How to Spot When Changing One PSLE Science Variable Automatically Changes Another

Wait, What? Sometimes you cannot change just one variable because changing it automatically changes something else.

Primary learners often hear a useful fair-test rule: change one factor and keep other relevant conditions the same. But real set-ups can contain linked variables. Moving a lamp farther from an object changes distance and can also change the light reaching it. Adding more water to the same container changes water amount and may change water depth. Using a larger object can change mass and surface area together. If two conditions are structurally linked, a learner must recognise that link instead of pretending the second condition stayed controlled.

Quick Answer

Before calling an investigation a fair test, ask: If I change X, does Y change automatically because of the way the set-up is built? If yes, X and Y are linked in this method. The investigation may still be useful, but you cannot automatically attribute the outcome to X alone. You may need a redesigned set-up, a different way of changing X, an additional measurement, or a narrower conclusion that acknowledges the linked change.

The Exact PSLE Science Learning Job This Guide Owns

This guide owns one inquiry job: detect linked co-change in an investigation and decide what that does to the causal claim. It does not replace the broader pages on variables, fair tests or several conditions changing at once. Its narrower job is the case where a second condition changes because the method used to change the first condition also changes it.

The 2026 PSLE Science assessment frame includes scientific inquiry, interpretation and analysis, and evaluation of methods. Recognising that a method couples two conditions is exactly the kind of method reasoning that prevents a neat-looking comparison from becoming a false causal conclusion.

Independent Changes and Linked Changes

Situation	What happens	Reasoning consequence
Independent change	X changes while Y can genuinely remain comparable	The effect of X may be isolated more cleanly.
Linked change	Changing X necessarily or mechanically changes Y too	The outcome may depend on X, Y or both.
Accidental second change	Y changes because the procedure was not controlled properly	The method may be repairable by better control.

Linked co-change is different from a simple mistake. Sometimes the method itself makes the two conditions inseparable unless it is redesigned.

Worked Example 1: Lamp Distance and Light Reaching the Object

A learner moves the same lamp farther from a set-up and says the investigation is testing only “distance”. But moving the lamp changes the spatial arrangement and can also change the amount or intensity of light reaching the target. If the measured outcome responds to light received, the scientific mechanism may be connected to the changed illumination rather than to distance as an abstract number.

The learner should identify what the investigation is really asking. If distance is being used as the practical way to alter the light condition, then the causal story should reflect that mechanism. If the goal is to isolate another effect of distance while keeping illumination constant, the method needs a different design.

Worked Example 2: Water Amount and Water Depth

Two identical containers receive different volumes of water. A student wants to test whether “amount of water” changes how much of an object is submerged. In the same container shape, changing water volume also changes water depth. Those are not independent conditions in that set-up.

A weak answer treats volume and depth as two unrelated causes. A stronger answer recognises the linkage: more water in the same container creates a greater depth, which changes the object’s surrounding condition. The investigation’s conclusion should match the actual scientific relationship the method creates.

Worked Example 3: Larger Object, More Than One Property Changes

A learner compares a small cube with a larger cube made of the same material. The larger cube may have greater mass and greater surface area. If the outcome could depend on both, “size” is not a clean single variable unless the question defines exactly what aspect of size matters and the method allows the alternatives to be separated.

This does not mean the comparison is useless. It means the conclusion must stay inside the evidence. Science can compare naturally linked packages of properties, but it should not pretend the experiment isolated one property when it did not.

Worked Example 4: Treatment Time and Total Exposure

If a tested condition is applied for longer, total exposure to that condition also increases. A learner should not describe “duration” and “total exposure” as though they were two independently controlled variables when one is created by the other in the method. The relevant scientific explanation depends on what quantity the process responds to.

The Link Test

Use three questions:

What did the experimenter deliberately change?
What else changes automatically when that action is performed?
Could that second change plausibly affect the measured outcome?

If the answer to Question 3 is yes, the second condition matters to interpretation. You cannot simply label it “controlled” because you wish it had stayed the same.

Linked Variable or Scientific Mechanism?

Be careful: not every downstream response is a second manipulated variable. Suppose temperature is deliberately increased and evaporation becomes faster. Faster evaporation is an outcome of the changed temperature, not necessarily a confounding variable. The key distinction is whether the second change is another condition of the test that could independently influence the measured outcome, or whether it is part of the causal response you are trying to explain.

Ask: Was Y already part of the conditions under which the outcome occurs, or is Y itself a response caused by X?

Failure Signatures

The learner says “only one variable changed” because only one action was performed.
A method changes object size while also changing mass or surface area, but only one is acknowledged.
A container level changes because volume changed, yet depth is called “controlled”.
A learner blames a second variable without asking whether it is a cause, a linked condition or an outcome.
The conclusion names one cause even though the method changed a coupled pair.
The learner tries to “keep Y the same” using a method in which Y cannot physically remain the same.

Earliest Weak-Link Diagnosis

If a learner misses linked co-change, ask them to describe the physical action used to set each test condition. Do not begin with vocabulary such as independent or controlled variable. Ask what actually happens to the set-up when the action is carried out.

What did you move, add, remove, enlarge, shorten or change?
What else physically became different because of that action?
Which of those differences can affect the outcome?
Can the method separate their effects?

Misconception Repair

“One action means one variable.” Not always. One action can alter several scientifically relevant conditions.

“If two variables are linked, the experiment is invalid.” Not necessarily. The experiment may answer a broader question about the combined change. The problem comes when the conclusion claims to isolate one variable that the method did not isolate.

“Every second change is a confounder.” No. Some second changes are outcomes or mechanism steps. Decide the scientific role before labelling them.

“A controlled variable is whatever I write under ‘controlled’.” Control has to exist in the method, not just in the answer.

The Linked-Variable Reasoning Protocol

NAME THE INTENDED CHANGE → DESCRIBE THE PHYSICAL ACTION → IDENTIFY AUTOMATIC CO-CHANGES → ASK WHICH CAN AFFECT THE OUTCOME → SEPARATE CONDITIONS FROM RESPONSES → DECIDE WHETHER THE EFFECTS CAN BE ISOLATED → REDESIGN OR LIMIT THE CONCLUSION → CHECK AGAINST THE EVIDENCE.

How to Repair the Method

Possible repairs depend on the scientific question. A learner might:

change X in a different way that leaves Y comparable;
measure Y so its co-change is visible rather than hidden;
add comparison conditions that separate plausible explanations;
choose a different specimen or apparatus geometry;
narrow the conclusion to the combined condition actually tested.

There is no universal repair. The correct improvement is the one that preserves the scientific question while making the evidence more discriminating.

Original Practice Set

Practice A: Increasing the amount of liquid in the same container also raises liquid depth. Name the deliberate change, the automatic co-change and a possible outcome that could respond to either.

Practice B: A student uses larger pieces of the same material. List properties that may change with piece size and decide which could matter to the measured outcome.

Practice C: A lamp is moved farther away. Identify what physical condition at the target may change together with distance.

Practice D: Temperature is increased and reaction speed increases. Decide whether reaction speed is a linked test condition or the measured response. Explain your choice.

Retrieval and Transfer Sequence

Take five investigation descriptions and mark every deliberate action.
For each action, list all automatic physical consequences before deciding variable roles.
Classify each consequence as linked condition, measured outcome, mechanism step or irrelevant detail.
Redesign one coupled investigation so the intended causal question becomes clearer.
Return several days later with a new context and perform the same analysis without prompts.

Unfamiliar Transfer Test

A learner changes the height from which an object begins moving down a track. The starting height changes, and the length of the path before the measurement point may also change depending on the track design. The learner should inspect the geometry instead of assuming “height” is the only changed condition. If both height and path length change, the causal claim must account for that coupling or the set-up must be redesigned.

Delayed Independent Return

After a delay, show an unfamiliar apparatus with one obvious manipulated condition. Ask the learner to find any automatic co-change before seeing the results. This is important: linked-variable reasoning should happen before the learner knows which result they want to explain.

Answer-Checking Receipt

I identified the deliberate action used to change the condition.
I checked what else changes automatically.
I separated linked conditions from outcomes.
I asked whether the second condition could affect the same measured result.
I did not claim a variable was controlled when the method physically changed it.
I know whether the method isolates one cause or tests a combined change.
My conclusion is limited to what the method can actually support.

Parent and Tutor Teaching Guide

Use real objects. Ask the child to change one feature and narrate every other thing that changes because of that action. For example, pour more water into the same glass: amount changes, water level changes, mass changes. Then ask which of those differences matters to a particular scientific question.

The teaching goal is not to make children suspicious of every experiment. It is to make them see that variable labels are descriptions of a physical method. Once that becomes visible, fair-test reasoning becomes causal reasoning rather than a memorised worksheet routine.

Useful Internal Routes

Authoritative References

Previous and Next

Previous: Direction of Change vs Size of Change
Next: Same Concept, Different Reasoning Job

Quiet Return

A fair test is not created by filling in three variable labels. It is created by understanding what the method physically changes. When one action changes two relevant conditions together, good science notices the coupling, protects the conclusion and redesigns the comparison when necessary.

Small Group Tutorials