Bringing Learning Science Into Classrooms – Dr. Stephen Kosslyn, President of Active Learning Science


The good news is researchers have established a tremendous amount about how human memory is acquired, organized, and deployed. The bad news, according to Dr. Stephen Kosslyn, is this information has pretty much stayed in technical journals and textbooks and not been applied in classrooms. Adding to the problem is that popular misconceptions about learning abound, so most of us are not learning nearly as effectively or efficiently as we could. Kosslyn, one of the world’s leading researchers on the science of learning, has long been concerned by the inadequacies of our education systems. Through public-facing books, and institutions he helped create like Minerva University and Foundry College, he has dedicated much of his life to bringing what researchers understand about learning into real world practice. Tune in to this fascinating conversation with host Dr. Rishi Desai to hear how our education systems could be improved by applying active learning and by teaching critical thinking skills, among other changes.





Hi, I’m Dr. Rishi Desai.  Our guest today, Dr. Stephen Kosslyn, spent decades teaching and doing research at two of the leading universities in the U.S. – Harvard and Stanford – and was involved in creating Minerva University and Foundry College.  He is also President of Active Learning Sciences, and is widely recognized as one of the world’s leading researchers on the science of learning. 


We’re going to explore active learning and the application of it to online education with him today.  Thanks for taking the time to join us, Dr. Kosslyn. 


I'd like to first start out with learning a little bit more about you. What got you first interested in education and learning science?


Stephen Kosslyn: It's actually when I was a high school student, believe it or not. It was in the '60s and like many high school students, I was just becoming socially aware and aware of what was going on around me more generally and decided that the world had many flaws and that the ultimate that will address those flaws was really going to involve education. 


So as a high school student I was already really oriented toward the idea that education was something need to be fixed and it could be done a lot better but couldn't know how. Then when I was in grad school, I became a developmental psychologist with this in mind, actually, and got diverted into just studying cognition more generally for decades. I always had in mind this idea that if we can understand how people think and how they have emotional responses to understand we would be in a better position to be able to convey information in ways that they'll not only understand but take the heart that is acted on.


Rishi: So, do you mind just walking me through how a high school student comes into thinking about this stuff? Was it a teacher, or was an experience you had as a student that made you feel that your educational experience could have been better?


Stephen: It was partly that sort of thing but is also the '60s. It is just kind of it was in the water, is in the air. This idea of social change and trying to make things better. So it was just my particular take on observing how ineffective the high school experience at least that I had was and just thinking about how it could be better that led me to start really thinking about this. It wasn't any particular teacher or anything like that. It was really kind of I suspect more a sign of the times than any particular way great insights that I personally have.


Rishi: Now we've obviously come a long way since the '60s in terms of how we learn. In terms of active learning, do you feel like there are principles that are misunderstood or misguided in terms of how they are carried out? 


Stephen: Yeah. So I think a lot of people that I've talked to think of active learning is learning by doing, which leads in to think that just having an unstructured discussion is going to be active learning or this so-called discovery learning where we let people muck around and hopefully get something out of it, that's active learning. I think a better way to look at it is learning by using, that is you need to have a learning objective. Some particular goal in mind which what the students to learn. 


Learning outcome would be a successful result of having had a learning objective and on the basis of that learning objective, tell them or convey to them some information that would allow them to achieve it. But more than that have them use that information. So just telling people something isn't going to work, people forget things at an alarming rate and even if they don't forget it, they're not going to do it unnecessarily. 


So I think the first step is to convey the information but the next step which I think is vital is to have them use it in some way. That is they might use it in a debate or roleplay or problem-solving, something where they develop some kind of a work product that you can then evaluate and see if, in fact, they were able to use the information you conveyed in a way that will help them achieve the learning outcome.


Rishi: With active learning defined that way, I'm just thinking of classically how people are taught. So, they go to a classroom, they may take in a lecture, a video, and then they answer some questions. Is answering a question enough to constitute a work product, or is that not how you tend to think of the application of what is learned?


Stephen: So I think it depends on what the question is. If the question were just asking them to repeat back something they were told, sort of answer by word, I think it's fairly useless. If the question had some subtlety and involved having to think through the information and see how it could be applied in some way, then I think answering the question could very well be an instance of active learning.


Rishi: That's interesting because the way medicine is taught, oftentimes, is rote memorization and recall. That's fascinating that you're defining it that way, as needing to be more robust than that. Am I interpreting that right?


Stephen: Yep.


Rishi: With that said, do you mind giving me a broad overview of active learning sciences… what you all do, and how you help resolve some of these problems that come up for students and teachers?


Stephen: Sure. So the first thing is that we take a step back and ask, what the goal is? What the point is? What is what we want students to learn? That's where the learning objectives come in and learning objectives are not the same thing as topics. So a topic, it's kind of a heading in a book or something. A learning objective has a verb at the beginning typically like analyze, identify, or synthesize whatever so that you can actually measure whether they achieve it or not. So that's a crucial first step in any kind of active learning. 


The second piece is I think when you design an active learning exercise, you should be drawing on the science of learning. An absolutely enormous amount is now known about memory, how you acquire information, organizes and store it using various ways, and so on. And that information is pretty much stayed in technical journals and textbooks. It's not systematically used widely in education and it can be. 


So, I've been co-author of 4 textbooks which led me to have to read a lot of this nature. Not all of it by any means but I can't, if anybody has, I doubt it. I ended up stealing five principles that I summarized in a book I published last year called, The Active Learning Online, but the principles in there can be used for more than online. It can be used for any kind of learning. I can walk through those principles if you like.


Rishi: Yeah, please do. I think that'd be great. 


Stephen: Okay, so they fall into two big categories. The first one is pay attention and think it through. So let me ask you a question, if I may. At the end of the day when you're laying in bed, can you reflect back on the events of the day and so remember what happened?


Rishi: I can. Yeah. 


Stephen: Okay. So here's the real question I want to ask. What percentage of what you recall at the end of the day do you think at the time it was happening during the day, you intentionally tried to memorize it so you'd be able to recall it later on. Was it 50%, do you think? 


Rishi: Oh, far less. Intentionally memorizing?  I mean maybe 1%...probably 0. 


Stephen: Okay. I've asked thousands of people literally in big, large groups so I ask them to raise their hands and so on. In the modal number is about 5%. So think about this for a second, that implies that 95 percent roughly say of what you recall you didn't try to learn at the time it was happening. This is a well-documented phenomenon called incidental memory. So there's a distinction between incidental and intentional memory, which is incidentals by far the majority of what we remember. 


So why, how's that happened? Well, it turns out that the more you focus on something to pay attention to it, the more likely it is you'll remember it even if you don't want to, that a lot of memory is a byproduct of processing it. So the first principle may be the most important I called deep processing that you need to focus on what's important because what you're focused on, what you think through, pay attention to, is which are likely to remember later on. So the principle of deep processing is really fundamental. It's one of these pay attention and think it through. Another one is called deliberate practice which you may have heard of. Are you a musician by any chance?


Rishi: I'm not. No. I do play sports


Stephen: Okay. So what sports do you play?


Rishi: I like to play basketball and badminton. 


Stephen: Okay, so I don't know about badminton. Have you ever had a coach?  


Rishi: I have.


Stephen: So what a coach would do is spot the things you need to draw deep on to do better. So this works for sports like golf is a good example, tennis another one, we have a coach but also certain things [?] like language and learning and many, many things. I taught for a year in France so when I'm learning French, I had a tutor, I'd say a word and she'd listen really carefully and then say it back but emphasize the part that I got wrong, whether it's a rolling R particularly, U vowel sounds, whatever. 


The point is that deliberate practice is not just doing it over and over again such as practice. It's identifying the hard spots where it's difficult for you and disproportionately allocating your attention to that. So deliberate practice is important. You need to focus on what's hard not just do the whole thing over and over again. 


The third principle in this general category of pay attention and think it through is dual coding. So it turns out if you give somebody the name of something or a description of any Chauvin picture, they're going to remember way better than either showing a picture alone or the description alone. That is having two different modalities, the visual and the verbal. It is two different kinds of memory that are set up. So it has you, it has your memory. 


So deep processing, delivered practice, dual coding. Those are all examples of more, general pay attention and think it through category. The other two examples are completely different they're about creating connections. So one of them is about chunking that is we can only hold in memory maybe 3 or 4 big organized units but each of those can be organized into smaller units. 


Usually, when I do this, I give a visual demo but there's a great study that was done where they brought in an undergraduate as a case study. One student 3 times a week on average for a year and a half. What they did is they read random digits to this guy, 1 per second. They started with a single digit. They asked him to repeat it back which he could do. Then they gave him 2 random digits, and read them back, he could do it. In that first session was 7, okay? A year and a half later, you know how many he could recall? What's your guess?


Rishi: A year and a half of doing this? I would say that he could do…..I don't know. I'm going to say 14 -- double what he ended with.


Stephen:  79. 


Rishi: Wow!


Stephen: Impressive. Random digits in 1 second and here's what he did...


Rishi: That's incredible! 


Stephen: It is incredible. So this is done by Anders Ericsson and Bill Chase and Russel Fuller and was published in Science magazine years ago. So, here's what he did. He turned out to be a long-distance runner, and he knew the amount of time it took him for different segments of different races. So when he heard digits like 4 digits, he convert those into a time for a particular segment of a race and he build up a kind of mental structure of a fictional race that have these different segments put together. 


By doing that, he was able to chunked organized things into sets of 4, 3 or 4 and then each of those could be organized in sets of 3 or 4, and so on. So by the end of this thing, he comes up with these techniques for doing this that are just stunning.  Then they tested them among letters. In the end, at the very end do you know how many letters he'd recall, random letters, read 1 every second?


Rishi: I'm scared to guess. I mean if he did 79 of the numbers, I'm not sure.


Stephen: Seven


Rishi: Seven? 


Stephen: Just like the first day with the digits. 


Rishi: Wow!


Stephen: Yep. The strategy only works for digits. It's all about running times, about numbers. So it's about forming connections - chunking but drawing it on what you know, okay? That's relevant. So chunking is really important because we can only -- it's into the bottleneck. We can only hold 3 or 4 of these chunks. But each of them could have chunks within them. 


The final principle is about associations. So I just drew on that. He was using his associations to races to help him chunk but you can use associations, not only to help you organize the information but also to help you store it. So for example, when I meet a new person, I'm terrible with names. So what I've discovered is if I look at them and I look for features that are similar to somebody else who has the same name, some other John I know or Sheila or something. I'm mentally coding, "Oh, those eyebrows, Sheila". "Oh, yeah that nose, whatever", and then later when I see them I'd scan their face and it triggers the associations and reminded the other person I know who has those features and I can remember their name. So I've integrated what is new in a new person's face into what I already knew. 


The last thing about associations is you can use them to help you retrieve information, to jar your memory. If I can tell you a lot more about that, but I'm talking too long already. So as you can see, I think that the science of learning has done hundreds of hundreds of studies. For each of these things I'm talking about, there are many studies that document them. They're not just anecdotes. They're really laboratory studies been done that have led to these 5 principles that pretty much cover the main things that I think are applicable and those can be used really quite easily in classroom settings and they're often not.


Rishi: I guess one thing that strikes me is that some of these techniques you’re describe are things that I could do myself.  But one deliberate practice you called out is requiring the help of a skilled coach to point out areas of weakness to work on disproportionately. 


So I guess my question is, in our current educational system do you feel that the teacher-student ratios in classrooms -- one teacher for say 20 students or 30 students or whatever -- lends itself to allowing that sort of deliberate practice or coaching? Or do we fundamentally need to think of learning differently -- like an apprentice model where maybe it's one-to-one or something like that? 


Stephen: It's a great question. So one of the things that I've worried about a lot is how to scale because I want to avoid -- so there's a distinction between growth and scaling. Growth is we have more and more students, we'd have to have more and more tutors. So that's not so great. Scaling is we're going to be, we have the resources grow way, way slower resources we need, and the number of students. So scaling is what we want. We wanted efficiency and efficacy so take deliberate practice.


It turns out there are techniques for every one of those principles where you can get them. We don't need to grow the number of faculty or resources. Let's take that one in particular, Benjamin Franklin, who was in some ways, in many ways actually, arguably America's greatest scientist decided he want to learn to write. So what he did is he would read a newspaper where sometimes they were columns that were written by people who were very good writers. He would pick one that he was impressed by and then a couple of days later, write it down in his own words, paraphrase it, and then he would compare what he had written to the original and notice the differences. 


So he was able to identify himself where the rough spots were, and where the weaknesses were that he needed to drill down on. Similarly, Winston Churchill before he give a big speech gave it to himself in the mirror and he gets direct feedback about how he looked and how it sounded which allowed him to identify the rough spots on his own and correct them. 


Similarly, if you have a clear learning objective, you can develop a rubric where you literally tell students what they should be doing. In order to do it not so well, little better, better still, really good. You can develop objective criteria as long as you know what it is you're trying to achieve that is with them as for learning objective comes at. 


So what we've done is we've had students produce a work product like say they're learning to write effective memos and then we have grouped them pair set [?] work on this. They get a lecture, information reservation, then they go into active learning where they actually have to use the information to, in this case, develop a memo to justify why speaker series would be a good idea or something to get them talking to everyone. Then what we do is we have pairs come together and use a rubric that has the characteristics of a good memo, concise to the point, all of the relevant information, none of the irrelevant, whatever. I mean, we have a list of characteristics and they essentially give each other formative feedback and this can be scaled massively, you don't need individual faculty involved in it. And the very act of using the rubrics is so far the learning experience. 


So, yeah, deliberate practice, you can do it on your own identifying where your own rough spots are. I do it all the time when I'm studying music. It's not settled or I'm having trouble figuring that out pretty quickly. You can also use objective criteria that somebody who really knows what the end goal ought to look like is created and used that to help identify where the rough spots are incorrect. 


Rishi: Has research been done comparing near-peer practice, which is kind of what you're describing, with expert-guided deliberate practice? And if so, what are the comparisons of the types of practice? 


Stephen: So, that's a really, really good question. I do not know of any such study, but I do know that the techniques such as prescribed will, in fact, allow students to learn the material because we've used them in Foundry College. And they really do work, and actually students like doing them but we've not done the kind of comparison, which I think is a really good question. I would look at two things. I look at both the speed and efficacy. If it turns out it takes a little longer, I can live with that but if it turns out that it's not as good, that's something we got to consider what to do about. To my knowledge, I'm not aware of any direct study of this. If anybody out there, who's listening to this knows of one, please send me an email and let me read it.


Rishi: Yeah, absolutely. And you kind of segued into my next big question, which is tell me more about Foundry College. What you're describing is fascinating and obviously has a wide-ranging impact in terms of how we better teach and learn. What are you all doing, and how is it going? 


Stephen: Yeah. So Foundry College sort of came out of Minerva, which was the first thing I did. I was the founding Dean there. The first academic faculty member that they hired and I ended up being the architect for their curriculum and that is a fantastic, absolutely fantastic curriculum, but it's really quote-unquote "elite". Very difficult to take less than one point five percent of the applicants. I mean, it's really hard to get into it. It's wonderful for those students. 


So after five and a half years there, I left to do sort of the anti-Minerva which was for working adults, where virtually anybody can get in as long as you're serious about it. The goal was instead of a liberal arts education is to teach skills and knowledge that would not be easily automated. So the idea was to help them get jobs that would not be automated out of existence in a few years. So we answer analyses than relying [?] on the sorts of jobs we have in mind but that's what we did. 


So part of what Foundry College is about was - an insight I got in with a single friend of mine talking about a non-profit here in New York in the Bronx where they work with high school dropouts and teach them to fix cell phones. So I asked I said, "Well cell phones change all the time. What happens?" Do you know what happens after 3 years or so to these students? 

They lose their jobs because they were taught narrowly and vocationally, very brittle, how to fix this particular model, what to do, here are the steps. What they needed was a broader foundation in more general problem-solving skills, critical thinking skills, and so forth. So that's what Foundry College is about. 


Foundry College gives students a foundation and very broad general skills and then on top of that puts courses like project management, and entry-level salesforce administrator. Things like that that will get them a good job. In fact, they earn a third-party certificate, not ours. They actually get it through Salesforce, their own certificate. We train them up, but we don't just merely train them on getting that job. We give them a foundation that will help them adapt and evolve as the world changes.


Rishi: So, you're helping this cohort in their college years. I'm just curious, is there sort of a Foundry K through 12 equivalent? Or is there a curriculum that you think folks ought to be thinking about to get their kids ready for jobs that aren't so brittle and vocational like you're saying?


Stephen: Yeah. So Foundry is mostly focused on working adults. So there are few in there who are relatively recent from high school that's not really the main group that it is designed for. In terms of high school, I think a lot of what those foundational courses are at Foundry could just as easily be taught in high school. So critical thinking for example is one of them. 


Critical thinking turns out not to be one thing. I've identified seven different forms of critical thinking which have almost nothing to do with each other. So when you evaluate, whether you should believe in an argument, the steps you go for that are really different than what you do when you're weighing a decision. Should you do this versus that? You look at the trade-offs and that's really different than whether you decide with you can believe a source, what's their agenda and so forth. So critical thinking could easily be a year-long course in high school and the trick is to break down its components and give them enough experience with different contexts so they can actually generalize.


The single greatest problem, hardest problem in the science of learning is the problem of transfer that is they say, "What happens in Vegas stays in Vegas". I don't know whether that's good or bad, but I'll tell you the educational equivalent which is "what happens in class stays in class" is definitely bad. That is they don't transfer it out. 


They learn it as a kind of academic thing they're going to do well in the test they don't use it is typical. So in order to get people to use it, they have to have a lot of experience in different contexts. It's got to become more automatic. So yeah, a year-long course in critical thinking will be easy, a year-long course in problem-solving, a year-long course in decision-making, and many different models and heuristics people can look at and encourage a weighing of alternatives. All these kinds of things, there's no reason you can't teach them in high school.


Rishi: So as a final question…a lot of our audience are early-career health students or professionals. Any thoughts on things they should consider as they go out into the world and chart their path?


Stephen: Well, an obvious thing is that the world is constantly changing and evolving. So you got to have a mindset that you're going to be learning all the time. So learning is not fun for many people. It takes work. So to the extent that you can figure out a heuristics and that is in techniques that you personally are comfortable with and make learning relatively easy, like constructing narratives and scenarios where something would be relevant, once they have you. That's going to be good. It's gonna be better because you're really going to have to keep learning. That's the way the world was shaping up.


Rishi: Well, that's a good place to end and gosh, I am really, really grateful for your insights and my mind is blown about what is possible when you sort of chunk out information. I’m going to ponder that for a long time. Thank you for sharing your ideas with me.


Stephen: My pleasure. Thank you for your interest.

Rishi: I’m Rishi Desai. Thanks for checking out today’s show. Remember to do your part to Flatten the Curve and Raise the Line - we're all in this together.