SkepticblogSkepticblog logo banner

top navigation:

The Science of Learning

by Steven Novella, Oct 14 2013

As a perpetual student and frequent teacher, I am very interested in the science and technology of learning itself. In medical school, for example, we have to transfer an incredible amount of information to students over a relatively short period  of time. The trick is to get students to maintain their attention, focus on the important bits of information, and understand and remember that information.

This is very challenging. Simply lecturing is not very effective. There are different figures as to what the average “attention span” is, depending on exactly how it is measured, but the figures are generally less than 20 minutes, and as low as 5 minutes. Of course, this depends on attention to what. I can pay attention to a 3 hour movie without difficulty, if it’s Lord of the Rings or similar quality. Try listening to a 3 hour lecture on a dry technical topic, and effectively process the information presented the whole time.

So how do we get students to spend the day in lectures and actually learn a significant portion of the material?

There are many answers to this question, but a recent study perhaps adds another technique to improve lecture efficiency. Robert Collins conducted a multi-year study in which he used video during his lectures, but for the first year the video had captions turned off, and for the second year he had captions turned on. He found that during the second year class discussion improved, and grades significantly improved.

While interesting, this should be considered a preliminary study. It was not well controlled, which means the simple introduction of a change – any change – might be responsible for the effect observed, rather than the specific variable of captions. Better controlled follow up studies are needed to see what effect the caption variable has specifically.

But if we take the results at face value, they are plausible and largely consistent with other research. One way to look at it is this – what demands are we making on the cognitive resources of students during a lecture? It takes mental energy to pay attention, to know where to focus your attention, and to process the information being presented. You can really only do one thing at a time, so while you are thinking about the information just given, you may miss the next bit of information. Further, while you are trying to make sense of a graphic or animation in a video, you may not fully process what the narrator is saying.

A good lecture, like a good movie, does not just impart information, it manages the experience of the student. That means drawing their attention to the important bits, without making them waste mental energy finding the important bits on their own. You want all of their mental energy going to perceiving, understanding, and remembering the critical information – not trying to find it or separate it from the less-than-critical information.

Captions to videos is one small technique, but can make it easier for students to get the information. Sometimes auditory processing can be resource intensive – that’s why you should not talk on the phone while driving. Adding captions might make it easier to understand what is being said, especially if new and unfamiliar terms are being used. Even seeing how a word is spelled can enhance comprehension.

I have found this myself, while watching science programs discussing species names, or chemical names, or other technical information, I find myself wanting to see the words in writing. Just hearing them is very unsatisfactory.  Seeing the name reinforces what I just heard, and adds new information (word roots, etc.) that enhance the information. Good science documentaries also know when to have graphics that grab your attention and present the critical bits of information in an easy to read manner or an infographic-type display that enhances comprehension.


Effectively transferring large amounts of complex information is extremely challenging, but it is a science and a skill worth developing. I do think that a good overall approach is to think about managing the experience of the student, rather than just imparting information.

I do think there is a good analogy to be made with movies, something with which most of us have experience. What is the difference between an awesome movie, and average movie, and a terrible B-move? Well, lots of things, but perhaps most important is the difference between just telling a story, and giving the audience a cinematic experience.

Brilliant directors manage that experience in tiny detail – to where you are looking on the screen, to how expectations are created and then fulfilled, giving the audience time to absorb one development before moving onto the next, and a hundred other ways that you are likely not consciously aware of but which dramatically affect your experience of the movie.

A good lecture is similar. There are many variables to consider, and no one magic solution to making a great lecture. I do think that knowing when to caption audio is an important detail, and this new study (while preliminary) supports that.

One Response to “The Science of Learning”

  1. Max says:

    The basic problem is the Curse of Knowledge, the false assumption that others know what you know. In an experiment, subjects who tapped the rhythm of a well-known tune estimated that 50% of listeners would identify the tune, when in reality only 2.5% of listeners could identify it.

    I see it all the time in presentations. Even colleagues in your field may not be used to your choice of terms or units. Like if you use Metric units and they use British units, they have to convert everything in real-time.

    That’s why a picture is worth a thousand words. If you catch yourself translating some mental image like a graph into words, then draw the damn picture, because it’s even harder for listeners to translate your words back into an image in real-time.

    Drawing and writing as you’re speaking helps to focus attention, and if you use viewgraphs, then point to the parts you’re talking about so viewers don’t have to guess. The worst slides have just text that’s different from what the speaker is saying.

    Captions are helpful for foreigners, but reading is also resource intensive – that’s why you should not read text messages on the phone while driving.
    I’ve tried to listen to text-to-speech (TTS) audio, and found the flat, robotic delivery very difficult to follow. At least when I’m reading text, it’s in my own voice. Fortunately, newer TTS voices are a lot less robotic than the old ones.