How to Tell Observational Evidence From Experimental Evidence in a PSLE Science Question Without Inventing a Changed Variable

Wait, What? Two Groups Can Be Different Even When Nobody Changed a Variable

A learner sees that plants in one part of a garden are taller than plants in another part and immediately writes, “The amount of light was the changed variable.” But the question never says that anyone moved the plants, controlled the light or designed two test conditions. It only says that the plants were observed where they were already growing.

That is a different kind of evidence. The data may show a useful relationship. They may support a scientific explanation when combined with relevant knowledge and other information. But the learner should not invent an experiment that never happened.

OBSERVE WHAT THE STUDY DID → ASK WHETHER A CONDITION WAS DELIBERATELY CHANGED → IDENTIFY WHAT WAS MEASURED → DESCRIBE THE PATTERN → DECIDE WHAT THE DESIGN CAN SUPPORT → ADD SCIENCE WITHOUT TURNING ASSOCIATION INTO EXPERIMENTAL PROOF.

Quick Answer

Observational evidence comes from watching, counting or measuring a system without deliberately assigning the scientific condition being compared. Experimental evidence comes from a design in which a condition is deliberately changed so the outcome can be compared under controlled conditions. In a PSLE Science question, do not decide from the presence of a table, graph or apparatus. Decide from what the method actually did.

If the question only describes observation, do not invent a changed variable. If the method deliberately changes one condition and manages relevant alternatives, you can reason about that intervention. In both cases, stay with the exact evidence and the relevant scientific mechanism.

The Exact PSLE Science Learning Job This Guide Owns

This guide owns one PSLE-specific reading job: deciding from a question stem whether the evidence is observational or experimental, and adjusting the learner’s interpretation and causal claims to match that design.

The broader scientific distinction between observational and experimental evidence already has a separate Science owner in the eduKate Sengkang ecosystem. This page does not replace it. This page teaches the examination-facing learner job: reading an unfamiliar scenario accurately before assigning variable roles or writing a conclusion.

The Current 2026 PSLE Science Frame

For examination from 2026, Standard PSLE Science assesses attainment in the 2023 Primary Science syllabus. SEAB’s assessment objectives include applying scientific facts, concepts and principles and applying scientific inquiry through predictions and hypotheses, interpreting and analysing information, evaluating observations, information and methods, and communicating explanations and reasoning.

Those objectives make design-reading important. A learner who calls every comparison a fair test can misread the evidence before the science explanation even begins.

The One Question That Separates the Two Designs

Did the investigator deliberately set or change the condition being compared?

If yes, the scenario may be experimental. You still need to inspect the comparison, controls and measurement. If no, and the investigator simply records naturally occurring groups, times, places or states, the evidence is observational.

Signal in the question	Likely evidence type	What to check next
“Pupils placed identical containers at three distances from a lamp”	Experimental	What was deliberately changed? What was measured? What stayed comparable?
“Pupils counted organisms at three naturally different locations”	Observational	What differs between locations? Which differences were measured rather than controlled?
“A class recorded temperature every hour for one day”	Observational time series	What changed naturally over time? What was actually measured?
“A pupil added different numbers of cells to otherwise similar circuits and measured bulb brightness”	Experimental	Was the number of cells the deliberate change? Were other relevant conditions comparable?

Worked Example 1 — Garden Plants

Original practice scenario: A learner measures the average height of the same plant species growing in four areas of a school garden. The four areas receive different amounts of sunlight during the day. No plants are moved.

Observation: The plants in areas receiving more sunlight are taller in the recorded data.

What did the learner deliberately change? Nothing about the plant locations. The learner observed existing conditions.

What can the evidence directly show? A relationship between recorded sunlight exposure and plant height in those observed areas.

What must not be claimed from the design alone? That sunlight was the only reason for the height difference. Soil, water, crowding, age and other conditions may also differ unless the question provides evidence about them.

Worked Example 2 — Deliberately Changing Light

Now change the scenario. A learner selects similar young plants, gives them the same amount of water, uses the same type of soil and places them under lamps at three stated light levels for the same duration. Plant growth is then measured.

Here the light condition has been deliberately changed. The investigation is designed to compare the outcome under different light conditions while keeping specified alternatives comparable. That makes experimental causal reasoning more defensible than in the garden survey—provided the rest of the method is suitable.

Notice what changed between the examples: not the topic, not the plant concept and not the presence of numbers. The design changed.

Worked Example 3 — Field Counts Are Not Failed Experiments

A class counts snails in shaded and open parts of a garden. A weak learner says, “This is unfair because they did not keep the amount of shade the same.” But shade is one of the naturally different conditions being compared. Keeping it the same would remove the comparison.

The better question is: What does this observational comparison let us say, and what other differences between the locations might also matter? Observational evidence is not automatically poor evidence. It answers a different kind of scientific question.

A Changed Variable Exists Only When Something Was Deliberately Changed

In school Science, learners often become so familiar with the phrase “changed variable” that they search for one in every table. That is dangerous. A variable is a quantity or condition that can take different values, but calling it the changed variable of an investigation implies a deliberate experimental role.

Temperature can vary naturally across a day. Soil moisture can differ naturally between sites. Population size can differ between habitats. Those are varying conditions, but they are not automatically manipulated experimental variables.

Observation, Association and Cause Are Different Claims

Claim type	Example	Evidence needed
Observation	Area A had more recorded organisms than Area B.	The counts themselves.
Association or pattern	Higher moisture readings occurred with higher organism counts in the observed sites.	Several relevant paired observations.
Causal explanation	A changed condition produced an effect through a scientific mechanism.	A suitable design plus relevant scientific knowledge; alternative explanations must be considered.

PSLE Science may still ask you to use scientific knowledge to explain an observational pattern. If so, make the distinction visible in your own thinking: the pattern comes from the evidence; the mechanism comes from relevant science. Do not pretend the observation alone proved the mechanism.

The Question-Reading Protocol

Identify the scientific object or system.
Find what was observed or measured.
Find the groups, places, times or conditions being compared.
Ask whether the investigator deliberately assigned or changed that condition.
If yes: identify the manipulated condition, measured outcome and stated controls.
If no: treat the differing conditions as observed features, not automatically as controlled experimental variables.
Describe the data pattern before explaining it.
Select the relevant scientific concept.
Build a mechanism only as far as the evidence and conditions allow.
Check the final claim against the design.

The PSLE Science Reasoning Law

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.

The extra design check is simple: Was the condition observed, or was it deliberately changed?

Observable Failure Signatures

Learner behaviour	Earliest weak link
Labels every difference as a changed variable	Deliberate intervention has not been checked
Calls a field survey an unfair experiment	Observational and experimental jobs are being merged
Writes “X caused Y” after seeing only naturally occurring groups	Association has been upgraded to causation
Refuses to explain observational data at all	Evidence type is being confused with uselessness
Uses apparatus presence as proof of an experiment	Design is being inferred from objects rather than procedure
Invents controlled variables that the stem never states	Familiar fair-test language is overriding the question

Find the First Weak Link

Did I identify what was actually done?
Did I find whether a condition was deliberately changed?
Did I identify what was measured?
Did I describe the pattern before explaining it?
Did I invent a control or manipulated variable?
Did I choose a concept that fits the observed conditions?
Did I state cause more strongly than the design allows?

Misconception Repair — “Observational Means Weak”

No. Observational evidence can be scientifically important because some systems are best studied where they naturally occur. The limitation is not “observation is bad.” The limitation is that naturally differing groups may differ in more than one relevant way, so causal conclusions need care.

Misconception Repair — “Experimental Means Automatically Conclusive”

No. An experiment can still have a poor comparison, inappropriate measurement, uncontrolled relevant conditions or too little evidence. Deliberately changing a condition gives the design a different causal job; it does not guarantee that the job was performed well.

Misconception Repair — “If Two Things Differ, One Must Be the Variable”

Two observed groups can differ in many ways at once. The word “variable” does not solve the causal problem. You need to know which condition was deliberately assigned, if any, and which differences were merely observed.

A Design-Switch Exercise

Take one context—temperature and evaporation, for example—and write it in two forms:

Observational version: measure evaporation from containers already located in different parts of a room.
Experimental version: deliberately place otherwise comparable containers at selected temperatures or controlled positions and measure evaporation.

The scientific topic stays similar. The evidence structure changes. Practising this switch is more powerful than memorising a definition because it teaches you to see the design inside unfamiliar wording.

Practice Sequence

Classification: Read ten short study descriptions and label each observational or experimental.
Evidence: Underline what was measured or observed.
Intervention: Circle only conditions deliberately changed by the investigator.
Claim control: Write one statement the evidence supports and one it does not yet support.
Mechanism: Add relevant scientific knowledge without pretending the design proved more than it did.
Transfer: Repeat across all five Primary Science themes.

Unfamiliar Transfer Challenge

A learner records the temperature of five playground surfaces at noon and notes whether each surface is shaded. The learner does not move the surfaces or create the shade. Another learner later places identical tiles made of the five materials under the same lamp for fifteen minutes and measures their temperatures.

Explain which data set is observational and which is experimental. Then state one conclusion each design can support and one conclusion that would go too far. If you can do this without relying on topic keywords, the design distinction is becoming usable.

Delayed Independent Return

Three to five days later, take four unfamiliar Science scenarios. Before solving them, write only two lines for each: “deliberate change: ___ / none stated” and “measured outcome: ___.” Then decide the evidence type. This delayed return tests whether you can recognise design structure after the wording changes.

The Design-Reading Receipt

I identified what the investigator actually did.
I know whether the compared condition was deliberately changed.
I did not invent a manipulated variable.
I know what was observed or measured.
I described the pattern before explaining it.
I kept association separate from causal proof.
I used scientific knowledge as explanation, not as invented evidence.
I checked whether relevant alternative conditions were controlled or merely unknown.
I did not dismiss observational evidence as useless.
My final claim matches the design.

Parent and Tutor Teaching Guide

Ask the child one question before naming any variables: “What did the investigator deliberately do to the system?” If the answer is “nothing; they only measured what was already there,” do not force fair-test vocabulary onto the scenario.

Next ask, “What does the evidence show even without an experiment?” This prevents the opposite error of treating observational evidence as worthless. Then ask what additional evidence would make a causal claim stronger.

Finally, give the same scientific topic in both observational and experimental forms. The learner should be able to explain why the scientific concept can remain the same while the evidence claim changes.

Useful Internal Routes

Authoritative and Research References

This learning guide does not prescribe one compulsory PSLE phrase. It teaches a reasoning boundary: the strength of the claim should match how the evidence was obtained.

Next in This Learning Sequence

Continue with the PSLE Science Learning Guide, or move to the feedback-timing guide in this publication batch after it appears.

The Quiet Return

A comparison is not automatically an experiment. A table is not automatically a fair test. A naturally different group is not automatically a changed variable.

Read what was done. Find what was measured. Notice what was deliberately changed—and what was merely observed.

Then make the strongest scientific claim the design has actually earned.

Small Group Tutorials