How to Tell Whether a PSLE Science Question Can Actually Be Answered by an Investigation

Wait, What? A Question Can Sound Scientific and Still Be Impossible to Answer With the Investigation You Planned

Students are often taught to turn Science questions into experiments. But not every interesting question can be answered by the apparatus, observations or measurements available. Sometimes the question is too vague. Sometimes the outcome cannot be observed. Sometimes several conditions change together. Sometimes the proposed evidence would not distinguish the possible answers.

The scientific skill is not “always do an experiment”. It is deciding whether the proposed investigation could actually produce evidence that answers the question.

BEFORE PLANNING THE METHOD, CHECK WHETHER THE QUESTION IS INVESTIGABLE WITH OBSERVABLE EVIDENCE.

Quick Answer

READ THE QUESTION → IDENTIFY WHAT MUST BE DECIDED → ASK WHAT OBSERVATION OR MEASUREMENT WOULD COUNT AS EVIDENCE → IDENTIFY WHAT MUST BE CHANGED, COMPARED OR TRACKED → CHECK WHETHER OTHER RELEVANT CONDITIONS CAN BE KEPT COMPARABLE → CHECK WHETHER THE EVIDENCE WOULD DISTINGUISH THE POSSIBLE ANSWERS → IF NOT, NARROW OR REWRITE THE QUESTION BEFORE BUILDING THE METHOD.

The Exact PSLE Science Learning Job This Guide Owns

This guide owns one learner job: deciding whether a proposed Primary 5/6 Science question can actually be answered by collecting observations or measurements in a bounded investigation.

It does not own general investigation planning, variable definitions or turning an observation into a testable question. Those have separate routes. Here, the learner is acting as a gatekeeper: does the question-and-evidence pair work before a full method is designed?

Why This Matters in Current PSLE Science

For examination from 2026, PSLE Science assesses the 2023 Primary Science syllabus. The official assessment frame includes scientific inquiry, prediction and hypothesis, interpretation and analysis, evaluation of observations, information and methods, and communication of explanations and reasoning. Those skills depend on asking questions that available evidence can meaningfully address.

This guide does not claim that “Is this investigable?” is a fixed national examination question type. It teaches a scientific inquiry habit that strengthens planning, evaluation and evidence reasoning.

The First Test: What Would Count as an Answer?

Consider the question: “Which material is best?”

Best for what? Keeping water warm? Allowing light through? Supporting a load? Absorbing water? Without a defined outcome, no measurement can settle the question.

A more investigable form might be: “Which of these materials keeps 100 mL of warm water at the highest temperature after 15 minutes under the same stated conditions?” Now the evidence job is visible.

The Six Tests of an Investigable Question

Test	Ask	If it fails
Object	What scientific object, system or specimens are involved?	The question may be too vague.
Outcome	What can be observed, counted or measured?	There may be no evidence route.
Comparison	What conditions, cases, times or groups are being compared?	A single result may not answer the question.
Control	Can other relevant influences be kept comparable or accounted for?	The evidence may not isolate the relationship.
Discrimination	Would different possible answers produce different observable evidence?	The method cannot decide between them.
Scope	Is the question small enough for the available method and evidence?	Narrow it before investigation.

Worked Example 1 — “Do Plants Like Light?”

This sounds like a Science question but “like” is not an observable scientific outcome. You need to decide what response will be measured or observed and under what conditions.

An investigable version could compare one defined plant response under different stated light conditions while keeping other relevant conditions comparable. The exact response and method must match the scientific concept being studied.

The repair is not to replace “like” with a complicated word. The repair is to define evidence.

Worked Example 2 — A Question With No Comparison

Question: “What is the temperature of the water after ten minutes?”

This can be measured, so it is answerable by observation. But it does not by itself test whether one condition caused a temperature difference. If the intended question is about the effect of insulation, at least one scientifically meaningful comparison is required.

Therefore distinguish two jobs: measure a state and test a relationship. Both can be investigable, but they need different evidence structures.

Worked Example 3 — Two Conditions Change Together

A learner asks whether warmer conditions make a process faster. Set-Up A is warm and uncovered. Set-Up B is cool and covered. The result differs.

The question itself may be investigable, but this proposed investigation does not isolate temperature because covering also differs. The repair is usually to keep the question and redesign the comparison, not to invent a conclusion from weak evidence.

Worked Example 4 — The Outcome Cannot Be Observed by the Method

A question asks whether a hidden internal process happened, but the proposed method records only an unrelated external feature. If that observation does not reliably indicate the process, the investigation cannot answer the question as designed.

You may need a different observable indicator, a direct measurement, or a narrower claim about what the existing evidence can show.

Worked Example 5 — Evidence That Cannot Distinguish Two Explanations

Two scientific explanations predict exactly the same result under the planned conditions. Even a perfectly conducted investigation will not decide between them because both explanations survive the evidence.

To make the question investigable in a useful way, choose a condition or observation where the explanations predict different outcomes. This is called a discriminating test in broader scientific practice.

Worked Example 6 — The Question Is Too Broad for the Sample

A learner tests two specimens and asks, “What is true for all organisms of this type?” The investigation may provide evidence about the tested specimens, but the broad universal question outruns the evidence.

One repair is to narrow the conclusion to the tested cases. Another is to design broader sampling if the scientific goal truly requires broader generalisation.

Investigable Does Not Mean Easy

A question can be scientifically investigable even if the method is difficult, time-consuming or impractical in a school setting. Distinguish:

scientifically investigable: suitable evidence could in principle answer the question;
practically feasible: the required method can be carried out safely and realistically with available resources;
answerable by the current data: the evidence has already been collected and is sufficient for the claim.

These are related but not identical.

Some Scientific Questions Need Observation Rather Than Deliberate Manipulation

Do not assume that an investigation always needs the learner to change a variable deliberately. Some questions can be addressed by observing naturally occurring differences or changes. The key is whether the evidence can answer the question and whether the conclusion respects the limits of the method.

Some Questions Are Better Answered From Established Scientific Knowledge

If a question asks for the accepted meaning of a term or a well-established fact, inventing a classroom experiment may be unnecessary. Scientific inquiry includes using reliable information as well as conducting investigations. The learner should match the evidence route to the job.

The PSLE Science Reasoning Chain

READ THE QUESTION → IDENTIFY THE SCIENTIFIC OBJECT OR RELATIONSHIP → DEFINE THE OBSERVABLE EVIDENCE → DISTINGUISH OBSERVATION FROM INFERENCE → IDENTIFY THE CONDITION OR COMPARISON → SELECT THE RELEVANT CONCEPT → CHECK WHETHER THE METHOD CAN TEST THE CLAIM → STATE ONLY THE OUTCOME THE EVIDENCE COULD SUPPORT.

Observable Failure Signatures

Failure signature	Earliest weak link	Repair
Starts listing apparatus before defining the outcome	Method before question	State what evidence would answer the question first.
Uses vague words such as “best”, “healthy” or “good”	Outcome not operationally defined	Replace vague judgement with observable criterion.
Only one result is collected for a causal question	Missing comparison	Define the reference or paired condition needed.
Two important conditions change together	Relationship not isolated	Redesign or narrow the conclusion.
Method measures something unrelated to the claim	Outcome mismatch	Choose evidence that directly bears on the question.
Two explanations predict the same observation	Evidence is non-discriminating	Find a condition where predictions differ.
Conclusion claims all cases from a tiny sample	Question scope exceeds evidence	Narrow the claim or strengthen sampling.

Earliest-Weak-Link Diagnosis

What exactly must be decided?
What observation or measurement could answer it?
What object or system will produce that evidence?
Do I need a changed condition, a comparison, repeated observation or simple measurement?
Could another relevant condition explain the same result?
Would different scientific possibilities give different evidence?
Is the question broader than the specimens or range I can test?
What is the smallest rewrite that makes the question answerable?

Misconception Repair — “Every Science Question Needs an Experiment”

No. Science uses observations, measurements, comparisons, existing evidence and well-established knowledge. An experiment is one evidence route. Choose it when deliberate manipulation helps answer the relationship being asked about.

Misconception Repair — “If I Can Measure Something, I Can Answer the Question”

The measurable quantity must be relevant to the claim. Recording an easy number that does not represent the outcome is not useful evidence merely because it is precise.

Misconception Repair — “A Fair Method Makes Any Question Answerable”

A beautifully controlled method cannot rescue a vague or non-discriminating question. Question quality and method quality are separate checks that must align.

Misconception Repair — “If the Result Is Different, I Found the Cause”

A difference is evidence only within the comparison actually made. If several relevant conditions differ, the result may show a difference without identifying its cause.

Practice Protocol: Question → Evidence → Feasibility

Write one proposed Science question.
Underline the outcome word.
Replace any vague outcome with an observable or measurable criterion.
State the minimum comparison or observation needed.
List important conditions that could also affect the outcome.
Ask whether the planned evidence separates the possible answers.
Check whether the question is too broad for the tested specimens or range.
Rewrite only what is necessary.
Then, and only then, plan the apparatus and method.

Unfamiliar Transfer Challenge

Write four original Science questions:

one that can be answered by one direct measurement;
one that needs a comparison between set-ups;
one that is too vague until an outcome is defined;
one where two possible explanations need a discriminating follow-up test.

For each, state exactly what evidence would count as an answer. If you cannot name the evidence, the question is not ready.

Delayed Independent Return

Three to five days later, take a fresh proposed investigation question. Without notes, decide whether it is answerable as written. If it is not, identify the earliest failure—vague outcome, missing comparison, uncontrolled alternative, non-discriminating evidence or excessive scope—and repair only that layer.

Investigability Receipt

I can state exactly what the question wants decided.
I can identify an observable or measurable outcome.
I know whether a comparison or deliberate change is needed.
I check other relevant conditions.
I ask whether the evidence would distinguish the possible answers.
I keep the question within the tested specimens, time and range.
I do not begin with apparatus before the evidence job is clear.
I can narrow an unanswerable question without changing its scientific purpose unnecessarily.

Parent and Tutor Teaching Guide

Give the learner a scientific-sounding but vague question such as “Which material is better?” Do not ask for apparatus. Ask, “What would we have to observe or measure before we could decide?”

Next, give a question with a clear outcome but a poor proposed comparison. Ask whether the problem is in the question or in the method. This separation is important: sometimes the question deserves to stay while the method is redesigned.

Finally, present two explanations that predict the same result. Ask what new observation would make them disagree. This trains evidence discrimination rather than answer guessing.

Useful Internal Routes

Authoritative References

Singapore Examinations and Assessment Board — PSLE Science syllabus, for examination from 2026
Ministry of Education, Singapore — Science Teaching & Learning Syllabus, Primary, 2023
National Research Council — A Framework for K–12 Science Education, used as broader scientific-practice context rather than PSLE marking policy.

Evidence and Boundary Note

This guide is a learning scaffold, not an official claim that PSLE Science requires one fixed wording for an “investigable question”. Different scientific questions legitimately need different evidence structures. The durable skill is to match the question, observable evidence and conclusion.

The Quiet Return

A method can collect data.

A good scientific question makes sure those data can actually mean something.

Small Group Tutorials