An Alternative to the Engineering Model of Personalized Learning

There is an article in EdWeek that quotes Larry Berger, CEO of Amplify, in his “confession” about personalized learning. The focus is on K-12 education but applies directly to higher ed as well.

Until a few years ago, I was a great believer in what might be called the “engineering” model of personalized learning, which is still what most people mean by personalized learning. The model works as follows:

You start with a map of all the things that kids need to learn.

Then you measure the kids so that you can place each kid on the map in just the spot where they know everything behind them, and in front of them is what they should learn next.

Then you assemble a vast library of learning objects and ask an algorithm to sort through it to find the optimal learning object for each kid at that particular moment.

Then you make each kid use the learning object.

Then you measure the kids again. If they have learned what you wanted them to learn, you move them to the next place on the map. If they didn’t learn it, you try something simpler.

If the map, the assessments, and the library were used by millions of kids, then the algorithms would get smarter and smarter, and make better, more personalized choices about which things to put in front of which kids.

I spent a decade believing in this model—the map, the measure, and the library, all powered by big data algorithms.

Here’s the problem: The map doesn’t exist, the measurement is impossible, and we have, collectively, built only 5% of the library. [snip]

So we need to move beyond this engineering model. Once we do, we find that many more compelling and more realistic frontiers of personalized learning opening up.

Larry is exactly right that there is a fundamental problem with the assumptions behind what he calls the engineering model of personalized learning. But there are alternate models that offer “more compelling and more realistic frontiers”. We have described this contrast in models at e-Literate, most directly in Michael’s post The Battle for “Personalized Learning”.

Phil and I have decided to claim this prime piece of linguistic real estate. We are asserting squatters’ rights.

We hereby decree, by the power vested in us by nobody at all, that “personalized learning” shall henceforth refer to a family of teaching practices that are intended to help reach students in the metaphorical back row. The ones who are bored, or confused, or tuned out, or feeling stupid. Personalized learning practices are almost always ones that teachers have been using for a very long time but that digital tools can support or enhance. Here are a few that we have identified so far:

Move content broadcast out of the classroom: In many disciplines, the ideal teaching format is a seminar, in which students spend class time engaged in conversation with a professor. In others, it is a lab. Both models have students actively engaged in academic practice during class time, when the professor, as the expert practitioner, is present to coach them. Every class spent lecturing is a wasted coaching opportunity.

Many disciplines have traditionally used assigned readings to move content broadcast out of the classroom, and some still do. But it is not always possible to find readings that capture what you want to cover, and in any case, it is becoming harder to persuade students to read. Luckily, there are tools that can help with this problem. You can record and post your lectures as videos, which students can watch as many times as they need to absorb what you’re trying to tell them. You can assign podcasts that they can listen to on the go, or find interactive content that keeps them more engaged.

Make homework time contact time: Good teachers help students see the direct connection between the work they do at home and the overall purpose of the class. They do this in a variety of ways. Sometimes they mark up and comment on the student work. Sometimes they ask the students questions in class that require them to build on the work they did at home. For a variety of reasons, which often boil down to professors’ having less available time per student, this has become harder to do. The great crutch that is now being used to limp along without actually solving this problem is robo-graded homework assignments. By itself, automated practice might help some students drag themselves through to the end of the semester. But it doesn’t often inspire them to think that maybe they are not destined to be the student in the back row forever. (There are important exceptions to this rule, which I address below.)

On the other hand, these same automated homework tools can also give teachers an easy view into how their students are doing and create opportunities to engage with those students. “Analytics” in these tools are roughly analogous to your ability to scan the classroom visually and see, at a glance, who is paying attention, who looks confused, who has a question. Nor are these the only tools available for making homework time feel less isolated and pointless. Any homework activity that is done electronically can be socially connected. Group work done on a discussion board can be read over by the professor when she has time. Highlights and margin notes on readings can be shared and discussed in class. This sort of effort on the professor’s part doesn’t have to be exhaustive (or exhausting). Sometimes a small gesture to show a student that you see her is all it takes.

Hire a tutor: You know what tutors are typically good for in your particular discipline. You also know that there generally aren’t enough good ones available, and that even when there are, it’s tough to get students to come into the tutoring center. One of the best uses of machine-graded homework systems, especially when they are “adaptive,” is to treat them as personal tutors that are available to students whenever they need them and wherever they are. They aren’t perfect, but what tutors are? Sometimes getting students out of the back row means helping them to believe that they are capable of learning. And sometimes students are willing to pose a question to a computer that they would be embarrassed to ask in person. In those cases, a little extra practice and feedback on the basics, without judgment, can make all the difference — even if the feedback comes from a machine. And if adaptive learning robo-tutors don’t fit the needs of your students and your discipline, technology also makes it possible to connect students with actual human tutors, who are available online to help them get through the rough spots.

We wrote more extensively about this description of personalized learning at EDUCAUSE Review in 2016 at “Personalized Learning: What It Really Is and Why It Really Matters”.

There is a battle for personalized learning, and the description of the engineering model is useful for understanding one approach (unfortunately the one most often used in marketing and by ed reformers). But there is an alternative and it is more compelling.