How to Tell Whether a PSLE Science Investigation Is Comparing Two Cases or Testing a Relationship Across Several Conditions

Wait, What? Two Set-Ups Can Show a Difference Without Showing a Trend

A PSLE Science investigation can place two set-ups side by side and ask what is different. Another investigation can test four or five conditions and ask how an outcome changes across the series. These may look similar because both contain a changed condition and a measured result. But they support different kinds of reasoning.

From two cases, you can make a comparison. From several suitably ordered conditions, you may be able to identify a relationship or pattern. The mistake is to turn one pair into a whole trend, or to reduce a multi-condition pattern to only one convenient pair.

TWO CASES ANSWER “HOW ARE THESE DIFFERENT?” SEVERAL CONDITIONS CAN ANSWER “HOW DOES THE OUTCOME CHANGE AS THE CONDITION CHANGES?”

Quick Answer

Use this route before describing the result:

READ THE INVESTIGATION QUESTION → IDENTIFY THE CHANGED CONDITION → IDENTIFY THE MEASURED OUTCOME → COUNT THE DISTINCT TEST CONDITIONS → ASK WHETHER THEY FORM A MEANINGFUL ORDER → CHOOSE COMPARISON OR RELATIONSHIP REASONING → CHECK THE WHOLE EVIDENCE SET → STATE ONLY WHAT THE DATA SUPPORT.

If there are only two conditions, do not invent what would happen between or beyond them. If there are several conditions, do not ignore most of the data and declare the relationship from one pair unless the question specifically asks for that pair.

The Exact PSLE Science Learning Job This Guide Owns

This guide owns one learner job: distinguishing an investigation that compares two cases from an investigation that tests a relationship across several conditions, then matching the conclusion to the structure of the evidence.

It does not own generic variables, graph drawing, fair tests, or the scientific concept being investigated. Existing guides remain the owners for those jobs. This page sits one level above them: it helps the learner recognise what kind of evidence structure is actually present.

Why This Matters for Current PSLE Science

For examination from 2026, SEAB states that PSLE Science assesses attainment in the 2023 Primary Science syllabus. The assessment objectives include interpreting and analysing information, evaluating observations, information and methods, and communicating explanations and reasoning. The official Primary Science syllabus also develops scientific inquiry through observing, comparing, analysing relationships and using evidence.

This means a learner should not merely read numbers. The learner must understand what the arrangement of those numbers allows the investigation to say.

Case Comparison and Relationship Testing Are Different Jobs

Evidence structure	Main learner job	Safe conclusion type	Common overclaim
Two cases	Compare A with B	A has a greater/lower/different outcome than B under the tested conditions	“As the condition increases, the outcome always…”
Several ordered conditions	Look for a relationship across the series	Within the tested range, the outcome increases/decreases/changes in the observed pattern	Using only the first and last point and ignoring a turning point
Several unordered categories	Compare categories	State differences among named categories	Inventing a numerical trend where no ordered scale exists

Worked Example 1 — Two Distances Are Still Only Two Cases

Imagine an original investigation in which a lamp is placed 10 cm from Set-Up A and 30 cm from Set-Up B. A measured response is larger in A.

What can you say?

Under these tested conditions, the response at 10 cm is greater than the response at 30 cm.
If the method is fair and the relevant mechanism is known, you may explain why the different distance could affect the response.

What can you not yet say from the two results alone?

that every extra centimetre produces a fixed decrease;
that the relationship is a straight line;
that 20 cm must give a result exactly halfway between;
that the same pattern continues beyond the tested range.

Worked Example 2 — Several Conditions Can Reveal a Relationship

Now imagine the investigation uses 10 cm, 20 cm, 30 cm, 40 cm and 50 cm, with a result recorded at every condition using the same method.

The learner should first order the conditions correctly, then inspect all the results. If the measured outcome falls steadily across the tested distances, it is reasonable to describe a decreasing relationship within the tested range.

The relationship statement still needs a boundary. “Within the conditions tested” is scientifically different from “this must continue forever”.

Worked Example 3 — Several Conditions Do Not Guarantee One Simple Trend

Suppose a learner tests five conditions and the results rise, rise again, then fall. If the learner compares only the first and last condition, an important turning region disappears.

Several conditions create the opportunity to analyse a relationship. They do not guarantee that the relationship is simple. The learner must let the pattern come from the full data set.

Worked Example 4 — Categories Are Not Automatically a Number Scale

Imagine a table compares four materials: P, Q, R and S. These are four conditions, but the labels P–S do not form a meaningful scientific numerical order. The learner should compare the material categories, not write, “As the material increases from P to S…”

Before looking for a trend, ask whether the changed condition has a genuine order: time, distance, temperature, amount, number, level or another scientifically meaningful sequence.

Worked Example 5 — A Multi-Condition Investigation Can Still Contain a Pairwise Question

A table may contain six conditions, but a sub-question may ask only, “Compare Set-Ups B and D.” For that part, answer the requested pair. Later, a different sub-question may ask for the overall relationship.

The evidence set may be large while the current learner job is small. Read the command before deciding how much of the table to use.

The PSLE Science Reasoning Chain

READ GIVEN INFORMATION → IDENTIFY THE SCIENTIFIC OBJECT OR RELATIONSHIP → DISTINGUISH THE TEST CONDITIONS → IDENTIFY WHAT WAS MEASURED → DECIDE PAIRWISE COMPARISON OR MULTI-CONDITION RELATIONSHIP → SELECT THE RELEVANT CONCEPT → EXPLAIN THE MECHANISM WHERE REQUIRED → CONNECT TO THE CONDITIONS → STATE THE OUTCOME → CHECK AGAINST ALL RELEVANT EVIDENCE.

Observable Failure Signatures

What the learner does	Likely weak link
Turns two points into “the more X, the less Y”	Pairwise evidence promoted into a general trend
Uses only the first and last values from a five-condition table	Middle evidence ignored
Treats P, Q, R, S as increasing numerical conditions	Category labels mistaken for an ordered variable
Describes a trend when one middle value reverses direction	Pattern simplified beyond the data
Lists every condition when the question asks only to compare two set-ups	Current response job not identified
Explains a mechanism before correctly describing the evidence structure	Concept knowledge outran data reading

Find the Earliest Weak Link

What is the scientific question?
What exactly changes between conditions?
What outcome is measured?
How many distinct test conditions are there?
Are they ordered scientifically or merely named categories?
Does the current sub-question ask about a pair or the whole series?
Do all relevant results support the same direction?
Is there a plateau, turning point, gap or exception?
How far can the conclusion travel beyond the tested conditions?

Misconception Repair — “Two Points Make a Trend”

Two points establish a difference between two tested cases. They may be consistent with many possible relationships between the points. More suitably spaced conditions are needed to reveal the shape of a relationship.

Misconception Repair — “More Conditions Mean I Must Draw a Graph”

No. A graph is useful when the representation suits the quantities and task. A table or verbal comparison may be enough. Choose the representation for the scientific job, not because there are many rows.

Misconception Repair — “A Relationship Means a Cause”

Describing how two measured quantities vary is not yet the same as establishing why. Causal explanation needs the investigation design, controlled conditions and relevant scientific mechanism.

Practice Protocol: Pair → Series → Boundary

Take a five-condition results table.
Choose one pair and write a strictly pairwise comparison.
Now inspect the whole series and write the relationship, if one is supported.
Circle any result that prevents a simple trend statement.
Add the tested-range boundary.
Explain the relationship only if the scientific mechanism and method justify it.
Create a two-condition version of the same investigation and notice which claims are no longer supported.

Unfamiliar Transfer Challenge

Create two original investigations around the same scientific idea. Investigation A has only two test conditions. Investigation B has five suitably ordered conditions. Use invented values.

Write three statements for each:

one statement definitely supported;
one statement that might be true but is not yet established;
one overclaim.

This forces you to notice how evidence structure changes the size of the conclusion.

Delayed Independent Return

Three to five days later, use a new investigation you have not seen before. Before reading the answer choices or model answer, label it PAIR, ORDERED SERIES or CATEGORIES. Then state the strongest conclusion the evidence supports.

Answer-Checking Receipt

I identified the changed condition and measured outcome.
I counted the distinct conditions correctly.
I checked whether the conditions form a meaningful order.
I know whether this sub-question needs one comparison or the whole pattern.
I used all relevant evidence before describing a relationship.
I did not invent a trend from two cases.
I did not ignore a turning point or exception.
I kept the conclusion inside the tested range and method.

Parent and Tutor Teaching Guide

Give the learner the same scientific relationship in two forms: first as two set-ups, then as five test conditions. Ask, “What can you say now that you could not say before?” The answer should refer to the structure of the evidence, not simply to having “more numbers”.

Next, deliberately insert a turning point into the five-condition data. If the learner still writes a simple increasing or decreasing trend, return to the whole-series check.

Finally, use four named categories with no numerical order. This tests whether the learner can resist inventing a trend merely because several rows are present.

Useful Internal Routes

Authoritative References

Evidence and Boundary Note

This is an eduKate learning guide, not an official answer formula. The official documents establish the current curriculum and assessment objectives; they do not require learners to label questions “pair” or “ordered series”. Those labels are temporary practice tools for seeing what the evidence can support.

The Quiet Return

Do not make the conclusion bigger than the experiment.

Two cases can tell you how two cases differ. A well-designed series can reveal how a relationship behaves. Read the structure first, and the Science becomes much harder to overclaim.

Small Group Tutorials