A student can be strong at arithmetic and still feel lost when a shape rotates.
Another can solve coordinate questions but cannot imagine what a net will become when folded.
A third can manipulate algebra accurately yet struggles with graphs because the equation, table and picture never feel like the same object.
These learners may not have a single “mathematics problem.” They may be meeting mathematics that asks the mind to represent space.
Spatial reasoning is the ability to notice, imagine, transform and reason about the location, shape, orientation, size and relationships of objects or representations. It includes mental rotation, spatial visualisation, perspective taking, recognising structure inside diagrams, composing and decomposing shapes, interpreting maps and coordinate systems, and translating between two-dimensional and three-dimensional forms.
That sounds like geometry. It is bigger than geometry.
Mathematics repeatedly turns abstract relations into spatial forms. A graph places covariation in space. A number line turns numerical order and distance into position. Area models show multiplication. Bar models make comparison visible. Vectors encode magnitude and direction. Coordinate systems allow algebra to become geometry. Fractions can be represented as regions, lengths or locations. Even symbolic expressions have structure that learners may begin to “see” spatially as they become fluent.
This makes spatial reasoning an important bridge between perception and abstraction.
Current Education Endowment Foundation research gives the topic unusual practical relevance. The SPACE project — Spatial Cognition to Enhance Mathematical Learning — completed its pilot in September 2026. The programme investigates whether deliberate spatial-reasoning activities in Key Stage 2 mathematics can be integrated feasibly into ordinary lessons and whether larger-scale impact evaluation is warranted. The pilot is not an efficacy trial, so it should not be presented as proof of attainment gains. But it reflects a wider research signal: spatial thinking is sufficiently connected to mathematics learning that major evidence organisations are testing curriculum-integrated approaches rather than treating spatial ability as a fixed gift.
The educational question is therefore not “Are some children naturally visual?” It is:
Which spatial operations matter for mathematics, how can learners practise them, and when does a visual representation clarify structure rather than decorate the page?
The 50-second answer
Spatial reasoning helps mathematics by making relationships visible or mentally manipulable.
A learner may need to:
- rotate a shape mentally without rotating the page;
- imagine what a solid looks like from another viewpoint;
- compose small shapes into a larger one;
- decompose a complex figure into familiar parts;
- track position and direction through a transformation;
- move between a 3D object and a 2D drawing;
- understand scale and relative size;
- use a number line or graph as a spatial model of mathematical relationships;
- translate between diagram, language and symbols.
The mechanism is powerful because an external diagram or internal image can reduce what the learner must hold verbally. Instead of remembering six relationships in sequence, the learner can inspect them together.
But spatial representations can also mislead. A diagram may not be to scale. Perspective drawings hide lengths. A rotated shape can look “different” while preserving its properties. A graph can make a tiny difference look dramatic if axes are truncated. Visual fluency therefore requires both seeing and checking what the representation actually preserves.
Strong teaching does not ask learners merely to look at pictures. It makes them act on spatial information: rotate, build, compare, sketch, predict, describe, map, translate, decompose and justify.
First distinction: spatial reasoning is not the same as being a “visual learner”
The popular idea that each student has one preferred learning style — visual, auditory or kinaesthetic — is not a sound basis for matching all teaching to a supposed style.
Spatial reasoning is different.
It is not a personality label. It is a family of cognitive operations demanded by particular tasks.
A learner solving a cube-net question has to reason spatially whether or not they “prefer visual learning.” A student interpreting a distance–time graph must coordinate position on axes with a changing relationship. A child working with fraction strips uses spatial extent to represent numerical magnitude.
The educational response is therefore not:
“Mia is a visual learner, so give her pictures.”
It is:
“This mathematical idea has an important spatial structure. What representation or spatial action will help Mia learn that structure?”
That difference protects the learner from a fixed identity while preserving the value of visual and spatial teaching.
Stage 1: space becomes a representation for relationship
Why does a graph help?
Because position in space can stand for a mathematical relationship.
On a number line, farther right often means greater value. Distance between points represents numerical difference. Zero becomes an origin. Negative numbers occupy the opposite direction.
On coordinate axes, horizontal and vertical position jointly specify a pair of values. A straight line can embody a constant rate of change. Parallel lines can reveal equal slopes. Intersections can represent simultaneous solutions.
In a bar model, length represents magnitude. Equal lengths represent equal quantities. Segment structure exposes part–whole relationships.
The learner is not just looking at a picture. Space is being used as a coding system.
That coding system can reduce abstraction by turning an invisible relationship into something inspectable.
But the mapping must be taught. A child does not automatically know that a longer bar means a larger quantity, that equal intervals on a number line must represent equal numerical changes, or that the steepness of a graph encodes rate.
Every representation contains a legend, even if the legend is implicit.
Teachers need to make that mapping explicit:
“What does position mean here? What does distance mean? What does direction mean? Which visual feature carries the mathematics?”
Stage 2: mental rotation preserves identity across orientation
One of the simplest spatial reasoning tasks is also one of the most revealing: recognise an object after it turns.
Imagine the letter-like shape:
XX
X
X
Rotate it ninety degrees. It looks different on the page, but its internal structure is preserved.
Mental rotation requires the learner to change orientation while keeping identity stable.
This matters mathematically because students must learn that orientation often does not change properties.
A square remains a square when tilted. A triangle does not become a different class because its base is no longer horizontal. A vector can move in a diagram while maintaining magnitude and direction under translation. Congruent figures may appear in unfamiliar orientations.
Children can develop prototype traps if every rectangle is presented horizontally and every triangle has a base at the bottom. Then a rotated example feels like an exception.
Spatial reasoning expands category flexibility.
A useful classroom move is therefore orientation variation. Present shapes in non-standard positions. Ask which properties survive rotation. Use physical manipulation at first, then ask learners to predict before turning the object.
The sequence is:
see → turn physically → predict → turn mentally → justify by properties.
The goal is not just faster mental rotation. It is understanding what rotation changes and what it does not.
Stage 3: composition and decomposition reveal hidden structure
Spatial reasoning often means seeing one thing as several things — or several things as one.
A complex polygon can be decomposed into rectangles and triangles. Two right triangles can compose a rectangle. Fraction pieces can compose a whole. Algebra tiles can combine into larger rectangular arrays that represent products.
This operation matters because many mathematical problems become easier when a representation is restructured.
Consider an L-shaped floor plan. A learner may stare at the irregular shape and search memory for an “L-shape formula.” None exists.
A spatially flexible learner can see:
- two rectangles to add; or
- one large rectangle with a smaller rectangle removed.
Now familiar area formulas become usable.
The mathematics did not change. The partition changed.
This is a good example of why spatial reasoning connects with problem solving. The learner is not merely perceiving. They are choosing a decomposition that exposes known relationships.
Teachers can build this by asking the same figure to be decomposed in two ways, then comparing which is more efficient.
“Where could you draw one line that turns this into shapes you already know?”
That question teaches the mind to search for latent structure.
Stage 4: perspective taking changes viewpoint without changing object
Three-dimensional objects look different from different positions.
A tower of cubes viewed from the front may hide cubes that become visible from above. A building plan, elevation and side view encode the same object using different projections.
Perspective taking requires the learner to imagine how the scene changes when the observer moves.
This is useful in:
- plans and elevations;
- maps;
- 3D geometry;
- engineering drawings;
- coordinate transformations;
- spatial data visualisation.
Young learners often benefit from physical objects because the viewpoint change can be enacted. But physical manipulation should not remain the final support forever.
A useful progression is:
- look at a block structure;
- walk around it;
- draw the front and top views;
- predict a side view before moving;
- work from drawings without the physical structure;
- reconstruct the possible 3D object from views.
The final step reverses the mapping. That reversal is demanding and valuable.
Stage 5: scaling requires relational rather than absolute vision
Maps, models and diagrams often shrink or enlarge objects.
Scale reasoning asks the learner to preserve proportional relationships while absolute size changes.
This sounds routine once formulas are known. It becomes difficult when linear, area and volume scales interact.
If a square’s side length doubles, the square does not merely become “twice as big” in area. Its area becomes four times as large.
If a cube’s linear dimensions double, volume becomes eight times as large.
Spatial reasoning can support this by helping students imagine replication:
A square with doubled side can be partitioned into four original-sized squares.
A doubled cube can be visualised as eight original-sized cubes.
The image makes the exponent relationship tangible.
But the learner still needs symbolic generalisation. Visual insight should eventually become:
linear scale factor \(k\)
area scale factor \(k^2\)
volume scale factor \(k^3\)
The visual is a bridge to the abstract rule, not a replacement for it.
Stage 6: coordinate systems turn movement into mathematics
Coordinates are spatially elegant because they turn location into ordered number.
A learner can describe:
- position;
- displacement;
- reflection;
- rotation;
- translation;
- gradient;
- distance;
- vector direction.
all through relationships between numbers and space.
But coordinate fluency depends on understanding the mapping.
A common error in transformations is procedural memory without spatial meaning: “change x to negative x” is remembered for reflection, but the learner cannot predict which axis the reflection crosses.
A better spatial route asks:
What stays the same distance from the mirror line?
Which coordinate changes sign?
Which coordinate stays fixed?
Now the symbolic rule grows from spatial constraint.
This is a recurring theme: spatial reasoning becomes mathematically powerful when visual behaviour and symbolic rule are linked in both directions.
Stage 7: graphs require spatial interpretation of covariation
Graphs look visual, but reading them well is not trivial visual perception.
Students must coordinate several mappings:
- horizontal position ↔ one variable;
- vertical position ↔ another variable;
- slope ↔ rate of change;
- shape ↔ changing relationship;
- intercept ↔ boundary or initial condition;
- area under a curve ↔ another quantity in some contexts.
Misreading occurs when the learner treats the graph as a literal picture.
A common example is a distance–time graph that rises and falls. Students may say the traveller “went up a hill” because the line rises on the page. They have interpreted spatial shape literally rather than as a representation of changing distance.
This is why graph teaching should ask both:
What does this shape look like?
and:
What does this spatial feature encode mathematically?
Spatial reasoning alone is not enough. The learner needs representational conventions.
Stage 8: 2D–3D translation is a reversible mapping
Nets are a classic example.
A cube net is flat. The cube is three-dimensional. The learner must imagine folds and preserve adjacency.
The key is not simply remembering which of eleven cube nets are valid.
The deeper operation is translating between dimensions.
Ask learners to:
- predict which faces will touch;
- mark opposite faces before folding;
- identify which edge pairs meet;
- build the solid physically;
- unfold it mentally again.
The reverse direction matters. If learners only fold provided nets, they may never learn to generate a net from the solid.
Physical manipulatives can be excellent here, but use them as temporary cognitive tools. The goal is eventually to internalise enough of the transformation that the learner can reason without the object in hand.
Spatial reasoning and number: the number line is not childish
Number lines are sometimes treated as an early-primary scaffold to be discarded once “real mathematics” begins.
That is a mistake.
The number line is a profound spatial model.
It represents:
- order;
- magnitude;
- distance;
- opposites;
- density;
- intervals;
- inequalities;
- absolute value;
- rational and irrational numbers.
For negative numbers, spatial position helps counter the misconception that “a bigger digit means a bigger number.” -8 lies left of -3.
For fractions, placing 3/5 between 1/2 and 2/3 reinforces magnitude rather than seeing fractions only as shaded pizza slices.
For inequalities, interval notation and solution sets become spatial.
For absolute value, distance from zero becomes visible.
The representation grows with the mathematics.
Spatial reasoning and algebra: structure can be seen
Advanced students often begin to describe algebra visually.
They “see” common factors.
They recognise a quadratic as a familiar shape.
They notice symmetry in an expression.
They treat terms as chunks rather than strings of symbols.
This is not necessarily literal mental imagery. It is structural perception supported by years of spatial and symbolic organisation.
Algebra tiles make this visible for beginners.
An expression such as:
\(x^2 + 5x + 6\)
can be represented as an area composed of one \(x^2\) tile, five \(x\) tiles and six unit tiles. Rearranging into a rectangle reveals dimensions \(x+2\) and \(x+3\), connecting factorisation with area.
This representation is not required for every learner or every problem. Its value is conceptual: it spatialises multiplication and factor structure.
Later, the learner should be able to move back into symbols efficiently.
Spatial reasoning and proof
Diagrams can suggest mathematical truths before formal proof establishes them.
A geometric diagram may make a theorem feel obvious. That feeling is useful and dangerous.
Useful because visual structure can generate conjectures.
Dangerous because a diagram is usually one case and may contain accidental features.
A triangle may look isosceles when it is not specified as isosceles. Lines may appear perpendicular. Lengths may look equal.
Therefore spatial reasoning must eventually pair with deductive discipline.
Ask:
- Which features are given?
- Which are only drawn?
- What would still be true if the diagram were distorted?
- Which relationship follows from a theorem rather than appearance?
A strong mathematician uses the picture to think and the proof to justify.
Why physical manipulation can help — and when it becomes a crutch
Blocks, tiles, paper folding, geoboards, tangrams and dynamic geometry software let learners manipulate structure externally.
This can reduce mental load.
A child who cannot yet rotate a shape mentally can turn the physical shape. A student learning symmetry can fold paper. A learner building a 3D solid can inspect hidden faces directly.
The physical action is valuable when it teaches an internal operation.
The sequence should therefore move:
act → predict → act to check → imagine → justify.
If the learner always needs to rotate the paper, mental rotation never has to develop. If the learner always needs algebra tiles, symbolic factorisation may remain dependent on the material.
Scaffolds should make cognition possible and then gradually transfer the work inward.
Why sketching matters even when the drawing is ugly
Some students refuse to sketch because they are “bad at drawing.”
Mathematical sketching is not art.
A good mathematical sketch may contain crooked lines and still be excellent if it preserves relevant relationships.
Teach students to sketch for information:
- mark known lengths;
- show direction;
- label angles;
- indicate equal parts;
- show containment;
- place variables.
The sketch is a thinking surface.
This is another reason spatial reasoning belongs in ordinary mathematics, not only specialist geometry enrichment.
Common failure mode: realistic diagrams that hide the mathematics
Visual material can be too visually rich.
A realistic picture of a swimming pool may contain reflections, people, tiles and perspective. The mathematical problem concerns a rectangular border.
The realism creates irrelevant detail.
A simplified rectangle may support reasoning better.
Good representations preserve the structure needed for the task and remove distraction.
This is the same principle engineers use in schematics. A circuit diagram does not look like the physical circuit because resemblance is not its job.
Common failure mode: decorative visualisation
Teachers may add pictures because “visuals help.”
Some visuals do. Some consume attention.
A picture is mathematically useful when the learner can answer:
What information does this visual encode that matters to the problem?
If the answer is “none,” the picture may be decoration.
Decorative images are not always harmful. They may increase interest. But do not confuse interest with representational support.
Common failure mode: one canonical orientation
If every parallelogram leans right, every triangle points up and every cylinder stands vertically, learners may attach category knowledge to orientation.
Vary examples.
Rotate them.
Stretch them while preserving defining properties.
Include non-examples that look similar.
Ask which features define the object and which features are incidental.
This builds visual discrimination and conceptual accuracy together.
Common failure mode: asking for mental imagery too early
“Just imagine the cube rotating” is not useful to a learner who cannot yet perform the transformation.
Start externally.
Turn the cube.
Narrate what remains facing the observer and what moves.
Pause halfway.
Ask the learner to predict the next face.
Gradually remove the object.
Spatial skills can be taught through graduated representational demands.
Common failure mode: visual answers without mathematical language
A student points to a graph and says, “It goes like that.”
The visual intuition may be correct. It needs language.
Translate:
“The gradient is positive and increasing.”
“The rate of change decreases.”
“The function is symmetric about this line.”
“The shape has rotational symmetry of order four.”
Mathematical education should connect spatial perception with precise terminology and symbols.
The learner needs to move both ways:
see → say → symbolise
and:
symbolise → say → visualise.
Common failure mode: treating spatial ability as fixed
A student says, “I can’t visualise things.”
That may describe current difficulty. It should not automatically become identity.
Spatial reasoning includes multiple component skills, and many can improve through practice. Mental rotation, spatial language, construction, perspective and visualisation are trainable to meaningful degrees, although individuals differ.
A teacher can therefore diagnose which operation is difficult rather than label the whole learner “non-visual.”
Can the student compare shapes?
Rotate them?
Remember locations?
Translate a plan into a solid?
Read axes?
Decompose a figure?
The problem becomes teachable when it becomes specific.
Spatial language is a tool
Words support spatial thinking.
Above. Beneath. Parallel. Perpendicular. Adjacent. Opposite. Clockwise. Between. Through. Across. Interior. Exterior. Horizontal. Vertical. Diagonal. Rotate. Reflect. Translate. Scale.
Students who lack this vocabulary may understand a spatial relation but struggle to hold or communicate it.
Teachers can pair action with language:
“Rotate the shape ninety degrees clockwise.”
“Reflect it across the vertical axis.”
“Which face is opposite the marked face?”
Precise language becomes an external control system for mental transformation.
This can be especially important for multilingual learners: do not mistake unfamiliar English spatial vocabulary for absence of spatial understanding. Show, act and name.
How spatial reasoning interacts with working memory
A spatial representation can offload information from working memory.
Imagine hearing:
A is north of B. C is east of B. D is south of C. E is west of D.
Holding all relationships verbally is possible but effortful.
A quick sketch externalises them. Now the learner can inspect the configuration.
This is why diagrams can improve problem solving even when the learner could technically solve without them.
But the reverse is also important: complicated diagrams can overload visual working memory. Good design simplifies.
The issue is not “visual is easier.” It is which representation distributes cognitive load most usefully for this task?
How spatial reasoning interacts with mathematical heuristics
Many classic problem-solving heuristics are spatial.
Draw a diagram.
Make a table.
Plot a graph.
Rearrange the figure.
Decompose the shape.
Change coordinates.
Our neighbouring article How Mathematical Heuristics Work | Why a Good Problem-Solving Move Is Not a Formula treats these as search strategies. The present owner goes deeper into the capability that lets those strategies work: reasoning through spatial representation and transformation.
A heuristic says, “Try drawing it.”
Spatial competence determines whether the drawing reveals anything.
Evidence, caveats and the SPACE project
The wider research literature has repeatedly found associations between spatial ability and mathematics performance, and experimental work has suggested that aspects of spatial thinking can be trained. However, translating improved performance on spatial tasks into durable gains across mathematics is not automatic. Near transfer is easier than far transfer.
This is why current curriculum-integrated research matters.
The EEF-funded SPACE pilot, completed in September 2026, examined whether spatial-reasoning activities could be implemented feasibly within the mathematics curriculum in Year 3 classrooms. The programme includes five broad spatial skill areas and teacher professional development. The pilot’s purpose was feasibility, implementation and evaluation design, not a definitive test of pupil attainment impact.
That distinction is essential.
We should not write:
“The SPACE project proved spatial training improves maths.”
It did not.
We can responsibly say:
Spatial cognition is a plausible and actively researched pathway for supporting mathematics learning, and current evidence organisations are testing integrated approaches at classroom scale.
Other caveats matter too.
Spatial reasoning is not one skill. Mental rotation, visualisation, perspective taking and spatial scaling are related but distinct.
Mathematics is not reducible to spatial ability. Language, number knowledge, symbolic fluency, reasoning, prior instruction and many other factors matter.
Visual representations can mislead. Learners need conventions and checking.
Individual differences are real. Training does not make every learner identical.
Transfer must be designed. If learners practise rotating shapes but never connect rotation to geometry, gains may remain narrow.
If you are a learner
When a mathematics problem feels invisible, ask whether space could carry some of the thinking.
Can you draw it?
Place values on a line?
Plot a few points?
Build it with cubes?
Rotate the page?
Then, importantly, rotate it back and try to do the transformation mentally.
Use sketches even if they are ugly. Label relationships.
When looking at a diagram, ask what is given and what merely appears true.
When using a graph, translate every important visual feature into mathematical language.
When you cannot imagine a 3D object, use a physical model first — then predict before touching it.
Your goal is not to become an artist.
Your goal is to make structure inspectable.
If you are a parent
Spatial reasoning is one of the easiest areas to support without turning home into another worksheet centre.
Build with blocks.
Follow maps.
Discuss routes.
Fold paper.
Estimate whether furniture fits a space.
Read plans.
Rotate puzzle pieces.
Cook with scaling and container size.
Ask:
“What will this look like from the other side?”
“How could we split this shape?”
“Which piece would fill the gap?”
“If every length doubles, what happens to the area?”
Keep the activity conversational. The value lies in describing, predicting and checking spatial relationships.
If a child struggles with diagrams, do not immediately say, “You are not visual.” Find the specific operation that is difficult and externalise it.
If you are a teacher
Audit your mathematics curriculum for spatial operations.
Where do students need to:
- rotate;
- reflect;
- compose;
- decompose;
- scale;
- change viewpoint;
- translate 2D ↔ 3D;
- read spatial mappings such as axes;
- use diagrams to reason?
Then teach those operations deliberately.
Vary orientation.
Ask learners to predict before manipulating.
Use spatial language precisely.
Compare representations.
Ask what each representation preserves and distorts.
Move gradually from concrete objects to drawings to mental and symbolic representations.
Do not add visuals merely because the page looks empty.
A strong visual has a mathematical job.
A classroom mini-sequence: from cube to imagination
Here is a short progression for nets.
Round 1: Give students a cube and a net. Fold it physically.
Round 2: Mark one face red. Ask students to predict which face becomes opposite it. Fold to check.
Round 3: Show a different net without a physical cube. Students mark predicted opposite pairs.
Round 4: Students explain how they tracked adjacency.
Round 5: Give one invalid net and ask where overlap occurs.
Round 6: Ask students to create a new valid net.
The sequence moves from perception to prediction to construction.
That is more powerful than memorising a catalogue of nets.
A classroom mini-sequence: the graph is not a mountain
Show a distance–time graph with a rising line, horizontal segment and downward line.
Ask learners to sketch the physical path they think the traveller took.
Many will draw a hill.
Now ask:
“What does vertical position actually represent?”
Distance from the start.
“What does a horizontal segment mean?”
Distance is unchanged, so the traveller is stationary.
“What does the downward segment mean?”
The traveller is getting closer to the starting point.
The lesson is representational: the graph’s spatial shape is a code, not a literal route.
Then reverse it. Tell a story of movement and ask students to create the graph.
Spatial reasoning becomes bidirectional.
A Clementi-floor test: does the learner still understand after the picture changes?
Use one concept in multiple spatial forms.
For fractions:
- area model;
- number line;
- set model;
- ratio table;
- symbol.
For linear relationships:
- story;
- table;
- graph;
- equation;
- diagram.
Ask the learner to translate.
If the concept disappears whenever the representation changes, knowledge is still attached to the picture.
If the learner can preserve the relationship across forms, representational competence is developing.
Frequently asked questions
Is spatial reasoning mainly geometry?
No. Geometry makes spatial operations obvious, but number lines, graphs, algebraic representations, scaling, vectors, functions and many problem-solving strategies also depend on spatial structure.
Can spatial reasoning be improved?
Research suggests that many spatial skills can improve with practice. The important educational question is whether that improvement transfers into the mathematical tasks we care about. Curriculum-integrated approaches are being actively studied.
Does strong spatial ability guarantee strong mathematics?
No. Mathematics also depends on numerical, symbolic, linguistic and conceptual knowledge, instruction, practice and many other factors.
Should students rotate paper in geometry?
Yes when it helps them access the problem, especially early on. But sometimes ask them to predict the rotated state before physically turning the paper so mental rotation develops too.
Are manipulatives always helpful?
No. They help when learners understand how features of the object map to the mathematics. A manipulative without explicit connection can become another thing to handle.
Why do some students misread graphs?
They may interpret the spatial shape literally rather than understanding what axes and visual features encode. Graph literacy requires learning the mapping between space and variables.
How is spatial reasoning different from visualisation?
Visualisation is one part of spatial reasoning. Spatial reasoning also includes relation, transformation, orientation, scaling, perspective, composition and interpretation of external spatial representations.
What is one easy way to strengthen spatial reasoning in maths?
Ask learners to predict a transformation before they physically perform it, then compare prediction with result and explain what stayed invariant.
Should visual representations replace symbols?
No. Strong mathematical learning connects them. Visual forms can build meaning; symbols provide precision and efficiency. Learners should move between the two.
What if a child says they cannot picture things in their head?
Use external representations first. Some people also differ substantially in mental imagery vividness. Mathematics does not require cinematic internal pictures; diagrams, language and symbolic relationships can all support spatial reasoning.
The deeper idea: mathematics gives space a second job
In ordinary life, space tells us where things are.
Mathematics recruits space to tell us something more.
Distance can represent numerical difference.
Position can represent value.
Slope can represent rate.
Area can represent multiplication.
Orientation can test whether a property survives transformation.
A folded net can connect two dimensions to three.
A graph can make an invisible relationship visible all at once.
Spatial reasoning becomes mathematically powerful when learners understand those mappings and can manipulate them deliberately.
That is why the goal is not “more pictures”.
It is better use of space as a thinking system.
A learner who can turn, build, decompose, map, scale and translate representations gains more than geometry tricks. They gain another way to hold structure still long enough to reason about it.
Sources and further reading
- Education Endowment Foundation. Spatial Cognition to Enhance Mathematical Learning (SPACE) – pilot. Completed September 2026. https://educationendowmentfoundation.org.uk/projects-and-evaluation/projects/spatial-cognition-to-enhance-mathematical-learning-space-pilot
- Education Endowment Foundation. Improving Mathematics in Key Stages 2 and 3. https://educationendowmentfoundation.org.uk/education-evidence/guidance-reports/maths-ks-2-3
- Institute of Education Sciences, What Works Clearinghouse. Improving Mathematical Problem Solving in Grades 4 Through 8, including recommendations on visual representations. https://ies.ed.gov/ncee/WWC/PracticeGuide/16
- eduKateSG. How Representational Competence Works | Read, Choose and Translate the Forms Knowledge Takes
- eduKateSG. How Mathematical Heuristics Work | Why a Good Problem-Solving Move Is Not a Formula