“Forgetting focuses remembering and fosters learning; remembering generates learning and causes forgetting; learning causes forgetting, begets remembering, and supports new learning.” Bjork (2015)
Our ultimate goal as teaching professionals is for children to acquire all the necessary learning, and more if possible, efficiently and securely. Having an understanding of the theories behind how the working memory works and how we build schema (assimilating and accommodating new learning) allows us to explore the most effective ways for the learner to obtain and consolidate new knowledge.
Daniel Willingham’s memory model provides a brilliant visual representation of how our mind retains and deals with new information; this links with Robert Bjork’s quote above; I understand it as learning causes forgetting, forgetting causes remembering, and remembering causes new learning. This remembering “part” is essential to ensure acquired knowledge can be easily brought back to memory: this is retrieval.
Retrieval strength and storage strength
Bjork (2015) refers to retrieval as a ‘memory modifier’ as using our memories can alter our memories. Think about how retrieving something more frequently than something else would alter the state of both of those items: one being more accessible and usable than the other.
Robert Bjork’s and Elizabeth Bjork’s Theory of disuse (1992) categorises an item in memory into two ‘strengths’: retrieval strength and storage strength. Retrieval strength refers to how quickly and accessible that item is to be recalled; the latter refers to how well that item has been learnt. So you might be thinking about the need to increase retrieval strength and storage strength; however, that is not necessary the case.
Bjork (2015) states that when something is accessed very often, little to no learning takes place. This could be largely down to recency. Therefore, if we use what we know about the benefits of forgetting and the importance of separating opportunities to retrieve through space and time, this will create desirable difficulties, thus, building storage strength which relates to permanent changes in learning.
Another consideration when planning retrieval tasks is to think about cues: context and state. Context refers to external cues (like environment and other external factors that supported the encoding of that knowledge) whereas state refers to internal cues (emotional, physical and mental state). Godden & Baddeley (1975) conducted an experiment where two groups had to memorise a series of words, one group being under water and the other on land. When recall took place in a different location (context) from where they had learnt (e.g. the scuba divers recalling on land), recall was poorer. Therefore, it’s worth considering the impact of their setting; for example, learning something in Science but then recalling it in Forest school, and vice versa. State cues could be how you’re feeling at the time of learning, considering your emotional, physical and mental state; therefore, returning to the same physiological and emotional state when a memory was formed can help recall that memory (White, 2003). Goodwin (1995) explains how it is not uncommon to hear stories of drinkers who stash alcohol or money while intoxicated and then can locate the hiding places once being intoxicated again.
Thinking about retrieval
Understanding the importance of prior learning and the pre-requisites is key for new learning to be built upon. Having this awareness better prepares us so that we can plan in specific questions, tests and tasks that activate a child’s prior learning. This can act as a desirable load demanding the individual to resurface the forgotten information, remembering it in more detail, and then applying it to the new learning. With this view, we should be seeing retrieval as ‘retrieval FOR learning’ as opposed to retrieval and learning; and what I mean by this is retrieval being seen as a separate entity, an isolated part of the lesson with no particular or evident links.
Now, in contrary to this, I do understand there is a lot of benefit in retrieving facts that are not linked with the lesson of that day; quite often, the use of ‘last day, last week, last term and last year’ quadrants can be useful in agreeing some kind of spacing algorithm across the school to support forgetting and remembering. However, the dangers following this approach is that this can constrain teachers to neglect many opportunities to retrieve prior learning as the pre-requisites for the day’s new learning intention. And I believe this premise greatly correlates with a strong, well-threaded curriculum.
Let me talk to you about my refined purpose for planning retrieval (aside from the obvious: to strengthen memory).
Consider the diagram above (Figure 1) as a representation of my thinking. The elongated bar at the bottom represents a year group from September to July; in this instance, let’s say Year 1. Now, during this period, they are taught a new concept (in yellow). If we could just focus on retaining this information, it would be wonderful to be able to spend all of the following year (Year 2) having hour long (just for argument sake) retrieval maths, reading, writing lessons so that learning is well embedded (enhancing the storage strength). However, of course, we need to teach new learning. So the question is, how do we incorporate retrieval (knowing how valuable it is) alongside new learning. This is where careful and meticulous curriculum design comes into play.
Let’s say the yellow box represents the learning of recognise and sort 2-D shapes. Well, for this knowledge to develop, the curriculum needs to consider how will this knowledge in Year 1 be built upon/ aid the new knowledge in Year 2, then develop further into Year 3 and so on. This is the thread of our spiral curriculum. The concept applies to all subjects, like in humanities, what is the thread of the curriculum? If one year group learns about Roman civilisation and then learns about the Great Fire of London, teachers and children may find it difficult to find links (thinking about the surface facts like dates, locations etc.); however, having a number of threads running through the curriculum that centres around a particular theme will help this (e.g. technological development or inspirational leaders), then these threads are what inform what should be retrieved, identifying similarities and differences across other eras (how does the impact of Trajan’s leadership in Rome compare with the indecisiveness of the Lord Mayor during the Great Fire of London); retrieving for learning.
This idea builds upon the children’s prior knowledge, often first retrieving key facts and dates but then requiring them to retrieve deeper meaning about what they know about the impact of say the leadership, thus, then using their working memory to explore how this memory links to new learning; modifying their memory and increasing their storage strength.
So looking at the arrows on Figure 1, a way to consider retrieval is through the following 4 steps (Figure 2) (this has a broad reference to Fiorella’s and Mayer’s Generative Learning SOI model):
- Select (facts and/or procedure) Select
- Practice Organise
- Application (within context) Integrate
- Application (across contexts) Integrate
This allows a good progression for learning to be remembered, forgotten, remembered and then strengthened as they broaden their schema making further and stronger links of meaning.
Here’s an example.
- Year 1’s Learning intention: Recognise and sort 2-D shapes
- Year 2’s Learning intention: Identify properties and draw 2-D shapes
Once the children have grasped how to recognise and sort 2-D shapes, let’s consult the 4 stages in preparation for Year 2’s learning intention.
If it is necessary to recap this surface level, then retrieval questions like “Name this shape,” or “Match the words to the correct shapes” will allow the children to just select from their long term memory.
Perhaps in the previous retrieval they were able to select the information but had low retrieval strength (needed the adult to give further clues), then you may wish to use the retrieval to encourage more practice of retrieving these facts like “Complete the words and then match to the shapes” or “Sort the 2-D shape names into 4 sides vs not 4 sides” (children needing to recall the shape picture and associate it to the written word).
Now, for me, this is where the thought around retrieval really takes some further thinking but is extremely valuable because we are now encouraging children to retrieve the same information but to use it to apply to something else. E.g. “Half of the shape is missing, only a corner is showing, what shape could it be” or “Here are some objects (box, Doritos tube, pencil pot etc.), what 2-D shapes can you see on the 3-D objects” or “Thinking of shape names, sort the shapes in alphabetical order, then in order of sides.”
4. Application (across contexts)
This is where we can encourage the children to retrieve the information, but retrieve its deeper meaning as developed in the previous step and that they see how it can be applied elsewhere, making it more secure and adaptable e.g. “Using double sided counters, what expression would make a square or a rectangle,” or “In forest school, use the sticks to create a triangle inside a square,” or “In P.E, dribble with the ball in the shape of a triangle.”
There are certainly links between this step and the concept of reasoning, and perhaps, this is a type of reasoning, but this process requires meticulous thinking around what exactly do they know and how much further we can deepen this learning through proximal transference: a contextual retrieval link that activates this specific part of learning
For your retrieval slots, thinking about questions that retrieve prior learning is of course the whole point, but it is worth giving consideration to see what parts of the learning you are encouraging the children to retrieve, thinking beyond just selecting the information but creating a desirable load that requires children to really thinking deeply about the meaning behind the learning in preparation for the next stage of their new learning. Ultimately, there is far more meaning that we intend to impart through teaching which only a small percentage of it is well and truly understood. By only planning opportunities for pupils to select this surface information means less time to retrieve the depths of meaning and increases more time for learning to be forgotten and become some-what inaccessible.
So for teachers, think about the pre-requisites and prior knowledge they have learnt the year before that needs to be retrieved to strengthen their storage, but after teaching the new learning, subsequently consider the 4 stages of retrieval throughout the rest of the year in preparation for this to be built upon again the following year or the next time you return to that thread.
Initial objective: sequencing numbers
- Complete the sequence of numbers.
- Complete the 3rd and 5th sequence of numbers.
- On a bus timetable, there are a 4 stops. It takes me 5 minutes to get to the stop. What is the latest time I can leave home if there was a 5th stop?
- Creating a graph in science, choose appropriate intervals to accommodate the sequence of data.
Initial objective: learning about the order of planets and gas giants
- Recap the mnemonic.
- Identify the 3rd planet from Earth (away from the Sun); identify the 2nd planet closest to the sun from Neptune.
- Which position are the gas giants? Which planets will orbit the sun faster than Saturn?
- Learning about forces, use your knowledge of north and south poles on magnets to represent practically the solar system demonstrating the gravitational pull.
Initial objective: long division
- Explain the process; or complete division questions.
- Divide with 2 digits, doubling the 2 digit divisor each time and noticing what happens with the quotient.
- Improving efficiency with the process, focusing on editing and improving the steps from an example e.g. to speed up the division process, work on strengthening multiplication strategies, so rather than only multiplying by 10 and then if not, doing repeated addition to get close to the first expression, children to practise finding 9 lots by multiplying by 10, minus 1, or doubling and halving, or using knowledge of constant ratio to find close multiples of the divisor in a more efficient way.
- Represent each step of the long division process through a bar model.
Further reading recommendations
- Bjork, R. & Bjork, E. (1992) A new theory of disuse and an old theory of stimulus fluctuation
- Bjork, R. (2015) On the symbiosis of remembering, forgetting and learning
- Enser, M. & Enser, Z. (2020) Fiorella & Mayer’s Generative Learning in Action (In Action series)
- Godden, D. & Baddeley, A. (1975) Context-dependent memory in two natural environments: on land and underwater
- Jones, K. (2019) Retrieval Practice
- Jones, K. (2021) Retrieval Practice 2