CIRCL perspectives offer a window into the different worlds of various stakeholders in the cyberlearning community — what drives their work, what they need to be successful, and what they think the community should be doing. Share your perspective.
How did you get started in learning technology?
After my undergraduate work at Penn in systems engineering, I came to Silicon Valley. I was in high tech for about 8 years, during the first crazy dot com bubble. At one point, I did a mental exercise of how I’ll feel if I’m doing this 10 years from now, and thought, “I’ll be a bitter old man!” So being an engineer, I said, well, the constant is I’m going to become an old man. Maybe I can do something about the bitter part. I saw a career coach and did some exploration. My wife suggested the Learning, Design Technology (LDT) program at Stanford, which is the one year master’s program. She was doing her master’s degree, and I saw how much fun she was having. Shelley Goldman was heading up the program, and my advisor was Dan Schwartz. I spent a lot of time at the d.school and studying learning sciences. It was an amazing year.
At that time, there were three things happening that were pretty interesting: First, Warren Buffett just announced that he was going to give 30 billion dollars to the Gates Foundation. So suddenly there are more resources to do things in education. Second, One Laptop Per Child was announced, which, for better for worse, at that point, was at least a huge drop in cost for technology. Third, online courses were starting up. When I was an undergrad, my summer job was working with a professor to put Penn’s first courses online — recording his lectures in digital audio, coding them in media files, and posting his presentations. Now, years later, I was looking at what was online and thinking “This is the same stuff I did 10 years ago.” At first I was a little depressed, but then I realized there was a big opportunity for innovation and impact.
These things gave me confidence to jump in and get my Masters. Afterwards, I spent a whole year trying to get into philanthropy, and failing miserably. I got a job offer in Philadelphia, but my wife and I just had our first kid, and didn’t want to move across country again. Then a friend forwarded a Hewlett Foundation opportunity to me, and I thought it was exactly what I had been trying to do all year. I talked with the directors, Mike Smith and Cathy Casserly, and got a position managing the technology and innovation grant making, especially around OER. During that time, I saw all kinds of different systems. Putting on my systems engineering hat, I could see so many connections and networks that could be built. I tried to be humble in my approach, knowing that I hadn’t been a teacher, and I definitely have blind spots. But it was a really exciting kind of shepherding. I got to work with just amazing people.
After about six and a half years, I stepped out for a year and a half to do independent consulting, and an opportunity around Learner Positioning Systems popped up — a new initiative to blend of learning science and leverage technology to support diverse and underserved learners. Karen Cator and I had talked about working together for a number of years, and so I came to Digital Promise.
And just to circle all the way back: While I was studying to get my masters at Stanford, my aunt who lives in Belgrade, Serbia, found my grandmother’s leather bound, handwritten Cyrillic script dissertation on the role of play in early childhood learning. My grandmother was the first woman psychology PhD from former Yugoslavia, in 1925. And I’m like, Oh, my God, this is an heirloom. I Skype with my aunt every once in a while; she’s getting the dissertation typed up. Another interesting thing is that my grandmother’s father was a science teacher in Croatia, and he noticed a girl who was incredibly talented in a middle school science, so he wrote a recommendation that she’d be allowed as the only exception to attend the boys school. And so she did, and he was her teacher. And then later, she went on and ended up marrying Albert Einstein. How did she influence the theory? And that brings us full circle.
Can you tell us about Learner Positioning Systems (LPS)?
Technology enables personalization and personalized learning, obviously, but what we need to understand is what are the key areas that can vary about how an individual learns, and how knowing that can impact specific outcomes.That’s actually where you gain a lot of leverage by personalizing, right? There are many different approaches you can take around pedagogy and strategies, but I felt there wasn’t enough that built off an understanding the whole child learner and using that as your center of gravity for design.
With LPS, we’re building research-based models that leverage technology and learning science to have a bigger impact in education. What’s unique about LPS is the framing of learner variability and what we know about a specific learner––it maps key factors that kids vary on with how they learn. Currently, we have detailed, research-based models of critical variables that impact learning success for grades preK-2 math, pk-3 reading, and 4-6 literacy.
Our first push is with product developers, to incentivize them to develop increased research-based supports that address the variability in all learners. Right now, product developers and designers are often “graded” on designing for the user. You look for things like someone falls off your website, or clicks the wrong thing, or is disengaged. How do you help developers become experts on designing for the learner? Knowing whether or not someone is learning is very difficult.
How do you work with product developers?
We use the LPS framework and describe strategies that product developers can understand and work towards designing into their product. For example, if you’re doing early math, working memory is tremendously integrated with learning, and so having some supports for kids with low working memory would go a long way to help more kids succeed using your product platform. Strategies that help include a graphic organizer, or breaking up problems into smaller steps with repetition.
I often say to product developers: You have some theory of impact of your product, right? There are some experiences that you consider optimal for a learner to go through in your product and come out with more knowledge or skills than they have before? What is the positive experience, and what percentage of your users have that experience? And the ones that drop off––what’s going on there? What if we could bump it up by using the research to support more learners using your product? Some of the developers have measures of learning, and many have a gut feel. Other developers are more engagement oriented, so it’s also about coaching them that engagement isn’t enough.
To help incentivize developers to go through this process, we’ve launched a mini grant program. Our starting goal is to see five major products in the market that take the LPS framework and make some product roadmap decisions based on it. For example, working with with Readworks through this program, we’ve helped them understand the learner from not only the reading comprehension standpoint, but also the cognitive and social-emotional standpoint––and they’ve changed their product roadmap as a result. They told us “we don’t have a research staff, and you have saved us two years worth of work here.”
How do you sustain this effort? What would you like LPS to look like in the next 5 years?
Right now we’re filling philanthropically funded, and will be for probably 5-6 years, so we have a runway to get it out there. Everything is open source, and we can always provide consulting services to support people using the framework.
In the longer run, LPS is kind of an ecosystem play. We hope people start to understand learner variability and see things more through that lens––for example, understanding that kids vary, and that a lot of things that are often labeled as negative, like dyslexia or autism, are actually just variability among factors. And we hope that infrastructure is designed to support learner variability, whether it’s through individual products, or kinds of data systems that districts use to capture information on how kids learn best and share that in safe ways to support learning.
Can you say more about how data systems and LPS could help teachers and students?
Here’s a personal example that shows the level of opportunity. My daughter is in fourth grade in public school. In second grade, i visited around the second month of classes and noticed that the kids had put writing samples on the walls. My daughter’s was like 2-3 sentences, while most others were longer. Her sentences were reasonably good quality, though, so the teacher wasn’t concerned. But production and practice is important for learning and growth, so I talked to the teacher about it and did some observing. I noticed that during writing time, my daughter got distracted super easily if some other kid did something. So I sat down with the teacher and we talked about some strategies––like don’t sit her with the three most rambunctious boys in the classroom (which she was), and if you’re talking in front of the class, have her towards the front so that if anyone distracts, it’s not in her line of vision. We could have also talked about headphones, or occupational therapy strategies with the dividers, if needed. But we just changed her seating and boom, she’s writing many sentences, just like the other kids.
But the bigger aha was the next year. Again, about two months into the class, my daughter was struggling with being distracted. I asked the teacher if he talked to last year’s teacher about her, and he hadn’t. So I said, “Oh, okay; well, let me tell you some strategies that are pretty powerful for my daughter.” He’s a good teacher, and he might have figured this out, but these nuances and strategies can take months with a new teacher––and over years of schooling, that can turn into years of fumbling time.
LPS can be applied in ways useful for teachers, for learners, and for parents. We can’t design for everybody all at once, so we picked products developers first because they help provide an infrastructure that starts to make easier to support teachers.You can’t just say to a teacher, “Hey, you should personalize on all these different dimensions.” But if a product can make suggestions, that can make it easier for teachers. There are tons of different things that matter, and you’re not going to be expert on all of them. But having a map that just shows you that this stuff exists is the first step. Then you can find an expert, or know to ask the question, or check this or that. Having a map helps make improvement possible.
Another place we’re seeing some usage is for coaches or for learning specialists, and we’re experimenting there, too. For example, the Friday Institute connected it’s Learning Differences MOOC with LPS and built a stack of microcredentials focusing on learner diversity and students’ learning differences around executive function factors. So we have microcredentials around understanding of working memory and sensory processing for attention, where a teacher can learn about these and submit evidence that they’ve implemented strategies to support these various factors.
What do you think the research community should doing?
A big eye opener I had in Dan Schwartz’s classes on research topics and learning transfer was the sensitivity of learning design. Sometimes when you look at contrasting cases, one little shift like showing one example versus two, can mean that transfer drops from like 70% to 30%. It’s hugely sensitive. But then the other piece, toward the end of average, is that even many established learning theories, which are powerful and great, when you look at the studies, it’s because it worked for 70% of the kids. What about the other 30%? If I take 20% of the K-12 population, that 16 million kids! I can’t just say, “Oh, let’s move on.” You have to start somewhere, and we are getting better tools, but it is having that mindfulness that let me go a little further than this and see if we can uncover another 10-20% with the goal of broadening impact. It’s more about what works for whom in what contexts. It’s going to be a handful of things.
LPS can apply to many different audiences, and for researchers, one thing that is popping up is identifying gaps in the research. We’ll do a learner model for K-3 reaching, and people will be like, “Oh, what about, say, growth mindset?” We’ll say, “Well, we haven’t seen much research with this age range and how it relates to reading. So we’re not going to make that claim; most of what we’re finding is middle school and high school. They are living models, so sometimes people will send us new research or research we missed and we’ll add it. But sometimes there isn’t much on a particular topic in a particular age range, or there are other issues. For example, a lot of the motivation and engagement work is tricky for the early years because much of it is self report, and some of the older research measured engagement as sitting quietly at your desk! So our work also highlights gaps in research, and we’d love the community to help us fill these gaps.