Thinking Like a Scientist: Reasoning and Models
How do scientists know about things they cannot see, like tiny particles? They use scientific models and scientific reasoning. You will learn what a model is, how scientists go from data to a general rule (and back again), and how plant experiments led to our idea of photosynthesis. Let's start!
Planning a fair investigation is covered in "Hypotheses, variables and controlled experiments". Sections 5 to 7 are an Extension topic in the 2025 curriculum.
1. What Is a Scientific Model?
A scientific model is a simplified way of showing how something in nature works. It can be a drawing, a diagram, a physical object, an equation or a computer program. Modelling is the process scientists use to build, test and improve models.
Think of the MTR route map. It leaves out buildings and the real bends of the tunnels, but shows what you need: which stations connect. A model keeps the important parts only.
Three jobs of a model: describe, predict, explain
Describe – show what something is like. A water cycle diagram shows evaporation, condensation and precipitation.
Predict – say what will happen next. A food web predicts what may happen to other animals if one species disappears.
Explain – say why something happens. The particle model explains why a liquid expands when heated.
Memory aid: a good model is a DPE helper – it Describes, Predicts and Explains.
Examples of scientific models
Particle model – all matter is made of tiny moving particles. It explains solids, liquids and gases.
Water cycle diagram – shows how water moves between the sea, the air and the land.
Food web – shows feeding relationships in an ecosystem, such as a woodland in a Hong Kong Country Park.
DNA model – a twisted ladder shape (double helix) showing the structure of DNA, which carries genetic information.
Maps and computer models – a map models the Earth's surface; forecasters run computer models to predict the weather.
Models can change
A model is the best idea scientists have so far. When new evidence does not fit, the model is revised or replaced. People once used a model with the Earth at the centre of everything. Observations of the planets later led to a model with the Sun at the centre of the Solar System. This is a strength of science!
Key takeaway: A scientific model is a simplified representation used to describe, predict and explain. Models are revised when new evidence appears.
2. Inductive Reasoning: From Data to a General Rule
Inductive reasoning (induction) means using specific data to reach a general conclusion, trend or model. You go from "these cases" to "all cases like these".
Example: heating a liquid
A class warms a flask of coloured water, fitted with a narrow glass tube, in a water bath and records the water level in the tube:
20 °C → 5.0 cm
40 °C → 5.8 cm
60 °C → 6.9 cm
80 °C → 8.2 cm
Pattern: each time the temperature went up, the level rose. General conclusion (induced): "When a liquid is heated, its volume increases." A line graph shows this clearly.
Building a model to explain the trend
The trend tells us what happens, not why. The particle model explains: when a liquid is heated, its particles gain energy, move faster and move further apart, so the liquid takes up more space.
Common mistake
"The particles get bigger when heated" is wrong! The particles stay the same size. It is the spaces between the particles that get larger.
Key takeaway: Induction goes from specific to general. Several observations lead to a trend, a conclusion or a model.
3. Deductive Reasoning: From a Rule to a Specific Result
Deductive reasoning (deduction) works the other way: start with a general rule, trend or model and use it to work out a specific result.
Examples of deduction
From a trend: What will the level be at 50 °C? Since 50 °C is between 40 °C and 60 °C, we deduce the level will be between 5.8 cm and 6.9 cm.
From a model: Liquids expand when heated, so we deduce that the liquid in a thermometer will rise when its bulb is put into hot water. That is how a thermometer works!
From a food web: Snakes eat frogs. If many snakes were removed, we deduce that the number of frogs would probably increase, at least for a while.
Induction or deduction? A memory aid
Picture a staircase. Induction climbs UP: from many small facts to one big rule. Deduction steps DOWN: from the big rule to one particular case.
Induction: specific data → general rule
Deduction: general rule → specific result
How the two work together
1. Observe and collect data.
2. Induce a trend or build a model.
3. Deduce a prediction from the model.
4. Test the prediction.
5. If the result matches, the model gains support. If not, revise the model.
Key takeaway: Deduction goes from general to specific. A model or trend predicts a particular result, which can then be tested.
4. Story Time: How Scientists Discovered Photosynthesis
Our idea of how plants make food was built step by step by different scientists, each reasoning from evidence.
Van Helmont and the willow tree
In the 1600s, the Flemish scientist Jan Baptista van Helmont planted a small willow tree in a pot of weighed, dried soil. For five years he gave it only water. The tree gained about 74 kg, but the soil lost only about 57 g. He reasoned that the mass did not come from the soil, so it must have come from the water.
Reasoning check: A clever use of evidence, but only partly right. He did not consider the air. Today we know much of a plant's mass is built from carbon dioxide from the air, together with water.
Did you know?
Van Helmont is also famous for inventing the word "gas"!
Priestley and the mint plant
In the 1770s, the English scientist Joseph Priestley found that a candle burning in a sealed jar soon went out, and a mouse in a sealed jar could not survive. But after he kept a sprig of mint in the "used" air for some days, a candle could burn in it again. He inferred that plants can restore air spoiled by burning or breathing.
Ingenhousz and the importance of light
In 1779, the Dutch scientist Jan Ingenhousz extended Priestley's experiments. He found that plants restore the air only in light, and only their green parts do this. Green leaves under water gave off small gas bubbles in sunlight, but not in the dark.
Putting the story together
Van Helmont: plants gain mass from water.
Priestley: plants also make used air fit to breathe again.
Ingenhousz: plants do this only in light, using their green parts.
Later scientists joined these ideas into today's model of photosynthesis: green plants use light energy to make food from carbon dioxide and water, and give out oxygen (studied in another unit).
Key takeaway: Scientific knowledge grows as scientists reason from new evidence and revise earlier models.
5. Inference to the Best Explanation (Extension topic)
When an idea cannot be tested directly, scientists, like detectives, infer the most likely explanation: they list possible explanations and choose the one that fits all the evidence best.
Example: the wet playground
The school playground is wet in the morning. Possible explanations: (a) it rained overnight, (b) the janitor washed the ground, (c) a water pipe burst.
Evidence: the roads, cars and trees outside are also wet. Best explanation: it rained, because only (a) explains all the observations.
The best explanation is the most likely one, not a proven fact. New evidence, such as a report of no rain last night, means we must think again.
Key takeaway: Choose the explanation that fits the most evidence, and stay ready to change your mind.
6. Making Decisions Based on Likelihood (Extension topic)
Science cannot always say for certain what will happen. It often tells us how likely something is (its probability), and we use this to make informed decisions.
Everyday examples
Weather: Forecasts, such as those from the Hong Kong Observatory, tell us how likely rain is. If the chance is high, bring an umbrella, even though rain is not certain.
Air quality: When the Air Quality Health Index (AQHI) is high, health risk is higher, so people with heart or breathing problems are advised to reduce outdoor exercise.
Typhoons: When a typhoon might come near Hong Kong, people secure windows early: the possible harm is serious even if a direct hit is not certain.
Tip: think about how likely an event is AND how serious it would be. A small chance of serious harm can still be worth acting on.
Key takeaway: Science gives us probabilities, not promises. Use likelihood to make sensible choices.
7. Limitations of Scientific Reasoning (Extension topic)
Scientific reasoning is powerful, but not perfect.
Small samples can mislead
For a long time, Europeans had only seen white swans, so they concluded "all swans are white". Then black swans were found in Australia. One new observation showed the rule was false. Induction from a few cases can support a rule, but never prove it 100%.
Correlation is not causation
Two things can change together without one causing the other. In summer, ice-cream sales and the number of swimmers at beaches both rise. Ice cream does not make people swim! Both are caused by a third factor: hot weather. To show real cause and effect, scientists need a controlled experiment.
Models are simplifications
Every model leaves things out. Real particles are not little coloured balls, and weather models get less accurate the further ahead they predict. A deduction is only as good as its model: a wrong or too-simple model can give a wrong prediction.
Key takeaway: Small samples, confusing correlation with causation and over-trusting simple models limit scientific reasoning.
Watch Out! Common Mistakes
Mistake 1: "A model must look exactly like the real thing." No – models are simplified on purpose. And a changed model is not a failure: revising is how science improves.
Mistake 2: Mixing up induction (specific to general) and deduction (general to specific).
Quick Review
1. Modelling builds models that describe, predict and explain natural phenomena.
2. Examples: particle model, water cycle diagram, food web, DNA model, maps and computer models.
3. Models are revised when new evidence does not fit.
4. Induction: specific data → general trend, conclusion or model.
5. Deduction: model or trend → specific result.
6. Van Helmont, Priestley and Ingenhousz each added evidence towards the model of photosynthesis.
7. (Extension) Pick the explanation that fits all the evidence, decide using likelihood, and know the limits of reasoning.
Great work! Every time you spot a pattern or make a prediction, you are reasoning like a real scientist. Keep asking, "How do we know?"