Imagine a scenario where students retain knowledge effectively and are active and engaged participants who are self-aware of what they know (and don’t know). Did you picture a lecture class, students taking a test, or students writing? In this episode, John discusses three ways in which he has been using peer-instruction in his classes: classroom polling, calibrated peer review writing assignments, and two-stage exams.
- Bruff, D. (2009). Teaching with classroom response systems: Creating active learning environments. John Wiley & Sons.
- Bruno, B. C., Engels, J., Ito, G., Gillis-Davis, J., Dulai, H., Carter, G., … & Böttjer-Wilson, D. Two-Stage Exams. April 11, 2017.
- Calibrated Peer Review
- Carl Weiman Science Education Initiative – Two-stage exams
- Collaborative Exams with Teddy Svoronos – Teach Better Podcast #64, October 9, 2017.
- The Active Learning Initiative at Cornell – Tea for Teaching Podcast #12, January 17, 2018
- Knierim, H., and R.K. Davis. 2015. Two-stage exams improve student learning in an introductory geology course: Logistics, attendance, and grades. Journal of Geoscience Education 63:157–164, https://doi.org/10.5408/14-051.1.
- Mazur, E. (1997). Peer instruction: A user’s manual. Pearson
- Mazur, E. (2014). “Confessions of a Converted Lecturer” – a recording of a presentation by Eric Mazur at SUNY Oswego, 5/19/14
- Newton, E. (1990). Overconfidence in the communication of intent: Heard and unheard melodies. Unpublished doctoral dissertation, Stanford University.
- Tien, L. T., Roth, V., & Kampmeier, J. A. (2002). Implementation of a peer‐led team learning instructional approach in an undergraduate organic chemistry course. Journal of research in science teaching, 39(7), 606-632.
Rebecca: Imagine a scenario where students retain knowledge effectively and are active and engaged participants who are self-aware of what they know. Did you picture a lecture class, students taking a test, or students writing? If not, stay tuned, this episode explores ways to use peer-instruction to transform the learning experience.
John: Thanks for joining us for Tea for Teaching, an informal discussion of innovative and effective practices in teaching and learning.
Rebecca: This podcast series is hosted by John Kane, an economist…
John: …and Rebecca Mushtare, a graphic designer.
Rebecca: Together we run the Center for Excellence in Learning and Teaching at the State University of New York at Oswego.
Rebecca: Today’s guest is my co-host John Kane. John is the director of the Center for Excellence in Teaching—that’s not even right…
John: …Center for Excellence in Learning and Teaching at SUNY Oswego.
Rebecca: Yeah, woops! Welcome to your own show, John!
John: Thanks, Rebecca.
Rebecca: Today our teas are…
John: Prince of Wales.
Rebecca: Oh, that’s a good one.
Rebecca: I have Golden Tipped English Breakfast today.
Rebecca: One of the areas you’ve been teaching experimenting in, and that I’m fascinated in, is peer instruction. Can you tell us a little bit about what peer instruction is and why you’re drawn to using this methodology in your courses?
John: Peer instruction involves using peers to assist with instruction, where students explain….
Rebecca: Thanks John.
John: …where students explain things to each other. One of the issues that we have is that, once we become experts in the field, it’s very hard for us to express things in terms that are easily understood by students. There’s a “curse of knowledge;” once you become adept at something, it’s really hard to explain things at a level that’s appropriate to the level of understanding that students may have. There was a classic study done in which a researcher gave people a list of songs, very well-known popular songs, and asked them to tap out the beats from that song.
Rebecca: Oh, I would fail…
John: …and then before actually seeing if people would recognize it (who had the same list), she asked them to make a prediction of what proportion of people would understand it based on their tapping… and they overestimated that by a factor of I believe, somewhere around 20 times. Basically, it was purely random if people happen to guess it. But the issue is, once you hear something in your own mind, it’s clear to you, but it may not always be clear to the people who don’t have the same rich net of connections. When students are explaining things to each other, they benefit from taking a position, arguing that position, trying to filling gaps and they’re also explaining in terms that are appropriate for people at their level of cognitive development for people who have a similar background in terms of what they know and their prior knowledge.
Rebecca: Sounds like a really good way to expand and refine mental models and also just develop better metacognition. Because, as soon as you go to explain it, you realize what you don’t understand.
John: …and if you don’t understand it yourself, your peers will often help you understand. they’ll say: “Well, you haven’t considered this…” and that sort of interaction is one that doesn’t work as well when it’s instructor to a large group of students. But, it does work very well one-on-one.
Rebecca: You’re known on our campus for teaching really large lecture sections. Implementing peer instruction in a large setting can seem pretty daunting, especially to someone who teaches smaller classes like I do. What strategies do you use?
John: The most commonly used one is to use clicker quizzes… and I use a methodology that Eric Mazur developed slightly over 20 years ago, where you ask the students a challenging question… you try to find questions that about half of them will get wrong… and over time you can develop that, you can come up with a pool of questions that fit somewhere in that range… and you let students first vote on the response themselves after they’ve had a little bit of time to process it, and then you look at the results. If you see that 90% or more of them got it correct or even 80% or more, you can just go over it and move on to the next topic, because most students understand it. But, if you see that somewhere around half of them get it right and somewhere around half of them get it wrong (plus or minus 20% or so), then the next stage is to let them explain it to each other, and that’s where the peer instruction comes in. When you have students argue it and take a stand and a position on it, we get a very significant gain and improvement when we then let them vote on it a second time… and the usual practice is not to reveal the poll results or the answer until after they’ve had that opportunity to engage in that discussion.
Rebecca: Just make sure, to make sure I understand correctly: you do the poll, you see the results as students don’t see the results…
Rebecca: …based on their answers or their responses when you decide whether or not they do the peer instruction piece. How long do they usually talk to each other about the topic?
John: It depends on the problem and normally I will have some undergraduate TAs and I’ll wander around the class and see what they’re talking about, listen in, answer some questions from them and the TAs will be doing the same thing…. and it’s usually pretty clear when they’re coming to a consensus. You can see them reaching for their clickers or their phones and getting ready to vote, so generally it may only be a minute or two, it could be longer… it depends on the complexity of the problems. Some of the problems require a bit of effort and require some calculations, but normally they’ve already done that… so, the second stage, where there’s a discussion, you can hear the volume build-up and then as they’re approaching solutions and consensus, it tends to drop back down again. It’s fairly easy to get a pretty good read on where they are and when they’re ready to vote again.
Rebecca: I imagine that you would really need to keep your ear to the ground, otherwise chaos could ensue. Because now, if they’re finished talking about the problem and there’s still time, then they could easily derail if you’re not quick to get back to the clicker question.
John: Right, and normally the time is generally held fairly tight. I suspect sometimes it’s only 30 seconds to a minute, other times it may go up to a couple minutes, but if I see them getting distracted and doing other things, the polling starts immediately.
Rebecca: Obviously technology is your friend in this particular situation. Can you talk a little bit about the technology you’re using to manage this many students all at once?
John: Here, we’ve adopted iClicker as a campus standard, so we use that in pretty much all of the classes where we’re doing polling and there’s both a physical radio frequency clicker that students may buy or they can buy an app and pay by the semester or over four years for the use of the app.
Rebecca: How do you make sure that the cost doesn’t get prohibitive to students?
John: That’s an issue, and it’s been a major source of concern…
Rebecca: They’re not very expensive, right?
John: Well, they can be expensive. A new clicker costs somewhere around $40. A used one can often be purchased for $15 to $20, sometimes less… and the apps I think, are somewhere around $12 to $15 for a semester and I think about $35 for four years.
Rebecca: …and you can use the clickers in all of the classes, right? So if multiple faculty member(s) are using all the same system, then the investment is a good one for students.
John: …and that’s why we have a campus adoption because in places where you don’t have that, students might have to buy two or three or four different clicker systems in different classes. So, once they buy the clicker for one as long as they hold on to it, they can use it in classes for the rest of their career. Almost everyone in the economics department, for example, now uses clickers, so if they’re economics majors or business majors, it’s very likely they’ll use them in multiple courses. The cost is much more tolerable when it’s spread out over multiple classes.
Rebecca: The other area where you do some peer instruction in these large classes is in writing. Which seems kind of crazy. You have all these students in this big classroom and somehow you manage to do writing assignments.
John: Yeah, my large class generally is somewhere between 350 and 420 students. At one time, for actually about a decade or so, I was giving weekly online discussion forums. But grading that or evaluating that and providing feedback was taking an awful lot of time…probably 30 to 40 hours a week. So, I pretty much…
Rebecca: A full-time job in and of itself…
John: I stopped that a few years ago and, a few years back, I replaced that with calibrated peer review assignments. The calibrated peer review system is something that Eric Mazur talked about while he was here… a visit in 2014… and when he mentioned it, a lot of people got excited. The way the system works is that you create an assignment, you store it on a central server at UCLA, and then it’s something that other people can adapt and use and modify—it’s released under a license, which is similar to a creative common license within the system… and you create the assignment… you create an evaluation rubric for the assignment… and you have to be really careful in designing that to make sure it’s one that students will be able to apply, because other ones that do that… and then you create three sample assignments yourself: a low-quality one, a medium quality one, and a high quality assignment… and you have students submit their own assignments first (according to the rubric and guidelines you provide to them)… then they go in and they evaluate the three that you’ve done. They’re given in random order, and they’re assessed in terms of how closely their evaluations match yours. That’s the calibration part. Students receive a calibration score based on how similar their evaluations are to the ones that you assigned to the sample responses. Then after they complete that stage, they evaluate each other, using the same rubric, and a weighted average of those scores is assigned as a component of the grade. They’re graded in a number of dimensions. One is based on the weighted average of the peers, where students who had a high calibration score will have evaluations that rate more highly in evaluating other students. They’re also rated in terms of how closely their evaluations match the others during that stage. So, if their evaluation is an outlier… much higher or much lower than other students… they lose some points on that… and then after they evaluate the other three students, they rate their own work… and one of the goals of that is so that they have improved metacognition. That, by the time they go back and look at their work again, they’ve rated three works by the instructor and three assignments done by their peers and then they’re asked to evaluate their own using the same criteria. What’s really interesting about the calibrated peer review process is their grade on this is tied not to whether they give themselves a high or low score on this, but it’s how close their self evaluation comes to the weighted average of their peer evaluations. So, they have an incentive to try to look at their work more objectively, and not try to game the system… because if they score their work too high or too low, they could end up with no points on the self-evaluation stage. So, the closer they get to the weighted average of their peer evaluations, the higher the score will be on that component.
Rebecca: I think that’s an area that we often see students struggling, is being able to effectively evaluate their work or other work. So, really training them to use a rubric and understand and think about what’s important or what’s not important about particular kinds of assignments or particular kinds of work could be really valuable to students in a way that we don’t really have other systems to do that.
John: The nice thing about this is it scales really easily. There’s a lot of upfront work in creating the assignments, creating the rubric, and a really good practice is to test them thoroughly before you give them out the first time. What I normally had done is asked some of my peers to look at that, some of my colleagues to look at it, and sometimes I’d have some upper-level students were…. and this does give students a little bit more reflective practice, where they get to look at their own work a bit more critically, perhaps, and reflect on it and see how they’re doing compared to how other students are doing in the course… and I think that’s helpful.
Rebecca: I think that the rubric would probably be a challenge to make but I think what would be more challenging is putting up those different assignments that are scored at different levels at the very beginning as your calibration tool. What strategies have you developed to make those in a way that it doesn’t take forever?
John: Well, I only do this three times in a semester, and once you’ve done it once, if you design it in a way so that it won’t go stale… and I generally have students, for example, find some articles in the news in the last six months that relate to a topic that we’ve talked about, or I ask students to find some examples in their own life to illustrate behavioral economics concepts in one of the assignments, for example…. where it’s not something that they could easily copy and paste from other people’s work. Because, there is always a concern with academic dishonesty and so forth. You don’t want these things showing up on Chegg or any of those other systems, where it would be easy to copy and paste good responses. So, I’ve tried to design assignments where once they’re done, they can be used for multiple years in one form or another. I modify them each year based on how they work. But perhaps a more serious problem is what happens when students really don’t like the evaluations. One of the things I’ve done when I’ve used this is to have three of these assignments, but I drop the lowest score… because, sometimes people will get some scores back that they didn’t expect or they may have neglected to look at the rubric I sent them and they may have omitted a major part of the assignment and ended up losing quite a bit of points all the way through that. But, as long as one of the scores is dropped, they have the opportunity to learn from their mistakes and do a little bit better. But, there are procedures built-in that make it easier to catch any outliers when you have someone who is just rating everyone extremely highly or rating everyone really poorly—inappropriately highly or poorly. There are tools in it which will give you a list of all the cases where there’s a high variance across reviewers or where someone happened to be evaluated by people who had very low calibration scores… so, if you end up with two out of the three peer reviews with low scores, that’s something that’s flagged by the system. I check all the cases where it’s flagged and I tell the students if they’re unhappy with their score or if they have any questions about it, to contact me, explain why they’re dissatisfied with their score, and then I’ll go in and look at it. In nearly all cases, it’s been an issue with the students submission and not with the peer reviews. Because, while some people tend to overrate things and some tend to underestimate some of it, compared to where I would evaluate the wok… on average, it’s been very close, typically, to what I would have scored or what I would have assigned as a score. But I do make, in rare cases, some adjustments when I see that something went wrong in the process.
Rebecca: Do you prevent students from seeing the score then, until you’ve reviewed all of the scores to make sure that you’re okay with what has happened before they have access or…?
John: In this system, that really can’t be done easily…
John:… because what happens is they get the results as soon as the last stage is completed. I’ll send a note out saying, “Now that the stage is completed, you can review your scores, you can read all the comments that your peers have provided, and you can see what your grade is at each component…” and we have gone over that in class so they know what they’ll be seeing.
Rebecca: What kind of workload do you end up with, dealing with problems?
John: In general, when I’ve used this in the class of 360 to 420 students, there’s usually 3 to 5 students who find their grade unreasonable, and sometimes, I found the grades perfectly fine. Occasionally one or two of those, I’ll make some minor adjustments to—if something went wrong where one of their peer reviewers didn’t show up, for example, one or two of them didn’t complete that stage of the assignment, and someone was overly harsh or perhaps overly harsh in their grading, but it’s rare.
Rebecca: Can that system be used for things other than writing? Like other kinds of documents?
John: It could be used for any type of document because basically students will either write something up or they’ll submit something and it could be an image, it could be used for peer review, or calibrated peer review, on pretty much anything as long as it can be disseminated in digital format. It could be used for websites, for example.
Rebecca: Well, that’s what I was getting at when I was asking.
You also teach some upper-level seminar courses with 30 or so students. This semester, you tried a two-stage exam after talking with Doug McKee when he was on campus about it. What is a two-stage exam and how did it work?
John: Backing up a bit, I was considering it even before Doug came here because I heard the episode of the Teach Better podcast where they discussed a two-stage exam and then when we were talking here and he was in one of our earlier podcasts and we discussed this very issue, I became more interested after we talked with Doug. A two-stage exam is one where in the first stage of the process, students take the exam by themselves and then in the second stage, they do some group work– either on a subset of the questions or on some very closely related questions. It’s being used quite a bit in the sciences and there’s a growing amount of research indicating that it has been successful. Some studies have found weak results, others are finding stronger results, but it’s still fairly early in the exploration of this. The Carl Wieman Science Education Initiative has quite a few resources associated with two-stage exams. This leverages peer instruction in the second stage.
The usual process, or the most common practice, is to take the exam period and have students work on this for the first two-thirds or so of the exam time slot and then they work in a group in the last third. I did it a little bit differently than this. In my case, I gave the exam on a Wednesday and I graded the exam but didn’t get them back to the students and then I selected a subset of the questions and I had them work on them in groups on that Friday… and that worked pretty well too, they had a little chance to review in between, they didn’t get to keep the exams, but there were only seven questions on it. They could go back and review things. I didn’t tell them which questions would be on the second stage in large part because I didn’t know. I told them that two of the questions would definitely be on it, but it would depend on how they did on the other part. So, I was able to look at the exam, find the parts where they had the most trouble, and assigned those as ones for the second stage… and in general, it was a remarkable experience. It was really nice to be giving an exam and to see students working in groups of three or four, actively discussing the issues, arguing over them, trying to explain things to each other and it was a really fun experience. It was very energizing to see that much effort being devoted to try to understand concepts that students had some difficulty with.
Rebecca: I remember seeing an image of your class being really actively engaged, really talking about the core class material that you shared during your test and I think the caption was: “This is during a test!”
John: Yes, I took a picture of it from my phone and I think I sent that to you during the exam because it was just so exciting to see that… and it was also a reminder for myself just how well this was working. I wandered around the room and listened in on the discussions and they were all very focused and coming up with much better explanations of these things then they would have likely been able to see if it was a whole class discussion… because they were very focused, they were arguing over what was the best approach to deal with some of these problems. I could see people making connections and suddenly understanding how things they had done before fit in and pulling together a lot of concepts that they might not have done as effectively if it had not been for those small group discussions.
Rebecca: Were you tempted to join in on those conversations because they were so lively?
John: I was, but I mostly just listened in and let them work it out themselves… and in general, they did quite a bit better… and what I should have mentioned before is that the overall grade for the exam is a weighted average of the first part and the second with most of the weight being on the individual part. One of the things that really appealed to me is that typically, when we give an exam and then grade it and return it, the students who did well generally just put it away and are happy with the results and they may glance at some of the things they got wrong (if they got many things wrong), but they’re not going to spend a lot of time actively processing it. The students who did poorly tend to get discouraged, some of them may give up a bit, but rarely are they likely to go back and try to put in the effort to correct their mistakes and to see where they went wrong. It was really nice to see that processing taking place by both groups. The students who did really well the first time deepen their understanding by explaining it to others and I suspect that should increase their long-term recall of this. The act of explaining it to others in some studies seems to be really helpful in encouraging transfer, where you can take concepts and apply them to other circumstances and when you’re in a course like econometrics, you have to be able to apply the same concepts in a wide variety of topics and areas. I think it was a very useful experience.
Rebecca: I think it’s a great method to allow some time and space for a reflective practice, because students tend not to do that on their own unless they’re asked to do it and if you do it as a homework assignment, I suspect that students don’t really spend that much time doing it, but this time they spent the whole class period doing the reflection. So, that seems really valuable.
John: Because I know a lot of people will do that. They’ll have an exam, they’ll give it back to students, and they’ll tell them they can make up part of the grade if they turn it in with corrections… and many students would do that, but I don’t think that would be as effective as having the group discussion on this. Some of them were able to make very clear what they didn’t understand and then they were able to get explanations from others and sometimes the explanations were right, sometimes they were wrong, but they had to process it much more actively and that’s always helpful, I think.
Rebecca: The grade weights is what seems most compelling to me in this situation because I’ve offered quizzes in my classes, more low-stakes assignments where I let students work on it for a while and I don’t tell them that they’re gonna get to do some peer instruction as part of it, but then they’re struggling with what they’re doing and then I say, “Oh, well, you have five minutes to work with your peers to revise anything you want to do before you turn it in.” And those generally result in some pretty active conversations as well, but there still are those few students who just copy down the answer and don’t engage in the conversation… but I think if there was that wait between before and after, that would really change that dynamic. So, I think that that’s a really compelling opportunity.
John: I thought it was useful and another reason why I didn’t do it all at one stage in one day is because I’m teaching on a Monday-Wednesday-Friday schedule and we only have 55 minutes and I have quite a few students in the class who are not native English speakers and they always take more time or they need more time to process and write information in a second language. So, I didn’t want to constrain the time and make both parts of it much shorter.
Rebecca: If you encourage people to practice and retrieve that information in extra time outside of class, there’s nothing wrong with that either.
Rebecca: I’d rather the students learn the material rather than just panic about a test. What do you recommend to our listeners to read to learn more about this evidence-based practice?
John: In terms of peer instruction, Derek Bruff has a really good book on using clickers. Eric Mazur’s original book on this, which is now slightly over 20 years old, is still very good… where he describes a process of developing this peer instruction technique. Eric Mazur also gave a talk here a few years ago and we have a recording of his presentation on this. There’s a really great example in there where he used peer instruction and what was most compelling about it, and Rebecca’s heard this before, but…
Rebecca: I was there!
John: …and Rebecca was there, was he used this example where he gave a really short presentation on what happens to the hole in a plate of metal if you heat it up… and people were asked to vote on that and then they had a chance to discuss it.
Rebecca: …and he never told us the answer!
John: …and then he noted how energized people were and he said, “You were so actively discussing these things…” When he tried to go on after making a point about how they suddenly were interested in something they normally wouldn’t have been interested in… he started to go on to the next topic. People were really upset, because they wanted the answer and he finally gave the answer, but he did that deliberately to show that this sort of thing… where the students don’t know the answer but they committed to a position and they want to know if they’re right… builds a sort of interest in learning that might not intrinsically be there otherwise.
…and that’s exactly what I saw, by the way, in my exam. They were so actively discussing things that normally they’d be bored out of their minds with. So, that environment can be very supportive of learning.
Rebecca: Yeah, it really gets people curious. I remember being in that room… dying to know what version was right? People had such compelling arguments.
John: Exactly, and that’s why it’s really good to pick questions, with any of these things, where it’s not going to be clearly obvious, where they have to process it, and they have to make connections, and you could build a case, correctly or wrongly, for different answers, and people want to know what the answers are.
Rebecca: I mean it was key that he finally gave the answer, right? So there was some corrective feedback there, so that people didn’t continue to mislearn the information.
John: And that was nearly four years ago, and we remember that very vividly. If that was just a point in a class that was given… say, four years ago, we probably wouldn’t be talking about that now.
Rebecca: I can’t believe it was that long ago.
John: I think it was.
Rebecca: It was a while ago.
Rebecca: …and so I’m dying to know, what are you gonna do next?
John: One of the next things I’m going to do is a follow-up to something we talked about in an earlier episode, when we talked to Judie Littlejohn about the metacognitive cafe. One of the things I’ve been observing is that the use of this process by having students work to improve their metacognition about how they learn and what they’re learning… Students, at least, perceive there is being some significant learning gains from that. That’s convinced me that I’d like to do something similar in a large class, but an online discussion forum for 400 students again doesn’t scale quite as well. So, I’m going to be doing some weekly activities and I’m working with Liz Dunne Schmitt who teaches our large macro class in the spring semester, and a couple of other people: Kris Munger, and Michelle Miller, who also who’s the author of Minds Online: Teaching Effectively with Technology (and was a guest here a while back). We’re going to try to put together an experiment where we use some evidence-based methods as weekly assignments, say for ten weeks in a semester…. that’s our current plan at least)… and students will be exposed to this… and they’ll engage in some sort of reflection or some practice with one of these activities… and then in terms of evidence-based methods of learning, such as retrieval practice, spaced practice, and interleaved practice, and similar things… and then we’re going to see how that exposure along with some reasonably easily assessed activity, which could be just some short responses in a forum or it could be perhaps some online quizzes, evaluating whether that impacts their actual behavior in the class, and their actual performance in the class. One-half of the group will be exposed to those types of interventions, and the other half will be exposed to some form of standard study skills module, because most of the students in this class of freshmen and basically what we’re looking at is, if we present students with evidence on what really increases our ability to learn, whether that will result in significant change in either their behavior, or in their performance. So, we’re going to try, at least the plan, is to try to see whether that affects the number of times they take quizzes that can be taken repeatedly, whether it affects the number of times they log in and view other materials, and whether it changes a perception of how we learn. so right now we’re at the…
Rebecca: And performance too, right?
John: …and their performance.
Rebecca: And is the plan to start collecting that data in the fall?
John: The plan is to put all this together the spring, I’m hoping and then to submit a proposal to the IRB, and then to conduct the study and the fall and the spring, at least for a first stage and then we’re hoping to be able to follow these students up, to see if this has a significant effect later in terms of their grades or their persistence.
Rebecca: Sounds pretty exciting. I’m looking forward to hearing how that goes.
John: It is. I’m looking forward to it being all together and actually being implemented. I think it’s an interesting study.
Rebecca: We’ll have to have you back, John.
John: I think we can manage that.
Rebecca: Well, thanks so much for sharing all this information about peer instruction. I know it’s something that I’m always kind of asking you about and like to hear about, and I’m sure others will too.
John: Well, thank you.
John: If you’ve enjoyed this podcast, please subscribe and leave a review on iTunes or your favorite podcast service. To continue the conversation, join us on our Tea for Teaching Facebook page.
Rebecca: You can find show notes, transcripts, and other materials on teaforteaching.com. Theme music by Michael Gary Brewer. Editing assistance from Nicky Radford.