This week’s blog post discusses how we can apply what we know about memory to maximise the effectiveness of revision. It’s divided into three sections, with each addressing a specific area:
1. Achieving mastery over a skill or an area of knowledge
2. Practice and consolidation
3. Retrieval practice for remembering
Each of the above is divided into several sub-sections.
This blog post focuses on three pieces of research. In particular, it focuses on Peter C. Brown, Henry L. Roediger III & Mark A. McDaniel’s seminal book on the science of learning, Make It Stick: The Science of Successful Learning (Cambridge, MA: Harvard University Press, 2014). It also draws extensively upon Barak Rosenshine’s highly influential ‘Principles of Instruction’ (2010) and Tom Sherrington’s book on Rosenshine’s ‘Principles’, Rosenshine’s Principles in Action (Woodbright: John Catt, 2019). References to quotes are included in the running text; other references are linked with endnotes.
Several previous CIRL blog posts have focused on the books and article above, which are linked at the relevant points below and at the end.
Jonathan Beale, Researcher-in-Residence, CIRL
1. Achieving mastery over a skill or an area of knowledge
1.1 Learning and memory
A learning model from cognitive science which underlies much educational research can be summarised as follows:
- There are two types of memory: working memory and long-term memory.
- We process information in our working memory to store it in our long-term memory.
- Working memory has very limited capacity and can only handle a few pieces of information at once, whereas long-term memory is unlimited.
- We organise information into schemata and build schemata in long-term memory (a ‘schema’ is a well-connected network of ideas).
- We retrieve information from long-term memory into working memory when needed.
- New information only becomes stored if we can connect it to knowledge we already have; therefore, prior knowledge significantly influences our capacity to learn.
- The more consolidated and interconnected our schemata are, the easier it is to make sense of and organise new information which relates to our existing schemata.
- As our schemata become more formed and interconnected, we can explore our knowledge and retrieve it more fluently.
- We forget information that we don’t initially store successfully in a meaningful schema or we don’t retrieve frequently enough.
- We quickly lose around 70% of what we have just heard or read. The remaining 30% goes away more slowly.
- To replace schemata, we need to re-learn; we can’t overwrite existing schemata with new learning. As Sherrington puts it:
‘If a schema contains incorrect information … we can’t simply overwrite it. A more primitive schema can return to dominate unless we unpick and fully re-learn a correct schema’ (Sherrington 2019, p. 11).[1]
For more on what recent evidence in the science of learning tells us about how we learn, see last week’s blog post, by Stuart Kime.
1.2 Automaticity and mastery
Automaticity is the stage of learning where recalling a particular area of knowledge has become effortless. Automaticity results from thousands of hours of deliberate practice: a systematic method of effortful, highly focused, goal-oriented practice which aims at improving performance.[2] It requires overlearning: learning beyond ‘initial mastery’, such that recall is automatic and skills are fluent. (For more on overlearning and deliberate practice, see this blog post.)
Since working memory is limited, if we use much of it for recall, we have less available for other mental activities. When we can recall knowledge automatically, it doesn’t take up space in working memory. This space can be used for other tasks, such as learning something new or problem-solving.
For example, in discussion of gaining mastery at mathematical problem-solving, Barak Rosenshine writes:
‘Mathematical problem solving is … improved when the basic skills (addition, multiplication, etc.) are overlearned and become automatic, thus freeing working-memory capacity’ (Rosenshine 2010, p. 13).
To reach the highest level of mastery over a particular skill or area of knowledge, students need to reach a level of automaticity in recalling some knowledge and skills in that area, and mastery over some others. Students can only be expected to reach automaticity over certain skills during their schooling. But they can reach a high level of mastery over many others, which is a stage on the way to automaticity. When we reach a high level of mastery over a skill, employing that skill uses less space in working memory.
1.3 Gauging mastery with time to prepare: the usefulness of formative testing
How can we gauge mastery of the skills and areas of knowledge students need to develop, early enough that these skills can be developed by the time of their exams? To respond to this question on the basis of the points above, we could answer that it is useful to gauge, early in the revision process, which skills students most need to improve their mastery over, and to focus on developing these as early as possible.
An effective way to gauge mastery is through formative testing: a form of formative assessment where students take tests to evaluate their learning during a course (as opposed to ‘summative assessment’, which takes place at the end). Formative tests have many benefits for learning. The retrieval practice involved in tests strengthens memory. Formative tests also enable the teacher and student to better identify the areas a student has mastered and those in which they most need to improve. Formative tests can take a variety of forms, from low-stakes quizzes to high-stakes tests.
We can use formative testing as a means of gauging which skills and areas of knowledge students most need to most improve upon. So it is useful to employ formative tests early in the revision process, so areas in need of improvement can be identified early enough such that mastery over those areas can be developed by the time of an exam period.[3] (For discussion of Brown, Roediger and McDaniel on formative testing and its benefits, see this blog post.)
2. Practice and consolidation
2.1 Avoid massed practice
Massed practice is the process of trying to learn something quickly by focusing on nothing but the area of knowledge or skill you’re trying to learn, and repeatedly practising it. It’s the ‘rapid-fire repetition of something you’re trying to burn into memory, the “practice-practice-practice” of conventional wisdom’ (Brown, Roediger & McDaniel 2014, p. 3). The paradigmatic example of massed practice is cramming for an exam.
Massed practice is one of the least productive learning strategies. It only has short-term learning benefits.[4] But there may be one time during the revision process in which massed practice could be useful. Employing massed practice the morning of an exam could perhaps be beneficial for performance in that exam, given its short-term learning benefits. But it’s vital to make clear that it may only have these potential benefits if until that point other, effective learning strategies have been employed. And even then, massed practice will not yield long-term learning gains.[5]
2.2 Spaced practice
Spaced practice is the process of spreading learning over time. Evidence suggests that spaced practice is more effective for long-term retrieval. The space needs to be longer than very brief intervals; research suggests that very brief intervals are no better for learning than massed practice.
When we space out revision, it often feels less productive. This is because the spaces cause memory to no longer be fresh in our minds. But allowing some forgetting to occur can have long-term learning benefits.[6]
2.3 Benefits of forgetting
The greater the effort required to retrieve knowledge or a skill, the more the learning is consolidated and the better long-term retention will be. The converse is also true: less effort results in less consolidation and retention. (For further discussion of this point, see this blog post.)
There are learning benefits to a small amount of forgetting, such that greater effort is required for recall. Forgetting can be utilised as a ‘desirable difficulty’: a short-term impediment which yields stronger learning. (For further discussion of desirable difficulties, see this blog post.)
How much forgetting? A small amount of forgetting has the advantage that retrieval requires greater effort and leads to better long-term recall. Too much and students have to re-learn material. So, to maximise the benefits of forgetting, a useful principle to follow is: allow a little forgetting to set in, but not so much that students have to re-learn material and you have to re-teach it.[7]
Spaced practice and interleaving support this process, as Brown, Roediger and McDaniel observe:
‘when you let the memory recede a little, for example by spacing or interleaving the practice, retrieval is harder … but your learning is deeper and you will retrieve it more easily in the future’ (Brown, Roediger & McDaniel 2014: 75).
For further discussion of the benefits of forgetting, see this blog post.
2.4 Consolidation
Spaced practice is more effective than massed practice because long-term learning requires consolidation: the process of strengthening mental representations in long-term memory:
‘[E]mbedding new learning in long-term memory requires … consolidation, in which [memories] are strengthened, given meaning, and connected to prior knowledge’ (Brown, Roediger and McDaniel 2014, p. 49).
This takes between several hours and several days, because
1. this period is required for consolidation to occur;
2. for learning to be durable, we need time in which we can mentally rehearse what we have learned, to remember it better and apply it in other contexts;
3. by incorporating time gaps, we have to make greater effort to retrieve knowledge, which retriggers consolidation and makes learning stronger.[8]
Sleep plays a crucial role in memory consolidation. Taking this into account is useful for estimating how long intervals between spaced practice should be to optimise its benefits. Practice with at least a day in between sessions is recommended to optimise memory consolidation, since you’ve then had time for a night’s sleep, which both supports the consolidation of memories and provides an adequate amount of time for spaced learning to take place.[9]
2.5 Structuring a revision period to make the most effective use of spaced practice
How can we structure the revision period to make the most effective use of spaced practice? Here are some pointers, in the context of teaching a class over several lessons:
- Several consecutive lessons could focus on practising the same specific skill, but there should be at least a day between each class.
- To optimise the benefits of spaced practice and forgetting, allow time between practising skills but not so much time that knowledge or skills have to be re-learned. Several weeks between practising skills is fine, but several months is not.
- To optimise the benefits of forgetting, regularly return to areas of learning, but not too regularly.
- To optimise the benefits of interleaving, over a week of lessons, focus on developing several skills and areas of knowledge rather than one or two.
- Focusing on two skills in a single lesson rather than one could also be beneficial. But, to reach mastery, focusing on too many skills in a single lesson or over a week would be detrimental to learning.
3 Retrieval practice for remembering
3.1 Memory and retrieval practice
Retrieval practice is the process of recalling knowledge or skills from memory. Retrieval practice strengthens memory and interrupts forgetting.
For learning to be retained long-term, we need to do two things:
1. embed learning in our long-term memory;
2. associate new learning with a varied set of cues that enable us to become adept at recalling knowledge later.
These ‘cues’ are known as retrieval cues: ways of cueing knowledge from long-term memory.[10]
Like our working memory, our retrieval capacity is very limited. But when we forget something, it’s often our retrieval cues that are forgotten rather than the knowledge itself. Or, sometimes, we stop paying attention to certain cues.
Two points about retrieval are particularly important to note:
1. First, retrieval cues enhance our capacity for retrieval. This involves creating cues that relate what we are learning to what we already know. If we can relate learning to what we already know, there is no limit to how much learning we can remember.
2. Second, retrieval practice of knowledge or skills from memory makes it easier to recall them again.This is because the act of retrieving a memory changes the memory, making it easier to retrieve.[11]
3.2 Memory and undesirable difficulties
Undesirable difficulties are impediments to learning that the learner cannot overcome:
‘If … the learner does not have the background knowledge or skills to respond to [learning strategies] successfully, they become undesirable difficulties’ (Bjork & Bjork 2009).
An example of an undesirable difficulty is anxiety while sitting a test. Anxiety takes up some working memory capacity, so its presence can cause a student to perform worse in an exam.[12] (For further discussion of these points, see this blog post.)
3.3 Reviewing
Daily, weekly and monthly reviewing of material is beneficial for consolidating learning in long-term memory. Reviewing helps to strengthen the connections between our areas of learning and builds mastery. It also helps to develop students’ fluency and confidence in a subject.
Reviewing activates prior learning. Activating relevant prior learning in the working memory is particularly important when teachers wish to introduce new learning.
The effort involved in recalling learning helps to embed it in long-term memory. The more often we do this, the greater ease we have when we try to connect any new material we learn with our existing knowledge.
More effort is required for weekly and monthly review than daily review. But, as Brown, Roediger and McDaniel have convincingly argued, the greater the effort, the greater the long-term gains.[13]
For discussion of Rosenshine and Sherrington on reviewing, see this blog post.
3.4 Reviewing and revising in small steps
Reviewing makes it ‘easier to be successful with problem-solving as less space in short-term memory is needed’ (Sherrington 2019, p. 37). This relates to the issue of cognitive load:
‘The cognitive load involved in a task is the cognitive effort (or amount of information processing) required by a person to perform this task. If the cognitive load needed for learning becomes excessive, little or no learning can occur’ (Reif 2010, p. 361).
Given the limited capacity of working memory, it is beneficial to learn material in small steps. To avoid cognitive overload – an undesirable difficulty – present information in small steps and only proceed to the next step once the previous steps have been mastered.[14] (For further discussion of this point, see this blog post.)
How can we structure the revision period to make the most effective use of reviewing and revising in small steps? Here are some pointers:
- Try to ensure that all students have mastered essential skills by the time of the revision period, so they do not need to spend time developing these during the revision period.
- If some students still need to master some essential skills by the time of the revision period, set these students particular tasks to work on for homework, independent study or as separate tasks during a lesson, while other students work on developing other skills.
- Incorporate weekly and monthly (or half-termly) review with classes.
- Encourage students to engage in independent, daily review. This could be monitored, for example, through online programs where the teacher can see the work the student posts on a course webpage, such as the OneNote ‘Class Notebook’, and/or through processes of peer assessment, between the members of a class. (On assessing learning when teaching online, see this blog post.)
Key points: teaching and learning strategies to employ and those to be avoided
On the basis of the above discussion, here’s an summary of teaching and learning strategies to employ and those to avoid.
Teaching and learning strategies to employ:
- Retrieval practice.
- Spaced practice.
- Ideally, use the two strategies abve in conjunction with one another – i.e., spaced retrieval practice.
- Formative testing.
- Revising material in small steps.
- Reviewing: weekly, with classes, and daily, independently by students.
Teaching and learning strategies to be avoided:
- Massed practice (except, perhaps, the morning of an exam, as long as this follows the use of effective learning strategies).
- Repetition, where this isn’t spaced or interleaved – for example, re-reading. This is a form of massed practice. Repetition is only effective for short-term memory. To be effective, repetition needs to be supplemented with other learning methods.
- Employing difficulties to the point where they become undesirable.
- Revising with ready-to-hand resources (for example, cue cards or open books). As Sherrington points out, we tend to think ‘we’ve learned something if the information is continually presented to us’. To avoid this, students need to generate versions of the knowledge to be recalled ‘from memory without looking at the source’ (Sherrington 2019, p. 38).
References
Bjork, E. L. & R. A. Bjork, 2009. ‘Making Things Hard on Yourself, But in a Good Way: Creating Desirable Difficulties to Enhance Learning’, in M. A. Gernsbacher, R. W. Pew, L. M. Hough and J. R. Pomerantz (eds.), Psychology and the Real World: Essays Illustrating Fundamental Contributions to Society (New York: Worth, 2009, pp. 56-64)
Brown, Peter C., Henry L. Roediger III & Mark A. McDaniel, 2014. Make It Stick: The Science of Successful Learning. Cambridge, MA: Harvard University Press.
Knight, Frances Le Cornu, 2020. ‘Sleepy teens in the classroom’ in J. Harrington, J. Beale, A. Fancourt & C. Lutz (eds.), The ‘BrainCanDo’ Handbook of Teaching and Learning: Practical Strategies to Bring Psychology and Neuroscience into the Classroom (London: Routledge, 2020).
Mascarenhas, Duncan R. D. and Nickolas C. Smith, 2011. ‘Developing the performance brain: decision making under pressure’. Performance Psychology (2011), 245-67.
Reif, Frederick, 2010. Applying Cognitive Science to Education: Thinking and Learning in Scientific and Other Complex Domains. Cambridge, MA: MIT Press.
Rosenshine, Barak, 2012 [2010]. ‘Principles of Instruction: Research-based Strategies That All Teachers Should Know’, American Educator.
Sherrington, Tom, 2019. Rosenshine’s Principles in Action. Woodbridge: John Catt.
Shibli, D. & R. West, 2018. ‘Cognitive Load Theory and its Application in the Classroom’. Impact: Journal of the Chartered College of Teaching, February 2018.
For detailed summaries of the article by Rosenshine and the books by Sherrington and Brown, Roediger & McDaniel, see various CIRL blog posts. For a discussion with and Brown, Roediger & McDaniel on some of the topics discussed above, see the recording from a recent webinar we hosted with them, available on the CIRL podcast.
[1] On the learning model summarised here, see Brown, Roediger & McDaniel 2014, p. 28; Rosenshine 2010; and Sherrington 2019, pp. 10-12.
[2] See Duncan R. D. Mascarenhas and Nickolas C. Smith, ‘Developing the performance brain: decision making under pressure’ (Performance Psychology 2011, pp. 245-67): deliberate practice, they write, involves effort and it aims towards improvement of performance (rather than personal enjoyment).
[3] On the points in this section, see Brown, Roediger & McDaniel 2014, p. 19 & chapter 2.
[4] Brown, Roediger & McDaniel 2014, pp. 3 & 9-10.
[5] For a discussion of this point with Brown, Roediger & McDaniel, see the recording of the webinar we hosted with them last year, available on the CIRL podcast.
[6] On the points in this section, see Brown, Roediger & McDaniel 2014, pp. 4 & 47.
[7] On these points, see Brown, Roediger & McDaniel 2014, pp. 68, 79 & 82.
[8] Brown, Roediger & McDaniel 2014, pp. 63 & 72-73.
[9] Brown, Roediger & McDaniel 2014, pp. 63-64. For a recent study of the importance of sleep for the education of adolescents which supports some of these points, see Knight 2020.
[10] Brown, Roediger & McDaniel 2014, pp. 3 & 75.
[11] Brown, Roediger & McDaniel 2014, pp. 28, 41 & 75-77.
[12] Bjork & Bjork 2009; Brown, Roediger & McDaniel 2014, pp. 91-92 & 99.
[13] Rosenshine 2010, p. 13; Sherrington 2019, pp. 35-36.
[14] Reif 2010, p. 361; Shibli & West 2018; Rosenshine 2010, pp. 16-17; Sherrington 2019, pp. 35-37.