How to Tell Observation, Inference, Prediction and Explanation Apart in PSLE Science

Wait, What? Four Correct Science Sentences Can Be Doing Four Different Jobs

Read these four sentences:

The water level is lower after 20 minutes.
Some of the liquid water has changed into water vapour.
If the surrounding air moves faster, the water level may fall more quickly over the same time.
The water level falls because liquid water at the surface changes into water vapour and leaves the container.

They may all be scientifically reasonable in the right situation, but they are not doing the same job. The first reports what was observed or measured. The second is an inference from evidence. The third is a prediction about what may happen under a changed condition. The fourth is an explanation: it connects evidence to a scientific mechanism.

Strong PSLE Science begins when you know not only what you are saying, but what kind of scientific claim you are making.

Quick Answer

Use this four-part distinction:

Observation: what you can directly see, measure, count, read or obtain from the information given.
Inference: what you reasonably conclude from observations using scientific knowledge.
Prediction: what you expect to happen under a stated condition, usually before the outcome is observed.
Explanation: why or how something happens, using a scientific mechanism linked to the conditions and evidence.

The safest reasoning order is: OBSERVE / READ → IDENTIFY EVIDENCE → INFER CAREFULLY → SELECT THE RELEVANT CONCEPT → EXPLAIN THE MECHANISM → PREDICT ONLY WHEN THE CONDITION JUSTIFIES IT → CHECK AGAINST THE EVIDENCE.

Owned PSLE Science Learning Job

This guide owns one specific learner job: helping a Primary 5/6 student recognise and use four different scientific moves without mixing them together. It does not replace the existing concept pages on evaporation, circuits, plants, forces, life cycles or other scientific topics. Those pages own the science concepts. This page teaches how the learner uses evidence and scientific ideas while answering PSLE Science questions.

The Current Official PSLE Science Frame

For examination from 2026, SEAB states that the PSLE Science paper assesses attainment in the 2023 Primary Science syllabus. The assessment objectives include knowledge with understanding and the application of scientific knowledge and scientific inquiry. Scientific inquiry includes making predictions and formulating hypotheses, interpreting and analysing information, evaluating observations, information and methods, and communicating explanations and reasoning.

MOE’s 2023 Primary Science syllabus organises scientific ideas through the connected themes Diversity, Cycles, Systems, Energy and Interactions. These are not meant to function as isolated boxes. That matters here because observation, inference, prediction and explanation are reasoning tools that can be used across all five themes.

The Four Jobs in One Table

Scientific move	Main question	Evidence status	Typical danger
Observation	What was actually seen, measured or given?	Directly available	Adding an interpretation and calling it an observation
Inference	What does the evidence suggest?	Reasoned from evidence	Treating one possible inference as certain
Prediction	What is likely to happen if this condition applies?	Not yet observed in that stated case	Predicting without connecting to a scientific reason
Explanation	Why or how did this happen?	Uses evidence plus mechanism	Repeating the observation instead of explaining it

1. Observation: Start With What the Question Actually Gives You

An observation is not limited to what your eyes see. In PSLE Science, an observation may be a measured temperature, a recorded mass, a water level, a number of organisms, a graph trend, a colour change, a bulb brightness shown in a diagram, a time taken, or another directly stated result.

Examples of observations include:

The temperature of liquid P decreased from 60°C to 45°C.
Plant A had more leaves after two weeks than Plant B.
The bulb in Circuit X was brighter than the bulb in Circuit Y.
Droplets were seen on the outside of a cold container.
The amount of water remaining in Container Q was less after one hour.

These statements do not yet explain why anything happened. They are the evidence floor.

The Observation Test

Ask: Could I point to the diagram, table, graph, reading or description and show exactly where this statement comes from? If yes, it is probably an observation. If the statement adds an unseen process, cause, category or hidden condition, it has moved beyond observation.

2. Inference: Move Beyond the Evidence, But Stay Anchored to It

An inference is a reasoned interpretation of evidence. It is not directly observed. It is built from an observation plus relevant scientific knowledge.

Suppose droplets form on the outside of a cold metal cup. “Droplets are present outside the cup” is an observation. “Water vapour in the surrounding air condensed on the cold surface” is an inference supported by scientific knowledge about condensation.

Notice what changed: the second sentence introduces a process that was not directly seen. It explains the evidence through a scientific model.

Inferences Can Be Stronger or Weaker

If several possible causes could produce the same observation, do not treat the first idea you remember as proven. A dim bulb could result from different circuit conditions. A plant with fewer leaves could have experienced different environmental conditions, damage or another cause. The question must provide enough evidence to discriminate.

A strong learner keeps this sentence available: “This evidence supports the inference, but does it rule out the alternatives?”

3. Prediction: Carry a Scientific Relationship Into a Not-Yet-Seen Case

A prediction states what you expect to happen under a given condition. It is not simply a guess. It should be connected to a scientific relationship or mechanism.

Suppose two identical wet cloths are placed under different airflow conditions. If the question asks what may happen when airflow increases while relevant conditions remain comparable, a prediction might be that the cloth in faster-moving air will dry sooner. The prediction should then be justified using the relevant process, not merely by saying “because wind makes it dry”.

A prediction is often about the future, but the more important distinction is logical: it concerns an outcome not yet observed in that exact stated case.

Prediction Is Not the Same as Hypothesis

In school science, a hypothesis usually proposes a testable relationship or explanation that an investigation can examine. A prediction states the expected outcome if the hypothesis or scientific relationship applies. The two can be related, but they are not identical.

4. Explanation: Connect Condition → Mechanism → Outcome

An explanation answers why or how. It should not merely restate the observation using different words.

Weak explanation: “Plant A grew better because it had more leaves.”

Problem: the sentence largely repeats an outcome and does not identify the relevant process or condition.

Stronger structure: given condition → relevant scientific process → resulting change → observed outcome.

Exactly which scientific process belongs in the chain depends on the question. Do not force a memorised mechanism into every situation.

A Complete Worked Example: Cold Cup, Warm Room

Imagine an original investigation in which a cold metal cup is placed in a warm room. After several minutes, droplets appear on the outside surface.

Observation: droplets appeared on the outside surface of the cup.
Inference: the droplets came from water vapour in the surrounding air rather than from the cup producing water.
Prediction: under otherwise comparable conditions, a colder surface may lead to more rapid formation of visible droplets initially.
Explanation: air near the cold surface cools; water vapour can condense into liquid water on the surface, producing the observed droplets.

Each sentence has a different job. The learner’s task is to match the job to the question.

A Second Worked Example: A Plant Investigation

Two similar plants are kept under different light conditions while relevant conditions are controlled. After a stated period, Plant P has accumulated more dry mass than Plant Q.

Observation: Plant P has a greater measured dry mass at the end.
Inference: under the stated conditions, Plant P produced or retained more material contributing to dry mass.
Prediction: if the same relevant relationship continues under comparable conditions, the plant receiving the more favourable light condition may continue to show a higher rate of food production until another factor becomes limiting.
Explanation: if light was the relevant limiting condition, the changed light condition affected photosynthesis; this changed the amount of food produced, which can contribute to differences in accumulated plant material.

The caution matters: do not say “more light always means more growth”. Scientific explanations are conditional.

A Third Worked Example: An Electrical System

Two circuits use identical bulbs, but the bulbs show different brightness.

Observation: Bulb X is brighter than Bulb Y.
Inference: the electrical conditions at the two bulbs differ.
Prediction: if the arrangement is changed so that the conditions become more favourable for the bulb, brightness may increase.
Explanation: the exact explanation must use the circuit arrangement and concepts provided by the syllabus; brightness cannot be explained from the word “electricity” alone.

The important skill is not to jump from observation directly to one cause without checking the circuit evidence.

Observation vs Inference: The Most Important Boundary

Many science errors begin here. Learners write an inference as though it were directly observed.

Example: “The plant lost water through its leaves.” If the question only shows that the mass of the plant-and-container system decreased, that sentence may be an inference. The decrease is observed; the pathway of water loss requires scientific interpretation and perhaps additional evidence.

Use the phrase evidence first, hidden process second.

Prediction vs Explanation: Same Science, Different Direction

Prediction moves from known relationship and condition toward an expected outcome. Explanation usually moves from an outcome and condition toward the mechanism that produced it.

Prediction: “If the exposed surface area of the water is larger, more water may evaporate over the same time under comparable conditions.”

Explanation: “More water evaporated because a larger exposed surface provided more water at the surface from which evaporation could occur during the same time.”

These sentences may use related scientific knowledge, but their jobs differ.

Why Learners Mix the Four Jobs

They memorise answer phrases without identifying the question type.
They treat every scientifically true sentence as an answer.
They use familiar keywords before reading the evidence.
They skip the difference between visible evidence and invisible mechanism.
They assume a prediction is correct merely because it sounds plausible.
They repeat an outcome when the question asks for a cause.

The Earliest Weak-Link Diagnosis

If your answer is wrong, diagnose the first break:

Evidence break: Did I misread what was actually observed?
Category break: Did I confuse observation with inference?
Concept break: Did I select the wrong science idea?
Mechanism break: Did I name the concept but fail to explain how it produced the outcome?
Condition break: Did I ignore the factor that changed?
Claim-strength break: Did I say more than the evidence supports?

Misconception Repair: “If It Is True, It Must Be an Observation”

No. “Water vapour is present in air” can be a scientific fact, but if a particular question only shows droplets forming on a cold surface, you did not directly observe individual water-vapour particles. You use scientific knowledge to infer the process.

Truth and observation are different categories.

Misconception Repair: “An Inference Is Just a Guess”

A good inference is constrained by evidence and scientific knowledge. A guess may be unsupported. The goal is not to avoid inference; science depends on inference. The goal is to make the evidence–inference connection visible.

Misconception Repair: “Prediction Means Future Tense”

Grammar alone does not define the scientific job. A sentence in future tense can still be a vague guess. A scientific prediction states an expected outcome under a condition and is supported by a relationship or mechanism.

Misconception Repair: “Explanation Means Write More”

Length is not mechanism. A short answer can explain well if it connects the relevant condition, process and outcome. A long answer can still fail if it only repeats facts.

A Student Protocol for Any PSLE Science Question

Underline the evidence that is directly given.
Write the observation in neutral language.
Ask what scientific idea could account for it.
Keep more than one inference alive if the evidence is incomplete.
Check the command word and reader job.
If predicting, state the changed condition and expected outcome.
If explaining, connect condition → mechanism → outcome.
Return to the evidence and ask whether your claim fits all of it.

Practice Set: Sort the Scientific Move

For each sentence, decide whether it is mainly observation, inference, prediction or explanation.

The mass of the wet cloth decreased by 8 g in 30 minutes.
Water from the cloth entered the surrounding air as water vapour.
If the cloth is spread out, it may dry faster under comparable conditions.
The spread-out cloth dries faster because more of its wet surface is exposed to the surrounding air.
The number of small organisms in Area A was lower in Week 4 than Week 1.
A changed environmental condition may have reduced survival or reproduction.

Suggested classifications: 1 observation; 2 inference; 3 prediction; 4 explanation; 5 observation; 6 inference. In real questions, the exact category can depend on what information was directly provided. That is part of the lesson.

Transfer Check: When the Surface Example Changes

You are shown an unfamiliar sealed device. A gauge reading falls, a small indicator light turns on and the casing becomes warmer.

You do not know the device name. You can still separate the jobs:

Gauge reading falls: observation.
Some internal process changed the measured quantity: inference.
If the process continues, the reading may fall further: prediction.
A valid explanation would need a mechanism supported by information about how the device works; you should not invent one.

This is genuine transfer because the reasoning survives without a familiar chapter picture.

Delayed Independent Return Test

Two days later, take a new Science question. Before solving it, write four small headings: Observation, Inference, Prediction, Explanation. Fill only the headings that the question actually requires. Then remove the headings and write the final answer naturally.

If you can still distinguish the four jobs without a teacher prompting you, the skill is becoming independent.

How to Check Your Answer

Did I clearly identify what was actually observed?
Did I add an unseen cause and accidentally call it an observation?
Is my inference supported by the evidence?
If I predicted, did I state the condition?
If I explained, did I include a mechanism?
Did I repeat the result instead of explaining it?
Did I claim certainty when the evidence only supports a possibility?
Does my final sentence answer the actual question asked?

Parent and Tutor Teaching Guide

Instead of immediately telling a learner that an answer is wrong, ask four short questions: “What did you directly observe?”, “What are you inferring?”, “What would you predict if the condition changed?” and “What mechanism explains the result?”

If the learner confuses categories, do not repair the whole topic at once. Use one tiny example and ask them to label each sentence. Once the evidence–inference boundary is secure, move to prediction and explanation.

The goal is not to make children speak like scientists performing a vocabulary ritual. The goal is to make the structure of scientific thinking visible enough that the learner can use it alone later.

Useful Routes From Here

Authoritative Reference Basis

Singapore Examinations and Assessment Board — PSLE Science syllabus, for examination from 2026.
Singapore Ministry of Education — Science Teaching & Learning Syllabus, Primary, 2023.
Education Endowment Foundation — A systematic review of approaches to primary science teaching, 2023.
Science-education research on observation versus inference and mechanistic reasoning, used to support the learning design rather than to invent PSLE marking rules.

The Quiet Ending

Science becomes much easier to control when every sentence has a job. Observe what the world gives you. Infer carefully. Predict only from a justified relationship. Explain with a mechanism. Then return to the evidence.

That four-part discipline is small enough to remember, but strong enough to travel across Diversity, Cycles, Systems, Energy and Interactions—and across questions you have never seen before.

Small Group Tutorials