BIOL 315 (Quantitative Biology I) is the introductory biostatistics course here at Calgary. It’s a big course that was working adequately as a conventional lecture+labs course. But yet we decided to flip the course—and so far it seems to be working. Want to know why we did it? Read on! I’ll cover how we did it in a future post.
To interpret any advice and figure out if it applies to you, you need as much contextual information as you can get. So here’s some context to help you understand why I thought flipping our intro biostats course was a good idea. Your mileage may vary!
The University of Calgary is a large (~25,000 undergrads) public research university. BIOL 315 is 132 students (max), mostly in their second year. Roughly half the students are required to take it as part of their major program (Ecology, Zoology, or Environmental Science) or because it’s a prereq for upper-level ecology courses they want to take. Most of the rest are premed students who take it to fulfill med school admissions requirements.* It’s offered both semesters (so 132 students/semester). Because of student demand and other constraints, there’s no possibility of making the course smaller by offering multiple sections/semester or lowering the enrollment cap. Not anytime soon, anyway.
So we’re stuck offering a big course to a diverse range of students, most of whom do not have a pre-existing interest in the topic. What to do about that, pedagogically?
Well, until this year I didn’t do anything special. In the semesters in which I was assigned to teach the course, I taught it as a conventional lecture (as did other faculty). Oh, I tweaked it around the edges every so often. Adjusted the topic coverage. Adopted a textbook (Whitlock & Schluter). Asked some clicker questions during lecture, to get the students thinking and give me and them a bit of instant feedback on what they had or hadn’t understood. Made a small part of the course mark dependent on answering a sufficient number of clicker questions, so as to encourage attendance. Gave a few of the lecture sessions over to doing exam-style practice problems. Added peer mentors, students who’ve taken the course previously and who want to help their classmates do well.
And it worked, more or less. Most students did fine. They attained an adequate understanding of the material, as evidenced by their performance on exams, assignments, and in subsequent courses. A minority did awesome. Only a very few failed. (Aside: you sometimes hear people arguing against lecturing because of a few studies suggesting it puts many students and/or certain groups of students at high risk of failure. That argument doesn’t apply here.) And while students weren’t thrilled with the course—student evaluation scores tended to run lower than the departmental and faculty averages for courses at the same level—I wasn’t too worried about that, for a few reasons. First, the student evaluation scores weren’t bad in an absolute sense. Second, I knew from surveys that students mostly don’t come into the course looking forward to it, and you’d expect student reaction to any course to partially reflect their pre-existing feelings about it. Third, the purpose of any course is to teach students something, not to make them happy. It’s nice if students like a course, and there are circumstances in which liking a course might help students learn. For instance, students might put more effort into studying for courses they enjoy (then again, maybe not–effort depends on lots of factors). But student learning is the ultimate goal. (Plus, students’ course evaluation scores mostly don’t correlate with learning, or even correlate negatively).
But all that tweaking wasn’t having any apparent effect on student learning. This is consistent with my experience of teaching: it’s not that hard to bring the class as a whole up to some adequate level, but very hard to bring the class as a whole up higher than that. Put another way, I seemed to be stuck on a “local fitness peak”, pedagogically. Finding a higher peak on the pedagogical “fitness landscape”, if one existed, was going to require a (directed) “macromutation”. The course was going to have to be totally revamped.
This had two obvious downsides. First, it was going to be a lot of work (but see). Second, there was some risk that it would make the course worse. But several considerations mitigated those downsides for me. First, I have an amazing colleague, Kyla Flanagan, with whom I currently share the job of teaching BIOL 315 (we alternate semesters). She has experience and background knowledge concerning a range of modern pedagogical approaches liked flipped classrooms. She wanted to restructure the course, and convinced me that I should teach the restructured version too. She also took the lead on the new prep, which totally spoiled me and made the decision to adopt a new structure much easier for me. Second, we had colleagues in our department who were happy to share their experiences with flipped classrooms. Third, our head of department cares about teaching and is very supportive of faculty who try to improve their teaching. Fourth, I felt myself getting stale as a teacher. I wanted to force myself to break out of my pedagogical rut (something Brian’s also talked about). Fifth, I didn’t feel any particular attachment to lecturing as an approach. I have every reason to think I’m a pretty good lecturer, but clearly I’m not so good as to be able to get 132 University of Calgary students all doing great in intro biostats. Bottom line: if you know how to cook, you should be able to follow a new recipe that others have already tried with success, without screwing it up too badly. And it’s fun to try a new recipe sometimes.
So this past fall, we flipped the BIOL 315 lectures.** Students now have to come to class having done background reading, and over half the class time is given over to quizzes and other “active learning” activities that can only take place when students are all in the same room as the instructor and each other. It seems to be working. The class average was up about 10% last fall if memory serves, even though the exams were broadly similar to those in the past (i.e. we didn’t sacrifice much in terms of breadth of coverage). N=1, obviously, and one success is no guarantee of future success, especially since Kyla taught it in the fall and now it’s me doing it this term.*** But so far I think things are going pretty well.
Now that I’ve whetted your appetite, you can look forward to a future post in which I’ll explain in much more detail how the course works. “Flipped classroom” is a broad term covering a bunch of pedagogical approaches, so just telling you that we flipped the classroom isn’t very informative…
*A few have taken a stats course in another department and take 315 in the belief that it will be an easy A because it covers material they’ve already had. This belief usually proves to be mistaken, for various reasons. And a few have taken a stats course in another department, but are taking 315 because they want to learn R. As far as I know we’re the only intro stats course on campus that uses R. The intro stats course in the Mathematics and Statistics Department uses Minitab (I know, it surprised me too when I learned this). And I think the stats courses in Economics use Stata, which only economists use as far as I know.
**The labs remained unchanged. The primary role of the labs is to teach the students how to do graphing and statistics using R. Even if we’d wanted to revamp the labs as well as the lectures, changing both the lectures and labs at once would’ve been both too much work and too risky.
***There is some data suggesting that some or even all of the apparent positive effect of active learning approaches is because those approaches have mostly been adopted by highly-trained teachers. Which I am not. No one teaching approach is always and everywhere superior to others, for all instructors. In teaching, as in life, you need to do what works for you. Seriously.
One small point I forgot to make in the post: One reason we flipped the class is that other easier ways of incorporating active learning into the lectures wouldn’t scale to a class this size. A while back I bought Andrew Gelman’s book on a “bag of tricks” for teaching introductory statistics. Lots of great ideas for active learning of statistics in there–none of which scale to a class of >60, according to Gelman himself.
I hope you address in your follow-up how “active learning” differs from what you do in lab, especially if you use lab to do little monte carlo simulations to gain a deeper understanding of basic (and advanced!) principles.
Oh, I definitely consider the labs a form of active learning! Didn’t mean to imply otherwise. But lectures + active learning in the form of labs was only getting the class as a whole up to an adequate level.
Also, for many students the experience of the lab is dominated by the challenge of learning R, especially in the first half of the term.
Via Twitter, Meg questions whether 132 students is “huge”. Fair point–the course has hundreds fewer students than our first-year bio courses, which have 500+ if memory serves. But it’s “huge” in a couple of other important senses.
First, it’s huge compared to every other intro biostats course I’ve heard of. In comments on old posts (such as the one I did asking readers for advice on textbook choice for this course), nobody ever referred to having taught or taken intro biostats with more than ~60 students. Commenters are a small and biased sample, of course. But I’m fairly confident saying that intro biostats at Calgary is much bigger than most intro biostats courses, even at big universities.
Second, it’s “functionally huge”, meaning that many teaching techniques one could use in a smaller course can’t be scaled to 132 students. For instance, I mentioned in the post that pretty much nothing in Andrew Gelman’s book of tricks for teaching statistics can be used with >60 students. Functional hugeness is a nonlinear saturating function of number of students in the class. For many pedagogical purposes, 132 student is *way* more than, say, 30. But 500 students isn’t *that* many more than 132. There aren’t many pedagogical approaches that are feasible for 132 students but infeasible for 500 (though there are some).
My big course at Georgia Tech (General Ecology) had 80-90ish students. That felt pretty different to me than when I first taught Intro Bio here (when I taught one section, with “only” ~250 students) and now when I teach both sections (~550 students). Teaching 550 doesn’t feel that different to me than teaching 250, with the exception that there are more emails and issues related to students to deal with (related to accommodations, illnesses, etc.). In terms of what I can do pedagogically, for me, there was no real difference between 250 and 550, but a notable difference between 90 and 250.
It would be interesting to know where the shift is, and how much it differs for different instructors and topics. Some of why the classes with several hundred feel so different for me is that, at 250, I have no hope of even recognizing all of my students. With 90, I could recognize them all (including telling that one student at an exam was one who was not enrolled in the course), and knew the names of a reasonable fraction of them. With 550, I know the names of only a handful of students. That is — I know the names of fewer students in my course of 550 than I did in my course of 90.
In terms of assignments, I’m guessing that the shift is fairly gradual, with certain assignments becoming infeasible at different numbers of students.
As I’ll discuss in the follow-up post, we flipped the class in such a way as to make it functionally smaller. Briefly, the students are assigned to permanent teams of 5-7, thereby effectively cutting the class size to ~1/6 of its actual size for some purposes.
Sorry this is late, but just in case you’re collecting data points, our Intro Biostats equivalent course (for 1st year biological sciences students) has 190 students this year, with plans to increase to 225 over the next couple of years. Limiting factor is the computer lab where we analyse data (using Excel for simple formulae and Tables of Critical Values), which only has 80 seats, meaning triple teaching ‘active sessions’ at the moment. Not cool!
Of course, it’s only after I put this post up that I hear about not one but two intro biostats courses bigger than ours. 🙂 (yours, and Owen Petchey’s which apparently will be 320 students).
Ours is going up to 170 next fall, because we’re only going to be offering it in the fall, but I think the plan is for it to drop back down to 132 the following year.
Re: computer labs, ours seat 24. The computer lab rooms are heavily used by various courses, so the course size is limited by the number of lab sections we can offer.
I’m not a fan of recording lecture sessions for students to view later (various reasons) – but I think videoing instructions for computer labs could be a really useful thing. Especially if you have to repeat a session 132 / 24 times! I lose my voice after the 3rd (and final) repetition.
Students working at their own pace to cover the basic material before the lab, then showing up to an open session with any questions would address a lot of the feedback I get; typically “this is too easy, why am I being treated like a child” vs “the lecturer went far too fast, I couldn’t keep up” [hmmm, note to self about flipping this part of the course…].
Andrew Jackson (@yodacomplex) has a great set of R intro videos on his website that I already get my postgrad stats students to watch before the course starts:
On the other hand, not everybody thinks highly of instructional videos. 🙂
Jeremy, our first year courses are around 1000-1200 for Energy Flow in Biological Systems and 750-800 for DNA, Evolution and Inheritance. So three sections and two of 350 respectively. Our biggest lecture halls seat 400 max.
Yes, yes, yes, I know. 315 is not “huge” compared to those courses. But it’s huge compared to other intro biostats courses I’ve heard of (though I just found out a friend of mine will be teaching a 320 student intro biostats course). More importantly, it’s functionally huge, meaning that it’s too big for many of the pedagogical techniques you might want to use. As Meg noted, that raises the question of what pedagogical techniques are feasible with any given number of students. I do think there are some things you can do in 315 that you can’t really do in our first year bio courses, which makes our first year bio functionally huger. But I think once you get up above 250-300 or so per section, I suspect that’s about as huge as it gets, functionally (?) Is there anything you can do teaching-wise in a section of 300 that you can’t do in a section of 600? (or 2000, which I’ve heard is done in Singapore!)
I wasn’t in any way questioning that 315 is huge from a biostats teaching perspective!!
Just giving the background of the rest of what we deal with, and I agree, actually, that over ~200 students it’s functionally the same as long as your room can hold them.
Sorry for misreading you Jessica. Meg already gave me a hard time about the whole “huge” thing, so I was primed for others to give me a hard time too. 🙂
Pingback: Teaching in an Active Learning Classroom: Pros and Cons | Arthropod Ecology
Pingback: Flipping our big intro biostats class didn’t work (for me) (yet) | Dynamic Ecology