How to Choose the Next PSLE Science Measurement When the Evidence Cannot Yet Decide

Wait, what? When the evidence is not enough, “measure more things” is usually a bad scientific plan.

Good scientific inquiry is not a shopping list of extra measurements. The useful next measurement is the one that reduces a specific uncertainty. If two explanations, two conclusions or two interpretations still fit the evidence, your job is to find the smallest additional observation that would make those possibilities behave differently.

Quick Answer

First state exactly what the current evidence cannot decide. Then list the remaining possibilities. Ask: “What one observation or measurement would be different if possibility A were true instead of possibility B?” Choose that measurement, keep the scientific question unchanged, decide in advance what different possible readings would mean, and only then collect or interpret the new evidence.

The PSLE Science Learning Job This Guide Owns

This guide owns a narrow inquiry job: choosing the next evidence item when the existing evidence leaves one important uncertainty unresolved. It does not own the whole design of a new investigation. It does not replace the initial job of deciding what outcome to measure. It begins after you already have some evidence and need to decide what would genuinely help next.

The reasoning chain is:

READ the evidence you already have.
IDENTIFY the exact question that remains unresolved.
DISTINGUISH what is known from what is still possible.
LIST the remaining scientific possibilities.
CHOOSE an observation or measurement that would separate those possibilities.
PREDICT what different possible results would mean.
INTERPRET the new evidence without moving the goalposts.
UPDATE the conclusion only as far as the combined evidence allows.

Why This Belongs in Scientific Inquiry

For the 2026 PSLE, SEAB identifies Science as revised and states that it assesses the 2023 Primary Science syllabus. The published assessment objectives include interpreting and analysing information, evaluating observations, information and methods, making predictions or hypotheses, and communicating explanations and reasoning. Choosing useful evidence is therefore not separate from Science reasoning: it is part of learning how evidence can and cannot support a claim.

This guide does not claim that one particular “next measurement” format is required in the national examination. It is a learning protocol for understanding inquiry and evidence.

The Core Distinction: More Data vs More Decisive Data

Extra evidence	Does it help?	Why?
Repeating a measurement already known very precisely	Maybe not	It may improve reliability but leave the scientific uncertainty untouched.
Measuring an unrelated quantity because the apparatus is available	Usually no	More numbers do not automatically answer the question.
Measuring the same outcome at a time when two possibilities predict different values	Often yes	The reading can discriminate between the possibilities.
Checking a condition that one explanation requires but another does not	Often yes	The evidence directly tests the difference between the explanations.
Adding a measurement that changes the original system	Potentially harmful	The new method may create a different scientific question.

The key idea is discrimination. A useful next measurement should make at least two remaining possibilities give different expected evidence.

Worked Example 1: Same Final Value, Different Possible Histories

Imagine an original practice investigation in which two set-ups begin with the same measured quantity and end with the same measured quantity after 20 minutes. The learner wants to know whether the quantity stayed constant throughout or changed and later returned to the same value.

The final measurement cannot decide between those histories. Taking the final measurement again may confirm the endpoint but still does not tell you what happened in between.

A more useful next evidence item would be a measurement at a clearly chosen intermediate time. Before measuring, write the predictions:

If the quantity remained constant, the intermediate reading should match the starting and final values within the measurement limits.
If the quantity changed and later returned, an intermediate reading may differ.

Notice the careful word may. A single intermediate reading does not prove the entire path. It is useful because it can rule out some simple interpretations if it differs. Scientific evidence has limits even when it is well chosen.

Worked Example 2: Two Explanations Fit One Observation

Suppose a set-up gives an unexpected result. Two explanations remain possible. Explanation A requires condition X to have changed. Explanation B does not require X to change.

The useful next question is not “What else can I measure?” It is “Can I observe or measure X well enough to decide whether Explanation A is still possible?”

If X is found to have remained within the intended condition, Explanation A becomes less supported. If X changed in the direction Explanation A requires, that does not automatically prove A, but it gives the explanation a piece of evidence it previously lacked.

This is scientific reasoning at the right strength: evidence can strengthen, weaken, eliminate or leave an explanation unresolved. It does not always produce certainty.

Worked Example 3: The Extra Measurement That Looks Useful but Is Not

Imagine two test conditions produce different outcomes. The unanswered question is whether the difference is caused by the stated test condition or by another condition that was not kept sufficiently similar. A learner suggests measuring the mass of the container because a balance is available.

If container mass does not affect either remaining explanation, that measurement is scientifically decorative. It may be accurate, but it is not decisive.

The repair is to ask what information would change the conclusion. If knowing container mass would not change what you believe, it is probably not the next measurement you need.

Four Kinds of Uncertainty

Uncertainty	Question to ask	Possible next evidence
State uncertainty	What is the system like now?	Measure or observe the relevant state directly if possible.
Time uncertainty	When did the change occur?	Add a strategically chosen observation time.
Condition uncertainty	Was an important condition actually different or controlled?	Measure or verify that condition.
Explanation uncertainty	Which of two mechanisms better fits the evidence?	Choose an observation for which the mechanisms predict different outcomes.

Before choosing the tool, choose the uncertainty. That prevents apparatus from driving the scientific question.

The NEXT Protocol

N — Name what cannot yet be decided

Write one sentence. Example: “The present data do not show whether ___.” If you cannot name the uncertainty, you are not ready to choose the measurement.

E — Enumerate the remaining possibilities

Keep them scientific and bounded. You do not need twenty imaginative stories. Usually two or three evidence-compatible possibilities are enough for the learning task.

X — Find the point where the possibilities separate

Ask: “If A were true and B were true, what would be different that I could actually observe or measure?” That difference is your candidate evidence.

T — Test the measurement before taking it

Imagine each possible result. Would result 1 support A more? Would result 2 support B more? Would some results leave both possible? If every possible reading leaves your conclusion unchanged, the measurement is probably not useful for this uncertainty.

Observable Failure Signatures

You choose apparatus before stating the scientific uncertainty.
You suggest measuring every quantity named in the question.
You repeat an already secure measurement while the unresolved issue concerns a different condition.
You collect more data but cannot say how any possible result would change the conclusion.
Your extra measurement introduces a new variable and changes the scientific question.
You choose the easiest measurement rather than the one most relevant to the uncertainty.
You treat one new result as proof when it only reduces uncertainty.

Find the Earliest Weak Link

If a learner proposes an irrelevant extra measurement, do not start by correcting the apparatus choice. Ask whether the learner can state what is currently unknown.

The sequence of diagnosis is:

Can the learner describe the existing evidence?
Can the learner say what that evidence supports?
Can the learner say what it does not support?
Can the learner name the remaining possibilities?
Only then: can the learner choose a measurement that separates them?

If step 3 fails, the problem is evidence limitation, not apparatus selection. Repair the earliest weak link.

Fair-Test Logic Still Matters

A useful next measurement must not quietly damage the comparison. Ask whether taking it changes the system, whether it is measured at comparable times or positions, whether the same instrument or method is needed across set-ups, and whether the reading actually represents the scientific quantity of interest.

For example, if opening a sealed container to measure something changes the condition that matters, the measurement procedure itself may alter the result. The evidence then needs to be interpreted with that limitation visible.

Measurement Is Not the Same as Explanation

A measurement tells you what was observed or measured. A scientific explanation connects relevant conditions and concepts through a mechanism to the outcome. Do not replace the explanation with a number.

The full chain remains:

READ THE GIVEN INFORMATION → IDENTIFY THE OBJECT OR RELATIONSHIP → DISTINGUISH OBSERVATION FROM INFERENCE → SELECT THE RELEVANT CONCEPT → EXPLAIN THE MECHANISM → CONNECT TO THE CONDITION → STATE THE OUTCOME → CHECK AGAINST THE EVIDENCE.

The new measurement becomes one more evidence input to that chain. It does not become the chain itself.

Practice Sequence

Evidence limit practice: read a small data set and write one thing it supports and one thing it cannot yet decide.
Possibility practice: generate two scientifically plausible possibilities that fit the same evidence.
Discrimination practice: name one observation that should differ between the possibilities.
Prediction practice: state what different possible readings would mean before seeing the result.
Transfer practice: repeat with a different Science theme and a different representation.
Delayed return: solve a fresh case after two or three days without the protocol visible.

Unfamiliar Transfer Challenge

Create an original scenario with two set-ups that produce the same final observation but could have reached it by different routes. Give only beginning and ending data. Your task is not to guess the route. Your task is to choose the single next measurement that would be most informative and explain what each possible result would tell you.

Then change the surface story completely—different objects, different units, different Science theme—and see whether your selection method survives. If it does, you have learned a reasoning operation rather than memorised a scenario.

Delayed Independent Return Test

On a fresh question after a delay, try to produce these four receipts without help:

Known: what the evidence already establishes.
Unknown: the exact unresolved issue.
Next evidence: one observation or measurement that targets that issue.
Interpretation: what different possible results would do to the conclusion.

If you can choose the measurement but cannot explain how its result would change the reasoning, the job is not finished.

Answer-Checking Receipt

Have I stated the exact uncertainty?
Does my chosen measurement address that uncertainty directly?
Would at least two possible readings lead to meaningfully different interpretations?
Am I measuring the right object, place, time and quantity?
Does the method preserve the original scientific question?
Have I avoided claiming more than the new evidence can support?

Common Traps

Apparatus trap: choosing what to measure because the tool is available.
More-is-better trap: adding many measurements instead of one decisive one.
Precision trap: improving precision on the wrong quantity.
Confirmation trap: choosing only a measurement expected to support your favourite explanation.
Goalpost trap: changing the question after seeing the new evidence.
Certainty trap: treating one additional reading as complete proof.

Parent and Tutor Teaching Guide

When a learner says, “We need more information,” do not immediately ask, “What should we measure?” Ask first: “What exactly can we not decide yet?”

Then ask the learner to give two possibilities that still fit the evidence. The best teaching question is often: “What would you expect to observe if the first possibility were true, and what would you expect if the second were true?”

If both possibilities predict the same reading, that proposed measurement is not discriminating enough. Let the learner discover that rather than simply labelling the measurement “wrong”. This builds scientific inquiry instead of answer imitation.

Useful Internal Routes

Authoritative References

Quiet Return

When evidence cannot yet decide, do not panic and do not guess. Name the uncertainty. Find the point where the remaining possibilities differ. Measure there. The scientific skill is not collecting the most data. It is knowing which evidence has a job.

Small Group Tutorials