Quick Read
Teaching changes something only if the change survives outside the moment of teaching.
That creates a simple problem.
How do we know whether the learner actually changed?
Not whether the explanation sounded good. Not whether the worksheet was finished. Not whether the student nodded. Not even whether one immediate answer was correct.
What evidence returned from reality?
The Evidence Loop is the return path: observe → form a cautious hypothesis → choose a useful test → teach or practise → observe again → update.
The One-Sentence Answer
Learning becomes trustworthy when our explanation of the learner is repeatedly allowed to meet new evidence and change when reality disagrees.
Why a Loop Instead of a Straight Line?
Education is often imagined as a straight line.
Teach → practise → test → finish.
That is convenient. It is also incomplete.
A test can reveal that the teaching did not land as expected. Practice can reveal that the original diagnosis was wrong. A changed question can expose a transfer weakness that familiar questions concealed. A delayed retest can show that immediate success was temporary.
So the system needs to return.
Teach → observe → revise → teach differently → observe again.
That is not inefficiency. It is correction.
Step 1 — Observe What Actually Happened
The first step is deliberately modest.
Describe the event before explaining it.
For example:
- The student selected the wrong operation in four of six word problems.
- The student answered correctly after the teacher underlined one phrase.
- The student could explain evaporation but did not refer to the experimental evidence.
- The student completed familiar algebra accurately but stopped when the form changed.
- The student began independently today after needing prompts in the previous two lessons.
These descriptions do not yet tell us why.
That restraint is valuable.
Step 2 — Form the Smallest Plausible Hypothesis
Now we ask what might explain the pattern.
The key word is might.
A student who repeatedly chooses the wrong operation may have weak language comprehension, weak representation, a concept error, or a learned habit of searching for keywords. A child who hesitates may be confused, careful, anxious, distracted or simply checking.
The first explanation is not automatically the true explanation.
A good hypothesis narrows the next question. It does not pretend the investigation is already over.
Step 3 — Choose the Smallest Useful Test
If two explanations are plausible, we do not necessarily need another full worksheet.
We need a question that distinguishes them.
Suppose a Mathematics student gets ratio word problems wrong.
We can separate the problem:
- Give the same relationship without difficult wording. If the student solves it, language may be part of the bottleneck.
- Ask the student to draw or represent the situation before calculating. If the representation is wrong, the break appears earlier than arithmetic.
- Give a clean ratio calculation. If that fails too, the concept itself may be unstable.
One discriminating question can be more informative than twenty repetitions of the original failure.
Step 4 — Teach or Practise the Weakest Justified Link
Once the evidence is strong enough, intervention becomes smaller and calmer.
Instead of “You are weak in Mathematics,” the instruction might be:
Before choosing an operation, represent the part–whole relationship. We are going to stabilise that step first.
Instead of “Your English is poor,” it might be:
You can infer the idea when the evidence is pointed out. Now we are going to practise locating the evidence independently.
The repair is tied to evidence rather than identity.
Step 5 — Retest Under a Slightly Different Condition
Immediate repetition is useful, but it is a generous environment.
The learner has just seen the method. The language is familiar. The teacher may still be nearby.
So the Evidence Loop asks the capability to survive a change.
- wait before retesting;
- change the wording;
- change the representation;
- mix the topic with another;
- remove a prompt;
- ask for an explanation;
- or add examination pressure later, when appropriate.
The purpose is not to catch the student out. It is to discover what the learning can now survive.
Step 6 — Update Only as Far as the Evidence Allows
If the learner succeeds once, we do not declare permanent mastery.
If the learner fails once, we do not declare permanent weakness.
We update proportionately.
Perhaps the capability has moved from blocked to fragile. Perhaps it is now stable in familiar questions but not yet transfer-ready. Perhaps the concept is sound and only examination execution remains weak.
This connects directly to Which Student Are You?, where state belongs to a capability at a moment in time, not to the whole learner forever.
One Mark Is Not a Loop
A mark is useful because it gives us an external return.
But it becomes more informative when we can compare:
- which topics changed;
- which error types disappeared;
- which mistakes remained;
- whether timing improved;
- whether unfamiliar questions became more manageable;
- and whether the learner required less external support.
The number is therefore part of the evidence, not the entire educational explanation.
A Wrong Answer Can Be Excellent Evidence
Imagine two students.
Student A gets the answer correct after three prompts.
Student B chooses the correct method independently, performs most of the reasoning correctly, then makes one arithmetic error.
The mark alone may prefer Student A.
The teacher may learn more from Student B.
The Evidence Loop therefore asks:
What produced the answer?
That question often reveals the difference between performance and capability.
Unfinished Work Can Be a Sensor
Homework is interesting because it often removes the teacher.
No facial cue. No underlined phrase. No reminder of the previous example. No instant rescue.
That can reveal the exact point where independence ends.
An unfinished question is therefore not automatically laziness or failure. Sometimes it is the clearest piece of evidence available.
The job is to investigate before judging.
Signal, Noise and Repetition
Not every mistake is meaningful.
Students misread. They are tired. A pencil slips. A question is unusually strange. A correct method is ruined by one arithmetic accident.
This is why repeated pattern matters.
| Return | Interpretation discipline |
|---|---|
| One isolated mistake | Notice it; do not overdiagnose |
| Same error across several similar tasks | Pattern worth investigating |
| Same weakness after explanation | Repair may not have landed |
| Success only with prompts | Capability may still be externally supported |
| Success after delay and context change | Stronger evidence of stability and transfer |
The dashboard should become more confident only when the evidence becomes more robust.
The Evidence Loop Feeds Tutor
Tutor is the student-facing dashboard.
The Evidence Loop is how that dashboard earns the right to update.
Without a return path, Tutor would slowly become a collection of stale assumptions.
So after a learning action, evidence comes back:
What changed? What survived? What failed? What remains unknown? Does the next move stay the same?
The dashboard changes only because reality gave it a reason to change.
The Evidence Loop Gives Tutorial a Purpose Beyond Attendance
Tutorial is the bounded learning room.
The loop asks what the room produced.
Perhaps the lesson repaired a missing concept. Perhaps it only identified the real weak link. Perhaps the learner became more independent. Perhaps the teacher discovered that the planned intervention was aimed at the wrong layer.
A Tutorial therefore does not have to solve everything to be useful.
It has to improve the learner or improve our understanding of what the learner actually needs next.
Voyage Changes the Meaning of the Evidence
The same behaviour can mean different things at different stages.
A young learner needing a concrete representation may be exactly where expected. A Secondary student who still cannot operate without that representation may need a bridge toward abstraction.
The Voyage Series therefore provides developmental context for the evidence.
Darwin Uses the Evidence to Decide Whether Adaptation Is Needed
Evidence may show that the learner has not forgotten the old method at all.
The old method may simply have expired.
The Darwin Series takes that return and asks whether the learning environment has changed enough to require a new strategy, representation or level of independence.
This prevents endless repair of a method that should actually be replaced or extended.
Evidence Loops Across Subjects
English
Observation: inference improves when evidence is highlighted.
Hypothesis: locating evidence independently may be the weak link.
Test: unseen passage without highlighting.
Repair: teach evidence-location cues and paragraph structure.
Retest: new text after delay.
Update: move from supported inference toward independent inference only if it survives.
Mathematics
Observation: correct method, repeated sign errors.
Hypothesis: signed-number control is fragile.
Test: isolate negative-number operations.
Repair: rebuild the earliest weak link.
Retest: return to algebra in changed forms.
Update: do not call the algebra repaired until accuracy survives inside the larger problem.
Science
Observation: facts are correct but explanation does not use the given evidence.
Hypothesis: the student is recalling rather than reconstructing.
Test: unfamiliar experiment.
Repair: observation → evidence → mechanism chain.
Retest: another system with different surface details.
Update: stronger confidence only when the reasoning transfers.
Retesting Is Not Punishment
If students experience every check as suspicion, the measurement changes the environment in a harmful way.
The purpose of retesting is not to prove the learner wrong.
It is to answer a shared question:
Did the learning become yours?
When students understand that, errors can become informative rather than humiliating.
A Parent-Safe Version of the Loop
Parents do not need to reproduce the full teaching process at home.
A simple version is enough:
- Notice: what seems to keep happening?
- Ask: what does the child think is difficult?
- Check: is there one piece of work that shows the pattern?
- Avoid identity labels: describe the present problem, not the child.
- Return: after help, ask what the child can now do independently.
That is enough to create better conversations with the learner and the teacher.
Boundaries: What the Evidence Loop Cannot Do
- It cannot make one piece of evidence certain.
- It cannot reduce a whole child to observable school performance.
- It cannot infer medical or psychological conditions from ordinary educational behaviour.
- It cannot guarantee that improvement will be linear.
- It cannot remove the need for teacher judgement.
- It cannot turn every useful quality into a number.
Its job is smaller and more defensible: keep educational explanations correctable by new evidence.
The RFE of the Evidence Loop
Let every important educational claim remain answerable to what the learner can actually do when the conditions change and the help is progressively removed.
The First Curie Loop Is Now Complete
The four foundations now fit together:
- Marie Curie Series — why invisible learning needs instrumentation.
- Tutor — how the learner sees the present state and next move.
- Tutorial — where high-resolution teaching, practice and observation occur.
- Evidence Loop — how reality returns and changes what we think we know.
Then the larger eduKate Sengkang system continues:
Voyage locates → Curie observes → Tutor explains the present picture → Tutorial intervenes → Evidence returns → Darwin adapts where necessary → the learner moves again.
Frequently Asked Questions
How often should a learner be retested?
Only as often as useful. Some capabilities can be checked immediately and again after a delay; others reveal themselves through normal schoolwork. The series does not require constant testing.
What if the evidence contradicts the teacher’s original view?
Then the view should change. That is the point of the loop. Good teaching is not weakened by correction; it becomes better calibrated.
Does every improvement need a higher mark immediately?
No. Earlier indicators may include reduced prompting, better method selection, stronger explanation, improved retrieval, fewer repeated errors or successful transfer. Marks remain important, but capability often changes before a large score movement appears.
Marie Curie Series · Tutor · Tutorial · Voyage · Darwin
eduKate Sengkang | Small groups of up to 3 | 1.5-hour tutorials | Properly Taught Kids Shine a Bright Light Into the Future.
eduKate Sengkang 