Secondary 1 G3 Mathematics Tutor Bukit Timah | Misleading Graphs, Axis Scales and Statistical Claims

Did you know? A bar graph showing a change from 95 to 100 can make the increase look as if it doubled—even though the actual increase is just five. How? By starting the vertical scale at 90 rather than zero. The first bar rises five units above the baseline, the second rises ten, so the visible bars appear to have a 1:2 relationship. But the values are 95 and 100, not 5 and 10.

For Secondary 1 G3 Mathematics students in Bukit Timah, a statistics question is not only about extracting a number from a chart. It also asks whether a picture gives readers an accurate sense of the data. A tidy design can still hide an inappropriate axis scale, an unexplained unit, a missing denominator or a conclusion the evidence does not support.

This guide helps students read bar graphs, line graphs and pie charts critically. All numerical scenarios are fictional teaching examples. The target is an everyday habit: read what the chart measures, how it displays the values and what the numbers really justify.

Quick answer: five things to check on any graph

• The title: what group, quantity and period does the graph describe?
• The axes: what does each axis measure and in which units?
• The scale: are equal numerical increments represented by equal distances?
• The denominator: does a percentage refer to a particular known total?
• The conclusion: does the evidence show what the caption or headline claims?

A graph may be useful without being perfect, and a visual choice is not automatically dishonest. But readers should notice when the design exaggerates, conceals or confuses the numerical comparison.

In a Secondary 1 lesson, it is often best to begin with a simple table, then compare two visual versions using the same underlying numbers.

Worked example 1: a truncated bar-chart axis

Imagine fictional survey counts of 95 participants in one year and 100 in the following year.

The actual increase is 100 − 95 = 5 participants. Relative to the starting 95, that is 5/95 × 100%, or approximately 5.26%.

Now draw bars from a baseline of zero. Their heights are 95 and 100, and the visual difference is appropriately modest.

Next begin the visible vertical axis at 90. The portions above the displayed baseline have lengths 5 and 10. The second visible portion looks twice the height of the first, even though the underlying total count increased only about 5.26%.

A reader who mistakes the visual height ratio for the true total-count ratio can come away with a dramatically distorted impression.

Does a vertical axis always have to begin at zero?

Not in every type of graph. For bar charts where bar length represents magnitude, beginning at zero is usually important because the reader compares lengths from a common baseline.

Line graphs often highlight variations over time. A carefully labelled narrowed vertical range can make small changes easier to see without necessarily misrepresenting the data.

The crucial distinction is whether the display makes its scale clear and whether its design suits the mathematical comparison being invited. If a line graph is labelled as showing tiny daily fluctuations, a narrow range may be legitimate. If an ordinary bar chart exaggerates a minor difference by visually doubling its bars, the effect deserves scrutiny.

Teach the student to ask what the intended comparison is before declaring every nonzero starting axis wrong.

Worked example 2: equal spaces need consistent numerical meaning

Imagine a vertical axis with labels 0, 10, 30 and 60 placed at equal physical distances. Without an explicit nonlinear scale, those spacings are not consistent with an ordinary linear axis.

On a linear axis, moving one centimetre up the graph should represent the same numerical increase everywhere. If one centimetre means ten units near the bottom and thirty units near the top, a straight visual comparison becomes unreliable.

A pupil may read a point as halfway between 30 and 60 and give 45. That interpolation depends on the axis interval actually being linear and correctly spaced.

To check a graph, compare at least three labelled tick marks, not just the first two. Ask what one grid step represents at different heights.

Worked example 3: graph units change the meaning

A line graph is labelled “distance” on its vertical axis, but the numbers have no unit. Are they metres, kilometres or centimetres?

Suppose a point has vertical value 8. It could represent 8 m, 8 km or 8 cm, vastly different distances.

Without the unit, we cannot interpret a speed, travel distance or physical scale confidently. A graph’s numerical value does not stand independently of its labels.

This applies to money, temperature, rainfall, time and percentages. A number 20 on a graph of temperature is different from a number 20 on a graph of sales.

Clear axes are part of correct mathematical communication.

Worked example 4: a count rises while its percentage falls

Suppose a fictional club records 15 participating students out of 30 eligible students in Year A. In Year B it records 18 participating students out of 60 eligible students.

The number participating increased from 15 to 18, a gain of three.

But the participation rate in Year A was 15/30 = 50%. In Year B it was 18/60 = 30%. The percentage of eligible participants fell by 20 percentage points.

Both statements can be true: participation count increased while participation rate decreased.

If a chart shows only the counts and a caption says “the club became more popular among eligible students,” it has not established that conclusion. The total eligible populations changed significantly.

Why denominators matter more than impressive numbers

Suppose School A has 40 students joining an activity from a cohort of 100. School B has 60 joining from a cohort of 300.

The absolute participation is higher in School B: 60 > 40. But the rates are 40% for School A and 20% for School B.

Neither comparison is automatically the right one. It depends on what the question asks. A staffing problem may care about the total number of participants; a proportional participation comparison cares about the relevant rate.

This example is fictional and not a claim about any actual schools. It illustrates why a good interpretation names the denominator instead of casually converting all counts into a ranking.

Worked example 5: pie-chart shares need a total

Imagine two fictional surveys. Survey A has 40 responses, of which 20 support an option. Survey B has 400 responses, of which 200 support the same option.

Both would show a 50% slice in a pie chart. Their proportions are equal, but their response counts differ by a factor of ten.

If two pie charts are displayed at identical sizes, readers may assume the underlying populations are similar. The figure needs a label indicating each survey’s total.

When a question asks for a frequency, you cannot get the count from a percentage alone without knowing the total number of responses.

For example, 30% of 40 is 12, while 30% of 400 is 120. Same percentage, very different count.

Worked example 6: convert a chart share back into an angle

A pie chart represents 60 valid survey responses. One category contains 18 responses.

Its percentage share is 18/60 × 100% = 30%. Its sector angle is 18/60 × 360° = 108°.

If someone draws a 180° sector for this category, that represents half the circle, or 50%—not 30%.

Checking that all sector angles add to 360° is a simple safeguard. Checking that all mutually exclusive and exhaustive shares add to 100% is another.

These checks help distinguish a drawing mistake from a wrong conclusion drawn from the underlying data.

A misleading headline can use true numbers

Imagine fictional attendance rising from 10 to 15 visitors. A headline says “Attendance increased by 50%.” That is mathematically correct: five extra visitors divided by the original ten is 50%.

But a headline that says “Five extra visitors transformed attendance” is an interpretation rather than a pure calculation. Whether the change is important depends on the setting and the underlying evidence.

A percentage by itself can sound large when the starting number is small. Reporting the original and final counts alongside the percentage gives readers a fuller picture.

The same discipline works in school data, household budgets and everyday online charts: ask for both the percentage and the number from which it was calculated.

Percentage change and percentage points are not the same

Suppose a fictional success rate rises from 40% to 50%.

The increase is 10 percentage points because 50% − 40% = 10 percentage points.

The relative percentage increase is (50 − 40)/40 × 100% = 25%. The base is the original rate of 40%.

Saying “a rise of ten percent” would be ambiguous and may be incorrect if the intended meaning is a percentage-point change. Precise descriptions should identify which measure they use.

The percentage-change guide works through more examples of changing percentage bases.

Worked example 7: line graphs do not always show continuous behaviour

Suppose a table contains fictional rainfall totals: Monday 0 mm, Tuesday 8 mm and Wednesday 4 mm. Connecting those daily totals with straight lines may help the reader see how the sequence changes between recorded days.

But the line segment between the points does not establish that rain fell evenly or that the total at some half-day moment was exactly 6 mm. Each measurement summarises a separate day.

The chart can help compare observations, but it does not fill missing measurements with genuine evidence.

A careful interpretation might say “The recorded daily total fell from eight millimetres on Tuesday to four on Wednesday,” rather than making unsupported claims about hourly rainfall patterns.

Worked example 8: a line’s slope only means something with axes

Consider points (1, 5) and (3, 13). The numerical gradient is (13 − 5)/(3 − 1) = 8/2 = 4.

But what does four mean? If the x-axis measures hours and the y-axis measures kilometres, the gradient is 4 km/h. If x measures printed items and y total dollars, it is $4 per item.

The arithmetic is identical, yet the interpretation differs. The graph’s labels turn a numeric slope into a meaningful rate.

Students studying linear graphs and gradient should practise writing those units when a context is provided.

Decorative three-dimensional bars can distort comparison

Suppose a designer draws two cylinders with their heights representing 10 and 20 units. The second is twice as tall.

If the illustrator also doubles the cylinder’s width, its pictured area and apparent volume may grow far more than twice as much. Readers may judge the larger object by its overall size rather than the intended height.

A plain two-dimensional bar with a clearly labelled axis often communicates the numerical comparison more reliably. Decoration is not a substitute for correct scale.

The teaching question is simple: “Which visible feature has been mapped to the number—the height, width, area or volume?” The figure should make that mapping unambiguous.

Overlapping categories can make percentages exceed 100%

A fictional survey asks students to tick every subject they enjoy. Twenty students tick Mathematics, eighteen tick Science and ten tick both.

These counts overlap. If we simply add 20 and 18, we count the ten students who chose both subjects twice.

The number choosing at least one of the two is 20 + 18 − 10 = 28.

If the survey allowed multiple answers, a chart of percentage selections may have category percentages whose sum exceeds 100%. That need not be an arithmetic mistake, but a pie chart dividing one whole into mutually exclusive slices would be inappropriate without redefining the categories.

Ask how responses were collected before assuming all categories form separate parts of one whole.

Sampling limits: forty answers do not speak for everyone

Imagine a voluntary online poll with forty responses. Even if the chart is drawn perfectly, it does not automatically represent every student in Bukit Timah or Singapore.

Volunteers might differ from those who did not respond. Respondents might come from only one class, school or interest group. The graph describes the collected responses unless the sampling design supports a wider inference.

A well-written conclusion acknowledges that distinction: “Among the forty respondents, fifteen selected MRT,” rather than “All Bukit Timah students prefer MRT.”

Statistical honesty begins with choosing words that do not claim more than the data can establish.

Eight questions for a quick chart audit

Take a fictional table: Year A has 95 participants, Year B has 100. A bar chart displays its vertical axis from 90 to 100. Another fictional survey has 18 of 60 respondents choosing Option B.

• A. By how many participants did the first count rise?
• B. What is the percentage increase from 95 to 100, approximately?
• C. Why might a bar chart beginning at 90 exaggerate the change?
• D. What percentage of the 60 survey responses chose Option B?
• E. What angle should Option B occupy in a pie chart?
• F. What extra information is needed to convert a bare 30% slice into a count?
• G. A rate rises from 40% to 50%. What is the percentage-point change?
• H. What is its relative percentage increase from the original 40%?

Answers and explanations

A. Five participants.

B. 5/95 × 100%, approximately 5.26%.

C. The visible bar portions above 90 have lengths five and ten, making a small relative change look like a doubling.

D. 18/60 × 100% = 30%.

E. 18/60 × 360° = 108°.

F. The total number of responses or observations.

G. Ten percentage points.

H. (50 − 40)/40 × 100% = 25%.

After marking, ask which answer would be unchanged if the entire graphic were drawn in a different colour. Mathematical values do not depend on decoration, but careless design can influence our visual impressions.

A short parent-and-student practice routine

For the first four minutes, find one simple graph in a schoolbook and read the title, axes, units and legend aloud. Do not interpret the trend until these are clear.

Then use five minutes to redraw a pair of fictional bars once with a zero baseline and once with a truncated baseline. Ask whether the underlying numbers changed. They did not; only the presentation changed.

For the final six minutes, compare a count and a rate with different denominators, then practise writing one cautious conclusion supported by the table.

A later retest using a different graphic is more informative than repeating exactly the same interpretation.

What Bukit Timah parents should ask a Mathematics tutor

Ask if the tutor encourages students to check the graph’s scales and the number of people represented, rather than only reading the height of a bar. A student who can compute an angle accurately may still misread what the survey actually measured.

The eduKateSG Secondary 1 Mathematics small-group reference explains close feedback and guided corrections near Sixth Avenue MRT. The Bukit Timah Secondary 1 G3 Mathematics tuition overview offers related local programme details. Confirm class availability and the child’s school syllabus sequence directly.

The official MOE G2 and G3 Mathematics syllabus sets out data handling and analysis within the G3 pathway. This article deepens introductory chart reading with everyday examples of potentially misleading displays; not every extension here is necessarily a separately tested Secondary 1 subtopic.

Frequently asked questions

### Is every graph with a nonzero vertical starting value misleading?

No. It depends on the graph type, purpose, labels and comparison being made. Truncated bar lengths can be especially deceptive when viewers read them as complete magnitudes.

### Can two surveys with equal pie-chart percentages have different counts?

Yes. Half of 40 is 20, while half of 400 is 200. Always check the survey total.

### Why should I check the axis labels before the bars?

The labels tell you whether the values represent counts, percentages, money, distance or some other quantity. Without them, the plotted numbers cannot be interpreted reliably.

### Is a bigger count always a bigger participation rate?

No. The relevant population size may also be bigger. Rates require a denominator.

### What is a safe way to describe a small survey?

Say what happened among the people who actually responded. Avoid extending the claim to an entire neighbourhood without suitable evidence.

Continue reading Secondary 1 G3 Mathematics

• Data handling, bar graphs, pie charts and pictograms
• Percentage change and percentage points
• Linear graphs, coordinates and gradient
• Significant figures and estimation

A good graph is an invitation to understand evidence, not simply believe a picture. Read its numbers, check its scale and ask whether its conclusion follows.