What is cognitive load theory?
Australian educational psychologist and academic, John Sweller published his research on Cognitive Load During Problem Solving: Effects on Learning (1988)
Most teachers are naturally interested in how learning happens. Research on retrieval practice, how memory is shaped, as well as cognitive load theory, I thought it would be helpful to publish this blog post in an attempt to explain Sweller’s original research as a beginners guide for all teachers.
Firstly, if you are serious about research and you’ve not read the original research paper, it is important that you do – it’s 29 pages in length – or at least have a good skim through the full paper. There is also the original book if you are interested, but it is costly, which is why I think the free research paper is probably your better alternative.
The second critical point to emphasise is that Sweller’s original paper is focused on problem-solving, which “requires a relatively large amount of cognitive processing capacity which is consequently unavailable for schema acquisition.”
A third point worth mentioning is that Sweller himself acknowledges that ‘discovery learning’ (Bruner, 1961) was spawned “by the perceived importance of fostering problem-solving skills.” As with all things in education, research leads to new ideas, regurgitated and filtered…
My interpretation will be subjective, which is why the above links are important references before you read on.
Thirty years later, where are we now, and what can teachers learn from the original research?
Sweller writes in the opening page of his paper, “that contrary to current practice and many cognitive theories, some forms of problem-solving interfere with learning.”
The first thing to distinguish is what does Sweller define as ‘some forms’ and ‘learning’. I have tried to do this below from his original paper.
Are pupils learning how to construct an argument for a history assignment on the Battle of Britain and why Britain’s planes outclassed the Germans? Perhaps we are learning how to construct a tower from Lego bricks, in pairs inside a year 3 classroom, to balance a tennis ball at the highest level?
Whatever is it, learning [what] needs to be defined. We also need to understand if different types of problem-solving exist and in what form. It is also worth mentioning that domain-specific knowledge is the acquisition of specialised structures.
For example, I may write the perfect blog post to explain what cognitive load is, but I may miss the details in return for a concise blog post for busy classroom teachers to read. Instead, I should consider that some teachers will read this blog post and opt for depth and breadth rather than an easy-to-access blog or a skim-over the surface attempt. I have tried to balance your reading time on this post whilst unpicking the 29-pages of research.
The above is an example of my domain-specific knowledge as a teacher (the theory and practice) versus my life as a blogger (the clicks). I have opted for a happy medium.
Sweller offers 3 categories from extensive research.
1.Memory of problems state configurations
This particular research on memory (Groot, 1966) derives from investigations between chess players who are masters versus less experienced players. Differences occurred in “chunk sizes with masters’ chunks being far larger than those of novices.”
In summary, expert players could remember larger sequences of moves…
2. Problem-solving strategies
Mathematics is given as an example here, where “problems can be solved using search techniques such as ‘means-end analysis’ which involves attempting to reduce differences between each problem.”
Strategies selected by expert and novice problem-solvers were different. Novices worked backwards from the goal, setting sub-goals, whilst experts “eliminated the backwards-working phase”.
Sweller offers an example and highlights that experts “work forward immediately” because they recognise each problem from previous experience.
Cognitive structures (schemas) allow experts to accurately recall the configuration of a given problem… (Sweller, 1998)
In essence, the person solving the problem can group the scenario when they possess an appropriate schema, memorising configurations compared to novices you may not have this prior knowledge.
3. Features used in categorising problems
Categorising algebra word problems is offered in this final example.
“If an expert has a schema which suggests that conservation of energy should be used to solve a particular problem, then that problem is likely to be categorised with other problems to which the same schema can apply.”
The expert-novice that “domain-specific knowledge, in the form of schemas, is the major factor distinguishing experts from novices in problem-solving skill.”
I know Sweller cannot offer endless case studies, but we can at least we are offered 3 problem-solving scenarios provided: a Chess game and two maths examples.
Sweller explains that little research has been carried out; that it is “commonly assumed by both theoreticians and those concerned with practical problem-solving issues that practice on a large number of conventional problems is the best way of gaining problem-solving skill.”
In his paper, he offers experimental evidence of interference between problem-solving and schema acquisition. This time ‘puzzle solving’ is provided as an example with “means-ends analysis” provided as further evidence of negative effects on learning.
Remember, means-ends analysis involves solving a problem by first considering the obstacles in the way, then attempting to solve the problem using sub-goals.
Sweller continues that means-ends analysis (solving one step at a time) may interfere with learning. Lots of examples are provided in the original paper to elaborate on how problems can be solved (or not) using means-ends production.
“A problem-solver must attend [to] the differences between a current problem and the goal.” If pupils, for example, where to use means-ends analysis – solving a problem one step at a time – the “relationship between problem and operators can be totally ignored…under most conditions.”
In comparison, pupils for example, who have the required schema acquisition, can recognise a problem as “belonging to a particular category of problem that requires particular moves.”
Cognitive Processing Capacity
Most teachers will have a general understanding of working memory – that it is limited and we can only manipulate a number of pieces of information at any one time – before additional information becomes redundant.
There is nothing to contest with here if we accept that cognitive load theory as presented, is under the umbrella of problem-solving. However, we know problem-solving in some respects is evident in all subject areas of the curriculum and that developing metacognition in our young people requires a degree of a pre-organised schema.
What is Cognitive Load?
We know that the brain is limited by what is can do with any new information at one time. All the research that I have read suggests that we can manipulate between four and nine pieces of information at any one time. We also know that the human brain has 100 billion neurons with 100 trillion possibilities – there are no known limits to how much information we can store.
What has evolved from Sweller’s research is that various instructional techniques are recommended (that fit in with the characteristics of working memory) and therefore unlock better learning potential.
I have written about cognitive apprenticeship which offers a paradigm for teachers.
I have also written about memory in the past. We know that our memory can only store small amounts of information for short periods of time. If you’re reading this blog post (still), I’m working hard to reduce unnecessary information (albeit, accepting that this sentence is a perfect example of the redundancy effect).
This is where we store large amounts of information waiting to be of service. Some of this information is either explicit (conscious) or implicit (subconscious). For example, reading these words and knowing what each word means, or riding a bike or tying your shoelaces.
Cognitive Load Theory
Sweller’s research highlights that knowledge is stored in our long-term memory in the form of a schema.
A schema (in some respects) is a mental mindmap of how one stores and processes information, deciding what to do and how the information will be used.
Our capacity to store endless information is unknown, and there is no limit to how complex the schemas can organise themselves. I guess where learning happens in our long term memory is evident, is when minimal conscious effort and a degree of automation kicks in, alongside developing expertise.
In an excellent paper summarising cognitive load, published by the Centre for Education Statistics and Evaluation (2017) a simple and effective example is provided:
- “Try to remember the following combination of letters: y-m-r-e-o-m.
- Each letter constitutes one item, so you are required to remember six items at once.
- Now try to remember the following combination of letters: m-e-m-o-r-y.
- This time, you are still required to remember the very same six items.
- However, because you have an [established] schema in your long-term memory for the word ‘memory’, you are able to chunk the letters into just one item” and your working memory is free to remember other items.
The paper continues with types of cognitive load. Something I have been using in my classroom and teacher training for several years which I would recommend all new teachers become familiar with.
Types of Cognitive Load
- Intrinsic load deals with information that needs to be processed; the number of elements that must be simultaneously processed in working memory and their interaction with each other. It’s affected by both the nature of the task and by levels of learner expertise… (The material itself.)
- Extraneous load is a form of working memory load that refers to the load imposed by information elements unrelated to the learning task, but related to how that task is carried out; how learning takes place. Essentially, these elements can be controlled by the person who designs the learning experience. That’s YOU! (How material is taught.) It can be relevant or harmful…
Something to be aware of is the term ‘Germane Load’ – the process of learning vs. working memory. In essence, a ‘healthy’ type of load.
In a recent paper, From Cognitive Load Theory to Collaborative Cognitive Load Theory by Paul A. Kirschner et al (including John Sweller, 2018), germane load is defined as “working memory resources devoted to dealing with intrinsic cognitive load.”
Of great importance is this research recongises that germane and intrinsic load are indistinguishable. Perhaps germane load is redundant, unless refined “as referring to the actual working memory resources devoted to dealing with intrinsic rather than extraneous load!”
In this context, the researchers recommend not to use collaborative learning even after a teacher considers their students’ expertise, the type of task and the composition of the class dynamics, alongside the learning goals and cognitive ability of the class!
Put simply for teachers? ‘Collective working memory’ hinders the success of the group.
In terms of germane load, teachers should consider well-designed activities and worked-example resources to help lower cognitive load. In return, specific tasks can help facilitate the development of schema and automation. If we make one final reference to intrinsic load, how a teacher develops these resources will be very important to how a student deals with any information.
I’ve produced a simple resource below to summarise the fine balance a teacher must work within.
There is also a great 3-minute video that explains the above types of load.
From first publication, Sweller’s research in mathematics and problem-based solving has helped the teaching profession understand how learning happens. He concludes that “there seems to be no clear evidence that conventional problem-solving is an efficient learning device.”
If you have the time, there is a fascinating 40 minute YouTube video, published in 2012. It’s probably one of the best things I could recommend you do next if you have read through all of this!
Cognitive load theory is not a theory of everything. It’s a theory of how we process information and how we facilitate instruction…
John Sweller provides much-needed context to help distinguish the nuances we all contemplate when discussing teaching and learning…
From what I can determine from the paper, Sweller highlights (1988) that means-ends analysis (solving one step at a time) imposes a heavy cognitive load and that schema acquisition may be substantially distinct. The “cognitive effort required by conventional problem-solving may not assist in this schema acquisition, and given that this is the most important component of problem-solving expertise, the development of expertise may be retarded by a heavy emphasis on problem-solving.”
Whilst Swellers original paper is tricky to unpick, I’d recommend that teachers return to the Centre for Education Statistics and Evaluation for a good overview.
A final thought…
What does cognitive load look like in a lesson when observing learning taking place in the mind of a 4-year-old, a 16-year-old or a child with learning difficulties? A mathematics or drama lesson, or a design and technology problem-solving design brief? Teaching and learning is highly complex…
It appears that problem-solving as an activity itself is not a good learning activity to use to help pupils learn how to solve problems! I conclude that ‘problem-solving’ is therefore only useful if pupils have prior knowledge and have been explicitly taught specific types of problem-solving strategies.
Where misconceptions are not challenged, or where knowledge is disorganised or left to chance, we know learning is less effective. We all know that. Now it is up to you to take what you know about cognitive load theory and apply it in your teaching…
N.b. the Centre for Education Statistics and Evaluation concludes this recommendation for maths, science and technology teaching.