Questions on Desirable Difficulties
Jonathan Firth's Memory & Metacognition Updates #101
Hello, and happy Lunar New Year to you – I hope the Year of the Snake is a fortunate one for you! 🎉 🐍
A couple of weeks ago, in my long post on desirable difficulties (see update #99), I mentioned a few questions on the concept that I wanted to return to. That’s what I’ll focus on today and next week, beginning with two questions that I’ve often been asked by teachers – exactly how difficult should tasks be, and is there a risk of discouraging students if the work feels too hard?
Level of difficulty
As an educator, we can’t avoid difficulty, but there is a question of how much. As with spices when cooking, there may be a balance that needs to be struck.
Bjork (1994, see below) makes it clear that the exact mix of difficulties in any given learning situation cannot be prescribed by researchers, as it will depend on the demands of that situation. However, he also goes on to say (p. 189) that whatever the mix, it will be a lot more than is currently used for most study and practice. In other words, desirable difficulties are under-used.
Still, there is a degree of judgement here, and this will depend on your knowledge of a specific group of students. Bjork & Bjork (2011) note that if “the learner does not have the background knowledge or skills to respond to [the difficulties] successfully, they become undesirable difficulties.”
Let’s think, for example, about how long the spacing (between two practice sessions) should be when we use the spacing effect.
The answer, unfortunately, is that it depends… on the students and on the material! The research can tell us that it’s best to wait until the re-study becomes challenging but not impossible. But the exact time delay will be longer for easier material and concepts, and shorter with struggling students. All of this means that a degree of teacher judgement is unavoidable (Firth, 2021).
With other difficulties, too, we might use our judgement and our knowledge of the students and of curriculum materials to gauge how much difficulty is too much.
Will this feel too hard?
A related issue that a secondary teacher recently asked me about is whether using desirable difficulties might cause their students to feel discouraged. They were concerned that applying the techniques could lead to students avoiding their subject and choose ones that feel easier.
Of course, there are times where we want to avoid classwork feeling too difficult. We don’t want students to feel overwhelmed, confused, or stressed.
At the same time, difficulty needn’t feel unpleasant. Many of us really enjoy the challenge of a workout, a hike or a stimulating lecture because we know we are being stretched. Easy work is dull in comparison!
I also think that when we connect successful learning techniques to metacognition, helping learners to understand what they are doing and why, it will really help.
Pooja Agarwal gave a good example in this blog post. She describes using low-stakes quizzes at the start of a new term to help start a conversation with learners. As she neatly explains, raising the issue of how much they have forgotten from their past learning can be shocking and embarrassing. This is a great motivation for them to start thinking about using more effective techniques.
Personally, I think that when learners see that they are making good progress and understand why the techniques work, they’ll be more likely to choose your class. I’d love to hear other perspectives or experiences on this issue, though.
Recommended study
If you want to read more about desirable difficulties, the chapter below by Bjork (1994) was the first to set out the argument in full. It had a massive impact on me when I first read it, and I have enjoyed revisiting it, too:
Bjork, R. A. (1994). Memory and metamemory considerations in the training of human beings. In J. Metcalfe & A. Shimamura, (Eds.), Metacognition: Knowing about knowing (pp. 185–205). MIT Press.
Thanks for reading – I hope that you enjoyed the challenge!
Jonathan
I hope you are enjoying Memory and Metacognition Updates. Subscribe for free to receive new posts and support my work – and consider sharing it with colleagues!! Thank you!
Please note that my slides and similar materials are shared under a CC BY-NC 4.0 license. This means you can use or adapt them with attribution for non-commercial purposes. If you wish to use my materials for other purposes, feel free to get in touch.
I believe predictive processing theory - the latest thinking in cognitive and neuroscience on how our cognition works - provides an immensely simplifying underlying theory and heuristic for managing desirable difficulty. Predictive processing theory posits that our brain creates a prediction of what our senses will experience next form a generative model stored in long term memory in the context of what has just happened. We compare this prediction to our actual sensory input and generate an error if they don't match. if they do match then our prediction is correct and the generative model used was successful and it is reinforced in our LTM. If there is an error then this is used to correct our generative model being used. Successful adjustments that minimise future error are then the signal to update the generative model in LTM ( i.e. learn). If errors are low ( low difficulty) then learning is low, if errors are too high we get overloaded and loose information, if errors are just right( desirable difficulty) we learn best. So desirable difficulty is about having the right level of predictability and surprise. Each individual will have varying past knowledge/experience/ generative models affecting how predictable things are for them. Each situation and context will also vary how predictable things are. Extraneous load means more things to predict - familiar things create less prediction error, unfamiliar create more contributing to the overall prediction error rate. As educators we want to get this overall predication error rate into the "just right" range high enough to be generating learning, but not so high our prediction error bandwidth is overloaded. Shifting the framing from "difficulty" to "predictability" allows for a clearer understanding and management task to get into the "just right" range. See my blog: https://predictablycorrect.substack.com/ - Adam W