 There are now ample statistics to show that flip classroom methods that for their self-paced materials use notes and textbooks can already provide superior grades for the top-end students, for the mid-tier students, and for the low-tier students. They also provide better passing rates for minorities. They also provide better attendance, better student responses, better conceptual development, better improvement of knowledge across a course, and better grades in subsequent courses. In other words, flip classrooms have already been pretty well tested to work well. Now because a lot of that research has been done on courses where the self-paced materials are based on textbooks and provided notes, it's clear that a lot of those benefits are based on what you could do within class activities when you don't have to introduce new content but just help students engage with it. However, we're still going to take a few minutes to look at what's known about making that self-paced learning efficient and effective. First of all, it turns out that we need the textbooks and the provided notes. Written material is typically the best way to present the most detail, and it's also the fastest way to skim to find a particular discussion of a specific topic, and it's often preferred by non-native speakers or students with other accessibility issues, and so it's essential to have these resources available for a course. On the downside, reading a textbook can often seem very impersonal, and that ultimately leads to a loss of student engagement. There are some things that are just naturally more engaging live, like inspirational discussion or demonstrations or movies and dynamic models, that kind of thing. One final problem with just using textbooks and provided notes is that it's very hard to embody uncodified knowledge in them, so it's much easier to model problem solving and your general approach to things live than it is in written form, and so if we want to do all these things live, then in a self-paced learning setting what that really means is we want to do it using videos, and of course the absolute cheapest way to do something like that is to simply record your current lectures or classes. And if you've ever actually given this a go, if you've ever recorded yourself giving a lecture or a class, and then watched it again afterwards, then you might understand why we do not recommend this method. Watching or listening to those recordings afterwards is typically very torturous, and not just because you see all those minor mistakes that you made, because while they might bother you, they won't bother the students, but mainly because it's just incredibly slow. The information density is really, really low and it drags on and on. And there's also typically a lot of repetition and filler precisely because you're trying to hold the attention of a range of people for a long time. So in other words, when you're giving live lectures or classes, you're optimizing your delivery for constraints that you don't have when you're actually delivering those videos online. In a real live lecture, students can't pause you and go get a snack or answer the phone. In a real live lecture, students have to turn up whenever the lecture is. They can't turn up when they're feeling awake and ready to learn something. In a real live lecture, if people are finding that it's going too slow, they can't just increase the speed of delivery. And in a real live lecture, if students are finding it really difficult or if they miss a point, they can't pause it, go back and watch it again and work through it and get to the point where they actually understood what you were saying in that previous section before listening to what you say after that. And all of these things mean that students interact with online videos very differently into the way they interact with lectures and classes, and we have to design our delivery of those videos slightly differently for that reason. So of course, the first key point is that the information density has to be very high. Don't leave large gaps. Try to resist the temptation to repeat things too much, and if you have to write things down or do a demonstration, then cut out all the craft and make it happen as fast as possible. The advantages you get from this is that the final result is enormously more rewatchable. When you watch it the first time going slow doesn't hurt too much, but if you're going to rewatch a section of some video, then having high information density is critical to have that experience be tolerable. And the second key point you have to take into account when you're making specially made video content for your self-paced learning is that you need to have regular breaks. You need to have chances for the student after they've watched something to respond to some question. And so what you do is you have regularly spaced short videos interspersed with questions. And of course, regularly asking students questions is standard best practice for live lectures and classes as well. But in this case, you're actually getting to ask every single student a question every single time. So between every video and the next video, you get every student to respond and make sure that they're really following what you're up to. Now the last few years have really seen the advent of MOOCs, the massive open online courses that deliver all of their content online and they have no in-class activities whatsoever. Now of course, because they're hamstrung that way, they've had to do a lot of research to find out exactly how to make this work. But one of their advantages is they have literally hundreds of thousands of students that they can do randomly controlled A-B testing on. And so they can tell what kinds of videos produce superior engagement. And the answer that is the videos ought to be between four and nine minutes long with a sweet spot around about five to six minutes. And that's all you absolutely need. All you really need is high information density, short videos with questions in between. If you want to go for all the points, then you should have some talking head, preferably in a collegial environment, and some talking over slides, but not at the same time. So not a little box with your head talking while you write over a slide, but one or the other so that they don't distract from each other. And finally, the talking over slides ideally should be handwritten, although of course there's a very small effect there. And so you should follow your own style, whatever feels most appropriate to you and your material.