“The art of remembering is the art of thinking . . . our conscious effort should not be so much to impress or retain (knowledge) as to connect it with something already there”
William James, 1890
“Memory is the residue of thought.”
Daniel Willingham, 2009
It’s interesting to put these two quotes side by side. They are separated by over a century of psychology but essentially saying the same thing. Whilst many of our ideas about memory have changed since the time of William James, some of the basic foundations are still there. Likewise, in education the distinction between ‘active’ and ‘passive’ learning has been a perennial one. Indeed, these ideas are couched in the same psychology of memory. Unfortunately, ‘active’ learning has lost something of this connection in some education discussions – often conflated with ‘engagement’ and ‘independent learning’. David Didau recently argued this point in his post “Is listening really passive?”
When I hear the term ‘active learning’, I immediately think of ‘levels of processing’ and ‘elaborative rehearsal’. Understanding the memory research that underlies the concept of ‘active learning’ may help teachers avoid mistaking engaged activity (which may only involve superficial levels of processing) for active processing.
Whilst there’s been a great deal of interest and attention paid to the working memory model, other areas of memory research should also be of interest to teachers. A really famous idea, which all teachers should know, was formulated out by Fergus Craik and Robert Lockhart in 1972.
They suggested that the strength of long-term memory was dependent upon the ‘depth of processing’ used during the encoding of the material. Shallow processing (e.g. Structural – attending to what words or letters look like) leads to short memory traces, deeper processes (e.g. Phonemic – attending to the sounds of a word) leads to slightly longer traces and the deepest processing (e.g. Semantic – attending to what the words mean) leads to the longest memory traces.
Craik and Tulving tested this hypothesis in a well-known experiment in 1975. They asked participants to answer a series of Yes / No questions.
“For example, in shallow processing, the subject answered questions concerning the word’s typeface (for example, is the word “HOUSE” written in capital letters?); in intermediate processing, the subject answered questions about rhyme (for example, does the word “house” rhyme with “pencil”?); and in deep processing, the questions were directed toward the word’s semantic content (for example, does the word “house” fit into this sentence: “The ______ has a beautiful window”?).”
As predicted, the items which required participants to answer semantic level questions tended to be recalled better than those requiring more superficial processing. Obviously the idea had critics, Alan Baddeley best known for the working memory model among them. A key criticism was the apparent circularity of the reasoning; deeper processing leads to better recall, but what is deeper processing? – that which helps recall!
One approach to overcome this criticism has been attempts to identify neurological correlates that might serve as a measure of ‘deep’ processing:
“In keeping with similar previous suggestions (32-34), we propose that when subjects process verbal stimuli in a semantic manner, either under experimental or real life conditions, this involves increased neuronal activity in the left inferior prefrontal cortex. Increased activity in this region, irrespective of the individual’s intention to remember, leads to a more readily retrievable memory trace.”
However, over time the core idea of ‘semantic processing’ has been joined by other processes that appear to make up ‘deep processing’.
Elaboration, distinctiveness and effort
Craik and Lockhart continued to refine and revise their ideas in light of new experimental evidence (much of it stimulated by the debate about their ideas). One fairly early refinement was the admission that elaboration processes were clearly working in addition to the semantic ‘depth’ of initial processing.
“This conception of a series or hierarchy of processing stages is often referred to as “depth of processing” where greater “depth” implies a greater degree of semantic or cognitive analysis. After the stimulus has been recognized, it may undergo further processing by enrichment or elaboration. For example, after a word is recognized, it may trigger associations, images or stories on the basis of the subject’s past experience with the word.”
With this addition, the basic idea of levels of processing still serves as a research framework within psychology.
“We argue that both notions of depth and elaboration are necessary: depth to talk about qualitative types of analysis and to capture the notion that some types of processing (typically sensory analyses) precede others and require few attentional resources to achieve them; elaboration to refer to richness or extensiveness of processing within each qualitative type.”
Material with high distinctiveness also tends to be recalled better as are materials which require more cognitive effort to process.
“So, in this instance, we tend to agree with Baddeley that the original notion of depth is too simple a formulation to provide an adequate analysis of situations involving further processing within a given domain, except perhaps in the sense that such further, more elaborate analyses often require more effort and more processing resources.”
Deeper processing may not require more time but, depending on the person’s subject knowledge, it does require more attention:
“Craik (2002) points out that deeper processing does not necessarily require more time than shallow processing. However, deeper processes require more attention (Treisman, 1964; Craik & Byrd, 1982). Divided attention results in shallower encoding, but this relationship is modulated by the meaning of the stimulus and by the expertise (i.e., the degree of content knowledge) of the person doing the processing.”
Links between working memory and levels of processing
There have been some interesting attempts to unite levels of processing with the working memory model. Whilst Baddeley was something of a critic of Craik and Lockhart’s ideas, Rose suggests that given long-term memory involvement in working memory, levels of processing manipulations that improve long-term recall should influence working memory task performance.
In a short sequence of experiments, Rose gets participants to process works based on their visual, phonological, or semantic features, then looks at the recall of the words. What’s interesting is that levels of processing effects were not found in an immediate recall test, but did occur when tested after a delay or when given a surprise test.
“The reason for this pattern of findings is likely due to the fact that on working memory tasks, participants intentionally encode to-be-remembered items and attempt to actively maintain those items in preparation for an upcoming memory test.”
This helps to clarify something of the relationship between working memory and levels of processing in long-term memory. It seems that when we’re expecting to be tested, we rely predominantly on working memory to maintain items for recall. However, where recall is delayed or unexpected we are unable to use this mechanism – instead we have to rely on long-term memory. Thus if you want to know what pupils recall from your lesson – the responses to a plenary quiz likely won’t give you too much insight.
So what counts as active learning?
The best answer is to try to ensure our students are cognitively busy in our lessons. Avoiding disruption to the lesson due to poor behaviour is key, otherwise attentional resources are diverted from the material to be learnt. Beyond this our best bet is to try to structure the lesson in a way that provides opportunities for semantic processing and elaboration, encourages optimal cognitive effort (i.e. within limitations of working memory) and makes the lesson and the ideas as distinctive as possible (but without diverting attention). This isn’t easy to do – and even harder to assess, however a good starting point is to simply avoid activities which encourage superficial processing of the core knowledge and ideas.
For example, word search puzzles are a great example of ‘structural processing’ – they can be completed with no understanding of the key words but simply pattern matching the first few letters. A better ‘quick quiz’ type activity might be to use a crossword – at least forcing students to attend to the semantic properties of the key words.
Other cognitively ‘easy’ activities (e.g. underlining and cloze tasks) are also probably best avoided, unless you can shape the task in such a way that the students are forced to really consider the meaning of the key words to complete it. If the choice of word (to underline or insert) can be guessed using superficial features of the task, then it’s likely very little learning is going on. The same is true for children working in groups who copy answers from one student – it’s important to consider whether ‘collaboration’ will genuinely assist learning or distract attention from it.
Mindmap type activities can potentially be used well – forcing children to consider the connections between key ideas and represent them showing these links – however, this requires considerable training and effort to get right. A spider diagram with a beautiful, bubble-written title and three key words likely indicates very little processing has taken place. Likewise, if we simply provide a ready-made mindmap, it’s unlikely to add much to the learning. If we use these things in lessons we need to really push for content – perhaps providing a very large number of key terms that students will struggle to organise into meaningful groups.
Activities inserted into lessons mainly to engage students may not always encourage ‘deep’ semantic level processing. Though fun, if these do not require students to consider the material very deeply then it’s unlikely to be learnt. It’s better to consider this issue the other way around – what are the key vocabulary and ideas I need them to think about? – then – how can I make this more distinctive?
Inside the black box
In essence, David Didau is right – listening could easily involve intense cognitive activity at a semantic level. On the other hand, it might not – a student could be thinking about something else entirely! The same is true for inquiry learning activities. In principle, encouraging students to use inductive inferences to solve questions about science should require considerable processing – yet, it’s also possible that students will attend to superficial aspects of the task and simply reproduce their own prejudices and misconceptions. There’s no easy way to tell (which was rather David’s point).
Using frequent, delayed assessments of learning may give you a glimpse into how much processing has occurred. Mix in some questions targeted at misconceptions or that require an application of some of the key ideas, and you may even get a glimpse of how students have connected up that knowledge. For me that was the point behind assessment for learning. The ‘black box’ in education was never really the classroom – but a student’s mind!
Great stuff – particularly like the way you move from abstract theory to real practice so that teachers can better understand the potential application of the research. Your careful use of English avoids the imperious certainty that can sometimes undermine lesser blogs at the moment.
Thanks for the comment. Yes – I think it helps to see examples of how some of these ideas can be applied to teaching.
Another terrific squirt of research into the rusty old machine; thank you.
“Thus if you want to know what pupils recall from your lesson – the responses to a plenary quiz likely won’t give you too much insight”
– but the responses to a quiz a few days down the line, might do?
My thinking here is that good questions in the lesson or plenary don’t demonstrate long-term learning but can expose misconceptions so a teacher has a chance to fix them. David has been emphasising that they don’t demonstrate long-term learning. I think this needs to be balanced against the AfL idea that identifying misconceptions might affect decisions about where to go next. There also seems to be quite good evidence for spaced retrieval practice improving long-term memory of knowledge and there is potential for killing two birds with one quiz if the retrival practice helps the children and the results help to show the teacher what has and hasn’t been learned. Does that make sense?
“- but the responses to a quiz a few days down the line, might do?”
Yes – would give some distributed practice too. Two birds – one stone indeed.
I agree that a misconception question might work well at the end of a lesson – especially for planning where the learning goes next. Though in my experience, kids often ‘get it’ in the lesson, then the misconception creeps back in over time – so a delay might work better there. Interesting question!
Pingback: The science of learning | Evidence into practice
Pingback: The Science of Learning | Blogs of the Month
Pingback: Germane load: The right kind of mental effort? | Evidence into Practice