Quick answer: an eduKate lesson should not be a fixed sequence of “teacher explains, student copies, worksheet ends”. A stronger learning cycle is simpler and more demanding: locate what the learner can currently do → explain only what is needed → attempt → observe the first breakdown → narrow the cause → repair → retrieve later → vary the task → test transfer → return control to the student. The cycle matters more than the amount of material covered in one sitting.
This page began in July 2017 as a broad description of what eduKate did in Punggol: syllabus teaching, small groups, understanding before practice, enjoyable lessons, customised materials, holistic education and enrichment outside the classroom. Those ideas remain useful, but several words had become too loose. “Holistic”, “customised”, “fun” and “understand” can all sound good while meaning almost anything. The 2026 rebuild turns those words into observable teaching actions and boundaries.
What this page is here to show
This article shows the learning cycle a student or parent should be able to recognise in practice: locate the learner’s current state, teach proportionally, let the learner attempt, diagnose the first breakdown, repair it, retest it and gradually return control to the student.
For the values that constrain these decisions, see What Makes eduKate? Values That Must Change Teaching Behaviour.
The cycle starts by locating, not assuming
A new topic does not automatically mean “teach from scratch”. A low score does not automatically mean “the student does not understand”. A correct answer does not automatically mean “the student has mastered it”.
Before teaching, locate the current state with a small amount of evidence:
- a recent marked question;
- one independent attempt;
- a short oral explanation;
- a retrieval question from earlier learning;
- a changed-form example.
The first decision is therefore not “what page are we on?” It is “what does this learner already own, and where does performance first become uncertain?”
Step 1: explain only the missing relationship
Over-explanation can be as inefficient as under-explanation. If a student already understands most of the structure, reteaching everything consumes time and hides the actual weak link.
A useful explanation should:
- connect to something the student already knows;
- make the relationship visible;
- show one or two carefully chosen examples;
- name the common confusion;
- stop before the tutor becomes the person doing all the thinking.
The explanation earns its place when it changes the next student attempt.
Step 2: the student attempts before the tutor rescues
The learner needs enough unsupported space to reveal what is actually available.
If the tutor supplies the first word, first equation, first formula or first interpretation every time, lesson performance can look excellent while independent capability stays weak.
A useful wait is not neglect. It is diagnostic space.
Watch:
- where the student begins;
- what they retrieve automatically;
- where hesitation begins;
- whether they can detect their own error;
- how they respond when the first method fails.
That behaviour is often more informative than the final answer.
Step 3: find the first breakdown, not the last wrong answer
A student may end with the same wrong answer for several different reasons.
| Visible outcome | Possible first breakdown |
|---|---|
| Wrong Mathematics answer | Concept, method selection, algebra, arithmetic or time |
| Weak English paragraph | Task relevance, idea development, sentence control or vocabulary deployment |
| Weak explanation | Question interpretation, evidence selection, concept connection or expression |
| Blank response | Missing knowledge, retrieval failure, overload, anxiety or no starting strategy |
The correction should begin where the chain first broke.
Step 4: narrow competing explanations
“The student is careless” and “the student does not understand” are often premature conclusions.
Suppose a learner repeatedly makes errors under time. Two explanations compete:
- the method is unstable;
- the method is stable but execution collapses under load.
The smallest discriminating test is simple: give a similar question without time pressure and ask the learner to explain each decision. If the method still fails, the first explanation becomes stronger. If performance becomes accurate, timing, fluency or cognitive load deserves more attention.
Good teaching tries to reduce uncertainty before increasing workload.
Step 5: repair the smallest useful unit
Once the bottleneck is clearer, repair the smallest unit that can change the larger performance.
Examples:
- one algebraic sign habit rather than the entire chapter;
- one comprehension inference step rather than another full paper;
- one paragraph-development problem rather than rewriting the whole composition immediately;
- one retrieval gap rather than rereading all the notes.
Small repair units make cause and effect easier to observe.
Step 6: correction must end in another performance
Seeing the correct answer is not the end of correction.
The learner should:
- explain what went wrong;
- repair the original attempt;
- complete a near example;
- complete a changed example;
- return to the skill later without the correction beside them.
The final step separates recognition from ownership.
Step 7: retrieve after a delay
Immediate success after teaching is useful but weak evidence. The example is still fresh, the tutor is nearby and the relevant method has just been activated.
Return later.
Ask whether the learner can:
- remember the key idea;
- reconstruct a forgotten detail;
- begin without a cue;
- notice the same old trap independently.
If the skill disappears after a short delay, it needs another retrieval cycle before being treated as stable.
Step 8: vary the surface
Students often become good at recognising worksheets rather than recognising underlying structure.
Change:
- wording;
- numbers;
- context;
- representation;
- the order of information;
- the surrounding topic;
- the amount of support.
Variation asks whether the student learned the structure or only the example.
Step 9: transfer beyond the original lesson
A repaired skill becomes more valuable when the learner recognises it in a different environment.
Transfer may mean:
- using a sentence principle in a new composition;
- using an algebra relationship inside a later topic;
- applying a checking routine in a full examination paper;
- using evidence discipline in a new subject or discussion;
- recognising when a familiar method no longer fits.
The student should not need the original worksheet to remember what to do.
Step 10: return control to the learner
The final direction of teaching is independence.
Support should shrink:
model → guide → hint → wait → independent attempt → changed task → delayed return.
If performance only exists when the tutor is present, the learning cycle is unfinished.
What “understand first” should mean
The original 2017 article said “Understand first, practice lots”. That remains broadly useful, but understanding should be tested rather than declared.
Signs of stronger understanding include:
- the student can explain the relationship in their own words;
- they can show the idea in another representation;
- they can identify why a plausible wrong method fails;
- they can use the idea after the surface changes;
- they can recover when one detail is forgotten.
Practice then stabilises and extends the understanding.
What “practice lots” should not mean
Volume is not automatically quality.
More practice stops being efficient when:
- the same error repeats without diagnosis;
- the learner is copying a method mechanically;
- fatigue is increasing error faster than learning;
- every question is the same surface form;
- the student cannot perform once examples are removed.
Practice should produce a return: better retrieval, accuracy, discrimination, transfer or independence.
What “enjoy learning” should mean
Enjoyment can reduce friction, increase willingness to attempt and make discussion easier. It is not a magical accelerator that makes all content easy.
A productive learning environment can include:
- humour;
- interesting examples;
- choice;
- peer discussion;
- visible progress;
- tasks that are difficult enough to feel meaningful but not impossible.
A simple check is this: if a “fun” activity produces excitement but no improved understanding, retrieval or transfer, entertainment has replaced the learning job.
What “customised” should mean
Customisation should follow evidence, not personality labels.
It can mean changing:
- the entry point;
- the amount of scaffolding;
- the sequence of examples;
- the difficulty;
- the practice mix;
- the delay before retest;
- the representation used to explain the idea.
Customisation does not mean every learner gets a completely unrelated curriculum or that preference automatically decides method.
What “small group” should mean
A small group is useful when it creates both access to the tutor and space for independent work.
While the tutor works closely with one learner, another can:
- attempt independently;
- retrieve earlier work;
- correct a previous error;
- compare a method with a peer;
- prepare a precise question.
A small group becomes weak when every learner is simply waiting for the tutor to rotate back and provide the next step.
What “holistic” should mean
Holistic education should not mean trying to teach everything everywhere.
It means remembering that academic performance interacts with:
- attention;
- sleep and recovery;
- confidence calibration;
- study habits;
- communication;
- responsibility;
- ability to ask for help;
- ability to work without help.
World connections and enrichment can be valuable when they clarify a concept, widen context or create transfer. They should not displace the core learning job merely because the activity is memorable.
The learning-cycle map
| Stage | Student evidence | Tutor decision |
|---|---|---|
| Locate | Independent attempt | What is already owned? |
| Explain | Student can restate relationship | How much explanation is necessary? |
| Attempt | Visible working/response | Where does performance first break? |
| Narrow | Small discriminating task | Which explanation best fits? |
| Repair | Corrected new attempt | Did the intervention change behaviour? |
| Retrieve | Delayed recall | Did learning survive time? |
| Vary | Changed-form task | Did learning survive surface change? |
| Transfer | Use in mixed/new context | Is the capability portable? |
| Independent return | Performance without tutor/tool | Can support shrink? |
When the learning cycle needs to change
We should change course when:
- the same intervention produces the same failure repeatedly;
- the learner can perform only with prompts;
- more practice increases fatigue but not transfer;
- the explanation becomes longer while student understanding becomes less clear;
- the learner can repeat the rule but cannot apply it to a changed task;
- the lesson feels positive but the delayed independent return does not improve.
A method is not protected because it is “our method”. It must continue to earn its place from learner evidence.
AI fits inside the cycle, not above it
AI can provide explanations, variants, feedback and examples. It is useful when it helps a learner move through a specific bottleneck.
A useful pattern is:
student attempts → assistance addresses one failure → student repairs → task changes → assistance disappears → student returns independently.
If the tool creates polished work while the student’s independent performance does not change, the cycle has measured the tool rather than the learner.
Historical eduKate classroom evidence
The original 2017 article documented small-group lessons, customised materials, student discussion and enrichment. Those classroom photographs are useful historical provenance. Old stream labels, old contact details, proficiency-award claims and broad outcome assertions are retired. The page now describes a durable learning process rather than a dated service advertisement.
What should become visible in the learner?
What should come back from the learner if the cycle is working?
- faster recognition of familiar structures;
- fewer repetitions of the same error;
- better performance on changed questions;
- stronger retention after delays;
- more accurate self-diagnosis;
- less prompting;
- greater ability to explain why a method or answer fits;
- more stable independent performance.
Those returns matter more than whether the lesson looked busy.
The deeper principle
What we do at eduKate should be visible as a cycle of increasingly independent capability: locate accurately, explain proportionally, let the learner attempt, diagnose the first failure, repair specifically, retrieve later, vary the surface, test transfer and keep reducing the amount of support needed.
First published 24 July 2017 as “What do we do at eduKate?”. Rebuilt in 2026 as a reader-facing learning-cycle article. The original emphasis on understanding, practice, enjoyment, customisation, small groups and holistic education is preserved but made operational, testable and bounded; obsolete service claims and unrelated travel imagery are retired.