Dr. Mazur’s particular contribution to higher education is peer instruction, a pedagogical model in which lectures are replaced by in-class discussion of case problems. In addition to his work in education research, Mazur is also an award-winning optical physicist. Mazur has founded or co-founded three companies, including one, Learning Catalytics, that offers educators technology that allows them to track and analyze their student’s comprehension of material.
Like many successful teachers, Professor Mazur originally had no intention of becoming a teacher. The son of two teachers – his father was also a professor of physics, and his mother was an art historian– he says that he “knew one thing for sure – I was never going to teach!”
After receiving a master’s degree (1977) and a Ph.D. (1981) in physics at Leiden University in the Netherlands, Mazur planned to take a position at Dutch technology giant Philips. However, his father convinced him to put his job at Philips on hold and go to Harvard as a post-doctoral student instead. Mazur told his future employers that during his time in the US he could learn more about lasers; Philips agreed to wait. Mazur ended up postponing the position yet another year, at the end of which he was offered an assistant professorship: he accepted and forgot about his job offer from Philips. Even though he had sworn he wouldn’t, he had become a teacher.
Mazur never asked himself how he was going to teach. As he puts it, instead of looking at other methods of teaching, he simply decided right away that: “I was going to do to my students what my professors had done to me.” In other words, he was going to stand in front of the class and lecture.
At first this strategy gave every sign of success. On end-of-semester questionnaires students gave him high marks. Even when he loaded his exams with extremely difficult problems – problems so hard that he was unsure that he could have solved them in his student years – his students passed them with flying colors. Mazur had gotten these results even though he was teaching not future physicists, but pre-med students – a group that, as a rule, was hostile both to its physics teachers and to physics as a whole, seeing it as yet another hurdle to jump on the road to med school. With time, however, Mazur began to sense that something was wrong. Students gave him high scores on course evaluations, but they were also writing things like “physics is boring” or “physics sucks.” Getting high marks from a set of students who had traditionally been hard to win over had lulled Mazur into a false sense of security, but these comments – and another important realization – snapped him out of it.
In 1990, after about seven years of teaching, Mazur discovered the Force Concept Inventory through an article discussing its use to test university students’ knowledge of physics. The Force Concept Inventory is a set of 30 multiple-choice questions designed to test a student’s understanding of Newton’s Third Law. One sample question from the Inventory goes something like this:
A heavy truck and a light car collide on the highway. Is the force exerted by the light car on the heavy truck – a) greater than that exerted by the heavy truck on the light car; b) equal to that exerted by the heavy truck on the light car; c) less than that exerted by the heavy truck on the light car; d) non-existent, i.e. the light car exerts no force on the heavy truck. The correct answer is b: they are equal. However, most students, both before and after taking a course in physics, intuitively answer “c”.
When researchers posed questions like these to students at several universities in the American Southwest, they discovered that students were unable to apply Newton’s Third Law to word-based problems – not only before taking a physics course, but also after taking one in which Newton’s Third Law had been covered several times. An impressive 80% of students surveyed could state Newton’s Third Law at the beginning of the course, but by the end of the course only 15% could successfully apply it to real-world problems. Even when those tested were broken down into sub-groups – students taught in large classes, students taught in small classes; students taught by award-winning teachers, students taught by teachers who scored poorly on student reviews – they all received nearly the same awful results. Both intrigued and challenged by these results, Mazur set out to prove that this was not true of his Harvard undergrads.
Mazur was soon proven wrong. During the very first test, a student raised her hand and asked: “Professor Mazur, how do I solve this problem? The way you taught us, or the way I usually think?” It turned out that his students had not really learned to apply physics to solve real problems. Mazur’s results were slightly higher than the results in the Southwest – this was Harvard, after all – but they were still abysmal. Some of the scores taken from Harvard students were truly shocking. In one interview, Mazur put it this way: “Some did barely better than a gorilla hitting random keys on a keyboard.”
Where was the problem? This was Harvard, so it couldn’t be the students. Mazur had been given high marks as a teacher. Since neither the teacher nor the students appeared to be at fault, Mazur decided that it had to be the tests. He soon discovered a paradox: his students excelled at tests in which formulas were expressed with groups of numbers, but they couldn’t answer questions in which those formulas became objects and situations. Once the figures became a heavy truck and a light car, the students were baffled. Both tests covered the same laws of physics, but posed the questions differently. The difference in the results showed that Mazur’s students had become very good at applying pre-made recipes to solve problems created to fit those recipes. The result was a closed circuit that had little to do with using physics in the real world.
Worried by their results, Mazur’s students asked him to hold a special class to go through the Force Concept Inventory’s problems one by one. Mazur agreed. He returned, once again, to the heavy truck and the light car; it took him two minutes to explain the problem using Newton’s Third Law. He was met with blank stares. Mazur returned to the chalkboard, and spent eight minutes explaining the problem using Newton’s Second Law. He was met with even blanker stares. In desperation, Mazur asked them to discuss the problem amongst themselves. In two minutes they had the answer. They had been able to explain to each other in two minutes what Mazur, a world-class physicist and Harvard professor, had been unable to explain to them in ten.
Mazur notes that another Harvard Professor, Steven Pinker, has a name for this phenomenon: “the curse of knowledge.” Mazur has this to say about it: “Imagine you have two students sitting next to each other: John and Mary. Mary has the right answer because she understands it; John does not. On average Mary is more likely to convince John than the other way around, simply by the force of logic. But here’s the important point: Mary is more likely to explain it to John than Professor Mazur in front of the class. Why? Because she has only recently learned it… She still knows what the difficulties are that the beginning learner has, whereas Professor Mazur in the front of the class learned it such a long time ago that it is simple and obvious.” If you think about it, what Mazur says is true: we’ve all been faced with the same situation in school, and we’ve all turned to the student next to us to ask for help instead of raising our hand to ask the teacher.
His students’ low scores on the Force Concept Inventory led Mazur to another sobering conclusion – his lectures didn’t work. After some thinking, Mazur realized that this was because his lectures were about transferring information, and not about assimilating concepts. Information was everywhere – in Mazur’s notes, on the blackboard, in the student’s notebooks – but there was nothing going on in the student’s heads. Merely receiving information is not enough, because the student must be able to internalize that information, to build mental models of it that can be used to solve real problems.
In the traditional classroom, the student listens to a lecture while surrounded by other students; after receiving that information in class, the student goes home and tries to assimilate it. Mazur realized that this model makes no sense. Lectures only made sense when dictating information to students was the only way of conveying it to them. Today each student can read the material at home, alone, thanks to the invention of the printing press, something that appeared over 500 years ago. There is no longer a need for the lecturer to transfer the information in class. In any case, the hardest part is not receiving the information – it’s assimilating it, truly coming to grips with it, and students are forced to do that on their own, outside of the classroom, without the help of their professors or classmates.
Mazur’s solution was simple: switch the two steps. Have students read about the concepts alone, at home, and have them work at understanding the concepts together, in class. Although Mazur did not coin the term himself, this model of individual reading at home and in-depth problem-solving in the classroom has since come to be known as the “flipped classroom.” In the flipped classroom, the key is not how to transfer the material to the student – the students have already read the material on their own – but how to help the student assimilate the material. To do this, Mazur stopped lecturing entirely and chose a model that has existed for at least 2400 years – the Socratic method, in which the teacher instructs not by giving answers, but by asking questions. Over time Mazur came up with a step-by-step process called “peer instruction” to formalize this method. The process has six steps:
1) Question. The instructor gives the students a question.
2) Think. The instructor gives the students time to think about the question. This is in stark contrast to traditional lecturing, in which the student barely has time to write the notes down, much less think them over.
3) Poll. The students are polled for their answers to the question (the question posed is usually multiple-choice.)
4) Discuss. If more than 30% of students have the correct answer, the instructor tells the students to discuss their answers with their neighbors. (If less than 30% of the students answered correctly, discussion is largely unproductive.) The idea is for students with different answers to discuss why they chose the answers they did.
6) Explain. The instructor or one of the students who answered correctly explains the problem and its solution to the class.
In 1997 Mazur published a book titled Peer Instruction: A User’s Manual, which outlined this methodology, as well as providing some material that teachers could use in class to implement the peer instruction strategy. Since then Mazur’s book has been translated into four languages, and nearly 1,500 articles on peer instruction have appeared in academic journals.
However, creating a new educational model is one thing, and trying it out in the real world is another. Professor Eric Mazur spoke with one of eRazvitie.org’s correspondents on what difficulties he met along the way.
– How did your colleagues and students respond to peer instruction? Was there any opposition?
In the beginning very much so. Don’t forget that I came up with the idea in 1990, so I’ve been at this for almost twenty-five years. I think in the beginning my colleagues at Harvard thought that I had gone out of my mind. I think in general change is difficult, because people – students, faculty, administrators – will typically try to stick with the status quo. Students thought, “Professor Mazur isn’t doing his job – we have to do all the learning ourselves.” Meanwhile my colleagues were saying that I had to lecture, because that’s what I’m paid to do. At first I think most of the opposition came from people who were thinking “why”? Why change? Things are fine. I think that before you change you have to believe there’s a reason for change. I think that quite a large fraction of instructors are not convinced that we should change.
Most professors and students live under the illusion that everything is fine. Students ask no questions. They ask no questions in class, because they think they understand everything – even though they haven’t had time to think about the material yet. When professors don’t get too many questions, they assume that they must have given a good lecture and everything must be clear. Our system of assessments, which is geared towards testing the regurgitation of information delivered to the students, only further reinforces the idea that everything is fine.
– If lectures are so ineffective, then why have they held on for so long?
Once again, it’s very easy to be fooled to think that everything is fine. Second, in a sense lectures are the lazy man’s approach to education. All you need to do is take information that’s already available in print and deliver it to students, who basically record it and don’t really get to think about it. No one goes “hmm – does this make sense?”, because there’s really no time to do that in a lecture. And both the professor and student are misled into thinking that the professor has actually taught something.
If we had an assessment process that’s more rigorous than the one that we have, we would find that most students don’t derive that much benefit from a lecture. What happens is that the person delivering the information is typically the person writing the exam, and the exam is written in such a way that students can pass it – otherwise the teacher is not doing his or her job. So the bar is lowered sufficiently low to guide the student through the exam – mainly by basically asking for information that has already been memorized, crammed. Which is something most students can do. If we were to ask our students to do something creative, something that requires higher-order thinking skills, I don’t think that most of our students would get there.
We’ve built this system where the people teaching are the people doing the assessment, too – which, frankly, just between you and me and your readers, is quite amazing. There’s no other human enterprise where things are that way. The people doing the assessment are usually not the people doing the work. Our getting away with it almost makes education seem like a cartel.
– When you and other instructors use peer instruction, do students miss the old tests with the cookie-cutter questions?
Some do. But I think many love it because nothing is more mind-numbing than sitting in four hours of lectures a day, just passively listening. Some lectures are interesting, but most are not. Generally students respond rather favorably to the different approach to instruction. Of course, the different approach to assessment scares the hell out of them. They know they can cram and pass the test by just regurgitating the information and applying rote procedures. If the answer to the question is four meters per second, and the student writes four meters per second, the student gets full credit. That can be evaluated completely objectively. If they need to create something, or they need to innovate, or they need to think critically, then the assessment is not as unambiguous as before. It’s scary. However, when students are asked to do something creative, they realize that there’s no way of knowing how other people are going to evaluate it.
However, students should really realize that that’s what real life is about. That’s how real life is. It’s not about applying things by rote or memorizing them, because that’s what computers can do. We need the workforce of the 21st century to be creative and innovative, to have the right thinking skills, no matter what the job. Any other kind of job will disappear.
– Are university administrations generally receptive to this new approach to education?
I think in general the administration of most universities at which I give presentations on peer instruction – including my own, Harvard—are receptive. However, courses are owned by faculty and ultimately the need to change needs to be determined by the faculty. You can’t push change from the top down; it also needs to come from the bottom up. I think in general there’s a lot of support – even from faculty. But, still, change is just scary for everybody involved, because you know what to expect under the old system, and you know you do well under it. Who knows what is going to happen after the change. Many junior faculty, especially when they are at research universities like mine, are worried to stick out their neck too much in terms of teaching. They think, “Let’s stick to the tried and true, let’s stick to the stuff that is supposedly – but that really isn’t – proven by time.”
– When asked about what peer instruction could be used to teach, you once said, “Learning is learning is learning.” Nevertheless, are there some subjects in which peer instruction is more effective, and others in which it is less so?
If you had asked me a few years ago, I would have said, well, I think it’s probably mostly useful in the sciences, where questions have right and wrong answers, and where you need a lot of critical thinking skills. But to my surprise I found that people in the humanities are also using it. If you go look at Monash University’s website, you’ll find an entire page dedicated to peer instruction in the humanities.
Even in corporate training that people are responding enthusiastically to peer instruction. I’ve seen airlines train flight attendants using it. When there’s an emergency in the air, you want people to be able to think on their feet. Most of the situations that happen in the air are probably not described in the playbook. If it’s not covered in the manual, then memorization won’t help them.
So that means that any discipline involving mental models, interpretation, or independent thinking skills, would be suitable for peer instruction. That’s why I said learning is learning is learning. Peer instruction is best not used in anything that involves rote memorization. But I would say that there’s probably very little we would want to teach that involves that.
– Does peer instruction change when it’s used outside the hard sciences?
Probably a little. One thing being that there is no right answer. Let’s say that we’re talking about the interpretation of a passage in a book or the interpretation of a painting from the Renaissance. There’s no right answer, so we can only speculate. There might be several valid answers within the historical context in which that sculpture or that piece of fiction was created. You can no longer say to students that this is the right answer and this is the wrong answer. That frames the discussion differently.
I’ve used peer instruction for a teaching module on ethics. Typically I have students vote on different outcomes of different actions that they might take in a given case scenario, and I show the distribution, which I don’t do in the sciences. And then I say, look, this many people voted for A, this many people voted for B. And then I tell them, look, forty percent of you said B, is there somebody who wants to articulate why they choose B? So before they launch into a discussion with each other, I have a few students articulate their point of view – which is much easier to do when there’s no right or wrong answer. Students are less afraid to take a risk. And then, after all the students in the class have heard these different points of view, I ask them to discuss these things with their neighbors and see if they change their minds. Total chaos. And then basically we vote again, and show the new distribution, which often changes: when people voted the first time, they were not aware of another way of thinking about it, so in a sense it opens the mind.
So the only modification there is that I have a few students articulate their reasoning before launching into the discussion. Other than that, it’s pretty much the same as in the sciences.
– The traditional model of higher education also includes seminar classes, which are supposed to involve discussions of real-world issues. In what way does peer instruction differ from the traditional seminar method?
Structure and scale are the two big differences. I think it’s much more scripted than the traditional seminar, which is much more free-flowing. And I think that a seminar by definition is small. Peer instruction is scaled up to three hundred people, nine hundred, a thousand – I’ve done it with an audience of three thousand.
– Do you think a teacher-free version of peer instruction is possible?
There is! Go to avantifellows.org and you’ll see a group of people in India who are doing just that. There’s a tremendous shortage of math teachers in India. There they have done something that I would have never thought of or realized was possible – they’re doing peer instruction without a content expert. (You said “teacher”, but I’ll say content expert, because you do need a facilitator.) There are just not enough math teachers, physics teachers, chemistry teachers in India. So what they do is take social workers who have no content knowledge, and have them facilitate instruction among students. I think it’s brilliant, because basically what happens is students teach each other, and improve their understanding of material. Any unresolved matters get transferred by the social worker to the “mothership” in New Delhi. They have results that are really stunning – much better results than when you take a traditional class and put a poor teacher in front of it. This year one of their students got admitted to MIT with full support!
– Do you think universities should have a single model of instruction? Or should there be various models in education?
I think no, there shouldn’t be a single model. There’s never a cookie-cutter approach to anything. Peer instruction is one of many different interactive approaches that work.
You mentioned the seminar method. For a small class the seminar method might work much better than peer instruction. Why structure it? A seminar is a great way of getting everybody involved, but you can only haven one when the group is relatively small. Likewise, different circumstances will require different activities. I think the one reason why peer instruction is getting so much attention now is that it lets you to turn a large lecture-based class, with minimal effort, into something that feels like a seminar. It allows you to take a class taught using an ancient, ineffective model, and turn it into something effective.
– In conclusion: how do you see models of education changing in the future?
If you were to survey all of the classes taught on planet Earth, probably in excess of 99 percent would be lecture-based classes. So in a sense we currently have a single educational model that is imposed on all of the classes that are taught. I think in the end we will see a mix of project-based, problem-based, team-based approaches, peer instruction, and the seminar method, depending on the context and the size of the class. I think that in the twenty-first century it’s time that we abandon the lecture as the predominant teaching method. Given that many of our jobs will be taken by robots and computers, we need to start training humans to do the types of jobs that computers and robots cannot do.