Key Takeaways

AI improves student learning when it’s built on learning science. The Doer Effect, which shows that students who practice while they read learn significantly more than students who only read, is the approach the research points to, and AI makes it possible to scale.

• Assign formative practice: AI-generated practice questions embedded in course materials drove a 3x increase in textbook usage at Iowa State University.
• Look for the Doer Effect: At Cal Poly Pomona, students who completed AI-generated practice questions before major exams outperformed their peers, with the largest gains among lower-performing students.
• Listen to students: 93% of Iowa State students said practice questions helped them learn, in unprompted, anonymous course evaluations.

How AI Can Scale the Doer Effect

Higher education leaders are navigating a crowded AI landscape, and most products claim to improve student learning. The real challenge is figuring out which tools actually do.

When it’s done right, AI gives students a better shot at retaining the material they study. But many of the tools that students are already reaching for weren't built on sound learning science, and they leave educators with no reliable way to know if they're working. The research points to one approach that consistently delivers: the Doer Effect. And AI is making it possible to scale this learning science principle.

We partnered with Higher Ed Dive to share what the research says about using AI to improve student learning. Here are three key takeaways.

Takeaway 1: Assigning formative practice means better use of course materials.

Assigning the reading and students actually doing the reading are two different things. When AI-generated practice questions are embedded directly into course materials, students are likely to engage more consistently and ultimately retain what they're reading. The data backs this up too. A VitalSource and Iowa State University study found that incorporating AI-generated practice questions drove a 3x increase in textbook usage.

"We know how much it costs for humans to write excellent formative practice questions. With AI, we can produce high-quality formative practice questions at a massive scale."
— Michael Hale, Chief Learning Officer at VitalSource

Takeaway 2: AI-generated questions achieve the Doer Effect.

Students who practice while they read learn significantly more than those who only read. VitalSource wanted to know if AI-generated questions could produce the same result. Working with Cal Poly Pomona across two semesters, our research confirmed that they did. Students who completed practice questions had better learning gains, and lower-performing students saw the largest gains.

"The big takeaway here is that when AI is used as a tool for research-based learning methods, all students benefit, but students who need the most support benefit most of all."
— Rachel Van Campenhout, Senior Research Scientist at VitalSource

Takeaway 3: Students can see the benefits of formative practice firsthand.

In a study with Iowa State University, 93% of students said practice questions helped them learn. They said so unprompted, in anonymous course evaluations sent directly to their instructors. When students are that candid about something that isn't graded, it tells you something real.

“Faculty are assigning reading because they want students to do it. Now they have a tool to not only make sure students are doing the reading, but to help ensure students are more effectively learning the material — a benefit for all students, but for struggling students most of all.”
— Rachel Van Campenhout, Senior Research Scientist at VitalSource

If you're interested in learning how to use AI effectively on your campus, download the full playbook here.