A Primary 6 pupil can know the Science and still misunderstand the investigation. The question says that a pupil changes the distance between a lamp and a plant, counts gas bubbles, repeats the test and draws a conclusion. The learner recognises photosynthesis immediately—but the mark may depend on something different: what was changed, what was measured, what had to stay comparable, and whether the method really supports the conclusion.
This is the inquiry layer of the Primary 6 Science Learning Hub. It develops the scientific-method work required across photosynthesis, forces, energy and environmental contexts. The investigations and questions here are original eduKate teaching material.
Quick answer: what is the investigation job?
Before naming the topic, reconstruct the investigation:
QUESTION → CHANGED CONDITION → MEASURED OUTCOME → CONTROLLED CONDITIONS → PROCEDURE → DATA → CONCLUSION → LIMITS.
This is an eduKate reasoning routine, not an official SEAB answer formula. It is useful because it separates eight jobs that pupils often blend together.
SEAB’s 2026 PSLE Science objectives include applying scientific inquiry: making predictions and formulating hypotheses, interpreting and analysing information, evaluating observations, information and methods, and communicating explanations and reasoning. Official reference: SEAB PSLE Formats Examined in 2026.
The experiment is a comparison with a purpose
An investigation is not a collection of apparatus. It is a designed comparison intended to answer a question.
Suppose a pupil tests whether surface type affects how far a toy car travels after leaving a ramp. The scientific purpose is not “to roll a car down a ramp”. The purpose is to compare the measured travel distance when surface type changes while relevant other conditions remain comparable.
That one sentence already reveals:
- the changed variable: surface type;
- the measured variable: distance travelled;
- the scientific relationship being tested: surface condition and motion, through friction;
- the need for controls: same car, ramp, release point and measurement method.
Changed variable, measured variable and controlled conditions
Different schools may use terms such as independent variable, dependent variable and controlled variables. At Primary level, pupils should be able to recognise the jobs even when wording varies.
The changed variable
This is the factor deliberately varied to test its relationship with an outcome.
Examples:
- distance between lamp and plant;
- surface type under a moving object;
- amount a spring is stretched;
- temperature of the environment;
- amount of water supplied to plants.
A common error is to name the values instead of the variable. “10 cm, 20 cm and 30 cm” are values of lamp distance. The variable is the distance between the lamp and plant.
The measured variable
This is the outcome observed or measured in response to the changed condition.
Examples:
- number of bubbles produced in five minutes;
- distance travelled by the toy car;
- extension of a spring;
- number of organisms observed;
- time taken for an event to occur.
Another common error is to name the instrument instead of the outcome. “Ruler” is not the measured variable. “Length of spring extension” is.
Controlled conditions
These are conditions that should be kept comparable so the changed variable is the most defensible explanation for differences in the measured outcome.
Not every imaginable factor needs to be controlled. The pupil should identify the conditions that could plausibly affect the outcome.
The fair-test question: what else could explain the result?
A useful way to find a control is to ask:
If this condition changed too, could it also change the measured outcome?
If yes, it may need to be controlled.
Example: light and photosynthesis
A pupil changes lamp distance and counts bubbles from an aquatic plant.
Potential controls include:
- same species and size of plant;
- same amount and composition of water;
- same duration of measurement;
- same carbon dioxide availability;
- same lamp except for distance;
- same temperature as far as reasonably possible.
Why might temperature matter? Moving a lamp can change not only light intensity but also heating. If temperature also changes substantially, the comparison becomes harder to interpret because more than one relevant condition has changed.
Fair test does not mean every result must be equal except one
A fair test is about design, not about forcing a desired pattern. Results can vary even in a well-designed investigation. Measurements have natural variation, instruments have limits and biological specimens are not identical machines.
That is why repeated trials, careful measurement and transparent reporting matter.
Repeated trials: why repeat?
Repeating a measurement can help identify variation and reduce the influence of an unusual single result. Repetition can also improve confidence in a pattern when similar outcomes are obtained repeatedly.
But “repeat three times” is not automatically a complete method improvement. Ask what is being repeated and why.
Example
A toy car travels 82 cm, 80 cm and 81 cm on Surface A. On Surface B it travels 42 cm, 41 cm and 43 cm.
The repeated results are close within each surface condition and clearly different between surfaces. This supports a stable comparison.
If Surface A results were 82 cm, 31 cm and 79 cm, the unusually low value would deserve investigation. Perhaps the car hit an obstacle, was released differently or the measurement was recorded incorrectly.
Average is useful, but it does not erase the trials
An average summarises repeated measurements. It does not mean every trial produced the average value.
For 82 cm, 80 cm and 81 cm, the average is 81 cm. That does not mean the car travelled exactly 81 cm three times.
When a question provides an average, keep the distinction between:
- individual results;
- summary value;
- pattern across conditions.
Prediction and hypothesis are not random guesses
A scientific prediction should follow from a concept, pattern or prior evidence.
Weak prediction: “The car will go 50 cm because I think so.”
Stronger prediction: “The car will travel a shorter distance on the rougher surface because the greater frictional effect will slow it more quickly.”
The numerical value may not be known in advance, but the predicted relationship can be scientifically grounded.
Hypothesis as a relationship statement
A hypothesis can propose a testable relationship between a changed condition and an outcome. At Primary level, avoid turning it into a complicated formal sentence if the pupil does not understand the relationship.
Example: “If the spring is stretched further, the object launched by the spring will travel farther, provided the other relevant conditions are kept the same.”
The investigation then tests whether the data support that proposed relationship.
Do not confuse the investigation question with the conclusion
Investigation question: Does surface type affect the distance travelled by the car?
Conclusion after data: In this investigation, the car travelled a shorter distance on rougher surfaces.
The question is what we want to find out. The conclusion is what the evidence supports after the investigation.
The control setup is not the same as a controlled variable
These terms sound similar and are often confused.
A controlled variable is a condition kept the same across comparison setups. A control setup is a comparison condition used to show what happens without the treatment or factor of interest, where appropriate.
Example
To test whether fertiliser affects plant growth, one group receives fertiliser and another similar group does not. The no-fertiliser group can serve as a control setup. Water amount, plant species and light exposure may be controlled variables.
Do not call the no-fertiliser group “the controlled variable”.
One specimen versus several specimens
Biological investigations create an important design choice. If the same specimen is used repeatedly, earlier treatment may affect later results. If different specimens are used, natural differences between specimens may affect comparison.
There is no universal answer that one is always better. The pupil should identify which source of variation matters in the given context.
Example: plant growth under light conditions
Using one plant first in bright light and then in dim light may be problematic because the plant’s history and growth during the first condition can affect the second. Using several similar plants for each condition may make comparison more practical, but differences among plants should be managed by choosing similar specimens and using sufficient replication.
Measurement quality: what exactly is being measured?
A result becomes scientific evidence only when the measurement corresponds to the idea being investigated.
Proxy measurements
Sometimes the desired process cannot be measured directly, so a related observable quantity is used.
For example, counting gas bubbles from an aquatic plant may be used as an indicator related to gas production, but bubble size can vary. Counting bubbles is not identical to measuring exact gas volume. A stronger method might collect and measure the volume of gas if the apparatus and level are appropriate.
This distinction is important because pupils should know when a measurement is an approximation.
Instrument resolution and reading
Every measuring instrument has limits. A ruler marked in millimetres cannot reliably distinguish changes much smaller than its scale. A timer operated by hand includes human reaction time. A measuring cylinder has scale markings that limit precision.
At Primary 6, pupils do not need a university treatment of uncertainty, but they should recognise that measurements are not infinitely exact.
Method limitation versus mistake
A mistake is an avoidable error in carrying out the intended method: reading the wrong scale, releasing the car from the wrong point or recording a value incorrectly.
A method limitation is a weakness in the design itself: for example, measuring plant growth from one specimen only when individual variation is large.
Calling every limitation “careless” prevents useful improvement. The pupil needs to identify whether the failure belongs to execution or design.
Original investigation workshop 1: Photosynthesis under different lamp distances
A pupil places equal lengths of an aquatic plant in three beakers. Each beaker contains the same volume of water with the same dissolved carbon dioxide concentration. Identical lamps are placed 10 cm, 20 cm and 30 cm from the plants. The number of bubbles produced in five minutes is counted.
Question A: What is the changed variable?
Answer: Distance between the lamp and the plant.
Question B: What is the measured outcome?
Answer: Number of bubbles produced in five minutes.
Question C: Name two conditions that should remain comparable.
Answer: Any two relevant examples such as plant species/size, water volume, carbon dioxide availability, measurement duration or lamp type.
Question D: Why might temperature be a concern?
Answer: Different lamp distances may also produce different heating, so temperature could change together with light exposure and affect the result.
Question E: Why might counting bubbles be a limitation?
Answer: Bubbles may not be equal in size, so bubble count does not necessarily measure exact gas volume.
Original investigation workshop 2: Friction and stopping distance
A toy car rolls down the same ramp from the same release point onto four horizontal surfaces. The distance travelled after leaving the ramp is measured three times for each surface.
Question A: Why use the same release point?
Answer: Changing release point could change the car’s motion when it reaches the horizontal surface, affecting stopping distance and confounding the comparison between surfaces.
Question B: Why repeat three times?
Answer: Repetition helps reveal variation and reduces reliance on a single possibly unusual result.
Question C: The pupil pushes the car slightly on one trial. What type of problem is this?
Answer: An execution mistake because the intended release method was not followed consistently.
Question D: The car’s wheels become wet before testing Surface D. Why is this a problem?
Answer: Contact conditions changed as well as surface type, so the comparison no longer isolates the intended variable cleanly.
Original investigation workshop 3: Elastic spring force
A spring is hung vertically. Masses are added one at a time, and the length of the spring is measured after each addition.
The pupil wants to find how load affects spring extension.
Question A: What should be calculated before comparing?
Answer: Spring extension, which is the measured length minus the original unstretched length.
Question B: Why is using spring length alone less clear?
Answer: The original length contributes to every measurement. Extension isolates the change caused by the load.
Question C: Why should the same spring be used for the comparison?
Answer: Different springs may have different properties, creating another relevant variable.
Question D: If the spring does not return to its original length after a very large load, what should the pupil do?
Answer: Recognise that the spring may have been permanently changed and should not be treated as unchanged for later comparable trials.
Original investigation workshop 4: Environment and survival
A pupil investigates where woodlice are more commonly found by placing equal numbers in a chamber offering a damp side and a dry side, then recording their positions after a fixed time.
Question A: What relationship is being investigated?
Answer: The relationship between moisture condition and where the woodlice are found/prefer to remain under the tested conditions.
Question B: Why should light and temperature be similar on both sides?
Answer: If they differ too, the pupil cannot tell whether moisture or the other environmental factor influenced the distribution.
Question C: Why is one group of woodlice better than observing one individual?
Answer: A group gives more evidence about the pattern and reduces dependence on one organism’s unusual behaviour.
Question D: Can the pupil conclude that all woodlice everywhere always choose damp places?
Answer: No. The evidence applies to the tested organisms and conditions. Generalisation should stay within the strength of the evidence.
Procedure order matters
Some investigations depend on sequence. If a measurement is taken after a treatment, reversing the order may change what is being measured.
Ask:
- What must happen before the measurement?
- How long should the system be allowed to respond?
- Must a baseline be recorded first?
- Could one step alter the system for later trials?
Start, throughout or end conditions
Pupils often see “same amount of water” and assume it is enough to set it once. But some conditions must remain comparable throughout the investigation.
For a plant investigation lasting several days, initial water amount may be equal, yet evaporation rates could differ if one setup is hotter. A condition that begins equal may not stay equal.
This is a higher-quality question to ask: Was the condition only equal at the start, or was it kept comparable during the whole process?
Same time versus same stage
Comparisons can be made after the same duration or at the same stage of a process. These are not always equivalent.
Suppose seeds germinate at different rates under two conditions. Measuring root length after exactly three days compares the same time point. Comparing root length when each seed first develops two leaves compares the same developmental stage. The choice depends on the investigation question.
Time to reach versus value after the same time
These two measured outcomes are often confused.
- Time to reach: How long does the water take to cool to 40°C?
- Value after same time: What is the water temperature after ten minutes?
Both can study cooling, but they are different dependent variables. Pupils should answer the one actually measured.
Connected setups: when one part feeds into the next
Some diagrams show multiple containers, tubes or components. Before identifying variables, decide whether the parts are independent comparisons or connected stages of one system.
If gas produced in Container A flows into Container B, a change in A can affect B. Treating them as separate unrelated setups would break the scientific model.
Use arrows or tracing with a finger to identify the pathway before reading labels.
Branching processes: different conditions, different outcomes
A flowchart may show one starting material followed by two possible pathways depending on a condition. Pupils should identify the decision point.
For example:
Same organism → Condition X → Outcome A
Same organism → Condition Y → Outcome B
The important variable is the condition at the branch, not the starting organism if that is held constant.
How to evaluate an investigation
Use five questions:
- Alignment: Does the method actually measure the question?
- Isolation: Is only the intended changed condition systematically different?
- Measurement: Is the outcome measured clearly and appropriately?
- Replication: Is there enough repeated evidence to avoid dependence on one unusual result?
- Inference: Does the conclusion stay inside what the data support?
Improving a method: improvement must target a weakness
“Repeat the experiment” is not a universal answer. If the weakness is that the ruler has large scale intervals, repetition does not improve measurement resolution. If the weakness is that two variables changed together, repetition reproduces the confounding.
Match the improvement to the problem:
| Problem | Possible improvement |
|---|---|
| One biological specimen only | Use several similar specimens and compare repeated results. |
| Two conditions changed together | Redesign so only the intended condition differs. |
| Outcome measured too coarsely | Use a more suitable instrument or measurement method. |
| Human release varies | Use a consistent release mechanism. |
| Observation made at different times | Standardise measurement timing. |
| Unexpected outlier | Check procedure, repeat and report transparently rather than deleting without reason. |
What a control cannot do
Keeping variables controlled does not prove the scientific theory by itself. It improves the fairness of the comparison. The conclusion still depends on the actual data.
Likewise, one clean experiment does not establish an unlimited universal rule. It provides evidence under tested conditions.
Conclusion language: relationship is not automatically cause
If an experiment deliberately changes one condition while controlling relevant alternatives, the design can support causal reasoning more strongly than a simple observation study.
But pupils should still write what the design and results justify.
Observation: “Plants receiving more light produced more bubbles in the tested period.”
Stronger causal interpretation, if the comparison is fair and the measure is appropriate: “Increasing light exposure increased the measured rate of bubble production under these conditions.”
Overclaim: “More light always makes every plant photosynthesise faster.”
Investigation error clinic
- Error: naming the values instead of the variable. Repair: ask what property the values belong to.
- Error: naming the apparatus as the measured variable. Repair: ask what quantity the apparatus measures.
- Error: controlling everything imaginable. Repair: control conditions that could plausibly affect the outcome.
- Error: adding more trials to a biased design. Repair: fix the design first.
- Error: treating one result as a pattern. Repair: repeat appropriately.
- Error: deleting an outlier automatically. Repair: investigate why it differs and report the reasoning.
- Error: confusing a prediction with a conclusion. Repair: prediction comes before evidence; conclusion comes after.
- Error: using “fair test” as a memorised phrase. Repair: state exactly what must be kept comparable and why.
A 12-question investigation drill
For any experiment diagram, answer these before doing the official question:
- What is the investigation trying to find out?
- What is deliberately changed?
- What values are used for that variable?
- What is measured?
- What unit is used?
- Which conditions must stay comparable?
- How is the system allowed to respond?
- When is the measurement taken?
- How many trials/specimens are used?
- What result pattern would support the hypothesis?
- What alternative explanation could threaten the conclusion?
- What specific improvement would address the largest weakness?
This drill should eventually become fast enough to perform mentally.
How parents can help without turning the kitchen into a laboratory
Use ordinary household examples to practise experimental thinking.
- Which paper towel absorbs more water? What changes? What is measured? How will sheet size be controlled?
- Which surface lets a toy slide farther? How will the starting push be standardised?
- Does ice melt faster in different locations? What other temperature or airflow conditions might differ?
- Which plant gets more light? What would count as a meaningful outcome over a reasonable period?
The aim is not to generate perfect research. It is to make variable control visible.
How teachers can use wrong answers diagnostically
If a pupil cannot name the changed variable, do not immediately reteach the topic chapter. First ask whether the learner can distinguish:
- object from property;
- variable from value;
- procedure from observation;
- measurement from conclusion;
- controlled variable from control setup.
Sometimes the Science content is intact while the inquiry grammar is broken.
Bridge to evidence
A good method creates data, but data do not interpret themselves. The pupil must decide what is observed, what pattern exists, what can be inferred and how strongly the evidence supports a conclusion.
That is the next guide.
Continue the Primary 6 Science Learning Guide series
- Guide 1: Concepts, Systems, Interactions & Energy
- Guide 3: Data, Graphs, Diagrams & Evidence
- Guide 4: PSLE Questions, Explanations, Error Analysis & Revision
Return to the Primary 6 Science Learning Hub.
The quiet return
A fair test is not a ritual. It is an argument built into a method: if this is the main condition I change, and the other relevant conditions stay comparable, then a systematic difference in the measured outcome becomes meaningful evidence about the relationship I am testing.
Ask a clean question. Change one scientific job. Measure the right outcome. Control the alternatives. Let the evidence decide.
