Cognitive Science Approaches
Cognitive Science Approaches
Cognitive Science Approaches
We would like to thank all those who have contributed or provided advice in the production of the original
review, and this teacher facing summary.
Review team: Dr Thomas Perry, Dr Rosanna Lea, Clara Rübner Jørgensen (University of Birmingham),
Prof. Philippa Cordingley (CUREE), Prof. Kimron Shapiro, Prof. Deborah Youdell (University of Birmingham).
Additional support in writing this practitioner summary was provided by Jonathan Kay and Harry Madgwick (EEF).
We would also like to thank the advisory panel who assisted the review team.
Advisory panel: Dr Robin Bevan, Prof. Robert Coe, Dr Iroise Dumontheil, Dr Amy Fancourt, Dr Davinia
Fernández-Espejo, Julia Harrington, Dr Niki Kaiser, Mark Stow, Prof. Hillevi Lenz Taguchi, Sonia Thompson,
Prof. Sam Twiselton.
July 2021
Foreword 3
Executive summary 4
Introduction 9
1 Spaced learning 15
2 Interleaving 19
3 Retrieval practice 21
7 Embodied learning 42
Overarching findings 46
Glossary of terms 49
One of the most important As with much evidence, the key message here is
questions educational research the importance of nuance. Principles from cognitive
can ask is how children learn. science are neither myths to be discounted, nor silver
If we know how they process bullets that directly translate into accelerated progress.
and retain information, we can
adapt our approach to teaching There are still many questions to be answered on how
accordingly and in turn, principles from cognitive science can be applied in order
increase effectiveness. to make the biggest positive difference for all young
people. We need to know more about the effectiveness
At the Education Endowment Foundation (EEF), of approaches when used in different subjects, phases
our mission is to ensure that pupils, regardless of and their impact on disadvantaged pupils.
their background, are able to reach the full extent
of their potential. Developing our understanding of The future research work of the EEF can help fill
memory and how to balance cognitive load, and then some of these gaps. We will continue to work to
applying this understanding in the classroom, has the generate and accurately summarise evidence on the
potential to improve outcomes for all children. approaches that can make a difference for all pupils,
particularly the most disadvantaged.
This is why the EEF has produced this evidence
summary on the impact of cognitive science
approaches when applied in the classroom.
Our hope is that by providing a transparent
summary of the evidence that shows both the
strengths and weaknesses of the current research,
we can support schools as they consider how Professor Becky Francis
principles might make a difference to their pupils. Chief Executive
Findings from two areas of cognitive science have ‘It is, for example, becoming increasingly clear
been especially influential: cognitive psychology, that using spaced or distributed practice, where
which is underpinned by interpretive, behavioural, knowledge is rehearsed for short periods over a
and observational methods, and cognitive longer period of time, is more effective than so-
neuroscience, which is underpinned by brain imaging called massed practice.’
technologies. Many theories of effective learning have
been derived from these research areas, including: Our survey of teachers found that over 85% of
respondents said that cognitive science strategies
• spaced learning—distributing learning and retrieval were central to their own approach to teaching.
opportunities over a longer period of time rather In addition, all early career teachers will be taught
than concentrating them in ‘massed’ practice; about memory and cognitive load as part of the
Early Career Framework.
• interleaving—switching between different types of
problem or different ideas within the same lesson
or study session;
• Teachers looking to apply cognitive science • approaches that consider the balance between
principles in the classroom will need to consider didactic instruction and the pupil’s role in learning:
how, and in what conditions, approaches strategies for working with schemas, learner-
informed by cognitive science might improve led strategies within managing cognitive load,
learning. The effectiveness of strategies is likely cognitive theory of multimedia learning, and
to depend on factors including the age of embodied learning.
learners, learner prior knowledge, the nature of
the subject and learning outcomes, and whether Little research has effectively explored how distinct
the approach is practically feasible—and make and similar strategies interact but it is clear that
sure that they are successfully enacted. careful consideration is required to optimise mergers
between approaches, for example, combining retrieval
• Special care should be taken to make sure approaches with appropriate spacing.
that principles are successfully implemented,
avoiding ‘lethal mutations’ when a practice The survey of teachers, review of the underpinning
becomes disconnected from the theory. For science, and the applied evidence in areas such as
example, teachers have reported that dual coding embodied learning suggest that social, emotional and
sometimes means that irrelevant illustrations physical aspects to cognition are also important for
are added to presentations, which may be a teachers to consider.
distraction rather than a way of developing
schemas and optimising cognitive load.
• a fraction of the input enters the working memory 1. Learning requires information to be committed
(for example, 2 x 2 = 4); and to long-term memory.
• through attention and rehearsal, the information 2. Information is processed through the
moves from the working memory to the long- working memory.
term memory.
3. The working memory has limited capacity
Many of the strategies derived from cognitive science and can be overloaded.
focus on the crucial interactions between working
memory and long-term memory and the important
observation from cognitive science that our working
memories have limited capacity.
Rehearsal
Encoding
Input Sensory Attention Working Long term
(stimuli) memory memory Retrieval memory
R e h e ar s al
Forgetting
2. Failing to equip teachers to deliver the theory 4. Or… the theory doesn’t actually work
in schools?
The Ofsted inspection framework prompts senior
leadership to book some staff training on cognitive It might be that none or all of these factors are in
science. While the twilight session covers some of play and that the theory itself has only ever been
the principles behind spaced learning, staff are given shown to work in highly controlled laboratory
no guidance on how to apply the approach. Senior settings or for particular age ranges or subject
leadership are not keen on any changes to school areas. In medicine, treatments are often shown
timetabling or the curriculum. No follow-up training to reduce in efficacy as they move from lab-testing,
is provided. to larger trials, to the real world.
• Relations in the classroom (teacher-pupil, pupil-pupil) • Subject or curriculum area (e.g., general differences in
the nature of subject content and pedagogy)
• Culture of participation
• Nature of specific learning content (e.g., complexity/
• Emotional environment
element interactivity, novelty, connection with
• Disruption, noise, or distraction other learning)
• Decoration and information • Nature of specific learning activity (e.g., student-led,
length, structure, resources)
• Access to learning resources
This is not an exhaustive list. It is also important As with any evidence base, it is important to
to note that whether or not a specific strategy (for remember that lack of evidence or lack of quality
example, dual coding) works in practice may be evidence is not the same as evidence of no impact.
influenced by any one of these factors, and more likely It is also possible, despite the complexity of teaching
most of them. As we describe above, even applied and learning, to identify guiding principles and ‘best
studies rarely study much of this variation. The best bets’ for action. The evidence around cognitive
applied studies provide an authentic test of a strategy science is evolving quickly and areas that rely
in realistic conditions; they are designed to take these on lab-based studies or studies with researcher
into account and make sense of why and in what delivery today may be tested in realistic classroom
circumstances something works as well as providing environments tomorrow.
a strong test of whether or not it does. Unfortunately,
at present, well-designed applied studies are rare.
1. Spaced learning
2. Interleaving
3. Retrieval practice
4. Managing cognitive load
5. Working with schemas
6. Multimedia learning (including dual coding)
7. Embodied learning
Spacing across days or lessons might see pupils The evidence also highlights the value of considering
revisit a specific concept, idea, or topic several times how spacing can be informed and enhanced by
over the course of one week, or once or twice a classroom feedback and assessment. Teachers can
week for many weeks. The alternative to spacing is use assessment to decide how often learning material
usually referred to as ‘blocked’ or ‘massed’ practice, should be revisited and the best time to increase the
where content is studied in a single learning session. challenge or move on.
Spacing is frequently combined with retrieval practice,
often known as spaced retrieval practice.
What is the evidence based on?
What does the evidence say?
The review found 18 studies of spacing across
There are a significant number of studies showing that days and lessons.
spacing across days and lessons can have a small
• The ages of children in the studies ranged from
positive impact on learning outcomes.
6 to 17.
There are, however, some important limitations to the • Studies covered several different subject areas
evidence. Substantial variation is found between study including reading and vocabulary, science,
results with a significant number showing either no
and maths.
impact or a negative effect. Also, while spaced practice
principles were tested in real classrooms within these • In only six of the studies pupils were taught by
studies, the approaches were rarely delivered by teachers rather than researchers or computer
classroom teachers. Those that were designed and
packages. For these six, the training, guidance,
had the lesson materials provided by researchers.
or materials were provided by researchers.
An example approach
What is it?
What is the evidence based on?
Collaborative problem-solving activities are those in
which learners work together to complete a problem The review found nine studies that focused on the
or complex task. Collaborative problem-solving is a optimisation of cognitive load through collaboration
general teaching strategy and concept. The evidence when problem-solving.
here focuses on a subset of this, specifically designed
to optimise cognitive load. • Eight of the studies focused on learners aged
12 to 16. The other study took place with
children aged eight and nine.
Key things to consider compare, organise, and map concepts try to make
schemas clear and visible and are thought to support
• Teachers should consider pupils’ prior learners to organise and extend their ideas.
knowledge and how this might influence their
organisation of new knowledge and experience Working with schemas can prove challenging for
of misconceptions. teachers as schemas are formed in the minds of
pupils as they relate new content to prior knowledge:
• There is some evidence that concept mapping schemas are therefore shaped by every individual’s
and comparison can improve pupil learning, preconceptions, prior experience, and personal
although the variation in impacts suggests development of understanding.
challenges with implementation. Indicative
evidence suggests that it is how students engage Also, schemas are never fixed in their contents or
with information—and the extent to which they arrangements meaning that teachers are unlikely to
self-generate and organise information—rather know exactly how their pupils’ minds organise new
than any particular format, mapping, or organising content. However, the flexible nature of schemas does
information that matters. mean that specific teaching and learning strategies
may be able to foster a more desirable order to a
• When using concept mapping and comparison, learner’s pre-existing and developing knowledge.
it is important teachers know why they are
employing these strategies and have a plan for There are several approaches to working with schemas,
assessing pupils’ understanding to ensure key many of which focus on pupils organising, comparing,
conclusions have been taken from tasks. and elaborating on their ideas to develop more complex
mental structures. Some approaches include:
An example approach
After studying a text with her class, Ms Howarth uses her knowledge of the text and experience teaching
the same text with other classes to create knowledge organisers that collate the most crucial foundational
concepts and knowledge onto a single A4 page. These resources help pupils make links between ideas and
concepts, often grouping information by big overarching themes from the text, key quotes with annotations
on language devices, and relevant information on the social and historical context in which the text was
written. Ms Howarth often provides pupils with the organisers ahead of end-of-unit assessments to help
them self-assess existing knowledge and inform the writing of practise essays.
These ideas can, however, conflict with pupils’ preconceptions. Even if pupils know that the particles in
a gas have gaps between them, they often think that the space between them is full of other things such
as bacteria, pollutants, or oxygen.
Key things to consider Given the large variety of practice reviewed in this
area, we decided to frame the area using a slightly
• The evidence around the use of multimedia is broader theory: the cognitive theory of multimedia
mixed: while many studies are positive, others learning (CTML). This theory builds on ideas of dual
show low or null effects. coding, cognitive load, and generative learning (where
students actively integrate new ideas with existing
• Promising approaches use multimedia to support ideas). CTML describes organising, selecting, and
the learning of complex information, often with integrating multimedia information into coherent
older pupils. representations and combining them with prior
knowledge1. It has three core assumptions:
• Multimedia approaches can be challenging to
implement successfully. How are we to ensure that • there are two separate channels for information
the multimedia does not increase cognitive load? (that is, dual coding theory);
1 Mayer, R. E. (2005). Cognitive theory of multimedia learning. The Cambridge handbook of multimedia learning, 41, 31-48
What is it?
What is the evidence based on?
Using visual representations and illustrations
involves learners being presented with, or creating, The review found 34 studies relating to visual
an additional image, picture, or icon that symbolises, representation and illustration.
illustrates, or represents aspects of the content being
learnt. Studies in this group involved the presentation • There were no studies of Key Stage 1 or early
of additional visual information to a task or concept years, but a good representation of pupils aged
that could potentially be learnt without the visual. In 7 to 18.
some cases, the visual representation was provided,
in other cases produced by the learner. • Most of the studies took place in science
and maths, although other subjects were
What does the evidence say? represented, including geography, history,
and vocabulary learning.
While many of the impacts of the studies identified
are positive, results were mixed. Studies frequently • Most studies were not delivered in typical
reported no effect or harmful effects. The strategies school conditions: they were often small
included in this group were varied. It is likely that a interventions, delivered by researchers rather
regrouping that organised the studies by theoretically than teachers, and often delivered in one
important aspects (such as whether the images were short session.
created by, or provided to, students, or the information
content of the images) might have revealed a
clearer picture about the conditions in which visual
representations can support learning.
An example approach
What does the evidence say?
(Adapted from Coleman, McTigue and As with the wider evidence on visual aids, the evidence
on diagrams shows that while most studies have
Dantzler, 2018)
positive effects, many have null or negative effects.
When learning about the water cycle, pupils are
presented with diagrams with labels and process Similar to the evidence on visual aids, a key limitation
is that studies often were not delivered by teachers in
information that illustrate the different steps of the
usual classroom conditions. Most of the studies were
cycle. Evaporation is shown through an illustration delivered by workbooks or computer programmes.
of water rising from the sea to clouds, while a
ray from the sun shines onto the sea. An arrow Implementation
illustrates the direction of travel and a note explains
Diagrams are a subset of other visual representations
that the evaporation is caused by heat from the
and illustrations and many of the same lessons
sun’s energy and occurs in water in lakes, rivers, apply. One of the distinct factors in a diagram,
oceans, and on land. as opposed to an illustration, is considering how
concepts are structured or organised. It is important
that the visualisation does this accurately. It is
possible, for example, for inaccurate diagrams to
What is it? reinforce misconceptions.
In some studies, children are shown diagrams • Most of the studies were delivered by
to explain concepts while in others children are workbooks or computers.
encouraged to draw diagrams to summarise learning.
An example approach
Implementation
(Adapted from Barner et al., 2016, 2018) The limitations in the evidence base for spatial,
visualisation, and simulation approaches mean
When doing calculation tasks in maths, pupils are that there are limited insights on how teachers
taught to use the mental image of an abacus to might apply these approaches successfully in the
visualise the way the calculation takes place. As classroom. There is significant variation in how these
approaches have been applied. Teachers should
they complete sums, they imagine their mental
carefully consider how the visualisation strategy is
abacus and use it to support their calculations. being used, for example, is it inherent to the subject,
or is it a method of retrieval or scaffolding?
What is it?
Spatial, visualisation, and simulation approaches
support children to imagine learning content, or
representations of it, often in order to simulate,
manipulate, or organise concepts and schemas
across time or space. In some cases, visualisation
is inherent to the learning objective (as, for example,
in geometry); in others, it is used as a form of
retrieval and rehearsal (for example, imagining a
story); in other cases it is used as a scaffold for
problem-solving (that is, visualising as a way of
analysing or anchoring a learning object in memory).
What does the evidence say? What is the evidence based on?
There were only seven studies that focused on
While the studies identified have small to moderate spatial, visualisation, and simulation approaches.
positive effects, the limitations of the evidence
base make it difficult to assess the effectiveness • All of the studies focused on primary aged pupils.
of the strategies.
• All but one of the studies focused on mathematics.
The number of studies is small and there is real The final study focused on reading comprehension.
variation between the approaches that go beyond • Two of the studies were delivered by
practical differences in delivery.
regular teachers.
Embodied learning and physical factors refer to • incorporating physical activity into non-PE lessons;
strategies that engage and make use of movement
and the body to support effective learning. • breakfast clubs to increase nutrition; and
More generally, embodied learning often also
encompasses the role of emotions in learning—but • sleep hygiene education.
this is something we have not included in the scope
of this review. Most studies reviewed in this area Embodied learning approaches have most often
emanated from searches for visual representation been used in primary schools and use movement
and the coding of information. around the classroom while teaching other subjects,
for example, throwing beanbags to each other
In contrast to many of the other strategies explored in during counting tasks. Other examples are tracing,
this report that primarily focus on improving the mind’s the use of gestures when teaching, and play-based
ability to acquire, organise, and apply new knowledge learning approaches.
and skills, approaches that use the body and physical
movement to support teaching and learning work
under the assumption that the mind is closely related
to the body and its sensory experience.
The evidence in this area is consistently positive. This means that this evidence may not be a
Studies comparing embodied cognition approaches comprehensive summary of the approach.
were found to have larger impacts on learning than Fourteen studies were found in total.
normal curriculum delivery.
• The studies mostly focused on primary school
There are, however, a number of limitations to the children with ages ranging from 5 to 14.
evidence base. A key challenge in interpreting this
• There were studies of maths, science, and
evidence is that there may be missing studies as no
direct search of ‘embodied learning’ or similar terms language (including reading comprehension
took place. and vocabulary).
An example approach
When co-designing a new Year 8 curriculum, the head of English and Key Stage 3 lead think hard about the
sequencing of schemes of work and how key concepts and items of learning can be revisited over several
weeks (spaced learning) and can be built upon to support pupils in developing complex mental models
with a secure grasp of foundational knowledge (as seen in Working with Schemas). For example, when
considering how best to develop pupils’ ability to organise the structure of literary texts, they introduce some
learning around Freytag’s pyramid during the study of Macbeth at the start of the year before building on this
through a later scheme of work on the sequencing of sections in short story narratives.
To ensure that teachers are aware of the rationale for these curriculum design choices, the head of
English and KS3 lead provide time to develop and explain the curriculum during a departmental meeting.
Furthermore, to support teachers responsible for delivering the new curriculum, the English department also
make time to discuss the possible misconceptions pupils might experience while learning about narrative
structures and develop worked examples and scaffolds that will assist learners with differing needs.
• The evidence on mixed strategy programmes was It is thought that by increasing knowledge and use
hard to interpret as it was difficult to ascertain which of cognitive science strategies, teachers can harness
specific strategies were influencing pupil outcomes. the evidence base of how people think and learn to
support pupils more effectively.
• The seven programmes reviewed provided a
mixture of positive and null outcomes. Many of the
applied, mixed strategy programmes also suffered How do teachers implement
with implementation challenges.
mixed strategy programmes in
• It was unclear whether these challenges related the classroom?
to the organisational aspects of the programmes, Most of the mixed strategy approaches are
school improvement efforts more generally, or programmes that are implemented at the level of the
from the nature of the cognitive science strategies school, either through professional development or
in question. curriculum development.
Cognitive science principles of learning can only supported by a few studies that examine
have a significant impact on rates of learning their impact in everyday classroom conditions—
in the classroom. There is value in teachers delivered by teachers over long periods of time.
having working knowledge of cognitive
science principles. • With some approaches—like interleaving—there
are studies with promising results but they have
Theories from basic cognitive science imply almost exclusively been tested in one subject area
principles for effective teaching and learning. (mathematics). More generally, there are serious
Principles include ‘spacing’ learning out over gaps and limitations in the age ranges, subjects,
time, providing worked examples or ‘scaffolds’ to and learning outcomes studied for most of the
support problem-solving, and presenting information strategies we reviewed. These gaps make the
both verbally and visually. The applied evidence extent to which strategies apply across all age
summarised does provide support for many of the ranges and subject areas unclear.
principles of learning implied by basic cognitive
science, albeit in specific contexts. For most of • Some approaches—like combining verbal
the strategies included in this review, cognitive explanations with graphical representations,
science principles were significant factors affecting also known as ‘dual coding’—are possible to
rates of learning and retention of information in the implement poorly. While some studies show
classroom. Most of the results could be explained positive impacts on pupil outcomes, there are also
using theories from basic cognitive science and multiple studies showing null or negative findings.
practice-facing versions of these.
It is important to note that a lack of evidence is
The evidence for the application of cognitive not the same as evidence that an approach is not
science principles in everyday classroom successful. We should be cautious about concluding
conditions (applied cognitive science) is that because a principle is found to be ineffective in
much more limited, with uncertainties about the lab or in one classroom context that it cannot be
the applicability of specific principles across deployed effectively elsewhere.
subjects and age ranges.
Applying the principles of cognitive science
Applied cognitive science is far more limited and is harder than knowing the principles and one
provides a less positive, and more complex, picture does not necessarily follow from the other.
than the basic science. For many of the strategies, Principles do not determine specific teaching
the evidence was restricted to particular age groups, and learning strategies or approaches to
subject areas, or learning outcomes. Applications of implementation. Considering how cognitive
cognitive science outside of these, while plausible science principles are implemented in the
given the basic science, are yet to be tested and classroom is critical.
found effective in the classroom.
Some accounts of cognitive science principles
• Even approaches with indicative evidence of prescribe practice based on the strength of the
promise like retrieval practice, spaced practice, underlying theory and claim that the behavioural
and the use of worked examples are, as yet, or neuroscientific evidence justifies application
46
across subjects and ages (see previous). These are Principles of cognitive science interact
inevitably interpretations of how principles apply in and should not be considered in
the classroom. The fact that many strategies have isolation from each other, or without
not been consistently tested in everyday classroom taking into account wider practical
settings means that it is particularly important to and pedagogical considerations.
think carefully about if, and in what circumstances,
they are applicable and how we could implement There are clear links between the different
them in a way that has a positive impact on approaches summarised within this review.
pupil outcomes: In particular there are relationships between:
• Teachers looking to apply cognitive science • approaches that consider how to optimise retrieval
principles in the classroom will need to consider of information from the long-term memory: spaced
how, and in what conditions, approaches learning, interleaving, and retrieval practice;
informed by cognitive science might improve
learning. The effectiveness of strategies is likely • approaches that consider how to present
to depend on factors including the age of information effectively and support
learners, learner prior knowledge, the nature of students to work through information
the subject and learning outcomes, and whether without overloading working memory: dual
the approach is practically feasible—and make coding and multimedia theory of learning,
sure that they are successfully enacted. strategies to optimise cognitive load such
as scaffolding and use of worked examples,
• Special care should be taken to make sure that and embodied learning approaches; and
principles are successfully implemented, avoiding
‘lethal mutations’ when a practice becomes • approaches that consider the balance between
disconnected from the theory. For example, didactic instruction and the pupil’s role in learning:
teachers have reported that dual coding strategies for working with schemas, learner-
sometimes means that irrelevant illustrations led strategies within managing cognitive load,
are added to presentations, which may be a cognitive theory of multimedia learning, and
distraction rather than a way of developing embodied learning.
schemas and optimising cognitive load.
Little research has effectively explored how distinct
Schools should consider how—and in which and similar strategies interact but it is clear that
contexts—to give teachers high quality CPD around careful consideration is required to optimise mergers
cognitive science approaches, and enough time to between approaches, for example, combining
test and incorporate approaches appropriately into retrieval approaches with appropriate spacing.
their practice and for their subject and learners. Many
teachers report that their main form of engagement
with cognitive science is independent study.
47
Future work
Many of the classroom approaches inspired by
cognitive science are promising and knowledge
of cognitive science is an important part of
understanding how pupils learn. There is, however,
much more work to be done to understand in
which contexts findings from the basic science
are applicable and effective in everyday classroom
conditions, and how these can be implemented.
The EEF will continue to evaluate approaches that
provide scalable programmes or training for the
effective deployment of cognitive science principles.
We will also continue to fund teacher choices
trials, which seek to rigorously assess the non-
programmatic teacher techniques for deploying
cognitive science in the classroom.
48
GLOSSARY
49
50
Production and artwork by Percipio
https://percipio.london