Two pupils want to test whether a condition changes a plant.
Pupil A measures one plant before and after.
Pupil B uses two separate plants, one for each condition.
Which method is fairer?
The answer is not always the same.
The same specimen controls natural differences well, but it can suffer carryover. Different specimens avoid carryover, but they bring natural variation.
This guide belongs to the Primary 4 Science Learning Hub. It develops a key Primary 4 investigation decision: whether to compare one specimen with itself or compare separate similar specimens.
The goal is not formal experimental design terminology. It is to help a learner see the trade-off clearly enough to choose a method that matches the question.
Quick Answer: The Specimen-Choice Loop
QUESTION → DOES THE FIRST TEST CHANGE THE SPECIMEN? → CAN THE SPECIMEN RETURN TO THE START? → HOW LARGE IS NATURAL VARIATION? → SAME SPECIMEN OR PARALLEL SPECIMENS → RECORD BASELINES → COMPARE → STATE THE LIMIT
This is an eduKate teaching routine, not an official MOE marking formula.
Wait, What? “Use the Same Thing” and “Use Two Similar Things” Can Both Be Good Advice
Teachers often tell pupils to keep things the same.
That is sensible—but “same specimen” is only one way to reduce variation.
Sometimes the same specimen cannot be reused fairly.
Examples:
- a plant is damaged;
- a material becomes wet;
- an object remains warm;
- a sponge is stretched and does not return fully;
- liquid volume has been lost;
- a chemical or residue remains on a surface.
In those cases, using the same specimen may create more unfairness, not less.
1. Before–after design: one specimen serves as its own baseline
Example:
Measure Plant A on Day 1 and again on Day 7.
Advantages:
- same plant identity;
- starting height is known;
- natural differences between plants are removed from this comparison;
- change is easy to calculate.
Weakness:
time passes, and many other things may also change.
The learner must not claim that every observed change was caused by one chosen factor unless the method isolates it.
2. Parallel design: compare different similar specimens
Example:
Plant A receives Condition 1.
Plant B receives Condition 2.
Advantages:
- both conditions can happen at the same time;
- one treatment does not carry into the other;
- irreversible effects stay separate.
Weakness:
Plant A and Plant B may begin different.
3. Baselines make the trade-off visible
Suppose Plant A starts at 12 cm.
Plant B starts at 18 cm.
After one week:
- Plant A = 16 cm;
- Plant B = 20 cm.
Final height alone says Plant B is taller.
Change in height says:
- Plant A grew 4 cm;
- Plant B grew 2 cm.
The starting state changes the interpretation.
4. Same specimen is strong when the property can be re-measured without changing the system unfairly
Good same-specimen examples include:
- height of one plant over time;
- temperature of the same water at several time points;
- shadow width of the same object at different positions when geometry is reset correctly;
- volume of the same liquid before and after transfer when none is lost;
- mass of the same object before and after a safe reversible change if the question requires it.
The same specimen helps because identity remains constant.
5. Same specimen is weak when the first treatment changes the next starting state
Example:
A cloth is soaked for Trial 1, then immediately used for a “dry cloth” Trial 2.
That cannot test dry and wet conditions fairly.
The first trial changes the second starting state.
Use a fresh comparable cloth or restore the relevant condition fully if the material permits.
6. Different specimens are strong when treatment effects are irreversible
If a plant has roots damaged, it cannot become undamaged for Trial 2.
If a material is permanently bent or torn, it cannot serve as its original state later.
Separate comparable specimens may be necessary.
Then record important baselines so natural variation is visible.
7. Similar does not mean identical
Two cups can be same model but still differ slightly.
Two leaves can differ in age.
Two plants can differ in height, leaf count or root development.
Two sponges can absorb slightly different amounts.
The method should make the relevant comparability explicit.
8. Which baseline matters?
Baseline depends on the property:
- height for growth;
- starting temperature for cooling;
- starting volume for liquid transfer;
- starting shadow geometry for Light;
- initial mass for mass change;
- initial visible wilting state for plant observations.
Do not collect unrelated baselines simply because more numbers look scientific.
9. Parallel specimens should be assigned fairly
Do not place the healthiest plants in one condition and weakest in another.
Use a pre-decided, fair assignment rule where practical.
This connects to Sampling, Representative Cases and Avoiding Cherry-Picking.
10. Same specimen can reduce variation but hide time effects
Measure one plant before and after a week.
The week itself includes:
- growth;
- watering;
- light changes;
- temperature changes;
- ordinary development.
If the question asks only how much the plant changed, that is fine.
If the question asks which one factor caused the change, more control is needed.
11. Parallel specimens can reduce time-order effects
Condition A and Condition B can occur simultaneously.
This helps when the room itself changes over time.
Example:
If one cooling material is tested in the morning and another in a much warmer afternoon, time-of-day may become another variable.
Running comparable cups side by side can reduce that problem.
12. Same specimen + reversible reset can be powerful
Light example:
Use the same card at 10 cm, 20 cm and 30 cm from the source, returning to the defined baseline between checks.
The object itself does not need to be replaced if:
- its size and orientation remain unchanged;
- the source and screen stay fixed;
- distance reference points are consistent.
This reduces specimen variation.
13. Same specimen + irreversible change can be misleading
Plant root damage example:
Trial 1 damages 20% of roots.
Trial 2 attempts to test “undamaged” condition on the same plant.
Impossible.
The original state cannot be restored.
A separate comparable plant is needed for the undamaged reference.
14. Different specimens need enough similarity for the question
For a simple classroom comparison, similar specimens might mean:
- same cup model;
- same material and dimensions;
- same plant type and reasonably similar starting size;
- same volume of water prepared in separate containers;
- same kind of card cut to the same dimensions.
“Looks similar” should become “similar in the properties relevant to the test”.
15. Paired comparison
A useful middle route is to pair similar specimens before assigning conditions.
Example:
Four plants are arranged into two pairs based on similar starting height.
Within each pair, one receives Condition A and one Condition B.
Primary 4 pupils do not need formal matched-pair statistics.
The simple idea is enough:
start with comparable cases before comparing outcomes.
16. Same specimen vs different specimens in Matter
Question:
Does transferring water to a different container change volume?
Same water sample is useful because the quantity being tracked is the same liquid.
But if water spills, the same-sample advantage is lost because the amount has changed.
Then restart with a known volume rather than pretending the original sample remains intact.
17. Same specimen vs different specimens in Heat
Question:
Which wrapping reduces cooling more?
Using the same cup sequentially can create thermal carryover.
Using two comparable cups side by side avoids order effects but introduces cup-to-cup variation.
Good design chooses the smaller problem and records relevant baselines.
18. Same specimen vs different specimens in Light
Using the same card is usually convenient because its size stays constant.
But if the card bends, becomes wet or changes orientation permanently, a fresh matching card may be needed.
Again, specimen identity matters only through relevant properties.
19. Same specimen vs different specimens in plant observations
For growth over time:
same plant is often best.
For irreversible treatment comparisons:
separate comparable plants may be better.
For generalising to many plants:
several plants may be required.
One design does not solve every plant question.
20. Measurement history belongs to the specimen
If one plant has already been moved, trimmed, watered differently or measured destructively, its history matters.
Record relevant history rather than treating every observation as a fresh start.
Batch 23’s Order Effects, Carryover and Resetting Between Trials develops this history problem.
21. The specimen-choice decision table
| Question | Same specimen favoured | Different similar specimens favoured |
|---|---|---|
| Is natural variation large? | Often | Need matching/baselines |
| Is treatment irreversible? | Usually no | Often |
| Can starting state be restored? | Often | May not be needed |
| Must conditions occur at same time? | Harder | Often easier |
| Is change over time the question? | Often | Not always |
22. Avoid “same specimen is always more controlled”
It controls identity.
It may worsen carryover.
Scientific design is about the whole comparison, not one slogan.
23. Avoid “different specimens are always independent”
They may share:
- same environment;
- same measurement error;
- same biased selection;
- same apparatus problem.
Different identity does not remove every shared weakness.
24. Original Specimen-Choice Casebook
Case 1 | Plant growth over seven days
Choice: same plant for before–after height.
Case 2 | Root damage vs undamaged
Choice: different comparable plants because damage is irreversible.
Case 3 | Shadow position changes
Choice: same card if geometry resets correctly.
Case 4 | Cooling materials
Choice: often comparable parallel cups to avoid thermal carryover.
Case 5 | Volume before and after transfer
Choice: same liquid sample if none lost.
Case 6 | Wet vs dry sponge
Choice: separate comparable sponges if drying fully between trials is impractical.
Case 7 | Same specimen but altered baseline
Problem: before–after comparison starts from changed state.
Case 8 | Different specimens with unequal starts
Problem: natural variation confounds final comparison.
Case 9 | Parallel cups from same model
Strength: simultaneous conditions reduce time-order effects.
Case 10 | Different plants selected by appearance
Problem: biased assignment.
Case 11 | Same plant through normal development
Strength: identity controlled; conclusion should stay descriptive unless cause isolated.
Case 12 | Fresh replacement after spill
Strength: restores quantity, but new container/specimen comparability must be checked.
25. Original Practice Set
- What is the main advantage of using the same specimen?
- What is the main danger of using the same specimen?
- What is the main advantage of separate specimens?
- What is the main danger of separate specimens?
- Why are baselines important?
- When is same-plant before–after useful?
- When is same-plant before–after inappropriate?
- Why can side-by-side cup testing reduce order effects?
- Why can two cups still differ?
- What makes specimens “similar enough”?
- Why does irreversible change favour parallel specimens?
- How does sampling connect to specimen choice?
- How does carryover connect to specimen choice?
- Why can final value alone be misleading?
- Why might change from baseline be more useful?
- Does same specimen prove cause?
- Does different specimen guarantee fairness?
- What is a simple paired-comparison idea?
- What should happen if same-sample water is spilled?
- Write one sentence justifying a specimen choice.
26. Practice Answers
1. It controls natural differences between specimens.
2. The first trial can change the specimen and affect the next.
3. Conditions can be tested without carryover from one treatment to the other.
4. Natural variation may create starting differences.
5. They reveal whether specimens began comparably and allow change to be calculated.
6. Tracking growth or another continuing property over time.
7. When the first treatment creates irreversible change and the next trial requires the original state.
8. Both conditions can occur under the same time/environment period.
9. Small manufacturing or condition differences remain.
10. Similarity in the scientifically relevant properties, not just appearance.
11. The original specimen cannot return to the comparison baseline.
12. Specimens should be selected fairly rather than chosen to favour an outcome.
13. Carryover determines whether same-specimen reuse remains fair.
14. Different starting states can create final differences.
15. It compares how much each specimen changed relative to its own start.
16. No. Other time-varying factors may explain the change.
17. No. Specimens, apparatus or assignment can differ.
18. Pair specimens with similar baselines, then assign different conditions within each pair.
19. Restart with a known amount rather than pretending the original sample is unchanged.
20. Example: “Separate comparable cups are used so both wrapping conditions can begin at the same time without thermal carryover from the first test.”
27. The Specimen-Choice Diagnostic
| If the learner… | Likely weak link | Repair |
|---|---|---|
| always chooses same specimen | carryover awareness | ask if treatment is reversible |
| always chooses different specimens | natural variation | record baselines |
| compares finals only | baseline reasoning | calculate change where relevant |
| selects extremes | assignment bias | predefine selection rule |
| ignores time of testing | order/environment effects | consider parallel conditions |
28. A 40-Minute Specimen-Choice Lesson
Minutes 1–5: sort same-specimen vs parallel-specimen scenarios.
Minutes 6–10: identify natural variation risks.
Minutes 11–15: identify carryover risks.
Minutes 16–20: compare final values with baseline changes.
Minutes 21–25: design a fair plant comparison.
Minutes 26–30: design a fair cooling comparison.
Minutes 31–35: justify one specimen choice.
Minutes 36–40: transfer to Matter or Light.
29. What Parents and Tutors Can Ask
- “What advantage do we get from using the same specimen?”
- “What might carry over?”
- “Can the original state be restored?”
- “How different could two fresh specimens be?”
- “Which baseline matters?”
- “Would side-by-side testing be stronger?”
- “Does final value or change answer the question?”
- “What limitation should the conclusion keep?”
30. Continue Batch 23
- Order Effects, Carryover and Resetting Between Trials
- Treatment Duration, Observation Time and Measurement Intervals
- Rate of Change, Change Per Unit Time and Fair Time Comparisons
The Quiet Return
The learner no longer asks only, “Should I use the same object?”
The question has become more scientific:
“Which choice creates the fairest starting conditions for the relationship I am trying to test?”
