Primary 3 Science Learning Guide | Classification with Yes/No Questions

Classification becomes much stronger when a pupil can ask a sequence of clear yes/no questions instead of guessing a group from appearance.

Primary 3 Science asks pupils to compare and classify living things and materials using shared observable characteristics. A simple yes/no classification tree is one useful way to make the reasoning visible. Each question divides the examples according to one stated characteristic. The next question divides one of those groups again.

The goal is not to teach formal taxonomy. The goal is to help pupils organise evidence, apply one rule consistently and understand that a good classification system can be followed by someone else.

Wait, What? “I Know It When I See It” Is Not a Classification Rule

A pupil may correctly identify a bird because the picture looks familiar. But if an unfamiliar bird appears, the learner may fail. A classification rule is stronger because it can be stated and checked.

For example:

Does the animal have feathers? Yes → bird group. No → ask another question.

The question uses an observable characteristic rather than familiarity.

What Is a Yes/No Classification Tree?

A yes/no classification tree is a simple branching sequence. At each step, ask one question that can be answered yes or no. The answer determines which branch to follow.

Example with four objects:

Does the object absorb water?
If yes, place it in one branch.
If no, place it in another branch.
Within one branch, ask: Can it be seen through?

This is classification through a sequence of consistent decisions.

Rule 1: Every Question Must Use One Clear Characteristic

“Is it big and furry and found in water?” combines several ideas and creates confusion. A better question uses one characteristic at a time.

Does it have feathers?
Does it produce flowers?
Does it absorb water?
Is it attracted to a magnet?

Each question creates one clear split.

Rule 2: The Characteristic Must Be Relevant

Colour can be observable but may be useless for the scientific job. If the task is to distinguish flowering and non-flowering plants, “Is it green?” does not separate the intended groups because both can be green.

A relevant question is “Does the plant produce flowers?”

Rule 3: The Same Rule Applies to Every Example at That Branch

Do not classify one animal by feathers, another by habitat and another by size while pretending they were grouped by one rule. At each branch, every example must answer the same question.

This consistency makes the classification reproducible.

Rule 4: Do Not Ask a Question the Evidence Cannot Answer

If the diagram does not show whether the plant produces flowers, a flowering/non-flowering split may not be possible. The correct move is not to guess. Use another available characteristic or state that more information is needed.

Classification must stay inside the evidence.

Worked Tree 1: Plants and Animals

Suppose a set contains a fern, hibiscus, pigeon and cat.

Is it a plant? Yes → fern, hibiscus. No → pigeon, cat.
For plants: Does it produce flowers? Yes → hibiscus. No → fern.
For animals: Does it have feathers? Yes → pigeon. No → cat.

The tree uses different questions at different branches, but each branch applies one clear rule.

Worked Tree 2: Animal Groups

Suppose a set contains a pigeon, butterfly, frog and cat.

Does it have feathers? Yes → pigeon.
If no: Does it have six legs? Yes → butterfly.
If no: Is it identified by the question as an amphibian? Yes → frog.
Remaining example → cat, a mammal in the given set.

In real assessment work, the question must provide enough information for every branch. The learner should not invent hidden traits.

Worked Tree 3: Materials by Properties

Four samples have test results for waterproofness and transparency.

Does it absorb water? No → waterproof branch. Yes → absorbing branch.
Within waterproof branch: Can objects be seen through it? Yes → transparent waterproof. No → opaque waterproof.
Within absorbing branch: apply transparency if needed.

The classification is built from test evidence rather than material names.

Worked Tree 4: Magnetic and Non-Magnetic Materials

Question: Is the object attracted to a known magnet?

Yes → magnetic material branch.
No → non-magnetic material branch.

This is a simple two-way classification. A second question could then divide the magnetic branch by object type or another stated characteristic if the task requires it.

The First Question Matters

Some first questions are more useful than others. A good first question creates a meaningful split. If all four animals have legs, asking “Does it have legs?” creates one group and leaves the other branch empty.

A more useful question may be “Does it have feathers?” if that separates one or more examples from the rest.

A Good Tree Can Have More Than One Valid Form

There may be several valid sequences of yes/no questions that correctly classify the same set. One pupil may begin with “plant or animal”. Another may begin with “produces flowers?” if the set and evidence allow it.

The important conditions are consistency, relevance and successful separation of the examples.

Classification Is Not Guessing the Teacher’s Tree

If two trees use different valid observable characteristics and both classify the examples correctly, both can be scientifically reasonable. The pupil should be able to defend the rule used.

This is a strong lesson in scientific representation: different models can organise the same information while preserving the evidence.

Do Not Use Opinion Questions

“Is it cute?” “Is it scary?” and “Would you like to touch it?” are not suitable scientific classification rules. Different observers may answer differently and the characteristics do not support the syllabus grouping.

Use observable or provided scientific characteristics instead.

Do Not Use Circular Questions

“Is it a bird?” can be useful only if the group is already known. If the purpose is to determine whether the animal is a bird from evidence, ask about the relevant characteristic instead, such as feathers.

A classification question should reveal the group rather than merely rename it.

When a Branch Contains One Example

Once a branch contains one remaining example, no further question may be needed. Extra questions do not improve the classification if the item is already uniquely identified within the set.

Efficient classification asks only as many questions as necessary.

When a Branch Still Contains Several Examples

Ask another yes/no question that separates the remaining members. The second question should use a different relevant characteristic from the first.

This is how a branching tree becomes multi-step without becoming complicated.

Unknown Examples Test the Tree

A strong classification tree should work on an unfamiliar example if the necessary characteristics are given. If a new animal has feathers, the same “has feathers?” branch should classify it with birds even if the species name is unfamiliar.

This is concept transfer through classification structure.

Classification Trees and Evidence Limits

A tree can only be as good as the available evidence. If the question requires “Does it reproduce?” but no reproduction information is provided, the tree cannot legitimately answer that branch.

Good Science sometimes stops and says: more information is needed.

The Build-Test-Repair Routine

List the examples.
Choose one relevant yes/no characteristic.
Split the set.
Repeat on branches with more than one example.
Test every example through the tree.
Try one unfamiliar example.
Repair any question that is vague or unsupported.

Common Classification-Tree Errors

Asking two or three characteristics in one question.
Using colour or size when the intended classification is different.
Changing the rule within one branch.
Using opinion rather than scientific evidence.
Asking questions the provided data cannot answer.
Using a group name as the question instead of a characteristic.
Continuing to ask questions after an example is already uniquely identified.
Assuming only one valid tree can exist.

How to Practise

Start with four familiar pictures. Ask the pupil to build the shortest clear yes/no tree. Then change one example and test whether the same questions still work. Later use material test results instead of pictures so the child learns that classification can be based on measured evidence as well as visible features.

A Mini Diagnostic

Explain why “I recognise it” is weaker than a stated rule.
Write one useful yes/no question for flowering plants.
Write one for magnetic materials.
Build a two-step tree for four simple examples.
Explain why opinion questions are unsuitable.
Give one case where more information is needed.
Explain why two different trees can both be valid.

Primary 3 Science Checkpoint

I can classify with a stated rule.
I ask one characteristic at a time.
I choose relevant observable characteristics.
I apply the same question to every example in a branch.
I can build a simple yes/no tree.
I know when a branch needs another question.
I know when no more questions are needed.
I can test my tree with an unfamiliar example.
I do not use opinion or unsupported information.
I understand that more than one valid classification tree can exist.

Continue the Primary 3 Science Learning Guide

Return to the Primary 3 Science Learning Hub.

Source and Syllabus Alignment

This guide supports the comparison and classification practices in the Singapore Ministry of Education Science Teaching & Learning Syllabus: Primary Three to Six.

Small Group Tutorials