Primary 4 Science Learning Guide | Physical Model-Making, Testing and Revision

A model can be made from cardboard, string, cups, clay, labels, arrows, blocks or almost nothing at all.

What makes it scientific is not the material.

What makes it scientific is whether the model represents a relationship clearly enough that the learner can use it, test it, explain it and notice where it stops being useful.

A physical model is not a miniature reality. It is a deliberately simplified object built to make one scientific relationship easier to see, manipulate or explain.

This guide belongs to the Primary 4 Science Learning Hub. Its instructional job is specific: tangible model-making, testing and revision. The existing Scientific Models and Their Limits guide owns the general idea of models and limits; the Scientific Diagrams guide owns two-dimensional visual representation. This page asks what happens when the learner must actually build a model with objects and then interrogate it.

Why Physical Model-Making Belongs in Primary Science

The current Singapore Primary Science syllabus recognises model-making as one of the valid modes through which pupils can demonstrate learning. That matters because some understanding is easier to reveal by building and manipulating a representation than by writing one sentence.

Official reference: MOE Science Teaching & Learning Syllabus — Primary.

Quick Answer: The Model-Making Loop

PURPOSE → SELECT PARTS → ASSIGN MEANING → BUILD → MANIPULATE → TEST → FIND LIMIT → REVISE → EXPLAIN

This is an eduKate teaching routine, not an official MOE marking formula.

Wait, What? A Beautiful Model Can Still Be Scientifically Weak

A digestive-system model can be colourful, symmetrical and carefully decorated—and still be poor if food cannot be traced from mouth to gullet to stomach to small intestine to large intestine.

A shadow model can have a perfect cardboard torch—and still be poor if the object, source and screen positions cannot be changed meaningfully.

Model quality begins with scientific purpose, not decoration.

Step 1 | State the Model’s Job

Before building, complete:

“This model is designed to show…”

Examples:

• how food moves through the digestive route;
• how source, object and screen position affect a shadow;
• how a liquid can keep volume while changing shape;
• how root damage can affect water uptake and wilting;
• how heat moves from a hotter object to a cooler object.

If the job is not clear, the model will usually collect unnecessary parts.

Step 2 | Choose Only the Necessary Parts

A model should include enough to explain the relationship, but not so much that the relationship disappears.

Digestive route model:

• mouth;
• gullet;
• stomach;
• small intestine;
• large intestine.

It does not need detailed blood vessels, enzymes or microscopic structures at Primary 4.

Step 3 | Assign Meaning to Every Object

Suppose five labelled cards are placed on a table.

Each card represents an organ.

A marble represents a piece of food moving through the route.

The model-maker should be able to answer:

• What does the marble represent?
• What does each card represent?
• What does movement from one card to another mean?
• What does the model not represent?

Step 4 | Build for Manipulation

A strong physical model can often be changed deliberately.

For Light:

• move the blocker;
• keep the screen fixed;
• change one distance;
• observe what the model predicts about blocked paths.

For Heat:

• use arrows or tokens to represent transfer direction;
• reverse which object is hotter;
• check whether the arrows reverse.

The physical movement should correspond to a scientific change.

Model 1 | Digestive Route Cards

Purpose: preserve biological sequence across different layouts.

Build: five organ cards.

Test: arrange them in a straight line, circle, zig-zag and vertical stack.

Invariant:

Mouth → Gullet → Stomach → Small Intestine → Large Intestine.

Learning: page position changes; biological route does not.

Model 2 | Digestive Function Overlay

Add function cards:

• chewing / saliva mixing;
• transport;
• further digestion / mixing;
• further digestion + absorption of digested food;
• main water absorption from remaining material.

Ask the learner to match function to organ.

Then remove organ names and identify organs from function clues.

Model 3 | Root–Plant Cause Chain

Use cards:

roots → water absorption → severe root damage → reduced uptake → wilting.

Now remove one card.

Ask:

“What reasoning step is missing?”

The physical gap makes the causal gap visible.

Model 4 | Anchoring vs Absorption

One root card can connect to two function cards:

• anchors plant;
• absorbs water and mineral salts.

Now present two question cards:

• why plant resists uprooting;
• why root damage contributes to wilting.

The learner must connect each question to the relevant function.

Model 5 | Matter With Shape Cards

Use one quantity card labelled 100 mL water.

Move it between container shapes drawn or represented by cups.

The learner should state:

• container changes;
• liquid shape changes;
• liquid height can change;
• volume remains 100 mL if none is lost.

The model represents conservation while appearance changes.

Model 6 | Solid, Liquid and Gas Property Sort

Create cards for:

• fixed shape;
• no fixed shape;
• fixed volume;
• no fixed volume;
• occupies space.

Build three state columns and place property cards.

Then add tricky examples such as sponge, rubber, cooking oil and air.

This tests whether classification is property-based rather than appearance-based.

Model 7 | Air Occupies Space

A simple physical demonstration can use an inverted plastic cup in water under suitable supervision.

The model job is not to “show air visually”.

The job is to show evidence that something invisible occupies the cup space because water does not fill it until air escapes.

Model 8 | Source–Object–Screen Rail

Build a simple track using paper strips or marked positions.

Place:

• light source token;
• object token;
• screen token.

Move only one token at a time.

The learner must state exactly which distance changed.

Model 9 | Straight-Line Light Strings

Use straight strings or rulers to represent boundary paths from source past object toward screen.

Important:

The strings are representations of selected light paths. They are not actual rays and do not imply the source emits only two paths.

Model 10 | Seeing a Book

Use three cards:

Lamp → Book → Eye.

Ask the learner to explain why the arrow does not mean the book “sends stored lamp light”.

It represents light reaching the book and reflected light entering the eye.

Model 11 | Heat-Transfer Tokens

Use:

• hotter-object card;
• cooler-object card;
• arrow token labelled heat transfer.

Place arrow:

hotter → cooler.

Then swap temperature labels.

The learner should reverse the arrow.

This exposes the misconception that heat direction depends on object type or page position.

Model 12 | Heat vs Temperature

Use separate cards:

• Heat Transfer = process arrow;
• Temperature = measured number such as 60°C.

The learner physically cannot place a temperature card as the arrow.

This separation can repair language such as “temperature flowed”.

Model 13 | Conductor Path

Create a chain of connected blocks representing a metal spoon.

Place a “hot soup” card at one end.

Move heat-transfer tokens along the chain toward the handle.

The model simplifies conduction. It does not represent particles or advanced microscopic mechanisms.

Model 14 | Investigation Variables Board

Use three labelled zones:

• CHANGE;
• MEASURE;
• KEEP COMPARABLE.

Place condition cards for a shadow or cooling investigation.

The model turns variable logic into something manipulable.

Model 15 | Baseline and Change Track

Place cards:

• start = 70°C;
• end = 58°C;
• change = 12°C.

Ask the learner to physically separate:

• state;
• state;
• difference.

This is useful for final-value vs change errors.

Model 16 | Evidence Ladder

Use cards labelled:

• direct measurement;
• repeated measurement;
• direct observation;
• indirect clue;
• guess.

Give a claim and ask the learner to rank evidence.

The ladder is an instructional model, not a universal scientific hierarchy.

Model 17 | Claim–Evidence–Reasoning Bridge

Build three physical zones:

CLAIM | EVIDENCE | REASONING.

The learner places sentence strips into each zone.

If the reasoning bridge is missing, the model visibly exposes the gap.

Model 18 | Concept Map With Movable Nodes

Instead of drawing a permanent concept map, use movable cards.

Why?

Because misconceptions can be repaired by moving a wrong link rather than redrawing everything.

Example:

Move “absorbs digested food” from large intestine to small intestine.

What Makes a Physical Model Testable?

A physical model becomes testable when the learner can ask:

• Does it preserve the correct sequence?
• Does changing one part create the expected consequence?
• Can I use it to predict an unfamiliar case?
• Does it accidentally suggest something false?
• Can another person interpret it without guessing?

Test 1 | Can the Model Survive Rotation?

Digestive cards are rearranged vertically.

If the learner now changes the biological sequence, the model reveals surface dependence.

Test 2 | Can the Model Survive Reversal?

Heat model:

swap 70°C and 20°C labels.

If the heat arrow does not reverse, the learner is following memory instead of temperature relation.

Test 3 | Can the Model Survive a New Object?

Replace metal spoon with metal rod.

Does the conductor model still apply?

If yes, the relationship transfers.

Test 4 | Can the Model Explain Evidence?

A model should not only display parts.

Ask:

“Which part of your model explains why the result happened?”

If the learner cannot point to the mechanism, the model may be decorative.

Test 5 | Can the Model Produce a Prediction?

Move the blocker closer to the source.

Ask:

“What should happen to the shadow in this fixed-screen arrangement?”

The model should support a reasoned prediction.

When a Model Misleads

All models simplify.

A physical model becomes dangerous when the learner forgets the simplification.

Examples:

• digestive organs are not literally separate cards;
• heat is not a pile of tokens;
• light rays are not strings;
• water does not become a rigid block because a volume card is fixed;
• roots are not only one function card.

The Model-Limit Card

Every physical model should include one final card:

“This model does NOT show…”

This simple habit prevents over-literal thinking.

Example Limit | Digestive Route

Model shows:

sequence and selected functions.

Model does not show:

• exact organ shape;
• actual relative size;
• microscopic structures;
• complete biochemical digestion.

Example Limit | Heat Tokens

Model shows:

direction of heat transfer.

Model does not show:

heat as physical beads moving through a spoon.

Example Limit | Shadow Strings

Model shows:

selected straight-line boundary paths.

Model does not show:

every path of light emitted by the source.

Revision: When Should a Physical Model Change?

Revise when:

• labels are ambiguous;
• sequence is wrong;
• arrows suggest the wrong relationship;
• scale accidentally misleads;
• essential variable is missing;
• the model cannot explain the evidence;
• a new counterexample exposes a bad rule.

Revision Is Not Decoration

Adding colour does not repair a wrong causal link.

Making the model larger does not fix a missing variable.

Scientific revision targets meaning.

Original Revision Case 1 | Digestive Model

Problem:

large intestine card placed before small intestine.

Revision:

restore route order.

Test:

identify small intestine from function clue.

Original Revision Case 2 | Heat Model

Problem:

arrow labelled “temperature flow”.

Revision:

change arrow to “heat transfer”; keep temperature as measured state labels.

Original Revision Case 3 | Light Model

Problem:

two strings imply only two rays exist.

Revision:

add note: “selected boundary paths only”.

Original Revision Case 4 | Matter Model

Problem:

wide bowl card says “less water”.

Revision:

replace with “same 100 mL, different height/shape”.

Model-Making and Scientific Communication

A model should support explanation.

Ask the learner to present:

1. purpose;
2. meaning of each part;
3. what can be changed;
4. what the model predicts;
5. one limitation.

Model-Making and Peer Critique

A classmate can ask:

• What does this piece represent?
• Why is this arrow pointing that way?
• What happens if I move this part?
• What does your model leave out?
• Which feature could mislead someone?

This turns model-making into scientific discourse rather than craft.

Model-Making and Portfolios

A powerful portfolio item is not only the finished model.

Include:

• first version;
• peer critique;
• revision note;
• final model photograph or diagram;
• reflection on what changed.

The evidence of growth lies in the revision.

Model-Making and Projects

A physical model can become one artefact inside a larger Science project.

But a project should still include:

• question;
• evidence;
• reasoning;
• reflection.

A model alone is not automatically a complete project.

Low-Cost Materials

Useful materials include:

• index cards;
• paper cups;
• string;
• paper clips;
• reusable blocks;
• sticky notes;
• paper arrows;
• safe modelling clay;
• recycled cardboard.

Scientific quality does not depend on expensive equipment.

Safety

Physical models should avoid:

• sharp cutting tools without adult supervision;
• hot liquids for dramatic demonstrations;
• electrical disassembly;
• small hazardous components;
• materials that create unnecessary risk.

When the model is meant to represent Heat or Light, the learner does not need the dangerous real phenomenon to understand the relationship.

The Model-Making Checklist

Question	Check
Is the scientific purpose clear?	□
Does every part have defined meaning?	□
Can one variable or relationship be manipulated?	□
Does the model preserve the correct Science?	□
Can it generate a prediction or explanation?	□
Is one model limit stated?	□
Can another learner interpret it?	□

Original Practice Set

Question 1

What makes a physical object a scientific model rather than a craft?

Question 2

Why should the purpose be stated before building?

Question 3

What does an arrow token in a Heat model represent?

Question 4

Why should a model include a limitation?

Question 5

How can a digestive model test transfer?

Question 6

What is wrong with adding decoration instead of fixing a wrong relationship?

Question 7

Why is manipulation useful in a model?

Question 8

What evidence should be kept in a portfolio after a model is revised?

Practice Answers

1. It represents a defined scientific relationship that can be explained, manipulated, tested and critiqued.

2. The purpose determines which parts are necessary and which details are irrelevant.

3. The direction of heat transfer, not a literal object moving through the material.

4. Every model simplifies; stating the limit prevents the representation from being mistaken for complete reality.

5. Rearrange or rotate the cards while preserving biological sequence and functions.

6. Decoration changes appearance but not the scientific error.

7. Changing one part lets the learner test predictions and see which relationships are invariant.

8. First version, critique, revision note and final version.

The Physical Model Diagnostic

If the learner…	Likely weak link	Repair
decorates heavily but cannot explain	purpose	state the model job first
treats tokens literally	model/reality boundary	add “does not show” card
cannot change one variable	manipulability	rebuild with movable parts
fails after rotation	surface dependence	identify invariant roles
does not revise after critique	model testing	target misleading relationship

A 40-Minute Model-Making Lesson

Minutes 1–5: state one model purpose.

Minutes 6–15: build only essential parts.

Minutes 16–20: define what each part represents.

Minutes 21–25: manipulate one variable.

Minutes 26–30: ask a peer to critique.

Minutes 31–35: revise one misleading feature.

Minutes 36–40: explain one prediction and one model limit.

What Parents and Tutors Can Ask

• “What is this model for?”
• “What does this piece represent?”
• “What can I move without changing the scientific identity?”
• “What prediction does the model make?”
• “What does the model leave out?”
• “What part should be revised after this counterexample?”

Continue Batch 18

• Science Learning Trails and Field Observation
• Science Project Planning, Evidence and Presentation
• Science Portfolios, Evidence of Growth and Reflection

The Quiet Return

A model earns its place by helping the learner think.

Build only what the question needs. Give every part a meaning. Manipulate the relationship. Test the prediction. Notice what the model hides. Revise the misleading feature. Then explain the Science without confusing the model with the world.