How to Tell Whether a PSLE Science Difference Existed Before the Test or Appeared During It

Wait, What? A bigger final difference does not prove the test created the difference.

Imagine Set-up A ends with a value of 18 and Set-up B ends with a value of 12. It is tempting to say the tested condition caused A to be 6 units higher. But what if A already began 5 units higher before the test? The final gap then contains two pieces of history: a difference that existed before the investigation and any additional difference that developed during it. PSLE Science learners need to separate those histories before assigning a cause.

Quick Answer

Look for a starting measurement or starting state before interpreting the final difference. Compare each set-up with its own starting point, then compare the changes. If the set-ups started differently, do not treat the entire final gap as an effect of the tested condition. If no starting evidence is given, state only what the final comparison supports and avoid inventing an equal baseline.

The PSLE Science Learning Job This Guide Owns

This guide owns one precise learner job: distinguish a pre-existing difference from a difference that arose during a PSLE Science investigation. It does not own the underlying science concept. It does not turn every investigation into a before-and-after test. It teaches how to use baseline evidence, matched comparisons and time order to decide what the test may reasonably have changed.

The PSLE Science syllabus for examination from 2026 includes interpreting and analysing information, evaluating observations, information and methods, and communicating explanations and reasoning. Starting conditions matter because evidence about change is only interpretable when you know what the compared systems were like before the tested condition acted.

Final Difference and Change Are Not the Same Quantity

Suppose:

Set-up A starts at 10 and ends at 18.
Set-up B starts at 5 and ends at 12.

The final difference is 6 units. But A changed by 8 units and B changed by 7 units. The difference in change is only 1 unit.

If the scientific question asks which set-up changed more, comparing 18 with 12 answers the wrong question. The starting values matter.

Worked Example 1: A Difference That Was Mostly There Already

Two similar specimens are placed under different conditions. Before the test, Specimen P has a measured value of 14 and Q has a value of 10. After the test, P is 20 and Q is 15.

A learner says, “Condition P caused a 5-unit advantage because P ended at 20 while Q ended at 15.” That statement ignores the starting gap of 4 units. During the test, P increased by 6 and Q increased by 5. The evidence shows P changed 1 unit more over the test period, not that the test created the entire 5-unit final difference.

The correct scientific interpretation may still depend on the mechanism, natural variation and method. But the arithmetic of the evidence already tells us the final gap cannot simply be credited to the test.

Worked Example 2: Equal Starts Make a Final Difference Easier to Interpret

Now suppose P and Q both start at 10. After the test, P is 18 and Q is 12. If the set-ups are otherwise suitably comparable and differ in the intended tested condition, the final difference also reflects the difference in change: P changed by 8 and Q by 2.

Equal starting values do not prove the test is perfect. Other relevant conditions still need checking. But the baseline no longer explains the final difference.

Worked Example 3: No Starting Measurement Means a Limit on the Conclusion

Two living specimens are measured only after ten days under different conditions. Specimen A is taller than Specimen B. The question does not provide their starting heights.

You may accurately state that A is taller at the end. You cannot automatically conclude that A grew more during the ten days because A may also have been taller at the start. To support a claim about growth, a starting measurement or another valid design feature would be needed.

This is an evidence-boundary lesson: final state is not automatically evidence of amount of change.

Worked Example 4: A Starting Difference Can Be Scientifically Important

A starting difference is not always a mistake. Sometimes an investigation deliberately begins with different starting conditions because that difference is the variable being studied. For example, different starting temperatures may be intentionally tested. The learner must ask whether the starting difference is the intended condition or an unwanted mismatch.

Do not apply “starting values must be identical” as a universal rule. Ask what relationship the investigation is designed to test.

A Baseline–Change–Final Table

Set-up	Starting value	Final value	Change
A	10	18	+8
B	5	12	+7

The final values differ by 6. The changes differ by 1. Those are different scientific comparisons. Which one matters depends on the question.

Ask What the Question Is Actually Comparing

Final state: Which set-up has the larger value at the end?
Change: Which set-up increased or decreased more?
Rate: Which changed faster over a stated interval?
Effect of a tested condition: What difference can reasonably be connected to the condition after accounting for starting state and other relevant factors?

Do not answer one comparison with another merely because the same numbers appear in the table.

Baseline Evidence Can Be Descriptive or Numerical

A baseline does not always have to be a number. A question may tell you that two specimens initially had the same colour, size category, state, number of leaves or other relevant feature. That starting observation can still help establish whether a later difference appeared during the test.

What matters is whether the starting evidence is relevant to the outcome being interpreted.

When Baseline Differences Confuse Cause

If two set-ups differ before the tested condition begins, that pre-existing difference may be an alternative explanation for a later result. The learner should keep both possibilities alive until the design or evidence discriminates between them.

For example, if two plants start at different heights and receive different amounts of light, the final height difference cannot be attributed cleanly to light without accounting for the starting height difference. The test has changed more than one scientifically relevant starting condition unless the analysis focuses on growth from baseline.

Fair Test Does Not Mean Every Starting Value Must Be Identical

Real specimens often vary. A fair comparison aims to prevent relevant alternative differences from being confused with the tested condition. Sometimes this is achieved by choosing similar specimens. Sometimes by using a starting measurement and comparing change. Sometimes by repeating across several specimens. There is no single method improvement that fits every investigation.

The Baseline Reasoning Protocol

Identify the measured outcome.
Find the starting evidence for that same outcome or a relevant starting condition.
Ask whether the set-ups already differed before the test.
Calculate or describe the change within each set-up.
Compare the changes on the same basis.
Check other relevant conditions and method limitations.
State only the effect the evidence supports.

How This Appears in Graphs

Two graph lines can begin at different vertical values. If you compare only the final endpoints, you may confuse starting separation with later change. Read the first point for each series, then examine how much each line changes, whether their slopes differ and whether the question asks about final value, change or rate.

A line starting higher does not automatically mean the tested condition made it higher.

How This Appears in Before-and-After Diagrams

Diagram sequences can hide baselines because the learner focuses on the striking final picture. Compare each object with its own starting state first. Then compare the amount or type of change between objects. Do not assume two different-looking final diagrams began from identical conditions unless the question shows that they did.

How This Appears in Investigation Evaluation

If a method compares different specimens only at the end, ask whether a starting measurement would help distinguish initial variation from test-related change. But do not recommend a starting measurement automatically. It must be relevant, measurable and compatible with the scientific question.

Failure Signatures

The learner compares final values when the question asks about change.
The entire final gap is attributed to the tested condition even though the set-ups began differently.
A missing baseline is silently assumed to be equal.
The learner subtracts final values from each other but never compares each set-up with its own start.
A starting difference that is intentionally the tested variable is incorrectly “controlled away”.
The learner says one specimen grew more simply because it ended taller.
A graph’s higher starting line is mistaken for evidence of a stronger treatment effect.

Earliest Weak-Link Diagnosis

Ask the learner to point to the starting value for each set-up before discussing the final result. If they cannot find one, ask whether the question provides a baseline at all. If it does not, the learner should be able to name the conclusion that becomes unsafe.

If the starting values are available but the learner still compares only endpoints, the weakness is comparison selection. If they correctly compare changes but then assign a mechanism incorrectly, the weakness is scientific explanation rather than baseline reasoning.

Misconception Repair

“Bigger final value means bigger increase.” Not when starting values differ.

“If there is no starting measurement, the set-ups must have started the same.” Absence of evidence is not evidence of equality.

“Every investigation needs identical baselines.” No. Some starting differences are intentionally tested. What matters is whether the design answers the scientific question.

“A starting difference makes the whole investigation useless.” Not necessarily. The method may allow change-from-baseline analysis, matched specimens, repeated evidence or another valid comparison.

The PSLE Science Reasoning Chain With Baseline Control

READ THE STARTING STATE → IDENTIFY THE TESTED CONDITION → READ THE FINAL STATE → CALCULATE OR DESCRIBE CHANGE → COMPARE CHANGES → SELECT THE RELEVANT CONCEPT → EXPLAIN THE MECHANISM → CHECK THAT THE CLAIM DOES NOT CREDIT THE TEST FOR A PRE-EXISTING DIFFERENCE.

Original Practice Set

Practice A: A starts at 12 and ends at 20. B starts at 7 and ends at 16. Which ends higher? Which changes more? Are those the same answer?

Practice B: Two specimens end at different heights, but no starting heights are given. Write one conclusion the evidence supports and one stronger conclusion it does not yet support.

Practice C: Two set-ups intentionally begin at different temperatures because starting temperature is the variable being tested. Explain why “make the starting temperatures the same” would destroy the scientific question rather than improve it.

Practice D: A graph line begins above another line but rises less steeply. Explain why final ranking and amount of change can point in different directions.

Retrieval and Transfer Sequence

Practise distinguishing starting value, final value and change.
Move to pairs with equal starts, then unequal starts.
Use tables, graphs and before-and-after diagrams.
Mix cases where the starting difference is a flaw with cases where it is the intended variable.
After a delay, ask the learner to decide whether a final difference was created during the test without prompting them to look for baseline data.

Unfamiliar Transfer Test

Give the learner a new context with two specimens, unequal starting measurements and unequal final measurements. The learner passes if they independently separate final difference from change, identify the baseline as part of the evidence history and avoid attributing the entire endpoint gap to the tested condition.

Delayed Independent Return Test

Several days later, present a graph in which one line starts higher but changes less. Ask which condition caused the greater change. Do not mention “starting values”. The learner should inspect the first points, reconstruct the changes and explain why the final endpoint alone is insufficient.

Answer-Checking Receipt

I know what each set-up was like before the test.
I know whether the question asks about final value or change.
I did not assume missing baselines were equal.
I compared each set-up with its own start where appropriate.
I checked whether the starting difference was intentional or unwanted.
I did not credit the tested condition for a difference that existed beforehand.
My conclusion is no stronger than the design and evidence allow.

Parent and Tutor Teaching Guide

Give the child two runners who start at different positions on a line. Ask who finishes farther ahead and who actually moves farther. This non-science warm-up makes final position versus change visible. Then return immediately to Science and replace position with temperature, length, mass, count or another measured outcome.

When correcting a mistake, ask, “How much of this difference was already there before the test?” That question often exposes the earliest weak link. If the child automatically says “none”, ask them to show the baseline evidence.

Do not teach subtraction as the main goal. The arithmetic serves the scientific reasoning. The real skill is preserving the history of the set-ups.

Useful Internal Routes

Authoritative References

Quiet Return

An investigation does not erase the history that existed before it began. Starting states travel into later evidence. Once you separate what was already different from what changed during the test, a final result becomes much more informative—and much harder to overclaim.

Small Group Tutorials