important step for student retention, progression, and completion: "Dr. Bernacki's research has offered a great contribution to the department. His collaborative research affords students in the STEM fields a crucial and systematic model of learning skills to ensure student success. The Learning to Learn model is especially timely for the highly diverse student population at UNLV and furthers the UNLV mission of student retention and completion."