Good support in Primary 2 Mathematics should make itself less necessary over time. A prompt, diagram, worked example or guiding question is useful when it helps the learner see the next mathematical relationship. The same support becomes unhelpful when it quietly performs the thinking the child needs to learn to do.
Scaffolding is therefore temporary architecture. It gives a learner access to a task that is currently just beyond independent control, then it changes as the learner changes. Sometimes the right move is to explain more. Sometimes it is to ask one discriminating question. Sometimes it is to remove a hint, wait, and see whether the child can now reconstruct the step.
Return to the Primary 2 Mathematics Learning Hub.
The purpose of help is not to make today’s question easy. It is to make tomorrow’s comparable question more independent.
Why This Guide Exists
Top learning platforms are very good at giving students examples, hints, visual models and practice sequences. Current mathematics research and professional guidance also gives strong attention to worked examples, metacognitive talk and guided practice. The harder operational question is what a parent or tutor should do moment by moment: whether to tell, model, prompt, wait, simplify, represent, or step away.
This guide owns that support decision. It is not a general article about encouragement, and it is not another worked-example page. It is a control system for assistance: identify the barrier, give the smallest useful support, observe the learner’s response, fade deliberately, retest later and verify transfer.
The Scaffolding Loop
| Stage | Teaching move | Question |
|---|---|---|
| Observe | Watch the actual attempt. | Where does control first weaken? |
| Choose | Select the smallest support likely to help. | What does the learner need, not what can I tell? |
| Prompt | Give one bounded cue, example or representation. | Can the learner now make the next move? |
| Fade | Remove or reduce assistance. | Can the step be reconstructed independently? |
| Return | Use a new problem later. | Did learning survive delay and variation? |
1. Support the First Weak Link, Not the Whole Question
A child may be stuck on a two-step problem because the comparison language is unclear, not because addition and subtraction are weak. If the adult solves the whole problem, the visible task is completed but the diagnostic opportunity disappears.
Ask what the first unstable dependency is. Repair only enough of that dependency for the learner to continue.
2. Scaffolding Is Not Giving the Answer Slowly
A sequence of increasingly obvious hints can become answer delivery disguised as teaching. Useful scaffolding keeps the mathematical decision with the learner whenever possible.
3. The Smallest Useful Prompt
Suppose a child reads, “Mei has 18 more stickers than Arun. Mei has 62.” Instead of saying “Subtract 18”, ask: “Who has more?” If the learner answers correctly, ask: “Which amount are we trying to find?” One well-chosen question may restore the relationship without supplying the operation.
4. Wait Time Is a Form of Support
Adults often fill silence too quickly. A learner may need several seconds to retrieve a fact, inspect a diagram or reorganise the problem. Waiting preserves ownership. If the child is still productively working, immediate rescue can interrupt learning.
5. Productive Pause Versus Unproductive Stall
A productive pause contains visible activity: rereading, trying a fact, sketching, checking, or explaining uncertainty. An unproductive stall may look like repeated guessing, blank inactivity without a plan, or cycling through unrelated operations. Support should respond to evidence, not simply elapsed seconds.
6. Worked Examples as Temporary Support
A worked example removes some performance demand so attention can shift to structure. The learner can inspect why a method was chosen, how quantities were represented and how the answer was checked.
The example becomes more educational when students compare methods, explain a step, predict what comes next or identify why a choice was made. EEF guidance highlights this use of worked examples to expose problem-solving strategies and metacognitive decisions rather than merely present a polished solution.
7. Do Not Keep the Worked Example Open Forever
If every practice question is completed beside an identical example, the child may copy surface steps without retrieving the underlying relationship. After initial study, partially cover the example, then remove it and ask for reconstruction.
8. Completion Problems | Remove One Step
A useful intermediate scaffold is a partially completed solution. Show the setup but leave the operation blank, or show the first calculation and ask the learner to label what it means. Completion problems reduce load while keeping an important decision active.
9. Errorful Worked Examples
Once a concept is reasonably stable, show a worked solution containing one deliberate mistake and ask where the reasoning first fails. This develops error detection and prevents worked examples from becoming passive reading.
10. Think-Alouds Make Invisible Decisions Visible
A teacher can model not only calculation but decision-making: “I see ‘more than’, but I’m not choosing an operation yet. First I’m finding who has the larger amount.” This demonstrates monitoring and strategy selection.
Current EEF work with Year 2 mathematics is explicitly testing think-alouds, worked examples and debriefs as a way to support metacognitive problem solving. The evidence is still developing; the useful design lesson is to make reasoning moves inspectable while keeping claims appropriately bounded.
11. A Think-Aloud Should Not Become a Lecture
Long adult monologues can overload a young learner. Model the small decision that matters, then hand the task back: “I checked who has more. Now you decide what the bars should show.”
12. Guided Practice | We Do, but the Learner Still Thinks
Guided practice is not the teacher doing half the arithmetic. It is shared control. The teacher may choose the first representation while the learner labels it, or ask the learner to choose the operation while helping with a difficult regrouping step.
13. Separate Concept Support From Calculation Support
A student may understand the word-problem structure but make a basic fact error. Do not reteach the whole model if only arithmetic retrieval is unstable. Conversely, perfect arithmetic does not repair a misread relationship.
14. Separate Reading Support From Mathematics Support
If vocabulary blocks access, clarify the language without solving the mathematics. Explain what “remaining” means, then ask the learner to decide the quantity relationship.
15. Representation Support
If a problem is too difficult to hold mentally, offer a representation category rather than a finished model: “Would a bar, number line or array help?” If that is too open, narrow further: “Can you show the larger and smaller amounts with two bars?”
16. Fact-Retrieval Support
For 4 × 7, instead of giving 28, ask: “What is 2 × 7? Could doubling that help?” The prompt reconnects the unknown fact to a known relationship.
17. Place-Value Support
If 603 − 278 breaks down, ask the child to represent 603 as hundreds, tens and ones. The support targets renaming across zero rather than explaining every subtraction step verbally.
18. Fraction Support
When a learner says 1/8 is greater than 1/4, ask: “Are the wholes the same size? What happens to each piece when the same whole is split into more equal parts?” If needed, move to fraction strips.
19. Money Support
If $4.05 is read as $4.50, separate the units: “How many dollars? How many cents?” Play money or a dollars-and-cents table may be the next support.
20. Time Support
If a child subtracts 320 − 245 to find a duration, ask: “How many minutes are in an hour?” Then offer a timeline crossing the next full hour. The support targets the 60-minute structure.
21. Data Support
If a child counts picture symbols and stops, point to the key and ask: “What does one symbol represent?” Do not immediately multiply for the learner. The key should trigger the equal-group relationship.
22. A Prompt Ladder
| Level | Example support | Adult restraint |
|---|---|---|
| 1. Wait | Give time to inspect and retrieve. | Do not fill silence immediately. |
| 2. Metacognitive cue | “What are you trying to find?” | Do not name the operation. |
| 3. Relationship cue | “Who has more? What is the difference?” | Keep calculation with learner. |
| 4. Representation cue | “Would two comparison bars help?” | Do not draw the completed model. |
| 5. Partial model | Draw one bar or label one quantity. | Leave a meaningful decision open. |
| 6. Worked micro-example | Show a simpler analogous case. | Return immediately to original problem. |
| 7. Direct explanation | Teach the missing concept. | Follow with an independent reconstruction. |
23. Move Up the Ladder Only as Needed
If a Level 2 question restores progress, there is no reason to jump to a completed model. The smallest successful prompt preserves the most learner ownership and gives better diagnostic information.
24. Prompt Dependency
A child can become very skilled at waiting for a familiar adult cue. If every comparison problem begins with “Who has more?”, the learner may outsource the decision to the prompt. Fade repeated cues by turning them into a self-check card, then remove the card later.
25. Fade the Prompt, Not the Expectation
The mathematical standard remains the same while support decreases. The child still needs to identify quantity roles, choose a representation if useful, calculate accurately and check. What changes is how much of that structure the adult supplies.
26. Fading by Removing Words
Begin with “Who has more? What is the difference? What are you finding?” Later reduce to “What are the roles?” Eventually ask only, “Talk me through your plan.”
27. Fading by Removing Diagram Detail
First provide a completed model, then a partially labelled model, then only an empty frame, then ask the learner to choose whether a model is needed at all.
28. Fading by Increasing Variation
A child who succeeds only when the new question looks nearly identical to the example may still depend on surface cues. Change names, contexts, number sizes and unknown positions while preserving the underlying relationship.
29. Fading by Increasing Delay
Immediate success may reflect short-term memory of the explanation. Return later without the cue. A delayed independent attempt gives stronger evidence that the learner can retrieve the method.
30. Fading by Withholding Confirmation
Some students repeatedly ask, “Is this right?” after every step. Instead of confirming immediately, ask how they can check. Use inverse operations, estimation, units or the original context to shift verification back to the learner.
Independence grows when the learner gradually inherits planning, execution and checking.
31. Worked Example | Support Without Solving
Problem: “Arun has 42 cards. Mei has 17 more cards than Arun. How many cards do they have altogether?”
- Child: “42 + 17 = 59, so 59.”
- Adult: “What does 59 represent?”
- Child: “Mei’s cards.”
- Adult: “What did the question ask for?”
- Child: “Both altogether.”
- Adult: waits.
- Child: “59 + 42 = 101.”
The adult did not reteach comparison or supply the second operation. One state-label question exposed that the first answer was intermediate.
32. Worked Example | A Prompt Is Not Enough
Problem: 402 − 185. The learner repeatedly tries to subtract 5 from 2 and writes random corrections.
Questions such as “What should you do next?” may not help because the underlying place-value renaming is missing. Move to base-ten blocks or a place-value chart, rebuild 402, rename one hundred as ten tens, then one ten as ten ones. After understanding is restored, return to the written algorithm.
33. Worked Example | Fade After Success
After modelling three picture-graph questions with the cue “Read the key first”, give the fourth graph without the cue. On the next day, use a horizontal graph with a different scale. The learner’s independent first action is now evidence: does the child inspect the key without prompting?
34. Debrief After the Problem
Once a problem is solved, ask a short reflective question: “What helped you get unstuck?”, “Which clue mattered?”, “What would you do first next time?” The debrief helps turn an adult prompt into a learner-owned strategy.
35. Build a Self-Prompt Library
- What am I trying to find?
- What does each number mean?
- What relationship connects them?
- Would a diagram help?
- What is my intermediate answer?
- What unit should the answer use?
- How can I check without asking someone else?
Initially these may appear on a card. Later reduce the list and eventually ask the learner to generate the questions internally.
36. Guided Practice Should Become Independent Practice
A lesson is incomplete if every successful problem occurred under active adult guidance. Include at least one independent attempt before concluding that the learner can perform the skill.
37. Independent Does Not Mean No Tools
A learner may independently decide to draw a number line or use counters. That can still be independent problem solving because the choice and use of the tool belong to the learner. Independence is about control, not austerity.
38. Transfer Is Stronger Than Repetition
If a child learns “more than” with stickers, then succeeds with money or lengths, the relationship is travelling. If performance collapses when the context changes, return to representation and language rather than simply adding more sticker questions.
39. The Independence Test
| Evidence | Question |
|---|---|
| Immediate attempt | Can the child do a similar problem after support is removed? |
| Delayed attempt | Can the method be retrieved later? |
| Variation | Does the relationship survive changed wording or representation? |
| Self-correction | Can the learner detect and repair some errors? |
| Strategy choice | Can the learner decide when a tool or model is useful? |
40. Repair Path
Use the repair path when the child cannot explain the concept even with support, repeatedly guesses, or performs a procedure without understanding. Return to the prerequisite and choose a representation that exposes the relationship.
41. Stabilise Path
Use the stabilise path when the child succeeds with one familiar scaffold but not yet independently. Vary examples, fade cues, ask for explanation and revisit after a delay.
42. Extend Path
Use the extend path when the learner performs independently and accurately. Introduce non-routine problems, multiple methods, changed representations and situations where the learner must choose the strategy without being told the topic.
43. Common Error | Adult Prompts Too Quickly
The child learns that uncertainty will be removed externally. Repair by allowing productive wait time and asking for the learner’s current plan before offering help.
44. Common Error | Hint Contains the Operation
“Should you subtract?” can turn operation choice into recognition of adult expectation. Ask about quantity roles or relationship instead.
45. Common Error | Support Never Changes
A worksheet scaffold that was useful in Week 1 can become unnecessary in Week 4. Review support based on current evidence rather than continuing it because it is familiar.
46. Common Error | Support Removed Abruptly
Removing all assistance at once can make a learner fail for reasons unrelated to the concept. Fade one dimension at a time: fewer prompts, less diagram detail, greater variation or longer delay.
47. Common Error | Help Is Measured by Amount, Not Effect
More explanation is not automatically better. The relevant measure is whether the learner can make a productive next move and later perform with less assistance.
48. Parent Routine | Before Helping
- Ask the child to read the question again.
- Ask what the problem wants.
- Ask what has already been tried.
- Identify whether the barrier is language, concept, fact, representation or calculation.
- Give one small prompt.
- Wait for a response before adding another.
49. Tutor Routine | Track Support Level
For important skills, record not only accuracy but support required: full explanation, worked example, representation cue, verbal prompt, self-prompt, or independent. Improvement can appear as the same correct answer reached with less assistance.
50. A Visible Progress Ladder
| Stage | Student capability |
|---|---|
| Access | Can complete with direct modelling. |
| Guided | Can complete with prompts or partial models. |
| Supported choice | Can choose a method after a broad cue. |
| Independent | Can solve and check without adult prompts. |
| Transfer | Can apply the relationship in a changed context. |
51. What Parents Should Look For
- Does the child start independently?
- What is the first point where help is requested?
- Can one question restart the reasoning?
- Does the same prompt need to be repeated every time?
- Can the learner explain what the prompt helped them notice?
- Can a similar question be completed later without the prompt?
52. What Teachers Should Diagnose
| Observed behaviour | Next diagnostic |
|---|---|
| Needs operation named | Relationship and unknown-role recognition. |
| Needs diagram completed | Representation construction. |
| Needs fact supplied | Fact retrieval and derived strategies. |
| Needs constant confirmation | Verification routines and confidence in checking. |
| Succeeds immediately, fails later | Delayed retrieval and scaffold dependence. |
| Succeeds familiar form, fails variation | Transfer and representation switching. |
53. FAQ | Should I Let My Child Struggle?
Allowing time for productive thinking is different from leaving a child indefinitely stuck. Observe what the learner is doing. If there is a plausible plan, give space. If the child is cycling through random guesses or the missing concept is clear, intervene with a bounded prompt or explanation.
54. FAQ | Is Showing a Worked Example Cheating?
No. A worked example can be an excellent learning tool when the student studies the reasoning and then attempts a new problem independently. The risk appears when examples remain permanently available and substitute for retrieval.
55. FAQ | When Should I Stop Helping?
Reduce help when the learner can reliably perform the supported step. Do not wait for perfection, but do not infer mastery from one successful guided question. Fade, test independently and return later.
56. What Mastery Looks Like
A strong Primary 2 learner can use adult or material support when genuinely needed, but increasingly owns the planning, representation, calculation and checking. The student can identify a problem type without a topic label, retrieve a method after a delay, adapt it to a changed context and recover from some errors without immediate rescue.
The best scaffold leaves evidence of itself in the learner’s thinking, not permanent dependence on the scaffold.
57. Primary 3 Bridge
Primary 3 increases topic load, multi-step reasoning and the need to select methods independently. Students who have learned to use and then outgrow scaffolds enter that transition with a more important capability than mere worksheet familiarity: they know how to get unstuck without surrendering ownership of the problem.
Evidence and Scope Notes
Worked examples, strategy comparison and metacognitive talk are supported in professional mathematics guidance, but teaching decisions still depend on learner response, task difficulty and context. Current EEF work includes a Year 2 pilot of think-alouds, worked examples and debriefs; a pilot tests feasibility and promise and should not be reported as settled proof of a universal effect.
- Education Endowment Foundation: Making the Most of Worked Examples
- Education Endowment Foundation: Problem Solving, Worked Examples and Scaffold Fading
- Education Endowment Foundation: Current Cognitive Science Research Agenda
- Math Learning Center Grade 2: Number Talks and Strategy Representation
