Checked: 3 September 2026. This supporting Primary Science guide is aligned to MOE’s Primary Science Syllabus 2023 and the revised 2026 PSLE Science framework. This legacy Holland Village URL does not claim a current eduKateSG Holland Village or Marina Bay branch.
A scientific conclusion should be no stronger than the experiment that produced it. Students often see one result and write a universal rule. Others merely copy the observation and call it a conclusion. Strong reasoning sits between those extremes: identify what was changed, what was measured, what pattern the evidence supports, and what the design still cannot tell us. For Holland Village families, this old 2017 URL now owns one precise job: move from experimental result to a defensible conclusion without overclaiming.
Quick read
- State the relevant result first.
- Connect the measured outcome to the changed condition.
- Use the experiment’s design to judge how strong the conclusion can be.
- Keep the conclusion within the tested conditions and range.
- Separate evidence from mechanism unless the question asks for both.
- Name important limitations where they affect confidence.
Observation, result and conclusion are different
Observation: what was seen or measured.
Result: the pattern or outcome produced by the data.
Conclusion: the claim the evidence supports about the relationship being tested.
These layers should connect, but they are not identical.
Example
Question:
Does water temperature affect dissolving time?
Result:
Sugar dissolved in 35 s at 20°C and 18 s at 40°C under the stated conditions.
Conclusion:
In this experiment, increasing the water temperature from 20°C to 40°C reduced the time taken for the sugar to dissolve.
The conclusion is bounded to the tested range and conditions.
Step 1: restate the scientific question
Ask:
- What factor was deliberately changed?
- What outcome was measured?
- What relationship was the experiment designed to test?
The conclusion should answer that relationship—not a different question.
Step 2: identify the relevant evidence
Use:
- measurements;
- patterns;
- comparison;
- repeated readings;
- observed change.
Do not include unrelated details simply because they appear in the diagram.
Step 3: state the pattern
Useful forms:
- as X increases, Y increases;
- as X increases, Y decreases;
- Y remains approximately unchanged;
- Setup A produced a greater/smaller Y than Setup B;
- no clear relationship was observed.
A conclusion needs the pattern before explanation.
Step 4: match claim strength to design
If the experiment controlled relevant competing variables and directly manipulated the factor, a causal conclusion may be more defensible.
If it merely observed two naturally different groups, stronger causal wording may not be justified.
Students should learn that design controls conclusion strength.
Step 5: keep the claim inside the tested range
If the experiment compares 20°C and 40°C, avoid:
The hotter the water, the faster sugar always dissolves at every temperature.
Safer:
Within the tested conditions, the higher-temperature setup had a shorter dissolving time.
Do not extrapolate farther than the evidence supports.
Step 6: distinguish conclusion from explanation
Conclusion:
The covered leaf showed less evidence of the tested process than the exposed leaf.
Explanation:
Then use the relevant concept—such as the role of light—if the question asks why.
Students should know whether the task wants what the evidence supports or why the outcome occurred.
Repeated trials and confidence
If repeated trials show similar results, confidence in the pattern increases.
If results vary widely, a single average may hide instability.
Students should inspect:
- consistency;
- anomalies;
- measurement method.
Reliability is part of conclusion quality.
An anomaly does not cancel everything
One unusual result may:
- suggest measurement error;
- reflect natural variation;
- show an uncontrolled condition;
- reveal a more complex relationship.
Repeat or investigate rather than deleting it automatically.
Fair test and conclusion
A fairer test reduces alternative explanations.
If two setups differ in both temperature and amount of water, the conclusion “temperature caused the difference” is weakened because volume is another possible cause.
Conclusion strength depends on control quality.
Sample size in simple terms
One organism or one trial may not represent natural variation well.
Where appropriate, more repeated observations can produce stronger evidence.
Primary students need not master statistics to understand the principle: more consistent evidence can increase confidence.
Instrument limits
If the instrument cannot resolve a tiny difference, the conclusion should not treat that difference as definitely meaningful.
Measurement precision and conclusion strength are linked.
“No difference measured” versus “no effect exists”
Suppose two readings are the same to the instrument’s resolution.
Conclusion:
No difference was detected under these conditions.
That is not automatically proof that absolutely no difference exists.
This distinction is a valuable early scientific habit.
Negative results
An experiment may not support the predicted relationship.
That is still a result.
Students should report it honestly rather than forcing the expected answer.
Science values evidence over expectation.
Conclusion versus recommendation
A conclusion states what the evidence supports.
A recommendation suggests what should be done next.
Example:
Conclusion: material A reduced heat loss more than material B.
Recommendation: use material A for the tested insulation purpose.
The second may require practical considerations beyond the experiment.
Conclusion versus prediction
A conclusion interprets existing results.
A prediction uses a supported relationship to anticipate a new result.
Students should not write future guesses as though they were measured outcomes.
Conclusion versus hypothesis
A hypothesis is proposed before testing.
The conclusion says whether and how the evidence supports the tested relationship.
The conclusion should not be copied from the hypothesis regardless of results.
Use evidence language
Useful forms:
- the results show that…;
- under these conditions…;
- the data suggests…;
- the measured value increased/decreased…;
- no clear difference was detected…
Choose wording that matches certainty.
Avoid “proves” casually
One simple school experiment rarely proves a universal claim.
Use supports, shows under these conditions, or the direct bounded relationship.
Scientific confidence should be earned.
What the experiment cannot conclude
Ask:
- Was only one range tested?
- Were other variables uncontrolled?
- Was measurement rough?
- Were there very few trials?
- Does the conclusion require a mechanism not measured?
One limitation can define the evidence boundary clearly.
Example: plant growth
Suppose one plant under more light grows taller than one plant under less light.
A cautious learner asks:
- Were starting sizes similar?
- Same species?
- Same water/soil?
- Was “growth” measured only by height?
- Was one plant enough?
The conclusion may still be useful—but its strength depends on design.
Example: material strength
If a paper strip tears under a lower load than a plastic strip, the experiment supports a comparison under the specific dimensions and conditions tested.
It does not automatically prove “plastic is always stronger than paper” in every form and thickness.
Common failure: observation copied as conclusion
Repair:
Connect the measured difference to the changed condition.
Common failure: universal claim from two setups
Repair:
Bound the claim to the tested conditions/range.
Common failure: mechanism invented from data alone
Repair:
Separate what the experiment measured from the concept used to explain it.
Common failure: expected answer overrides data
Repair:
Report the actual result first. Then consider why it differed from expectation.
Common failure: limitation listed mechanically
Repair:
Name a limitation only if it changes confidence or offers an alternative explanation.
Three students and conclusion strength
In a 3-pax class, give one data set and ask:
- Student A: result statement;
- Student B: conclusion;
- Student C: limitation/evidence boundary.
Then rotate and compare which wording is too strong or too weak.
Catch Up, Keep Up, Move Ahead
Catch Up: changed variable → measured outcome → simple conclusion.
Keep Up: fair-test reasoning, repeated results and bounded conclusions.
Move Ahead: confidence language, limitations, negative results and evidence boundaries.
A 90-minute conclusion lesson
- Identify experimental question.
- Read results.
- State data pattern.
- Write bounded conclusion.
- Separate explanation.
- Identify one limitation.
- Rewrite an overclaim.
- Finish with an unfamiliar investigation.
How parents can help
- Ask “What exactly did the experiment measure?”
- Ask “What can you honestly conclude?”
- Ask whether the claim goes beyond the tested range.
- Ask what else could explain the result.
- Praise honest “the evidence does not show enough” answers when justified.
Current official sources
MOE’s Primary Science Syllabus 2023 emphasises evidence, investigation, reasoning and communication. SEAB’s 2026 PSLE formats page lists revised Science as subject 0009.
What we removed from the old 2017 page
The historical Holland Village Science page used Marina Bay claims, stale phone details, old programme labels and promotional galleries.
This rebuild gives the URL one distinct job: conclusion strength and evidence boundaries.
Holland Village route and boundary
This guide is written for Holland Village families and other Singapore Primary Science learners. It does not claim a current eduKateSG Holland Village or Marina Bay branch. It complements the separate Holland Village Science Diagram article by focusing on experimental conclusions rather than diagram interpretation. Current locations are on the contact page.
Frequently asked questions
Should a conclusion include the actual numbers?
Often a concise comparison is enough, but include key values where they strengthen the evidence and the question requires them.
Can a conclusion be “the hypothesis was wrong”?
State what the results show about the relationship; then note whether that supports the prediction.
Why avoid “proves”?
Because simple experiments usually support a bounded relationship rather than establish a universal truth.
Should every answer mention limitations?
No. Include them when the question asks or when they materially affect what can be concluded.
What is the long-term goal?
A learner who can look at evidence and stop the claim exactly where the evidence stops.
The larger point
A conclusion is scientific not because it sounds certain, but because its confidence fits the experiment that earned it.