The Fundamental Question of Edtech
Remarks to the Bill & Melinda Gates Foundation senior leadership this week.
Earlier this week I traveled to Seattle, WA, to talk with the Gates Foundation leadership about education technology. I had five minutes. Here is what I said.
There is a fundamental question in edtech, one that’s rarely asked out loud even though every edtech company has an answer. What are the humans good for and what is the tech good for? And by humans I mean both the teachers and students.
One answer to that question has been dominant for decades and is very well represented by Khan Academy. That answer says that what the computers are good for is for instantaneously evaluating certain forms of student thinking; for instantaneously serving up a learning resource if the student is evaluated as wrong, like a video, a worked example, feedback, a nudge; and, if the student is evaluated as right, for instantaneously serving up a star or some other kind of positive feedback.
What are teachers good for? The teachers are good for backup. They come along to fill in whatever needs the computer can’t meet.
What are the students good for? The answer, almost explicitly in these models, is “not much.” This edtech model sees other students as inhibitors of your learning, which is why they call it “personalized learning” rather than “socialized learning.”
The costs of this model are becoming more broadly understood. For example, the Gates Foundation has been studying personalized learning nationwide since 2015 and has found that students in personalized learning schools feel diminished feelings of belonging and safety than students in the comparison group.
My team and I at Amplify are working on a different set of answers to the question of what humans and computers are good for. In short:
Students are useful because they have lots of early mathematical ideas, ideas that are not very well captured by multiple choice items, ideas they need to offer for the sake of their learning and the learning of their classmates.
Teachers are useful because they can recognize the value of those early ideas and build on them in effective ways.
And computers are good for supporting that interaction. They aren’t backup. They aren’t primary. They’re partners. They can elicit that early thinking. They can collect it for the teacher. They can connect student thinking and teacher thinking together. This is a different kind of answer to the fundamental edtech question—a centaur model. And a little abstract. So here’s what I mean.
In this lesson, which is pretty typical of our curriculum, the computer shows the student something that doesn’t feel particularly “mathy” at first—a person getting shot out of a cannon at a carnival—which is what helps elicit that early, concrete thinking from students
.We ask students to sketch the graph of height over time here and students graph lots of different things. Straight up. Straight down. Not a lot of precision yet.
What do we do with this thinking? We don’t evaluate it as right or wrong because early student thinking is always a mix of both. Instead, we interpret student thinking back into the context. Here is the cannon person you meant to graph and here is the cannon person you graphed. Students can revise and resubmit.
This is nice but I want to suggest that where this edtech model hits a new gear is when we get the teacher involved. We give teachers a view of this early student thinking in real time. Here the teacher has lots of options.
They can say, “Everyone stop what you’re doing. Check this out. This is so interesting.” And pull one student graph up. “What are two things you like about this graph? What are two things you’d change? Talk to your neighbor and then I’ll share some of my own thoughts.”
Or they can pull up two student graphs and say, “How are these two graphs the same? How are they different? Talk to your neighbor and then I’ll share some of my thoughts.”
I was on this screen not long ago in a class in San Jose, CA, where I saw one particular student response that surprised me.
Who is this kid?
I want to offer the possibility that this is a protest. This is a student who maybe feels a low sense of belonging or engagement in math class, a student who maybe feels stupid and is going to make me, the teacher, feel stupid as well.
So what does it do for that student to hear the teacher say, “Everyone stop what you’re doing. Check this out. This is so interesting. What do you think will happen when we press play on this graph? Talk to your neighbor and then I’ll share some of my thoughts.”
What did that student learn about mathematics here? What did the class learn about mathematics here, and about their classmates?
These two answers to the fundamental question of edtech can complement one another to a limited degree but they have many more differences than similarities. And those differences result in students learning very different ideas about mathematics and about themselves and their classmates as math learners.
2023 Mar 16. I corrected the RAND study link.
Thank you so much for existing on this planet! Reading this is exactly what people need to understand about the role of teaching and learning. Interpreting the smily face response as a student reaching out for belonging is something only a human teacher could do. Teaching as an act of love is what you modeled here, and not something computers can provide.
I am a current user of the Desmos/Amplify curriculum and LOVE it. I appreciated your perspectives on EdTech and wanted to share this information with my administrators, so I clicked on the linked Gates Foundation research. It brought me to a paper from 2015 that only talked about the scholastic gains of students experiencing personalized learning, but didn't mention "diminished feelings of belonging and safety". Is there a different study you could direct me to?