How to Infer What a Part Does in PSLE Science by Comparing the System With and Without It

Wait, What? Removing a Part and Seeing the System Fail Does Not Tell You Everything About That Part

A learner looks at a system with several parts. One part is removed. The system no longer produces the same result.

The learner concludes:

“That part does everything.”

The experiment has shown something important—but not that.

If removing a part changes the outcome while other relevant conditions remain comparable, the part is evidence-linked to that outcome. It may be necessary for the system to work normally. It may contribute one step in a larger chain. It may support another part. It may provide a path, force, material, signal or structure.

But removal evidence does not automatically reveal every detail of the mechanism, prove the part is the only cause, or show that adding more of the part will always increase the effect.

Comparing a system with and without a part can reveal what changes when the part is unavailable. The scientific job is to turn that difference into a careful inference about the part’s contribution—not a bigger claim than the evidence earns.

Quick Answer

When a PSLE Science question changes, removes, blocks or restores one part of a system, use this route:

IDENTIFY THE PART → IDENTIFY THE SYSTEM OUTCOME → COMPARE WITH PART / WITHOUT PART → CHECK WHAT ELSE CHANGED → STATE THE OBSERVED DIFFERENCE → INFER THE PART’S CONTRIBUTION → SELECT THE RELEVANT SCIENTIFIC CONCEPT → EXPLAIN THE MECHANISM → CHECK WHETHER THE EVIDENCE SHOWS NECESSITY, CONTRIBUTION OR SOMETHING STRONGER → TEST THE INFERENCE WITH RESTORATION OR ANOTHER CONTROL WHERE POSSIBLE.

A useful conclusion often sounds like:

“When Part P was absent or blocked, Outcome Y decreased/disappeared/changed while the other stated conditions were kept comparable. This supports the inference that P contributes to ______ under these conditions.”

That is a reasoning scaffold for practice, not an official PSLE marking phrase.

The Exact PSLE Science Learning Job This Guide Owns

This guide owns one learner job: how a Primary 5 or Primary 6 learner infers the function or contribution of a part by comparing a system when that part is present, absent, blocked, changed or restored, while keeping observation, conclusion and mechanism separate.

It does not replace the broader Systems-theme guide. It does not replace the page on function versus mechanism, the fair-test owner, the control-setup owner or the necessary-versus-sufficient guide. Those remain canonical for their own jobs.

This page owns a particular evidence move:

What can I learn about a part from what happens when I change its availability in an otherwise comparable system?

The Current 2026 PSLE Science Frame

For examination from 2026, Standard PSLE Science assesses the 2023 Primary Science syllabus. The official assessment objectives include applying scientific facts, concepts and principles; interpreting and analysing information; evaluating observations, information and methods; and communicating explanations and reasoning.

The Primary Science syllabus also treats Systems as a connected way of understanding parts, functions and interactions. That does not mean every “part question” is a memory test. A learner may need to infer a part’s role from experimental evidence.

Function and Mechanism Are Related but Different

Function asks what contribution a part makes to the system.

Mechanism asks how that contribution is produced.

Example:

Function: a switch can complete or break a conducting path in a simple circuit.
Mechanism: changing the switch position changes whether there is a complete conducting path through the circuit.

An intervention can reveal a function before the learner has described the full mechanism. Do not merge the two jobs too early.

The Core Comparison: Present Versus Absent

The cleanest version of this reasoning compares two otherwise similar systems:

Setup	Part P	Other relevant conditions	Outcome
A	Present	Comparable	Outcome Y occurs
B	Absent	Comparable	Outcome Y decreases or does not occur

If P is the important deliberate difference, the change in Y supports an inference that P contributes to Y.

But the exact strength of the inference depends on the design.

Removal, Blocking and Disconnection Are Different Interventions

A part can be made unavailable in several ways:

physically removed;
blocked so material cannot pass;
disconnected from the rest of the system;
covered so light or another input cannot reach it;
temporarily prevented from moving;
replaced with another material or part.

These interventions are not always equivalent. Removing a part may change structure as well as function. Blocking a pathway may create pressure or accumulation elsewhere. Replacing a part can introduce new properties.

Ask exactly what the intervention changes.

Worked Example 1 — Simple Circuit: Remove the Path, Lose the Response

Original practice situation: A working simple circuit contains a cell, wires and a bulb. One connecting wire is removed while the other components remain in place. The bulb no longer lights.

Observation: bulb lit before; bulb did not visibly light after the wire was removed.

Inference: the connecting wire contributes to providing a complete conducting path.

Mechanism: removing the wire breaks the conducting path, so the circuit no longer has the complete path required for the stated bulb response.

What should the learner not conclude?

the wire is the only part that matters;
the wire “creates all the electricity”;
any wire of any material would work identically;
the bulb itself is definitely faulty or definitely healthy without further information.

Worked Example 2 — Block a Pathway: The Outcome Changes

A system has a tube through which a liquid normally moves. In Setup P the tube is open. In Setup Q the same tube is blocked. The other stated conditions are kept comparable.

Liquid reaches the far side in P but not Q.

The evidence supports the contribution of the open pathway to liquid movement through that route.

It does not prove the tube itself pushes the liquid. Another part of the system may produce the pressure or driving force. The pathway’s function and the force-producing mechanism can belong to different parts.

Worked Example 3 — Support Structure: Removal Changes Shape

A model structure contains several supports. One support is removed and the structure bends more under the same load.

The fair inference is that the removed support contributes to resisting deformation or distributing load in that model.

Do not jump to “this support carries the entire load”. Other supports can still share the load.

Worked Example 4 — Covering a Part: Input Versus Part Function

A part of a system is covered so it no longer receives light. The system’s measured response decreases.

What changed?

the part is still physically present;
its access to light changed.

The result can support the importance of light reaching that part under the tested conditions. It does not prove that simply removing the part and simply removing light from the part are scientifically identical interventions.

Worked Example 5 — Restoration Makes the Inference Stronger

Suppose a system works normally with Part P. P is removed, and the outcome decreases. P is then restored in the same position and the outcome returns close to the original state.

This restoration comparison is stronger than removal alone because the effect changes in the expected direction twice:

P present → normal outcome; P absent → altered outcome; P restored → outcome returns.

It still does not reveal every microscopic detail of how P works, but it reduces some alternative explanations such as an unrelated permanent change during the experiment.

Worked Example 6 — Replacement Tests a Property, Not Merely Presence

A component made of Material A is replaced with a same-shaped component made of Material B.

If the outcome changes, the learner should not say only “the part is important”. The stronger scientific question is which property of the material is relevant.

Replacement can therefore shift the learner job from “Does this part matter?” to “Which property of this part matters?”

Worked Example 7 — Same Outcome After Removal

Part P is removed but the measured outcome remains unchanged.

Can the learner conclude P has no function?

No.

Possibilities include:

P is not needed for this particular measured outcome;
another part compensates;
the measurement is too coarse to detect the change;
P affects a different outcome;
the test duration is too short;
the system already contains stored material or energy that hides the effect temporarily.

“No observed change” is evidence, but its interpretation must remain bounded.

Contribution Is Not the Same as Only Cause

Suppose removing P stops Outcome Y.

This can show P is necessary under the tested conditions. It does not show P is sufficient by itself.

A bulb may need a cell, complete path and working components. Removing any one required part can stop the response. That does not make any one part the whole system.

Necessary means the outcome does not occur normally without it under the tested conditions. Sufficient means it can produce the outcome with the other required conditions in place. Those are different claims.

Removal Evidence Is Strongest When Only One Relevant Thing Changes

If Part P is removed and the temperature, material, light level and starting amount also change, the learner cannot know which difference caused the outcome.

The investigation should preserve the comparison architecture:

same system type;
same starting state where relevant;
same measured outcome;
same measurement method;
same duration;
one deliberate availability change where practical.

Removal Can Create Side Effects

A sophisticated learner asks whether removing a part changes more than its intended function.

Examples:

removing a cover changes airflow as well as protection;
removing a structural part changes spacing and alignment;
blocking a tube changes pressure elsewhere;
disconnecting a component can alter the entire path;
removing a leaf changes both surface area and total leaf number.

This does not make removal experiments useless. It means the learner should identify competing changes before claiming the mechanism.

Function Inference Versus Function Memory

Some questions ask learners to recall a taught function. Others give an unfamiliar system and expect the learner to infer a contribution from evidence.

Use the evidence when the question supplies it.

Do not force a memorised textbook function onto a new object simply because the shape looks familiar.

Function Inference From Increase, Decrease or Loss

Intervention	Observed result	Careful inference
Part removed	Outcome disappears	Part may be necessary for that outcome under tested conditions
Part removed	Outcome decreases	Part contributes to the outcome; other contributors may remain
Part blocked	Material no longer reaches destination	Open part/path contributes to transfer along that route
Part restored	Outcome returns	Strengthens link between availability of part and outcome
Part removed	No measured change	No detected effect on that measured outcome under these conditions; function not disproved universally

How Do We Know It Is the Part and Not the Procedure?

Ask whether the act of changing the part introduces a new disturbance.

Was the system handled more in one setup?
Was extra time allowed to pass?
Was another part accidentally moved?
Was the measurement taken at a different position?
Was the restored part returned to the same state?

Method changes can masquerade as part-function evidence.

The Present–Absent–Restored Protocol

Present: observe the normal system.
Absent/blocked: change the availability of the target part while holding other relevant conditions comparable.
Measure: record the same outcome using the same method.
Compare: state exactly what changed.
Infer: identify what contribution the part likely makes.
Restore where possible: return the part or condition and see whether the outcome returns.
Explain: use the relevant scientific mechanism.
Bound: state what the evidence cannot prove.

The Earliest-Weak-Link Diagnostic

Failure signature	Earliest weak link	Repair
“The system stopped, so P does everything.”	Contribution exaggerated into sole cause.	List other required parts/conditions.
“P is necessary, therefore P alone is sufficient.”	Necessary and sufficient confused.	Ask whether P by itself can produce the outcome.
“Removing P proves the exact mechanism.”	Function inference jumped beyond evidence.	Separate observed contribution from mechanism model.
“No change means P has no function.”	Measurement and condition limits ignored.	Consider compensation, another outcome or insufficient resolution.
“P was removed, but three other conditions changed too.”	Confounded intervention.	Repair fair comparison before causal inference.
“Restoring P had no effect, but I ignored that.”	Contrary evidence discarded.	Re-evaluate the original inference.
“I memorised a function from another diagram.”	Shape/topic cue replaced evidence.	Read the actual intervention and outcome.

Misconception Repair — “If Removing It Stops the System, It Must Be the Most Important Part”

Many systems contain several necessary parts. Removing any one can stop the whole chain. “Necessary” does not create a ranking of importance.

Misconception Repair — “Function Is the Same as Mechanism”

Function describes the contribution. Mechanism describes how the contribution occurs. A removal comparison often supports the first more directly than the second.

Misconception Repair — “Restoration Always Proves the Claim”

Restoration strengthens evidence only if the restored system is genuinely comparable. If the intervention caused permanent damage or other changes, the system may not return fully.

Misconception Repair — “A Part Has Only One Function”

A part can contribute to more than one outcome. An experiment testing one outcome does not map every function.

How This Appears in Multiple-Choice Questions

Identify the target part.
Find the setup where its availability changes.
Check what other conditions remain comparable.
Read the measured outcome.
State the observed difference before explaining.
Reject options that claim the part is the only cause without evidence.
Reject options that confuse function with mechanism.
Choose the claim whose strength matches the intervention evidence.

How This Appears in Structured Answers

A useful thinking shape is:

When ______ was removed/blocked, ______ changed from ______ to ______ while ______ was kept comparable. This supports the conclusion that ______ contributes to ______ because ______.

Use only the clauses the actual question needs. This is not a compulsory marking sentence.

How This Connects to the PSLE Science Reasoning Law

Observe: what changed when the part changed?
Object: which part and which system outcome?
Observation versus inference: “bulb off” is observed; “wire provides part of the conducting path” is inferred from the model and evidence.
Concept: select the relevant system mechanism.
Mechanism: explain how availability of the part affects the system.
Condition: state the intervention and fair-comparison conditions.
Outcome: connect mechanism to the measured result.
Check: ask whether another change could explain the result.

Practice Sequence

Start with simple present-versus-absent comparisons.
Move to blocked-versus-open pathways.
Add restoration as a third condition.
Add one confound and ask why the inference weakens.
Use necessary-versus-sufficient contrasts.
Give a “no observed change” case and generate several explanations.
Use replacement rather than removal to test a property.
Move across circuits, structures, transport systems and unfamiliar models.
Return after several days with no function labels given.

Unfamiliar Transfer Challenge

A mystery system contains Parts A, B, C and D. Under the same starting conditions:

Test	System state	Measured output
1	All parts present	10 units
2	Part C blocked	2 units
3	Part C restored	9 units

What can you infer?

C strongly contributes to the measured output under the tested conditions.
Blocking C is associated with a large drop in output.
Restoring C and recovering close to the original output strengthens the inference.

What can you not yet infer?

C is the only part required.
C alone can create 10 units.
the exact microscopic mechanism;
twice as much C would produce twice the output.

The topic is hidden. Intervention reasoning still works.

Delayed Independent Return

Three to five days later, take a new system question and ask:

Which part was changed?
How exactly was it changed—removed, blocked, disconnected or replaced?
What outcome was measured?
What else stayed comparable?
What changed in the result?
What contribution can I infer?
Is the part necessary, sufficient, both or neither based on this evidence?
What mechanism is supported by Primary Science?
What alternative explanation remains?
Would restoration strengthen the inference?

The Answer-Checking Receipt

Did I identify the correct part and outcome?
Did I state the observed difference before explaining?
Did I check what else changed?
Did I avoid turning contribution into sole cause?
Did I distinguish necessary from sufficient?
Did I distinguish function from mechanism?
Did I consider side effects of removal or blocking?
Did I use restoration evidence if available?
Did I treat no-change results cautiously?
Did I keep the conclusion within the tested system and conditions?

Evidence and Model Limits

Intervention is a powerful scientific strategy because changing one part can reveal causal contribution. Real scientific research uses much more sophisticated perturbation, control and restoration designs.

At Primary level, the durable principle is enough: compare the system before and after a well-defined change, keep competing conditions under control, observe what changes, and match the strength of the conclusion to the evidence.

A removal result supports a contribution claim more directly than a complete mechanistic story. If several explanations still fit, more evidence may be needed.

Useful Internal Routes

Parent and Tutor Teaching Guide

Build the learner’s reasoning with three questions:

“What changed when the part changed?”
“What does that show about the part?”
“What does it still not prove?”

The third question is crucial. It prevents evidence from expanding into an unsupported story.

Use simple physical systems first. Remove a connecting piece, block a route or take away one support. Ask the child to separate observation from inference.

Then add restoration. If replacing the part makes the outcome return, ask why that strengthens the link without proving that the part acts alone.

Next use a no-change case. The child should learn that “no detected effect” can mean the part does not affect that outcome under those conditions, but can also reflect compensation, measurement limits or a different function.

Finally, remove the familiar Science context. Use a mystery system labelled A, B, C and D. Mastery is shown when the learner can infer contribution from intervention evidence without guessing from object shape or memorised function labels.

Authoritative and Research References

The research references support broader scientific reasoning about variables, intervention and evidence. They do not create an official PSLE function-inference formula.

The Quiet Ending

To understand a system, sometimes change one part and watch what the system loses.

Then be careful.

The difference tells you the part matters. The Science tells you how. The evidence boundary tells you how much you are allowed to claim.

Small Group Tutorials