The “flipped classroom” pedagogical model is the hot ‘new’ idea in my industry. The idea is to move didactic delivery of knowledge (especially facts) out of the classroom to venues better suited to it, like the web and books, and use live meetings of students and prof for coached use of that knowledge in discussion, exercises, and the like. New, as in, “this is how art, mechanic arts, and sports have been taught for thousands of years”. I have pretty much drunk this Kool-Aid, partly on the evidence that active learning is the universal practice everywhere the task is to acquire a skill, including very high-level skills. No-one begins teaching the piano with a reading assignment or a lecture.
Another way to “flip” the learning experience is to move testing from a big-deal end-of-the-course high-stakes assessment to frequent, low- or no-stakes events throughout the course. There’s a lot to like about this on its face, starting with the near-complete dissimilarity between what conventional exams measure and what education should be making people better at. I also dislike the infantilizing affective tone of exams that can be graded with an answer key: grownups who quiz each other as a social convention “Hey, guess what I read in the paper this morning. No, really, guess!” don’t have a lot of friends.
It turns out that low- or no-stakes testing right after we get a dollop of knowledge helps us retain the knowledge, and better than restudying or reviewing the material. This seems to be the key paper , and why haven’t I known about this for eight years? Active learning of course has a lot of this built in: after you play a passage on the piano and get some coaching on what you did, you actively ‘recall’ what you learned by playing it again, or playing something else. It seems that for enduring recall, the task of retrieving something from memory is as important to improve with practice as the act of committing to memory.
The effect size here is notable, and look at those error bars:
Remarkably, you don’t even have to actually ‘take the test’ in the sense of writing down your answers. And it seems that learning your score on these tests doesn’t matter, never mind whether they ‘count for your grade’.
I think, on the strength of this research, I have to make space in every class session for quick testing , and figure out how to design such tests for the kind of material I teach. Students are mistaken about this stuff, and incorrectly predict that they will remember material better after a week when they study it repeatedly than when they study it once and are tested on it. (Of course, students are mistaken (or misinformed) about a lot of learning technique, like highlighting (lose it).) They will grouse, but at least I have some real research to reassure them with.