The Tutor Handbook · Volume 0204 · Series ID THB-0204
The Tutor Handbook: Complete Series Index
The learner understands the blocks. The equation still looks foreign.
A learner can build three equal groups with counters.
They can point to the total.
They can say, “There are three groups of four.”
Then the tutor writes \(3 \times 4 = 12\), and the learner hesitates.
Another learner can manipulate the equation confidently but cannot explain what the symbols represent in a word problem. A third can copy a bar model after seeing one but cannot decide when a bar model is useful. A fourth can describe a Science process accurately in words and then misread the corresponding graph.
These are not simply “visual learner” problems or evidence that one representation is better.
They are translation problems.
A representation carries selected features of an idea. Counters make quantity and grouping visible. A diagram can expose relation and structure. Words can name conditions and causal links. Symbols can compress a relationship so efficiently that the learner no longer sees the concrete objects at all.
The Representation-Translation Gate asks whether the learner can preserve the important structure while the representation changes.
It helps a tutor decide when to introduce another representation, when to keep two forms visible together, when to ask the learner to translate between them, and when fluency in one form should not yet be treated as understanding that will travel.
The educational goal is not to march mechanically from concrete to picture to symbol.
It is to build connections strong enough that the learner can use the form that the task requires without losing the concept underneath it.
Quick answer
Use a representation when it makes an important relation easier to see, reason about or communicate.
Do not assume the representation teaches the relation by itself.
Name what the representation preserves.
Ask the learner to connect corresponding parts across forms.
Move in both directions.
Concrete to diagram.
Diagram to words.
Words to symbols.
Symbols back to a concrete or contextual meaning.
Vary the surface while keeping the deep relationship stable.
Then remove one form and test whether the learner can reconstruct the relation independently.
If the learner succeeds only while two representations remain side by side, the connection may still be supported rather than owned.
If the learner can translate but only by following a memorised conversion routine, test a changed case.
If the real task requires one particular representation—formal notation, graph construction, algebraic manipulation or a written explanation—return to that target form before claiming readiness.
Representation is a thinking tool.
Translation is evidence that the idea is becoming portable.
The ownership boundary
Several existing eduKate owners sit nearby.
How Explanation Works in Teaching owns the general problem of making invisible structure visible without replacing learner thinking.
The Example-Variation Gate asks what should vary across examples and what must remain invariant.
The Response-Modality Equivalence Gate asks whether oral, typed, handwritten, selected or visual responses provide evidence about the same target capability.
The present gate owns a different object: the representation of the idea itself.
A learner may respond in writing while translating a concept from graph to equation. A learner may respond orally while explaining how a diagram maps onto a scientific mechanism. Response mode and conceptual representation can change independently.
That distinction matters.
The tutor is not merely asking, “How can the learner answer?”
The tutor is asking, “Can the learner preserve the structure when the idea is expressed differently?”
What strong external guidance says
The What Works Clearinghouse practice guide Organizing Instruction and Study to Improve Student Learning, released September 2007, gives two directly relevant recommendations with moderate evidence in that guide: combine graphical presentations with verbal descriptions, and connect abstract representations of a concept with concrete representations of the same concept.
The WWC practice guide Assisting Students Struggling with Mathematics: Intervention in the Elementary Grades, released March 2021, gives a strong-evidence recommendation to use well-chosen concrete and semi-concrete representations to support mathematical concepts and procedures.
A current Regional Educational Laboratory resource based on that guide, Assisting Students Struggling with Mathematics in Grades 3–6, makes an important implementation point: explicitly connect concrete, semi-concrete and abstract representations, use questioning to help students explain how the model represents the concept, and avoid treating concrete-then-picture-then-abstract as a rigid one-way sequence.
The WWC guide Improving Mathematical Problem Solving in Grades 4 Through 8, released May 2012 and revised October 2018, includes a strong-evidence recommendation to teach students to use visual representations.
These sources support deliberate representation use, connection and explanation. They do not establish that every representational translation should be taught in the same order, or that success in one representation guarantees transfer to another subject or task.
This Tutor Handbook gate turns those broad recommendations into a tutoring decision.
A representation is selective
Every representation highlights some features and hides others.
A number line makes magnitude and distance visible. It does not automatically show part-whole area.
A bar model can make comparison and total-part relations visible. It can make a simple problem look more complicated if the learner already sees the equation directly.
An algebraic expression compresses structure. That compression is powerful because details disappear.
A graph can reveal covariation. It can hide the mechanism that generated the relationship.
A Science diagram can show components and direction. It may simplify scale, motion, three-dimensional form or time.
Language can make conditions explicit. It can also bury a simple relation inside vocabulary and syntax.
The tutor should therefore ask a representation question before teaching with the representation:
What job is this form doing that the learner cannot yet do reliably without it?
If the answer is “because this is the method we always use”, the representation may be becoming ritual.
Composite case: fraction tiles that never become fractions
This case is fictional and constructed for teaching.
Alicia is learning fraction equivalence.
With fraction tiles, she can place two one-quarter pieces over one-half and see that the lengths match. She can repeat the same comparison with four eighths.
Her tutor concludes that equivalence is understood.
On a worksheet, Alicia says \(2/4\) and \(1/2\) are different because “two is not one and four is not two”.
The tiles produced correct performance without a stable symbolic connection.
The tutor changes the task.
Alicia places the tiles again. This time, she must point to each physical quantity while the tutor writes the corresponding fraction. The tutor asks what the denominator is counting in each case and what remains the same when the number of parts changes.
Then the tutor removes the tiles and shows two bar diagrams.
Alicia links the diagrams to \(1/2\) and \(2/4\).
Next she is shown \(3/6\) without a diagram and asked to sketch a representation that would justify whether it is equivalent to \(1/2\).
The translation now runs both directions.
The tile was not the concept.
The connection among quantity, partition and symbol was the educational job.
Do not worship the concrete
Concrete materials can be valuable because they make relations manipulable and perceptually available.
They can also create their own surface dependency.
A learner may know that the blue rods “go with” one kind of question without understanding the relation.
They may count blocks when a more efficient grouping relation should be emerging.
They may be unable to solve the same problem when the colours or shapes change.
Concrete is not automatically conceptually pure.
The tutor should check what the learner attends to.
Ask:
“What does this piece stand for?”
“What would stay the same if the pieces looked different?”
“Can you draw what these blocks are showing?”
“Can you write the relationship without the blocks?”
“Can you rebuild the model from the equation?”
The goal is not to remove manipulatives quickly.
It is to prevent the manipulative from becoming the only place where the relationship exists.
Do not worship the abstract either
Symbolic fluency can hide shallow mapping.
A learner performs \(3(x+4)=21\) by a familiar sequence and gets \(x=3\).
Then the tutor asks what the equation says about three equal groups.
The learner cannot answer.
Or the learner manipulates a formula accurately but substitutes the wrong physical quantity because the symbols have become detached from meaning.
Abstract notation is powerful precisely because it compresses.
The learner needs enough conceptual connection that the compression can be unpacked when a new problem changes the surface.
A useful tutoring move is reconstruction.
“Show me what this equation could mean.”
“Draw a relationship that would produce this graph.”
“Tell me what each symbol controls.”
“Create a word problem whose structure matches this expression.”
If the learner cannot reconstruct meaning, the tutor may be looking at procedural familiarity rather than portable understanding.
Two representations side by side can still create borrowed understanding
Placing a diagram beside an equation feels like connection.
Sometimes it is merely co-presence.
The tutor points from the bar to the number. The learner follows.
The tutor colours matching parts. The learner nods.
The tutor narrates every correspondence.
Later, without the tutor, the learner cannot recreate the mapping.
A strong connection task gives the learner some of the mapping work.
Hide one label.
Ask the learner to place it.
Give two equations and ask which matches the diagram.
Show a correct and an almost-correct mapping and ask what breaks.
Remove the diagram and ask the learner to reconstruct it from the symbolic form.
Return after a delay.
Connection must eventually become an action the learner can perform.
Translation should be bidirectional
One-way routines are fragile.
A learner may be trained to turn word problems into equations and remain unable to interpret an equation in context.
They may convert a table into a graph but not read a graph back into a relationship.
They may turn a Science diagram into prose but fail to sketch the mechanism from prose.
Bidirectional translation tests whether the correspondence is genuinely understood.
This does not mean every lesson needs every direction.
Choose the direction that reveals the current uncertainty.
If the learner can calculate from the equation but not explain the graph, move equation to graph.
If the learner can model with blocks but not identify the symbolic form, move concrete to symbol.
If the learner can read the diagram but cannot generate one, move words to diagram.
A translation direction is a diagnostic choice.
Composite case: the bar model that becomes a drawing exercise
This case is fictional.
Beatrice has been taught to use bar models for comparison word problems.
Her diagrams are neat.
Her labels are accurate.
Her answers are often wrong.
The tutor initially thinks she needs more practice drawing bars.
A closer look shows that Beatrice begins drawing before deciding which quantities are being compared. She has learnt the representation as a required product rather than as a model of a relationship.
The tutor reverses the sequence.
Before drawing, Beatrice must state the relation in words:
“Rina has 8 more than Mei.”
Then she chooses between three candidate diagrams.
Next she explains why one wrong diagram represents “8 times as many” instead.
Only after discrimination does she draw.
Finally, she is given a symbolic relation and asked to invent a matching story.
The bar model becomes a thinking tool again because the relation precedes the drawing.
The problem was not poor diagram skill.
It was representation without relational ownership.
Translation can reveal a misconception that one form conceals
Different representations expose different errors.
A learner may verbally say “the graph gets steeper” and reveal that they are confusing height with rate of change.
A symbolic answer may hide the same misconception because a memorised formula still produces the correct number.
A diagram may reveal reversed cause and effect that fluent prose disguised.
A table may expose that the learner thinks an increase is additive when the context is multiplicative.
The tutor can deliberately change representation to stress the model.
This is especially useful when the current form has become over-familiar.
Do not change representation merely to add variety.
Change it because a different form makes a consequential relation observable.
Translation is not the same as decoration
A colourful diagram beside a paragraph is not automatically instruction.
A graph placed beside an equation does not teach the mapping if nobody directs attention to corresponding features.
The WWC recommendation to combine graphics with verbal descriptions is often misunderstood as “use more visuals”.
The educational job is integration.
What does the arrow correspond to in the sentence?
Which segment of the graph shows the condition described in the paragraph?
Which number in the equation corresponds to the quantity represented by this area?
What feature appears in one representation but not the other?
The tutor’s language should direct attention to structure, not merely announce that two forms are related.
Use contrastive representations
Sometimes the strongest translation task contains a near-miss.
Show two diagrams that look similar but represent different relationships.
Show one equation that matches a model and one that contains the common misconception.
Show a graph where the intercept changes and another where the slope changes.
Ask what feature makes one match and the other fail.
Contrast can make the mapping rule visible.
The Erroneous-Example Gate remains relevant when deliberately incorrect examples are used. The learner needs enough knowledge for the wrong form not to seed confusion.
Representation choice should depend on the learning job
A representation can serve at least four different jobs.
Expose. Make an invisible relation visible.
Reduce. Remove irrelevant complexity so the learner can focus on one structure.
Coordinate. Keep several quantities or conditions visible at once.
Transfer. Change the surface so the learner must recognise the same deep relationship elsewhere.
The tutor should know which job is active.
A representation that is excellent for exposure may be poor for final verification because it gives away the structure the learner is meant to select independently.
A diagram that helps a novice organise a word problem may become an unnecessary scaffold for a learner who now needs to choose among methods without cues.
The representation can therefore move from support to evidence condition to something that must be faded.
Translation can become its own overload
More representations are not always better.
Blocks, diagrams, tables, equations, colour coding and verbal explanation can produce a lesson in which the learner spends more effort coordinating forms than understanding the idea.
The tutor should add a representation only when its information value justifies the coordination cost.
Watch for signs of representational overload.
The learner keeps looking back and forth without making a decision.
They copy labels mechanically.
They ask which colour means which quantity.
They can solve with one representation but become confused when another is introduced.
They spend most of the task drawing rather than reasoning.
The response may be to simplify.
Use two forms, not five.
Stabilise the relation.
Add another representation later if a new decision requires it.
Language is also a representation
Tutors sometimes treat words as neutral explanation around the “real” Mathematics or Science.
Language is carrying structure too.
“Three more than” is not the same relation as “three times as many”.
“Because” marks a causal relation.
“If” introduces a condition.
“Per” carries a rate relation.
A learner can understand a diagram and fail to map the language.
This does not mean the subject concept is irrelevant. It means the translation between disciplinary language and another representation may be a legitimate learning target.
The Task Purity Check remains useful when language demand may be distorting evidence of another target.
Sometimes the right route is parallel: keep teaching the concept through a clean representation while also teaching the language needed to access the real task.
Composite case: a Science graph that is treated as a picture
This case is fictional.
Ciara can explain that increasing light intensity raises the rate of photosynthesis up to a point where another factor limits the process.
Shown a graph of rate against light intensity, she describes the line as “going up and then becoming straight”.
Her verbal concept is stronger than her graph interpretation.
The tutor asks her to annotate the graph with three phrases from her explanation.
“Light is limiting here.”
“Rate increases here.”
“Another factor limits further increase here.”
Ciara then draws a second graph for a different limiting condition and explains what changed.
Later, she receives only the graph and must reconstruct the mechanism.
The tutoring route does not label her “bad at graphs”.
It identifies a representation translation that needs explicit teaching.
The target representation still matters
Alternative representations can reveal understanding.
They do not remove the need to perform in the form required by the curriculum or examination.
A learner may explain an algebraic relationship with a diagram and still need to manipulate the algebra.
A learner may understand a Science mechanism through a sketch and still need to write a precise explanation.
A learner may reason correctly with a table and still need to construct a graph.
Use the alternative representation to diagnose and teach.
Then return to the target form.
This prevents a helpful representation from becoming a permanent bypass.
Translation as a transfer test
A changed representation can be a high-value transfer probe.
If the learner studied a ratio in a table, can they recognise the same relation in a graph?
If they learned a fraction with area models, can they interpret it on a number line?
If they studied a process in prose, can they sequence a diagram?
If they solved an equation symbolically, can they identify which word problem has the same structure?
Success across representation changes strengthens the claim that the learner owns more than one surface routine.
Keep the probe fair.
Do not introduce a new representation so unfamiliar that failure tells you only that the learner has never seen the format.
The changed form should stress the concept, not create an unrelated format puzzle.
A practical representation-translation protocol
Start with the target relation.
Write it without naming a representation.
“Part-whole equivalence.”
“Constant rate.”
“Cause under a condition.”
“Comparison difference.”
Choose the representation that currently makes the relation easiest to see.
Ask the learner to explain what each important feature stands for.
Introduce a second representation.
Make the correspondence explicit enough that the learner can begin, but leave some mapping work to them.
Ask the learner to translate.
Reverse the direction.
Vary the surface.
Remove one support.
Return after a delay.
Then test the target representation under realistic conditions.
Record the conclusion at the right scale.
“Can translate between bar model and equation for additive comparison.”
Not:
“Understands representations.”
Failure modes
The concrete-equals-understanding failure. Successful manipulation is treated as proof that the symbolic or relational meaning is owned.
The abstract-equals-mastery failure. Correct symbolic procedure is treated as proof that the learner can interpret or transfer the relationship.
The rigid ladder failure. Concrete, pictorial and abstract forms are taught as a one-way age sequence rather than flexibly connected tools.
The decoration failure. Visuals are added without directing attention to the structural correspondence.
The co-presence failure. Two representations appear side by side, but the tutor performs all the mapping.
The one-way translation failure. Learners practise converting in one direction and are assumed to understand the reverse.
The representation overload failure. Too many forms compete for working attention.
The ritual diagram failure. The learner produces a model because the method requires it, not because it represents a chosen relation.
The bypass failure. A helpful alternative representation becomes a permanent substitute for the target form.
The visual-style failure. A temporary representational strength is turned into a fixed learner identity.
Translation should preserve the relationship, not the decoration
When learners move between representations, tutors need to decide what must survive the move.
A bar model, number line, equation and verbal explanation may look completely different. Their visual features are not the invariant.
The invariant is the relationship they encode.
If three equal groups contain twelve objects in total, the important structure includes equal grouping, group count, group size and total. A learner who redraws the blocks accurately but writes an equation that reverses the relationship has copied the surface without preserving the structure.
The tutor can therefore ask a translation question that points to the invariant:
“Where is the ‘three groups’ idea in your equation?”
“Which part of the diagram represents the unknown?”
“What does this symbol correspond to in the story?”
“Which relationship stays true when the picture disappears?”
These prompts do not ask the learner to decorate one representation with labels from another. They ask the learner to map meaning.
That mapping is the educational work.
As expertise develops, the tutor can reduce the mapping prompts and ask the learner to justify the correspondence independently.
Use representation mismatch as diagnostic evidence
A mismatch between forms can be more informative than a wrong final answer.
A learner draws the situation correctly but cannot express it symbolically.
That suggests one kind of bottleneck.
A learner writes a correct equation from a familiar diagram but cannot explain what the terms mean.
That suggests another.
A learner can manipulate symbols successfully but cannot recognise the same relationship in a graph or story.
That suggests fragile transfer or over-specialised procedural knowledge.
The tutor should avoid collapsing all three into “doesn’t understand the topic”.
Instead, record the boundary.
“Concept is available in concrete and verbal form; symbolic translation is not yet independent.”
“Symbolic procedure works; relation to the represented quantities is weak.”
“Graph reading is secure when axes are familiar; translation from context to graph needs support.”
These statements create targeted next moves.
They also prevent unnecessary reteaching of what is already secure.
Representations should eventually become selectable tools
At the beginning, the tutor often chooses the representation.
Later, the learner should increasingly decide whether a representation is useful.
A diagram is not automatically better than an equation.
Concrete materials are not automatically more supportive.
A table is not always the clearest route.
Expertise includes choosing the form that exposes the relationship efficiently for the current problem.
The tutor can therefore shift the question from:
“Use this model.”
to:
“Would a model help here? Which one, and why?”
Then verify whether the learner can proceed when no representation is suggested.
The end state is not abandonment of diagrams or concrete materials.
It is representational agency.
The learner can enter a problem through one form, translate when useful, detect when two forms disagree, and choose a representation because it serves the reasoning rather than because the tutor always supplies it.
Evidence boundaries
The WWC guide Organizing Instruction and Study to Improve Student Learning, released September 2007, is an evidence-based practice guide whose recommendations include combining graphical and verbal information and connecting concrete and abstract representations. Its evidence ratings apply to the recommendations as reviewed there, not to every tutoring routine proposed in this article.
The WWC guide Assisting Students Struggling with Mathematics: Intervention in the Elementary Grades, released March 2021, gives strong-evidence recommendations around mathematical language and representations for elementary intervention. Applying those recommendations to older learners, other subjects or different tutoring contexts requires judgement.
The REL resource Assisting Students Struggling with Mathematics in Grades 3–6 operationalises the WWC recommendation and explicitly warns against rigid concrete-to-semi-concrete-to-abstract sequencing. It is professional-learning guidance, not a trial of this Tutor Handbook protocol.
The WWC guide Improving Mathematical Problem Solving in Grades 4 Through 8, released May 2012 and revised October 2018, supports teaching visual representations in mathematical problem solving.
The broader principle used here—that translation should preserve deep structure while representation changes—is consistent with these sources. No claim is made that a specific number of translations, a fixed sequence of forms or one representational taxonomy is universally optimal.
The end state
A learner should not need to live permanently inside the tutor’s favourite representation.
Concrete materials can make an idea touchable.
A diagram can make a relation visible.
Language can make a condition explicit.
Symbols can make a structure compact and powerful.
The educational achievement is connection.
The learner sees that the forms are not separate lessons. They are different ways of carrying the same important relationship.
They can move between them when the task demands it.
They can notice when one form hides something another reveals.
They can reconstruct meaning when notation becomes compressed.
And when the final task requires a particular representation, they can perform in that form without needing the tutor to translate first.
That is the Representation-Translation Gate.
The representation helps the learner think.
The translation shows whether the thinking can travel.
Representation choice should fade from tutor-owned to learner-owned
Early in instruction, the tutor may choose the representation because the learner does not yet know which form will expose the relation.
That is legitimate support.
Later, the learner should increasingly choose.
A tutor can ask, “What would help you see this relation?” and watch the choice itself as evidence.
A learner who automatically reaches for a bar model on every word problem may have learned a routine rather than a representational judgement. A learner who refuses all diagrams because they can often manipulate symbols quickly may be losing a useful error-checking tool. A learner who chooses a table for a covariation problem and can explain why the table helps is showing a different level of control.
Representation selection can therefore become part of independent problem solving.
The tutor can build this gradually.
First, choose the representation and explain why.
Then offer two plausible representations and ask the learner to select.
Later, ask the learner to generate a useful form.
Finally, include problems where no additional representation is needed and see whether the learner can decide not to create one.
This matters because expertise is not maximal representation use. It is economical representation use.
The learner needs a repertoire and a selection rule.
Keep representational errors separate from concept errors
A wrong diagram can have several causes.
The learner may not understand the concept.
They may understand the concept but not the conventions of the representation.
They may know both but make a local execution error.
They may misread the task language before the representation begins.
The tutor should not collapse these into “doesn’t understand”.
Use a small differential.
Ask the learner to explain the concept in the strongest available form.
Ask them to interpret a completed correct representation.
Ask them to construct the representation from a known relation.
Ask them to use the representation to make a prediction.
The pattern matters.
If they can explain and interpret but cannot construct, teach representation construction.
If they can construct but cannot reason from it, the form may be procedural.
If they fail across forms, the concept itself may need repair.
This is another reason translation is valuable: it separates the idea from the vehicle carrying it.
