Modelling learning

Left is the slightly more complex model of working memory I now want to consider. It does not include one important element of the process. That is the sensory data ALL goes into the brain and causes neuron changes. But it is good enough, I think.

Below, a simpler version of working memory, (although it has lots of arrows!)

Any model has limitation and will eventually fall down if pushed to explain some part of the system it is modelling. All models are attempts at simplifying reality. Think of a model train set that runs in my large loft. It models the timetable from Kings Cross to Hatfield almost perfectly. Great model for the timetable function of the real thing. But when a train, on my model, falls off the track I just lift it up and put it back. NO massive response from the emergency services as there would be for a derailed train in the real world. My model train layout does not even try to model this aspect of the railway.

The clever thing is to use the model with just the right amount of complexity/simplicity to be able to model the features then one needs to have to explore the system.

There are several models of working memory and the constraints and opportunities offered by these models critically affects the way we can think about learning. Cognitive load is a critical feature and one which some teachers who have certain beliefs about learning, I will not label them but if you are not using the ideas of working memory explicitly or implicitly to help you decide on how learning can happen efficiently and effectively then you are, I think, ignoring a critical feature of how we learn. I think you will get your own model of learning wrong.

I am indebted to Sue Gerrard for firstly putting up with me on twitter and then spending a little time with me, live, at ResearchEd in Birmingham to explain why the model that Willingham, and I, had been using was more limited and might then have some distortion in our understanding of learning. I now think I understand the model she prefers. A version of the more sophisticated appears above.

You might like to compare this with the simpler model that Willingham and other use. It is the role of attention, for me in my present very novice state of thinking, that is making me rethink working memory and it’s role in learning.

The features that are different from the possibly too simplistic Willingham and others model are:

  • Information from the environment into the brain is NOT limited by working memory. All the information that we collect, sense, through our senses enters our brain and changes the brain neuron structure. In one sense our brain ‘learns’ continuously.
  • We also sense internal changes in our body continuously and these changes, feelings, reactions also change the brain structure. This all goes on the background.
  • Although all our sensory inputs enters our brain is is not, by any means, accessible or retrievable. As teachers we are critically interested in learning that is retrievable.
  • Attention is the device that can select particular items from the massive stream of input from our senses.
  • It is NOT true that all the sensory input goes through working memory. This is, I think, a critical issue which is not shown by the Willingham model.
  • There are different types of memory. I have not got to grips well enough to write about this, yet. But I will.

    It is the function of this attention that is making me rethink whether the Willingham model is complex enough to allow us to think about learning (tests type) as teachers. And the different types of memory will also have an impact, though that is just a gut feeling at the moment.

Willingham has a lovely phrase, ‘Memory is the residue of thought’ which is a great phrase and evidently obvious. But it might be that the phrase needs to have some reference to attention being the trigger that then allows us to think about the sensory input. The things that can trigger that attention can be emotional, other knowledge from long term memory, novelty, threat and probably many more. Willingham’s statement is true but is, possibly, a little too reduced. We also need to be clear about how to get a learner’s attention to attend to that which we need them to learn. 
  • How to ensure we do not distract them? 
  • How do we ensure that their attention does not focus on something which is not important for the learning we want to happen?
I get slightly uneasy at this as I don’t want folk to think I am saying that we should design novel, attention grabbing activities for children as these will support their learning. My worry is that some might think I am asking for what Katie Ashford calls ‘fireworks’ lessons. Where the novelty and excitement is evident but the effect is to distract the learner.
My thoughts are that we need to think a little more about the role of attention in learning, my version of learning, than we have done. I think it has some differences to the thrust of the Willingham quote about thinking causing memories. I don’t know what they are but I am minded to be more cautious about using a simpler model as it might hide some needed complexity. Might not, though!

As a note: The word ‘learning’ seems to be used differently from the way I and some teachers would use it. For me learning is the thing that gets tested when children do an exam, or use to solve problems etc. In the brain world learning seems to be any change to the brain that is as a result of sensory input. I’ll try and be clear which version of the word I am using.

There is a choice. My attention is being pulled in two directions. Rugby or more blogging.

Rugby won! More later.

4 thoughts on “Modelling learning

  1. Since you know physics let's use a physics analogy. Newton's second law can be summarised as F=ma. Let's say Working memory is the cognitive equivalent to F=ma. How useful is F=ma in practice? On one level, very but on another useless without a lot of caveats. Even in a school lab with controlled conditions we have to stretch the imagination a bit to verify it mainly because of friction. Try firing a bullet at a target 800 metres away in a cross wind. How easy to calculate the exact trajectory using F=ma for the bullet? This is why I'm not arguing against theories of working memory, I'm arguing that the simplified theoretical model is not going to work unmodified in all contexts. No doubt it will in some especially if it is replacing some completely erroneous theory such as F=M^2a or something. However the allying of the science to some political dogma over traditionalists vs progressives is more likely to get the science rejected than it is to get it adopted. It seems that few people who do not already believe it are suddenly seeing the light. This in itself shows your “attention” factor is significant. Why have al the teachers suddenly not seen the light and the logic and come on board? It's because there is more to learning than just exposing people's memory to “the facts”.Ironically DT Willingham seems to understand this rather better than many of his devoted followers. If you read his work, it is cautious and measured and it talks about things being difficult to do not impossible. He doesn't demonise and alienate the very people he wants to encourage to pay attention. If we extend your attention parameter to include the influential modulators of working memory you might well be on to something.

  2. I'm arguing that the limitations of models should be very clearly stated when they are used in real world contexts. That is just basic good science. Some simple models are useful but simplistic inferences and conclusions from them are if anything counter-productive. eg this model is likely to work well for a space ship in a vacuum, not at all well for a submarine at sea.

  3. I think you’ve made some really good points, Peter. Especially about simplifying but not oversimplifying and about constraints and opportunities.

    One point that keeps coming up about factual information is that teachers without a background in the natural sciences seem to be very focussed on the importance of ‘concrete facts’ – the more the better, it seems. Teachers with a background in the natural sciences (like commenter ’42’ for example) tend to focus on the information that can be abstracted from the concrete facts.

    Often what can be abstracted from a mass of confusing concrete facts (data) is a handful of general principles (information or knowledge) that don’t overtax the capacity of working memory and that cover 80% or more of the cases you’re likely to meet. Here’s a diagrammatic representation of what I mean

    Brings us back to your point about simplifying but not oversimplifying. Often four or five rules of thumb are enough to guide and appropriately constrain professional practice on a day-to-day basis.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s