Carl Wieman (2016) - A Scientific Approach to Learning Physics

It's always a real pain to be at the end of a session with a bunch of Nobel-prize-winners speakers, because they always run over time and the session chairs never have the nerve to cut them off. But anyway, I’ll try and keep this on time. So I wanted to just go over, it’s a completely different topic here, of how you can – you, I'm really aiming this for anybody whose post-doc level and below – how you can learn to think like a good scientist, physicist, as quickly and effectively as possible. Now, the way I got into this actually was about 25 years ago: looking at the graduate students coming in my atomic physics lab and seeing how they can go through these many years of great success in courses but really not being able to do physics. And yet, after a few years they would learn quite quickly how to become physicists, although the very best students in courses never turned out all that great. I started to notice this as so consistent I wanted to just figure it out. So I tackled this really as a science question, looked at what we knew about learning. What we knew about, particularly, how to think scientifically, learning science. (Interruption, microphone gets adjusted) What we knew about learning science. And I found there was quite a lot. In fact, that one could really go beyond opinions, of which in teaching and learning there’s countless numbers of, and really approach this as a science. And by that I mean doing controlled experiments, measuring the learning that takes place, and having data and real fundamental principles that come out of those, just like in physics, that you can use to make sense of. And this kind of research and learning to think scientifically, it started in physics several decades ago, but now it's spread throughout undergraduate sciences and engineering. I’ve been doing research in this area myself for now almost 3 decades and have quite a few publications. So I'm going to start to give you an example of some of the kind of things one can find out doing these sorts of experiments. And so in this experiment they had 3 groups of students tested to be equivalent. They ran them and sort them into these 3. The first group goes to lecture, they take notes, and they try and learn as much as possible while they are doing it. The second group goes to lecture, doesn’t take any notes but tries to focus on learning as much as they can. And the third group doesn’t go to lecture, they just stay home or in a side room. They get the instructors' notes for the lecture, and they spend the same equivalent amount of time, trying to learn as much as they can from them. And then all these groups are given the same test on the material covered in this lecture. And so now you are going to have to make a prediction, on how these 3 treatments rank in terms of the amount of learning. And so the choices, I'm going to have you vote on by raising your hand here, are a) means that the "going to lecture and taking notes" learnt the most and "staying home" learnt the least. b) Those are flipped, etc. So I'm going to give you a few seconds to think about these and then I'm going to call on you to make your vote. Okay, everybody ready to vote? So raise your hand if you think a). A bunch of people think a). b) A bunch of people think b). c) Lots of people think c). d) Lots of people think d). Finally e) A lot of people think e). (Incomprehensible heckling) Okay, so it's pretty - (Incomprehensible heckling) you're not supposed to interrupt me. (Laughter). The correct answer is the one that the fewest people actually chose. Interestingly enough, a lot more of you choose this than I usually get with a group of students. It's almost never chosen, but it's still obvious the minority here. That, actually, people who learn the least are the ones going to lecture and taking notes. As I say, students normally, when I pose this to them, don’t choose that. But when I give them the chance to talk to each other and see this answer, they very quickly come to the deeper reasons, principles of learning, that tell you why those actually work. That says that these aren’t that hard and non-intuitive. It's basically that they have been brainwashed by the teachers to believe that going to lecture and taking notes is supposed to be more effective. Now, you might wonder why that actually can happen. The reality, I think, is that going to lecture and taking notes - the original design happened for a very good reason and it served a very clear purpose. That purpose became obsolete with modern technology; in this case the modern technology was the invention of the printing press. But professors are slow to adapt to using the technology, they are kind of lazy. So it's just easier to keep doing the same thing and tell students that must be good for them. Moving on to exactly why - so you can understand what we do know about learning and why the results are the way they can. The reason, 'not taking notes' is better than 'taking notes' is because the brain has very limited capacity to pay attention and process different things. And so the more it's called upon to do at the same time, the less effective it can be. So aiming to take notes just adds these extra demands or distraction to the learner over just listening as carefully as they can. And so that’s why 2) is better than 1). The reason that 3) is better than 2) is just that if they are sitting at home, looking at these lecture notes, there are several advantages to that. First, they can go through them at their own pace. So they can stop and ponder things and think about it and not lose complete track because it's being controlled by the speakers pace. When things are laid out like that, the organisation and the structure are much clearer when it's on paper than when you are trying to just follow a verbal lecture. And all of these things allow the learner to have more processing of the material and that’s always what turns into more learning. I'm going to give you a second experiment here. Now, this is much closer to and more quantitative, looking directly at physics and learning of physics. Here one has - in an experiment I was involved in - 2 large sections of introductory physics, carefully measured, that student populations are very nearly identical. And both of these sections are trying to learn exactly the same set of learning objectives, covering an exact same amount of class time in one week of lectures. And then they are given a pop quiz, right at the end of the last lecture, to test how much they learnt from those lectures. The 2 sections, the comparison is, one of them is taught by somebody who’s is a very experienced teacher, with material of a fairly traditional lecture, but has very high student ratings at giving these lectures, this material, to students. And then the second section is taught by somebody who is a post-doc who had been trained in these methods of scientific teaching and learning. So the question is, How do you think the results come out for how these two sections compare? Well, probably you are going to have some guess as to what I just showed you before. But here’s the histogram of the number of students here versus their score in the test, with the experienced traditional lecturer in red and the scientific teaching in white. So you can see that distributions are dramatically shifted. But, in fact, the differences are bigger than what you might think at first impression here, because this was actually a carefully developed multiple choice test to keep it objective. And so, on average, just by guessing, you would get 3 on this. So you really have to look at how much better than 3 the students are as a measure of their learning. Then you realise, how dramatically different really the amount of learning was between these 2 sections. I want to emphasise - and you can see the whole distribution shifted up. So, basically, the top students and the bottom students, all learnt much more in this teaching here. And the amount of learning from the traditional lecturer is then a very tiny fraction, here about 16%. Just to make a snarky comment: if you compare that actually with something like the format of these Nobel Prize lectures, there are several reasons I would confidently predict these would be a lot worse than that. At this point in my talk I either have somebody leap up and interrupt me or, at least, a lot of people are thinking this. I'm always asked to comment on something to the effect of: "But these lectures can’t be as bad as what you are claiming. Look at all these Nobel Prize winners who were taught by going to traditional lectures and how well they turned out." But this is where you need to think about approaching this as a science. Because, in fact, it's exactly this same argument, same reasoning, that was responsible for the fact that bloodletting remained the medical treatment of choice for about 2,000 years. Somebody gets sick, you let out a bunch of their blood, most of the time they recovered – so, gee, obviously bloodletting was effective. But, what we learnt in medicine a 150 years ago is really what we've now been doing in teaching and learning. Which is if you want to do this scientifically, you have to do a proper comparison group. So what I would argue, if we could do a proper comparison group, we'd probably find, if these Nobel Prize winners had actually had a better education, they would have been more successful, more prizes, younger ages etc. Now, there’s a quick caveat, getting into the technical details: people actually can learn from lectures under very special circumstances, which aren’t usually adopted. I can talk more about that later today. So, there are actually thousands of studies on learning and about 1,000 that look quite specifically at the university level and sciences and engineering, comparing the standard lecture method with these scientific teaching principles. These show always consistent differences in the amount students learn. And the effects are particularly big if you measure, how well the students are starting to think like experts in the field, physicists, chemists etc. And so I'm just going to rush through a few examples – not to show you anything to convince you, except that I’ve lots of data I could show. Here’s an example looking at introductory physics, testing how well the students are able to look at novel situations and apply, correctly apply, like a physicist would, the concepts of force and motion. And they measured this with all these different instructors were lecturing, they were down around 0.3. Then they all switched to a version of scientific teaching. And, basically, across the board there was a doubling of the amount students learn. How much they could think properly about these concepts. Just to show it’s not all introductory physics, this is a fourth year modern optics course for physics majors. Here they switched. And in particularly challenging final exam problems there was about a 1 standard deviation improvement. Different ways of having students learn this. This is from computer science just looking at the drop & failure rates. Again a bunch of different instructors converted how they teach. And across the board big decreases in the drop & failure rates down to about a third. That's just to give you a sample that we have lots of data. But you don’t care so much about that, what I want to focus on is something be useful. So I hopefully convinced you that I know something what I'm talking about. And now I'm going to give you some principles and methods that we see coming out of these kinds of studies, on how you can better learn to think like expert physicists. And so the rest of my talk is going to start with really the nature of expert thinking and how it's learnt. And then how this applies to physics in particular and how you can use it in your own learning. Cognitive psychologists have studied a lot about how experts think and develop their expertise, across all these different fields, musicians, chess players etc. And they find that there are certain basic components of expertise and a certain basic process for how that expertise is learnt. And so the first component of expertise is one everybody could guess: Experts know a whole lot about their subject. The second and third aren’t nearly as obvious. The second is that experts have, unique to their subject area, a particular mental framework by which they organise all that information. And it's that particular organisational framework that allows them to be very efficient and effective at finding and applying the appropriate knowledge to solve a problem. That means looking for and recognising certain patterns and complicated relationships. Most of what we talk about as scientific concepts, is really just the way scientists in a particular field have taken a whole bunch of different pieces of information and seen how they can organise that into sort of one piece. And then quickly decide if that piece is going to be helpful or not in solving a problem. The third general feature of expertise is ability to monitor one's thinking. And so by that I mean, as an expert, working through a problem in their subject, they are able to actually keep asking themselves: Do I understand this? Is this a sensible way of me solving this problem? And actually test that and then change what they are doing accordingly as they are working. Now what the research shows is, these are fundamentally new ways of thinking. And to develop them, everyone requires many hours of intense practice to develop these capabilities. And very recently it's becoming clear that – well, first I should say 'many hours'. But to reach a high level of expertise, sort of university professor level, it's many thousands of hours of intense practice. And what we are realising now is, that this is basically a biological determination, or limit. That, as a result of this 'many hours in intense practice', the brain is actually changing in substantial ways the wiring and so on. So it's actually within this rewired brain that the expertise lies. It's really a very close analogy to what’s involved in building up a muscle. If you want to build up a muscle you've got to use it very strenuously over a long period of time, and the body responds by making it bigger and stronger. In the case of the brain it says, "Oh, it's going to keep making me do that hard physics problem. I'm going to make the connections and the neurons stronger to make that easier." Now, the research says that, okay it's many hours in intense practice, but it's also a very specific kind of practice that’s required. It has to be hard problems exercising the brain, but they also have to be giving it practice in exactly the kind of expert thinking that you want the brain to learn. That’s what the neurons have to be reinforcing. But, of course, it's not enough to just practice hard, you have to know that what you are doing is successful. And so you also need to have feedback on that practice, to guide you what you are doing and see how to improve it. And then when you’ve got this, you just keep doing this enough hours and you become an expert. So that’s a very general thing. Now I'm going to get a little more specific about what are some of these basic thinking components of expertise? And I’ll just start here with some that are quite generic across the sciences and engineering. For example, every area has a set of concepts and mental models that it uses. And more importantly, expertise lies in a very sophisticated selection criteria for identifying which of those models or concepts apply, and don’t apply, in any specific situation. There’s recognising what information is relevant to solve a problem and what is irrelevant. When experts have come up with an answer in every field, they have very specific-to-that-field criteria by which they know how to check, if that answer makes sense. Or if there might be a better way to solve this problem. And then, finally, an extremely common one is, experts in every science and engineering area, there are very specialised ways to represent information. And experts can move very fluently between these different representations and get new insights as they do for problem solving. So there’s a bunch of others, but these give you the basic ideas of some components that need to be practiced. Now, when people are talking about learning to be a physicist or a chemist and so on, those discussions are almost always in terms of 'oh, we need to have them cover all these sets of topics', but it's really essential to realise that, okay, that topic, that knowledge, that’s important. But it's really only important if it's really embedded in these broader aspects of expertise, of expert thinking, that really govern the understanding about when and how to use that knowledge. The research shows lots of examples, where students can pass tests, saying they know something. But then, when given a real problem, they can never use that effectively. So getting down to more specific things you can do: If I summarise the research on learning, there are 5 basic components that are really important in this. But the other things then I think you can get some help on. The first is connecting with prior thinking. As I said, you are rewiring the brain, but you got to - the brain doesn’t start blank, it's got a lot of wring in there. So you need to connect and build on what you already know and think. There are some basic things about how the memory works in terms of initially processing information and then how to retain it long-term. And then there’s this idea of practicing authentic expert thinking and getting feedback on it. So the things I'm going to talk about are a bunch of examples that give you things you can be doing on your own to practice these to learn better. By the way, if you are going to be teaching, these are the things you need to have your students doing as well, for them to learn. Now, I want to emphasise what it’s not - and this is a good test. If you think about the way most people study: they read over the text book and at the lecture notes and go over the problem solutions, or passively sit and listen to a lecture. And one of the immediate ways you can test that it isn’t very effective is, it's too easy. It's not pressing the brain to work hard and that basically means the brain doesn’t squirt the chemicals in there to make the effects. Okay, what are some hard things you can do that will be effective? The first is, you need to really study intensively and focused, where you really are putting full attention into it. Or just don’t brother, it's just not, you don’t really get anything out of paying half attention to studying. When you get some new material, a topic here, you need to sit and think and reconcile that with your past knowledge. Think how the topics, how this is connected with things you’ve already learned, situations in the world you know about and so on. And do that in a very active, deliberate way. When you learn about some new concept, sit down and write down the criteria: when this is going to apply, in what kind of situations, and, very importantly, when and why it won’t be useful. You always want to be looking at ways to test your thinking and catch where you might have weaknesses. And so there’s a whole bunch of different ways to do that: seeing how your ideas would apply in a variety of different situations; checking, arguing with other students, talking to them, professors etc. One particularly effective way is to try teaching this to somebody else – we in our instruction do this with our generic 'grade 10 sibling'. It requires taking university material, thinking how to process that so you could explain it to your 10th grade brother or sister. Actually, it makes you process it in a very special and effective way for learning. And it turns out in these studies, you don’t actually have to teach it to anybody, you just have to imagine you are teaching it to somebody and you get most of the benefits. There’s solution planning. This is something experts spend a lot more time on than beginners. And so be very explicit when you have a new problem, trying to work out what the solution is, as to how it's broken down into pieces and so on, before you ever start diving into it. When you solve a problem, think of alternative ways you might be able to solve it - experts do this all the time. And, very importantly, when there are simplifying assumptions, try and ask yourself, what would happen if those assumptions weren’t right? Or are there other possibilities of assumptions you could make? And, getting close to the end here, if you have something wrong, don’t just look at what the right answer is. This is a big mistake most teachers make, that they see a student get something wrong and say, It turns out that doesn’t accomplish much learning. The learning really takes place when the learner not just knows they are wrong but understands why they are wrong and how they need to change their thinking in the future to be right. And that’s where almost all the effective learning happens. Getting back to long-term retention. It turns out if you want to remember things for a long time, you have to test yourself on them repeatedly and spaced in time. And so by testing, it's not a formal test, you just retrieve and apply those ideas. And it's really true that you either use that thinking process or you lose it; the brain takes those neurons to do other things with them. Finally one, I really wish I’d known better when I was a student: it's really important to sleep. There are 2 reasons for that. One is that you get stupider if you don’t get as much sleep. But, less obvious than that, when you are learning something, it's during the sleep process that the brain actually consolidates a lot of that learning and makes a bunch of the necessary wiring changes more permanent. So it's really an essential part of learning to sleep. Okay, so I'm going to just finish up quickly here another way how you can quite specifically improve your learning-from-homework problems. First, you have to think about: what is the expertise practiced and given feedback on from the typical text book problem or exam problem, if you look at the list of components of expert thinking I talked about earlier, and then you think about what’s in the normal problem. First, it gives all the information needed and only that information tells you what assumptions, things to neglect. So right away you’ve lost practicing those parts of expertise. Didn’t ask for any answer, just get an answer, you don’t have to justify why it's reasonable - that goes away. Usually it only calls for one representation, whatever they put into that problem - that goes away. And, more often than not, it's possible to solve the problem just by looking and using the concepts or procedures that were just covered in the text book or class - and so that means those go away. So with your normal homework problems, you are actually left with very little practice of expert thinking. You are really predominantly just learning facts and procedures, but not the more useful expert thinking. I'm not going to go into details for this, you can get copies of my slides and think about it. But these are just ways you can take your standard homework problem and rewrite it yourself into a problem that really then gives you the practice of expert thinking, that helps you actually learn much more effectively from working through those problems. Okay, so I'm just about on time here. I realise that this talk - I realise better than almost anyone because I study this - there’s too much stuff here, it's gone too fast for you to really process it. But I’ll make sure you can get copies of my slides and so you can ponder those and that will allow you to learn something. And coming to the afternoon session today where we will discuss these in more details, that will be useful. And just when you do look at the slides I have a few references. If you want to learn more about this effective learning and learning to think expertly there are references there. Thank you.

Carl Wieman (2016)

A Scientific Approach to Learning Physics

Carl Wieman (2016)

A Scientific Approach to Learning Physics

Abstract

Bloodletting was the treatment of choice for 2000 years, and doctors and patients believed it was effective because most people recovered after treatment. The learning and teaching of physics is similar to pre-1800 medicine. The pedagogical equivalent of bloodletting is the standard method and has comparable “evidence” of its effectiveness. I will discuss new controlled experiments on the teaching and learning of physics, equivalent to the modern tests of the effectiveness of different medical treatments. These studies show much more effective alternatives to traditional approaches. They are also providing empirically based general principles for learning, particularly learning to think like an expert physicist. This research is starting to shift the teaching and learning of science away from methods based on folklore and opinion to those based on science.

Content User Level

Beginner  Intermediate  Advanced 

Cite


Specify width: px

Share

COPYRIGHT

Content User Level

Beginner  Intermediate  Advanced 

Cite


Specify width: px

Share

COPYRIGHT


Related Content

Associated Data