What the Science of Learning Says About Improving Online Courses
Editor’s note: Al-Fanar Media first published this article about a year ago but it seems highly relevant to teaching efforts now underway. This article is the second in a two-part series and focuses on improving online courses once they have been set up. The first article was “3 Key Lessons for Online or Blended Higher Education.”
Universities setting up online programs can turn to a growing body of research on what works when teaching in a digital format. One important takeaway from the research is that the process of designing courses is never really finished. Online courses can be gradually improved over time based on how well each component works with each new batch of students.
In other words, each online course can be viewed as a learning experiment, where data is collected (such as how much time students spend on each video or reading, or how well they do on quizzes and tests) and adjustments can be made as the amount of student data collected grows over time.
While that kind of “iterative” improvement already happens in face-to-face classes, moving online can provide richer data sets, argues Michael Feldstein, an educational-technology consultant and director of the Empirical Educator Project, which seeks to encourage such approaches. “It works particularly well in online courses because everything that happens in class is in the environment,” he added, meaning that it happens digitally and can be measured. “You have the data on every interaction in there.” Feldstein and like-minded experts argue that teaching can become more of a science and less of an art as a result of studying online education.
Here are some lessons learned from researchers and professors working over time to craft their online courses for maximum impact:
Design Multimedia With Care
Richard E. Mayer is a superstar when it comes to the science of learning—the journal Contemporary Educational Psychology ranked him the most productive educational psychologist in the world a couple of years ago. He recently published a paper looking back at 30 years of findings, and highlighted a series of key principles in how to design multimedia for online courses:
- Coherence principle: Eliminate unneeded material.
- Redundancy principle: Do not add on-screen text to narrated graphics.
- Segmenting principle: Break lessons into user-paced parts.
- Pre-training principle: Provide names and definitions of key concepts before the lesson.
While most of his principles sound like common sense, he says strict adherence to them can substantially improve the quality of learning. For instance, he advises “eliminating unneeded material” from charts or videos, under the coherence principle. But he stresses that that kind of careful editing isn’t just a nicety—it is proven to help learners get to key details that will stick. Citing what psychologists have dubbed cognitive load theory, he points out that the working memory of any learner is limited, so instructors should reduce the strain that can be caused by distracting and unnecessary information.
In the classic book he co-wrote with Ruth C. Clark called e-Learning and the Science of Instruction, he admits that teaching is creative and that research studies alone can’t tell any instructor exactly what to do. But he stresses that there are key practical findings based on evidence that show that, for instance, using a mix of words and pictures leads to better results than words alone, provided the illustrations are relevant to the material and not merely decorative.
Tests Can Be Key Teaching Moments
Research and practice have also shown that watching videos or reading online materials—even if they are well designed—is not enough to gain mastery of academic subjects.
Mayer and others say there’s a growing consensus that so-called “active learning” is essential, meaning that students must practice and try to demonstrate what they’ve learned. The active learning doesn’t have to be elaborate hands-on activities. Sometimes simply stopping a video lesson and asking students a short question about what they have just watched is enough.
Jay Parkes, a professor of educational psychology at the University of New Mexico at Albuquerque, is among those who argue that educators should stop thinking of tests and quizzes as things that are given after the learning happens—as a way to measure whether information has stuck or not. Instead, he argues that learning also happens during testing, as students think through answers, or even get some wrong and are shown the right answer and an explanation. He laid out his thinking in a book he co-wrote, called Learning and Assessing With Multiple-Choice Questions in College Classrooms.
In an interview with EdSurge, Parkes said that carefully written, multiple-choice questions can give professors a clear sense of where students are in their understanding. For instance, professors can intentionally include a “distractor,” a wrong answer that isn’t correct but that has some rationale based on the material. The point isn’t to trip up students, but to check their true mastery of the material.
‘Data Dashboards’ Can Inform Course Design
When professors teach in classrooms, they’re accustomed to being able to “read the room” and watch students’ reactions to see if they are comprehending what is being taught. In online courses, professors have to find other means to gauge student attention.
The learning-management systems that most universities use for online courses increasingly give professors the ability to see at-a-glance visualizations of how much time students have spent viewing course materials, how many times they’ve participated in online discussion, and similar details.
Utah State University is one example of an institution that has experimented with designing dashboards for professors that can give them a continuously updated perspective on students’ progress through course materials. Utah State professors have found that seeing how students use the materials also helps them improve the design of courses.
For instance, when Courtney Stewart, an assistant professor of teaching education and leadership, peeked into the data on an online course he taught about learning strategies, he noticed that only about half of the students ever visited the home page for the course. Instead they had jumped right into homework assignments. The students only clicked on links to reading assignments or video lectures when they got stuck and wanted to look up an answer. That insight led him to consider moving some of the teaching materials right onto the homework assignment pages.
In all these examples, the goal is to bring a science-and-engineering mind-set to the creation of courses. That can be a big culture shift for some professors and course administrators, and it makes the shift to online more complicated than just moving lectures from a classroom to a webcam.
While the shift to online may be complex, however, over time it opens up the possibility that the collection and sharing of student-learning data could make teaching a much more efficient process—in any part of the world.
Jeffrey R. Young is an editor and reporter focused on technology issues and the future of education. He is a senior editor at EdSurge, covering the intersection of technology and education. He previously spent 20 years at The Chronicle of Higher Education as a reporter and editor. In 2014 he spent a year as a Nieman Fellow at Harvard University, where he was also a fellow at the Berkman Center for Internet and Society.