Primary 6 Science is full of models. A circuit diagram is a model of electrical connections. A food web models feeding relationships. A graph models how quantities change. A life-cycle diagram models stages through time. A fair-test setup models a relationship between variables. These representations are powerful because they simplify the world—but every simplification also has limits.
This guide develops a higher-order Primary 6 capability: using models, spotting assumptions, recognising limits, comparing competing explanations and judging scientific claims against evidence.
Return to the Primary 6 Science Learning Hub.
The model rule
Use this route:
WHAT DOES THE MODEL SHOW? → WHAT DOES IT LEAVE OUT? → WHAT ASSUMPTIONS ARE BEING MADE? → WHAT EVIDENCE SUPPORTS THE CLAIM? → HOW FAR CAN THE CLAIM TRAVEL?
This is an eduKate reasoning routine, not an official SEAB answer formula.
Part I — A model is useful because it is incomplete
A model simplifies reality so that one relationship becomes easier to see.
A circuit diagram does not show the exact physical shape of wires or components. It shows functional connections.
A food web does not show every organism, every feeding event or exact population size. It shows selected feeding relationships.
A graph does not show every microscopic event. It shows measured quantities and their relationship.
The pupil should ask what job the model is designed to perform.
Do not confuse model appearance with reality
Objects in diagrams may not be drawn to scale. Distances may be exaggerated. Arrows may show direction rather than force. Colours may be symbolic.
Use labels, legends, stated measurements and relationships rather than visual impression alone.
Part II — Assumptions are hidden conditions
An assumption is something treated as true for the purpose of reasoning even if it is not directly measured in the question.
Examples:
- two plants are similar enough for comparison;
- the release method is consistent;
- the sensor is working correctly;
- the diagram connections represent the actual setup;
- other relevant environmental conditions remain similar.
Some assumptions are stated explicitly. Others must be noticed when evaluating the strength of a conclusion.
Assumptions can be reasonable without being certain
A good scientific answer does not need to distrust everything. It needs to know when a conclusion depends on conditions that were not directly verified.
Example: if two “identical” plants are used, the design assumes biological differences between them are small enough not to dominate the result.
Part III — Evidence supports claims, not stories
A claim should match what the observations or measurements support.
Evidence: a car travels shorter distances on rougher tested surfaces.
Supported claim: under the tested conditions, rougher surfaces produced greater frictional effects on the car.
Unsupported extension: every rough surface everywhere will stop every object faster.
The second statement travels far beyond the evidence.
The claim ladder
Think of conclusions as having different strength.
- Observation: what was directly measured.
- Pattern: relationship among observations.
- Inference: scientific interpretation of the pattern.
- Generalisation: claim extended beyond the exact tested cases.
The higher the claim climbs, the more evidence and caution it usually needs.
Part IV — Correlation and causation
If two quantities change together, that is a relationship. It does not automatically prove one caused the other.
Observation study: places with more plants also have more insects.
Possible explanation: plants may provide food or habitat.
But other conditions such as water, shade or temperature may also differ.
A controlled experiment that deliberately changes one factor while holding relevant alternatives comparable provides stronger causal evidence.
Part V — Competing explanations
Strong scientific reasoning considers more than one possible explanation and tests them against the evidence.
Example: a plant grows less in Setup B.
Possible explanations:
- less light;
- less water;
- different plant health;
- temperature difference;
- measurement error.
If the experiment changes only light and controls other relevant conditions, the light explanation becomes stronger.
Eliminate explanations with evidence
Do not ask only, “What could cause this?” Ask, “Which possibilities does the evidence rule out?”
This is especially powerful in MCQ questions where distractors are scientifically possible in general but contradicted by the setup.
Part VI — Model limits in graphs
A trend line is not a promise that the pattern continues forever.
If spring extension increases for loads 1–4, predicting at load 5 may be reasonable if the question asks. Predicting at load 100 assumes the same relationship far beyond the measured range.
Recognise extrapolation risk.
Plateaus and thresholds
A model may show a relationship that changes shape.
A photosynthesis-related graph may rise and then level off. The plateau shows that increasing the horizontal-axis factor further no longer produces much additional measured effect under those conditions.
Do not force the early straight-line trend through the plateau.
Part VII — Model limits in food webs
A food web shows feeding relationships, not exact population dynamics.
If one prey decreases, a predator may be affected. But the web may show alternative prey.
A food web also does not automatically show:
- how much of each food is eaten;
- population sizes;
- reproductive rates;
- disease;
- weather changes;
- migration;
- every organism in the ecosystem.
Predictions should remain conditional when the model omits relevant information.
Part VIII — Model limits in circuits
A circuit diagram shows connections, not exact wire length, resistance or device efficiency unless specified.
If two diagrams are functionally identical but drawn in different shapes, they may represent the same circuit.
Do not infer electrical behaviour from artistic layout.
Part IX — Model limits in biological diagrams
Organ diagrams simplify real anatomy. Arrows may show flow. Colours may distinguish regions. Sizes may be exaggerated.
Use the functional relationships taught at Primary level.
A diagram of the circulatory system is not a literal map of every blood vessel. A plant-transport diagram simplifies pathways. The model’s purpose is to show movement and function.
Part X — Method claims
A fair test can support a relationship only if the changed variable is isolated well enough and the measured outcome is appropriate.
A conclusion can fail because:
- two variables changed together;
- the measurement did not represent the intended outcome;
- the sample was too narrow;
- results were not repeated;
- the instrument was unsuitable;
- the conclusion extended beyond tested conditions.
Reliability is not validity
A method can give the same answer repeatedly and still measure the wrong thing.
Example: counting plant leaves may be very consistent, but if the question asks for exact photosynthesis rate, leaf count may not be a valid direct measure.
Part XI — Healthy scepticism
Scientific scepticism means checking whether the evidence supports the claim. It does not mean distrusting every result automatically.
Ask:
- Was the method appropriate?
- Were the measurements consistent?
- Were relevant variables controlled?
- Is there enough evidence?
- Does the conclusion match the data?
If the answer to these is strong, confidence should increase.
Original claim workshop 1: spring
Data: 1 unit load → 2 cm extension; 2 units → 4 cm; 3 units → 6 cm.
Claim A: “Extension increased with load in the tested range.”
Supported.
Claim B: “Every spring in the world will extend exactly 2 cm per unit forever.”
Unsupported. It overgeneralises across springs and beyond the tested range.
Original claim workshop 2: food web
A food web shows hawks eating rabbits and birds. Insect numbers fall; birds eat insects.
Claim: “Hawks must decrease immediately.”
Too strong. Hawks have another prey source, and population effects may be delayed.
Original claim workshop 3: plant experiment
Two plants are compared. Plant A receives light and water. Plant B receives less light and less water. A grows more.
Claim: “Light caused the growth difference.”
Weak causal conclusion. Water changed too, so the design does not isolate light.
Original claim workshop 4: cooling cup
One trial shows Cup Q stays warmer after ten minutes than Cup P.
Claim: “Cup Q is always a better insulator.”
Too broad. The evidence supports a difference in this trial under these conditions; repeated comparisons would strengthen the generalisation.
Part XII — Negative results
“No observed effect” is still information.
If changing a variable produces no measurable difference under the tested conditions, that result can challenge the proposed relationship or show that the effect is below the measurement method’s resolution.
Do not treat negative results as failed experiments automatically.
Missing evidence is not evidence of absence
If a gas was not measured, we do not know its value. If a population was not counted, we cannot assume zero.
A blank graph region or untested condition is unknown, not automatically absent.
Part XIII — Updating explanations
Scientific explanations should change when new evidence appears.
Initial evidence may support Explanation A. A later result may contradict it. The pupil should revise the model rather than defend the first answer because it was familiar.
This is scientific thinking: explanations are accountable to evidence.
Part XIV — Confidence language
Use wording that matches evidence strength.
| Evidence strength | Useful language |
|---|---|
| Direct measured comparison | shows, demonstrates in this setup, supports |
| Reasonable but indirect inference | suggests, is consistent with, may indicate |
| Prediction beyond range | if the relationship continues, would be expected |
| Insufficient evidence | cannot conclude, not enough evidence, remains unknown |
Part XV — Evaluating a scientific claim
Use the CLAIM test:
- C — Condition: Under what conditions was the evidence obtained?
- L — Link: What relationship connects evidence to claim?
- A — Alternatives: What other explanations could fit?
- I — Information: What evidence is actually available?
- M — Model limit: How far can the conclusion be extended?
This is an eduKate teaching routine.
Part XVI — MCQ use
For each option, ask:
- Does it fit every stated condition?
- Does it introduce an unmeasured property?
- Does it overgeneralise?
- Does it confuse observation with cause?
- Does it require an assumption not supported by the question?
The best option is the one that survives the evidence, not the one that sounds most scientific.
Part XVII — Open-ended use
A strong evaluation answer names the weakness and explains its consequence.
Weak: “The experiment is not fair.”
Stronger: “Both light and water availability changed between the two plants, so the growth difference cannot be attributed confidently to light alone.”
The second answer identifies the violated assumption and the limit it creates.
Part XVIII — What parents can ask
- What does this diagram show?
- What does it not show?
- Which part of your answer is directly observed?
- Which part is inferred?
- What else could explain the result?
- What evidence would rule out that alternative?
- How far can this conclusion be extended?
Where to connect
- Data, Graphs, Diagrams & Evidence
- Investigations, Variables, Fair Tests & Method
- Primary 5 Models, Assumptions, Simplification & Limits
Retrieval checklist
- I can explain what a model shows and leaves out.
- I can identify assumptions in a comparison.
- I can distinguish observation, pattern, inference and generalisation.
- I can avoid turning correlation into causation.
- I can compare competing explanations.
- I can recognise extrapolation beyond the tested range.
- I can explain why reliability does not guarantee validity.
- I can treat negative results as evidence.
- I can update an explanation when new evidence appears.
- I can match confidence language to evidence strength.
The quiet return
A scientific model earns its value by simplifying reality enough to reveal a relationship. Scientific judgement begins when the pupil also understands the model’s boundaries.
Use the model. Check the assumptions. Test the alternatives. Respect the evidence. Stop the claim at the boundary.
Return to the Primary 6 Science Learning Hub.
