Researchers are gaining a better understanding of how people learn—both what works and what doesn’t go so well—in the classroom. The next step is to apply that research in actual college instruction.
One person pushing to put learning science into practice on college campuses is Sanjay Sarma, vice president for open learning at the Massachusetts Institute of Technology.
Sarma is familiar with large data sets and tough technical challenges—he helped develop the RFID tags that track inventories in libraries and big box stores. But he says his research into learning science has led him to better value how fundamental teaching and learning are to being human.
He shared his thoughts on the future of learning during a recent episode of our monthly EdSurge Live series. Listen to highlights below, or read the transcript that follows, which has been lightly edited and condensed for clarity.
EdSurge: When MOOCs started a few years ago, researchers were excited to learn from the data generated from all of these online learners. What has been the biggest lesson learned from this data trove?
We tend to think that if you have a lot of data, insights will fall out magically. But if you only have a lot of big data, it's surprisingly difficult to extract insights if you don't have an intentional experiment.
If you do intentional experiments with data, insights come out. So for example, I could in a video and change my background to blue, and see if people like it more. That's an intentional experiment. But if I just take all the logs of this video, it's very hard for me to figure out whether people want blue.
When we [started using] edX, there was enthusiasm that somehow magic would fall out of the data. We did learn a lot. We learned, for example, the demographics. We learned about our users. We learned about their career choices. But it started plateauing some time ago. We stopped getting a lot of insights out of the data.
We're also learning how people tend to transition between courses. What are the more popular topics? Where do they tend to gravitate? Where do we lose them?
At MIT you have taken MOOCs and free materials that were generated for the masses and used them in courses. Can you give an example of that?
First I'm going to give you the example of my daughter. She did all of her Spanish classes online.
How old is she now?
Sixteen, but at the time she was 12 or 13. And she learned Spanish on Quizlet actually. She'd walk into the classroom, and the poor teacher wouldn't know what to do because the kids walked in knowing Spanish. So what did he do? He invented a whole new methodology where they would do skits. They would write skits and they would act them out. Without planning it, somehow education evolved. He's an award winning teacher now, but he had to morph.
In 2002 MIT put our courses up on OpenCourseWare. We started giving our material away for free to the world, and students started taking the material for free. [Then later we started building MOOCs.]
Meanwhile, one of our professors said, “Sanjay, can I use the MOOC on campus?” It surprised me how rapidly [MOOCs] penetrated campus. At this point 99 percent of our students will have seen the edX platform on campus regardless of whether they ever take it on edX because we have a private instance of edX on campus for our students. And what do professors do with it? They flip the classroom. They give instant feedback on exams. A lot of exams are going online.
You have been researching learning science and digging into that. Can you tell us something surprising that you've found?
When we did this research we thought it was going to be very clinical, sort of cold-blooded. But I've been struck by how warm and intuitive and human the results are.
We're learning animals. We're born to learn. We're teaching animals. We're born to bring up children. That is our instinct. So what all the learning science tells us is the parent-child dynamic is closer to the truth about learning than a lot of the dogma we take for granted in colleges and schools.
So what does that mean? When I talked to kids, not just mine but other kids, I know in ten minutes their eyes glaze over and I know to back off. I know to remind them a day later, a week later, a month later. I know to make things interesting. I know that I need to change topics, mix things up. The technical terms for these are “retrieval learning,” “testing effect,” “space practice,” and interleaving.
[Audience question] Have you had experience where you've been able to gather data around the efficacy of learning materials to the point where it actually influenced those materials? So materials were removed, replaced or modified based on the data feedback you were able to gather?
A little bit. You can see for example where students are struggling and that gives you a sense of where you need to rethink the material. Actually, there's some beautiful research that shows that the most important thing one can really do is make the student curious—really focusing on the motivation. We've done some of that and it's impacted us for sure.
If you want to hook the student of course, you need to make the early parts a little bit easy so they have an on ramp. We really need to spend time thinking about how to motivate students, how to make them curious. It's like saliva. You're hungry, saliva flows. You're motivated, dopamine flows. So we need to somehow work on that, it is very human.
That's another key point. The humanization also involves things like arts and the humanities and having fun. I really think that we have to look at it from that perspective. The lens ought to be, in my view, a completely different perspective than just tweaking the existing system. I'm not suggesting blowing up schools. There are many things we do that's absolutely right, but we need to look at it from a different perspective.
