How to Infer the Scientific Question From a PSLE Science Investigation Set-Up

Wait, What? Sometimes the Question Is Hidden Inside the Method

A diagram shows three identical containers. The amount of light is different in each set-up. After two days, the same outcome is measured in all three. But the scientific question is not written above the diagram.

Many learners respond by guessing the topic: “This is about plants.” That may be true, but it is not yet the investigation question.

A scientific investigation is organised around a relationship. One condition is deliberately changed. An outcome is observed or measured. Other relevant conditions are kept comparable. If you can identify those roles, you can often reconstruct the question the method was built to answer.

Quick Answer

Find what the investigator deliberately changes. Find what is measured or observed as the outcome. Check which conditions are kept comparable. Then phrase the question as a relationship without adding a conclusion that has not yet been shown.

A useful reconstruction is: WHAT IS CHANGED → WHAT IS MEASURED → WHAT IS HELD COMPARABLE → WHAT RELATIONSHIP CAN THIS METHOD TEST?

Owned PSLE Science Learning Job

This guide owns one specific learner job: inferring the scientific question being investigated when a set-up, method or results structure is shown but the question itself is unstated or indirect.

It does not replace the existing guide on planning an investigation from a scientific question. That is the forward direction: question → variables → method. This page runs the reasoning backwards: method → variable roles → scientific question.

It also does not replace the general fair-test guide. Variables remain important here, but the dominant job is reconstructing the question the investigation is capable of answering.

Why This Matters in the Current PSLE Science Frame

The 2026 PSLE Science examination assesses the 2023 Primary Science syllabus. SEAB’s assessment objectives include scientific inquiry, interpreting and analysing information, evaluating observations, information and methods, and communicating explanations and reasoning.

Those skills require more than recognising apparatus. A learner must understand what a method is doing scientifically: what relationship it is trying to reveal and what the evidence can actually support.

The Core Pattern: Change X, Observe Y

At Primary level, many investigations can be understood through a simple structure:

X: the condition deliberately changed between set-ups or trials.
Y: the outcome observed or measured.
Comparable conditions: other relevant factors controlled so the comparison is meaningful.

The inferred question often has a form such as: “How does X affect Y under these conditions?” But do not force that wording when the method is only comparing categories or testing whether a condition matters. The wording should match the evidence structure, not a memorised sentence frame.

Worked Example 1: Different Amounts of Light

Three identical plants of similar starting size are placed under different light levels. They receive the same amount of water, use the same type of soil and are observed over the same number of days. Increase in height is measured.

Changed condition: light level. Measured outcome: increase in plant height. Comparable conditions include water, soil, observation duration and suitable starting similarity.

A reasonable inferred question is: How does the amount of light affect the increase in height of these plants over the stated period?

Notice what is not included. We do not write “Does more light make plants grow faster?” before seeing the results. That wording already suggests a direction and may confuse increase in height with rate of growth.

Worked Example 2: Different Materials, Same Test

Four strips of different materials have the same dimensions. The same load is added to each strip. The amount each strip bends is measured.

Changed condition: material type. Measured outcome: amount of bending under the stated load. A useful question is: How does the type of material affect how much the strip bends under the same load?

Do not turn this into “Which material is strongest?” unless the test and definition of strength in the question genuinely support that claim. One measured property under one method should not silently become a broader material label.

Worked Example 3: Before–After Is Not Always a Changed Variable Between Set-Ups

A cup of water is measured at the start and again after 20 minutes in the same environment. The method records temperature at both times.

Here, time is part of the observation structure. The question may be about how the water’s temperature changes over time, not about comparing two independent set-ups.

This is why the learner must inspect whether rows represent times, trials, specimens or different conditions. Do not label every first column “independent variable” without understanding what one observation represents.

Worked Example 4: A Control Set-Up Reveals the Question

Set-up P contains a seed with water, air and suitable warmth. Set-up Q is identical except that water is absent. The outcome is whether germination occurs.

The deliberate difference is water availability. The observed outcome is germination. The question can be reconstructed around whether water availability affects germination under the stated conditions.

The control or reference set-up helps reveal the relationship because the two cases differ in the factor being investigated. The control is not simply “the set-up where nothing happens”.

The Seven-Step Reverse-Engineering Protocol

1. Name the scientific object. Plant? Material? Circuit? Water sample? Organism?
2. Compare the set-ups. What is deliberately different?
3. Identify the outcome. What is actually measured, counted or observed?
4. Find the common conditions. Which relevant conditions are kept comparable?
5. Check timing and location. When and where is the outcome measured?
6. Write the relationship question. Keep it neutral enough not to assume the result.
7. Test the wording against the method. Could the method genuinely answer the question you wrote?

Do Not Infer a Question the Method Cannot Answer

Suppose two plant set-ups differ in both water amount and light level. Plant P grows more. The method cannot isolate which of those two differences caused the outcome.

A learner should not reconstruct a neat single-variable question such as “How does water amount affect growth?” because the method does not provide a fair comparison for water alone.

The correct reconstruction may be broader: the set-ups compare two combinations of conditions. The evidence can show that the outcomes differ under those combinations, but not which individual condition is responsible.

Observation, Question and Conclusion Must Stay Separate

Layer	Example
Method	Different exposed surface areas; same starting volume; measure water remaining after one hour.
Question	How does exposed surface area affect the amount of water lost over one hour under the stated conditions?
Result	Set-up A lost 12 mL; Set-up B lost 5 mL.
Conclusion	Under these tested conditions, the set-up with the larger exposed surface lost more water.
Explanation	A scientific mechanism accounts for why the difference occurred.

If you write the conclusion into the question before examining the result, you blur what was tested with what was found.

What If the Investigation Uses Categories Rather Than Numbers?

The changed condition does not have to be a number. It could be material type, surface type, object type or another category.

If four materials are tested for electrical conduction using the same simple circuit, the scientific question may compare whether or how the material type relates to the observed circuit response.

Do not invent a numerical trend merely because the results are placed in a table. Categories do not automatically have a meaningful low-to-high order.

What If More Than One Outcome Is Measured?

An investigation may record both plant height and number of leaves. Those are two distinct measured outcomes.

The scientific question may explicitly include both, or the task may later ask about only one. Keep them separate. Do not merge different measurements into an invented single idea called “plant health” unless the question defines such a measure.

Failure Signatures and Earliest Weak-Link Diagnosis

Failure signature	Likely weak link	Repair
“It’s about plants.”	Topic recognised, relationship missing.	Name what changed and what was measured.
You write the result inside the question.	Question/conclusion confusion.	Use neutral relationship wording before looking at outcome direction.
You choose the easiest thing to see as the outcome.	Measured-variable confusion.	Read what is actually recorded or compared.
You infer one cause when two conditions changed.	Fair-comparison failure.	List every difference between the set-ups.
You call a repeated time point a separate set-up.	Observation-unit confusion.	Identify what one row or point represents.
You write a question the method cannot answer.	Method-question alignment failure.	Ask what claim the measurement and comparison can genuinely support.

Misconception Repair: Apparatus Does Not Define the Question

The same thermometer can be used in many investigations. The same lamp can appear in questions about light, plant growth or heating. The same beaker can hold different substances.

Do not identify the question from the most familiar piece of apparatus. Identify the relationship from what is deliberately changed and what is measured.

Misconception Repair: “What Is Being Tested?” Has Two Meanings

Students sometimes say “The thing being tested is the plant.” In everyday language, the plant is indeed the specimen. In investigation reasoning, however, the scientific relationship being tested may be how light affects growth.

Be precise about whether the question asks for the specimen, the changed condition, the measured outcome or the relationship under investigation.

Reverse Engineering a Results Table

Sometimes the set-up is not shown, but the table is enough to reveal the investigation structure.

If the first column lists “distance from lamp: 10 cm, 20 cm, 30 cm, 40 cm” and the second column lists “temperature after 5 minutes”, you can identify a deliberately varied distance and a measured temperature outcome. Then ask whether all other relevant conditions are stated or implied to be comparable before inferring a causal question.

The table gives you variable roles. It does not automatically prove the cause.

A Scratch-Question Method for Difficult Set-Ups

On scrap paper, write:

CHANGE: ______ → MEASURE: ______ → COMPARE: ______ → QUESTION: ______

Fill the first three blanks before writing the fourth. This is not an official exam template. It is a compact learner aid for stopping topic familiarity from replacing variable reasoning.

Practice Sequence

Start with a simple two-set-up fair test and infer the question.
Use a categorical changed condition such as material type.
Use a time-series investigation and distinguish time points from separate set-ups.
Use a flawed investigation with two conditions changed and explain why a single-variable question cannot be supported.
Use a results table without a diagram and infer the relationship cautiously.
Finally, compare your inferred question with the method and ask whether every part of the wording can actually be tested.

Unfamiliar Transfer Challenge

A fictional object called a “Luma tile” produces a reading on a sensor. Four identical tiles are placed at different distances from a light source. Sensor reading is recorded after the same time.

You do not need to know what a Luma tile is. You can still infer that distance from the light source is deliberately varied and sensor reading is measured. The scientific question must preserve that relationship and avoid inventing a mechanism the question has not supplied.

Delayed Independent Return Test

Several days later, give the learner three unfamiliar investigation diagrams with the question title removed. One should be a fair test, one a time-series observation and one a flawed comparison with two changed conditions.

The learner should identify what the method can test, what it cannot isolate and what additional evidence would be needed. That is stronger than simply labelling variables from a memorised diagram.

Investigation-Question Checking Receipt

What exact object or system is being studied?
What condition is deliberately different?
What outcome is actually measured or observed?
Which relevant conditions are comparable?
Are the observations separate set-ups, repeated trials or time points?
Have I written the question without assuming the result?
Can this method genuinely answer the question I wrote?
Have I avoided claiming a cause if the comparison is confounded?

Parent and Tutor Teaching Guide

When a child says “I don’t know what they are testing”, do not immediately name the topic. Ask two questions: “What did they deliberately change?” and “What did they measure?”

If the child can answer those, ask them to write a neutral relationship question. Then test it against the method: would the data actually answer that question?

Use deliberately flawed methods occasionally. They teach an important boundary: not every diagram deserves a clean causal question. Scientific inquiry includes evaluating methods, not merely decoding them.

Useful Internal Routes

Authoritative References and Evidence Boundary

The reverse-engineering protocol is a learner scaffold, not an official SEAB answer format. Some authentic scientific investigations are observational, multivariable or exploratory and do not fit a simple one-variable fair-test structure. This guide stays within common Primary Science reasoning while explicitly preserving those limits.

The Quiet Return

An investigation diagram is not a collection of objects. It is a question made physical.

Find the deliberate change. Find the measured response. Protect the comparison. Then ask what relationship the method is truly capable of revealing.

Once you can do that, even an unfamiliar apparatus begins to speak the language of scientific inquiry.

Small Group Tutorials