Originally published 25 May 2017 as a Punggol Primary Mathematics tuition page. Rebuilt in 2026 as a noindexed guide to moving students from supported examples toward independent problem solving. Dated schedules, broad promotional claims and generic tutor language have been retired.
Quick answer: guidance should decrease as capability increases. A strong Primary Mathematics sequence is worked example → explained example → completion problem → faded prompt → independent attempt → variation → delayed retrieval → mixed transfer. If students can succeed only while a tutor is beside them, the teaching has produced assisted performance rather than independent mathematical control.
This page is intentionally noindex. Its reader job is the transition from guided Mathematics practice to independent problem solving.
Support is useful only if it can eventually be removed
A tutor can make a difficult question feel easy by asking the right questions in the right order. That is valuable during teaching, but it creates a measurement problem.
Who solved the problem?
- the student;
- the tutor;
- or the student–tutor pair?
The distinction matters because examinations, homework and future learning eventually require the student to operate without the same prompt sequence.
Begin with a clear worked example
Worked examples can reduce unnecessary search when a student is meeting a new idea.
A useful example should make visible:
- what information matters;
- what relationship is being represented;
- why the method fits;
- what each line of working does;
- how the answer is checked.
The example should not merely show a finished sequence of symbols. The student should be able to explain the logic connecting the steps.
Ask the student to explain the example back
Recognition is easier than reconstruction. A student may nod through an example because every step looks reasonable once it is visible.
- Why did we choose this representation?
- What would make another method unsuitable?
- What is the purpose of this intermediate value?
- How could we check the result?
If the student cannot explain the route, they may be following rather than learning.
Move next to completion problems
A completion problem gives part of the structure and asks the student to finish it.
- a diagram is provided but not labelled;
- the first equation is given but the next step is not;
- a bar model is started but not completed;
- the method is named but the execution is left to the learner.
This bridges the large gap between watching a full example and solving a blank-page problem.
Fade prompts deliberately
Prompts should become less specific over time.
| Support level | Example prompt |
|---|---|
| High | “Draw two bars and show the fixed difference.” |
| Medium | “What representation could make the relationship visible?” |
| Low | “What do you know?” |
| Independent | No prompt |
If the tutor continues using high-support prompts after the student no longer needs them, the student may never practise selecting the next move independently.
Do not remove support faster than the learner can carry the load
Fading is not abandonment. If the student repeatedly collapses when a prompt disappears, restore the smallest amount of support needed and identify which capability is missing.
- concept;
- representation;
- method selection;
- retrieval;
- execution;
- checking.
The aim is a gradual transfer of control.
Independent problem solving begins before the student writes the first line
The student must decide how to begin.
A useful internal sequence is:
read → identify target → identify relationship → choose representation → generate method → execute → check.
If the tutor always supplies the representation or method, these early decisions remain untrained.
Use variation before declaring mastery
A student can memorise the surface of a worked example. Change the surface while preserving the mathematical structure.
- change the numbers;
- change the context;
- change the diagram;
- reverse what is known and unknown;
- remove a cue;
- combine the idea with another topic.
If the method survives variation, the student is learning the relationship rather than the template.
Delayed retrieval matters
Success immediately after an example is expected because the method is still active in memory.
Return later without announcing the method. Ask whether the student can:
- recognise the structure;
- retrieve the relevant idea;
- select the method;
- complete the solution;
- check independently.
Delayed success is stronger evidence than immediate imitation.
Mix topics so the student must choose
Blocked practice is useful while a method is being stabilised. But if every question on the page is the same type, the worksheet gives away the method.
Mixed practice asks a different question: Can the student identify which method fits?
Tutor questions should reveal thinking, not supply it
Compare two interventions.
Answer-supplying: “Use a bar model, then subtract these two values.”
Thinking-revealing: “What relationship do you see? How could you represent it?”
The second approach exposes the student’s model and creates a chance to diagnose the real weak link.
Track the prompt, not just the answer
A correct answer with heavy prompting is not equivalent to a correct answer produced independently.
| Outcome | Interpretation |
|---|---|
| Correct after tutor models | Exposure achieved |
| Correct after strong prompt | Partial control |
| Correct after narrow cue | Near-independent |
| Correct with no prompt | Independent performance |
| Correct after delay and variation | Stronger evidence of learning |
Allow useful struggle
If help arrives the instant a student hesitates, the learner never practises recovery.
Useful struggle can include:
- trying a representation;
- rejecting an unsuitable method;
- checking an assumption;
- finding an earlier mistake;
- restarting with a simpler case.
The tutor should intervene when the struggle has stopped generating information, not merely when the student feels discomfort.
But do not romanticise confusion
Leaving a child lost for a long time is not independence training. The tutor should distinguish productive search from dead time.
- Is the student testing ideas?
- Are they narrowing possibilities?
- Are they learning from failed attempts?
- Or are they simply waiting?
Good support restores movement while preserving as much student thinking as possible.
Corrections should also fade
Early on, the tutor may need to point directly to an error. Later, the student should be expected to locate it.
A useful progression is:
tutor identifies error → tutor identifies line → tutor identifies error class → student locates error → student predicts likely error → student checks before submission.
Checking is part of independence
- Is the answer reasonable?
- Are units correct?
- Does the result satisfy the original condition?
- Can the method be checked another way?
- Did any line break mathematical equivalence?
A student who can solve but never check still depends on an external quality-control system.
Use explanation to test control
After solving, ask the student to explain:
- why the representation was chosen;
- why the method works;
- where another student might go wrong;
- what would change if one condition changed.
Explanation makes hidden reasoning visible and reveals whether the method is understood or merely reproduced.
A useful release-of-responsibility cycle
| Stage | Tutor role | Student role |
|---|---|---|
| Model | Make expert thinking visible | Observe and question |
| Explain | Probe understanding | Reconstruct reasoning |
| Complete | Provide partial structure | Finish solution |
| Fade | Reduce prompts | Select more steps |
| Independent | Observe | Solve and check |
| Transfer | Change conditions | Adapt method |
| Delay | Remove recent cues | Retrieve independently |
The group can help—but must not replace independent thinking
In a small group, students can compare methods and explain ideas to one another. That can deepen understanding. But require an independent first attempt or an independent final transfer task so one fast student does not become the method-selector for everyone else.
Historical classroom context
The original 2017 page emphasised tutors guiding students when they did not understand and checking work closely before examinations. That remains useful, but the stronger long-term purpose is the transfer of responsibility: the tutor should gradually become less necessary for starting, selecting, checking and recovering.
A practical independence test
- teach with a worked example;
- ask the student to explain it;
- give a completion problem;
- fade the prompt;
- give a fresh independent problem;
- change the surface or representation;
- return after a delay;
- mix the topic with others.
At each stage, record how much help was needed.
What parents, tutors and students can measure
- Is the student initiating the first step more often?
- Are prompts becoming less specific?
- Can the learner explain why a method fits?
- Can they solve changed versions?
- Does the method survive after a delay?
- Can they recover from a stall without immediate rescue?
- Can they check their own work?
What not to conclude
- A correct answer with heavy prompting is not the same as independent mastery.
- Worked examples are useful, but permanent imitation creates dependency.
- Removing support too quickly can overload the learner.
- Leaving a student confused is not automatically productive struggle.
- Blocked practice does not prove method selection.
- The strongest evidence is delayed, varied, independent performance.
Related Mathematics routes
- Why Primary Mathematics Age Problems Are Really About Invariants and Differences
- How Primary Mathematics Weaknesses Grow
- What Small-Group Tuition Should Actually Make Possible
For current programme information, use the eduKate contact page.