Mathematics Tutor Singapore | Secondary 4 cumulative frequency diagrams become difficult when a student confuses the height of the curve with the number of pupils in one interval. A cumulative frequency is a running total. It counts how many observations have occurred up to a specified boundary, not merely how many lie in the latest class.
Did you know that the median of thirty observations sits around the fifteenth observation on a cumulative frequency curve, while the first quartile is read near 7.5 on the cumulative scale? The horizontal coordinate read from the curve gives the estimated data value; the vertical coordinate gives a count.
This Singapore G3 E-Maths tutor guide covers frequency tables, cumulative frequency curves, estimated median and quartiles, interquartile range, percentiles and box-and-whisker diagrams. All examples are original, with a clear distinction between exact results from raw data and estimates from grouped data, aligned with SEAB’s 2027 K310 Statistics strand.
Explore cumulative frequency and box-plot learning
- A cumulative frequency is a running total
- Build cumulative frequencies from class intervals
- Draw and interpret an increasing curve
- Estimate the median from the curve
- Estimate lower and upper quartiles
- Worked example: grouped scores and linear interpolation
- What the 80th percentile means
- Five-number summary and box-and-whisker plot
- Compare spread and median between two groups
- Range versus interquartile range
- Why grouped statistics are often estimates
- Eight statistical misunderstandings to diagnose
- Sixteen worked cumulative frequency questions
- A ten-session SEC E-Maths revision route
- Questions about cumulative frequency tuition
- 2027 K310 syllabus and linked Maths lessons
A cumulative frequency is a running total
Imagine pupils completing a short assessment. A frequency table might tell us that six pupils scored between ten and twenty marks. But a cumulative frequency at twenty includes everyone with a score below that boundary, including pupils in earlier intervals.
The word cumulative signals accumulation. Each new class contributes its count to all previous counts. For example, if four pupils are in the first class and six in the next, the cumulative number at the second upper boundary is ten.
For an ordered list of nonnegative class frequencies, the cumulative frequency cannot decrease as the upper boundary increases. A graph that slopes downward has been drawn incorrectly or uses a different quantity on the vertical axis.
- Read both axes and their units.
- Identify whether a table contains individual frequencies or cumulative totals.
- Add class frequencies successively to create the cumulative column.
- Start a cumulative curve with zero observations at the initial lower boundary when appropriate.
- Mark upper class boundaries against accumulated counts.
- Read median, quartile and percentile levels on the vertical cumulative axis.
- Remember that grouped-data readings are estimates, not exact hidden scores.
Build cumulative frequencies from class intervals
Consider a fictional class assessment for thirty pupils. The grouped scores are in five intervals of width ten: 0≤score<10 has four pupils, 10≤score<20 has six, 20≤score<30 has ten, 30≤score<40 has eight, and 40≤score<50 has two.
| Score class | Frequency | Cumulative frequency at upper boundary |
|---|---|---|
| 0≤x<10 | 4 | 4 |
| 10≤x<20 | 6 | 10 |
| 20≤x<30 | 10 | 20 |
| 30≤x<40 | 8 | 28 |
| 40≤x<50 | 2 | 30 |
| Total | 30 | Final total 30 |
At x=30 the cumulative frequency is twenty, meaning twenty pupils have scores below thirty within the stated class convention. It does not mean twenty pupils scored between twenty and thirty; that individual class contains ten.
The last cumulative value must equal the total frequency. If the class counts total thirty but the final cumulative frequency is twenty-eight, two observations have been missed or one addition is wrong.
Draw and interpret an increasing curve
Place upper-boundary points at (10,4), (20,10), (30,20), (40,28) and (50,30), and include the initial point (0,0) in this example. A smooth or suitable connecting cumulative frequency curve rises through those points.
A cumulative graph is not a bar chart. A bar height can fall from one category to the next. A cumulative total cannot fall if the frequencies are nonnegative. The steepness of each section indicates how quickly observations accumulate per unit of the horizontal quantity.
| Point on curve | Interpretation |
|---|---|
| (10,4) | Four scores are below ten |
| (20,10) | Ten scores are below twenty |
| (30,20) | Twenty scores are below thirty |
| (40,28) | Twenty-eight scores are below forty |
| (50,30) | All thirty scores are below fifty |
A student may draw a curve starting at (0,4), as though the first class’s four pupils were already counted at the lower boundary. The correct starting cumulative value for this example is zero because there are no pupils with scores below zero under the grouped-score range.
Estimate the median from a cumulative frequency curve
For thirty observations, the exact median of an ordered raw-data list is the average of the fifteenth and sixteenth values. On a cumulative frequency curve, we typically estimate the central location at cumulative count N/2=15. We then read the corresponding score value on the horizontal axis.
Our graph rises from cumulative ten at score twenty to cumulative twenty at score thirty. With linear interpolation within that class, cumulative fifteen lies halfway through the ten-mark interval. The estimated median is 25 marks.
The value twenty-five is not guaranteed to be the exact median of the unseen thirty individual marks. The grouped table does not say exactly where each of the ten scores in 20≤x<30 lies. The straight-line assumption is a convenient graphical approximation.
Estimate lower and upper quartiles
The first quartile Q1 corresponds to about one-quarter of observations, so we read cumulative level N/4=7.5 for N=30. The third quartile Q3 corresponds to about three-quarters of observations, so we read cumulative level 3N/4=22.5.
On our example, level 7.5 lies in the 10≤x<20 class because cumulative frequency rises there from four to ten. Level 22.5 lies in the 30≤x<40 class because cumulative frequency rises there from twenty to twenty-eight.
The interquartile range is Q3−Q1. It describes the spread between the lower and upper quartiles—roughly the middle half of the observations—not the overall difference between maximum and minimum.
| Statistic | Cumulative frequency level | Class |
|---|---|---|
| Q1 | 30/4=7.5 | 10≤x<20 |
| Median | 30/2=15 | 20≤x<30 |
| Q3 | 3×30/4=22.5 | 30≤x<40 |
Worked grouped-score interpolation
For a grouped interval of width ten, linear interpolation assumes observations accumulate approximately uniformly through that interval. We can calculate the corresponding score by adding a fraction of the interval width to its lower boundary.
Q1: cumulative rises from four at score ten to ten at score twenty. The target is 7.5, which is 3.5 of the six cumulative counts into the interval. Estimated Q1 = 10+(3.5/6)×10≈15.83 marks.
Median: target fifteen lies five of ten into interval 20–30, so estimated median=20+(5/10)×10=25 marks.
Q3: target 22.5 is 2.5 of eight cumulative counts into interval 30–40. Estimated Q3=30+(2.5/8)×10=33.125 marks.
Thus the estimated interquartile range is 33.125−15.8333…≈17.29 marks. A calculator might show slightly different decimals depending on rounding; avoid rounding intermediate quartiles prematurely.
| Measure | Interpolation | Estimate |
|---|---|---|
| Q1 | 10+(7.5−4)/6×10 | 15.83 |
| Median | 20+(15−10)/10×10 | 25 |
| Q3 | 30+(22.5−20)/8×10 | 33.125 |
| IQR | Q3−Q1 | ≈17.29 |
A graph-read question may not expect this level of numerical precision. When the task says “estimate from the cumulative frequency diagram”, follow the graph’s scale and the question’s requested accuracy rather than present exact-looking numbers the source data cannot support.
What the 80th percentile means
The 80th percentile is a value below which roughly 80% of observations fall, subject to the data and convention used. With thirty observations, 80% corresponds to cumulative level 0.8×30=24.
Our cumulative graph has count twenty at score thirty and count twenty-eight at score forty. A linear read at level twenty-four is halfway between these counts, so the estimated 80th-percentile score is 35.
A student should not confuse the percentile rank 80 with a score of eighty marks. An 80th percentile score can be thirty-five or another value depending on the distribution.
Five-number summary and a box-and-whisker plot
A box-and-whisker plot represents the minimum, lower quartile Q1, median, upper quartile Q3 and maximum. The box extends from Q1 to Q3, with a median mark inside it. The whiskers in a basic school box plot extend to the minimum and maximum under the stated convention.
Consider an original ordered data set of twelve scores: 2,3,4,4,5,7,8,9,10,12,14,16. The median is (7+8)/2=7.5. Under the median-of-halves method, the lower half is 2,3,4,4,5,7 and its median is (4+4)/2=4. The upper half is 8,9,10,12,14,16 and its median is (10+12)/2=11.
| Five-number summary | Value |
|---|---|
| Minimum | 2 |
| Q1 | 4 |
| Median | 7.5 |
| Q3 | 11 |
| Maximum | 16 |
| Interquartile range | 11−4=7 |
| Range | 16−2=14 |
Quartile conventions can vary across textbooks and statistical software. For school exercises use the method, graph or calculator convention specified by the teacher and examination; the given method above is explicit so the numerical result is reproducible.
Compare spread and median between two groups
Suppose Group A has median 7.5, Q1=4 and Q3=11. Its IQR is seven. Group B has median eight, Q1=6 and Q3=10, so its IQR is four. Group B has a slightly higher central median and a narrower middle-half spread.
It would be inaccurate to conclude that every pupil in B scored higher than every pupil in A; box plots do not contain all individual scores. Nor can we infer that the groups have the same sample size unless the information is supplied.
| Comparison | Group A | Group B | Supported statement |
|---|---|---|---|
| Median | 7.5 | 8 | B’s median higher |
| Q1 | 4 | 6 | B’s lower quartile higher |
| Q3 | 11 | 10 | B’s upper quartile lower |
| IQR | 7 | 4 | B’s central 50% less spread out |
Range versus interquartile range
The range is maximum minus minimum. It is sensitive to very small or very large observations. The interquartile range measures the middle-half spread and is less affected by extreme values.
In the twelve-score example, range is 16−2=14, while IQR is 11−4=7. Neither value alone tells us the exact arrangement of every score between the quartiles.
A narrow IQR does not necessarily imply a narrow full range. A group could have most pupils close together yet still include a few extreme marks. That is why a useful statistical comparison often reports both location and spread.
Why grouped statistics are often estimates
A grouped frequency table does not specify exact scores. It only states how many lie in each interval. Different raw data sets can produce the same frequency table but have different exact medians and quartiles within those intervals.
A cumulative frequency curve typically joins class-boundary points smoothly or approximately. Reading a point between boundaries introduces an interpolation assumption. This is useful for estimating central values, but students should not present the result as if every individual score were known.
The distinction matters beyond school. In real survey summaries, the data aggregation and measurement choices limit what can be inferred. Mathematical accuracy includes an honest statement of uncertainty.
Eight statistical misunderstandings to diagnose
| First mistaken move | Likely gap | Next diagnostic |
|---|---|---|
| Writes individual frequency ten as cumulative count at 30 | Running total ignored | Add all earlier classes |
| Draws a decreasing cumulative curve | Cumulative meaning misunderstood | Can collected observations disappear? |
| Reads horizontal coordinate as number of pupils | Axes swapped | Label counts versus marks |
| Uses n/2 as an exact raw median value | Position confused with score | Read x-coordinate at cumulative level |
| Takes Q1 directly as score 7.5 | Count level confused with data scale | Project across curve then down |
| Claims grouped median 25 is exact | Interpolation limitation ignored | Compare possible raw scores within interval |
| Calls range the same as IQR | Different spread measures confused | Mark min/max and Q1/Q3 |
| Says smaller IQR implies every score higher | Overinterpretation of box plot | What individual data is missing? |
A pupil may know the formula for IQR but be unable to read the quartile scores from a curve. Another may read the curve accurately but subtract the wrong values. These require different exercises.
Sixteen worked cumulative frequency questions
1. First cumulative total
If the first frequency is four, the first cumulative value is four.
Name the axis, unit and whether the result is exact or estimated. For grouped data, explain the interpolation assumption and use the precision requested by the examination question.
2. Second cumulative total
Frequencies four and six produce cumulative ten.
Name the axis, unit and whether the result is exact or estimated. For grouped data, explain the interpolation assumption and use the precision requested by the examination question.
3. Third cumulative total
Adding ten to cumulative ten gives twenty.
Name the axis, unit and whether the result is exact or estimated. For grouped data, explain the interpolation assumption and use the precision requested by the examination question.
4. Final frequency total
4+6+10+8+2=30 observations.
Name the axis, unit and whether the result is exact or estimated. For grouped data, explain the interpolation assumption and use the precision requested by the examination question.
5. Class frequency from cumulative counts
At upper boundaries 20 and 30, counts are 10 and 20; class 20–30 contains ten.
Name the axis, unit and whether the result is exact or estimated. For grouped data, explain the interpolation assumption and use the precision requested by the examination question.
6. Scores below 40
Cumulative at boundary 40 is 28.
Name the axis, unit and whether the result is exact or estimated. For grouped data, explain the interpolation assumption and use the precision requested by the examination question.
7. Scores from 30 to below 40
28−20=8 pupils.
Name the axis, unit and whether the result is exact or estimated. For grouped data, explain the interpolation assumption and use the precision requested by the examination question.
8. Approximate median position
For n=30, median cumulative level is 15.
Name the axis, unit and whether the result is exact or estimated. For grouped data, explain the interpolation assumption and use the precision requested by the examination question.
9. Estimated median value
Linear read between (20,10) and (30,20) gives 25.
Name the axis, unit and whether the result is exact or estimated. For grouped data, explain the interpolation assumption and use the precision requested by the examination question.
10. First quartile level
N/4=7.5.
Name the axis, unit and whether the result is exact or estimated. For grouped data, explain the interpolation assumption and use the precision requested by the examination question.
11. Lower quartile estimate
10+(7.5−4)/6×10≈15.83.
Name the axis, unit and whether the result is exact or estimated. For grouped data, explain the interpolation assumption and use the precision requested by the examination question.
12. Upper quartile level
3N/4=22.5.
Name the axis, unit and whether the result is exact or estimated. For grouped data, explain the interpolation assumption and use the precision requested by the examination question.
13. Upper quartile estimate
30+(22.5−20)/8×10=33.125.
Name the axis, unit and whether the result is exact or estimated. For grouped data, explain the interpolation assumption and use the precision requested by the examination question.
14. Estimated IQR
33.125−15.8333≈17.29.
Name the axis, unit and whether the result is exact or estimated. For grouped data, explain the interpolation assumption and use the precision requested by the examination question.
15. The 80th percentile
At cumulative 24, linear interpolation gives approximately 35 marks.
Name the axis, unit and whether the result is exact or estimated. For grouped data, explain the interpolation assumption and use the precision requested by the examination question.
16. Box plot exact IQR
For 2,3,4,4,5,7,8,9,10,12,14,16 under median-of-halves, IQR=11−4=7.
Name the axis, unit and whether the result is exact or estimated. For grouped data, explain the interpolation assumption and use the precision requested by the examination question.
A ten-session G3 E-Maths revision route
| Session | Skill | Evidence of success |
|---|---|---|
| 1 | Read grouped frequency table | Understands each interval |
| 2 | Build cumulative totals | Final equals N |
| 3 | Plot cumulative boundaries | Curve nondecreasing |
| 4 | Find median level N/2 | Reads horizontal value |
| 5 | Read Q1 and Q3 | Distinguishes count from score |
| 6 | Compute interquartile range | Subtracts scores |
| 7 | Read a percentile | Identifies cumulative level |
| 8 | Build five-number summary | Makes box plot |
| 9 | Compare two distributions | States location and spread accurately |
| 10 | Mixed unseen graph after a delay | Explains without tutor prompts |
Where a learner struggles with percentage conversion, repair that prerequisite briefly. Where the issue is reading the axes, practise transferring from horizontal lines to curve intersections and back down to the data axis before using more complicated questions.
Questions about cumulative frequency tuition
What is cumulative frequency?
A running total of observations up to a stated value or upper class boundary.
Why must a cumulative frequency curve not decrease?
Counts of accumulated observations cannot reduce as the boundary rises.
How do I locate the median?
Use cumulative level about N/2 and read the corresponding horizontal data value from the curve.
How do I find Q1 and Q3 on the graph?
Read horizontal values corresponding to cumulative levels N/4 and 3N/4.
What is interquartile range?
Q3−Q1, measuring spread across the middle half of observations.
Why are grouped quartile values estimates?
The grouped table hides exact raw scores within each interval, so interpolation is approximate.
What does a box plot show?
A five-number summary: minimum, lower quartile, median, upper quartile and maximum.
Does a smaller IQR guarantee every observation is closer to the mean?
No. It describes the middle-half spread only and does not show every observation.
Are these topics examined in the SEC G3 syllabus?
Yes. The 2027 K310 Statistics topic S1 lists cumulative frequency diagrams, box-and-whisker plots, quartiles and percentiles.
Can tuition guarantee SEC performance?
No. Accurate data interpretation is one important topic within a wider Mathematics syllabus.
SEAB syllabus and connected Maths lessons
The official 2027 SEC G3 Mathematics K310 syllabus lists cumulative frequency diagrams, box-and-whisker plots, quartiles, percentiles, range, interquartile range, mean and standard deviation within topic S1. Exact teaching order can vary by school, and G1/G2 levels use different specifications.
Continue through the Mathematics Learning Hub. For introductory statistics read Secondary Statistics: Mean, Median and Outliers. For Primary data interpretation use PSLE Pie Charts. For a later H2 probability and statistics application, see JC H2 Hypothesis Testing and p-Values.
The original Secondary 1 Mathematics Tutor Clementi teaching reference remains unchanged. Contact eduKate Singapore with a student’s original grouped table and box plot to identify whether the difficulty is cumulative totals, reading the curve or interpreting quartiles.