How to Keep Your Own Earlier Answer Separate From Given Evidence in a PSLE Science Structured Question

Wait, What? Your Answer Is Not Automatically Part of the Question

Imagine an original PSLE Science practice question with several parts. In part (a), you predict that the temperature will decrease. In a later part, the question gives the actual result: the temperature increased.

A learner who has become attached to the earlier prediction may start explaining the later result as though the temperature decreased. The learner is no longer reasoning from the question’s evidence. The learner is reasoning from their own earlier answer.

Your earlier answer is part of your working. It becomes evidence for a later part only when the later question explicitly tells you to use it or when its result has been independently established by the question.

This is a subtle PSLE Science control skill. In a long structured question, the page can contain printed information, diagrams, tables, later supplied results and your own written responses. They can sit close together and still have different scientific status.

Quick Answer

When you move from one sub-part to the next, keep three sources separate:

QUESTION-GIVEN: facts, diagrams, measurements, conditions, tables and later information supplied by the question.
YOUR WORK: your prediction, calculation, inference, label, explanation or choice.
NEW QUESTION EVIDENCE: additional results or information supplied in a later sub-part.

Then use this control chain:

READ THE CURRENT SUB-PART → IDENTIFY WHAT THE QUESTION GIVES → LABEL WHAT CAME FROM YOUR OWN EARLIER WORK → USE ONLY THE SOURCES THE CURRENT TASK ACTUALLY AUTHORISES → IDENTIFY THE SCIENTIFIC OBJECT OR RELATIONSHIP → SELECT THE RELEVANT CONCEPT → REASON TO THE CURRENT OUTCOME → CHECK YOUR CLAIM AGAINST THE QUESTION’S EVIDENCE.

The Exact PSLE Science Learning Job This Guide Owns

This guide owns one dominant job: preserving the provenance of information in a PSLE Science structured question so the learner can tell question-given evidence from their own earlier answer.

It is not a guide to official follow-through marking, and it makes no claim about how marks are awarded when an earlier part is wrong. It is not a guide to the chronology of when information becomes available as a separate examination concept. It is not a general note-taking method.

The student job is simpler and more fundamental: do not silently promote your own answer into evidence.

For the wider route, return to the PSLE Science Learning Guide. For keeping prediction and conclusion in different roles, use How to Keep a PSLE Science Investigation Question, Prediction and Conclusion in Different Jobs.

Why This Matters in Current PSLE Science

Standard PSLE Science from 2026 assesses the 2023 Primary Science syllabus. The official learning and assessment frame includes interpreting and analysing information, evaluating observations, information and methods, making predictions or formulating hypotheses, and communicating explanations and reasoning.

Those jobs depend on information discipline. A prediction is not the same thing as an observation. A learner-generated calculation is not the same thing as a printed measurement. A conclusion should be checked against actual results rather than against what the learner hoped would happen.

The Core Distinction: Source and Scientific Status

Statement on the page	Source	Scientific status	Can you automatically treat it as later evidence?
“The water started at 80°C.”	Question stem	Given condition / measurement	Yes, while relevant and not superseded by later stated information.
“I predict the water will cool faster in cup P.”	Your answer	Learner prediction	No.
“After 10 minutes, P was 62°C and Q was 68°C.”	Later results supplied by question	Given evidence	Yes for later reasoning where relevant.
“Therefore P is made of metal.”	Your inference	Learner-generated claim	No, unless independently established.
“Use your answer from part (b).”	Current instruction	Explicit task direction	Use the earlier response because the question tells you to.

This table is a learning tool, not an official marking rubric. Its purpose is to make source provenance visible.

Why the Error Happens

Once you write something down, it becomes visually present on the same page as the printed question. The eye can stop remembering where it came from. This is especially dangerous when:

the structured question has several sub-parts;
a prediction is followed by actual results;
you calculate a value in one part and later see a related table;
you assign an identity to P, Q, X or Y and then continue as though that identity was stated;
you make an inference from a diagram and later forget it was an inference;
the question adds new evidence that conflicts with your earlier answer.

The scientific repair is not “be more careful”. It is to keep the source of each claim attached to the claim.

The Provenance Protocol

1. Re-read the current sub-part as a new job

Do not begin from the sentence you just wrote. Begin from the current prompt. Ask: “What am I being asked to produce now?”

2. Point to the evidence source

Is the information in the stem, diagram, results table, caption, method or a new statement in the current part? Physically pointing during practice can help. The aim is eventually to perform the check mentally.

3. Mark your own earlier contribution mentally

Was it your prediction, calculation, inference, explanation or classification? Give it the label MY WORK until the question independently establishes it.

4. Check whether the current task explicitly refers back to your answer

If the question says “using your answer to part (a)” or otherwise explicitly directs you to use an earlier response, follow that wording. If it instead supplies new evidence, reason from the new evidence.

5. Rebuild the Science from authorised information

Identify the scientific object, relationship and conditions again. Select the relevant concept. Do not force the current reasoning to defend an earlier answer merely because you wrote it.

6. Check for contamination

Ask: “If I erased my previous response, would this later answer still be supported by the printed or explicitly referenced information?” If not, you may have turned your own work into evidence.

Worked Example 1: Prediction vs Actual Result

An original practice investigation compares two identical containers of warm water with different covers. Before seeing the results, part (a) asks you to predict which container will have the higher temperature after 20 minutes. You predict Container A.

A later table supplied by the question shows that Container B actually has the higher temperature.

If the next task asks you to explain the actual result, your earlier prediction is not evidence. The table is. You must now reason from B’s measured result and the relevant properties or conditions given by the question.

This does not mean the earlier prediction was useless. Comparing prediction with result is valuable scientific learning. It means the prediction and the observation have different jobs.

See How to Compare a PSLE Science Prediction With the Actual Result for that separate job.

Worked Example 2: Your Calculation vs a Later Printed Value

Part (a) gives measurements and asks you to calculate the increase in a quantity. You calculate 12 units. Later, the question provides a summary table containing a value of 10 units for the same named quantity and asks you to compare two cases.

First check whether the two values genuinely refer to the same quantity, interval and case. If the later table explicitly supplies 10 units as the evidence for the later comparison, do not overwrite it with 12 merely because 12 is written above on your page.

Your calculation remains your work. The printed table remains question-given information. The current prompt determines which source is relevant.

Worked Example 3: Your Label for an Unknown Object

A diagram labels an unfamiliar object X. From one observation, you infer that X may be a particular type of material. You write that material name in part (a).

In part (b), the question asks how X behaves under a new condition. Do not automatically treat your material identification as a printed fact. Return to the properties and relationships actually supplied for X. If the identification was not uniquely supported, the later answer must not depend on it as certainty.

This protects you from an especially damaging chain: one early over-inference becomes a false premise for several later answers.

Worked Example 4: Your Variable Label vs the Method

You read an investigation and write that “amount of water” is the changed variable. Later, another part asks you to evaluate the method. The method itself shows that amount of water was kept the same while the surface area was changed.

Do not defend the earlier label. Re-read the method. Scientific evidence and method structure outrank your own mistaken annotation.

A learner who can revise the internal model when evidence disagrees is doing better Science than a learner who stays consistent with a wrong answer.

When Your Earlier Answer Is Explicitly Needed

Some structured tasks legitimately refer back. A later prompt may say “Using your answer in part (a)…” or may ask you to compare your prediction with the observed result. In that case, your earlier answer has a defined role in the current task.

The important distinction is explicit reference. Do not assume every later part wants your earlier answer just because it is nearby.

What This Guide Does Not Claim About Marking

This guide does not tell you whether an examiner will award follow-through marks after an earlier error. It does not describe a hidden marking policy. It does not tell you to deliberately contradict an instruction that explicitly asks you to use a previous answer.

Its purpose is scientific reasoning control: know which statements are question-given evidence and which statements are products of your own reasoning.

The “Erase Test”

During practice, imagine erasing your answer to the previous part while leaving the printed question untouched. Can you still justify the current answer from the remaining evidence and instructions?

If yes, your current reasoning is likely anchored to the question. If no, inspect whether the current part explicitly told you to use your earlier answer. If it did not, you may have created a false dependency.

The “Source Tag” Method for Training

For a week of practice, put tiny source tags beside important statements on your scratch paper:

G = given by question
O = directly observed or measured in supplied evidence
C = your calculation
I = your inference
P = your prediction
E = your explanation

This is a temporary learning scaffold, not an examination requirement. Its purpose is to make provenance visible until you can track it without writing the tags.

How This Protects the PSLE Science Reasoning Chain

The series repeatedly trains:

OBSERVE / READ GIVEN INFORMATION → IDENTIFY THE SCIENTIFIC OBJECT OR RELATIONSHIP → DISTINGUISH OBSERVATION FROM INFERENCE → SELECT THE RELEVANT CONCEPT → EXPLAIN THE CAUSAL MECHANISM → CONNECT TO THE QUESTION’S CONDITION → STATE THE OUTCOME → CHECK AGAINST THE EVIDENCE.

Source confusion damages the very first step. If “given information” quietly includes your own unsupported earlier answer, every later step can be beautifully written and still rest on the wrong foundation.

Observable Failure Signatures

What happens	Likely weak link	Earliest repair
Learner keeps defending a prediction after contradictory results appear	Prediction became pseudo-evidence	Put prediction and actual result in separate boxes.
One wrong identity for X causes several later wrong answers	Learner-generated label became a fact	Return to the printed properties of X before each new job.
Later answer uses a calculated value even when a new value is supplied	Source provenance is lost	Label values by source and quantity.
Learner rereads only their previous response, not the new prompt	Current task is not re-anchored	Read the current command before looking upward.
Learner changes the question’s evidence to stay consistent with an earlier answer	Consistency is valued over evidence	Use evidence as the fixed reference and revise the learner model.
Learner refuses to use earlier work even when the prompt explicitly asks for it	Rule applied too rigidly	Follow explicit cross-reference wording.

Misconception Repair: Five False Rules

False rule: “If I wrote it in part (a), it is now true for part (b).” Repair: your answer keeps its original status unless the question establishes or explicitly reuses it.
False rule: “I must stay consistent with my first answer.” Repair: Science stays consistent with evidence; your model should change when evidence changes.
False rule: “A calculation is more reliable than a printed result because I worked it out.” Repair: source and quantity determine relevance, not personal ownership.
False rule: “Every later part starts from zero.” Repair: carry forward question-given context that still applies, but preserve source status.
False rule: “Never use a previous answer.” Repair: use it when the current prompt explicitly requires it.

How to Recover When You Notice Source Contamination Mid-Question

Stop defending the current chain.
Return to the current sub-part.
Underline only information supplied or explicitly referenced by the question.
Put a mental bracket around your previous answer.
Re-identify the scientific object and relationship.
Rebuild the answer from the authorised evidence.
Check whether the new answer conflicts with your earlier response. If it does, accept the difference unless the current prompt requires consistency with that response.

This is not “starting over”. It is removing one contaminated premise.

Practice Sequence: Provenance Before Speed

Choose an original three-part Science question.
Before answering, highlight all given information in one way.
Answer part (a).
Mark your answer as MY WORK.
Read part (b) and identify what it uses from the original question.
Add any new supplied evidence to a separate list.
Answer part (b) without silently borrowing unsupported claims from part (a).
Repeat for part (c).
Remove the highlighting on the next practice item and do the provenance tracking mentally.

Unfamiliar Transfer Challenge

An original question shows three opaque boxes P, Q and R. Part (a) asks you to infer which box contains a magnet based on the movement of a nearby object. You choose Q. Part (b) then gives a new observation: only box R attracts an iron nail. The current task asks which box is supported by the combined evidence.

Your earlier choice Q is not equivalent to the new observation about R. Treat Q as your earlier inference and R’s attraction as new evidence. Re-evaluate. The skill transfers because the scientific topic changed, but the provenance job stayed the same.

Delayed Independent Return Test

After a gap, complete a multi-part original Science task without source tags. At the end, choose three statements from your working and answer:

Was this given by the question or produced by me?
If produced by me, was it a calculation, inference, prediction or explanation?
Did any later answer rely on it as though it were direct evidence?
If yes, did the later prompt explicitly permit that use?

If you can reconstruct the source history accurately without seeing the scaffold, the skill is becoming independent.

Answer-Checking Receipt

What does the current sub-part actually ask?
Which evidence is printed or explicitly supplied by the question?
Which statements came from my own earlier work?
Did the current prompt explicitly ask me to use an earlier answer?
Have I promoted a prediction into an observation?
Have I promoted an inference into a fact?
Have I promoted a calculation into a printed measurement?
If new evidence disagrees with my earlier model, did I update the model?
Can I point from my current conclusion back to authorised evidence?

Parent and Tutor Teaching Guide: Ask “Where Did That Come From?”

When a child produces a later answer that looks strangely detached from the question, do not immediately correct the Science. Ask, “Where did that information come from?”

The child may point to a previous answer rather than to the stem, diagram or results table. That is a valuable diagnosis: the earliest failure is provenance, not necessarily concept knowledge.

During teaching, use different boxes for given evidence and student thinking. When new evidence arrives, place it with the given evidence. Ask whether the student’s earlier model survives. Then fade the boxes.

Do not teach the child to fear changing an answer or model. Scientific reasoning improves by revising claims when stronger evidence appears.

Useful Internal Routes

Authoritative References and Evidence Boundary

The provenance routines in this guide are eduKate learning tools. They are not official examination annotations, marking rules or claims about error-carried-forward credit. The durable scientific principle is that observations, given information, predictions, calculations, inferences and explanations should retain their correct roles.

The Quiet Return

Good Science is willing to change its mind without changing the evidence.

Keep the question’s information where it belongs. Keep your own thinking where it belongs. Then let evidence—not attachment to an earlier answer—decide what you should say next.

Small Group Tutorials