Learning analytics embedded into digital textbooks may be more effective at predicting student outcomes than prior academic achievement, a metric long thought to be the strongest predictor of success.
Predictive and learning analytics has been touted for its ability to “read” students, providing educators with a host of data about studying habits, problem areas, and background information, according to a recent study commissioned by eTextbook publisher CourseSmart.
With analytics installed directly into a digital textbook, that analysis can happen seamlessly while students are doing their own reading, researchers said.
“Seeing how a student’s engaging with a book, how much of the book they’re consuming, how they’re using the tools in the book, is an indication of how successful they’ll be,” Cindy Clarke, CourseSmart’s vice president of marketing, recently told Education Dive.
CourseSmart provided researchers with data on about 230 students who used its digital course materials during a pilot program in 2013.
The information included final course grades, prior academic achievement, and what CourseSmart calls an engagement index score, which measures how much a student is engaging with eTextbooks. This includes the number of pages viewed, highlighting passages, and taking notes.
The researchers found that they were able to more accurately predict student outcomes and better identify points of intervention based on the engagement index than traditional methods like attendance, class participation, and prior achievement.
The report found that students read an average of 551 pages throughout the semester, and spent an average of 442 minutes on reading the digital textbooks.
Each student created about 4 highlights, 42 bookmarks, and 16 notes – and the more a student engaged with the eTextbook, the better the final course grade.
But not all types of engagement were necessarily indicators of better outcomes said Reynol Junco, an associate professor of library science at Purdue University who co-wrote the research paper.
“What was especially interesting was that highlighting was related to student course outcomes, although not in the way that you might think,” Junco wrote. “Those students who were in the top 10th percentile of number of highlights had significantly lower course grades than students in the lower 90th percentile.”
This, Junco said, is due to the fact that low-skill readers often highlight more text than high-skill readers.
Having such data come right from a textbook in real time could help instructors identify students who are not only struggling to absorb information, but having difficultly reading it in the first place.
With even traditional textbook publishers transitioning into digital learning companies, simple e-versions of print textbooks may soon be seen as old-fashioned as the physical books themselves.
Over half of Wiley’s annual revenue now comes from digital course material, according to a White Paper released earlier this month by IXXUS that urged publishers to adopt a “digital-first” approach.
In October 2013, McGraw-Hill Education’s former CEO predicted that nearly 40 percent of its offerings would be digital products by the end of the year, with a particular focus on adaptive learning and analytics.
“No matter what a student’s prior academic ability, which may not be specifically known, the course instructor can have an unobtrusive real-time method to identify students at risk of academic failure that is not tied to activity on a learning content management system,” the CourseSmart researchers wrote.
Follow Jake New on Twitter at @eCN_Jake.
- What does higher-ed look like in 2023? - January 5, 2015
- Are ed-tech startups a bubble that’s ready to burst? - January 1, 2015
- Are MOOCs really dead? - August 28, 2014