There’s a lot of talk about memory in education. Until fairly recently, no one really seemed to be talking about schemas in relation to teaching; now they’re everywhere! The only time I encountered them formally was when looking at Piaget as part of my training. As a Psychology teacher they’re very familiar to me, although it’s taken me a remarkably long time to meaningfully apply this knowledge and understanding to my own teaching (see here for a similar discussion about working memory).
Understanding the role schemas play in learning is powerful, possibly one of the most fundamentally important things we need to know about as teachers. Having strong schemas allows students to free up working memory capacity, complete tasks more effectively due to having some elements automated, interpret and learn new information more effectively. They allow us to make predictions about the world, and to be able to adapt to new situations or deal with information in novel contexts without getting flustered (cognitive overload). But there’s a flipside too. And one that barely seems to get a mention in all the articles, books, blogs and discussions around student memory. We’re all familiar with the idea of forgetting (and there are a number of different explanations for it) but I want to focus specifically here on those times when we don’t completely forget something, but still don’t remember it correctly. How do our schemas cause us to misremember?
Frederick Bartlett (1932) proposed the theory of reconstructive memory. He argued that memory cannot be seen as a passive process of simply recording, objectively, information from the environment. The notion that when we learn something we simply file it away in a mental box, to be able to retrieve it from there, in its original form, is wrong. Instead, memory is an active process, and a subjective one.
Schemas can be considered to play two important roles in this theory. Firstly in how we encode information based on what’s in our schemas already, and secondly in how schemas shape the way we retrieve information.
When we perceive new information, our immediate interpretation of it is shaped by our prior knowledge (van Kesteren, 2012). In some cases it may be that a particular set of prior knowledge is cued which leads us to a particular interpretation. For example, in the image below, consider the central character/figure:
Whether we interpret it as the number thirteen or a capital letter B depends on what schema has already been activated – most likely here it would depend whether you noticed the other numbers or letters first.
Bartlett suggested that we make effort after meaning; we intentionally (though not necessarily knowingly) try to link new information with something we already know in order to make sense of it. For example when someone teaches us a new concept we use analogies to say “Oh that’s like…” in order to help make it feel more familiar and less ambiguous.
So right from the outset, we aren’t ‘seeing’ the world objectively, but through the lens of our schemas. The process of encoding is an active one, driven by the content of our schemas (see a more detailed and neuroscientific explanation of this here). Failure to recall some information correctly might not be a failure of retrieval, but because it was never encoded ‘correctly’ in the first place.
When we recall information we aren’t simply reproducing information from memory. Schemas don’t house neat parcels of information (although that is often how they are portrayed in psychology textbooks) but they are formed by a network of connected areas of activation in memory. When you think about something from memory you are activating a range of connected pieces of information which are drawn together to create the experience of remembering something. Crucially, the act of reconstruction happens every time you recall that set of information, and memories retrieved after consolidation may have a different neural architecture to when they were encoded (Dudai et al, 2015).
Most people will be familiar with the idea of Chinese whispers; Bartlett famously showed this process in a series of experiments involving story-telling. British participants were shown a story from Native American culture – the War of the Ghosts – and then subsequently asked to retell the story to another participant. This happened via serial reproduction – A tells B, B tells C, C tells D and so on. Bartlett found that each time the story was retold it underwent significant reconstruction. Importantly, these changes were not random, but details were altered to make the stories more meaningful to the participants, to fit better with their own experiences. For example the word canoe was changed to boat because that was a more culturally common term for the participants. In other words, their schemas interfered in the reconstruction, changing information to fit better with their existing mental model of the world.
The effects of schemas and reconstructive memory has been investigated in a range of contexts, and a huge body of research relating to the reliability (or not) of eyewitness testimony has found that memories can be influenced by a range of different factors. Post-event information is information received after encoding of the original information. For example watching a football match live, and then later watching the highlights on TV. The two sources of information combine and integrate to produce one coherent memory, and we are not able to distinguish between different sources. When it comes to retrieval we may believe we are simply recalling what we saw happen for ourselves, but are actually including information we have received at another time. So we may recall something that happened in the football match and believe we saw it happen live, when in reality we only saw it during the analysis on TV.
Sources of post-event information might include post-event discussion where we talk about what we’ve experienced with someone else before being asked to recall it. Gabbert et al (2003) found that participants who watched a video and were then allowed to discuss it with someone who’d watched a video showing the same events but from a different perspective, reported details that they couldn’t possibly have seen. Loftus & Zanni (1975) showed that simply changing the wording of a question from “did you see a broken headlight?” to “did you see the broken headlight?” was enough to cause changes in participants’ recall. Loftus & Pickrell (1995) showed that you can deliberately plant false memories, for example people recounting a vivid memory about being lost as a child in a shopping mall even though it never actually happened.
All of which paints a clear, if rather pessimistic view. Memories are fallible, and it may seem like something of a miracle that our students actually remember anything correctly at all! But understanding how memory works is key if we are to help our students learn better.
What are some of the implications of this for our teaching?
We need to carefully consider the schemas our students hold before we teach them something new. If they don’t have the relevant knowledge, their knowledge holds misconceptions, or we simply fail to activate it appropriately, then students will likely struggle to encode the new information correctly, or will add faulty knowledge from their existing schema when they try to retrieve the information.
We need to consider that when students make mistakes, misremembering key information, or making overly broad generalisations, it’s not entirely their fault! They are working with (or against) a memory system that has evolved to interpret new information in light of what has already been learned and this means students might not necessarily discern key differences in apparently similar concepts or sets of ideas. As Sarah Cottingham discusses here, highlighting this ‘prediction error’ when we give corrective feedback helps students to update their schemas with correct information.
We should be mindful of sources of post-event information that could change what students remember. A common example is the post-exam ‘debrief’ where students describe what was on the paper and what they wrote. But this is usually preceded by a more immediate debrief with their peers as they leave the exam hall. Trying to prevent this is about as effective as trying to nail jelly to a wall but we should, at the least, be aware that what they report to teachers, and later what they tell themselves about how they answered, may well be quite removed from what they actually wrote. The same may apply to any account either we or a student gives about something that happens at school. By the time someone has spoken to other students, colleagues or parents about it, it’s entirely possible that recall will have significantly changed, without any intentional dishonesty.
A final thought is whether we can harness this for our own purposes. With careful questioning and a bit of artistic licence, we have the power to deliberately change what students might have previously remembered by adding our own post-event information. “Do you remember when we looked at… and we agreed that…” may plant an idea in their minds even if it’s not what actually happened in the original lesson. Asking students “why did we come to that conclusion?” takes advantage of choice blindness (eg Johansson et al 2013) which means that we like to justify and rationalise our decisions, even if they were not the decisions we actually made in the first place! Of course, there is an ethical question here, to what extent is it acceptable for us to wilfully change our students’ memories in this way? As teachers one could argue that that is essentially our job anyway, but perhaps some might consider this crosses a line. I’ll leave you to decide that for yourself.
Bartlett (1932). Remembering: A Study in Experimental and Social Psychology. University Press.
Dudai, Karni & Bjorn (2015). The consolidation and transformation of memory, Neuron, 88, 20-32.
Gabbert, Memon & Allen (2003). Memory conformity: can eyewitnesses influence each other’s memories for an event? Applied Cognitive Psychology. https://doi.org/10.1002/acp.885
Johansson, Hall, Tärning, Sikström & Chater, N.(2013). Choice Blindness and Preference Change: You Will Like This Paper Better If You (Believe You) Chose to Read It! Journal of Behavioral Decision Making. DOI: 10.1002/bdm.1807
Loftus, E. F., & Zanni, G. (1975). Eyewitness testimony: The influence of the wording of a question. Bulletin of the Psychonomic Society, 5(1), 86–88. https://doi.org/10.3758/BF03336715
Loftus & Pickrel (1995). The formation of false memories. Psychiatry Ann. 25, 720-725
Van Kesteren & Meeter (2020). How to optimize knowledge construction in the brain. Nature Partner Journals: The Science of Learning