A child asks, “Why does this happen?” Science begins there—but a useful investigation needs one more step.
The question must become specific enough that evidence can answer it.
Scientific curiosity becomes powerful when the learner can turn a broad wonder into a testable question with a clear changed condition, a measurable result and a fair way to compare.
This guide develops question formation inside the Primary 4 Science Learning Hub.
Quick Answer: What Makes a Question Testable?
A useful testable question usually has:
- a factor that can be changed;
- a result that can be measured or observed;
- a comparison that can be carried out safely;
- conditions that can be kept reasonably similar;
- a scope small enough for evidence to answer.
A useful eduKate routine is:
WONDER → NARROW → CHANGE → MEASURE → PREDICT → TEST → REVISE
This is a teaching routine, not an official MOE marking formula.
Curiosity Is Broader Than a Testable Question
“Why are shadows interesting?” is a broad curiosity question.
“How does the distance between an object and a torch affect shadow width?” is testable.
The first can lead to discussion.
The second can lead to a controlled investigation.
Broad Question → Narrow Question
Broad:
“What affects cooling?”
Narrow:
“How does wrapping material affect the temperature decrease of hot water over 15 minutes?”
The narrow version tells us what changes, what is measured and over what time.
Changed Condition and Measured Result
A testable question often contains two roles:
- changed condition: what the learner deliberately varies;
- measured result: what the learner records.
Question:
“How does object–torch distance affect shadow width?”
Changed condition: object–torch distance.
Measured result: shadow width.
Do Not Ask Two Investigations at Once
Weak question:
“How do object distance and object size affect shadow width?”
This may be a valid larger study, but it is difficult for a simple Primary 4 fair test because two factors are being changed.
Better:
- first test distance;
- then separately test object size.
Hypothesis: A Reasoned Expected Relationship
At Primary 4 level, a hypothesis can be taught simply as an expected relationship supported by a reason.
Example:
“If the object is moved farther from the torch while the screen remains fixed, the shadow is expected to become smaller because the blocked-light geometry changes.”
The hypothesis is not a fact until tested.
Prediction vs Hypothesis
A hypothesis states an expected relationship.
A prediction applies that relationship to a particular case.
Hypothesis: “Foam wrapping will reduce cooling more than no wrapping.”
Prediction: “After 15 minutes, the foam-wrapped cup should show a smaller temperature decrease.”
A Hypothesis Can Be Wrong and Still Be Useful
If evidence does not support the prediction, the learner should not hide the result.
Instead ask:
- Was the question fair?
- Were the measurements reliable?
- Did another variable change?
- Does the hypothesis need revision?
The point is to test the idea, not protect it.
Original Case 1: Light
Wonder: “Why do shadows change size?”
Testable question: “How does object–torch distance affect shadow width?”
Hypothesis: “Greater object–torch distance will produce smaller shadow width under the same arrangement.”
Evidence: measure shadow width at several distances.
Original Case 2: Heat
Wonder: “Which material keeps water warm?”
Testable question: “How does wrapping material affect temperature decrease over 15 minutes?”
Hypothesis: “A poorer conductor such as foam will reduce the temperature decrease.”
Evidence: compare temperature changes under controlled conditions.
Original Case 3: Plants
Wonder: “Do roots really matter?”
Testable classroom question: “How does severe root damage affect wilting under otherwise similar conditions?”
Hypothesis: “A plant with severe root damage will wilt more because it is less able to absorb sufficient water.”
Evidence: compare similar plants under carefully controlled and ethically appropriate conditions.
Original Case 4: Matter
Wonder: “Does water volume change when the container changes?”
Testable question: “Does 100 mL of water still measure 100 mL after being poured into a differently shaped container without loss?”
Hypothesis: “The volume remains 100 mL because a liquid has fixed volume when none is added or removed.”
Evidence: measure before and after.
Questions Must Be Safe
A testable question is not automatically a suitable classroom investigation.
Do not create dangerous heat, electrical, chemical or biological procedures merely because they are measurable.
Scientific curiosity operates inside safety boundaries.
Questions Must Be Measurable or Observable
“Which shadow is nicer?” is subjective.
“Which shadow is wider?” is measurable.
“Which plant looks happier?” is vague.
“Which plant has more drooping leaves under a defined observation scale?” is more operational.
Operational Definitions
If the result is “wilting”, define how it will be judged.
Example teaching scale:
- 0 = leaves firm;
- 1 = slight drooping;
- 2 = many leaves drooping;
- 3 = severe drooping.
This is an eduKate example, not an official MOE scale.
The principle is that the result should be observable consistently.
Questions Must Be Narrow Enough
“How does everything affect plant growth?” is not practical.
“How does water amount affect plant height over one week?” is narrower.
A narrow question does not mean unimportant. It means answerable.
Questions Can Be Comparative
“Which wrapping results in the smaller temperature decrease?”
This is a comparison question.
It can still be scientific if starting conditions and timing are controlled.
Questions Can Be Descriptive
Not every scientific question must test cause.
“How does water temperature change over 20 minutes in this room?”
This asks for a pattern.
The learner records observations without necessarily changing one variable deliberately.
Questions Can Be Classification-Based
“Which of these substances behave as solids, liquids or gases under the classroom conditions?”
This is answered by criteria rather than a causal experiment.
Science includes more than fair tests.
Questions Can Be Model-Based
“Which digestive organ is represented by X if it comes after the stomach and absorbs digested food?”
This is not an experiment. It is a model-inference question.
The broader lesson is that not every scientific question requires hands-on testing.
Good Questions Reveal Variables
Weak: “What happens to the shadow?”
Better: “How does object–torch distance affect shadow width?”
The better question identifies:
- changed variable;
- measured variable;
- relationship to investigate.
Question Quality Check
Before testing, ask:
- Can I change the factor safely?
- Can I measure the result?
- Can I keep important other conditions comparable?
- Can I complete the test with available apparatus?
- Will the evidence actually answer the question?
Original Weak-to-Strong Question Set
Weak: “Do materials matter?”
Stronger: “Which material keeps water warm?”
Testable: “How does wrapping material affect the temperature decrease of equal volumes of water over 15 minutes?”
Each step narrows ambiguity.
Hypothesis Should Match the Question
Question: “How does distance affect shadow width?”
Weak hypothesis: “Light travels in straight lines.”
That is a model statement, not an expected relationship between the variables.
Better hypothesis: “As the object is moved farther from the torch under the same set-up, shadow width will decrease.”
Reasoning Behind the Hypothesis
A strong hypothesis can include the scientific reason.
“The foam-wrapped cup is expected to cool less because foam is a poor conductor of heat and reduces heat transfer to the cooler surroundings.”
The reason shows that prediction comes from a model rather than guessing.
Evidence Can Support, Fail to Support or Challenge
After testing, avoid forcing a yes/no mindset.
Evidence can:
- support the hypothesis;
- fail to show a clear difference;
- challenge the hypothesis;
- suggest the method was too weak to decide.
These are all useful scientific outcomes.
Original Data Example
| Wrapping | Start | After 15 min |
|---|---|---|
| Cloth | 70°C | 57°C |
| Foam | 70°C | 61°C |
Temperature decrease:
- cloth = 13°C;
- foam = 9°C.
The evidence supports the hypothesis that foam reduces cooling more than cloth under these conditions.
Hypothesis and Confidence
One result can support a hypothesis modestly.
Repeated consistent controlled results can increase confidence.
Primary 4 pupils do not need advanced statistics to understand that stronger evidence supports stronger confidence.
Unexpected Results
If the foam cup cools more in one trial:
- check starting conditions;
- check thermometer use;
- check wrapping thickness;
- repeat;
- consider whether the hypothesis needs revision.
Do not erase unexpected evidence.
Curiosity Questions That Are Not Yet Testable
“Why is Science useful?”
“How do plants know what to do?”
“Why does light exist?”
These can be valuable questions for discussion and future study. Not every curiosity question must be forced into a Primary 4 experiment.
Questioning Without Over-Teaching
A good Primary 4 question should stay inside the learner’s current model unless the purpose is clearly enrichment.
Do not require advanced plant transport tissues or formal optics equations to make the question seem more scientific.
Depth can come from reasoning, not only advanced vocabulary.
Common Question-and-Hypothesis Errors
- question too broad;
- two changed variables hidden in one question;
- result not measurable;
- hypothesis merely repeats a fact;
- prediction has no scientific reason;
- unsafe design;
- hypothesis treated as something that must be proven right;
- unexpected results ignored;
- question cannot be answered by the planned evidence.
Original Practice Set
Question 1
What two variable roles should a simple testable question usually reveal?
Question 2
Why is “What affects shadows?” too broad for one simple fair test?
Question 3
What is the difference between a hypothesis and a prediction?
Question 4
Why can a wrong hypothesis still be scientifically useful?
Question 5
Why is “Which plant looks happier?” weak as a measurement question?
Question 6
What should happen if evidence does not support the hypothesis?
Question 7
Must every scientific question be an experiment?
Question 8
What makes a question appropriately narrow?
Practice Answers
1. Changed condition and measured/observed result.
2. Many factors could change at once, making the question difficult to isolate.
3. A hypothesis states an expected relationship; a prediction states what should happen in a specific case.
4. Testing it can reveal that the model or expectation needs revision.
5. “Happier” is subjective and not operationally defined.
6. Check method and evidence, then revise the explanation or hypothesis if needed.
7. No. Science also uses classification, models, observation and descriptive questions.
8. It focuses on a manageable relationship that available evidence can answer.
The Question-Design Diagnostic
| If the learner… | Likely weak link | Repair |
|---|---|---|
| Asks huge questions | Scope | Narrow to one relationship |
| Changes two factors | Variable control | Split into separate tests |
| Cannot measure outcome | Operational definition | Define observable result |
| Guesses prediction | Model connection | Add scientific reason |
| Protects hypothesis | Evidence mindset | Treat testing as possible revision |
A 25-Minute Question-Design Lesson
Minutes 1–5: convert broad wonders into narrower questions.
Minutes 6–10: identify changed and measured variables.
Minutes 11–15: write one hypothesis with reason.
Minutes 16–20: decide what evidence would test it.
Minutes 21–25: inspect one weak question and redesign it.
What Parents and Tutors Can Ask
- “What exactly are you changing?”
- “What will you measure?”
- “What do you expect to happen?”
- “Why?”
- “What evidence would challenge your idea?”
- “Can this be tested safely?”
- “Is your question small enough to answer?”
Continue the Primary 4 Science Series
- Primary 4 Science Learning Guide | Claims, Evidence and Reasoning
- Primary 4 Science Learning Guide | Constructing Tables, Graphs and Data Displays
- Primary 4 Science Learning Guide | Scientific Communication and Presenting Findings
For investigation design, use Procedure Writing and Investigation Design.
The Quiet Return
Science does not begin with the answer.
Start with curiosity. Narrow the question. Identify what changes and what can be measured. Predict from a model. Test fairly. Then let the evidence improve the question you ask next.
