As I said last week, I recently wrapped up a term experimenting with “active learning” techniques in the two intro courses I was teaching. The diagnostic test results were a mixed bag– one section showed really good improvement in their scores, the other was no better than the same class with traditional methods– and the exam scores weren’t really any different. There was probably some slight improvement in the multiple choice– fewer students got trapped by the Newton’s First Law questions than usual– but in the free-response problems, they did about the same as usual.
So, having turned in my grades, what did the students have to say about the course on their end-of-term surveys? Again, this was very much a mixed bag.
Really, this experiment was a great demonstration of the problem with student course evaluations as a method of evaluating anything, because the signal-to-noise ratio sucks. I got a whole bunch of comments complaining that I was unavailable for help, some of them nastily personal, which it’s really hard not to resent, because my unavailability was directly related to the whole new baby thing.
I also got a bunch of complaints about the fact that I didn’t spend enough class time lecturing about the basic concepts. Which, you know, was the whole point of the exercise. As I explained in detail on the first day of class. So, again, it’s hard to give this much weight, other than as support for the prevailing conventional wisdom that asking our students to read the textbook will be regarded as a gross imposition.
The main substantive complaint was that I didn’t do examples in class that were like the homework problems. This, at least, is a fair point– I didn’t, in large part because I was expecting them to ask more questions about the homework. This has been a problem with traditional-style classes in recent years as well– I used to be able to rely on students asking questions about the homework every day, but the last few years, they’ve just sat there mutely whenever I ask if there are questions. I probably need to start designating one homework problem per assignment to go over in class whether they ask for it or not, because silence clearly does not indicate understanding.
This is somewhat problematic, though, because doing example problems takes up class time, and it’s not clear that it does any good. That is, there is evidence from the Physics Education Research community that going over example problems in class or in a recitation session does not substantially improve problem-solving ability. Which makes sense– watching someone else do a problem is great for producing the illusion that you understand what they’re doing, but doesn’t necessarily help you solve a similar problem on your own.
A better tactic is to get them to actively solve problems in class, but that’s been really difficult to do. Again, several years ago, I could get decent results from just putting a problem up on the board, and asking the class to work on it while I circulated to give advice and answer questions, but in the last few years, this has been a really dismal failure– at least half of the students just sit there staring at a blank sheet of paper, saying they have no idea how to start.
A possible solution is to try to integrate this with the in-class polling stuff, but the attempts I made at that had only mixed success. Putting up a whole problem hits the same “I don’t know how to start” issue, while trying to break it into manageable chunks that can be asked as multiple choice questions only sort of works– the individual steps are often rendered trivial by the transition to multiple choice format.
There was also a significant technology issue here, in that the text-message-based polling system I was using to do the clicker questions was often really, really slow to record responses. The students would have reasonable discussions about the questions, text in their answers, and then have a minute or more of down time to talk about miscellaneous gossip before enough results had come in for me to be comfortable moving on. On a couple of early occasions, I just went ahead with early results that mostly had the right answer, only to find that a big group of late responders had talked themselves into wrong answers, so I had to wait for all of them to come in. In the future, I think I’ll have to go with an in-class clicker system (ITS has them, but I thought the text-based system had some advantages, which were mostly cancelled out by the slow responses), to avoid those delays and keep things moving.
This experiment has made me less happy with Matter and Interactions, as well. I was drawing “clicker questions” from the really excellent collections at the University of Colorado, but those are based on a traditional text (Halliday, Resnick, and Walker, I believe), so the ordering and the presentation were somewhat different. There are also huge swathes of material that the Matter and Interactions book just wipes out. They’re so intent on keeping everything in terms of momentum (rather than acceleration) that there are hardly any problems involving multiple forces acting on systems that are moving. They have no block-on-an-incline problems at all, which removes a huge chunk of traditional intro physics, the only tension force problems involve suspended static masses (i.e., no pulleys, no falling masses on ropes), the way they talk about energy dissipation and friction doesn’t really allow stopping distance problems, and there are no quantitative problems involving one-dimensional elastic collisions.
The publisher does provide a collection of “clicker questions” to go with the book, but they mostly suck. Far too many of them require numerical calculations, which is just deadly– you would not believe how long it takes a class full of students to multiply three numbers, and there’s absolutely no discussion involved– and a good chunk of the remaining questions are too simple to be good for anything beyond checking to see that everyone’s awake. The Colorado questions are way better, but require a good deal of adaptation to work.
The handful of really pissy comments aside, I still think this method has promise, and will probably try it again next year in one of the off-sequence classes (so I don’t have to worry about the “we had to do way more work than the other sections…” problem). We have a visiting faculty member this year who was experimenting with whiteboarding, which he seemed to think was working well, so I may also look at that as a way to deal with the doing-problems-in-class issue. First, I’ll have to find out how that went over with his classes.