Wait, What? “Be More Specific” Is Not Yet a Repair
A teacher writes “explain”, “use evidence”, “be specific”, “wrong concept” or “answer the question” beside a PSLE Science response.
The learner reads the comment, copies the corrected answer neatly, and moves on.
Two weeks later, a different question produces the same error.
The feedback was seen. The wording was corrected. The underlying learner operation was never repaired.
Feedback becomes learning only when you can turn it into a specific change in your scientific reasoning and then prove that the change survives a new question.
A useful PSLE Science correction therefore has two parts:
- Repair: What scientific link must change?
- Receipt: What later question will show that the repair now works?
Quick Answer
Use this feedback-to-repair route:
READ THE ORIGINAL QUESTION → READ YOUR ORIGINAL ANSWER → READ THE FEEDBACK → LOCATE THE FIRST SCIENTIFIC MISMATCH → NAME THE REPAIR IN ONE SENTENCE → REDO WITHOUT COPYING → CHANGE THE SURFACE QUESTION → RETEST INDEPENDENTLY → RETURN AFTER A DELAY.
Do not translate feedback into a vague instruction such as “write better”. Translate it into a scientific action such as:
- identify the changed condition before choosing the concept;
- use one decisive table comparison;
- restore the missing causal mechanism;
- name the scientific object instead of writing “it”;
- keep the conclusion within the tested evidence;
- distinguish the result from the explanation;
- remove an irrelevant but true fact.
The Exact PSLE Science Learning Job This Guide Owns
This guide owns one learner job: how a Primary 5 or Primary 6 learner converts teacher, tutor or reference-answer feedback into one precise scientific repair, applies the repair without copying corrected wording, and proves that the repaired reasoning survives changed-context and delayed retesting.
It does not replace the correction-book guide, model-answer guide, error-type diagnosis guide or self-check guide. Those remain canonical for their own jobs.
This page owns the handoff:
Feedback received → scientific weak link identified → repair performed → independent evidence of repair.
Why This Matters in the 2026 PSLE Science Frame
For examination from 2026, Standard PSLE Science assesses the 2023 Primary Science syllabus. Official assessment objectives include knowledge with understanding, application of scientific knowledge, interpreting and analysing information, evaluating observations and methods, and communicating explanations and reasoning.
Feedback can therefore point to very different failures. “Wrong” may mean the concept was wrong. “Explain” may mean the mechanism was missing. “Use data” may mean the learner ignored evidence. “Too broad” may mean the conclusion travelled beyond what was tested.
One correction method cannot repair every type of error.
Feedback Is Evidence About a Performance, Not a Complete Diagnosis
A comment tells you something about one answer. It does not always tell you the full cause of the error.
For example, “use evidence” could arise because the learner:
- did not notice the table;
- read the table incorrectly;
- noticed the right evidence but omitted it in writing;
- used the wrong comparison;
- thought evidence was unnecessary because the concept felt obvious.
The learner should therefore inspect the original reasoning before deciding the repair.
The Six Main Feedback Families
| Feedback family | Possible scientific problem | Repair question |
|---|---|---|
| Concept | Wrong scientific model | What relationship should replace my current one? |
| Evidence | Ignored/misread data or observation | Which exact observation supports the answer? |
| Mechanism | Outcome restated without why | What process connects condition to outcome? |
| Condition | Generic fact not bound to question | Which stated condition makes this concept relevant? |
| Representation | Correct thinking lost in writing/diagram | Which scientific meaning disappeared? |
| Scope | Overclaim / irrelevant information | What is the strongest conclusion the evidence actually supports? |
Worked Example 1 — “Use the Data”
Original practice situation: A table compares the mass of water remaining in two identical dishes after the same time. Dish P has 62 g remaining and Dish Q has 81 g.
The learner writes:
“P evaporates faster because it has a larger surface area.”
Feedback: “Use the data.”
Weak repair:
“Remember to use data next time.”
Testable repair:
“Before writing the mechanism, I will identify the one comparison that shows the outcome: after the same time P has less water remaining than Q.”
Retest with a new table on a different topic. If the learner independently selects the decisive comparison and connects it to the mechanism, the repair has transferred.
Worked Example 2 — “Explain Why”
The learner writes:
“The bulb is brighter because the circuit makes it brighter.”
Feedback: “Explain why.”
The problem is not sentence length. The answer repeats the result.
Testable repair:
“When the question asks why, I will identify the scientific process between the stated circuit condition and the bulb outcome instead of repeating the outcome.”
Retest with a heat, force or plant question so the learner cannot memorise the circuit wording.
Worked Example 3 — “Be More Specific”
The learner writes:
“It moves faster because it has more force.”
Feedback: “Be specific.”
That comment could refer to the object, comparison, condition or mechanism.
Inspect the answer. If “it” could refer to two objects, the repair is reference precision:
“At the causal step, I will name the exact object if a pronoun could be ambiguous.”
Do not translate “be specific” into “write more”.
Worked Example 4 — “Wrong Concept”
A learner uses heat-transfer language in a question whose decisive relationship is evaporation.
Feedback: “Wrong concept.”
Testable repair:
“Before choosing the concept, I will identify what was directly changed and what outcome was measured, then ask which scientific relationship connects those two.”
The retest should hide the topic label and change the surface context.
Worked Example 5 — “Too General”
An investigation tests three temperatures and the learner concludes:
“Higher temperature always makes the process faster.”
Feedback: “Conclusion too broad.”
Testable repair:
“I will state the relationship only for the tested range unless the question and scientific mechanism justify a wider claim.”
Retest with a graph that contains a plateau or turning point so the learner must control conclusion scope.
Worked Example 6 — The Corrected Answer Is Not the Repair
The teacher supplies a scientifically complete model answer.
The learner copies it perfectly.
What has been proven?
Only that the learner can copy the correction.
To prove learning, remove the model answer and ask the learner to:
- state what was wrong in the original answer;
- name the scientific repair;
- answer a changed question;
- return after a delay.
Convert Comments Into “Next-Time Rules”
A useful next-time rule is specific, observable and scientifically meaningful.
| Weak rule | Stronger testable rule |
|---|---|
| “Read carefully.” | “Before solving, underline the changed condition and circle the measured outcome.” |
| “Explain more.” | “Add the process that connects the condition to the outcome.” |
| “Use keywords.” | “Use the scientific term only inside a relationship that explains what it does here.” |
| “Don’t be careless.” | “Check the comparison reference and unit before writing the conclusion.” |
| “Be precise.” | “Name the scientific object where a pronoun could refer to more than one thing.” |
| “Use evidence.” | “Choose the smallest table/graph comparison that directly supports the claim.” |
The First-Mismatch Method
When feedback is broad, compare the answer against the PSLE Science reasoning chain:
GIVEN INFORMATION → OBJECT/RELATIONSHIP → OBSERVATION VS INFERENCE → CONCEPT → MECHANISM → QUESTION CONDITION → OUTCOME → EVIDENCE CHECK.
Find the first place your answer stops matching the question or Science.
Repair from there. Later errors may disappear automatically.
Feedback Can Point to Different Causes Even When the Comment Is the Same
Two learners both receive “use evidence”.
- Learner A never read the graph correctly.
- Learner B read the graph correctly but omitted the data from the final answer.
Same comment. Different repair.
This is why feedback should be connected to the learner’s original reasoning, not only the final corrected sentence.
The Repair Sentence
After reviewing feedback, complete:
“My earliest error was ______. Next time, before I ______, I will ______ so that ______.”
Example:
“My earliest error was choosing a familiar concept before checking the changed condition. Next time, before I select the concept, I will identify the changed condition and measured outcome so that the concept answers the actual relationship being tested.”
This is a learning scaffold, not an exam sentence.
The Immediate Redo Is Necessary but Not Sufficient
Redo the original question after understanding the feedback. But recognise that the original context is now familiar.
A stronger sequence is:
- Redo original question.
- Explain what changed in your reasoning.
- Answer a new question with the same learner job.
- Change representation: text → table → graph → diagram where appropriate.
- Return after several days.
How to Build the Changed Retest
Keep the weak link constant while changing the surface.
If the feedback was “wrong comparison”, the retest should require the learner to choose the correct comparison in a different topic.
If the feedback was “no mechanism”, the retest should ask for a causal explanation using a different scientific concept.
If the feedback was “overclaim”, the retest should provide limited evidence that tempts the learner to generalise too far.
Do not simply replace names P and Q with X and Y while leaving the entire wording identical.
When Feedback Should Trigger Relearning
- The learner cannot explain the relevant concept in a simple example.
- The mechanism is wrong even without time pressure.
- The same misconception appears across several contexts.
- Prompting helps only after the teacher supplies the missing scientific relationship.
In these cases, fix the scientific model before practising answer phrasing.
When Feedback Should Trigger More Application Practice
- The learner can explain the concept independently in a simple case.
- The error appears only when the representation changes.
- The learner chooses the wrong concept from several plausible concepts.
- The reasoning works after the question is simplified but fails in unfamiliar contexts.
Now the concept may be known, but selection or transfer needs strengthening.
When Feedback Should Trigger Representation Practice
- Oral reasoning is complete but written reasoning loses a link.
- Evidence is identified but not included in the answer.
- A correct scratch model is not translated into words.
- Pronouns or vague comparatives hide the scientific object.
Observable Failure Signatures
- The correction book contains many copied model answers but the same errors recur.
- Every feedback note becomes “be careful”.
- The learner can explain what the teacher wrote but not what changed in their own reasoning.
- The original question can be redone immediately, but changed questions still fail.
- The learner remembers the corrected sentence more strongly than the scientific relationship.
- Feedback is sorted by topic only, so the same inquiry or explanation error remains hidden across chapters.
The Earliest-Weak-Link Diagnostic
| Feedback pattern | Possible earliest weak link | Best next test |
|---|---|---|
| “Wrong concept” repeatedly | Concept selection from conditions | Hide topic label and choose concept from evidence |
| “Use evidence” repeatedly | Evidence extraction or written inclusion | Ask learner to point to decisive evidence before writing |
| “Explain” repeatedly | Mechanism missing | Require condition → process → outcome chain |
| “Be specific” repeatedly | Object/reference/comparison ambiguity | Identify exactly which noun or reference is unclear |
| “Too broad” repeatedly | Conclusion scope | Use limited tested range and ask what can/cannot be claimed |
| “Careless” only under time | Execution rather than knowledge | Compare timed vs untimed reasoning |
Misconception Repair — “Copying the Correction Means I Learned It”
Copying can help you inspect a complete answer, but it does not prove that you can generate the reasoning independently.
Misconception Repair — “Every Teacher Comment Is a Precise Diagnosis”
Feedback is often deliberately brief. Use the original question and answer to identify the exact scientific meaning behind the comment.
Misconception Repair — “More Feedback Is Always Better”
A page covered in comments can overwhelm the learner. Prioritise the earliest meaningful scientific weak link and repair that first.
Misconception Repair — “A Corrected Answer Should Sound Exactly Like the Model Answer”
The scientific relationship must be correct. Exact wording can vary unless the question requires a specific term or official guidance establishes a requirement.
The Five-Minute Feedback Conversion
- Read the question again.
- Read your original answer exactly as written.
- Read one piece of feedback.
- Find the first mismatch in evidence, concept, mechanism, condition, representation or scope.
- Write one next-time rule.
- Redo the question without copying.
- Choose one changed retest.
The Weekly Feedback Audit
At the end of a week, group feedback by learner job rather than chapter.
| Feedback cluster | What it may reveal |
|---|---|
| Evidence errors across heat, plants and circuits | Cross-topic evidence-reading gap |
| Mechanism missing across several open-ended items | Explanation-construction gap |
| Correct concept but wrong condition | Question-condition binding gap |
| Conclusions too broad | Evidence-boundary gap |
| Only late-paper errors | Possible time/execution pattern |
This turns feedback from isolated red marks into a map of recurring scientific operations.
Unfamiliar Transfer Challenge
Previous feedback said: “You described the trend but did not explain the mechanism.”
New question: a completely different setup produces a graph showing one quantity increasing as another condition changes.
Can the learner:
- state the observed pattern;
- select the relevant scientific concept;
- explain the mechanism;
- connect it to the changed condition;
- return to the measured outcome?
If yes, the feedback has become transferable learning.
Delayed Independent Return
Three to seven days later, use a new question with the same learner job and no visible correction notes.
- Does the learner remember the repair principle?
- Can they apply it without the original teacher phrase?
- Does the same error return under a new representation?
- Can they explain what they checked before answering?
The delayed result is a stronger receipt than a perfect immediate correction.
The Feedback-Repair Receipt
- Can I state what my original answer did wrong scientifically?
- Can I identify the earliest weak link?
- Did I turn the feedback into one specific next-time action?
- Can I redo the original question without copying?
- Can I solve a changed question that tests the same weak link?
- Did I preserve evidence, object, concept, mechanism, condition and outcome?
- Can I do it again after a delay?
- Has the same feedback stopped recurring across topics?
Evidence and Model Limits
Feedback quality depends on the task, the learner’s prior knowledge, the clarity of the comment and what the learner does next. A short teacher comment may be highly useful when the learner understands the context; a long comment may still fail if it replaces the learner’s reasoning.
This guide does not claim one feedback method is always best. Its narrower principle is that feedback should lead to an observable repair in future independent performance.
Do not infer a clinical or learning disorder from recurring Science feedback patterns. They are educational evidence that should be interpreted cautiously.
Useful Internal Routes
- How to Use a PSLE Science Correction Book as a Scientific Learning Tool
- How to Learn From a Model Answer Without Copying Its Wording
- How to Compare Your First and Corrected Answers
- How to Decide Whether a Mistake Needs Relearning or More Application Practice
- How to Find the First Broken Link in an Explanation
- How to Retest a Corrected Mistake After a Delay
- Primary Science | Complete P1–P6 and PSLE Science Guide
Parent and Tutor Teaching Guide
When giving feedback, the most useful next question is often not “Can the child copy the correction?” but “Can the child identify what scientific relationship must change next time?”
Ask the learner to explain:
- what the question asked;
- what their original answer did;
- what the feedback is pointing to;
- what one next-time rule they will use;
- what new question would test whether that rule works.
If the learner cannot identify the scientific mismatch, the feedback may need unpacking. If they can identify it, do not over-support the next attempt. Let the changed question provide the receipt.
One precise repair that survives three new contexts is more valuable than ten beautifully copied corrections.
Authoritative and Research References
- Singapore Examinations and Assessment Board — PSLE Science syllabus, for examination from 2026.
- Singapore Ministry of Education — Science Teaching and Learning Syllabus, Primary, 2023.
- Education Endowment Foundation — Teacher Feedback to Improve Pupil Learning.
- Education Endowment Foundation — Metacognition and Self-Regulated Learning.
- Hattie & Timperley — The Power of Feedback.
- Pedaste and colleagues — Phases of Inquiry-Based Learning.
The research supports broader principles of feedback, monitoring and inquiry. It does not prescribe a PSLE marking scheme or require one specific correction-book format.
The Quiet Ending
A red mark is not the end of an answer.
It is evidence about where the last performance broke.
Find the first scientific mismatch. Turn it into one next-time rule. Remove the old wording. Change the question. Return later.
When the repair survives, the feedback has finally done its job.
