Cumulative Frequency
On this page
How to draw a cumulative frequency graph, then read the median, quartiles and interquartile range from it, count values above or between two values, test a claim, draw a box plot and spot what is wrong with a graph.
Builds on: Frequency Tables, Comparing Distributions
The method
Draw the graph. The lengths of 120 fish are grouped in the table.
1. Add the frequencies as you go down. 8, then 8+20=28, then 28+44=72, and so on. The last cumulative frequency is the total, 120.
2. Plot each cumulative frequency at the upper end of its class: (20, 8), (30, 28), (40, 72), (50, 104), (60, 120). Start at (10, 0), the lower end of the first class: no fish are shorter than 10 cm.
3. Join the points with a smooth curve. Straight lines from point to point also score; a line of best fit does not.
Plotting at the middle of each class (at 15, 25, …) is the commonest lost mark. A cumulative frequency graph never goes down.
Reading the graph, for n values:
- Median: across from \frac n2 to the curve, then down. Lower and upper quartiles: from \frac n4 and \frac{3n}4.
- Interquartile range = upper quartile - lower quartile.
- Up from a value and across gives how many are up to that value.
- Every reading is an estimate.
Now try this
The table gives information about the times, in seconds, that 75 people took to answer a quiz question.
On a cumulative frequency graph, what are the coordinates of the point plotted for the class 40<t\le60?
- Cumulative frequency for 0<t\le60: 9+23+25=57
- Plot it at the upper end of the class, 60: the point is (60, 57).
Worked example 1: Median and interquartile range
The cumulative frequency graph gives information about the masses, in grams, of 200 apples from an orchard.
(a) Find an estimate for the median mass.
(b) Find an estimate for the interquartile range.
Apples from a second orchard have a median mass of 126 g and an interquartile range of 24 g.
(c) Compare the distributions of the masses of the apples from the two orchards.
- (a) Median: across from \frac{200}{2}=100, then down: about 132 g.
- (b) Quartiles: across from \frac{200}{4}=50 and \frac{3\times200}{4}=150: about 116 g and 148 g.
- Interquartile range \approx148-116=32 g
- Subtract the masses, not the cumulative frequencies. Read at half the total, not at 50.
- (c) The first orchard's apples have the higher median, 132 g against 126 g, so they were heavier on average. The second orchard's apples have the smaller interquartile range, 24 g against 32 g, so their masses were more consistent.
Now try this
The cumulative frequency graph gives information about the lengths, in minutes, of 60 phone calls.
Find an estimate for the interquartile range.
- Quartiles: across from \frac{60}{4}=15 and \frac{3\times60}{4}=45: about 9 minutes and 17 minutes.
- Interquartile range \approx17-9=8 minutes
Worked example 2: More than a value, and between two values
The cumulative frequency graph gives information about the distances, in km, that 150 cyclists rode in a charity event.
(a) Find an estimate for the number of cyclists who rode more than 42 km.
One of the cyclists is chosen at random.
(b) Find an estimate for the probability that this cyclist rode between 18 km and 42 km.
- (a) Up from 42 km and across: about 114 cyclists rode 42 km or less.
- More than 42 km: 150-114=36
- The reading counts the cyclists up to 42 km: subtract it from the total.
- (b) Up from 18 km: about 24. Between: 114-24=90
- Probability \approx\frac{90}{150}=\frac35
Now try this
The cumulative frequency graph gives information about the times, in minutes, that 70 students spent on homework one evening.
Find an estimate for the number of students who spent more than 70 minutes on homework.
- Up from 70 minutes and across: about 50 students spent 70 minutes or less.
- More than 70 minutes: 70-50=20
Worked example 3: Is the claim correct?
The cumulative frequency graph gives information about the speeds, in mph, of 160 cars on a road.
Owen says, "75% of the cars had a speed of 40 mph or less."
Is Owen correct? You must show how you get your answer.
- Up from 40 mph and across: about 112 cars.
- \frac{112}{160}\times100=70\%
- 70% is less than 75%, so Owen is not correct.
- Or: 75% of 160 is 120. Across from 120 gives about 41 mph, not 40 mph. Either way, write the figures and then the decision.
Now try this
The cumulative frequency graph gives information about the masses, in grams, of 60 eggs.
Find an estimate for the percentage of the eggs with a mass of more than 60 g.
- Up from 60 g and across: about 42 eggs. More than 60 g: 60-42=18 eggs.
- \frac{18}{60}\times100=30\%
Worked example 4: A box plot from a cumulative frequency graph
The cumulative frequency graph gives information about the lengths, in mm, of 72 leaves.
The shortest leaf was 28 mm. The longest leaf was 90 mm.
Draw a box plot for the lengths of the leaves.
- Read across from \frac{72}{4}=18, \frac{72}{2}=36 and \frac{3\times72}{4}=54: lower quartile 42 mm, median 52 mm, upper quartile 60 mm.
- The shortest and longest come from the question: the graph cannot show them.
- Read at 18, 36 and 54, not at a quarter, a half and three quarters of the length axis.
Worked example 5: What is wrong with the graph?
The table gives information about the heights, in cm, of 40 seedlings.
Ria drew the cumulative frequency graph below for this table. Give two mistakes Ria has made.
- 1. The points are plotted at the middle of each class (10, 30, …). They should be at the upper ends: 20, 40, 60, 80, 100.
- 2. The cumulative frequency for 0<h\le80 is plotted at 30. It should be 34.
- Say exactly what is wrong: "the points are plotted at the midpoints", not "the points are wrong".
- The same kind of question asks why a range can't be found from the graph (the shortest and longest values are not shown), or what is wrong with a cumulative frequency table (every class must start at the lowest value: 0<h\le40, not 20<h\le40).
Next: Box Plots
More on this topic: Cumulative Frequency worksheet with full solutions