Secondary 4 G2 Mathematics Tutor Bukit Timah | Cumulative Frequency, Box Plots and Data Spread. Did you know? Two classes can have exactly the same average score but very different results. One class may have everyone clustered around the mean; the other may include some much lower and some much higher scores. The word average cannot tell that whole story.
Statistics becomes genuinely useful when Secondary 4 G2 Mathematics students in Bukit Timah can read a cumulative-frequency curve, understand a box-and-whisker plot and compare both the centre and spread of data. This reader-friendly guide covers the underlying vocabulary, original worked examples, graph traps and ways parents can check learning without another pile of worksheets.
The official 2027 SEC G2 Mathematics K210 syllabus includes S1 Data Handling and Analysis. Several specific topics—including cumulative-frequency diagrams, box plots, quartiles, percentiles, spread measures and calculation of standard deviation—are underlined in that document. Underlined content is assessed only in Paper 2 Section B for the 2027 K210 examination. Not every statistics topic is limited to Section B. Candidates graduating in 2026 should use their applicable N(A)-Level Mathematics Syllabus A 4045 entry; SEAB 2026 and SEAB 2027 distinguish the two national examination years.
Quick answer: analyse data in three layers
- Centre: What is typical? Mean and median offer different views, especially when extreme values are present.
- Spread: How much do results vary? Range, interquartile range and standard deviation describe different aspects of variation.
- Context: Which comparison is justified? An average alone, a truncated axis or an imprecise graph may create a misleading claim.
If two classes share a mean of 60 marks, the class with smaller variation is not automatically “better” or “worse”; it is simply more consistent under that measure. Always state the comparison the data actually supports.
The five numbers behind a box-and-whisker plot
| Feature | What it tells you | Useful check |
|---|---|---|
| Minimum | Smallest recorded observation | Must be at or below Q1. |
| Lower quartile Q1 | Lower quarter reference point under the prescribed convention | Quartiles require an ordered dataset. |
| Median Q2 | Middle of the ordered observations | Half the observations lie on either side, subject to ties. |
| Upper quartile Q3 | Upper quarter reference point under the prescribed convention | Must be at or above the median. |
| Maximum | Largest recorded observation | Must be at or above Q3. |
| Interquartile range | Q3 − Q1, spread of the central 50% | Never confuse with the full range, maximum − minimum. |
There can be different quartile conventions, especially for small datasets. In the worked lists below, the middle observation is excluded from each half for an odd-sized dataset, and halves are averaged for an even-sized dataset. For actual examination work, follow the question’s and school’s stated convention.
Seventeen original data-spread and cumulative-frequency examples
1. Mean from individual observations
For 2, 4, 6, 8, the total is 20 across four observations. Mean = 5. The calculation uses all data values, so unusual extremes can influence it.
2. Median in an odd-sized ordered list
For 3, 5, 8, 9, 11, 14, 18, the fourth of seven numbers is the middle. Median = 9. Order the data first.
3. Quartiles of that seven-number list
Using the stated split-halves convention, lower half is 3, 5, 8 and upper half is 11, 14, 18. Thus Q1 = 5 and Q3 = 14.
4. Range versus interquartile range
For 3, 5, 8, 9, 11, 14, 18, range = 18 − 3 = 15, while IQR = 14 − 5 = 9. The two statistics describe different spreads.
5. Median of an even-sized ordered list
For 2, 4, 5, 7, 9, 10, 12, 15, median = (7 + 9)/2 = 8. Neither middle observation alone is the median.
6. Quartiles of that eight-number list
Using the specified split-halves convention: lower half 2, 4, 5, 7 gives Q1 = (4 + 5)/2 = 4.5; upper half 9, 10, 12, 15 gives Q3 = (10 + 12)/2 = 11. IQR = 6.5.
7. Two datasets with the same mean
Group A = 4, 4, 4, 4; Group B = 1, 3, 5, 7. Both have mean 4, but Group A has no variation and Group B is visibly spread out.
8. Standard deviation of the constant group
For A = 4, 4, 4, 4, each difference from the mean is zero. The population standard deviation is 0. Every observation equals the mean.
9. Standard deviation of the varied group
For B = 1, 3, 5, 7, mean = 4. Squared differences are 9, 1, 1, 9, sum = 20. Divide by 4 and take the square root: σ = √5 ≈ 2.24, to three significant figures.
10. An even simpler standard-deviation calculation
For 2, 4, 6, mean = 4; squared deviations are 4, 0 and 4. Population variance = 8/3 and σ = √(8/3) ≈ 1.63. Do not confuse variance with standard deviation.
11. Grouped-data estimated mean
Suppose class intervals 0–<10 and 10–<20 have frequencies 2 and 3. Use midpoints 5 and 15. Estimated mean = (2×5 + 3×15)/5 = 11. Because original individual data are unavailable, it is an estimate.
12. Grouped-data estimated standard deviation
For that same example, estimated mean = 11 and Σfx²/Σf = (2×25 + 3×225)/5 = 145. Estimated standard deviation = √(145 − 11²) = √24 ≈ 4.90. It is based on class midpoints.
13. Building a cumulative-frequency table
Four ordered equal-width classes 0–<10, 10–<20, 20–<30, 30–<40 have frequencies 3, 5, 8 and 4. Cumulative frequencies at upper class boundaries are 3, 8, 16, 20. Add each new frequency to the previous total.
14. Recovering a class frequency
In the cumulative table above, 8 observations are below 20 and 16 are below 30. The frequency from 20 to below 30 is 16 − 8 = 8. A cumulative value is not a single-class frequency.
15. Percentage below a threshold
In the example of 20 observations, 16 are below 30. Thus 16/20 × 100 = 80%. The denominator is all 20 recorded observations.
16. Approximating the median from grouped cumulative frequency
There are 20 observations, so the approximate median is read near cumulative frequency 10. A straight-line interpolation between cumulative 8 at 20 and 16 at 30 gives 20 + (10 − 8)/8 × 10 = 22.5. It is an estimate from the grouped curve, not an exact known individual value.
17. Approximating quartiles from the same grouped curve
On a 20-observation cumulative-frequency graph, Q1 is read near cumulative 5 and Q3 near 15. Straight-line interpolation gives Q1 ≈ 10 + (5 − 3)/5 × 10 = 14 and Q3 ≈ 20 + (15 − 8)/8 × 10 = 28.75, giving estimated IQR ≈ 14.75. Actual hand-drawn curves and individual data may differ.
What a cumulative-frequency graph can and cannot tell you
- Can show: approximate counts below thresholds, median and quartile positions, and the proportion of observations within an indicated interval.
- Can support: a comparison of how distributions accumulate when curves use compatible axes and samples.
- Cannot necessarily show: exact individual values when plotted from grouped classes or a smooth drawn curve.
- Cannot prove: what caused a data pattern or that a small sample exactly describes a larger population.
Did you know? Reading the cumulative frequency backwards can be useful: first find the horizontal position of the median or quartile on the vertical cumulative axis, then move across to the curve and down to the measurement axis. Many students incorrectly choose the midpoint of the horizontal measurement scale instead.
Two classes: how to compare centre and spread fairly
| Measure | Class A: 4,4,4,4 | Class B: 1,3,5,7 | What follows |
|---|---|---|---|
| Mean | 4 | 4 | Same average. |
| Median | 4 | 4 | Same middle location. |
| Range | 0 | 6 | B has larger extreme-to-extreme spread. |
| Population standard deviation | 0 | √5 ≈ 2.24 | B has larger spread around the mean. |
It would be fair to say Class B is more variable under these measures. It would not be fair to claim Class A was taught better, studied harder or will perform better next year from this dataset alone. Statistical interpretation stops where evidence stops.
The graph-reading trap: a chart can exaggerate differences
Suppose two groups have mean scores of 98 and 100. On a bar graph whose vertical axis starts at 95, the bars may appear dramatically different. The actual difference is 2 points. A truncated axis may be useful to reveal small changes, but the reader must notice it before interpreting the visual magnitude. Always inspect the axis, units, scale, sample size and graph title.
Five common errors and their corrective questions
| Error | What went wrong | Question that repairs it |
|---|---|---|
| Quartiles are read before arranging observations | The ordered-position meaning was lost | Is the original list already sorted? |
| IQR is calculated as maximum minus minimum | Full range was confused with middle spread | Which values are Q1 and Q3? |
| A cumulative total is used as the frequency of one class | The student forgot accumulation | What was the earlier cumulative total? |
| A graph-estimated median is described as an exact individual value | Precision and grouped-data limits ignored | Can you recover the original twenty observations from the graph? |
| A larger standard deviation is called a higher average | Centre and spread were confused | Do the means differ, or only the variability? |
A five-question self-test without notes
- For 1, 4, 6, 8, 9, 12, 15, use the split-halves convention to find the median, Q1, Q3 and IQR.
- A frequency table has class frequencies 2, 6, 5 and 7. List the cumulative frequencies.
- In that table, how many observations are in the third class?
- Group X consists of 5,5,5. Group Y consists of 3,5,7. Which has the greater population standard deviation?
- Explain why a graph with its vertical axis starting at 90 can exaggerate the visual difference between 96 and 98.
Answers: (1) median 8, Q1 4, Q3 12, IQR 8; (2) 2, 8, 13, 20; (3) 5; (4) Y, because X has zero variability whereas Y does not; (5) the visual scale shows a two-point change against a much shorter vertical span. State the numerical difference before interpreting the picture.
The 2027 Paper 2 Section B distinction
In 2027 K210, Paper 2 Section B requires one answer from two questions: one from specified underlined Geometry and Measurement content, one from specified underlined Statistics and Probability content. The cumulative-frequency, box-plot and spread topics addressed here include the underlined S1 material. This is an examination-structure fact, not advice to neglect the wider statistics syllabus. A 2026 candidate must follow the 4045 paper structure applicable to their own examination year.
An illustrative 90-minute small-group data lesson
| Time | Teaching move | Evidence of independent skill |
|---|---|---|
| 0–10 | Closed-book recall of mean, median, range and IQR | Each learner explains centre versus spread. |
| 10–25 | Short unseen cumulative-frequency table | Tutor identifies the first incorrect running total. |
| 25–50 | Draw a curve and interpret a box plot side by side | Student can locate quartiles with their data limitations. |
| 50–75 | Individual comparison of two unfamiliar datasets | Student distinguishes averages, variation and what is unsupported. |
| 75–90 | Exit problem, error note and delayed retest | Student explains the method without using a copied template. |
The unchanged eduKateSG three-pupil Mathematics tutorial reference describes premium small groups of three and 1.5-hour weekly lessons near Sixth Avenue MRT. This article illustrates how the learning method might be adapted to Secondary 4 G2 data interpretation; no present Secondary 4 G2 availability, fees or teacher assignment is implied.
FAQ for Bukit Timah families
Why does my child know the mean but struggle with box plots?
Mean is one calculation; a box plot requires ordering, locating quartiles, reading the median and comparing spread. Practise those steps separately before combining two plotted datasets. Do not assume every quartile convention is identical unless the school or question defines it.
Is a lower standard deviation always better?
No. It describes greater consistency around the mean, not quality by itself. Whether high or low variation is desirable depends on what is being measured. Compare the mean and the context as well.
Are cumulative-frequency and standard-deviation questions part of G2?
Yes. They are explicitly included in the 2027 K210 Statistics and Probability S1 content, with specified underlined portions restricted to Paper 2 Section B. Confirm the correct syllabus code and candidate year before choosing practice papers.
Can a student improve by doing more statistical charts every day?
Practice helps when each chart tests an identified weakness. If a student repeatedly confuses cumulative frequency and class frequency, correct that distinction with changed tables and a delayed retest rather than simply adding another graph.
Continue the Secondary 4 G2 Mathematics pathway
- Statistics, probability and graphs overview
- Probability trees and combined events
- Estimation, rounding and calculator accuracy
- 2026 N(A) and 2027 SEC revision roadmap
Official references: 2027 Mathematics K210 syllabus PDF · 2027 SEC G2 school-candidate listing · 2026 N(A) school-candidate listing. For tuition enquiries, see the immutable original small-group Mathematics guide or eduKate Singapore. This article uses original illustrative datasets, not official SEAB examination questions.