Reading a graph is only half the skill. Primary 6 pupils also need to understand how scientific data should be organised so that another person can see the relationship clearly. A poor table can hide a pattern. A poor graph scale can exaggerate a difference. Missing units can make correct measurements unusable.
This guide develops construction of tables, graphs and scientific data displays for PSLE Science. The focus is not artistic presentation. It is disciplined representation: choosing the form that makes the scientific relationship easiest to inspect.
Return to the Primary 6 Science Learning Hub.
The display rule
QUESTION → VARIABLE TYPES → RAW DATA → TABLE → SUMMARY IF NEEDED → GRAPH/DISPLAY → PATTERN → CONCLUSION.
This is an eduKate reasoning routine, not an official SEAB formula.
Part I — Start with the scientific question
Question: “How does lamp distance affect bubble count?”
This immediately suggests:
- changed variable = lamp distance;
- measured variable = bubble count;
- table needs both quantities;
- graph should show how bubble count changes with distance.
Part II — Put variables in logical places
A common scientific table places the changed variable in the first column and measurements in later columns.
| Lamp distance (cm) | Bubble count, Trial 1 | Trial 2 | Trial 3 | Average bubble count |
|---|---|---|---|---|
| 10 | 41 | 42 | 43 | 42 |
| 20 | 30 | 32 | 31 | 31 |
| 30 | 20 | 19 | 21 | 20 |
Part III — Headings need quantities and units
Weak heading: “Distance”.
Stronger: “Lamp distance (cm)”.
Weak heading: “Temperature”.
Stronger: “Water temperature (°C)”.
The quantity and unit tell the reader exactly what each number represents.
Part IV — Keep raw data visible
If repeated trials were performed, record them individually before calculating an average.
Raw data reveal:
- variation;
- anomalies;
- consistency;
- possible method problems.
An average alone can hide all four.
Part V — Decide whether an average is appropriate
Averages are useful when several measurements represent repeated trials of the same condition.
Do not average values that represent different experimental conditions.
For example, do not average results from 10 cm and 30 cm lamp distances into one number if the purpose is to compare distance effects.
Part VI — Identify categorical versus numerical variables
Examples of categorical changed variables:
- surface type;
- material type;
- insulation type;
- habitat type.
Examples of numerical changed variables:
- distance;
- time;
- temperature;
- load;
- number of cells.
The variable type helps determine the most appropriate display.
Part VII — Bar charts for categories
If comparing separate categories such as Surface P, Q and R, a bar chart can show differences clearly.
Bars should:
- have equal width;
- use a consistent scale;
- be separated when categories are distinct;
- have labelled axes;
- show units.
Part VIII — Line graphs for ordered numerical relationships
If the changed variable is numerical and ordered, such as lamp distance or time, a line graph may show the pattern across values.
The graph should reveal whether the measured outcome:
- increases;
- decreases;
- plateaus;
- turns;
- changes at different rates.
Part IX — Axis placement
In many school investigations:
- horizontal axis → changed variable;
- vertical axis → measured variable.
This is a useful default when the question follows that structure.
Part X — Scale choice matters
A poor scale can make a graph hard to read.
Choose a scale that:
- uses most of the available plotting area;
- has equal intervals;
- includes all data values;
- is easy to interpret;
- does not distort the pattern.
Part XI — Truncated axes can exaggerate differences
If a vertical axis begins at 90 instead of 0, a change from 95 to 100 can appear visually huge.
Truncated axes are not automatically wrong, but the reader must inspect the scale before judging size.
Part XII — Plot accurately
Use the exact coordinate:
(changed variable, measured variable)
Example: lamp distance 20 cm, average bubble count 31 → plot (20,31).
Do not swap the axes.
Part XIII — Do not connect unrelated categories
Surface P, Q and R are separate categories. Connecting them with a continuous line can imply meaningful intermediate surfaces that were never tested.
The representation should match the variable type.
Part XIV — Error bars are beyond the core requirement, but variation still matters
Primary 6 pupils may not need advanced statistical displays. They should still understand that repeated results can vary.
A table of raw trials is often enough to show that variation clearly.
Part XV — Use graphs to inspect anomalies
An anomalous point can stand out visually.
If three points follow a trend and one lies far away, check the corresponding raw data and method.
Original workshop 1 — spring extension
| Load (units) | Extension (cm) |
|---|---|
| 0 | 0 |
| 1 | 2 |
| 2 | 4 |
| 3 | 6 |
A line graph shows a steady increase over the tested range.
Do not extend the line indefinitely without evidence.
Original workshop 2 — cooling over time
| Time (min) | Cup P temperature (°C) | Cup Q temperature (°C) |
|---|---|---|
| 0 | 80 | 80 |
| 5 | 65 | 70 |
| 10 | 52 | 61 |
Two lines on the same axes allow direct comparison at equal times.
Original workshop 3 — habitat counts
| Habitat | Average insects per 1 m² |
|---|---|
| Shaded | 18 |
| Open | 8 |
A bar chart is appropriate because the habitat types are categories.
Part XVI — Titles should identify the relationship
Weak: “Science Graph”.
Stronger: “Average bubble count at different lamp distances”.
A useful title or caption tells the reader what the data display represents.
Part XVII — Data displays should answer a question
Do not create a graph merely because graphs look scientific.
Choose the display that makes the required comparison easiest to inspect.
Part XVIII — Tables can be better than graphs
If exact readings are important, a table may be best.
If the trend across many values is important, a graph may be best.
If category comparison is important, a bar chart may be best.
Part XIX — Graph construction and fair comparisons
If two lines are compared, make sure:
- same axes;
- same units;
- same time basis;
- same scale;
- same starting reference where relevant.
Part XX — Graph construction and averages
If plotting averages, label the axis or caption accordingly.
Do not pretend an average is a raw trial.
Part XXI — Graph construction and missing values
If a value was not measured, do not invent a point.
Interpolation may estimate between known points only when the question permits it. The display itself should distinguish measured data from inferred values.
Part XXII — Graph construction and uncertainty
Visual neatness does not remove uncertainty.
A perfectly drawn graph based on a confounded experiment still represents weak causal evidence.
Part XXIII — The DISPLAY test
- D — Data: raw values correct?
- I — Independent/changed variable: placed logically?
- S — Scale: equal and sensible?
- P — Plot: accurate coordinates or bars?
- L — Labels: quantities and units?
- A — Appropriate form: table, bar or line?
- Y — Yield: does the display reveal the relationship?
This is an eduKate teaching mnemonic.
Part XXIV — Common construction errors
- Missing units.
- Unequal scale intervals.
- Axes reversed.
- Points plotted inaccurately.
- Categories connected with a line unnecessarily.
- Raw trials hidden before checking anomalies.
- Different conditions averaged together.
- Graph title too vague.
- Missing values invented.
Part XXV — Why construction matters for PSLE
Building a data display forces the pupil to identify variables correctly, organise measurements, maintain units and notice the pattern before writing a conclusion. Representation construction is therefore a reasoning task, not just presentation.
Where to connect
- Data, Graphs, Diagrams & Evidence
- Practical Planning, Data Recording & Conclusions
- Primary 4 Constructing Tables, Graphs & Data Displays
Retrieval checklist
- I build a table from the scientific question.
- I label quantities and units.
- I keep raw trials visible.
- I average only comparable repeated trials.
- I distinguish categorical and numerical variables.
- I choose bar charts and line graphs appropriately.
- I select a sensible scale.
- I plot accurately.
- I recognise how display choices can distort interpretation.
- I can construct a data display that makes the scientific relationship visible.
The quiet return
A good scientific display is a transparent window onto the evidence. It should not decorate the data or hide uncertainty. It should make the variable, measurement and pattern easier to see.
Organise the data. Label the quantities. Choose the right display. Plot honestly. Let the pattern appear without distortion.
Return to the Primary 6 Science Learning Hub.
