How Representational Competence Works | Read, Choose and Translate Representations

eduKateSG Learning Node Series · 0129

A student can know the formula and still not understand the graph. They can read the graph and still fail to explain the diagram. They can recognise the diagram and still choose the wrong representation when the problem changes.

This is one of the quiet reasons a learner can appear knowledgeable in one format and suddenly become lost in another. School subjects do not live only in sentences. Mathematics uses equations, tables, graphs, geometric figures and diagrams. Chemistry moves between observable phenomena, particle models, formulae and equations. Geography uses maps, cross-sections, satellite imagery, statistics and prose. Science uses models, data displays and causal diagrams. English uses text, structure, annotation and sometimes visual or multimodal evidence.

Representational competence is the capability to work intelligently across those forms. It includes knowing what a representation shows, what it hides, how it relates to the underlying phenomenon, when it is useful, how to construct it and how to move from one representation to another without losing the structure that matters.

Representational competence works when the learner stops treating graphs, equations, diagrams and models as separate school objects and starts using them as different instruments for seeing the same underlying structure.

The 50-Second Read

Representations are tools for thinking, communicating and acting on things that may be difficult or impossible to inspect directly.
Competence is more than decoding a symbol. It includes selecting, generating, comparing, integrating and transforming representations.
Experts often read structural relationships where novices notice surface features.
A learner may understand a concept in one representation but fail when the same concept appears in another.
Translation between forms is therefore a powerful test of understanding.
Multiple representations help only when learners understand the relation between them; more diagrams can otherwise create more load.
Every representation has affordances and limits: it makes some features visible and suppresses others.
Strong teaching explicitly connects representations instead of assuming students will infer the links.
Current chemistry-education research continues to treat representational competence as central to expert reasoning, especially when learners must coordinate molecular, symbolic and macroscopic forms.
The destination is representational choice: knowing which form makes the current problem easier to see.

Canonical Owner Boundary

This Learning Node owns the learner capability to interpret, construct, choose, connect and translate representations across a domain. How Cognitive Flexibility Theory Works owns advanced learning through multiple cases, perspectives and representations in ill-structured domains. How to Think Properly | Change the Representation When the Problem Will Not Open owns the tactical move of switching representation during problem solving. This page owns the deeper literacy that makes such switching possible in the first place.

1. Knowledge Does Not Arrive Naked

A relationship in the world is almost always mediated by a form.

Velocity may appear as a sentence, a number with units, a slope on a graph, an arrow in a vector diagram or a variable in an equation. Population change may appear as a table, age pyramid, map or time series. A chemical reaction may appear as a colour change in a beaker, a symbolic equation, a molecular diagram or an energy profile.

The underlying phenomenon is not identical to any one of these. Each representation is a deliberately selective interface.

Learning a subject therefore means learning its objects and learning the representational system through which the discipline makes those objects thinkable.

2. A Representation Is a Tool With an Agenda

Representations foreground some relationships and background others.

A map preserves spatial relationships while discarding most visual detail. An equation compresses a relationship among quantities while hiding the concrete story. A diagram may reveal parts and connections while ignoring scale. A bar chart makes comparison easy but may conceal individual variation.

Representational competence begins when the learner asks not only “What does this show?” but “What is this form good at showing?”

3. Decoding Is the First Layer, Not the Last

Novices first need basic decoding. What does the x-axis mean? What does the arrow denote? What is the unit? What does the coefficient in front of a chemical formula mean?

But expert use goes further. The learner should infer relationships, test implications, compare cases and use the representation to decide what to do next.

Reading a graph is not merely naming axes. It is seeing rate of change, turning points, intervals, proportionality, outliers and uncertainty where relevant.

Competence turns notation into reasoning.

4. The Research Idea: Representational Competence

Representational competence has been developed especially strongly in science and chemistry education. Work associated with Richard Kozma and Joel Russell describes it as a set of skills and practices that allow people to use multiple representations reflectively to think, communicate and act on phenomena that may not be directly perceptible.

Later research extends the idea beyond simple symbol reading. Representationally competent learners can interpret representations, connect them to disciplinary concepts, transform among them, construct new representations and evaluate which form is appropriate for a purpose.

A 2017 review by Prajakt Pande and Sanjay Chandrasekharan described representational competence as critical in science, mathematics and engineering because external representations often stand in for entities and processes that cannot be directly perceived. Their account emphasises integration between external representations, internal imagination and operations performed on those representations.

5. Chemistry Makes the Problem Easy to See

Chemistry is a near-perfect example because learners must coordinate several levels at once.

You may observe a precipitate forming. You may describe particles rearranging. You may write a balanced equation. You may inspect a molecular model. You may analyse an energy graph.

These are not five unrelated facts. They are five windows on one event.

A 2026 Journal of Chemical Education paper on students’ difficulties reading and drawing chemical structures continues to frame representational competence as the ability to use representations purposefully for inquiry, communication and judgment. A 2026 JACS Au study likewise describes underdeveloped representational competence as encouraging surface-feature reasoning and algorithmic symbol use, while stronger competence supports explanation, prediction and problem solving.

6. Novices Often Read the Surface

Experts see meaning through the marks. Novices can become trapped by the marks themselves.

A novice sees a graph with a steep line and thinks “big.” An expert asks whether the slope represents a rate and which quantities define it. A novice sees two equations with different letters and treats them as different kinds of problem. An expert sees the same relational structure.

This difference matters because school tasks frequently change surface appearance. If learning is attached to colours, positions or familiar templates, transfer becomes fragile.

7. Structural Reading Is the Upgrade

Representational competence moves attention from decoration to relation.

On a graph: which variable changes with which? In an equation: what is invariant? In a diagram: which arrows represent causal influence and which merely label movement? In a map: what spatial relation is preserved? In a table: which comparison is possible because values share a common unit?

The learner starts asking what information architecture the representation carries.

8. Translation Is a Stress Test for Understanding

If you really understand a relationship, you should often be able to express it in more than one valid form.

Take a linear relationship. Can you move from a table to a graph? From graph to equation? From equation to a sentence explaining rate and intercept? From sentence to a real-world example?

Translation exposes hidden gaps because each move requires preserving structure while changing surface form.

A student who can calculate from an equation but cannot recognise the same relationship in a graph may possess procedural control without representational integration.

9. Translation Is Not Mechanical Substitution

Students sometimes learn conversion routines without understanding.

“Find two points, plot them, draw the line.” This can produce a graph. It does not guarantee that the learner sees how slope, intercept and equation relate.

A stronger task asks what feature of the equation becomes visible as slope, what the intercept means in context, and how changing one coefficient transforms the graph.

The translation should preserve meaning, not merely format.

10. Multiple Representations Are Not Automatically Better

Teachers often assume that adding a diagram to text must improve learning.

Sometimes it does. Sometimes it creates a second decoding problem.

If students do not understand how the diagram maps onto the text, they must process two systems and invent the connection. This can increase rather than reduce cognitive load.

The instructional job is therefore not simply to provide multiple representations but to teach the relation among them.

11. Ainsworth’s Multiple-Representation Problem

Shaaron Ainsworth’s DeFT framework for learning with multiple representations highlighted that representations can serve different functions: they may complement one another, constrain interpretation or help learners build deeper understanding.

The important point is functional. Two representations should not be added merely because two look richer than one.

Ask what the second representation contributes. Does it make an invisible relationship visible? Does it prevent a predictable misreading? Does it force integration?

12. Every Representation Has Affordances

An affordance is what a form makes easier to see or do.

A graph affords visual comparison of trends. An equation affords exact manipulation. A diagram affords spatial organisation. A table affords lookup and precise comparison. Prose affords qualification, causal explanation and context.

Representational competence includes matching affordance to question.

If the question is “Where is the maximum?”, a graph may outperform an equation. If the question is “What exact value occurs at x = 3?”, an equation may be better. If the question is “Why does the relationship hold?”, prose plus diagram may be necessary.

13. Every Representation Also Has Blind Spots

A map that preserves scale may omit social meaning. A statistical average may conceal distribution. A molecular model may suggest rigid balls where quantum descriptions are more complicated. A flowchart may imply linear sequence when real processes loop.

Strong learners ask what information was discarded when the representation was created.

This becomes especially important when representations acquire authority. A chart looks objective. A model looks scientific. A dashboard looks complete.

No representation is the world. It is an engineered view of the world.

14. Constructing a Representation Is Different From Reading One

A student may understand a finished diagram yet struggle to create one.

Construction requires deciding what to include, what to omit, how to encode relationships and how to arrange the elements so another person can read them.

This is why drawing, graphing, modelling and diagramming can be cognitively productive when the learner must make representational decisions rather than merely decorate notes.

15. Choosing a Representation Is a Higher-Level Skill

Many school tasks tell students the form in advance: “Draw a graph.” “Write an equation.” “Complete the table.”

Real problems often do not.

A representationally competent learner asks: What would make the structure easiest to inspect? Should I sketch the situation? Create a table? Write a symbolic relationship? Build a timeline? Draw a causal diagram?

Choice is where representation becomes strategy.

16. Mathematics: Learn the Same Relationship in Several Languages

Mathematics routinely asks learners to coordinate verbal, symbolic, numerical and graphical forms.

Consider direct proportion. A learner should recognise it in a sentence, table, graph and equation. They should know what feature remains invariant and what feature changes when the constant of proportionality changes.

If one form is always taught first and another is treated as a separate topic, students may never build the bridge.

17. Mathematics: Draw Before You Calculate

Word problems often become solvable once the learner chooses a useful representation.

A ratio table can expose multiplicative structure. A bar model can expose part-whole relationships. A coordinate sketch can expose geometry. A number line can reveal signed movement.

The drawing is not extra work if it reduces the search space of the problem.

18. Additional Mathematics: Representation Determines Method

Advanced mathematics frequently rewards the learner who changes form before calculating.

A trigonometric expression may simplify after rewriting it in another identity. A coordinate problem may become easier after representing a line parametrically. An algebraic relationship may become visually obvious when graphed.

Method selection is often representational selection in disguise.

19. Physics: The Free-Body Diagram Is a Thinking Device

A free-body diagram is not a picture of the object. It is a selective representation of forces acting on it.

Students who treat the diagram as art may draw motion arrows as forces or omit interactions that are not visually obvious.

Competence means knowing the representational rules: what belongs, what does not, and how the diagram connects to equations of motion.

20. Chemistry: Move Between Macroscopic, Submicroscopic and Symbolic

A learner sees magnesium burn. At the macroscopic level, there is bright light and a white product. At the particle level, atoms rearrange with oxygen. At the symbolic level, the reaction becomes a balanced chemical equation.

Mastery means coordinating these, not memorising each independently.

Recent chemistry education research continues to show that difficulty translating among representations is a genuine learning barrier, not merely a cosmetic notation issue.

21. Biology: Scale Changes the Representation

Biology moves between organism, organ, tissue, cell, molecule, population and ecosystem.

A diagram of a heart, a pressure graph and a cellular account of muscle contraction operate at different scales. Learners need to understand not only each representation but how explanations connect across levels.

Representational competence includes knowing when a shift in scale has occurred.

22. Geography: Maps Are Arguments About Space

Maps feel neutral because they look technical.

But every map chooses projection, scale, categories, colours and boundaries. A choropleth map can make one pattern obvious while hiding population density. A transport map may distort distance deliberately to emphasise connectivity.

Representational competence means using the map while understanding the choices embedded in it.

23. English: Structure Can Be Represented Too

Representational competence is not only a STEM idea.

An argument can be mapped as claim → evidence → warrant → counterclaim. A narrative can be represented as timeline, tension curve or scene sequence. A sentence can be diagrammed to expose clause relationships.

The point is not that every English task needs a diagram. The point is that changing representation can make hidden structure visible.

24. History: Timelines Are Useful and Dangerous

A timeline is excellent for chronology. It is weak at representing multiple causal layers unless deliberately extended.

Students may mistake temporal sequence for causation because events that appear next to each other look causally connected.

Use different representations for different historical questions: timeline for sequence, causal network for interaction, comparison table for cases, map for spatial constraint.

25. Economics: A Curve Is a Model, Not a Photograph

Supply-and-demand diagrams are powerful precisely because they remove most of reality.

The learner must know what is held constant, what shifts a curve, what movement along a curve means and which real-world complications sit outside the representation.

Representational competence includes resisting the temptation to treat a simplified model as a literal picture of the economy.

26. Statistics: Graph Choice Can Change the Story

A mean, box plot, histogram and scatter plot answer different questions.

Students who ask only “Which graph am I supposed to draw?” remain dependent on task wording. Students who ask “What relationship am I trying to see?” are beginning to choose representations strategically.

That distinction matters in the real world because misleading data displays often exploit representational weakness.

27. Representational Competence Is Also Verification

When two representations disagree, something useful has happened.

An equation predicts one trend but the graph shows another. A verbal explanation contradicts the diagram. A table total does not match the stated percentage.

Instead of treating the mismatch as annoyance, use it as an error detector.

Cross-representation consistency checking is a powerful way to catch mistakes because one form can expose assumptions hidden in another.

28. AI Raises the Stakes

Generative systems can produce graphs, diagrams, tables, equations and fluent explanations rapidly.

The ability to verify whether these forms actually correspond becomes essential. A beautiful graph can use the wrong scale. An equation can be dimensionally inconsistent. A diagram can imply a causal relation the text does not support.

Representational competence becomes one layer of AI literacy: do the forms agree with the phenomenon and with one another?

29. Teach the Legend, Then Teach Beyond the Legend

Beginners need explicit conventions.

What do arrows mean? What do dashed lines mean? What counts as a node? Which axis is independent? How is uncertainty shown?

But once conventions are known, teaching should move toward interpretation: what relationship does this represent, why was this form chosen, and what would change if we represented the same situation differently?

30. Model Expert Translation Out Loud

Teachers often switch representations so fluently that learners miss the reasoning.

Say the bridge aloud: “This steep section of the graph corresponds to the large rate in the equation.” “This arrow in the diagram becomes the negative sign here because we defined the opposite direction as positive.” “This phrase in the passage is evidence for the relationship drawn in this causal map.”

The bridge is part of the lesson.

31. Ask Learners to Explain Why a Representation Was Chosen

Do not only ask students to produce a graph.

Ask why a graph is more useful than a table for this question. Ask what the diagram makes visible that prose does not. Ask what would be lost if the equation were replaced by a verbal description.

These questions teach the meta-level: representations have jobs.

32. Use Representation Pairs

Show the same phenomenon in two forms and ask learners to map features explicitly.

Which part of the equation corresponds to this graph feature?
Which sentence in the text corresponds to this arrow in the diagram?
Which row of the table becomes this point on the graph?
Which particle-level change explains this macroscopic observation?
Which part of the map explains the travel-time pattern?

This prevents the learner from treating the pair as two simultaneous but unrelated displays.

33. Use Incomplete Representations

Give a graph with one missing label, a diagram with one missing connection, a table missing one row or a model missing one component.

Ask students to complete it from another representation.

Now translation is required. The learner cannot simply copy the visible form.

34. Use Contradictory Representations

Present two representations that cannot both be correct.

Which one is wrong? What evidence reveals the contradiction? What would have to change to make them consistent?

This develops a habit of coherence checking instead of passive acceptance.

35. Cross-Domain Comparison: Maps and Navigation

A road map, subway map, satellite image and elevation map can describe the same city.

No single map is best. The useful map depends on the task.

Finding the fastest rail interchange is easier on a network map. Estimating walking distance may require geographic scale. Understanding flooding may require elevation.

Learning works similarly: the right representation depends on what relationship must become visible.

36. Cross-Domain Comparison: Engineering Drawings

An engineering system may have plan views, sections, schematics, wiring diagrams and exploded views.

Experts know that each view answers a different class of question. Confusing one view for another is not a small literacy error; it can produce a wrong build or unsafe maintenance decision.

School representations are training for this broader representational world.

37. Cross-Domain Comparison: Finance Dashboards

A financial dashboard can display revenue, margin, cash flow and growth in one screen.

The dashboard is useful only if the reader knows which measure answers which question and how changing aggregation changes interpretation.

More visualisation is not more understanding. Competence lies in interrogating the representation.

38. A Practical Representational-Competence Protocol

Name the phenomenon: what thing or relationship is being represented?
Decode the conventions: symbols, axes, units, arrows, scale, labels.
State the affordance: what does this form make easy to see?
State the blind spot: what does it hide or simplify?
Read the structure: identify relationships rather than decorative features.
Translate: express the same structure in another representation.
Map features: identify exactly what corresponds across the two forms.
Construct: generate a representation from raw information.
Choose: select a representation for a new problem and justify the choice.
Check coherence: look for contradictions among forms.
Vary context: use different surface cases so the representation does not become tied to one example.
Fade prompts: move from teacher-selected forms to learner-selected forms.

39. Failure Mode: Representation as Decoration

The diagram is colourful but does not change what the learner can infer.

Repair: ask what intellectual job the representation performs. If none, remove it.

40. Failure Mode: One Representation Becomes the Concept

Students recognise one familiar graph or diagram but fail when the same relationship appears differently.

Repair: vary form while preserving structure and make correspondences explicit.

41. Failure Mode: Translation Without Meaning

Students learn steps for converting a table to a graph but cannot explain what the graph reveals.

Repair: require explanation of invariant relationships and affordances after translation.

42. Failure Mode: Too Many Representations at Once

The slide contains diagram, text, graph, animation and equation before learners understand any one of them.

Repair: sequence, integrate and explicitly connect. Multiple representations should reduce ambiguity, not create a representational traffic jam.

43. Failure Mode: The Teacher Makes Every Representational Choice

Students become excellent at reading teacher-selected diagrams but freeze when faced with an unfamiliar problem.

Repair: ask learners to choose the form and justify why it is useful.

44. Failure Mode: The Representation Is Trusted More Than the Evidence

A polished chart or model appears authoritative even when its scale, assumptions or source are weak.

Repair: make provenance, assumptions and representational limits part of interpretation.

45. The Missing-Node Scan

If students can solve equations but cannot read graphs, can label diagrams but cannot draw them, can follow maps but cannot explain scale, can repeat a model but cannot state its limits, or can answer only when a problem arrives in a familiar format, the missing node may be representational competence.

Look for these clues: students asking “Which formula is this?” before identifying the relationship; graphs treated as pictures; diagrams copied without interpretation; answers changing when notation changes; multiple representations displayed with no explicit bridge; and learners unable to say why one form is more useful than another.

The knowledge may already exist. What is missing is portability across forms.

46. The Return Path

Return to the student who knew the formula but not the graph.

The formula was never the whole knowledge. It was one encoding.

Once the learner can see the same relationship in symbols, graphs, diagrams, language and real cases, something important changes. The representation no longer dictates the thinking. The learner can choose the representation that serves the thinking.

That is the deeper move from school notation to disciplinary capability.

Representational competence works when knowledge survives a change of form: the learner can read the representation, see through it to the structure, move that structure elsewhere and choose a better form when the problem demands it.

Research and Further Reading

eduKateSG Learning Node Series · 0129 · Previous: 0128 — How Cognitive Flexibility Theory Works.

How Representational Competence Works | Learn to Read, Choose and Translate the Forms Knowledge Takes