What does the research say about designing video lessons?
Posted on 30-04-2020
Education technology is really powerful. The problem is that it is just as easy to use that power badly as to use it well.
You can see this with video lessons – clearly video allows you to do all kinds of cool things, but how many of these cool things will help students to learn better?
The best guide I have come across is the work of Richard Mayer, in particular The Cambridge Handbook of Multimedia Learning. It is truly exhaustive – it details hundreds of studies, establishes a couple of dozen principles, and makes it clear where the research is limited or in flux, and where the boundary of each principle lies. In chapter 3 of my book Teachers vs Tech I write about quite a few of Mayer’s principles. I also say that given just how complex this research is, it’s unrealistic to expect teachers to be solely responsible for designing lessons that accord with it. And if I thought that before Covid-19, I think it even more now. Right now, the focus is rightly on getting content out there, even if the format and design are not perfect. Still, here are a few more or less effortful research-backed tips for designing video lessons.
You can always just share slides
Mayer’s ‘image principle’ (ch 14 of the Cambridge Handbook) is that adding an image of the instructor to a lesson doesn’t improve learning. Basically this is because the image of the speaker is usually not relevant to the content you are trying to teach, so it can function as a distraction.
More recent eye-tracking studies show that when a talking head speaks to a set of slides, the learners spend more time looking at the talking head than at the content on the slides – even when the talking head is referring to the content on the slides.
A number of famous online resources don’t feature the speaker’s image at all, just the lesson content and the speaker’s voice. Look at Khan Academy, for example, or Hegarty Maths – they feature the teacher’s voice talking you through a worked example that appears in stages on a slide.
Hang on a minute, you might say, what about the classroom? Students deal with a human presence there all the time. That’s true, but it is offset by the fact that in a physical classroom, it is much easier to follow a teacher’s gaze and gesture – and of course for the teacher to monitor the students’ gazes. The way in which different technologies allow you to follow the teacher’s gaze is another important area of study.
Go for a plain background
Still, it is worth pointing out that Mayer’s research doesn’t show learning is worse if it features an image of the speaker – just that it isn’t any better. If you don’t have slides to share, or you do just want students to be able to see you, then go for a plain background that minimises distraction. Bookshelves, pictures, elaborate patterns, pot plants…all these can be distracting! I think we are all realising this as we watch people being interviewed from home and spend more time wondering about their book and art choices than about what they are saying. I am certainly guilty of this – I delivered a webinar a few weeks ago with a collection of lovely yellow Wisdens in the background.
Integrate the speaker and the content
This is potentially very powerful, particularly for subjects like science, but it is also quite hard to achieve and probably needs a technique like green screen. With this approach, the speaker can walk around and interact with the slides or the content they are presenting. This allows you to take advantage of other multimedia principles: the signalling principle, for example, which is about the benefits of drawing a student’s attention to specific information. Integrating the speaker and content like this also prevents the ‘split attention effect’, which is when learners struggle to integrate information from two different sources. Barbara Oakley and Terrence Sejnowski’s online course, Learning How to Learn, provides some great examples of how this can work in practice. They have also written up their design approach in a Nature paper here.