A plant measures 15 cm on Monday.
It measures 15 cm again on Tuesday.
A child writes:
“The plant did not grow.”
Maybe.
But another possibility exists: the plant changed by less than the ruler and method could detect.
Or perhaps its height returned to the same measured value after some intermediate change.
Or perhaps the selected measurement—height—was not the part of the plant that changed.
A result showing no measured change is evidence. It is not permission to claim that nothing happened in every possible sense.
This guide belongs to the Primary 4 Science Learning Hub. It develops a scientific habit that is often missing from school work: treating an expected effect that fails to appear as meaningful evidence rather than as an embarrassment to erase.
The page uses “null result” in a simple classroom sense: a result showing no clear measured difference or change under the tested conditions. It does not teach formal statistical null-hypothesis testing.
Quick Answer: The No-Change Loop
EXPECTED CHANGE → MEASURE → NO CLEAR DIFFERENCE → CHECK TOOL + METHOD → REPEAT / VERIFY → ASK WHAT THE METHOD COULD DETECT → WRITE A BOUNDED CONCLUSION → CHOOSE THE NEXT TEST
This is an eduKate teaching routine, not an official MOE examination formula.
Wait, What? “Nothing Happened” Is Usually Too Big
Suppose two cups both finish at 58°C.
Can we say:
“The wrappings made no difference.”
Not yet.
We need to know:
- starting temperatures;
- water volumes;
- measurement precision;
- time;
- whether the thermometer could distinguish a small difference;
- whether the wrappings were actually different;
- whether repeated trials agreed.
A no-difference observation becomes interpretable only after the measurement system and comparison are understood.
1. What is a no-change result?
A no-change result occurs when the measured value remains the same, or when the difference is too small to establish clearly with the method.
Examples:
- plant measured 15 cm on both days;
- two cups show the same final temperature;
- shadow width appears unchanged after moving the object by a tiny amount;
- liquid volume is 100 mL before and after transfer;
- a repeated observation category remains score 1.
These cases do not all mean the same thing scientifically.
2. No change can be the expected scientific result
Question:
“Does pouring 100 mL water into a different-shaped container change its volume if none is lost?”
Measured result:
- before = 100 mL;
- after = 100 mL.
Here, no measured volume change supports the intended Matter concept.
The no-change result is not disappointing.
It is the evidence.
3. No change can also challenge a prediction
Prediction:
“Moving the card 1 mm closer to the torch will make the shadow visibly larger.”
Measured shadow width appears unchanged.
Possible interpretations:
- the predicted effect may be too small;
- the ruler or shadow boundary may not resolve the change;
- the movement may have been too small;
- the geometry may not have changed as intended;
- the prediction may be wrong for the tested arrangement.
A null result creates a question.
It does not automatically choose the answer.
4. “No detected change” is often better wording
Compare:
Too strong: “There was no change.”
More careful: “No change was detected with this measurement method.”
The second statement keeps the instrument and method visible.
This is especially important when the possible change is near the scale interval.
5. No detected difference is not exact equality
Two leaves both measure 8.4 cm to the supported ruler precision.
They may not be exactly identical in length.
The method supports:
“No clear length difference was detected at this measurement precision.”
It does not support:
“The two leaves have exactly identical length.”
6. The instrument may be insensitive to the change
A ruler with 1 cm markings is used to check one hour of very small plant growth.
Both readings are 15 cm.
The ruler may simply be too coarse for the expected change.
Batch 22’s Small Differences, Scale Intervals and Honest Precision develops that problem.
7. The wrong property may have been measured
A plant changes by producing a new leaf.
Height remains 15 cm.
If the investigation measures only height, the record can correctly show no height change while the plant changed in another way.
Therefore:
no change in one measured property is not no change in the whole system.
8. Final value equality can hide different paths
Temperature starts at 25°C and ends at 25°C.
Can we conclude the temperature never changed?
No.
It may have:
- remained stable;
- increased and returned;
- decreased and returned.
If only start and end were measured, the intermediate path is unknown.
Do not invent it.
9. Zero net change vs no process
Derived change = 0.
This means:
final measured value − initial measured value = 0.
It does not necessarily mean:
- no process occurred;
- nothing moved;
- no heat transferred;
- no intermediate variation occurred.
The data support endpoint equality, not the complete history.
10. A negative result is still a result
Suppose a class expects material A to reduce cooling more than material B.
Repeated trials show no clear difference.
The temptation is to:
- repeat until a preferred difference appears;
- choose only the one trial that supports the prediction;
- change the conclusion to match the expectation;
- hide the no-difference result.
Do not.
The no-difference finding may be the most important result of the investigation.
11. Repeat before making a strong no-effect claim
One no-difference trial can occur by chance or due to measurement noise.
Repeat the planned comparison consistently.
Example:
| Trial | Foam decrease / °C | Cloth decrease / °C |
|---|---|---|
| 1 | 11 | 11 |
| 2 | 10 | 11 |
| 3 | 11 | 10 |
These small alternating differences do not show a stable winner.
A cautious conclusion:
“The three trials did not show a clear consistent difference between foam and cloth under these conditions.”
12. Do not repeat only the unexpected condition
If the expected winner fails once, pupils may want to repeat only that condition until it “works”.
This can create biased evidence.
If the whole comparison needs repeating, repeat the comparison—not only the inconvenient side.
13. Null results can expose poor measurement design
Question:
“Does changing distance by 1 mm affect shadow width?”
Ruler interval = 1 cm.
No difference detected.
The no-change result may say more about the method than the Light relationship.
Next action:
- choose a larger safe change;
- use a more suitable measurement method;
- define the shadow boundary;
- repeat.
14. Null results can reveal a weak manipulation
A learner wraps one cup with one thin sheet of material and leaves another unwrapped.
No clear cooling difference appears.
Possibilities include:
- the material effect is small;
- the wrapping was not applied consistently;
- the time was too short;
- other heat losses dominated;
- measurement was too coarse.
The result should guide method improvement rather than be discarded.
15. Null results can support conservation
Matter example:
- 100 mL before transfer;
- 100 mL after transfer.
No measured volume change under the condition “none lost”.
This is strong conceptual evidence for fixed liquid volume despite changed shape.
No-change results are therefore not all “failed experiments”.
16. Null results can support control stability
Control cup measured at:
- 20°C;
- 20°C;
- 20°C.
If the control is expected to remain stable, no change can strengthen confidence that the environment or measuring process was stable enough for the comparison.
But only within the measurement resolution.
17. No change and missing data are different
Table entry:
0 cm change.
This means the calculation produced zero from measured values.
Blank cell:
no result recorded.
These must never be treated as the same.
Batch 22’s Missing Data, Blanks, Zeros and Exclusions develops that distinction.
18. No change and raw data
Keep both values that produced zero change.
Raw:
- 15 cm;
- 15 cm.
Derived:
- 0 cm change.
Without the raw pair, another learner cannot verify how the zero was obtained.
19. No change and observer expectations
A pupil expects a difference.
The scale is ambiguous.
They may be tempted to choose slightly different readings so a difference appears.
Use fixed reading rules, independent checks and honest precision.
Batch 21’s Observer Expectations, Confirmation Bias and Independent Checks protects this part of the method.
20. No change and instrument checks
Repeatedly identical readings from a stuck instrument are not evidence of a stable phenomenon.
Example:
thermometer display stays at 25°C even when moved between clearly different safe conditions.
Now the no-change pattern suggests an instrument problem.
Use Instrument Checks, Zeroing and Reference Tests.
21. No change and simulations
A virtual model produces the same output after a tiny input change.
Possibilities:
- the model rounds the display;
- the effect is below its output resolution;
- the input did not actually change;
- the model predicts no change under those settings.
Read the model and settings before claiming a real-world null effect.
22. No change and field observations
A shaded surface and sunlit surface are both recorded at 28°C.
Field conditions may differ in many uncontrolled ways.
The result is still worth recording.
But the conclusion should say:
“At the recorded times and locations, no temperature difference was detected with the method used.”
Do not generalise that shade never affects temperature.
23. No change and living things
A plant may show no visible change over one day.
Living processes can continue without a visible difference in the chosen measurement.
The correct conclusion is about the observed property and period.
Example:
“No visible change in measured height was detected over this one-day interval.”
24. No change and heat transfer
Two objects begin at the same temperature and remain at the same measured temperature.
Can we conclude no microscopic energy exchange ever occurred?
That level is beyond the evidence and P4 scope.
The appropriate conclusion is:
“No temperature difference was detected between the objects during the measured period.”
25. No change and control groups
A control or reference condition is sometimes expected not to show the tested change.
If it changes unexpectedly, the experiment may contain another factor.
If it remains stable, that no-change result can strengthen the design.
26. The Null-Result Decision Tree
| Question | If yes | If no |
|---|---|---|
| Was the property measured appropriately? | Continue | Improve measurement |
| Could the tool detect a plausible small change? | Continue | Use better resolution or larger safe contrast |
| Were relevant conditions controlled? | Continue | Repair comparison |
| Did repeats also show no clear difference? | Increase confidence in null result | Investigate variation |
| Does the claim stay within tested conditions? | Report bounded conclusion | Narrow conclusion |
This is an eduKate teaching scaffold, not a statistical test.
27. Original Null-Result Casebook
Case 1 | Same volume after pouring
100 mL before and after.
Meaning: supports unchanged volume if none was lost.
Case 2 | Same plant height
15 cm on consecutive days using coarse ruler.
Meaning: no detectable height change, not proof of no growth process.
Case 3 | Same final temperature
Starts missing.
Meaning: cannot compare cooling amounts.
Case 4 | Same shadow width after tiny movement
Change smaller than method resolution.
Meaning: method may be insensitive.
Case 5 | Same sensor value everywhere
Instrument does not respond to obvious condition changes.
Meaning: check instrument before interpreting phenomenon.
Case 6 | Control unchanged
Expected stability observed.
Meaning: can support control stability.
Case 7 | Prediction fails repeatedly
No consistent material difference across three trials.
Meaning: report no clear difference under tested conditions.
Case 8 | Final value returns to start
25°C → unknown intermediate → 25°C.
Meaning: zero net change; path unknown.
Case 9 | Blank recorded as zero
No measurement taken.
Wrong: missing data is not a null result.
Case 10 | Selected property unchanged
Height stable but leaf count increases.
Meaning: no height change, not no system change.
Case 11 | Field comparison equal
Two stations both 28°C once.
Meaning: no detected difference at those observations; no universal claim.
Case 12 | Simulation plateau
Output unchanged across settings.
Meaning: inspect model range, display resolution and input state.
28. How to write a good no-change conclusion
Use this structure:
PROPERTY + METHOD + TESTED CONDITIONS + NO CLEAR DETECTED CHANGE / DIFFERENCE + LIMIT
Example:
“Across the three trials, the two wrappings produced no clear consistent difference in temperature decrease using the thermometer and timing method used. This result applies to the tested materials and conditions.”
29. Avoid “proves no effect”
One classroom null result rarely proves that an effect cannot exist.
Better language:
- “did not detect”;
- “showed no clear difference”;
- “remained the same at the measured resolution”;
- “no consistent effect was observed in these trials”.
30. What this guide does not teach
Primary 4 pupils do not need:
- null hypotheses;
- p-values;
- statistical significance;
- effect-size calculations;
- power analysis;
- confidence intervals.
The foundational habit is enough:
an absent measured difference deserves honest interpretation, not forced disappearance or forced certainty.
31. Original Practice Set
- What is a no-change result?
- Why is “nothing happened” usually too strong?
- When can no change be the expected result?
- What does “no detected change” mean?
- Does no detected change prove exact equality?
- How can instrument resolution create a null result?
- How can measuring the wrong property create a null result?
- What is zero net change?
- Why should a null result be preserved?
- Why should a whole comparison be repeated rather than only the inconvenient condition?
- What can a null result reveal about method design?
- Why can a stuck instrument create false stability?
- How is a blank different from zero change?
- Why can a control’s no-change result be useful?
- How should a null result be worded?
- Why should a no-difference conclusion be bounded to tested conditions?
- What does repeated no difference add?
- Why might a plant change even if height does not?
- What should happen when an expected effect is not detected?
- Write one careful null-result conclusion.
32. Practice Answers
1. A result showing no clear measured change or difference under the tested conditions.
2. The method may not detect small changes, the wrong property may have been measured, or intermediate changes may be unknown.
3. When the scientific model predicts stability, such as liquid volume remaining constant during a shape change with no loss.
4. The method did not reveal a clear change at its measurement capability.
5. No.
6. A real difference smaller than the scale interval or measurement ambiguity may appear unchanged.
7. Another part of the system may change while the chosen measurement remains stable.
8. The final measured value equals the starting value; it does not prove no intermediate process occurred.
9. It is evidence about the tested conditions and may challenge the prediction or reveal a method limit.
10. Repeating only one side can bias the evidence.
11. The manipulation may be too small, timing too short or measurement too coarse.
12. The instrument may fail to respond even when the phenomenon changes.
13. A blank means no value is recorded; zero change is a derived result from measurements.
14. It can show that the reference condition remained stable enough for the comparison.
15. Use language such as “no clear difference was detected with this method”.
16. The result does not establish what happens for untested materials, times or conditions.
17. It increases confidence that the no-difference pattern is not one accidental trial.
18. Leaf number, leaf position or other properties can change while height does not.
19. Check the method, instrument and prediction, preserve the result and plan the next discriminating test.
20. Example: “Across three trials, no clear difference in measured shadow width was detected between the two positions with this ruler and boundary definition; smaller differences may not be resolved by the method.”
33. The Null-Result Diagnostic
| If the learner… | Likely weak link | Repair |
|---|---|---|
| writes “nothing happened” | claim width | name measured property and method |
| forces expected winner | prediction attachment | preserve null result |
| uses coarse tool | sensitivity | improve resolution or contrast |
| calls blank zero | data semantics | separate missing from null |
| generalises universally | scope control | bound to tested conditions |
34. A 40-Minute Null-Result Lesson
Minutes 1–5: classify expected-change vs expected-stability questions.
Minutes 6–10: compare “no change” vs “no detected change”.
Minutes 11–15: inspect a coarse-scale plant case.
Minutes 16–20: analyse a conservation example.
Minutes 21–25: diagnose a stuck-instrument case.
Minutes 26–30: inspect repeated no-difference trials.
Minutes 31–35: write a bounded null-result conclusion.
Minutes 36–40: propose the next measurement that would reduce uncertainty.
35. What Parents and Tutors Can Ask
- “What exactly failed to change?”
- “Could a smaller change be hidden by the scale?”
- “Did you measure the right property?”
- “Did the instrument respond to a reference condition?”
- “Did repeated trials also show no difference?”
- “What does zero net change actually mean?”
- “Can you say ‘no detected difference’ instead of ‘nothing happened’?”
- “What next test would distinguish the possibilities?”
36. Continue Batch 22
- Small Differences, Scale Intervals and Honest Precision
- Raw Data, Derived Values and the Evidence Trail
- Missing Data, Blanks, Zeros and Exclusions
The Quiet Return
The learner expected a difference.
The difference did not appear.
That is not where Science ends.
Keep the result. Check the method. Ask what the instrument could detect. Repeat fairly. Then write the smallest conclusion the evidence truly earns.
