Ep#24| Unpacking the Mind: A talk with John Sweller

Hello everyone, my name is Laura and welcome to a new episode of Edotech XP. Today's episode is very special. We have with us the renowned cognitive psychologist Dr. John Swaller, who is a professor at the University of New South Wales in Australia and has made groundbreaking contributions to the field of cognitive psychology through his work on the cognitive law theory. He will talk with Michelle about how educational technology relates to his theory and many more questions. Stay tuned. Welcome Mr. Swaller to our podcast today and thank you for being part of the Edotech XP podcast experience. It's a pleasure to be here. Thank you for inviting me. Could you please to get started, tell us a bit about yourself, about who you are and also your academic path? Right. I'm a cognitive psychologist who developed an interest in education. So I'm concerned with human cognitive architecture, how we learn, how we think, how we solve problems. And that led automatically and naturally and easily to an interest in instructional design. So basically I'm concerned with human cognition, human cognitive architecture and its consequences for instructional design. I see. And even now during our studies and reading a lot of papers while doing our masses, we read your name quite often already in citations or authors and author. And so it liked also to talk about one of your biggest research achievements I would call it and the cognitive law theory, which was correct me from wrong published in the late 80s. Right. The easy way to think about the theory is with respect to three branches. The first branch is the theory is concerned with how we learn, how we think, how we solve problems. In other words, in general terms, human cognitive architecture. What's the, how do our cognitive processes work? So that's the first branch. Once we've isolated the relevant processes of human cognition, the next thing is can we derive hypotheses that flow from that cognition, hypotheses with respect to instructional design. So that's the second branch using what we know about human cognition, using what we know about human cognition to tell us about to generate hypotheses, which might be of useful instruction design. And the third branch is the testing of those hypotheses. And all of the hypotheses without exception are tested using randomized controlled experimental designs. And that means we look at the way in which instructional design works now. We look at what the theory says we ought to be doing. And then we run an experiment in which we compare what we're doing now with what the theory says we ought to be doing. And if the experiment works, if it indicates yes, the hypothesis that's been generated by cognitive load theory results in better instruction, we have what we call an instructional effect. So the instructional effects are the third branch. And in some ways the most important branch, because ultimately there that's the branch that tells us, okay, this is the way you ought to engage in instruction. Or this is the way you should not engage in instruction. So those three branches constitute cognitive load theory. And obviously there's a lot more detail to that. And for example, relatively recently I guess about 10 to 15 years ago we introduced evolutionary educational psychology into the human cognitive architecture because that turns out to be really important. And when cognitive load theory began, we weren't aware of that importance. It's a relatively new discipline in any case. And that has resulted in different instructional effects and it's resulted in some explanations. And bits of the jigsaw puzzle that we're all working on began to fill in as a consequence of bringing that particular aspect in. But there are other aspects as well that have been brought in over the decades. Yeah, I guess that's also especially that fear that a human cognition is developing constantly because we know nowadays that we don't know everything so that it's getting more and more over time, right? That's correct. And in a way more importantly, when I first started out, human cognition, it had some importance, but we really didn't know enough about human cognition for it to have much of an effect or consequence for instructional design. Most instructional designers even today tend to ignore human cognitive architecture and that's understandable for historical reasons because human cognitive architecture, there wasn't enough known for it to be of any use. That changed literally in the last 10, 15, 20 years. There's a lot more known about human cognitive architecture and we now can use human cognitive architecture to generate instructional procedures which we really couldn't do terribly well 30, 40 years ago. I see. So, you know, I said that it's relatively new field and relatively new research direction. Now it would be interesting knowing what led you to the idea then back then to the cognitive lot theory and what inspired you to work in that field, to the research in that field. Because cognitive load theory, you know, since it wasn't invented, it was developed over a lot of the period of time. It's difficult to pinpoint exactly when it started, but I guess these days when I think back probably the learest point was I was running an experiment on human problem solving. I'm interested in human problem solving and learning and the relations between them. So, I'm running an experiment on human problem solving and I was using an artificial problem which went as follows. I gave the people I was testing a start number, let's say it was 14 or something like that. And I gave them a goal number, a number they had to reach and let's say that number and I'm making these numbers up now so it won't work with these particular numbers. But let's say the goal number was 53. And I told people the only manipulation you can do is multiplying by 3 or subtracting 27. Okay? The only moves you could make, you could multiply by 3 or subtract 57 as many times as you wanted. And I gave them those problems and they went very difficult problems. Most people could solve those problems. They could obviously solve two move problems but they could solve anything up to 10 moves without great difficulty. But one of the characteristics of the solutions was that I set it up so that the only possible solution was by alternating, multiply by 3, subtract 27, multiply by 3, subtract 27. So you had to do that either two times or four times or six times or ten times. Every single solution involved that alternation process. And as I said, there's people who I was testing. They didn't have any great difficulty solving the problems. But I noticed that when they reached the end, almost every single one of them had no idea that they had alternated the two moves. And I gave some people anything up to 16 of these problems. They did 16 problems. Every one of them they solved by alternating. And when I tested them at the end to see if they knew about alternation, no, I didn't know I was alternating. And I thought, well, that is educational consequences because we get kids to solve problems. What are they actually learning? Even if they solve the problem. And you would have sometimes, if you like doing puzzle problems, you may have come across this yourself. You work on it and you work on it and you finally get to the solution and you look at it and say, wow, how did I do that? You don't know how you did it. You know you did it, but if you asked to do it again, you may be able to do it a little bit more easily and more rapidly, but you don't really know how you did it. And I thought, why is that? And of course, I knew from my background about working memory. And I knew that working memory when it's dealing with novel information is exciting. extremely limited in capacity and duration. And what was happening is students were alternating, but all they were looking at is, okay, I'm here now, this is where I want to get, which move should I make? I'll look all multiplied by three here, or maybe I'll subtract by 27 here. And they don't look back and realize, okay, I'm getting closer and closer to the solution, but I'm alternating because they weren't concentrating on that and their limited working memory didn't allow them to attend to that particular aspect of the problem. And of course, in education, that's exactly what they ought to be attending to. What you really want in education, if you give somebody a problem, the reason for giving them a problem unless it's a test. The reason you solve a problem in order to learn is so that you learn the procedure. And they weren't learning the procedure. I think that's probably the closest I can get to the beginnings of cognitive loan theory. There are other aspects of the theory that are important because most people are unaware of the huge size of long-term memory. Now, a lot of us feel, oh, I've got a poor memory, I can't really remember very much. In fact, we can remember enormous amounts of information. We learn you're well of those bits of information which our working memory is attending to. And so that makes us feel we don't really know much, but we really do know a lot. I sometimes use the following example when I'm talking to people about the size of long-term memory. If I give you the following few sentences, the man went into a restaurant, the waiter spilt a soup on his lap. The waiter didn't get a tip. You understand that perfectly. It's not as though you're thinking what's going on here. That's strange. You understand it all perfectly. If you think about what has to be held in long-term memory to understand that, it's enormous. You have to understand a restaurant is a place you go to eat. A waiter serves you in the restaurant. A waiter is someone who serves you in a restaurant. The waiter gets paid for that service. Soup is a liquid form of food. People don't like that food dropping on them. They don't like it on their lap. A lap is a funny thing because it's only there when you're sitting down. If you're standing up or lying down or walking or doing almost anything but sitting, you don't have a lap. A tip is given to a waiter for good service. A tip is a financial reward. This is what we mean by the financial system. The list goes on and on and on. All of that's held in long-term memory. The reason you understood those two or three sentences is because of the enormous amount of information that you hold in long-term memory. I labeled the theory cognitive load theory because of its emphasis on the limitations of working memory. I could just as easily and it may well have been better if I'd labelled it something like, I don't know, long-term memory. Long-term memory is at least as important and possibly more important of the theory than working memory. As I say, we tend not to be aware of the immeasurable amounts. We don't know what the limits are of long-term memory. It's immeasurable. We tend to be unaware of those limits. Those two aspects of the theory, working memory and long-term memory, they go together. Talking a lot about working memory, I think I would also then be interesting to connect to that is that I personally now in university also learned a lot about the three different types of loads such as intrinsic load, extraneous load, and domain load, which are effect than our working memory. Could you please explain also to our audience what exactly is the difference between those three different types of loads? OK, let's start with intrinsic cognitive load. Some material that you learn is difficult to learn because it has a whole lot of elements that have to be processed simultaneously in working memory. If you're solving an equation, and let's say a simple algebraic equation, let's say a plus b all over c equals d sulfur a. Now, to solve that equation, you have to hold all of the elements together. You have to work out what the change is. And if you change one side of the equation, you change the other side of the equation, all has to be done simultaneously. That can easily exceed working memory capacity. And if it exceeds working memory capacity, you've got something that we call it high in elemented reactivity. Intrinsic cognitive load is very high. OK? Other things have very low intrinsic cognitive load. They may be very difficult, but they have a low intrinsic cognitive load. If you're, as you've obviously done very well, learning the some of the words of a second language, you can learn each one of those words independently of the other words. In other words, when you're learning the translation of the word cat, you don't have to think of the translation of the word dog. You can learn them separately. And then you can learn the next word and the next word and the next word. It's an immensely difficult task. And I don't have to tell you what takes years to do. But it doesn't impose a heavy working memory load. It's a difficult task for different reasons. There's a lot of information. There's a lot of elements, but they don't interact. So intrinsic cognitive load are associated with learning the vocabulary of a second language or learning the chemical symbols, ion is FE and copper is CU, learning that. That's low in elemented reactivity. Not an easy task, but it's low in elemented reactivity. Learning how to solve an equation such as A plus B all over C equals D solve for A. That's a lot of elements and you can't just learn one element at a time. You have to learn them simultaneously. So intrinsic cognitive load is high. So that's intrinsic cognitive load. Extraneous cognitive load also talks about elemented reactivity. But it's determined by the way in which you present the material. Cognitive load theory is mainly, not entirely, but mainly concerned with reducing extraneous cognitive load, unnecessary cognitive load. The cognitive load associated with learning in one way or being taught in one way as opposed to another way. Most of the cognitive load effects reduce extraneous cognitive load. So you can be taught by studying worked examples or you can be taught by solving problems. If you're taught by studying worked examples, that reduces extraneous cognitive load. If you're taught by having to solve novel problems, that increases extraneous cognitive load. So they're really the two divisions of cognitive load. Few words about germane cognitive load. That was introduced because we thought there was a cognitive load associated with the actual learning process. I don't use germane cognitive load anymore. When you look at the one and a half dozen or so cognitive load effects, most of them due to extraneous cognitive load, a few of them are due to intrinsic cognitive load. Not a single one, we haven't come up with a single one due to germane cognitive load. I no longer use that as a source of cognitive, as an independent source of cognitive load. So the two versions of cognitive load that I talk about, extraneous and intrinsic, I don't talk about germane anymore. Okay, that's also very interesting to hear because we still do hear about that and we still learn about it. It's also very, very nice and very interesting to see your opinion on that topic. Look, I still talk about germane cognitive load theory in a paper I published in 2010 in Educational Psychology Review. You can look that up. I talk about germane resources. I certainly would not be talking about germane cognitive load as being an independent source of cognitive load in that paper. I stop talking about that some time ago. Yeah, look, this is part of the development of cognitive load theory. Mostly we add things on very rare occasions. We realized, look, we put something in here, which we don't have any germane cognitive load effects. It has no function, so we tend to eliminate it. I also have another question regarding intrinsic and then extraneous load because we now also talked about reducing extraneous load, for example, how is it possible to measure that or is it even possible to measure cognitive load? Okay, look, it's possible, but it's not easy. Here's why cognitive load consists essentially of two or three different interacting segments or branches. Firstly, it's a function of the nature of the information, otherwise intrinsic cognitive load. Secondly, it's a function of how the information is presented. So the same material presented in one way has a lower cognitive load and present another way has a higher cognitive load. Thirdly and most importantly, it's a function of what you know, what's in long-term memory. For example, if I give you an equation like A plus B called C solve for A, you know immediately, okay, subtract B from both sides. Okay, A equals C minus B, you can do it in your head. It's a single element for you. For a junior high school student who's just learning algebra, there's a lot of elements there. That's high in elemented activity. So we have to take indoor count when measuring elemented activity, the nature of the material, what the person knows, how it's being taught. Simultaneously, currently it's not possible to do that in any accurate, simple way. What we can do and this works reasonably well is we can ask people how difficult did you find that problem or how difficult did you find it to learn this material. Fred Pass in the Netherlands came up with a scale. It's a very simple scale where you simply ask people exactly that on a one to nine scale. How difficult was this? Is it a two, which is pretty easy? Is it a seven, which is pretty hard or nine, which is really hard? People can subjectively give you an idea of how difficult it is, quite easily, quite readily and the scale seems to be reliable and valid. So when I measure cognitive load, I use the pass scale. There are other ways that are in development, psychophysiological measures, things like pupil size, things like link rate, that there's a variety of physiological measures that have been tried. Some of them seem to work reasonably well. They're difficult to organize. I mean, it's much easier to simply ask people who look on this one to nine scale, circle the number where indicating the difficulty. You can do that in 20 seconds. You don't need apparatus, but there are other measures which are being studied, I haven't personally studied in myself, but other people are working on those measures. And one never knows when easy to use simple measure, which is highly accurate, comes along. But as I say, we have that difficulty of determining cognitive load simply because it's not a single entity, it's several entities. From the beginnings of the cognitive load theory and the consequences of the human cognitive architecture for instructional design, to the multimedia learning principles and the possible future directions of this field, this interview was full of very valuable insights for the data tech community. Remember that you can follow the conversation about these another topics on Instagram, Facebook, on LinkedIn, thanks again, Professor Swaller and Michelle. And thanks to you, dear audience, for joining us. Until the next time, in Editech XP.