Did you know? The same group of responses can be shown accurately in a bar graph, a pie chart or a pictogram—but not every graph answers a reader’s question equally well. If you want to compare category counts, bars are often easiest to read. If you want to show shares of a whole, a pie chart may be useful. If you want to show changes across ordered times, a line graph may tell the story more clearly.
For Secondary 1 G3 Mathematics students in Bukit Timah, choosing a statistical diagram is a reasoning skill, not just an artwork choice. A learner who can draw every chart mechanically may still struggle to explain which one communicates the data most accurately.
This guide offers quick choices, worked examples, practical limitations and a short self-test. All survey data below is fictional, and no results are claims about actual Bukit Timah schools or students.
Quick answer: decide whether your question asks about exact numbers, category comparisons, a time trend or parts of one whole. Then select the representation that communicates that relationship clearly, with a correct title, axes, scale and legend.
Choose a chart type, compare worked examples or jump to the practice answers.
- The four chart types in plain English
- Start with the question, not the drawing tool
- Worked example 1: use a bar graph to compare counts
- Worked example 2: turn the same table into a pictogram
- Worked example 3: make a pie chart from the survey
- When is a pie chart a poor choice?
- Worked example 4: a line graph needs ordered observations
- What line graphs can and cannot tell us
- Worked example 5: two surveys of different sizes
- Which diagram answers which question fastest?
- Why an axis that starts at 11 might exaggerate a bar difference
- A pictogram must keep symbol size consistent
- A bar graph and a histogram are not interchangeable
- What does the data actually prove?
The four chart types in plain English
- Bar graph: compare frequencies or measured quantities across separate categories. Bar length communicates magnitude.
- Line graph: show a pattern or trend across naturally ordered measurements, often time. Interpret connecting segments according to what was measured.
- Pie chart: show how a single whole is divided into nonoverlapping categories. The sector angles together form 360°.
- Pictogram: use repeated pictures or symbols to represent counts, with a key stating how many observations each symbol means.
A frequency table is not a failure to make a chart. It is often the clearest way to find exact values and check that the total has been recorded correctly.
The right choice depends on what the reader needs to know and whether the data genuinely supports that display.
Start with the question, not the drawing tool
Suppose a fictional survey asks 30 students to select exactly one favourite activity. The results are:
- Reading: 12 responses.
- Sport: 9 responses.
- Art: 6 responses.
- Music: 3 responses.
The total is 12 + 9 + 6 + 3 = 30.
If we ask “How many more selected Reading than Music?”, a table or bar graph supports that comparison: 12 − 3 = 9.
If we ask “What proportion selected Reading?”, a pie chart or percentage table can highlight its share of the total: 12/30 = 40%.
If we ask “Did Reading become more popular across three months?”, this one survey does not answer it. We would need observations across time. Choosing a line graph cannot invent missing trend data.
Worked example 1: use a bar graph to compare counts
For the 30-response survey, draw four separated bars labelled Reading, Sport, Art and Music. Put “Number of responses” on the vertical axis and use a consistent scale beginning at zero.
The bar heights are 12, 9, 6 and 3. Reading has the highest frequency, Sport is second, Art third and Music fourth.
The bar format makes it easy to compare 12 and 9, or 6 and 3, without converting to percentages first.
If the vertical axis instead says “Percentage of respondents,” the bar heights should be 40%, 30%, 20% and 10%. A bar height of 12 on a percentage axis would not communicate the same meaning as 12 responses.
The axes and units are part of the mathematics.
Worked example 2: turn the same table into a pictogram
Set a key in which one full symbol represents three responses.
Reading then needs four symbols, Sport three symbols, Art two symbols and Music one symbol.
The key matters. Four symbols for Reading means 4 × 3 = 12 responses, not four respondents.
Using a key of three works especially well here because every frequency is divisible by three. If Music had four responses instead, one full symbol plus one-third of a symbol would be needed under the same key, and the fractional representation would need to be clear.
A pictogram can be inviting for quick communication, but repeated images may be harder to compare precisely than a bar graph when counts are large or the key produces awkward partial symbols.
Worked example 3: make a pie chart from the survey
A pie chart represents a complete group as a circle of 360°. The angle for each category is its fraction of the 30 total responses multiplied by 360°.
Because 360° ÷ 30 = 12° per response:
- Reading: 12 × 12° = 144°, or 40%.
- Sport: 9 × 12° = 108°, or 30%.
- Art: 6 × 12° = 72°, or 20%.
- Music: 3 × 12° = 36°, or 10%.
Check the angles: 144 + 108 + 72 + 36 = 360°. Check percentages: 40 + 30 + 20 + 10 = 100%.
The circle emphasises the composition of one whole. It is not necessarily the easiest picture for deciding whether 12 responses are three more than nine; a bar graph may show that difference more directly.
When is a pie chart a poor choice?
A pie chart becomes less useful when there are many tiny categories whose sectors are difficult to distinguish. A bar graph with labels may communicate small differences more clearly.
A conventional pie chart also expects the categories to account for one whole without overlapping. If survey participants are allowed to select multiple favourite activities, the category totals may exceed the number of people.
For example, one person selecting both Reading and Music belongs to two response categories. Those counts should not be inserted into one simple “share of respondents” pie chart without redefining the categories to be mutually exclusive.
Before selecting a chart, ask how the observations were collected.
Worked example 4: a line graph needs ordered observations
Imagine fictional daily temperature observations taken at the same time each day: Monday 24°C, Tuesday 25°C, Wednesday 23°C, Thursday 27°C and Friday 26°C.
Put the ordered days along the horizontal axis and temperature in degrees Celsius on the vertical axis. Plot the five measurements in their stated order.
A line joining consecutive points can help readers see a rise from Monday to Tuesday, a fall to Wednesday, a rise to Thursday and a slight fall on Friday.
The values are ordered by day, which makes a trend display natural. A pie chart would be an unsuitable way to show these temperatures as portions of one whole: 24°C, 25°C and 23°C are measurements, not exclusive shares of a common total.
The line is a visual connection between observations. It does not prove the exact temperature at every moment between the daily measurement times.
What line graphs can and cannot tell us
A line graph of daily measurements is useful for comparing changes through time. Its direction and slope can reveal an increasing or decreasing trend across the recorded points.
But don’t assume that joining two observations means the real quantity changed at a perfectly constant rate between them. The graph may be a convenient visual summary of discrete readings.
And don’t automatically claim that two values are causally connected because their lines rise together. A chart may display a relationship without establishing why it occurred.
The strongest school interpretation states the observed change first. A broader explanation requires supporting information.
Worked example 5: two surveys of different sizes
Imagine Survey A has 30 responses with 12 choosing Reading. Survey B has 60 responses with 21 choosing Reading.
Survey A has a Reading share of 12/30 = 40%. Survey B has a Reading share of 21/60 = 35%.
The count rose from 12 to 21, but the percentage fell from 40% to 35%.
If you use bar graphs of raw counts, Survey B’s Reading bar is taller. If you use percentage bars, Survey A’s Reading share is higher. Neither is intrinsically wrong, but the graph must be labelled according to the comparison being made.
Two side-by-side pie charts also require their individual survey totals to be labelled; equally sized circles do not tell readers that one survey contained twice as many responses.
Which diagram answers which question fastest?
Ask “Which activity received the most responses?” A bar graph or frequency table is a strong first choice.
Ask “What fraction of respondents chose Sport?” A pie chart or percentage table can reveal Sport’s share: 9/30 = 30%.
Ask “How did daily rainfall vary across a week?” A line graph may help with the chronological pattern, provided the observation definitions are clear.
Ask “Can younger readers count groups using a simple icon key?” A pictogram may work well, especially with modest whole-number frequencies.
Ask “What was the exact number for every category?” A numerical table may communicate the most precisely. The best representation is not always a graphic.
Why an axis that starts at 11 might exaggerate a bar difference
Suppose fictional category counts are 12 and 13. Their difference is only one response.
If a bar chart begins at zero, the bars are close in height. If the axis begins at 11, the visible bar portions are one and two units high, making the second appear twice the first.
The actual count ratio is 13/12, approximately 1.083, not two. A truncated baseline can therefore exaggerate an ordinary bar-length comparison.
For detailed explanation, see misleading graphs, axis scales and statistical claims.
Different graph types have different display conventions; a narrowly scaled line graph can be legitimate when clearly labelled and appropriate to its purpose. Do not declare every nonzero axis wrong without understanding the comparison.
A pictogram must keep symbol size consistent
Suppose one symbol represents three people. Making the symbol for a category with nine responses twice as tall and twice as wide as other symbols can distort perception, even if you print the correct key.
Readers may judge the apparent area of a larger picture instead of counting the number of equal symbols.
Use consistent-sized symbols and a clear key. If fractional pictures are required, explain whether half a symbol represents half the count of a full symbol.
The purpose of a pictogram is to make values easier to interpret, not to replace accurate numerical representation with decoration.
A bar graph and a histogram are not interchangeable
A simple bar graph compares distinct categories such as Reading and Sport. These categories are separated because they represent different labels, not consecutive numerical intervals.
A histogram is a different statistical representation used for grouped numerical intervals. Its construction and interpretation depend on class intervals and sometimes frequency density.
For Secondary 1 data handling, the important immediate question is which display represents the given category or ordered observations clearly. Do not label every set of touching or separated bars incorrectly just because the pictures look similar.
The later histogram concept is useful background, not a reason to force advanced statistics into an introductory chart lesson.
What does the data actually prove?
If 12 out of 30 fictional respondents select Reading, the supported statement is “40% of these 30 respondents chose Reading.”
It does not follow that 40% of all students in Bukit Timah prefer Reading. The sample was fictional, and a real sample would need suitable selection and evidence to justify claims about a wider population.
Good chart interpretation distinguishes observed data, calculated proportions and conclusions beyond the sample.
A carefully chosen statistical diagram helps communication, but it cannot repair incomplete or biased data collection.
Four different weaknesses behind a wrong chart
- Data-count problem: frequencies don’t match the raw observations.
- Representation problem: a line is used for unordered categories or a pie chart is used for overlapping choices.
- Scale problem: bar heights don’t correspond to their labels or units.
- Interpretation problem: a caption claims more than the data supports.
A good tutor checks the first wrong decision. A learner who calculated every angle correctly may still have used an unsuitable pie chart; another who chose the right graph may have miscounted the raw responses.
These are different problems and should be taught as such.
Ten-question mini diagnostic
Use the fictional survey: Reading 12, Sport 9, Art 6, Music 3; total 30.
- A. Which chart is usually clearest for comparing these four category counts?
- B. Which chart is designed to highlight their shares of the one whole?
- C. What angle should Reading occupy in the pie chart?
- D. What angle should Music occupy?
- E. If one pictogram symbol equals three responses, how many symbols represent Sport?
- F. What percentage selected Art?
- G. Which type is naturally suited to measurements ordered by day?
- H. Does a line graph connecting daily observations prove the exact value between measurements?
- I. Can a simple one-whole pie chart directly represent overlapping multiple-choice responses?
- J. If one survey has 12 out of 30 choosing Reading and another 21 out of 60, which survey has the higher Reading percentage?
Answers with a reason
A. A bar graph or frequency table; the counts belong to separate categories.
B. A pie chart, assuming categories are exhaustive and mutually exclusive.
C. 12/30 × 360° = 144°.
D. 3/30 × 360° = 36°.
E. Three symbols, since 9 ÷ 3 = 3.
F. 6/30 × 100% = 20%.
G. A line graph often helps display an ordered trend through time.
H. No. The line connects observed points; without further information, it does not establish precise intermediate measurements.
I. Not without first redefining the categories to remove overlap or changing what the whole represents.
J. Survey A, at 40%, compared with Survey B at 35%.
Ask the learner not only for an answer, but also for the purpose of the representation. That is the heart of this lesson.
A focused fifteen-minute tutorial
Begin with a small frequency table and ask the child to identify two questions: one about the biggest count and one about the share of the whole.
Use five minutes to sketch a bar graph and five to calculate pie-sector angles from the same table. Confirm the frequencies sum to the total and the pie angles sum to 360°.
Then supply several measurements across time and ask whether a line or pie chart would communicate them better. The child should give one sentence of reasoning.
Finish by presenting a misleading axis or a multiple-choice survey, and ask why the display might be unsuitable. The changed situation reveals whether the graph choice was understood rather than memorised.
What Bukit Timah parents should ask a tutor
A productive Secondary 1 G3 Mathematics tutor checks whether students can organise data, choose a suitable representation, read axis scales and explain why the chart fits the question.
The immutable eduKateSG Secondary 1 Mathematics small-group reference describes close feedback in three-student tutorials near Sixth Avenue MRT. The Bukit Timah G3 Mathematics tuition overview offers local programme context. Parents should verify current availability and the student’s school chapter sequence directly.
The MOE G2 and G3 Mathematics syllabus includes tables, bar graphs, pictograms, line graphs, pie charts, and their purposes, advantages and disadvantages at Secondary 1 G3 level.
Frequently asked questions
### Is a pie chart always better than a bar graph for percentages?
No. Percentage bars can make comparisons clearer, especially when categories have similar shares. A pie chart is particularly useful for the composition of one whole.
### Should a line graph always be used for anything measured over time?
Not automatically. It depends on the nature of the observations and whether connecting them communicates the intended trend accurately. A table or set of bars can sometimes be more appropriate.
### Why does a pictogram need a key?
Because one picture may represent several observations. Without the key, the reader cannot translate symbols into counts.
### Can different graphs use the same data?
Yes, as long as each faithfully communicates the correct quantities and the meaning of the axes, sectors or symbols. The best choice depends on the question.
### Can a good chart prove what every Singapore student thinks?
No. Wider claims depend on valid data and sampling, not on the attractiveness of the graph.
Continue with these connected guides
- Build tally charts and frequency tables from raw data
- Data handling: bar graphs, pie charts and pictograms
- Misleading graphs and statistical claims
- Fractions, decimals and percentages
- Linear graphs, coordinates and gradient
Before your child draws a chart, ask just one question: “What do we want the reader to understand?” Choosing the graph begins with that answer.