Here is a brief report on a somewhat less brief meta-study that claims to show lectures are less effective than active learning approaches in STEM subjects. The report makes a claim that this is an important study. Well, kind of - it will no doubt be cited a lot. This is one of those studies that most educational researchers, including me, would very much like to be correct. Most of our theories and models suggest that lecturing is a truly dumb way to teach. Unfortunately, however, this study does not really show that. We know, and this study confirms in its data, that some lectures work really well some of the time for some students. We also know that active learning approaches do not always work better and sometimes work worse. The study makes the claim that, on average, more active trumps more passive (there are very few blacks and whites in this, it is all on a spectrum, which the study does not very effectively cater for). Sadly, however, it uses a skewed sample because (as this perceptive article highlights very nicely) active approaches typically rely on great, passionate teachers who know how to educate in order to implement them well. And guess who gets to write studies comparing active learning and lecture approaches. Even comparative studies, such as the sort that typically compare a previous run of a course with a newly minted active version, are bound to show this bias. Compare what it is like to teach a course that feels stale with one that is not only new, based on ideas you believe in (which the study, to its credit, attempts to control for) but also (as active learning approaches do) allows your own skill as a teacher to shine. This is systemically biased data.

A message I have been trying to hammer home for some time is that it ain't what you do, it's the way that you do it, that's what gets results. The softer, more malleable, more student-led a course becomes, the more that is required of the skill and passion of the teacher to be responsive and helpful, which means that such courses can be truly awesome or terrifically awful, depending on the people involved. Published research tends to come from those who are more skillful and interested in teaching because those that are not do not normally publish papers about it, so the results available to meta-studies are inevitably skewed.

Lectures are hard, rigid teaching technologies that, together with technologies like textbooks, exams, designed outcomes and timetables, are built to cater for mediocrity in teachers. Great teachers can surpass the limitations an educational system imposes, but some learning happens as a result of process design even when the teacher is truly dire. A dire teacher in an active learning context will leave students confused, lacking direction and demotivated even more than one who just lectures badly.  Furthermore, lectures can have genuine value. I have no great objections in principle to occasional lectures as long as they take up no more than a few minutes of a learner's day, and as long as lecturers start with the assumption that no direct learning will result from them. Lectures can be good catalysts, a reason to get together, and can help structure studies and thinking, even if they are almost useless for directly learning facts or skills. Just occasionally, not enough to justify their use, they can even inspire and motivate. It ain't what you do, it's the way that you do it.