How to Spot When the Order of Testing Changes a PSLE Science Investigation

Wait, What? The Second Test May Not Start With the Same Object You Had in the First Test

A learner tests the same cup of water under Condition A. Then, without resetting anything, the learner tests that same cup under Condition B.

That sounds efficient. But there is a hidden question:

Did the first test change the starting state of the second test?

If the first test warmed the water, cooled it, removed some water, added material, stretched an object, changed a surface, depleted a resource or otherwise altered the specimen, the second condition is not being tested on the original starting state.

The order of testing has become part of the experiment.

This is called a carryover effect in broader experimental language: something from an earlier test carries into a later one. Primary learners do not need the jargon. They do need the scientific habit of asking whether yesterday’s or the previous step’s change is still inside today’s result.

Quick Answer

When one specimen, object or setup is tested under several conditions one after another, check whether each new condition begins from a comparable starting state.

Identify what the first test could change.
Ask whether that change remains before the next test begins.
If it remains, the second result may reflect both the new condition and the inherited state from the earlier test.
Decide whether the system can be reset reliably to the same starting condition.
If not, consider comparable fresh specimens or separate setups.
Keep all other relevant conditions as similar as possible.
Interpret the result only within what the method can support.

Use this reasoning route:

READ THE INVESTIGATION → IDENTIFY THE TEST ORDER → TRACK THE STATE AFTER EACH TEST → ASK WHAT CARRIES FORWARD → CHECK WHETHER THE NEXT TEST STARTS COMPARABLY → REPAIR THE METHOD → COMPARE RESULTS → CONNECT TO THE SCIENTIFIC MECHANISM → STATE THE LIMIT.

The Exact PSLE Science Learning Job This Guide Owns

This guide owns one PSLE Science learner job: how a Primary 5 or Primary 6 learner detects when the order of testing changes the starting state of later trials and therefore weakens a scientific comparison.

It does not replace the general guides on variables, fair tests, repeated trials, measurement interference or apparatus. Those owners remain separate. This page focuses on one particular method failure:

The previous test leaves the system in a different state, so the next condition is not being tested fairly.

The Current 2026 PSLE Science Frame

For examination from 2026, Standard PSLE Science assesses the 2023 Primary Science syllabus. The official assessment objectives include applying scientific knowledge and carrying out scientific inquiry through interpreting and analysing information, evaluating observations, information and methods, and communicating explanations and reasoning.

The 2023 Primary Science syllabus also expects learners to select and use apparatus appropriately, compare observations and data, and investigate scientifically. Evaluating whether a test truly compares conditions therefore belongs inside scientific inquiry.

Why Order Can Matter

A fair comparison tries to make the tested condition the important difference between setups. But when the same specimen is reused, the first condition may alter the specimen itself.

Possible inherited changes include:

temperature;
amount of water or another material;
wetness or dryness;
shape or extension;
surface condition;
stored energy;
position or orientation;
concentration or composition;
available resource;
damage, fatigue or permanent deformation;
time already spent undergoing the process.

If any of these persist, Condition B is not acting on exactly the same starting system as Condition A did.

State Before Condition: The Hidden Variable

A useful way to think is:

Result = what the new condition does to the state that already exists.

If the state entering Test 2 differs from the state entering Test 1, then the two results are not directly comparable unless that difference is part of the investigation.

Worked Example 1 — Heating the Same Water Twice

Original practice situation: A learner wants to compare two methods of warming equal amounts of water. The learner heats one cup using Method A for five minutes, then immediately uses Method B on the same water for another five minutes.

Problem: Method B begins with warmer water because Method A already changed the starting temperature.

The second temperature change cannot be compared fairly with the first unless the water is returned to the same starting temperature and other relevant conditions are restored.

Possible repair:

use separate equal amounts of water starting at the same temperature; or
allow the same setup to return reliably to the same starting condition before testing the next method.

The best choice depends on the actual setup. There is no universal rule that fresh material is always required.

Worked Example 2 — Wet Cloths Tested One After Another

A learner wants to compare two air conditions using the same wet cloth. The cloth is exposed to Condition P for 20 minutes and loses water. It is then moved to Condition Q for another 20 minutes.

The second test begins with a drier cloth and a smaller amount of water than the first.

If the question is “Which condition causes more water loss from the same starting wetness over 20 minutes?”, the design is not fair.

A better design could use two comparable cloths prepared with the same starting amount of water, or restore the same cloth to a defined starting wetness before each test if that can be done consistently.

Worked Example 3 — Stretching the Same Material

An object is loaded under Condition A and stretches. Then a greater load is applied under Condition B. The learner treats the two tests as though both began from the original shape.

That may be unsafe scientifically. If the object returns fully to its starting shape, the carryover may be small. If it remains stretched, damaged or altered, the second result includes the effect of the first test.

Before comparing, inspect whether the property being measured is reversible under the tested conditions.

Worked Example 4 — Cooling After Heating

A learner heats an object, records a result, then immediately starts a “cooling” condition using the same object. Is that automatically a carryover problem?

Not necessarily. If the scientific question deliberately asks what happens when a warmed object is then cooled, the changed state is part of the intended sequence.

Carryover is a problem only when the inherited state is unintended for the comparison being claimed.

This distinction prevents the learner from labelling every sequence as “unfair”.

Worked Example 5 — Reusing the Same Water Sample

Suppose an investigation tests how two conditions affect evaporation from water. The same sample is used for Condition A first and then Condition B, with no replacement.

After A, less water remains. The amount and perhaps temperature may have changed. Those inherited differences can affect the next observation.

If the goal is to compare the conditions from equal starting states, use equal fresh samples or restore the sample to the same measured starting state before each condition.

Worked Example 6 — Same Surface, Changed Surface

A learner rubs or treats a material in the first test and then tests another property on that same surface. If the first procedure makes the surface smoother, rougher, wetter, drier or otherwise altered, the second test inherits that change.

The learner should ask whether the first procedure changed a condition that matters to the second mechanism.

Worked Example 7 — When Reuse Is Fine

A non-destructive measurement is made, and the object demonstrably returns to the same starting state before the next condition. Reusing it may be perfectly reasonable.

For example, if a simple observation does not alter the object or its relevant conditions, reuse by itself is not an error.

Do not memorise “same object = unfair”. Diagnose whether the earlier test changes what matters.

Three Questions That Expose Carryover

What was the starting state before Test 1?
What did Test 1 change?
Was that change still present when Test 2 began?

If the answer to Question 3 is yes, ask whether the carryover is intended by the scientific question. If not, the comparison needs repair.

Carryover Versus Measurement Interference

These are related but distinct.

Problem	What changes the system?	Example
Measurement interference	The act of measuring or observing	Opening a container repeatedly changes its conditions.
Carryover effect	An earlier test or condition	Heating in Test 1 changes the starting temperature of Test 2.

An investigation can have either problem, both problems or neither.

Carryover Versus Repeated Trials

A repeated trial means running the planned test again. If the system is not reset, the second “trial” may not really repeat the same starting conditions.

So before calling something a repeat, ask whether the initial state has been restored.

Carryover Versus Repeated Measurements Over Time

Repeated measurements over time deliberately follow the same changing system. The fact that later readings depend on earlier history is not a flaw—it is the process being observed.

The problem appears only if the learner tries to treat those later readings as independent tests beginning from identical starting states.

How to Repair the Investigation

The correct repair depends on the mechanism.

Reset: return the same setup to the same measured starting state.
Wait: allow a reversible change to return to baseline, if that is reliable and appropriate.
Replace: use comparable fresh specimens or separate setups.
Prepare in parallel: create several setups with matched starting conditions.
Record the starting state: verify rather than assume that it is the same.
Change the research question: if sequence itself is the intended object, treat order as part of the condition rather than pretending it is absent.

Why “Use a New Specimen Every Time” Is Not a Universal Rule

Fresh specimens can solve some carryover problems, but they can introduce natural variation between specimens. A new leaf, seed, cloth or object may not be perfectly identical to the previous one.

The learner therefore needs to balance two evidence problems:

reusing the same specimen may carry an altered state forward;
using different specimens may introduce specimen-to-specimen variation.

A strong design uses comparable specimens, consistent preparation and enough repetition where appropriate. There is no one-line method phrase that solves every investigation.

Order Can Be the Variable on Purpose

Sometimes the scientific question is about sequence itself.

For example: “Does doing A before B produce a different result from doing B before A?”

Then order is not a hidden error. It is the changed condition being investigated. The method should deliberately compare different orders while keeping other relevant conditions controlled.

This distinction is central:

Uncontrolled order is a confound. Controlled order can be the experiment.

The Earliest-Weak-Link Diagnostic

Failure signature	Earliest weak link	Repair
“I used the same object, so the test is fair.”	Identity confused with same starting state.	Measure whether the relevant state was restored.
“I repeated the experiment immediately.”	Repeat assumed without reset.	Check starting conditions before each run.
“The second condition caused a smaller change.”	Inherited state ignored.	Ask what Test 1 already changed.
“Use a different specimen” is written for every method problem.	Generic repair used without mechanism.	Identify exactly what carries over first.
“Any sequence is unfair.”	Intended process history confused with method error.	Check the scientific question.
“The final result belongs only to the last condition.”	History of the system ignored.	Track state across the sequence.

Misconception Repair — “Same Specimen Means Better Control”

Using the same specimen controls individual differences only if the specimen’s relevant state does not change in a way that contaminates later tests.

Misconception Repair — “Fresh Specimen Means Perfect Fairness”

Fresh specimens can vary. They must be comparable and prepared consistently.

Misconception Repair — “Reset Means It Looks the Same”

Visual appearance is not enough. The relevant starting quantity or condition—such as temperature, mass, wetness, position or extension—must be restored if it matters to the mechanism.

Misconception Repair — “Later Means Caused by the Later Condition”

A later result may contain effects inherited from earlier conditions. Time order alone does not assign cause.

Question-Reading and Inquiry Protocol

What is the scientific question?
Is the same specimen or setup reused?
What is done first, second and later?
What does each step change physically?
Which of those changes persist?
Does the next test begin from the same relevant starting state?
If not, is that history intentional?
If not intentional, choose the smallest justified repair.
Check that the repair does not introduce a new major difference.
State only the conclusion the repaired or original method can support.

A State Ledger for Multi-Step Investigations

Stage	Starting state	Condition applied	State after	Does it carry forward?
Test 1	Known baseline	A	Changed state	Check
Test 2	What is actually present now?	B	New state	Check

This small table makes hidden history visible.

Practice Sequence

Obvious carryover: heating, wetting or removing material.
Reversible change: decide whether reset is possible.
Permanent or uncertain change: decide whether fresh comparable specimens are needed.
Repeated measurements: distinguish process tracking from repeated independent trials.
Intended sequence: identify when history is the scientific question.
Competing method repairs: compare reset versus replacement.
Unfamiliar context: diagnose the state history without relying on topic memory.
Delayed return: revisit after several days.

Unfamiliar Transfer Challenge

A mystery material begins in State 5. Condition P changes it to State 8. The same material is then exposed to Condition Q and ends at State 10.

A learner concludes: “Q changes the material by 5 because it ended at 10 and the original state was 5.”

What is wrong?

Q did not begin from State 5. It began from State 8. The history matters.

Now ask whether P’s effect is reversible, whether a fresh sample is available, and what starting state is required for a fair comparison of P and Q.

The material is fictional, but the carryover reasoning transfers.

Delayed Independent Return Test

Three to five days later, take a new investigation with three tests performed in sequence. Without notes:

write the starting state before each test;
identify what the previous test changed;
mark any inherited condition;
decide whether the inheritance is intended;
propose one justified method repair if it is not;
explain how the repair strengthens the comparison;
state one remaining limitation.

The Answer-Checking Receipt

Did I identify the test order?
Did I track the system state after every step?
Did I assume “same object” means “same state”?
Did I check whether the change is reversible?
Did I distinguish a repeated trial from repeated measurements over time?
Did I decide whether order is intended or uncontrolled?
Did my repair restore a comparable starting state?
Did the repair introduce specimen variation?
Did I connect the method problem to the scientific mechanism?
Did I keep the conclusion within the evidence?

Parent and Tutor Teaching Guide

When a child sees the same object reused, ask:

“Is it the same object in the same scientific state?”

This wording is powerful because it separates identity from condition.

Use household-scale thought experiments rather than elaborate practical work: a warm cup, a wet cloth, a stretched band, a partially emptied container. Ask what each first test changes and whether the next test inherits it.

Then include counterexamples where reuse is fine. The child should not learn a mechanical rule that “same specimen is bad”. The learning target is to diagnose whether the first test changes a condition relevant to the second.

Finally, reverse the task. Give a sequence where order is deliberately the variable. Ask why the same history that would be a confound in one investigation becomes the object of study in another.

Useful Internal Routes

Authoritative and Research References

Singapore Examinations and Assessment Board — PSLE Formats Examined in 2026.
Singapore Examinations and Assessment Board — PSLE Science syllabus, for examination from 2026.
Singapore Ministry of Education — Science Teaching and Learning Syllabus, Primary, 2023.
Education Endowment Foundation — systematic review of approaches to primary science teaching. Used as broader science-education evidence, not PSLE marking policy.

Evidence and Model Limits

Real experimental design can deal with order effects using randomisation, counterbalancing, washout periods and statistical methods. Primary Science does not require that technical machinery here.

The durable Primary-level principle is simpler: when tests happen in sequence, track what state the system carries from one test into the next. A fair comparison needs comparable starting conditions unless the sequence itself is what you intend to investigate.

The Quiet Ending

An experiment has a memory whenever the first test changes what the second test receives.

Do not look only at the condition being applied now. Look at the state that arrived from before. Reset it, replace it, measure it—or make the history part of the question.

That is how the order of testing stops being invisible.

Small Group Tutorials