The n-Category Café

I’ve never done this, but I’ve been thinking for a while about trying it, so I’m interested to hear your experience! The main thing holding me back is a worry that the students won’t do the reading. Has that been a problem for you?

As for activities, are you trying to think of something new and different every week? I wouldn’t worry about repeating myself; in fact I’d be inclined to stick with only a few activities that work (once I’d found out which those were). Can you share your experience of which activities worked well and which didn’t?

The main new thing I’ve been doing this year in the classroom is “voting questions”(aka “clickers”, although we use a free smartphone app rather than actual clickers). I’ve been very happy with this overall, although it’s reinforced my fears that students won’t do assigned reading: even the easiest of voting questions based on the reading don’t get very good responses.

I’ve been using the flipped classroom for about five years now, in subjects ranging from calculus to CS courses. I blog about some of those experiences over at Casting Out Nines, which is hosted by the Chronicle of Higher Education (it should be linked to my name above).

The more I’ve refined the process, the better it’s worked (no big surprise there) but the most gratifying thing I’ve found is that my students are more engaged, better able to learn things on their own, and are becoming very enthusiastic about this approach to college work. They are coming to the conclusion, which I think is correct, that in order to really learn something, you have to struggle with it and experience it for yourself, rather than just sit in a lecture. (Although lecture has an important place at the table and always will.) I give about 4-6 workshops a year on the flipped classroom at various college campuses, and the more I work with faculty on this technique the more exciting it becomes. I really think in five year, we won’t be talking about “the flipped classroom” but rather just “the classroom”, as the flipped structure becomes normative.

I invite everyone to come on over to Casting Out Nines – I’m wrapping up a series where I debrief the completely flipped calculus 1 class that a colleague of mine and I did last year.

I tried a variant of that partially flipped format in some Operational Research lectures some years ago. (The idea worked because of the type of subject and might not be adaptable to other topics.) I set questions on problem sheets (before they had met the method in lectures) that required them to try out some ideas as to how one might attack the (`real life’) problem, e.g. a transportation problem. Usually one can get a certain way in these problems using basic common sense and some earlier material (e.g. linear progamming). The problems are often combinatorics at least to get initial solutions, then iteration of some process gets the optimal solution.

The format I used would typically sketch out some ideas for them to try. It seemed to work well. It meant that when I talked about the algorithm later on in the lectures the students seemed to appreciate the way that the taught algorithm got around some of the difficulties that they had encountered when attacking things in the problem class. Once or twice students came up with really good points about why a particular method worked, insights that I had not met before. The lectues felt a bit more like a shared learning experience and they seemed to like that. I also wanted to emphasise that when you go to a maths class you should not leave your common sense outside!

The one thing I’m missing from your description above is a way for the students to assess their learning as they go through the material before class.

That is, either questions for them to answer as they go through the material, or (easy!) opportunities for them to ask questions or just say “I didn’t get this!”. This should be done in such a way that you get the chance to see their responses well in advance of the lecture so that you can take them into account during the lecture.

Then there is a definite incentive for the students to go through the material beforehand: if they do so then their questions will be answered in the lecture time. If they don’t, then they have to endure other people’s questions being answered without necessarily knowing all the background.

The problem I see with how you’ve done it is that half lecture at the start. I’m not sure that it serves any purpose. Either you assume that they’ve all read the material, in which case your best choice of activity is to find out what they didn’t understand and go directly for the jugular, so to speak. Or you assume that they haven’t, in which case those that have wonder why they bothered.

But it can still be useful to have a time when the expert speaks. So I would have a half lecture before they read the material giving them an overview of what’s coming up. Then they read the material and ask questions/indicate stuff they don’t understand. Finally, you take that up in the flipped part of the class.

In Sheffield we have started teaching our maths courses for first year engineers in this fashion. We seem to be calling it ‘blended learning’ rather than ‘flipped classrooms’, I forget why.

We realised there was a need to change the way we were teaching the engineers as the traditional approach did not seem to be inspiring the students. We also have a massive spread of ability in these courses: some of the engineering students are at least as good at maths as our best maths students and some of the engineering students are quite a bit weaker at maths. This resulted in more extremes of feedback of the form “the lectures go way too slow” and “the lectures go way too fast”.

The traditional format was two hours of lectures and one hour of problem class, where the students sat and work on the standard problem sheets, putting up their hand to ask for help when they got stuck. Attendance at the problem classes tended to drop off rather quickly, leaving an enthusiastic core, but indicating that many were not getting much out of these classes.

We decided to switch to a mix of videos and two problem classes (of a more interactive nature) per week. Based on the local Catster knowledge, we opted for ten minute videos. The students have a couple of days before the problem class in which they must watch three ten-minute videos.
A group of about seven of us recorded the videos between us. Generally the videos are in the Catster format of one person talking in front of the board. Some people edited in some computer graphics, but in general there was minimal editing.

At the end of each video the students must answer three online questions. These are generally quite simple questions that test mainly that the students have watched the video. Our online testing system is built on maple, so the questions can be appropriately randomized and require mathematical input, as opposed to just multiple choice. As an encouragement to do the questions, the marks for them count for a small amount of the final grade.

The problem classes begin with the person leading the class giving a five-minute summary of the videos. They then go through a couple of simple questions with input from the class. The students then get a sheet of three-ish questions based on the videos they’ve watched, but quite varied in scope: from a little bit of basic material to some more interesting and challenging questions. A lot of effort went into trying to think up good, balanced, thought-provoking questions. The students can work on these in groups and in any order. This leads to more cohesion between the students, given that they are working on roughly the same problems, rather than some racing ahead. In the old system the problem classes included a lot of drill problems which the students are generally better working on at home. Indeed they still have the standard problems which they are encouraged to work on at home.

Unfortunately, because I’ve been teaching other courses I haven’t lead any of the problem classes, but I hear that most students seem to be enjoying them and initial feedback seems to be very good indeed. We’ve been testing this on a relativity small cohort of engineers this year, but intend to expand out to most of the first year engineers next year.