Observation, experiment and the method of science

What this heading covers

How science is actually done, and therefore how it should be taught. The syllabus's whole argument is that these are the same question: science is a method before it is a body of facts, so a classroom that delivers only the facts has taught something other than science.

What makes this heading answerable rather than vague is that each process skill has a precise meaning and the paper tests the boundaries between them. Observing against inferring is the most-set distinction in the whole science pedagogy syllabus.

Observation, and why it is not obvious

Observation is what the senses report. Inference is what you conclude from it.

Statement	Which it is
"The leaf is yellow at the edges."	Observation
"The plant is not getting enough water."	Inference
"The liquid turned blue."	Observation
"A chemical reaction took place."	Inference
"The ice cube got smaller over ten minutes."	Observation
"The ice melted because the room is warm."	Inference
"There are bubbles on the leaf under water."	Observation
"The plant is giving off oxygen."	Inference
"The seed in the dark did not grow."	Observation
"Seeds need light to germinate."	Inference — and a wrong one

That last pair is worth sitting with, because it shows why the distinction matters pedagogically rather than merely terminologically. A child who records the inference instead of the observation has already decided what the experiment showed, and the record can no longer be used to check them.

Observation is trainable

Children's untrained observation is dominated by expectation: they see what they were told would happen. The syllabus treats observation as a skill to be taught, not a capacity to be assumed. What training it looks like:

• Ask for quantity, not quality. "How many bubbles in a minute?" rather than "are there bubbles?"
• Ask for change against a baseline. "What is different from yesterday?" is a better question than "what do you see?"
• Ask for the negative. "What did not happen?" catches what expectation edits out.
• Use more than sight. Smell, sound, touch and temperature are observations too, with the obvious safety rules about tasting.
• Require the record at the time, not from memory afterwards.
• Compare records. Two children observing the same thing and disagreeing is the most productive five minutes available in a science lesson.

Measurement is observation made precise — which is why the metre rule, the balance and the thermometer belong in this heading rather than in a separate one about apparatus.

The other process skills

The paper lists these and asks you to identify one from a described activity:

Skill	What the child is doing
Observing	Reporting what the senses give
Classifying	Grouping by a stated property, and naming the property
Measuring	Quantifying with a unit
Inferring	Explaining an observation
Predicting	Saying what will happen before it does, from a pattern or a rule
Hypothesising	Proposing a testable explanation
Experimenting	Testing it with controls
Communicating	Recording and reporting — table, graph, drawing, words

Two that get confused: inferring explains something that has already been observed; predicting states what has not yet happened. "The soil is wet, so it rained" is an inference. "Those clouds mean it will rain" is a prediction.

Experiment: the anatomy of a fair test

An experiment is not a demonstration with a known ending. Its structure is examinable:

1. A question.
2. A hypothesis — a testable proposed explanation, phrased so it could be wrong.
3. A design that changes one factor at a time.
4. A control — the comparison case in which that factor is not changed.
5. Observation and measurement, recorded as they happen.
6. A conclusion, which may contradict the hypothesis.

The three variables

Variable	What it is	Also called
Independent	The one thing you deliberately change	Manipulated
Dependent	What you measure to see the effect	Responding
Controlled	Everything else, kept the same	Constant

A fair test changes the independent variable and holds every controlled variable constant. A question describing an experiment and asking "why is this not a fair test?" is asking which controlled variable was allowed to vary.

Worked: does a plant need light?

• Question: does a plant need light to stay green and grow?
• Hypothesis: a plant kept in the dark will lose its green colour and grow poorly.
• Independent variable: light — one plant on the windowsill, one in a dark cupboard.
• Dependent variable: colour of the leaves, and height, measured every two days.
• Controlled: same species, same age and size at the start, same pot and soil, same amount of water, same temperature.
• Control group: the plant in the light.

The controlled variable children most often miss is water — a plant shut in a cupboard is easily forgotten, and then the experiment has tested light and water at once and shows nothing about either. That is the question the paper sets about this experiment.

Worked: does sugar dissolve faster in hot water?

• Independent: temperature of the water.
• Dependent: time until no sugar is visible.
• Controlled: same volume of water, same mass of sugar, same grain size, same stirring — or no stirring at all, in both.

Stirring one and not the other is the classic spoiled test. So is using a lump in one glass and powder in the other, because grain size is a second independent variable smuggled in.

Worked: which surface slows a rolling ball most?

• Independent: the surface — tile, carpet, cloth.
• Dependent: distance rolled before stopping.
• Controlled: the same ball, released from the same height on the same ramp (so it starts with the same speed), measured from the same starting line.

Releasing by hand instead of from a fixed ramp height is the usual failure: the push becomes an uncontrolled variable and the result measures the child's hand.

Why "practicals" are often not experiments

A practical in which every learner follows a printed procedure to a stated result practises technique — handling, measuring, recording — which is worth having and is not the same as science. The hypothesis is absent, the outcome is known in advance, and nobody could be shown wrong.

The syllabus's distinction, and a question sets it: a verification confirms what has been taught; an investigation answers a question that is open at the start. Both have a place; only the second teaches the method.

Discovery and inquiry

• Discovery method — learners arrive at the principle themselves from evidence, rather than being told it and shown a confirmation.
• Inquiry method — learners pursue a question through their own investigation, including deciding how to investigate it.

Inquiry is broader: discovery usually has a destination the teacher has in mind, while inquiry may not. The syllabus prefers both to telling, and the paper's options reflect that.

Guided discovery is the practical middle: the teacher frames the question and supplies the materials, and the learners do the reasoning. This is what most classroom "discovery" actually is, and saying so is not a criticism.

The honest costs

A good answer names these rather than presenting inquiry as free:

• It is slow. Covering a syllabus by inquiry takes longer than covering it by telling.
• It needs materials and planning.
• Children can discover the wrong thing, which then has to be unpicked.
• It is harder with large classes, which is the ordinary condition.

The syllabus's position is that the cost is worth paying for the concepts that matter most, not that every lesson must be an investigation.

Approaches: disciplinary and integrated

• Disciplinary — physics, chemistry and biology taught as separate subjects. Appropriate at the secondary stage and beyond.
• Integrated — science taught as one subject organised around themes. This is the upper primary stage, and it is what the syllabus means by the integrated approach.

The justification, which a question asks for directly: a real phenomenon does not belong to one discipline, and children do not divide the world that way. Rusting is chemistry and economics; a leaf is biology, chemistry and physics at once; rain is all three plus geography. Separating the disciplines early teaches that the divisions are in nature rather than in the university.

A further argument the syllabus makes: integration is what allows science to connect to health, environment, technology and daily life — the connections that make the subject matter to a child who will not become a scientist.

A question of the kind the paper sets

A teacher asks children to find whether sugar dissolves faster in hot or cold water. One child uses a spoonful of powdered sugar in hot water and a sugar cube in cold water. The experiment is not valid because: (a) the child did not repeat it (b) the amount of water was not measured (c) more than one factor was changed (d) the child did not state a hypothesis

(c).

Both temperature and grain size were changed, so if the hot water dissolves its sugar first, nothing can be concluded about temperature — the powder would have dissolved faster in any case.

The others are real weaknesses and none is the reason the result cannot be read. (a) and (b) would make it less reliable; (d) is a procedural omission. Only (c) makes the comparison meaningless, and the stem asks why it is not valid.

Notice how to read such questions fast: find the independent variable the teacher intended, then look for a second thing that moved with it.

Mistakes that cost marks here

• Recording an inference as an observation. "The plant needs water" is not what was seen.
• Confusing inferring with predicting. One explains what happened, the other states what will.
• Calling a demonstration an experiment, or a verification an investigation.
• Forgetting the control, or naming the controlled variables as "the control group" — they are not the same thing.
• Letting a second variable change and still drawing a conclusion.
• Presenting inquiry as cost-free. Name the time, materials and class-size costs.
• Treating the integrated approach as "mixing subjects" rather than as organising science around phenomena.
• Saying observation is passive. It is a trained skill, and expectation distorts it.

In one line

Observation reports and inference explains; an experiment changes one variable, holds the rest and keeps a control, which is why the dark-cupboard plant must still be watered; a practical with a known answer teaches technique rather than method; and the integrated approach exists because phenomena do not come divided into physics, chemistry and biology.

In the classroom

The quickest way to teach the observation–inference distinction is two columns on the board during any practical: what we saw and what we think it means. Children move statements between the columns and argue about them, and after a week they stop writing conclusions in the first column without being told.

For fair testing, a deliberately spoiled experiment is better than a correct one. Run the sugar test with a cube in one glass and powder in the other, ask what it proves, and let the class find the flaw. A class that has caught an unfair test will design fair ones afterwards without the term being defined for them.