Landing : Athabascau University

What really impacts the use of active learning in undergraduate STEM education? Results from a national survey of chemistry, mathematics, and physics instructors

This is a report on an interesting study by Naneh Apkarian et al, that asked a large-ish number (3796) of in-person American STEM profs (college and university levels) about the effects of various known factors on their use of active learning approaches. To a large extent it seems that 'active learning' is mainly taken to mean 'not lectures' (which is both unfair to a minority of lectures and over-kind to a majority of alternative teaching methods). Photo of a lecture (credit to Sam Balye) It's a good paper but the study itself has some gaping flaws (there are many chicken-and-egg issues here, lots of confounding factors, massive fuzziness, loads of systemic biases, and great complexity hidden in the details), which are, in fairness, very well recognized by the authors. Wisely, they largely avoid making causal connections and, when they do, they use other evidence beyond that of their findings to support them. Flaws aside, it's a good contribution to our collective story, and a thoroughly interesting read. This is what they found:

1) Though active and inactive(TM) learning approaches are used across the board, lectures are far more likely to be used when class sizes are large (notably so at 60+ class sizes, predominantly so at 100+ class sizes). Depressing, but not surprising: big class sizes massively exaggerate the dominant role of the teacher, and controlling teachers faced with the scary prospect consequently tend focus on what they want to indoctrinate rather than what students need to do. It doesn't have to be that way, but it's how lecturing began in the first place, so it has a bit of a history.

2) If you schedule classes in lecture theatres, most people use them for lecturing. This could  be seen as useful supporting evidence for my own coparticipation model, which predicts this on theoretical grounds (large and slow technologies influence smaller and faster ones more than vice versa, defaults harden). However, it actually shows no causal relationship at all. In fact, the reasons are likely much more mundane. From my dim recollections of in-person teaching, if the course design involves lectures then you get classes scheduled into lecture theatres. If you are stuck with a lecture theatre because of dimwitted/thoughtless timetablers but want to do something different then you have a (fun and challenging) problem, but that's not what the results here tell us.

3) There's a small correlation between how teachers are evaluated/the perceived importance of teaching in those evaluations, and how they teach. Those who perceive teaching to be less valued tend to lecture more. This doesn't seem very useful information to me, without a lot more information about the culture and norms of the institutions, relative weightings for research or service, and so on. Even then, it would be hard to find any causal relationships. It might show that teachers who don't like or have time for teaching tend to lecture because it is the easiest thing for them to do, but I'd need more evidence to prove that. It might show that extrinsic motivation drives compliance (a little), but, again, it's not even close to proven. Much more context needed.

4) Perceived job security has no obvious effects on teaching practice. This might be seen as a little surprising as there is a fairly widespread perception that people give up on doing good things when they get tenure, but it doesn't surprise me, given the multiple factors that affect it. Whether active or not, you can always teach badly or well. The implied assumption that active approaches are riskier and more experimental is not actually true much of the time, and there's nothing in the survey that draws out whether people are taking risks or not anyway. Most teachers continue to teach in ways that seemed to work before, and tenure makes little or no difference to that.

5a) Very active researchers tend to lecture quite a bit more than quite inactive researchers. Indeed. See 3 - if you are a researcher but not engaged in the scholarship of learning and teaching then you probably have less interest and/or time to spend on teaching well, not to mention the fact that many universities compete to get the best researchers and couldn't care less whether they can teach or not. There is a happier corollary...

5b) those who engage in educational research of any form lecture a lot less. This speaks to common sense, to what educational research has consistently shown for about 100 years, and to the dominant educational doctrine that lectures are bad. Personally, I kind-of agree with that doctrine, but I think the problem is much subtler than simply that lectures are bad per se - lectures can play a useful role as long as you don't ever try to use them to impart information, as long as you always remember the rest of the learning assembly into which they fit (and in which most of the learning happens), and as long as you never, ever, ever, whether implicitly or explicitly, mandate attendance. The fact that most institutional lectures fail on all three counts, and virtually all falter on at least the most important two, does indeed make them very bad, but it's not inherent in the technology. Tain't what you do, it's the way that you do it.

6) People who have experienced active learning as learners are far more likely to use such approaches. Well, yes. It would be quite a surprise if, having discovered there are better ways to learn that are more satisfying and effective for all concerned, people did not then use them.

None of this is novel, all of it reconfirms (but doesn't prove) what we already know, especially in the hard disciplinary areas of STEM. However, it will still be a useful paper to lend support to other research, or when thinking about what needs to change if institutions are trying an intervention.  I expect that I will cite it some time.

I'm more interested, though, in what lessons might be drawn for online teaching, especially in an institution like Athabasca University, where teaching is explicitly distributed, where roles in that distributed assembly are well defined and, too often, mutually exclusive, and where lecturing is almost unheard of. 

Inactive online learning

For AU courses, I think the nearest equivalent to a lecture is a heavily content-oriented course (typically greatly reliant on a textbook) with over-controlling, easily-marked assignments, and a proctored exam at the end. That's the 'don't think about it' too common default. It's not quite that simple, because the involvement of experienced and well-educated learning designers, editors, and media experts tends to make the content quite well written and at least somewhat informed by theory. Also, compassionate tutors can fill in a lot of gaps: good tutoring is often the saving grace of an otherwise yawn-inducing pedagogical model. It's efficient and well-honed, like the lecture, and it works most of the time because our students are wonderful and do much of the teaching themselves (despite  attempts to control them), but it's not a great way to teach anyone. Better than lecturing, for sure, but it has to be because there's not so much of the other stuff that teaches in in-person institutions. We do of course have a great many courses that do not follow this pattern, that involve far more active learning: it's far from ubiquitous, even in STEM teaching.

I think that part of the reason for a preponderance of inactive approaches at AU can be found in the paper's second finding. In our case, an LMS is the functional equivalent of a lecture theatre (with a similar emphasis on teacher control, structure, and content), especially as our self-paced model limits the options for using its already impoverished social features. There's also a lot of rigidity in our course development processes, with a laser-sharp focus on measurable outcomes or, worse, clearly defined objectives, that tends to make things more content-driven. Perhaps a bigger part of the reason, though, relates more closely to finding 6. It's not that our teachers aren't engaged and interested in producing good stuff: they really are. It's more that they don't have a great many role models and examples to call on. This is compounded by:

  • again, the stupidity of LMS design (courses are enclosed and hidden, for the most part),
  • a lack of sharing of tacit knowledge between teachers (we tend to only meet and communicate with a defined purpose, leaving little time for incidental and passing exchanges), and
  • our contact with students tends to be similarly instrumental and formal so we don't usually learn as much about how they feel about other courses as in-person teachers.

All in all, though it does happen, and we are constantly getting better at it, good ideas still do not spread easily enough. In fairness, that's also true of many in-person institutions, but at least they have serendipity, greater visibility of teaching, and simpler ways to connect socially for free, because physics. We have to actively design our own social physics, and the results of doing are seldom particularly great. As we move towards become a near-virtual institution (or even nearer-virtual) we are really going to have to work much harder on that.

On the bright side, we are fortunate to have a vast number of faculty (around 40%) who fall into the 5b category. If only we could do a better job of sharing their learning. That, of course, is a lot of the reason I am writing this, and it was a big impetus behind why we created Athabasca Landing in the first place.