Wait, What? A Conclusion Can Sound More Scientific and Still Be Less Correct
Suppose an experiment shows that, in one tested set-up, a plant exposed to more light produced more bubbles per minute. A learner writes: “Plants always grow faster when there is more light.” The sentence sounds confident. It also says much more than the evidence showed.
Good Science does not reward the biggest claim. It rewards the claim that fits the evidence.
A strong scientific conclusion is not the boldest sentence. It is the narrowest sentence that fully answers the question and is justified by the data.
Quick Answer
To write a strong PSLE Science conclusion, identify what was changed, what was measured, what pattern the data actually shows, and the exact conditions under which that pattern was observed. Then state only the relationship supported by that evidence. Do not automatically add “always”, “proves”, “causes”, “best” or “all” unless the investigation genuinely justifies such a claim.
Owned PSLE Science Learning Job
This guide owns a specific inquiry-and-answering job: controlling the strength of a PSLE Science conclusion so it matches the tested conditions and evidence. It does not replace the general evidence owner, the variables-and-fair-tests guide, the method-evaluation guide or the unexpected-results guide.
The Official PSLE Science Frame
SEAB’s 2026 PSLE Science syllabus assesses application of scientific inquiry including interpreting and analysing information, evaluating observations, information and methods, and communicating explanations and reasoning. A conclusion therefore needs two things at once: correct Science and correct evidence discipline.
The Conclusion Reasoning Chain
WHAT WAS CHANGED? → WHAT WAS MEASURED? → WHAT PATTERN IS ACTUALLY SHOWN? → UNDER WHICH TESTED CONDITIONS? → WHAT RELATIONSHIP DOES THIS SUPPORT? → WHAT DOES IT NOT PROVE?
Stage 1 — Separate the Question From the Result
An investigation begins with a question. The result is what happened. The conclusion connects the result back to the question.
Example question: “How does the amount of exposed wet surface affect the time taken for the same volume of water to evaporate under the same surrounding conditions?”
Possible result: Water in the wider shallow dish took less time to evaporate than water in the narrow deep container.
A suitable conclusion should answer the relationship tested: under those conditions, a larger exposed surface area was associated with faster evaporation, shown by the shorter time taken.
Stage 2 — Name the Changed and Measured Variables Clearly
A conclusion becomes vague when it loses the variables. “The bigger one was faster” is not enough. Bigger in what way? Faster in what measured outcome?
- Changed condition: exposed surface area.
- Measured outcome: time taken for a fixed amount of water to evaporate, or another stated measure in the question.
- Observed pattern: larger exposed area corresponded with shorter evaporation time in the tested set-ups.
Stage 3 — Describe the Pattern Before Explaining It
A conclusion may require the relationship shown by data. Do not skip directly to mechanism before establishing the pattern.
If a graph rises as temperature rises, the observation is a pattern between the measured quantities. The explanation of why the relationship occurs is a separate reasoning job.
Observation and explanation may both be needed, but they should not be confused.
Worked Example — Fertiliser and Plant Height
Three groups of similar seedlings receive 0 g, 2 g and 4 g of a fertiliser. After the same time, the average heights are 10 cm, 13 cm and 12 cm.
Weak conclusion: “More fertiliser makes plants taller.”
Why it fails: the data do not show a simple increase across the full tested range. The 4 g group is shorter than the 2 g group.
Better conclusion: “In this investigation, the plants given 2 g of fertiliser had the greatest average height among the three tested amounts.”
If the question asks about the relationship, the learner can state that increasing fertiliser from 0 g to 2 g was associated with greater average height, but increasing it further to 4 g did not produce a further increase.
The data control the conclusion. The learner does not get to force the data into the pattern they expected.
Stage 4 — Keep the Tested Range
If the experiment tested temperatures of 20°C, 30°C and 40°C, the conclusion should not automatically make claims about 5°C or 90°C. If it tested three materials, do not generalise to all materials.
This is a powerful habit: say “among the tested conditions” when the evidence is limited to the tested set.
Stage 5 — Association Is Not Automatically Cause
If more than one relevant condition changed, the investigation may show a difference without identifying one unique cause.
Suppose Set-up A is brighter and warmer than Set-up B, and a plant in A grows more. The result does not allow the learner to conclude that greater light caused the extra growth, because temperature also differed.
A scientifically disciplined conclusion would state the observed difference and recognise that the design cannot isolate which changed condition produced it.
Stage 6 — “Proves” Is Usually Too Strong
Primary Science investigations are models of evidence. One small school experiment rarely proves a universal law by itself. Safer phrases include:
- “The results show that, in these set-ups…”
- “The data support the conclusion that…”
- “Among the tested conditions…”
- “As X increased from ___ to ___, Y…”
- “The investigation suggests…” when uncertainty or design limits matter.
Do not use cautious language mechanically. Use it because the evidence has limits.
Stage 7 — Do Not Replace Data With a Memorised Fact
A learner may know a correct textbook fact and still write the wrong conclusion if the data differ from the expected pattern.
If an experiment gives unexpected results, the conclusion must describe those results honestly. You can later discuss possible explanations, anomalies or method limitations. The first duty is to the evidence given.
Worked Example — Light and an Aquatic Plant
An original investigation places an aquatic plant at three distances from a lamp and counts bubbles produced in one minute. The counts are 8, 14 and 15 bubbles per minute as the lamp moves closer.
Weak conclusion: “The closer the lamp, the more bubbles are always produced.”
Better: “Across the three tested distances, moving the lamp closer was associated with an increase in bubble count, but the increase between the two closest distances was small.”
If the investigation is being used as a model for photosynthesis, the learner can connect the bubble count to the stated experimental interpretation, while remembering that bubble count is a measured proxy in the setup rather than a direct measurement of every aspect of plant growth.
Stage 8 — Distinguish “No Evidence of a Difference” From “Exactly the Same”
Suppose two set-ups both produce a measured value of 12 units. It is fair to say no difference was observed in the measured result under the stated conditions. It may be too strong to say the two systems are identical in every way.
Conclusions should stay attached to what was measured.
Stage 9 — Conclusions Can Be Correct but Incomplete
“Plant A grew more” may match the data, but if the question asks how the changed condition affected growth, the answer should name both the changed condition and the measured response.
Use this structure when suitable:
WHEN [changed condition], [measured outcome] [increased/decreased/remained similar] UNDER [relevant tested conditions].
Stage 10 — Explanation Comes After the Conclusion
If the question asks “conclude”, the main job is evidence-to-claim. If it asks “explain”, add the relevant scientific mechanism. Do not assume every conclusion requires a long mechanistic paragraph, and do not assume every explanation can ignore the data.
PSLE Science Conclusion Checklist
- What exactly was changed?
- What exactly was measured?
- What pattern do the data actually show?
- Is there an exception or plateau?
- Did more than one relevant condition change?
- Am I staying within the tested range?
- Did I write “always”, “all”, “best”, “proves” or “causes” without enough evidence?
- Does my conclusion answer the investigation question?
- Did I accidentally replace the data with a memorised fact?
Observable Failure Signatures and Repairs
“My conclusion is true Science but does not match the graph.” Repair: describe the pattern first, then use scientific knowledge to interpret it.
“I use ‘more X means more Y’ automatically.” Repair: inspect every data point for plateaus, reversals and exceptions.
“I write ‘caused’ whenever two things differ.” Repair: check whether only one relevant condition changed.
“I write a huge paragraph.” Repair: identify the exact relationship the investigation tested and state it first.
“I keep losing the variables.” Repair: write X changed → Y measured before writing the sentence.
The Earliest Weak-Link Diagnosis
A weak conclusion can begin much earlier than the final sentence. The learner may have misidentified the changed variable, misread the graph, ignored a control difference, assumed the expected pattern or confused mechanism with observation.
Fix the earliest error. Rewriting the final sentence alone will not repair the reasoning if the learner still reads the evidence incorrectly.
Unfamiliar Transfer Example
Three identical toy cars roll down ramps set at different heights. Average travel distances are 120 cm, 165 cm and 164 cm. A learner has never seen this exact problem.
The evidence supports an increase between the first and second tested height, but no meaningful further increase is visible between the second and third measured values. The learner should not invent a perfectly increasing rule simply because it feels neat.
Delayed Independent Return Test
Several days later, give a new table with a non-linear pattern, one repeated value or one exception. Without hints, can the learner identify the changed variable, measured outcome, actual pattern, tested range and limit of the conclusion?
Parent and Tutor Teaching Guide
Before correcting the sentence, ask the learner to point to the data that supports each phrase. Useful prompts are: “Which result lets you say that?”, “Did the experiment test all cases?”, “Did only one condition change?”, “Are you describing a pattern or explaining a cause?” and “What is the strongest sentence the data can defend?”
A powerful teaching move is to present two conclusions: one too weak and one too strong. Ask the learner to build the evidence-matched version in between.
Authoritative References
- SEAB — PSLE Science syllabus for examination from 2026.
- MOE — Primary Science Teaching & Learning Syllabus.
- EEF — systematic review of approaches to primary science teaching.
- EEF — feedback evidence summary.
The Quiet Ending
Science becomes trustworthy when the conclusion is answerable to the evidence. That discipline begins long before laboratories and research papers. A Primary 6 learner can practise it now.
Read the data. Respect the conditions. Say what the evidence supports. Stop where the evidence stops.
