How to Answer “How Do You Know?” in PSLE Science With Evidence Instead of a Mechanism

Wait, What? A Scientifically Correct Explanation Can Still Fail to Answer “How Do You Know?”

A results table shows that Set-up P lost 18 mL of water while Set-up Q lost 7 mL. The question asks how you know that more water was lost from P.

A learner writes: “P had a larger exposed surface, so more water evaporated.” That may be a reasonable mechanism if the conditions support it. But it does not answer the evidence question. The learner was asked how the conclusion is supported by the given information, not why the scientific effect occurred.

PSLE Science requires both kinds of thinking. Strong learners know when the question wants the evidence and when it wants the mechanism.

Quick Answer

When a PSLE Science question asks “How do you know?”, “What evidence supports…?” or an equivalent evidence-justification prompt, identify the exact claim first. Then cite the decisive observation, measurement or comparison that supports that claim. Keep the units, object and reference clear. State what the evidence establishes. Add a mechanism only if the question also asks why.

The learner chain is: CLAIM → DECISIVE EVIDENCE → VALID COMPARISON → WHAT THE EVIDENCE SHOWS → LIMIT → MECHANISM ONLY IF REQUIRED.

Owned PSLE Science Learning Job

This guide owns one response job: answering evidence-justification questions with the observation, data or comparison that supports the claim rather than substituting a causal explanation.

It does not replace the broader guide on choosing decisive evidence, which teaches evidence selection across many answer types. It also does not replace the result–conclusion–explanation guide. Here the focus is the student-facing question boundary: when the examiner or practice task asks how you know, what should actually appear in the answer?

The Current PSLE Science Frame

The 2026 PSLE Science examination assesses the 2023 Primary Science syllabus. SEAB identifies interpreting and analysing information, evaluating observations and information, and communicating explanations and reasoning as part of scientific inquiry.

Those objectives make evidence use central. But they do not create one universal school phrase or marking template. A good response must fit the exact question and the information supplied.

Evidence and Mechanism Have Different Jobs

Question job	What the answer should mainly contain	Example
How do you know P lost more water?	Relevant measurements and comparison.	P lost 18 mL while Q lost 7 mL, so P lost 11 mL more.
Why did P lose more water?	Relevant scientific mechanism tied to the condition.	The larger exposed surface allowed more water to leave the liquid surface over the same period, under the stated comparable conditions.
What can you conclude?	Evidence-bounded relationship.	Under the tested conditions, the larger exposed surface was associated with greater water loss.

All three answers can be scientifically connected. They are not interchangeable.

Worked Example 1: Evidence From a Table

An original practice table shows that Plant P increased from 9 cm to 15 cm while Plant Q increased from 10 cm to 13 cm.

Claim: P increased in height more than Q.

Evidence: P increased by 6 cm while Q increased by 3 cm.

A weak “how do you know?” answer says: “P had better conditions for growth.” That is not evidence from the table. A stronger answer cites the actual changes and lets the comparison support the claim.

Worked Example 2: Evidence From a Diagram

A diagram shows a bulb lit in Set-up X and unlit in Set-up Y. The question asks how you know current can flow through the complete path in X.

One relevant piece of evidence is that the bulb in X lights under the stated set-up. If the question asks why X forms a complete path, then the learner must move beyond the observation and explain the connections.

Keep the observable indicator and the inferred circuit condition separate. The lit bulb is observed. The complete conducting path is the scientific interpretation supported by the set-up and concept.

Worked Example 3: Evidence From a Graph

A graph shows temperature falling from 70°C to 48°C in Cup A and from 70°C to 59°C in Cup B over the same 10-minute interval.

If asked how you know Cup A experienced the larger temperature decrease, cite the values: A decreased by 22°C while B decreased by 11°C over the same interval.

If asked why the decrease differed, you need additional information about the set-ups and a scientific mechanism. Do not invent that mechanism from the graph alone.

Worked Example 4: Evidence Can Be Qualitative

Evidence does not always contain numbers. Suppose two indicators are observed after the same test. Set-up P becomes cloudy while Set-up Q remains clear.

If the question asks how you know the two set-ups produced different observable outcomes, the qualitative observation itself is evidence. Do not invent a numerical amount of cloudiness unless the method provides one.

The scientific meaning of the cloudiness may require a concept or indicator relationship, but the observation remains distinct from that interpretation.

The Five-Part Evidence Sentence

For practice, a useful scaffold is:

OBJECT + MEASURED/OBSERVED QUANTITY + VALUE OR STATE + REFERENCE/COMPARISON + WHAT IT SUPPORTS.

Example: “Set-up P lost 18 mL of water compared with 7 mL in Set-up Q over the same period, so the data show that P lost more water.”

This is not an official required sentence pattern. Its purpose is to prevent evidence from becoming vague.

Do Not Copy the Whole Table

An evidence answer should be sufficient, not enormous. If a conclusion depends on the difference between P and Q, you may need only the two decisive values and the correct reference.

Copying every number can hide the comparison the question actually needs. Ask: which observation or pair of values does the logical work?

But Do Not Under-Cite the Evidence Either

“P is higher” may be too vague if the graph contains several quantities or time points. Higher than what? At which time? In which unit?

Strong evidence language keeps the reference visible: “At 10 minutes, P was 62°C while Q was 54°C.”

Evidence Is Not the Same as a True Fact From Memory

Suppose the question gives data about plant growth and asks what evidence supports the conclusion that P grew more.

“Plants need water to grow” may be scientifically true. It is not the evidence for this particular conclusion unless the question’s data and conditions make water availability relevant to the claim.

Use the given evidence first. Add scientific knowledge when the job shifts from support to interpretation or explanation.

Evidence Is Not the Same as a Conclusion

Evidence: P had 14 leaves and Q had 8 leaves after the stated period.

Conclusion: Under those tested conditions, P had more leaves than Q after the period.

Explanation: A mechanism may account for why the difference occurred.

If you cannot point from the conclusion back to the evidence, the answer may be relying on expectation instead of the question.

When One Piece of Evidence Is Not Enough

Some claims require more than one comparison. A trend across four tested conditions should not be justified by quoting only one pair if the broader pattern matters.

Likewise, a claim that a method is reliable may require repeated measurements or consistency evidence rather than a single correct-looking result.

Match the amount of evidence to the claim. Do not count evidence pieces mechanically; ask what must be shown for the claim to be justified.

When Two Pieces of Evidence Say the Same Thing

A table and a graph may display the same underlying measurements. Repeating both does not automatically create two independent pieces of evidence.

Use multiple representations when they help interpretation, but do not double-count the same data simply because it appears twice.

Failure Signatures and Earliest Weak-Link Diagnosis

Failure signature	Likely weak link	Repair
You explain why instead of showing how you know.	Response-job recognition.	Underline the claim and locate the observation or values that support it.
You copy the whole table.	Evidence selection.	Choose the smallest decisive comparison.
You say “P is more” without a reference.	Comparison precision.	Name the quantity, reference and time/condition.
You use a textbook fact instead of question data.	Evidence-source control.	Separate what the question gives from what you know.
You invent numbers for descriptive results.	Qualitative-evidence handling.	Use the categories or observations exactly as recorded.
You cite one pair to justify a whole trend.	Claim–evidence scope.	Use enough of the ordered data to establish the broader pattern.

Misconception Repair: “Because” Does Not Automatically Mean Evidence

Students sometimes begin every Science answer with “because”. But the word does not decide the reasoning job.

“I know P lost more water because P lost 18 mL while Q lost 7 mL” uses evidence. “P lost more water because evaporation was faster” offers an explanation. Both contain “because”; their logical roles differ.

Read the relationship between the statements, not merely the connector word.

A Three-Question Self-Check

What exact claim am I supporting?
Which observation, measurement or comparison directly supports it?
Am I adding a mechanism the question did not ask for?

For dense data, add a fourth question: “Does the claim require one comparison or a pattern across several points?”

Worked Example 5: “How Do You Know?” With an Investigation

Four identical pieces of material are tested with the same load. Their bending distances are 2 mm, 4 mm, 7 mm and 3 mm. The question asks how you know Material C bent the most.

Evidence answer: Material C bent 7 mm, which is greater than the bending distances recorded for A, B and D.

Do not answer “because C is the most flexible” unless flexibility has been defined and the method supports that broader interpretation. The direct evidence is the measured bending distance.

Worked Example 6: Evidence With a Before–After Comparison

A fruit has mass 120 g before drying and 86 g afterward. Another fruit has mass 115 g before and 101 g afterward.

If asked how you know the first fruit lost more mass, compare change, not final values. The first lost 34 g; the second lost 14 g.

The evidence answer depends on the correct reference. This is why “how do you know?” still requires scientific and mathematical care.

Practice Sequence

Start with two values and identify the decisive comparison.
Use a graph and cite time, object, value and unit.
Use qualitative observations without inventing numbers.
Separate an evidence question from a mechanism question about the same situation.
Use a four-point trend and decide how much evidence is needed to justify the pattern.
Use a flawed claim and state that the evidence is insufficient rather than inventing support.

Unfamiliar Transfer Challenge

A fictional device records “response units”. X gives 12 units, Y gives 19 and Z gives 14 under the stated comparable test. You do not know what the device measures.

You can still answer: “How do you know Y had the highest recorded response?” The evidence is the recorded value of 19 units compared with 12 and 14.

You cannot explain why Y produced that response without additional scientific information. Evidence use can therefore be sound even when the mechanism is unknown.

Delayed Independent Return Test

Several days later, mix six short prompts: two asking for evidence, two asking for conclusions and two asking for mechanisms. Remove obvious headings.

The learner should first identify the response job, then answer. Success is not merely knowing the Science; it is selecting the right logical layer without being told which template to use.

Evidence-Answer Receipt

What claim am I supporting?
Which exact object or set-up is involved?
Which observation or measurement is decisive?
What is the correct comparison or reference?
Have I preserved units, time and conditions?
Does the evidence actually support the strength of the claim?
Have I avoided replacing evidence with a mechanism?
If the question also asks why, have I then added the relevant scientific mechanism?

Parent and Tutor Teaching Guide

When a child gives a mechanism to an evidence question, do not say the Science is completely wrong. Say, “That explains why. Now show me how the data tell you the claim is true.” This preserves the correct knowledge while repairing response selection.

Ask the child to point to the evidence before writing. For graphs and tables, have them name the object, quantity, reference and relevant values aloud. This often exposes reversed comparisons before they reach the answer.

Do not turn evidence answers into keyword hunts. The goal is a defensible link between a claim and the information that supports it.

Useful Internal Routes

Authoritative References and Evidence Boundary

The evidence-answer scaffold here is not an official national marking phrase. Actual responses depend on the question wording and the evidence provided. Science education also distinguishes direct and indirect evidence, observations and inferences, and conclusions of different strengths; this guide keeps those boundaries visible without claiming that every “how do you know?” prompt has one fixed answer form.

The Quiet Return

Science explanations tell us why a result makes sense. Evidence tells us why we are entitled to say the result or conclusion is supported.

When a question asks “How do you know?”, do not rush past the data to the chapter you remember. Stay with what was observed. Name the decisive comparison. Let the evidence earn the claim.

Small Group Tutorials