Wait, What? Two Experiments Can Point in the Same Direction Without Answering the Same Question
A learner sees two investigations and thinks, “Great—more evidence.” Sometimes that is true. But two investigations can use different objects, different variables, different measured outcomes or different ranges. If the scientific questions are not aligned, simply pooling all the data can produce a conclusion neither investigation actually supports.
COMBINE EVIDENCE ONLY AFTER YOU CHECK WHAT EACH INVESTIGATION ACTUALLY TESTS.
Quick Answer
When two PSLE Science investigations are presented together, use this route:
WRITE THE QUESTION FOR INVESTIGATION A → WRITE THE QUESTION FOR INVESTIGATION B → ALIGN OBJECTS, CHANGED CONDITIONS, MEASURED OUTCOMES, RANGE AND TIME → DECIDE WHETHER B REPEATS, CORROBORATES, NARROWS, EXTENDS OR TESTS A DIFFERENT RELATIONSHIP → COMBINE ONLY THE CLAIMS THAT SHARE A SCIENTIFIC BASIS → KEEP NON-COMPARABLE DATA SEPARATE.
The Exact PSLE Science Learning Job This Guide Owns
This guide owns one learner job: synthesising evidence from two separate investigations without pretending that different scientific questions, methods or data sets form one interchangeable experiment.
It does not own combining text, diagrams and data from one question, qualitative and quantitative evidence, or replication versus extension by itself. Those are separate jobs. Here the learner must decide how two investigation-level evidence packages relate.
Why This Matters for PSLE Science
The 2026 PSLE Science assessment objectives include interpreting and analysing information, evaluating observations, information and methods, and communicating explanations and reasoning. When more than one investigation is described, the learner must preserve the question and evidence boundaries of each before drawing a combined conclusion.
Five Possible Relationships Between Two Investigations
| Relationship | What it means |
|---|---|
| Replication | Tests essentially the same question again under comparable conditions |
| Corroboration | Uses different evidence or method that supports the same claim |
| Narrowing | Tests a smaller region or condition to resolve uncertainty |
| Extension | Tests whether the relationship also holds in a wider range or different case |
| Different question | Looks related but measures a materially different relationship |
Worked Example 1 — Same Relationship, Different Range
Investigation A tests how a changed condition affects an outcome over values 1 to 4. Investigation B uses the same objects and outcome but tests values 5 to 8.
The two investigations may jointly extend the tested range, but only if the methods and variable meanings remain sufficiently comparable. Do not automatically fit one trend across all eight values if the method changed or the relationship behaves differently in the second range.
Worked Example 2 — Different Outcome, Related Concept
Investigation A measures the time taken for a process. Investigation B measures the final amount produced after a fixed time. Both involve the same scientific concept, but the dependent quantities are different.
You may use both to reason about the same broad mechanism, but you should not merge the numerical data into one graph or treat “shorter time” and “greater amount after five minutes” as the same measurement.
Worked Example 3 — Different Specimens, Same Claim
Investigation A uses one type of specimen and Investigation B uses another. Both show the same directional relationship under otherwise similar conditions.
That may strengthen confidence that the relationship is not unique to one specimen type, but only if the question asks for that broader interpretation. It does not prove the relationship holds for every possible specimen or condition.
Worked Example 4 — Two Investigations That Should Stay Separate
Investigation A asks whether changing light level affects one plant response. Investigation B asks whether changing water amount affects a different response. Both are about plants, but the scientific questions are not the same.
Topic similarity is not evidence alignment. Keep the conclusions separate.
The Alignment Grid
| Check | Investigation A | Investigation B |
|---|---|---|
| Scientific object/system | ? | ? |
| Changed condition | ? | ? |
| Measured outcome | ? | ? |
| Controlled conditions | ? | ? |
| Time basis | ? | ? |
| Tested range | ? | ? |
| Method quality | ? | ? |
You do not need to draw this table in an examination. Use it in practice until you can perform the alignment mentally.
Evidence Can Agree at Different Levels
Two investigations may agree on direction without agreeing on size. They may both support “increases” while showing different amounts of increase. They may support the same mechanism but under different conditions. They may agree only within overlapping ranges.
State the level of agreement precisely. Do not turn “both support an increase” into “both produced the same result”.
Do Not Double-Count the Same Evidence
If Investigation B simply redraws or reanalyses data from Investigation A, it is not independent evidence. Trace where the data came from. Separate a genuinely new investigation from a new representation of an old one.
The PSLE Science Reasoning Chain
READ INVESTIGATION A → IDENTIFY ITS RELATIONSHIP → READ INVESTIGATION B → IDENTIFY ITS RELATIONSHIP → ALIGN VARIABLES AND CONDITIONS → DISTINGUISH SHARED EVIDENCE FROM DIFFERENT EVIDENCE → SELECT THE RELEVANT CONCEPT → STATE THE COMBINED CLAIM → CHECK THAT BOTH INVESTIGATIONS SUPPORT IT.
Observable Failure Signatures
| Failure signature | Likely weak link |
|---|---|
| All values from both investigations are placed in one graph immediately | Variable alignment not checked |
| Same topic is treated as same question | Surface similarity replaced scientific comparison |
| Two outcomes with different units are averaged together | Measured quantities confused |
| A second investigation is called stronger evidence even though it reuses the same data | Evidence provenance lost |
| A relationship found in one specimen is generalised to all cases after one extension | Conclusion travels beyond tested scope |
| Different ranges are treated as if both tested the same interval | Range conditions ignored |
Find the Earliest Weak Link
- What question does Investigation A answer?
- What question does Investigation B answer?
- Are the changed conditions the same variable?
- Are the measured outcomes the same quantity?
- Are time, units and range comparable?
- Does B repeat, support, narrow or extend A?
- Which claim is supported by both?
- What must stay separate?
Misconception Repair — “More Investigations Mean Stronger Evidence”
Only if the additional investigation contributes relevant evidence to the claim. An unrelated investigation adds information, not necessarily support.
Misconception Repair — “Same Concept Means Same Data”
Two investigations can involve the same concept while measuring different quantities. Preserve the measurement role before combining conclusions.
Misconception Repair — “If Both Trends Rise, They Are the Same Relationship”
Direction alone is not enough. Check the variables, conditions, ranges and meaning of the axes before deciding how far the agreement extends.
Practice Protocol: Two Investigations, One Claim?
- Write one sentence for the question of Investigation A.
- Write one sentence for the question of Investigation B.
- Underline the changed variable in each.
- Circle the measured outcome in each.
- Mark shared conditions.
- Decide the relationship between the investigations.
- Write one combined conclusion that both support.
- Write one tempting combined conclusion that goes too far.
Unfamiliar Transfer Challenge
Create two original investigations around the same Science concept. Make them look similar but change one important evidence role: outcome, range, specimen or timing. Then decide whether they should corroborate, extend or remain separate.
Delayed Independent Return
Three to five days later, use two fresh investigation summaries. Before reading any model explanation, classify the relationship between them and write the strongest claim both can support. If you resist merging merely because the topic looks similar, the skill is transferring.
Cross-Investigation Receipt
- I identified the question for each investigation.
- I aligned objects, variables, outcomes, time and range.
- I know whether the second investigation repeats, corroborates, narrows, extends or differs.
- I did not pool non-comparable measurements.
- I did not double-count reused evidence.
- I stated the level of agreement precisely.
- My combined conclusion is supported by both investigations.
Parent and Tutor Teaching Guide
Give the learner two short investigation summaries and hide the results at first. Ask them to state each scientific question. If the learner cannot do that, do not move to combined conclusions yet.
Next, reveal the results and ask, “What can both investigations support?” Then ask, “What does only one investigation support?” This prevents one evidence set from silently lending authority to the other.
Finish with a transfer task where the investigations share a topic but not a measured outcome. The learner should learn that connection does not mean merger.
Useful Internal Routes
- PSLE Science Learning Guide
- Tell replication from extension
- Decide whether two data sets show the same relationship
- Avoid double-counting the same evidence
- Keep conclusions within what was tested
Authoritative References
- SEAB — PSLE Science syllabus, for examination from 2026
- MOE Singapore — Science Teaching & Learning Syllabus, Primary, 2023
Evidence and Boundary Note
This guide is a learning framework for evidence synthesis, not an official PSLE answer template. Individual questions determine whether two investigations should be compared, combined or kept separate.
The Quiet Return
Two investigations can make one conclusion stronger.
But first you must prove that they are speaking about the same scientific relationship.
