How to Read a PSLE Science Graph Whose Axis Does Not Start at Zero Without Exaggerating the Difference

Wait, What? The Bars Look Huge — but the Numbers Are Only Four Units Apart

Imagine a bar graph comparing two temperatures. Bar A reaches 96°C. Bar B reaches 100°C. The vertical axis begins at 90°C instead of 0°C.

On the page, Bar B may look almost twice as tall as Bar A above the visible baseline. But the actual temperature difference is only 4°C.

The graph is not necessarily wrong. The danger is reading the visual gap as though it were the same thing as the numerical difference.

When a graph axis does not start at zero, trust the labelled scale and actual values before trusting how dramatic the picture looks.

This is an important PSLE Science data-reading skill because the scientific conclusion must come from the measured quantities, not from the apparent height, length or steepness created by the visible plotting window.

Quick Answer

Use this route whenever a graph axis begins above zero, includes a break, or shows only a narrow range:

READ THE AXIS START → READ THE SCALE INTERVAL → IDENTIFY THE UNIT → RECOVER THE ACTUAL VALUES → FIND THE ACTUAL DIFFERENCE OR CHANGE → DESCRIBE THE SCIENTIFIC RELATIONSHIP → IGNORE EXAGGERATED VISUAL IMPRESSION → CONNECT TO THE RELEVANT CONCEPT → CHECK THE CONCLUSION AGAINST THE NUMBERS.

A restricted axis can be useful when scientists or exam writers want small differences to be visible. But it changes how large the differences look. The learner must separate appearance from measurement.

The Exact PSLE Science Learning Job This Guide Owns

This guide owns one PSLE Science learner job: how a Primary 5 or Primary 6 learner reads a graph whose axis does not begin at zero, recovers the true numerical meaning of the plotted values and prevents a restricted visual scale from exaggerating or distorting the scientific conclusion.

It does not replace the general graph-reading guide, the units-and-resolution guide, or the concept page behind the data. It owns the narrower visual-reasoning problem created by a non-zero baseline or restricted plotting range.

The Current 2026 PSLE Science Frame

For examination from 2026, Standard PSLE Science assesses attainment in the 2023 Primary Science syllabus. The official assessment objectives include application of scientific facts, concepts and principles, interpretation and analysis of information, evaluation of observations and information, and communication of explanations and reasoning.

Graphs are therefore not decorative. A learner must read the representation accurately, identify what the values mean, and connect those values to the scientific relationship being tested.

Why Would a Graph Axis Start Above Zero?

A graph may use a restricted range because the important values are all close together.

Suppose four temperatures are 95°C, 96°C, 98°C and 100°C. If the axis runs from 0°C to 100°C, the bars may look almost the same height. If the axis runs from 90°C to 100°C, the differences become easier to see.

That can be useful. It can also make a small difference look visually large.

The correct response is not “all non-zero axes are misleading”. The correct response is:

Read the scale before interpreting the picture.

The Three Quantities You Must Keep Separate

Quantity	Meaning	Example
Actual value	The measurement represented by the top of the bar or point	96°C
Actual difference	The numerical difference between two values	100°C − 96°C = 4°C
Visible height above displayed baseline	How far the bar appears to extend within the shown plotting window	6 scale units above 90 versus 10 above 90

The third quantity is a property of the drawing window. It is not automatically the scientific quantity you should compare.

Worked Example 1 — Temperatures of Two Cups

Two cups are measured after heating.

Cup	Temperature
P	58°C
Q	62°C

The graph axis runs from 55°C to 65°C.

Above the visible baseline, P extends 3°C while Q extends 7°C. Q therefore looks more than twice as tall within the displayed region.

But scientifically:

P = 58°C.
Q = 62°C.
Actual difference = 4°C.
You cannot say Q is “more than twice as hot” from the bar heights.

The correct comparison uses the temperature values and the question conditions, not the height of the bars above 55°C.

Worked Example 2 — Plant Height With a Narrow Axis

Plant A is 101 cm tall. Plant B is 103 cm tall. A graph begins at 100 cm.

Visually, Plant B’s bar may appear three times the height of Plant A above the displayed baseline:

A: 1 cm above the shown baseline.
B: 3 cm above the shown baseline.

But Plant B is not three times as tall as Plant A. The actual heights are 101 cm and 103 cm.

The scientific conclusion is that B is 2 cm taller under the measured conditions.

Worked Example 3 — A Line Graph With a Restricted Vertical Range

A line graph records a temperature change from 94°C to 97°C. The vertical axis shows only 92°C to 98°C.

The line may look very steep because the entire vertical range is only 6°C.

To interpret it:

Read the first value: 94°C.
Read the second value: 97°C.
Find the actual change: +3°C.
Check the time interval.
Only then discuss rate or scientific mechanism.

Do not compare steepness across two different graphs unless the axes use comparable scales. A line can look steeper simply because the graph uses a narrower vertical or horizontal range.

Worked Example 4 — Same Data, Two Different Pictures

Suppose two objects have measured masses of 98 g and 102 g.

Graph A uses an axis from 0 g to 110 g. Graph B uses an axis from 95 g to 105 g.

The data are identical. Graph B makes the difference visually dramatic. Graph A makes it look small.

Neither picture changes the measurements:

Object X = 98 g.
Object Y = 102 g.
Difference = 4 g.

This is why a careful learner reads numbers before impressions.

Worked Example 5 — A Broken Axis

Some graphs show a zigzag or break symbol on an axis to indicate that part of the scale has been omitted.

If the axis jumps from 0 to 80 and then continues 80, 85, 90, 95, the visual distance across the break does not represent the missing 80 units in the same way as the later equal intervals.

When you see a break:

read the labels on both sides;
do not count the physical gap as one normal scale interval;
recover the actual numeric value from the labels;
base comparisons on the values, not on bar length.

Worked Example 6 — When the Non-Zero Axis Is Actually Helpful

Three experimental set-ups produce values of 49.8, 50.1 and 50.4 units.

A 0-to-60 axis would compress the differences so strongly that the pattern is difficult to inspect. A 49-to-51 axis may make comparison much easier.

The narrower axis is useful if the reader recognises that it magnifies the visual separation. The graph becomes a microscope for the difference, not a claim that the quantities are enormously far apart.

Bar Height Ratios Are Especially Dangerous With a Non-Zero Baseline

Suppose two bars represent 95 and 100 units, and the graph begins at 90.

Above the visible baseline:

95 is shown as 5 units tall;
100 is shown as 10 units tall.

The second bar appears twice the height of the first, but 100 is not twice 95.

This is why statements such as “twice as large” or “three times as much” must come from the actual values or a valid ratio calculation—not from apparent bar lengths on a truncated axis.

Line Graphs Need a Different Caution

For line graphs, the main risk is often apparent steepness.

Steepness depends on both the vertical scale and the horizontal scale. If one graph compresses time and expands the measured-value axis, the line can look much steeper even when the underlying rate is unchanged.

If the question asks about rate, compare actual change over actual time:

rate comparison depends on change and time, not on how steep the printed line looks across differently scaled graphs.

Do Not Subtract From the Axis Minimum Unless the Question Asks for That

If a bar reaches 98 on an axis beginning at 90, the scientific measurement is 98—not 8.

The 8 is only the displayed height above the chosen baseline. It may help draw the bar, but it is not automatically the measured quantity.

Actual Difference Versus Percentage Difference

Sometimes learners react to a dramatic-looking graph by inventing a percentage claim.

Example: 96 and 100 differ by 4 units. The percentage difference depends on the comparison reference chosen and is not simply “the second bar is 67% taller above the visible baseline”.

At Primary level, unless the question explicitly asks for a percentage, stay with the measured values and scientifically relevant comparison.

The Axis-Reading Protocol

Read the axis title. What quantity is measured?
Read the unit. °C, cm, g, mL, seconds or another unit?
Find the starting value. Does the axis begin at zero?
Check for a break symbol.
Find the scale interval. How much does one grid step represent?
Read each actual plotted value.
Calculate or state the real difference/change.
Describe the relationship using the numbers.
Only then interpret the scientific mechanism.
Ignore visual drama that is not supported by the values.

Observable Failure Signatures

Failure signature	Earliest weak link	Repair
“Bar B is twice Bar A because it looks twice as tall.”	Visible bar height confused with data value.	Read the numerical labels and compare actual values.
“98 means 8 because the axis starts at 90.”	Axis minimum wrongly subtracted from measurement.	Treat 98 as the plotted value; 8 is only display height above baseline.
“The graph is dishonest because it does not start at zero.”	Restricted scale treated as automatically invalid.	Check whether labels and intervals are clear; then interpret cautiously.
“This line is steeper, so the rate is faster.”	Printed slope compared across different axis scales.	Compare numerical change over numerical time.
“The difference is huge.”	Visual impression used instead of magnitude.	State the actual difference with unit.
“The axis break is one interval.”	Break symbol misunderstood.	Read values on both sides of the break.

The Earliest Weak-Link Diagnosis

Quantity: Do I know what the axis measures?
Unit: Did I read the unit?
Baseline: Did I notice where the axis starts?
Interval: Did I decode the scale correctly?
Value: Did I recover the actual measurement?
Difference: Did I compare the numbers rather than the bar lengths?
Rate: If relevant, did I compare change over time?
Mechanism: Did I connect the data to the correct science only after reading the graph accurately?

Misconception Repair — “All Graphs Must Start at Zero”

No. A restricted axis can be legitimate and useful, especially when small differences need to be visible. The learner’s responsibility is to notice the baseline and avoid reading visual ratios literally.

Misconception Repair — “If the Difference Looks Large, the Scientific Effect Is Large”

No. A small numerical difference can fill most of a graph if the displayed range is narrow. Describe the measured magnitude first.

Misconception Repair — “A Bar’s Printed Height Is the Measurement”

The top of the bar aligns with the measurement on the axis. The physical number of centimetres of ink on the page is not the scientific quantity.

Misconception Repair — “Steeper Always Means Faster”

Only after scale is controlled. If axes differ, visual slope cannot be compared directly. Use the actual change and actual time interval.

Question-Reading Protocol for a Non-Zero Axis

Cover the bars or line briefly and read only the axes.
Say the quantity and unit aloud.
State the axis minimum and maximum.
State the value of one interval.
Now uncover the data.
Read each value numerically.
Write the actual difference beside the graph.
Answer the question from the values.
Use the visual shape only as a support after the numbers are understood.

Practice Sequence

Baseline spotting: inspect ten graphs and identify which axes start at zero.
Value recovery: read actual values from restricted axes.
Difference practice: calculate or state real differences.
Same data, different scales: compare how the picture changes while the data stay constant.
Bar-ratio traps: reject “twice as large” claims based only on printed bar height.
Slope comparison: compare rates only after checking both axes.
Mixed transfer: practise temperature, mass, height, volume, time and unfamiliar quantities.

Unfamiliar Transfer Challenge

A mystery sensor gives readings of 502 units and 506 units. The graph axis runs from 500 to 508.

The second bar looks three times as tall above the visible baseline:

502 is 2 units above 500.
506 is 6 units above 500.

Can you say the second measurement is three times the first?

No. The actual values are 502 and 506. The actual difference is 4 units. The visual 2-to-6 ratio comes from the chosen plotting baseline, not from the measurement ratio.

Delayed Independent Return

Three to five days later, take a fresh graph and answer without notes:

What does each axis measure?
Where does each axis begin?
What does one interval represent?
What are the actual plotted values?
What is the real difference or change?
Does the picture exaggerate or compress that difference?
What scientific conclusion is justified?
What claim would be visually tempting but numerically wrong?

The Graph-Scale Receipt

I read the axis labels before the data.
I noticed whether the axis started at zero.
I checked for a break.
I decoded the interval correctly.
I recovered the actual values.
I used the actual numerical difference.
I did not infer ratios from bar heights on a non-zero baseline.
I did not compare printed slopes across differently scaled graphs.
I connected the data to the scientific mechanism only after the graph was read correctly.
I kept the conclusion within the evidence.

Evidence and Model Limits

Research on graph perception shows that truncated or non-zero baselines can influence how large differences appear to viewers, especially in bar graphs. That does not mean every restricted axis is deceptive. It means the reader must rely on labelled values and scale rather than visual magnitude alone.

PSLE learners do not need advanced data-visualisation theory. They need a durable scientific habit: read the quantity, unit, baseline and interval before judging the size of a difference.

Useful Internal Routes

Parent and Tutor Teaching Guide

A useful teaching move is to redraw the same data twice: once with a zero baseline and once with a restricted baseline.

Ask the child:

“Did the science change, or only the picture?”

Then require three statements:

The actual values are ______ and ______.
The actual difference is ______.
The restricted graph makes the difference look ______, but it does not change the measured values.

For line graphs, show the same two data points on axes with different vertical ranges. Ask whether the apparent steepness changed. Then calculate or compare the actual change over time.

Finally, remove the scaffolding. Give an unfamiliar graph and ask the learner to identify the baseline before reading any bar or line.

Authoritative and Research References

Singapore Examinations and Assessment Board — PSLE Science syllabus, for examination from 2026.
Singapore Ministry of Education — Science Teaching & Learning Syllabus, Primary, 2023.
Truncating Bar Graphs Persistently Misleads Viewers. Used as broader evidence about graph perception, not PSLE marking policy.

The Quiet Ending

A graph can zoom in on a small difference.

That zoom can help you see.

It can also make the difference feel larger than it is.

Read the axis. Recover the values. Let the numbers carry the Science.

Small Group Tutorials