It’s been pleasing to see that, thanks in part to the work of a Scottish born paediatric surgeon based in Sheffield, more and more people are using images as opposed to text in their slide-sets when presenting. The message seems to be getting through that this is the way to go (at least in EM and critical care presentations) so I thought I would look at the educational theory that underpins this mini revolution.
[for more from Ross (@ffolliet) visit his P Cubed website here]
Unfortunately, educational practice does not, for the most part, rely on research findings. Instead, we tend to rely on our intuitions about how to teach and learn – with detrimental consequences.Yana Weinstein: Understanding How We Learn: A Visual Guide
I am not a big educational theorist but I think it’s important we try to understand the basics about why something works as this may help us to realise what won’t work.
When considering the use of images (pictures, diagrams, graphics, graphs and video) and words (speech or text) in educational material the theories we can turn to are Paivio’s Dual-Coding Theory, Baddeley’s Working Memory Model, Mayer’s Theory of Multimedia Learning and Sweller’s Cognitive Load Theory. Ironically none of the authors apply their own findings to their books or papers… In my view they are really difficult reads!
The details of the differences between the theories is beyond the scope of this blog but if you are interested in knowing more click here for a paper that gives a good overview . For the purpose of simplicity I am going to bundle them together and call them all dual-coding, which is the name most people are familiar with.
A simple memory model
Before looking at how dual-coding works it is probably worth reviewing a simple memory model again.
Whatever information comes to our attention is encoded in the working memory by creating schema’s with previously held knowledge from our long-term memory. The amount of encoding that can take place to generate learning is limited by the intrinsic cognitive load (the complexity of the information coming in) and the extraneous cognitive load (the complexity of the method of instruction – we aim to keep this low). The higher either one of those is, the less space for encoding and learning. We seem to be able to cope with no more than four distinct pieces of information in our working memory at one time. Anything doesn’t come to our attention, or is not encoded, will be lost.
Encoded information is then passed to our long-term memory where it is consolidated for storage. Poorly encoded information, or information that its not used, is likely to be forgotten with time. In order to apply learning we need to be able to recall the information from long-term memory when needed; this brings it back to our attention and the cycle begins again.
According to the theories, dual-coding improves all parts of this process; it helps attention; it reduces cognitive load; it improves encoding, allows for better storage and aids recall.
Working memory capacity can be effectively increased, and learning improved, by using a dual mode presentationJohn Sweller: Cognitive Load Theory
In simplified terms it works like this:
Imagery and speech are processed by separate channels in the brain; the visual channel for what we see and the auditory channel for what we hear. Two distinct mental models are then created, spreading the cognitive load in working memory.
The process is slightly different when dealing with imagery and text because all the information initially enters via the visual channel. The images are then processed to create the visual model whilst the text is diverted off to the auditory channel as a sort of internal dialogue to create the auditory model.
As long as the image matches up with the words, this gives the learner twice the chance of recalling the same information. Either the verbal model or the visual model can be recalled, triggering the other.
Another advantage of using diagrams with words is that the brain doesn’t have to use valuable cognitive resources creating its own mental model – it’s there already.
Problems occur if an image and the words don’t relate to each other. A verbal model will be created but it will be difficult to produce a visual model. You may be able to recall the image from memory but not be able to connect it with the verbal model so less learning occurs. This can be seen when educators (with all the right intentions) use a slightly obscure picture in a presentation, because they feel they must use one, but couldn’t find quite the right one for their message. This leaves the learner looking at the image going “why have they used that picture… what’s the connection?”. If a learner is doing that then they are not attending, encoding and processing information.
Too many images on a slide or page, or complex imagery all increase extraneous cognitive load as the working memory tries to process those images. This uses up valuable cognitive capacity leaving less space to generate learning.
If imagery, speech and text are all used together this also increases the cognitive load and leads to less learning; Either the text or the speech is redundant because they convey the same message. Both will need to be processed by the auditory channel adding to the extraneous cognitive load .
Cramming text and diagrams on to a page increase cognitive load and reduces any effects of dual-coding. Whilst it is a good idea to try to convey a message on a single side of A4, reducing the font size and line spacing in order to do this is counterproductive.
What does this mean in practice?
Dual-coding is one of the six strategies for learning (along with retrieval practice, spacing, interleaving, elaboration and concrete examples) recommended by cognitive psychologists .
Use photographs as an aide-memoire to make connection and improve recall. This is typical of the stuff we see in good presentations; an image with minimal text to associate with the words the presenter is saying. We may not be able to recall the exact words, but we can remember the image and the basics of the message as long as the image is relevant.
Avoid using an picture for the sake of using an picture. If it’s not directly relevant to the message then the connection will be lost; learners may be able to recall the image but what was the message?
If we were talking about the number of alcohol related admission to the ED we could use a picture of a beer glass and a percentage – a clear connection, good for memory. However, if we were teaching about how alcohol is metabolised, using the same image would be much less effective. Explaining how ADH & ALDH metabolise alcohol into water and carbon dioxide will not be helped by having a photograph of a glass of beer – it will help with attention but will not improve learning. What would improve learning would be a clear diagram of the process, highlighted at the relevant parts to match the words.
An image is worth 1000 words so use diagrams to explain a concepts in visual form in a concrete way. Keep the material clear, use few colours, and lots of ‘white space’ with only necessary text for understanding to keep the extraneous cognitive load low. Any explanatory text should be as close to the image as possible so that the learner doesn’t have keep moving between text and diagram. An example would be the diagrams next to the description of dual-coding above.
To give you an idea of what is possible with educational material, in @MFT_UGME we have recently reworked one of our iBooks used to teach MSK examinations from text heavy description to multiple step-by-step simple images with text light descriptions (along with pop-ups for more information). Not only is it easier for students to follow, it’s easier to remember the images and is easier on the eye.
[If you are interested in seeing it in full, it can be downloaded for free from the Books section in St Emlyns].
Photographs can also be used to grab attention. Here at St. Emlyns we use it for every blog because people are more inclined to read an article if there is a image associate with it. The key is not to have so many that they create extraneous load.
Finally (this one isn’t directly related to learning) I feel we should be making more use of graphics in emergency instructions.
If you look at this sign from the building site you can make out the key areas and messages without having to search through all the text to find the bit you are looking for. High levels of stress affects our ability to process information and words swim before our eyes as we try to read quickly, searching for the relevant bit. Graphics allow us to scan information quickly, looking for cues; they are much easier to process. It may also help if English isn’t a persons first language. If you think of Lego or Ikea instructions, they use no text at all but it’s still easy to see what each step is.
[We have over 100 graphics for medical related instructional material that you are welcome to download for any non-commercial use here]
We can use dual-coding when we are learning by making our own simple drawings . This works because in order to produce the drawing we have to understand the information we are learning; it’s a way of summarising, and in order to summarise we need to have understood the subject matter. It can also help provide a concrete example of an abstract theory which is good for encoding and recall. If you don’t think you can’t draw it doesn’t matter; as long as you know what it represents thats fine. Our brains aren’t so picky as to reject an image because it’s a ‘bad’ representation.
There is no one size fits all situations when it comes to dual-coding; we come up against boundary conditions, situations where it fails to improve learning, usually for those people at the ‘higher’ end of the learning spectrum.
We need to balance the need to gain attention using visuals for decoration against the increased cognitive load they cause; we need to use the right visuals, at the right time, to make use of the dual-coding effect.
But in the most part dual-coding grabs people’s attention, reduces cognitive load and produces stronger encoding allowing for easier recall later if we use it effectively – what’s not to like.
- Clark JM, Paivio A. Dual coding theory and education. Educational psychology review. 1991 Sep 1;3(3):149-210.
- Baddeley A. Working memory, thought, and action. OUP Oxford; 2007 Mar 15.
- Mayer, R. Multimedia learning. Camberidge: Cambridge University Press; 2009.
- Sweller, J., Ayres, P. and Kalyuga, S. Cognitive Load Theory. New York, NY: Springer New York. 2011
- Reed SK. Cognitive architectures for multimedia learning. Educational psychologist. 2006 Jun 1;41(2):87-98.
- Mayer, R. Multimedia learning (2nd Ed). Camberidge: Cambridge University Press; 2009. p118
- Weinstein Y, Sumeracki M, Caviglioli O. Understanding how we learn: A visual guide. Oxford: Routledge; 2019. p83
Special thanks to @OliCav who has helped me with some of the finer points of dual-coding recently. In a lot of cases I have simplified peoples works so any errors are mine not his. You can find lots of information and free infographics on cognitive load and dual-coding at his amazing website www.olicav.com and he has a book coming out in May 2019 called Dual Coding with Teachers. Also thanks to @allibobs83 who corrected my dreadful spelling, grammar and punctuation!