Percentiles and Interpercentile Range
A percentile is the value below which a given percentage of a data set falls when the values are arranged in order, so the 30th percentile has 30% of the data below it. The interpercentile range, often the 10th to 90th, is the difference between two chosen percentiles and measures spread while ignoring extreme values. It is estimated from a cumulative frequency graph on GCSE Statistics Higher tier.
Before you start
Make sure you're comfortable with these topics first:
Method
- Draw or read a cumulative frequency diagram, with cumulative frequency on the vertical axis and the variable on the horizontal axis.
- To find the pth percentile, calculate its position as (p/100) x n, where n is the total frequency.
- Find that position on the cumulative frequency axis (or table), then interpolate within the class it falls in to estimate the value.
- To find an interpercentile range, find both percentile values in this way, then subtract the lower one from the upper one.
- Give the answer to a sensible precision for the scale of the data, with units, and check it is smaller than the full range of the data.
Common mistakes
MistakeUsing the cumulative frequency after the percentile class as the interpolation denominator instead of the frequency in that class.
InsteadUse the class frequency for the fraction of the class traversed: (target position - cumulative frequency before the class) divided by class frequency.
MistakeReading the percentile from a class boundary without first locating its position in the cumulative frequencies.
InsteadCalculate the position from the total frequency, find the class containing it, and interpolate from that class's lower boundary using its frequency.
MistakeSubtracting the lower percentile from the upper percentile in the wrong order, producing a negative spread.
InsteadFor an interpercentile range, subtract the lower percentile value from the upper percentile value and include the variable's units.
Worked examples
Example 1
The table shows the finishing times, in minutes, of 100 runners in a fun run: 0 < t <= 10, frequency 5; 10 < t <= 20, frequency 20; 20 < t <= 30, frequency 25; 30 < t <= 40, frequency 30; 40 < t <= 50, frequency 20. Estimate the 10th percentile, the 90th percentile, and the 10-90 interpercentile range.
- n = 100, so the 10th percentile is at position 0.10 x 100 = 10th value, and the 90th percentile is at position 0.90 x 100 = 90th value.
- The cumulative frequencies are 5, 25, 50, 80, 100 at the class boundaries 10, 20, 30, 40, 50.
- The 10th value falls in the class 10 < t <= 20 (cumulative frequency goes from 5 to 25 there): 10th percentile = 10 + (10 - 5)/20 x 10 = 12.5 minutes.
- The 90th value falls in the class 40 < t <= 50 (cumulative frequency goes from 80 to 100 there): 90th percentile = 40 + (90 - 80)/20 x 10 = 45 minutes.
- Interpercentile range (10th to 90th) = 45 - 12.5 = 32.5 minutes.
- Final answer: 10th percentile = 12.5 minutes, 90th percentile = 45 minutes, 10-90 interpercentile range = 32.5 minutes.
Example 2
The grouped table gives the journey times, in minutes, for 160 commuters: 0 < t <= 8, frequency 12; 8 < t <= 16, frequency 28; 16 < t <= 24, frequency 36; 24 < t <= 32, frequency 44; 32 < t <= 40, frequency 25; 40 < t <= 48, frequency 15. Estimate the 15th percentile, the 85th percentile and the interpercentile range between them, giving answers to 2 decimal places.
- The total frequency is 160, so the 15th percentile is at position 0.15 x 160 = 24 and the 85th percentile is at position 0.85 x 160 = 136.
- Cumulative frequencies are 12, 40, 76, 120, 145 and 160. Position 24 is in 8 < t <= 16, so P15 = 8 + (24 - 12) / 28 x 8 = 11.428... minutes.
- Position 136 is in 32 < t <= 40, so P85 = 32 + (136 - 120) / 25 x 8 = 37.12 minutes.
- Interpercentile range = 37.12 - 11.428... = 25.691... minutes.
- Final answer: P15 = 11.43 minutes, P85 = 37.12 minutes and the interpercentile range = 25.69 minutes (all to 2 d.p.).
Example 3
Two fictional groups record the same measurement, 42 cm. In Group A, 18 of 48 measurements are below 42 cm. In Group B, 36 of 120 measurements are below 42 cm. No measurements equal 42 cm. Compare the percentile positions of 42 cm in the two groups. Show why comparing the counts below 42 alone would give the wrong conclusion.
- A percentile position describes the percentage of measurements below a value. Because the totals differ, divide the number below 42 by each group total before comparing.
- For Group A, 18 / 48 x 100 = 37.5, so 42 cm is at the 37.5th percentile position.
- For Group B, 36 / 120 x 100 = 30, so 42 cm is at the 30th percentile position.
- Group B has twice as many measurements below 42 as Group A, but it also has a larger total. Its proportion below 42 is smaller: 30% compared with 37.5%.
- Final answer: 42 cm has the higher percentile position in Group A, by 7.5 percentage points (37.5th compared with 30th).
Practice questions
Q1[2 marks]In a data set of 40 values arranged in order, which position corresponds to the 25th percentile?Show answer
Answer: 10th value (25/100 x 40 = 10)
Q2[2 marks]A cumulative frequency graph shows 60 values. Which position corresponds to the 80th percentile?Show answer
Answer: 48th value (0.80 x 60 = 48)
Q3[2 marks]A data set has an interquartile range of 18. Would the 10-90 interpercentile range usually be smaller or larger than this? Explain.Show answer
Answer: Larger - the 10-90 interpercentile range covers 80% of the data (10th to 90th percentile) while the IQR covers only the middle 50% (25th to 75th), so it usually spans a wider gap
Q4[2 marks]Heights of 50 plants (cm): 0 < h <= 10, freq 5, cumulative 5; 10 < h <= 20, freq 20, cumulative 25; 20 < h <= 30, freq 15, cumulative 40; 30 < h <= 40, freq 10, cumulative 50. Estimate the 20th percentile.Show answer
Answer: 12.5 cm (position 0.2 x 50 = 10th value; 10 + (10-5)/20 x 10 = 12.5)
Q5[2 marks]Weekly household spending for 40 households has cumulative frequencies: up to 50 pounds, 10; up to 100 pounds, 25; up to 150 pounds, 35; up to 200 pounds, 40. Estimate the 75th percentile.Show answer
Answer: 125 pounds (position 0.75 x 40 = 30th value; 100 + (30-25)/10 x 50 = 125)
Q6[2 marks]Class A has a 10th percentile of 42 marks and a 90th percentile of 88 marks. Class B has a 10th percentile of 55 marks and a 90th percentile of 79 marks. Compare the spread of the middle 80% of marks in each class.Show answer
Answer: Class A's interpercentile range is 88 - 42 = 46 marks; Class B's is 79 - 55 = 24 marks, so Class B's middle 80% of marks are much less spread out (more consistent) than Class A's
Exam-style questions
Written in the style of a GCSE Statistics exam paper, with a full mark scheme.
Exam tips
- Find the total frequency and calculate each percentile position before looking for its class in the cumulative frequencies.
- Show the lower class boundary, cumulative frequency before the class, class frequency and class width in every interpolation.
- Subtract the lower estimate from the upper estimate, round to a sensible precision for the graph or table, and include units.
Q1[3 marks]Delivery times, in minutes, for 60 parcels have cumulative frequencies: up to 10, 5; up to 20, 25; up to 30, 45; up to 40, 55; up to 50, 60. Estimate the 30th percentile.Show mark scheme
Tick each line you got.
Nothing ticked yet - 3 available
Q2[4 marks]The ages, in years, of 120 gym members have cumulative frequencies: 16 <= a < 20, 15; 20 <= a < 30, 60; 30 <= a < 40, 100; 40 <= a < 60, 120. Estimate the interpercentile range between the 20th and 80th percentiles.Show mark scheme
Tick each line you got.
Nothing ticked yet - 4 available
Q3[2 marks]The 10th percentile of the prices of second-hand cars at a dealership is 3200 pounds and the 90th percentile is 11800 pounds. Calculate the 10-90 interpercentile range and briefly interpret what it shows about the spread of prices.Show mark scheme
Tick each line you got.
Nothing ticked yet - 2 available
See real GCSE Statistics past-paper questions, with official mark schemes
This topic in real past papers
We have tagged some boards' past papers question by question (not every board or year yet). This topic comes up in:
- AQA GCSE Statistics past papers: 2 questions across 2 papers, latest tagged June 2023
- Edexcel GCSE Statistics past papers: 1 question across 1 paper, latest tagged June 2019
Free printable worksheet
Want more practice on paper? Download the Percentiles and Interpercentile Range worksheet pack - 30 pages of exam-style questions with a full mark scheme. Free to download, no card or account. Print it for personal and classroom use.
This topic is chapter 2 of GCSE Statistics Higher Essentials 2, the whole course as one free printable PDF. Its worksheet is also chapter 2 of GCSE Statistics Higher Topic Workbook 5 and chapter 12 of GCSE Statistics Higher 10-Minute Tests 2, free printable practice books.
Other versions of this worksheet: