Cumulative frequency is a running total. You add the frequencies class by class, plot each total against the upper end of its class, and join the crosses with a smooth curve. From that curve you can read the median, the quartiles, the interquartile range and any percentile, and you can answer "how many were below (or above) this value?" In Cambridge IGCSE Mathematics 0580 this is E9.6, and it is Extended content only: Core students do not study it.
In our experience, of the two Extended-only statistics sub-topics, this is the one that gives up marks most readily once the method is secure, because every answer is read from a graph you drew yourself. The other is histograms. Both are placed in context on our Statistics strand overview for 0580.
What E9.6 asks for
The syllabus lists two skills:
- Draw and interpret cumulative frequency tables and diagrams.
- Estimate and interpret the median, percentiles, quartiles and interquartile range from cumulative frequency diagrams.
Its notes add that plotted points should be clearly marked, for example as small crosses, and joined with a smooth curve. Note the word "estimate": values from a curve are estimates, because the curve is drawn from grouped data. If you would rather practise these graphs with someone checking each reading, our Cambridge IGCSE tutors work through original data sets with Extended students.
Worked example 1: the table and the plot
120 people completed a puzzle. Their times, t minutes, are shown.
| Time (t min) | Frequency | Cumulative frequency | Plot at |
|---|---|---|---|
| 0 < t ≤ 10 | 8 | 8 | (10, 8) |
| 10 < t ≤ 20 | 22 | 30 | (20, 30) |
| 20 < t ≤ 30 | 40 | 70 | (30, 70) |
| 30 < t ≤ 40 | 30 | 100 | (40, 100) |
| 40 < t ≤ 50 | 14 | 114 | (50, 114) |
| 50 < t ≤ 60 | 6 | 120 | (60, 120) |
Running totals. 8, then 8 + 22 = 30, then 30 + 40 = 70, and so on. The last total must equal the number of people, 120. If it does not, an addition has slipped.
Why the upper end? The cumulative frequency 30 means "30 people took 20 minutes or less". That statement is only true at the end of the class, t = 20, so you plot (20, 30), not (15, 30).
Start point. Nobody took 0 minutes or less, so the curve starts at (0, 0). Plot it.
Axes. Time across (0 to 60), cumulative frequency up (0 to 120). Mark each point with a small cross and join them with one smooth curve, drawn freehand, not with a ruler. The curve should never go down, because a running total cannot decrease.
Worked example 2: median, quartiles and IQR
Use the curve to estimate the median, lower quartile, upper quartile and interquartile range.
For a cumulative frequency curve, use n/2 for the median and n/4 and 3n/4 for the quartiles. With n = 120:
| Measure | Read across at | Estimate from the curve |
|---|---|---|
| Lower quartile | 120 ÷ 4 = 30 | 20 minutes |
| Median | 120 ÷ 2 = 60 | about 27.5 minutes |
| Upper quartile | 3 × 120 ÷ 4 = 90 | about 36.7 minutes |
Method on the graph. Go across from 60 on the vertical axis to the curve, then down to the horizontal axis. Draw these guide lines; they show your method.
IQR = 36.7 − 20 = about 16.7 minutes.
The lower quartile reads exactly 20 because (20, 30) is a plotted point. The others depend on the shape of your curve, so a reading of 27 or 28 for the median is the kind of small variation you would expect from a hand-drawn curve. What matters is reading at the right height and showing it.
Interpretation. The median says half of the people finished in about 27.5 minutes or less. The IQR says the middle half of the times spread over about 17 minutes.
Worked example 3: percentiles and "more than"
(a) Estimate the 90th percentile. (b) Estimate how many people took more than 45 minutes.
(a) The 90th percentile is the value below which 90% of the data lies. 90% of 120 = 108. Read across at 108 and down: about 45.7 minutes, so roughly 46.
(b) This time you start on the horizontal axis. Go up from 45 minutes to the curve and across: a cumulative frequency of about 107. That is the number who took 45 minutes or less. So the number who took more than 45 minutes is 120 − 107 = about 13 people.
This subtraction is the step most often missed. A cumulative frequency reading always counts "up to" a value; for "more than", subtract from the total.
Worked example 4: comparing two groups
A second group of 120 people did the same puzzle after a practice session. Their median was 24 minutes and their IQR was 10 minutes. Compare the two groups.
- On average, the second group was faster, because their median (24 minutes) is lower than the first group's (about 27.5 minutes).
- The second group's times were more consistent, because their IQR (10 minutes) is smaller than the first group's (about 16.7 minutes).
Same shape as any E9.2 comparison: one average, one spread, each with a reason and the context. Note that "lower median" means "faster" here, because the data is a time. Read what the variable means before you write "better".
Why n/2 and not (n + 1)/2?
For a list of individual values, the median of n values sits at position (n + 1)/2. A cumulative frequency curve is different: it treats the data as continuous, and the curve itself is an estimate, so the usual IGCSE convention is to read at n/2, n/4 and 3n/4. With 120 people the difference between 60 and 60.5 is far smaller than the accuracy of a hand-drawn curve anyway. Our averages and spread guide covers the list method for quartiles.
Non-calculator and calculator papers
- Paper 2 (no calculator): totals are usually chosen so that n/4 and 3n/4 are whole numbers, such as 80, 120 or 200. The work is reading the graph, so take care with the scale on each axis: a square may be worth 2, 4 or 5 units.
- Paper 4 (calculator): the same graph work, with percentages to calculate (for example, the 85th percentile of 160 people is the 136th value) and perhaps a percentage answer: 13 out of 120 is 10.8%.
In both, the graph is drawn on the question paper, so a sharp pencil and a clear ruler for guide lines help.
Mistakes our tutors see
- Plotting at the midpoint or the lower end of each class instead of the upper end.
- Forgetting the start point (0, 0) or the lowest class boundary.
- Joining points with straight lines when the syllabus asks for a smooth curve (or, the reverse, a curve that wobbles and dips).
- Reading the quartiles at n/4 on the horizontal axis instead of the vertical axis.
- For "more than X", giving the cumulative frequency itself instead of the total minus it.
- Misreading the axis scale, especially when one small square is worth 2 or 4.
- Giving the IQR as the two quartiles instead of their difference.
When to get help
- Your curve looks right but your readings disagree with the answer by a whole class width.
- You are never sure whether to plot at 10, 5 or 0 for the first class.
- Percentile questions stop you, even though you can find quartiles.
- "More than" and "less than" questions get mixed up under time pressure.
Self-check
- A cumulative frequency curve shows data for 200 students. At what heights do you read the lower quartile, median and upper quartile?
- A class is 20 < m ≤ 40 with cumulative frequency 35. Where do you plot the point?
- Cumulative frequencies for masses are: up to 20 kg, 12; up to 40 kg, 35; up to 60 kg, 70; up to 80 kg, 90. How many masses were between 40 kg and 60 kg? How many were over 60 kg?
- For 150 values, which value do you read for the 80th percentile?
- The quartiles read from a curve are 18 and 31. Find the IQR.
Answers. 1: 50, 100 and 150. 2: at (40, 35). 3: 70 − 35 = 35 between 40 and 60 kg; 90 − 70 = 20 over 60 kg. 4: 0.8 × 150 = the 120th value. 5: 31 − 18 = 13.
Where it leads
Cumulative frequency often sits next to a histogram question on the same data, so continue with our frequency density and histograms guide. For the other half of the strand, see scatter diagrams and correlation for IGCSE. Box-and-whisker plots left 0580 in 2025, but reading and comparing distributions does not go away: it continues in "Statistics and probability" in the IB Diploma, and "Representation of data" opens Cambridge 9709's Probability & Statistics 1. If you are moving on to the IB, see our IB Maths AI SL tutoring.
How Maths Bodhi helps
We ask students to draw the curve in the lesson, on real graph paper or a shared screen, because most lost marks come from plotting and reading rather than from understanding. The tutor then sets two readings, one up and one across, until both directions are automatic. Families with a student in the board year can see how Grade 10 IGCSE maths help is planned around mocks.
Home tuition is available in Gurgaon (Gurugram). Students elsewhere in India and abroad learn with us online. You can book a free demo class.
Written and checked by the Maths Bodhi team. Facts checked against official Cambridge International and Pearson sources, November 2026.
Frequently asked questions
- Is cumulative frequency in IGCSE Core maths?
- No. In 0580, cumulative frequency diagrams (E9.6) are Extended content only. The Core syllabus marks C9.6 as "Extended content only".
- Where do you plot points on a cumulative frequency graph?
- At the upper class boundary of each class, against the running total. Also plot the starting point at the lowest boundary with cumulative frequency 0.
- How do you find the median from a cumulative frequency curve?
- Read across from n/2 on the cumulative frequency axis to the curve, then down to the horizontal axis. The value there is an estimate of the median.
- How do you find the interquartile range from a cumulative frequency graph?
- Read the lower quartile at n/4 and the upper quartile at 3n/4, then subtract: IQR = upper quartile − lower quartile.
- Should a cumulative frequency graph be a curve or straight lines?
- The 0580 syllabus notes say plotted points should be joined with a smooth curve. Mark the points as small crosses first.
- How do you find how many values are more than a given number?
- Read up from the value to the curve and across to get the cumulative frequency, then subtract that from the total frequency.
