Category Archives: *Cognitive: Science of Learning

Better Learning: The art and science of teaching. From effective content delivery to supportive relationships.

John Sweller Interview 9: CLT – misconceptions and future directions

This is part of a series of blogs detailing a discussion that I had with John Sweller in mid 2017. See all parts of this series on this page

OL: What do you think is the biggest misconception about cognitive load theory that people have that you would really like to clear up?

JS: It’s only been very recently that people started taking notice of Cognitive Load Theory. For decades I put papers out there and it was like putting them into outer-space, you know, they disappeared into the ether! So, the issue of misconceptions in cognitive load theory didn’t arise. I guess the most common one is that really all I’m talking about is: ‘Don’t give students too much work to do at once’. That’s not really what I mean. It’s true; don’t give them so much work that they can’t get through it all. But that’s not what Cognitive Load Theory is about. What Cognitive Load Theory is about is you can teach the same stuff by reducing working memory load or by increasing working memory load and the issue is, how do you decrease it? And the whole purpose of decreasing it is so you can give them more information that is important. Cognitive Load Theory does not say: ‘Don’t teach them very much’. Cognitive load theory says: ‘Teach them as much as you possibly can because it’s important in any advanced industrial commercial society. But teach them in such a way that they can take it in without overwhelming working memory’.

OL: What are you most excited about in terms of where CLT is going at the moment?

JS: It changes whenever we have a brand new area and in the last few months we have a brand new area. There’s no publications I can point you to yet because there’s nothing out there but we’ve got experimental data. We’ve always assumed that working memory capacity was essentially fixed. The only thing that changes is what’s in long term memory. If you’ve got a lot of information in long term memory, bring it into working memory and you’ve got a huge increase in working memory. But other than that, working memory is fixed. It has become clear recently that we have what we called working memory resource depletion effects and that means, if you’ve been using your working memory, especially in a particular area, heavily for a while, after a while, and you would have experienced this yourself, your working memory keeps getting narrower and narrower and narrower and after a while it just about disappears. You may need to rest. You may even need a rest until the following day. Get sleep in between. That means that at rest, your working memory comes back.And we’re getting some data on that now. (Since this interview was conducted, some research has been published on this. See Greg Ashman’s summary of the article here, and the original article here)

OL: That’s interesting. And that relates to something else I’ve heard about. I’m sure you’ve heard about cognitive bandwidth obviously?  But the work of Sendhill Mullainathan, do you know that work? They did this really interesting experiment (see the paper here). They work with the impacts on working memory of being under financial stress. They did this experiment in a mall in New Jersey. They had two experimental conditions. In the first, they said: “Ok John, imagine that you just crashed your car and it’s going to cost $1500 to fix” and the other condition they said: “John, imagine you just crashed your car, it’s going to cost $150 to fix.” And then after they’d done that, they got you to complete some cognitive tasks, such as Raven’s Matrices.

JS: Oh, I think I see where this is going.

OL: What they found was that in those who weren’t under financial stress, the $150 vs. $1500 made no difference, but for those who were under financial stress, it made a big difference. So, for those with less money, their brain suddenly decided to process in its subconscious: “Oh, how am I going to do this? Where’s the money going to come from?’ And it had a big effect. So it’s kind of similar. But the other study Mullainathan did,  was they tested the working memory, or performance on Raven’s matrices, of farmers in developing countries. And they tested them before harvest, when they were still waiting and unsure of whether their crop was going to come to fruition, and then after, when they had the money. And they saw big impacts of that as well. So there’s all this stuff in the back of our heads that’s actually using up our cognitive bandwidth (working memory) and we’re just not aware of it at all.

John Sweller: And it may not be going on in the back of our heads. It might be going on right in the front of our heads. Two or three years ago I had an academic from Canada, Kris Fraser, who came on sabbatical to visit me and she was a medico, and she was looking at emotion and she got some really interesting results. She tested medical students practicing on plastic models. She had one group of people who had to learn to give treatment for whatever the condition was, but during practice, the patient died.  The other group, learned to give exactly the same treatment, but in this case, the patient lived and recovered. What they then they did was look at how much they had learned, and what they found out was that the people whose model had died had learned less.

OL: Yeah, because they were all stressing out about how they’d killed someone.

JS:  Exactly. So, as I was saying before when we were talking about motivation, you know, these things shouldn’t be mixed up. At that time, I began to think, this is still valid, ‘Well maybe motivation and emotion can be connected in this way?’. I obviously still haven’t decided whether these factors can be related.

OL: It’s occupying a working memory slot or number of slots of the seven plus or minus slots available.

JS: Exactly. If you’re worrying about the idea: ‘My patient died’. You’re not learning. If you’re worried about: ‘How do I afford 1500 dollars’, you’re not going to learn as much.

OL: If you worry about: ‘If I don’t do well on this tests I’m not going to be able to get into the uni course I want to do, etc etc’.

JS: Tell me about it!

OL: Well thanks you for your time today John.

JS: Oh good! It’s good talking to you!

You might also like to check out:

Education Research Reading Room Podcast Episode 9 with Andrew Martin on Load Reduction Instruction, Motivation and Engagement, also available on iTunes.

All posts in this series:

  1. Worked Examples – What’s the role of students recording their thinking?
  2. Can we teach problem solving?
  3. What’s the difference between the goal-free effect and minimally guided instruction?
  4. Biologically primary and biologically secondary knowledge
  5. Motivation, what’s CLT got to do with it?
  6. Productive Failure – Kapur (What does Sweller think about it?)
  7. How do we measure cognitive load?
  8. Can we teach collaboration?
  9. CLT – misconceptions and future directions

John Sweller Interview 8: Can we teach collaboration?

This is part of a series of blogs detailing a discussion that I had with John Sweller in mid 2017. See all parts of this series on this page

OL: Okay, Raj’s final question is: ‘In your 2011 book with Paul Ayres and Slava Kalyuga, you presented preliminary evidence for the collective working memory effect. How has the research on this effect developed since then? What instructional design implications does this effect suggest for designing group work?’.

JS: Yeah. One of my and Paul Ayres’ ex PhD students, Endah Retnowati who’s an academic back in Indonesia now, did a PhD on the collective working memory effect. It’s interesting because that work was started by our Dutch colleagues and she found interesting results that have just come out in the Journal of Education Psychology.

Most commonly, collaborative engagement occurs in an environment where you’ve got people with different knowledge backgrounds who are getting together. So, in a business environment, you might get an engineer and an economist who need to work together on a particular project. One knows about the engineering element the other knows about economic things that they can bring together, and they can’t do it without each other so they’ve got to collaborate. That seems to happen less often in a classroom. In a classroom, you would hope, everybody has a reasonably similar level of knowledge so collaboration doesn’t work in the same way there. We found that, for example, if people are problem solving, collaboration may work because one person may be able to help another person. But if they’re all studying worked examples, it has negative effects. You’re better off studying a worked example by yourself. You’re getting the information you need from the worked example. In problem solving, you may need information and the only place you can get it from is somebody else. So you have those sorts of complexities.

OL: That’s not a distinction I’ve thought about a lot in the past. Because a widely held view at the moment is that ‘We need to teach students how to be collaborative problem solvers’ (see the Education Research Reading Room podcast episode on this with Jan Owen here). But you’re suggesting that, if we want to talk about what happens ‘in the real world’,  there’s different experts coming in from different fields. I think that’s really valuable to consider.

JS: Yeah, look, that’s another issue related to the importance of the distinction between biologically primary and biologically secondary knowledge. I’m not sure to what extent you can teach somebody to act collaboratively, because humans have evolved to be social animals. We work collaboratively. We know how to do it. In a sense, what you and I are doing now is working collaboratively. Nobody sat down with either of us and said: ‘Ok, when you engage in this sort of collaboration, this is what you do and this is how you do it.’ Nobody did that or needed to do it.

OL: But there are things that you and I probably both learned over time that have enabled this collaboration to be a richer one. Like for example, I knew that if I came to you after doing a lot of thinking in my own time, with prepared questions and also asked other people their opinions, I’d be able to bring more high quality questions to this discussion. So there’s stuff like that I guess.

JS: Absolutely, and the point you’re making is an important one. Let me expand on it a little bit more. When we talk about biologically primary knowledge it’s not a suggestion that we don’t learn biologically primary knowledge. We certainly do and you certainly had to learn exactly what you just described. But you didn’t have to be told that in class. You didn’t have to have a,  ‘Today we’ll learn how to collaborate’ type thing. You do have to when teaching biologically secondary knowledge: ‘Look if you’ve got an equation (a+b)/c=d solve for a, the first move you ought to be making is multiplying out the denominator on the left hand side’. That you have to be told. You might pick it up yourself eventually but it’s a long slow process. What you just described that you learned, yeah you had to learn it, but my guess is you learned it pretty quickly without any explicit tuition, and that’s really what I mean by biologically primary. You know that if you’re going to have a discussion with somebody on a complex topic, it’s a good idea for you to prepare before and it may be useful for somebody to remind you occasionally: ‘Look, you really need to prepare beforehand. Before you go into this, if you’re being interviewed for a job, don’t just walk in there, prepare for it’. And people may need to be reminded of that periodically. You don’t really need to be told how to prepare. Make sure you know the topic that you’re going to be talking about sufficiently to talk about it. Make sure you give some thought to ‘What does the other person I’m talking to know and how does that relate to what I know and how does that affect how we’re going to hold the conversation?’ But you do all that automatically. There was never a module in class, and if somebody set up such a module at any level teaching ‘This is how you collaborate with people’, I think they would be wasting their time.

OL: Could it be the case for some people that this is more biologically primary than for others?

JS: Oh, absolutely! It’s not that it’s more biologically primary, but rather, for all skills, all biologically primary, probably secondary ones as well, but certainly biologically primary skills, you get a normal distribution. You get some people way on the left and they’re very poor at it, and some people may, for genetic reasons, be very poor at some skills. And they’re likely to be people who we feel may end up with societal problem because their skills are so poor in that area. Other people will pick up these skills easily. And you can say that for every single biologically primary skill. It varies. The fact that it’s biologically primary doesn’t mean we all have it to the same extent. Or even if we all have it. Some people don’t have it and they generally face challenges in life as a result.

OL: So in the same way that for some people who are on the bottom end of the distribution when it comes to being able to acquire speech, and it aids them to have a speech pathologist to work with, for example. Could it be argued that for those who are on the bottom end of the distribution with respect to group work or collaboration, or having discussions, or not hitting people in the face, or whatever it may be, that it may help them to have explicit instruction on how to relate to other people?

JS: Yeah that’s exactly right. Look, at the extreme end we’ve got kids on the autism spectrum. Those kids frequently have much greater challenges learning how to relate to other people. You know, most non-autistic people just pick it up automatically. Autistic kids won’t. And can they learn to do it? Yeah, they can learn to do it, via the biologically secondary system. It’s a slow clumsy way of learning it, but it’s the only way they may have. You can have some high performing autistic people whom you don’t really know they’re autistic because they spent years learning how to: “Oh, this is what I’m supposed to do in this social situation. Oh right!” They need to be explicitly told: ‘Don’t say this in this situation.’, ‘Do act in this way in this situation’. To them, it’s all new.

OL: Therefore, if it’s possible to take people who were at the bottom of this distribution and, through explicit instruction, move them more towards the centre, is it feasible that we could identify the skills and attributes of those at the top end of the spectrum and use explicit instruction to move those individuals in the middle of the distribution more towards the high end?

JS: People in the middle of the distribution, you probably can’t move towards the high end, it just means they have learned the skills a little bit slower or they’d be a little bit older when they learned the skills but they acquire the skills and they automate these skills once they got them, there’s nothing you can or need do. They’re already there and we know, for example, that somebody who’s not terribly socially competent in middle primary school, by the time they get to secondary school they’re better at it. Or somebody who has difficulty with relationships when they’re teenagers, by the time they’re in their 20’s they don’t have the same difficulty with relationships any more. Other people have difficulty with relationships all their lives.

OL: Yeah, and that’s my question I guess. Because, one could propose, the difference between a working marriage, where they stay together and it’s generally a positive experience, and a thriving marriage for example, might be a few daily actions. For example: expressing gratitude, acknowledging when you notice what the other person does a nice thing, ‘I notice you vacuumed the floor, thanks for doing that.’ Things like that. And I really think in situations like that, explicit instructions on those things, which is essentially going to therapy, could be beneficial.

JS: You’re absolutely right. Those little things, we can tell somebody something and they’d say ‘Oh right, why didn’t I think of that?’

OL: So I’m curious, with that as a background, I’m curious as to why you still think that it’s a waste of time for people to try to teach these little things that could measurably improve collaboration, because a relationship is essentially collaboration.

JS: Look, you probably can. But the amount that has to be learned and the difficulty of learning it is probably fairly minor. In other words, there’s no working memory load associated with it. You tell somebody ‘Look, you should say thank you a little bit more.’ I remember one of my daughters, when she was younger, she rarely smiled at anybody. We told her, ‘Look, people think you’re grim all the time. Just smile a little bit.’ And that’s all it took. What I just told you was the limit of it, and that was okay.

OL: I know what you mean. If you try to run a course on some of these skills, half the students are going ‘Oh my goodness this is the most obvious thing you could ever talk about’, and the other half might not be ready to learn it yet. These kind of lessons have to come to a person at the right time for them to take it on. That’s my impression.

JS: You’re right. We always have to remember that all skills follow a normal distribution, and some people, for genetic reasons, take a bit longer to learn and it might be better to teach them something. The only concern I have is some people place such an emphasis on these skills, and they can see that these skills are much more important than anything they teach in school, because you can completely wreck your life if you don’t have these skills. So they say, ‘That’s what we ought to be concentrating on’. I’m worried about it taking away too much from learning mathematics or other subjects because it is something which, for most people, is easily learned. And yeah, they might need a little bit of help, but that’s all you need. Sometimes, most often that help simply comes from your family but some kids are in the unfortunate situation where they don’t really have a family who can provide suitable information to them.

 

Next post:

9. CLT – misconceptions and future directions

All posts in this series:

  1. Worked Examples – What’s the role of students recording their thinking?
  2. Can we teach problem solving?
  3. What’s the difference between the goal-free effect and minimally guided instruction?
  4. Biologically primary and biologically secondary knowledge
  5. Motivation, what’s CLT got to do with it?
  6. Productive Failure – Kapur (What does Sweller think about it?)
  7. How do we measure cognitive load?
  8. Can we teach collaboration?
  9. CLT – misconceptions and future directions

John Sweller Interview 7: How do we measure cognitive load?

This is part of a series of blogs detailing a discussion that I had with John Sweller in mid 2017. See all parts of this series on this page

OL: These questions come from Raj who I met a few weeks ago. He lives up in Queensland. He’s just finished reading your book. I think it was that one (pointing to: Sweller J;Ayres PL;Kalyuga S, 2011, Cognitive load theory, 1st, Springer, New York)

JS: Yeah, okay, my sympathies to him

(both laugh)

OL: Yeah he’s a diligent guy. He used to be an engineer he’s just moved into education now.

(First question omitted due to overlap with previous content)

So the second question is: ‘Measuring cognitive load has been done by self-report, secondary tasks and physiological markers. What are the weaknesses of these as measures for cognitive load? And which measure (or combination of measures) is the most valid and reliable?’

JS: Ok. I routinely use self-report and I use self-report because it’s sensitive. That’s the first thing you need. You need something that’s reasonably sensitive to what the cognitive load is. You can have something that can measure cognitive load but you might need vast differences in cognitive load to use these tools. A cognitive load measure that only is sensitive enough to distinguish between someone who is almost asleep as opposed to someone who is concentrating is not much use.

OL: What questions do you ask?

JS: I use the Paas 1992 test which is ‘How difficult did you find this instruction?’ Paas uses his original version which is something closer to, ‘How much mental effort did you put into studying this material?’ They’re very heavily correlated. It’s quick, takes literally a few seconds, very sensitive.  You can easily get differences between conditions using that test. The other one that’s sensitive is secondary tasks. They’re just as good as using self-report but they’re messy to set-up. You’ve got to set-up some sort of system to allow the secondary task and sometimes the system that you set-up might involve apparatus and technology which in a normal classroom is pretty difficult to put in. The last one, physiological markers, have had endless work on them. They don’t work yet.  We don’t have physiological markers which are sensitive. We’ve got physiological markers which will demonstrate the difference between really studying something hard as opposed to staring out the window. That’s no use to us. We need something that really distinguished between, say, studying worked examples vs. solving a problem or studying a worked example in a split attention format as opposed to an integrated format. Paas’ question and secondary tasks will do this. The physiological markers, despite endless efforts by a lot of people around the world for years and years and years, don’t really work.

OL: Yeah, that makes sense. Ok, Following on from 2, ‘Which CLT effects do you have the most/least confidence in given the evidence base?’

JS: Oh, that’s unanswerable in a way. I mean, the ones that I list in the book, I’m confident in them. And I’m confident in them because my students and I have carried out several experiments which have demonstrated them. People around the world have carried out experiments. If you want a popularity contest then the one that people have most commonly been studying is worked examples. People all over have tested and done that. But split-attention is important, redundancy is important, the transient information effect, which is a very new one, that’s important. So ultimately my confidence is based on whatever data is available and I sort of have this vague subjective opinion as to which effects have been more studied than other effects. And if something has been studied a lot.

OL: It’s probably more robust.

JS: Yeah exactly.

Next post:

8. Can we teach collaboration?

All posts in this series:

  1. Worked Examples – What’s the role of students recording their thinking?
  2. Can we teach problem solving?
  3. What’s the difference between the goal-free effect and minimally guided instruction?
  4. Biologically primary and biologically secondary knowledge
  5. Motivation, what’s CLT got to do with it?
  6. Productive Failure – Kapur (What does Sweller think about it?)
  7. How do we measure cognitive load?
  8. Can we teach collaboration?
  9. CLT – misconceptions and future directions

John Sweller Interview 6: Productive Failure? – Kapur (What does Sweller think about it?)

This is part of a series of blogs detailing a discussion that I had with John Sweller in mid 2017. See all parts of this series on this page

OL: Now,  the next thing I want to ask you about is the work of Kapur, and Andrew alludes to this in his paper as well (as mentioned in the previous post). Some would argue that the results of Kapur suggest that in some case it is actually better to enable people to engage in approaches that lead to overburden on working memory and subsequent failure to achieve the intended outcome, if you then follow that up with a gap filling kind of instruction, so where are you at in terms of that at the moment? (Hear Andrew Martin’s take on productive failure in this podcast)

JS: Okay, here’s my problem. There are ways of testing the productive failure hypothesis legitimately. Now I have to talk about experimental design here. We’re talking about randomized controlled trials. Most people in Education, when they run a randomized controlled trial, get the randomization done fine. The control part of it they tend not to. And the whole point of using a randomized controlled design is to determine causality because you only alter one variable at a time. If you alter multiple variables simultaneously, you can’t determine causality and it’s a waste of time running that sort of experiment. So there are certain things you can’t do. For example, you can’t do something like: give one group of students, let’s say, a lecture which involves guidance; and another group of students problem solving, which reduces guidance; and then test them and find out which one is better. And the reason you can’t do that is because you are altering multiple variables simultaneously and so can’t know why you got an effect. Let’s concentrate on the lecture to begin with. Assume it’s a really brilliant lecture, magnificent lecture, very clear, students understand it. They are highly motivated. Is that going to be better than problem solving? Almost certainly the lecture is going to be better. If it’s a really poor lecture, students don’t understand what’s going on, it’s disorganized, they’re bored out of their minds, then studying a problem, no matter its deficiencies, is going to be better. You shouldn’t do those sorts of experiments.

There are experiments you can run, and they’ve been running in a different context without even attempting to test the productive failure hypothesis. The obvious one is to compare worked examples followed by problems to problems followed by worked examples. Well, on the productive failure hypothesis, you’d be better off giving people problems first, followed by worked examples. Now you’ve got exactly the same conditions for both except for one variable; you either start with problems or you end with problems. You either start with worked examples or you end with worked examples. So, you compare worked examples followed by problems to problems followed by worked examples. If you do that, and you’re using novices who really need the worked examples, the results are uniform. Worked examples followed by problems is always better.

If you’re using somebody who no longer needs the worked example, you see the expertise reversal effect. As expertise goes up, the advantage of worked examples go down, and as expertise continues to go up, eventually the relative effectiveness of worked examples and problems reverses and the problems are more helpful than worked examples. As you know, once you learn how to solve something, you need to automate it, you need to get practice, you need to be able to do it without thinking about it and at that point you’re better off solving problems rather than studying worked examples.

If you get people who are sufficiently knowledgeable, yeah you’ll get a different result, they’re better off with problems first. But, come what may, you have to have experiments that only alter one variable at a time. And that variable has to be the one that’s of interest to you.  Interestingly, in the old days, decades ago when computers first started appearing in education, people immediately asked the question: ‘Oh gee, is it better to give people a lecture or have them get computer assisted instruction?’ That’s a nonsense question! You cannot test that out. A good lecture is much better than bad computer assisted instruction and vice versa. As it happens, you can test out the productive failure hypothesis, but when you test it out properly, only altering one variable in your experiments at a time, you get different results to some of the ones reported. That’s my problem with productive failure.

Next post:

7. How do we measure cognitive load?

All posts in this series:

  1. Worked Examples – What’s the role of students recording their thinking?
  2. Can we teach problem solving?
  3. What’s the difference between the goal-free effect and minimally guided instruction?
  4. Biologically primary and biologically secondary knowledge
  5. Motivation, what’s CLT got to do with it?
  6. Productive Failure – Kapur (What does Sweller think about it?)
  7. How do we measure cognitive load?
  8. Can we teach collaboration?
  9. CLT – misconceptions and future directions

John Sweller Interview 5: Motivation, what’s CLT got to do with it?

This is part of a series of blogs detailing a discussion that I had with John Sweller in mid 2017. See all parts of this series on this page

OL : So, CLT has a complex relationship with motivation. This is out of some more of Michael Pershan’s writings on CLT. Pershan referenced your work with Merrienboer from 2005 and how that paper talked about four major developments in current CLT research. One of these developments was that we should take learners’ motivation and their development of expertise during lengthy courses or training programs into account. Indicating that that was something that CLT was starting to look at. But then, Pershan points out how, in 2012 when you were interviewed, you said: “One of the issues I faced with Cognitive Load Theory is that there are at least some people out there who would like to make Cognitive Load Theory a theory of everything. It isn’t. […] It has nothing to say about important motivational factors…It’s not part of CLT.”

 So in terms of right now,  where is your thinking in terms of bringing together CLT and motivation?

JS: Motivation is incredibly important. I want to emphasize that before saying what comes next. I don’t think- and in the earlier days I thought it may be possible-I don’t think that theories of motivation should be mixed up with theories of cognition. It doesn’t mean that motivation is not important. It just means that I don’t think you can turn Cognitive Load Theory into a theory of motivation which in no way suggests that you can’t use a theory of motivation and use it in conjunction with cognitive load theory. So it’s really another way of saying: Look, you can make anything either motivating or demotivating. You can give people problem solving, learning through problem solving and a lot of people say, ‘one of the reasons we use problem solving is it’s motivating’. It’s motivating sometimes, for some people, but it’s terribly demotivating for others at other times. A lot of people who become math-phobic become so because they’re demotivated. “Yeah I’ve tried to solve this problem but I couldn’t, I don’t want to see maths ever again!” Same with worked examples. You can organise worked examples in such a way that people are completely demotivated. “I don’t want to look at these worked examples. It’s boring!” You can make it interesting. Anything you teach can be made either interesting or boring. It could be motivating or non-motivating. And it’s got nothing to do with Cognitive Load Theory.

When you look at the alternatives we look at in Cognitive Load Theory, all of them could be either motivating or non-motivating and should they be motivating? Yeah! And motivation is absolutely critically essential to everything, like when you’re standing in your class and you’re using perfect instruction from a cognitive load perspective and its being done in such a way that the kids are staring out the window bored, it doesn’t matter how good the cognitive load instruction is. They tend to be separate and we haven’t really managed to do a great deal with it.

At times it looks as though we might be able to do something in that people have suggested, look, lack of motivation simply means that you’re using your working memory resources on something other than what you ought to be working on. In other words, instead of using your working memory resources on the Mathematics, you’re using your working memory resources on what’s going on outside the window. There may well be a connection there. But it hasn’t gone very far yet. If it goes somewhere then I’ll change my mind as you might gather just by looking at the history of Cognitive Load Theory. I’ve tend to flip flop on this. Right now, I’ve got my doubts simply because of the history of the sort of failure to be able to bring motivation into the cognitive structures that we use in Cognitive Load Theory. It suggests to me, yeah we ought to be looking at motivation, but it’s a separate topic. I understand from your email that you’re going to be talking to Andrew Martin (podcast available here). Andrew, as you know, is a motivation researcher. He’s talked about motivation and Cognitive Load Theory. I don’t think he’s attempted to sort of unify them into a single theory.

OL: That’s pretty much what he’s been doing. I’ll give you a picture of a diagram that comes from his work because I’m sure it will interest you.

Andrew Martin CLT Motivation and Engagement Ollie Oliver Lovell

OL: It’s from his paper called ‘Using load reduction instruction increase motivation engagement’. So, he is talking, in some ways, about potential causal relations but he does say several times throughout the paper ‘This is not prescriptive nor is it exhaustive’. This is simply suggested relationships between these various constructs and instructional approaches, all of which require further research. That’s what I took out of it.

JS: Yeah, I should ask him about some of the relations in the diagram that I don’t understand. For example, he’s got disengagement here. But I don’t know what it means to have the line connecting disengagement to using different modalities.

OL: I think he thinks that if you use different modalities if you avoid redundancy and if you allow appropriate instructional time; that can help to boost engagement.

JS: Yeah but I’ll need to find out why particular cognitive load effects are related to particular emotions. I need to ask Andrew why, for example, different modalities would address disengagement while mental practice, worked examples, guided practice wouldn’t. I’m sure there is a good answer.

OL: Let’s just jump up quickly…let’s just find ‘disengagement’ in the paper.

Disengagement is complex and can arise for many reasons (Finn & Zimmer, 2013). It may be that the student lacks particular skills in a domain such as literacy or numeracy, or self- regulation skills such as study and organisational skills (Covington, 1992, 2000). In some cases there are motivational problems such as low self-efficacy (Bandura, 2001), low valuing of the domain or tasks within it (Wigfield & Eccles, 2000), or uncertain control leading to helplessness (Abramson et al., 1978; Weiner, 1985). From a cognitive psychology perspective, it may be a function of the instruction or task itself that over-burdens some learners’ cognitive capacity or renders the instructional material uninteresting and repetitious, leading to abandonment of effort (Sweller, 2012). Approaches under the LRI (Load Reduction Instruction) umbrella can be a means of addressing many of these factors that can underpin disengagement. Here the discussion centres on using different modalities, avoiding redundancy, increasing coherence, and providing appropriate instructional time. (pg. 33)

JS: Yeah, I understand that.

OL: So here it is, ‘…can be a means of addressing many of these factors that can underpin disengagement.’ But I see what you’re saying. That could be said for all/most of the constructs in the motivation and engagement wheel pictured.

JS: Exactly. And I need to find out from Andrew why some constructs go together but others don’t.

OL: Ok that’s good. So at the moment you feel that they’re relatively orthogonal?

JS: I suspect so but I’m sure Andrew has some reason for connecting particular constructs.

Next post:

6. Productive Failure – Kapur (What does Sweller think about it?)

All posts in this series:

  1. Worked Examples – What’s the role of students recording their thinking?
  2. Can we teach problem solving?
  3. What’s the difference between the goal-free effect and minimally guided instruction?
  4. Biologically primary and biologically secondary knowledge
  5. Motivation, what’s CLT got to do with it?
  6. Productive Failure – Kapur (What does Sweller think about it?)
  7. How do we measure cognitive load?
  8. Can we teach collaboration?
  9. CLT – misconceptions and future directions

John Sweller Interview 4: Biologically primary and biologically secondary knowledge

This is part of a series of blogs detailing a discussion that I had with John Sweller in mid 2017. See all parts of this series on this page

OL: This is a question from Michael Pershan. He’s done a great summary of some of your work online . I’ll have to share with you because it breaks it down a lot of great ways. He’s a maths teacher in New York. He’s got a great blog. He wants to know: ‘Does the distinction between biologically, primary and secondary knowledge produce testable predictions?’

JS: Yeah it does. And we haven’t made nearly enough of that. The main prediction is that – it’s almost a description of the two – you acquire skills differently depending upon whether they’re biologically primary skills or biologically secondary skills. There’s a very clear difference in the way in which you acquire them. We’ve just been talking about explicit instruction and the real reason we had a problem with people insisting you don’t need explicit instruction is precisely because that distinction (between biologically primary and secondary learning pathways) was one they weren’t aware of.

OL: I remember I read of this distinction first in your Story of a Research Program paper. And when I read that distinction I was like, “This makes so much sense.” All those arguments people always making about ‘Yeah, but how did you learn to talk? But how did you learn to walk?’

JS: That’s exactly it! And we learn all those biologically primary things without explicit instruction and not only that, it would be absurd to provide somebody with explicit instruction. If somebody come along to you and said: ‘You want to learn English. This is what you do with your tongue. This is what you do with your lips. This is what you do with your breath. This is what you do with your voice. This is how you speak in English’ (Laughs). Well, you know, that’s just silly. But that’s exactly what we do with writing. You’ll say: ‘Okay, want to write the letter a? First you write a circle then you write a line on the right-hand side.’ That’s exactly what you do. We don’t do that for speaking. And that’s testable and you can build on it and say: In effect, all of the cognitive load effects are testing precisely that because we’re saying: ‘Okay, if you trying to teach biologically secondary material as though it’s biologically primary, it’s not going to work properly.’

A lot of people, including people working in Cognitive Load Theory, because that distinction came a little later, see the biologically primary-biologically secondary distinction as being some sort of optional extra that you don’t really need. You need it very badly. Until that distinction came along, there was a missing part of the jigsaw puzzle. I knew that, ‘ok people say that you can just learn naturally in the classroom’, but I thought, “Well, yeah but it doesn’t seem to work like that when I run an experiment. I can’t get data for it!” And  then all of a sudden, there was the jigsaw puzzle piece.

This biologically primary/secondary distinction came from David Geary. Once you have realized the distinction between biologically primary and secondary, you then need to look at the cognitive architecture associated with biologically secondary material because that’s the material that’s taught in class. You know, when you were teaching, you’re teaching biologically secondary stuff. And there’s a cognitive architecture associated with that (see this podcast with Andrew Martin for more on this cognitive architecture). And that cognitive architecture, those cognitive processes, are different from the processes that are used when acquiring biologically primary material. We’ve touched on some of those differences in the discussion so far, but beyond that, things like, for example, the limitations of working memory that we have when acquiring novel material, that certainly applies to biologically secondary information. I’m not sure it applies in biologically primary. My guess is that it doesn’t, or if it does apply, it doesn’t apply to the same extent. You can take in a lot more biologically primary information than secondary. You know, Miller’s magical number seven plus or minus two. I don’t know what it would be for somebody who’s learning biologically primary information, but I suspect there’s much more than that. I mean, consider our ability to learn and recognise faces. I don’t know how many pieces of information there are that we take note of when we’re looking at a face, but it’s a lot. And we do it simultaneously. Just… there it is! We recognise people. The same goes with language and with everything else that is biologically primary.

OL: There are two exceptions that I’ve thought of to what we’ve just been talking about. The first is—and I’ve just made this connection— I think that perhaps there are limits to working memory in acquiring some biologically primary skills. The example I can think of in this case is, have you heard of motherese? It’s the simplification of language by mothers in order to make it more comprehensible to their young. So that’s perhaps an example of how there could potentially be limits to working memory (on the part of the baby acquiring language) in the context of biologically primary skills.

JS: Yeah, that may be true but, notwithstanding that, you think of what, even in the simplified form, when a mother is saying to the child: “This is a pussycat.” The child can take all of those sounds in. For example, if you heard somebody, assuming you don’t speak Chinese, if you heard somebody say in Chinese: “This is a pussycat.” All you’re going to hear is a sort of sound modulation and if you are asked to repeat that you couldn’t do it. There’s a huge amount of information there. I suspect even if they can’t speak very well, imagine a young kid whose Mommy says: “This is a pussy cat.” They can imagine that speech in their head even if they haven’t quite got the tongue, lips, etc working to actually say it.

OL: Somehow they get it. That’s actually a perfect segue into the next point that I wanted to talk about. And that is, maybe there is a window of time for this specific biologically primary learning. Because for example, I have learnt Chinese, as an adult.

JS: Oh okay! (laughs)

OL: And in order to produce the sounds correctly, I actually literally had to study mouth positions. So it’s moved out of the realm of primary into the realm of secondary, because I’ve missed that developmental period for that language.

JS: That’s actually correct. Before our interview I talked with you about second language learning with my French colleagues, and that’s exactly the problem. People assume that we learn a second language as an adult in the same way as we do as a child, and immersion as a child works perfectly but we’ve evolved for it to work perfectly as a child. For an adult, learning language is biologically secondary. It’s not biologically primary.

OL: It’s funny isn’t it?  How it can switch category. It’s just amazing.

Note: Post 8 of this series also includes a discussion of biologically primary and secondary knowledge in the context of collaboration.

Next post:

5. Motivation, what’s CLT got to do with it?

All posts in this series:

  1. Worked Examples – What’s the role of students recording their thinking?
  2. Can we teach problem solving?
  3. What’s the difference between the goal-free effect and minimally guided instruction?
  4. Biologically primary and biologically secondary knowledge
  5. Motivation, what’s CLT got to do with it?
  6. Productive Failure – Kapur (What does Sweller think about it?)
  7. How do we measure cognitive load?
  8. Can we teach collaboration?
  9. CLT – misconceptions and future directions

John Sweller Interview 3: What’s the difference between the goal-free effect and minimally guided instruction?

This is part of a series of blogs detailing a discussion that I had with John Sweller in mid 2017. See all parts of this series on this page

OL: What are some of the fundamental differences between goal-free activities and minimally guided instruction?

JS: One of the reasons we switched from goal-free problems to worked examples is that goal-free problems work really well, but in a very limited number of areas. They don’t work well in all areas. The areas they work well in are areas where if you tell people: ‘Calculate the values of as many variables as you can’, there are only a very limited number, maybe 3, 4, 5 variables, that you can calculate. In other words, if you really know what you’re doing you can calculate everything very very quickly. The other areas, like in some areas of mathematics but also areas in other disciplines, if you tell people: ‘Do as much as you can’, they’ll be going from now to infinity! There are literally an infinite number of things you can do. So you just cannot do that. An example I use is to ask people to consider something like simple algebra. Give someone an algebra equation and say ‘manipulate this algebra equation in as many ways as you can’. There’s an infinite number of ways.

OL: You end up with the ‘x’ on one side by itself in only a small number of these combinations.

JS: Exactly. We’ve never run a goal-free problem using that sort of material. In most geometry areas it tends to be limited. It works in some areas of physics. Especially in physics, word problems like calculating velocity of something, acceleration, or time. If you tell someone: “calculate everything”, they’ll run out of things to calculate very, very quickly. They can do it and, all of a sudden, they find: ‘Oh  right, I’ve just calculated what this question was asking’. I’m thinking of the experiment you proposed a little while ago (see this post). But in other areas you can’t use it. Worked examples on the other hand, work everywhere. All the way from limited mathematics areas to…

OL: Shakespeare?

JS:  Shakespeare, yeah. That’s why we put the emphasis on worked examples rather than goal-free problem solving.

OL: Is there a key difference between—I understand what you’re saying in terms of how the goal-free effect is limited in its scope—but is there a fundamental difference between that and minimally guided instruction?

JS: Probably not. There is no instruction other than “calculate whatever you can”, but the reason it works is the reason I just outlined. There may not be any instruction, but there’s not much you can do anyway.

OL: Yeah okay. It’s within a kind of bounded region of exploration.

JS: It’s very very bounded, and generally speaking, if it’s using motion equations you just look at your 3 or 4 equations, you’ve got an unknown in each one of them and you just say, “Ok. I’ll try this, I’ll try this, I’ll try this.” And at the end of the day, not only that but everything you try will teach you something that you need to learn. For any of these equations, you have to be really good at it.  You have to be able to calculate any unknown at any time. You can always be given a problem which will require you to calculate this unknown and then the next one this other unknown, etc, etc.

OL: Got it. What I’m taking away from that is: Goal-free kind of approaches and minimally guided approaches can both be effective within a bounded set of examples if also students record what they’re doing in a clear way and then reflect upon it.

JS: Students need to reflect on it and can I reiterate strongly that the reason worked examples work is because you’re asking them to reflect on it. In effect you’re saying: “Study the worked example”. That’s another way of saying: “Here’s a problem solution, reflect on it.” Okay, “You didn’t calculate it yourself but it doesn’t matter”.

OL: It doesn’t matter who did it, as long as they’re reflecting afterwards?

John Sweller: Yeah.

Next post:

4. Biologically primary and biologically secondary knowledge

All posts in this series:

  1. Worked Examples – What’s the role of students recording their thinking?
  2. Can we teach problem solving?
  3. What’s the difference between the goal-free effect and minimally guided instruction?
  4. Biologically primary and biologically secondary knowledge
  5. Motivation, what’s CLT got to do with it?
  6. Productive Failure – Kapur (What does Sweller think about it?)
  7. How do we measure cognitive load?
  8. Can we teach collaboration?
  9. CLT – misconceptions and future directions

John Sweller Interview 2: Can we teach problem solving?

This is part of a series of blogs detailing a discussion that I had with John Sweller in mid 2017. See all parts of this series on this page

OL: The next little thought experiment I wanted to test out was based upon your goal-free effect. My understanding is that the initial work surrounding goal-free effect was on physics problems. You could have explored things like, for example, parabolic motion, and asked participants to determine the maximum height of a projectile given an initial launch angle and velocity or something like that. You essentially found that in that context, if you were to say to a group of students, ‘Here’s a launch velocity, here’s a launch angle, work out the maximum height’, then you took the other experimental group and you said: ‘Find out as much stuff as you can given this initial launch velocity and angle.” The group to whom you said ‘find out as much stuff as you can’ would be more likely to find out the maximum height than the group to whom you explicitly said ‘find the maximum height’. Even more surprisingly, if you subsequently gave each group the problem ‘Find maximum height’ in a similar context, the group who had the goal-free question in the first condition would also be more likely to find the maximum height under the goal-specific condition after the initial experimentation phase.

I was wondering if you think it would be at all possible to teach learners to turn problems from goal-specific problems to goal-free problems themselves? For example, perhaps we could conceptualise a general problem solving strategy as: ‘If you don’t think you can do the problem, what you should start doing is to just find stuff out’. Then say, ‘Every 5 or 10 minutes, if you see you’ve worked some stuff out, you should look back, and think about where you were trying to get to in the first place. If you can see how to get there, great. If you can’t, just keep playing around and find some more stuff out.”

JS: That’s smart. We haven’t done that, but it ought to be done. Good idea. You should try it out. Yeah, I can’t see anything but positives to that.

OL: Okay great. Because what I was trying to do in that thought experiment was bring together what a lot of people seem to want, which is to help support people to become competent problem solvers, with your goal-free effect, which is supported by a lot of experimental evidence. So yeah, that’s obviously something that you think could potentially work.  

JS: Yeah, in effect you’d have one group who were given a problem to find whatever it is they have to find and the other group would be told to find whatever it is, but that second group is also told that ‘if you have difficulty finding it, forget about what you’re trying to find, just calculate the values of as many things that you can think of and see how that goes’. (pauses) Hmm, Yeah. You try it.

OL: An area for future research…

It’s important to read the above section in conjunction with the next post, so don’t just stop here otherwise you’ll take away an incomplete picture!

Next post:

3. What’s the difference between the goal-free effect and minimally guided instruction?

All posts in this series:

  1. Worked Examples – What’s the role of students recording their thinking?
  2. Can we teach problem solving?
  3. What’s the difference between the goal-free effect and minimally guided instruction?
  4. Biologically primary and biologically secondary knowledge
  5. Motivation, what’s CLT got to do with it?
  6. Productive Failure – Kapur (What does Sweller think about it?)
  7. How do we measure cognitive load?
  8. Can we teach collaboration?
  9. CLT – misconceptions and future directions

John Sweller Interview 1: Worked Examples – What’s the role of students recording their thinking

This is part of a series of blogs detailing a discussion that I had with John Sweller in mid 2017. See all parts of this series on this page

OL: Let’s start off by considering that original experiment of yours, back in 1982, in which you required students to move from a source to a target number using only the operations: multiply by three or subtract 29. I want to do a little thought experiment based upon the original experiment. So, if we took that experiment and we ran that in tandem with another, which is exactly the same, except in the second condition, you ask the problem solvers, the undergraduate students, a different question. Instead of saying: ‘Try to get to this number from this number using these operation in whatever order’, you say: ‘Try to do that. But also, record every step clearly, show what you’ve done and after you solve every problem, I want you to look back and try to find some generalisation or some principle.’ What do you think would be the outcome of this experiment?

JS: That’s exactly how one learns when problem solving. In other words, you don’t learn a lot by problem solving, you’ll learn a lot more by looking back at your solution because at that point, it’s essentially a worked example.  You’ve created a worked example, and it occurred to me at the time of those original experiments that instead of having people look back at their own solution, why not just give them a solution? It’s easier, it’s quicker. That’s how we started working on the worked example effect, for precisely that reason so that’s a useful insight. It would work better, yeah.

OL: Let’s consider the political environment of education at the time that you were trying to introduce this information on the worked example effect. It was heavily in favour of problem solving. So, against that backdrop. I was wondering… If you had tweaked it and had, say, presented the worked example effect in a way that I just presented it just then (with students doing the problem solving themselves, then reflecting upon their own solutions), do you think that would have influenced the way that people reacted to it?

JS: Look, I’d be pessimistic about that. I’d be pessimistic because the entire research environment in those days was absolutely committed to problem solving. That you’d learn through problem solving. You can look at the problem later but it doesn’t really matter because there are two things you learn through problem solving, so the story went —the first thing you’ll learn is, you’ll learn how to solve problems in general. But the second thing you’ll learn is you’ll learn the content better. So you have that two path advantage. You’ll learn how to solve problems and you’ll learn content better. Now neither of those paths is true. The data show that quite clearly. But notwithstanding that, that’s what people assumed and based on that assumption, they were not going to change their minds.

OL: Okay. But if people are hell-bent on getting students to solve problems, then getting them to accurately record what they’re thinking at the time and then to reflect on it, is better than just getting people to solve problems?

JS: Oh yeah! No question about that. It’s a better way of doing it and indeed getting people to reflect and think about the solution is beneficial in a whole variety of points of view. So that would be useful.

OL: Cool. It’s interesting because I haven’t seen it in much of your work. Perhaps I have missed it- much of an emphasis on recording one’s own thinking and using a kind of external means of recording as a way for individuals to reduce their cognitive load.

JS: That is correct because my aim has always been: Let’s get teachers to do it properly in the first place. There’s an assumption that when somebody looks at a worked example and studies a worked  example, what you just said is precisely what they’re doing. It’s just that they didn’t provide the worked example themselves. When we ask an individual to ‘study the worked example’, which is the instruction that people were always given in a worked example experiment. My assumption has always been: That’s what they’re doing, they’re thinking about the worked example. They’re studying it. They’re trying to work out why it works this way and they’re doing it in the sense that it’s provided to them so they’re not wasting their time going into useless dead ends.

OL: Yeah and it’s just more time efficient. That makes a lot of sense.

Next post:

2. Can we teach problem solving?

All posts in this series:

  1. Worked Examples – What’s the role of students recording their thinking?
  2. Can we teach problem solving?
  3. What’s the difference between the goal-free effect and minimally guided instruction?
  4. Biologically primary and biologically secondary knowledge
  5. Motivation, what’s CLT got to do with it?
  6. Productive Failure – Kapur (What does Sweller think about it?)
  7. How do we measure cognitive load?
  8. Can we teach collaboration?
  9. CLT – misconceptions and future directions

A Conversation with John Sweller

In mid 2017 I met with John Sweller to ask him a few questions that I had about Cognitive Load Theory, and his work more broadly. John was gracious enough to let me record our conversation and turn it into the following series of blog posts. All together, these posts amount to almost 9000 words of text, so I have broken our discussion into nine different sub-posts for the convenience of readers. That said, as these are a transcript of a conversation, each section flows on logically from, and often references, previous sections.

  1. Worked Examples – What’s the role of students recording their thinking?
  2. Can we teach problem solving?
  3. What’s the difference between the goal-free effect and minimally guided instruction?
  4. Biologically primary and biologically secondary knowledge
  5. Motivation, what’s CLT got to do with it?
  6. Productive Failure – Kapur (What does Sweller think about it?)
  7. How do we measure cognitive load?
  8. Can we teach collaboration?
  9. CLT – misconceptions and future directions

Note: If you’re not already familiar with the terms ‘cognitive load theory’, ‘the worked example effect’, or biologically primary and secondary knowledge’, it might be a good idea to start by reading Sweller’s fantastic overview entitled Story of a Research Program, available here.

A big thank you to Bryn Humberstone for his help and suggestions during the editing of this series. Also thank you to Rajiv Bhar and Michael Pershan for contributing questions for use in this interview with Sweller.