How to Test a PSLE Science Explanation by Removing One Causal Link

Wait, what? One of the quickest ways to discover whether a sentence in your Science explanation is doing real work is to imagine taking that link away.

If your explanation says A leads to B, B leads to C, and C produces the outcome, ask: what should happen if B cannot occur? If your predicted outcome changes for a scientific reason, B may be a meaningful part of the mechanism. If nothing changes, B may be decorative, redundant, incorrectly placed—or the system may have another route to the same outcome.

This is not a trick for manufacturing an answer. It is a way to test the internal logic of an explanation during learning and practice.

Quick Answer

READ THE GIVEN EVIDENCE → BUILD THE CAUSAL CHAIN → CHOOSE ONE PROPOSED LINK OR NECESSARY CONDITION → IMAGINE A SCIENTIFICALLY POSSIBLE CASE WHERE THAT LINK IS ABSENT OR BLOCKED → KEEP OTHER RELEVANT CONDITIONS AS COMPARABLE AS POSSIBLE → PREDICT THE OUTCOME → COMPARE THAT PREDICTION WITH THE ORIGINAL EXPLANATION → DECIDE WHETHER THE LINK IS NECESSARY, CONTRIBUTORY, REDUNDANT OR STILL UNRESOLVED → RETURN TO THE ACTUAL EVIDENCE.

The final step matters most. The imagined case helps you test reasoning. It does not become new evidence about the original question.

The Exact PSLE Science Learning Job This Guide Owns

This guide owns one learner job: stress-testing a PSLE Science causal explanation by removing or blocking one proposed causal link or condition in a carefully controlled thought experiment, predicting what should happen, and using the result to inspect the explanation’s structure.

It does not own the underlying scientific concepts. Existing concept pages remain canonical. It also does not claim that PSLE candidates are formally required to perform counterfactual reasoning, to write “if this were removed” in their answers, or to use any fixed examination template. This is a learning and checking method for understanding mechanisms more deeply.

It is also different from merely asking whether the reverse of a relationship is true. Here, the learner keeps a proposed causal chain and removes one link to test what that link contributes.

Why This Fits the Current PSLE Science Frame

For examination from 2026, Standard PSLE Science assesses the 2023 Primary Science syllabus. Current SEAB assessment objectives include knowledge with understanding, application of scientific knowledge and scientific inquiry, including prediction or hypothesis, interpretation and analysis of information, evaluation of observations, information and methods, and communication of explanations and reasoning.

A learner who can test an explanation instead of merely repeating it is practising several of those habits at once: distinguishing evidence from inference, making a conditional prediction, checking whether a mechanism actually connects the condition to the outcome, and limiting conclusions to what the evidence can support.

An Explanation Is More Than a Row of True Facts

Consider this invented explanation:

The metal spoon becomes warmer because heat is transferred from the hotter water to the cooler spoon, causing the spoon’s temperature to increase.

The sentence has a structure:

temperature difference → heat transfer from hotter object to cooler object → spoon gains heat → spoon’s temperature increases

Now remove the transfer link in a scientifically meaningful imagined case. Suppose the spoon and water were perfectly prevented from transferring heat between them. Under that changed condition, the original mechanism no longer predicts the same warming through that route.

That tells you the heat-transfer link is not just a nice phrase added to the sentence. It is doing explanatory work.

The Causal-Removal Test

Use this test during practice, correction or revision.

Step 1: Separate evidence from explanation

Write only what the question actually gives: observations, measurements, diagram relationships and stated conditions. Do not let your imagined change rewrite those facts.

Step 2: Build the explanation as a chain

Expose the middle. Use arrows if necessary:

given condition → relevant scientific relationship → mechanism → changed state → observed outcome

If you cannot build the chain, the explanation may already contain a missing reasoning link.

Step 3: Choose one link to remove

Pick a causal link or condition that your explanation appears to depend on. Do not remove three things at once. If many conditions change, you will not know which change explains the different prediction.

Step 4: Check whether the imagined case is scientifically possible

A thought experiment is useful only if the changed condition makes scientific sense at the level of the problem. “Imagine gravity does not exist for this one object but everything else behaves normally” may be an unhelpful fantasy for a Primary Science reasoning task. “Imagine the conducting path is broken at this point” can be a valid changed case for a simple circuit.

Step 5: Keep other relevant conditions comparable

The point is to test one link. If you remove a conductor and also change the cell, bulb and wire arrangement, the predicted difference cannot be attributed cleanly to the chosen link.

Step 6: Predict before looking for a conclusion

Ask what the changed chain predicts. Does the outcome disappear, weaken, remain possible through another route, or become uncertain? The answer should come from the scientific model, not from what feels convenient.

Step 7: Classify what the link seems to do

Necessary in this model: without the link, the outcome cannot occur through the proposed mechanism.
Contributory: removing the link changes the outcome, but other routes or factors may still produce some effect.
Redundant in the explanation: removing the sentence changes nothing because it merely repeats another link or does not connect to the outcome.
Unresolved: the thought experiment does not discriminate clearly, so more evidence or a better model is needed.

Step 8: Return to the original question

Your imagined case is a reasoning probe. The final PSLE Science answer must still respond to the evidence and conditions in the actual question.

Worked Case 1: Break One Link in a Simple Circuit

Imagine an original question showing a cell, wires and a bulb connected in a complete circuit. The bulb lights.

A learner proposes this model:

complete conducting path → electric current can flow through the circuit → bulb receives the electrical effect needed to operate → bulb lights

Remove one link by imagining a gap in the conducting path while keeping the other components comparable. The model now predicts that current cannot flow through a complete path, so the bulb should not light through that circuit.

The exercise tests whether “complete circuit” is structurally connected to the outcome. It does not prove that every dark bulb in every situation must be caused by a gap. A faulty bulb or other issue could also prevent lighting. The same outcome can have different possible causes.

Worked Case 2: Remove the Screen From a Shadow Explanation

Suppose a light source shines toward an opaque object and a screen. A shadow is observed on the screen.

A learner explains:

light travels from the source → opaque object blocks some light → a region behind the object receives less direct light → the screen makes that region visible as a shadow

Now imagine removing only the screen. The blocked-light region in space may still exist, but there is no screen at that location on which to observe the familiar shadow image. This reveals something subtle: the screen is important to the observation “a shadow is seen on the screen”, but it is not what causes the opaque object to block light.

The causal-removal test can therefore separate the mechanism producing a condition from the condition allowing us to observe its result.

Worked Case 3: Remove Contact in a Heat-Transfer Situation

An original set-up places one end of a metal strip in contact with a warmer object while the other end begins cooler. After time passes, a point farther along the strip becomes warmer.

A proposed explanation includes heat transfer through the material from hotter regions toward cooler regions. Imagine inserting a suitable insulating break so the two metal sections no longer provide the same conducting path while keeping the heat source and observation time otherwise comparable.

If the original explanation depends on conduction through the continuous metal strip, breaking that pathway should change the predicted warming beyond the break. The imagined change tests the pathway, not merely the vocabulary word “conduction”.

Worked Case 4: When Removing One Factor Does Not Remove the Outcome

Now consider a more difficult case. A container of water loses heat to its surroundings through more than one route. A learner focuses on one route and assumes it is the sole cause of cooling.

Imagine reducing that one transfer route while other routes remain possible. The water may still cool, perhaps differently. Therefore the thought experiment does not support the statement “this was the only cause”. It may support the weaker statement that the route contributes to the overall change.

This is a crucial scientific distinction:

A CAUSE CAN CONTRIBUTE WITHOUT BEING THE ONLY CAUSE.

Primary learners do not need advanced causal terminology to use this idea. They need the habit of refusing to turn “matters” into “explains everything”.

Necessary Does Not Automatically Mean Sufficient

Suppose a bulb needs a complete circuit to light. A complete circuit alone may still not guarantee that a particular bulb lights if another necessary condition fails. Therefore:

if removing a condition stops the outcome, that condition may be necessary within the model;
but restoring that one condition does not automatically prove it is sufficient by itself;
other required conditions may also need to be present.

This protects learners from a common logical jump: “Without X, the outcome cannot happen; therefore X alone causes the outcome.” The second statement does not follow automatically from the first.

Observation, Counterfactual Prediction and Evidence Must Stay Separate

Reasoning object	What it means	What it cannot become
Given observation	What the question states was seen or measured	An imagined result
Proposed mechanism	The causal relationship used to explain the evidence	A fact simply because it sounds plausible
Counterfactual change	A scientifically possible imagined alteration used to test the mechanism	A new event that actually happened in the original investigation
Counterfactual prediction	What the model says should occur under the imagined change	Direct evidence about the original case
Final conclusion	What the real evidence plus appropriate scientific reasoning support	A claim broader than the evidence

The table gives the most important boundary in this guide: thought experiments test models; observations test the real case.

When the Removal Test Is Invalid

A counterfactual can be badly designed just as a practical investigation can be badly designed.

You changed several important conditions at once. Then the predicted difference cannot isolate the chosen link.
You imagined an impossible or incoherent state. The test no longer tells you much about the actual mechanism.
You removed an observation rather than a causal link. Deleting evidence from the page is not the same as changing the scientific system.
You assumed the outcome must completely disappear. Another pathway may still contribute.
You treated “no change” as proof that the link never matters. Your chosen test may simply be insensitive or the link may matter only under another condition.
You forgot the question’s timescale. A link might affect how quickly an outcome develops without changing the eventual state within the observed interval.

Failure Signatures in Student Explanations

Your explanation contains three Science keywords but removing one sentence changes nothing.
The mechanism jumps directly from the condition to the outcome.
You cannot say what would change if one proposed link were blocked.
You claim one factor is the only cause even though the evidence does not eliminate alternatives.
You invent an imagined result and later quote it as if the question supplied it.
You remove a link and silently change two other conditions.
You choose a counterfactual because it produces the answer you want rather than because it is scientifically valid.
You confuse “necessary in this model” with “sufficient by itself”.

Find the Earliest Weak Link

If the causal-removal test goes wrong, diagnose in this order.

Evidence failure: Did the learner misread what actually happened?
Object failure: Is the causal link attached to the wrong object, part or system?
Relationship failure: Is the selected concept actually relevant?
Chain failure: Is an intermediate mechanism missing?
Counterfactual validity failure: Is the imagined removal scientifically coherent?
Control failure: Were other important conditions changed at the same time?
Prediction failure: Does the predicted outcome really follow from the altered model?
Scope failure: Did the learner make a stronger causal claim than the test can support?

Repair the first broken stage. A beautifully worded prediction cannot rescue a counterfactual that changed the wrong part of the system.

Misconception Repair

“If removing X changes the outcome, X is the only cause.”

No. X may be necessary or contributory in the model while other conditions also matter.

“If removing X does not eliminate the outcome, X does not matter.”

No. X might affect degree, speed, efficiency or probability while another route still produces some outcome. Check what the scientific model actually predicts.

“The imagined case counts as evidence.”

No. It is a reasoning test. Only the observations, measurements and information actually supplied or obtained are evidence about the original case.

“Every explanation should pass a one-link removal test in the same way.”

No. Some systems have parallel causes, feedback, thresholds or multiple routes. The method is useful precisely because it can reveal that a simple chain is too simple.

Practice Ladder: From Obvious Link to Competing Routes

Level 1: One obvious necessary link

Use a simple original circuit, light-path or heat-transfer situation. Build a three-link chain. Remove one clearly necessary connection and predict the changed result.

Level 2: Remove a condition, not an object

Keep the same objects but alter one relevant condition. This trains the learner to see that scientific relationships depend on conditions, not only on named objects.

Level 3: Add a second possible route

Use a system in which two factors can contribute. Remove one. Ask whether the outcome should vanish, weaken or remain possible. This repairs the “one cause explains everything” habit.

Level 4: Give an explanation with one decorative sentence

Insert a scientifically true but irrelevant fact into an explanation. Ask the learner to remove each sentence in turn and decide which removal changes the causal chain. This teaches relevance, not just correctness.

Level 5: Change the representation

Present the same reasoning job through a diagram, table or graph. The learner must reconstruct the chain before performing the removal test. This checks whether the skill survives outside familiar prose.

Practice Set With Explained Answers

Practice A: Circuit

A bulb lights in a complete circuit. A learner says the complete conducting path is a necessary condition in the proposed explanation. What imagined change would test that link?

Answer: introduce a gap in the conducting path while keeping the other relevant components and conditions comparable. The model predicts the bulb should no longer light through that path. This supports the importance of the complete path in the proposed mechanism but does not prove that every non-lighting bulb is caused by a gap.

Practice B: Shadow

A shadow is visible on a screen behind an opaque object. What does removing the screen test?

Answer: it tests the role of the screen in making the blocked-light pattern visible at that location. It does not test whether the opaque object blocks light, because that mechanism can still occur without the screen.

Practice C: Two Contributors

An outcome can be produced by two pathways. One pathway is removed and the outcome becomes smaller but does not disappear. What is the safest conclusion?

Answer: the removed pathway appears to contribute within the model, but it was not the only route capable of producing the outcome. The evidence does not support calling it the sole cause.

Practice D: Bad Thought Experiment

A learner tests one explanation by changing the material, temperature, size and observation time together. Can the learner tell which causal link changed the outcome?

Answer: not cleanly. Too many relevant conditions changed. A useful causal-removal test should alter one target link or condition while preserving other important conditions as far as the reasoning task allows.

Unfamiliar Transfer

After a learner succeeds with circuits, do not give five more circuits. Move to a different PSLE Science context. The surface objects should change while the reasoning job stays stable:

reconstruct the evidence;
build the causal chain;
select one meaningful link;
design a scientifically possible removal or block;
predict the changed outcome;
decide what the result says about the link;
return to the real evidence and write the actual explanation.

Transfer is stronger when the learner can do this without being told which link to test.

Delayed Independent Return Test

Several days later, give one unfamiliar original question with a plausible causal explanation already written. Ask the learner to mark:

the given evidence;
the proposed causal chain;
one link worth testing;
one valid changed condition that removes or blocks it;
the predicted outcome;
the strongest conclusion the thought experiment permits;
one conclusion it does not permit.

If the learner performs the reasoning only after a tutor names the missing link, the skill is not yet independent. Return later with less support.

Answer-Checking Receipt

Before accepting a repaired explanation, check:

Evidence: Did I preserve what was actually observed?
Object: Is each causal link attached to the correct object or system?
Concept: Is the scientific relationship relevant?
Mechanism: Does one step genuinely lead to the next?
Counterfactual: Is my imagined removal scientifically possible?
Control: Did I avoid changing several important conditions at once?
Prediction: Does the changed outcome follow from the altered model?
Scope: Did I avoid claiming sole causation, proof or certainty beyond what the test supports?
Return: Does my final answer still answer the actual question rather than the imaginary one?

This receipt is a study tool, not an official marking rubric.

Parent and Tutor Teaching Guide

When a child gives a long explanation, avoid asking only, “Did you use the right keywords?” Instead point to one causal sentence and ask, “If this part could not happen, what would your explanation predict?”

If the child cannot answer, the sentence may be memorised rather than understood. If the child predicts confidently but changes several conditions at once, the weak link is causal control. If the child says the outcome must disappear completely, ask whether another route could still produce part of the result.

For younger learners, draw the explanation as four boxes connected by arrows. Cover one box with a card. Ask what can still happen and what can no longer happen. Then uncover the box and return to a written explanation. The physical action makes the reasoning visible without turning the lesson into abstract philosophy.

For stronger learners, provide two explanations that both fit the first observation. Ask them to design one changed condition that would make the models predict different outcomes. This trains discrimination between competing explanations rather than loyalty to the first plausible story.

Useful Internal Routes

Authoritative References

The Quiet Return

A scientific explanation should be able to survive questions about its own machinery.

So when an answer sounds correct, do not stop at familiarity. Expose its causal chain. Remove one link in a valid imagined case. Predict what should change. Notice whether the link was necessary, contributory, redundant or unresolved. Then return to the evidence that really exists.

The goal is not to become clever at imaginary scenarios. The goal is to understand why each part of an explanation is there—and to know when the evidence is asking you to change your mind.

Small Group Tutorials

How to Test a PSLE Science Explanation by Asking What Would Happen If One Causal Link Were Removed