Skip to content
Marlbridge

Study Guides

AQA GCSE Mathematics 8300: Statistics – Study Guide

Study guide for AQA GCSE Maths 8300 Statistics (S1-S6): sampling, charts, averages and spread, histograms, box plots and scatter graphs.

Subject
Mathematics
Level
GCSE
Topic
Statistics
Updated

Aligned to AQA GCSE Mathematics (8300), For first teaching 2015. Official specification .

Syllabus page (what it covers and how it is assessed): AQA GCSE Mathematics.

Syllabus topics and points this page is mapped to

8300

  • 6 Statistics (topic-level mapping)

"Topic-level mapping" means this page belongs to that topic; it does not claim to cover every point in it.

Found an error? Report a correction.

Need help with this topic? Request a free trial class for GCSE Mathematics (8300).

This study guide teaches Topic 6, Statistics (specification references S1 to S6), of the AQA GCSE Mathematics (8300) specification, for teaching from September 2015 with exams from May/June 2017 (version 1.0). Statistics can be tested on any of the three papers at Foundation or Higher tier. Most of it is for both tiers. Histograms, cumulative frequency graphs (S3), box plots, quartiles and interquartile range (part of S4) are Higher tier only and are labelled below. Exams run every May/June and November.

Use it with the Statistics revision notes and the Statistics practice questions. The course hub is AQA GCSE Mathematics, the printable checklist lists every statement, and the free 10-minute diagnostics show where to start.

What this topic covers

Spec What you must be able to do Tier
S1 Infer properties of a population from a sample, knowing the limits of sampling Both
S2 Read and draw frequency tables, bar charts, pie charts, pictograms, vertical line charts, and tables and line graphs for time series; choose a suitable diagram Both
S3 Draw and interpret histograms (equal and unequal class widths) and cumulative frequency graphs Higher tier only
S4 Compare distributions using graphs, averages (mean, median, mode, modal class) and range, including outliers Both
S4 Box plots, quartiles and interquartile range Higher tier only
S5 Apply statistics to describe a population Both
S6 Scatter graphs, correlation, lines of best fit, interpolation and extrapolation; correlation is not causation Both

Probability and statistics together make up about 15% of the marks at both tiers.

Types of data

You must know these four terms (S4 notes):

  • Primary data: you collect it yourself, e.g. your own survey.
  • Secondary data: someone else collected it, e.g. a website or a published table.
  • Discrete data: can only take separate values, usually counts, e.g. number of siblings.
  • Continuous data: measured, so it can take any value in a range, e.g. height, time, mass.
A tree splits data into qualitative (words or labels) and quantitative (numbers). Quantitative data is discrete (counted, such as goals, where 2.5 makes no sense) or continuous (measured, such as a time of 17.35 seconds).
Data is qualitative or quantitative; numbers are counted (discrete) or measured (continuous).

Populations and samples (S1, S5)

The population is the whole group you want to know about. A sample is the part you actually ask or measure. Sampling is quicker and cheaper, but a sample only gives an estimate.

A sample is biased if some members of the population are more likely to be chosen than others. Asking people at a gym about exercise habits will overstate how much the town exercises. A random sample gives every member an equal chance. A larger random sample gives a more reliable estimate.

Worked example 1. A factory makes 2000 phone cases a day. A random sample of 40 has 6 faulty cases.

Proportion faulty in sample = 6/40 = 0.15
Estimate for the population = 0.15 × 2000 = 300 faulty cases

The answer is an estimate. A different sample of 40 could give a different answer.

A factory makes 2000 phone cases a day. A random sample of 40 cases contains 6 faulty ones, which is 0.15. Scaling up, 0.15 times 2000 gives an estimate of about 300 faulty cases a day.
A sample proportion scaled up estimates the population, but another sample could give a different answer.
A town of 40 dots, 8 of whom exercise often (20 per cent). Asking only at the gym picks 5 of those 8, so all 5 say they exercise often: 100 per cent, which overstates the town. One random sample of 5 gives 1 who exercises often: 20 per cent, close to the town value.
Asking only at a gym is a biased sample; a random sample gives everyone an equal chance.

Tables, charts and diagrams (S2)

Diagram Use it for
Bar chart Categories (e.g. favourite sport); bars the same width with gaps
Pictogram Categories; needs a key saying what one symbol stands for
Pie chart Showing each category as a share of the whole
Vertical line chart Ungrouped discrete data (e.g. shoe sizes)
Line graph Time series: how a value changes over time

Worked example 2 (pie chart). 72 people were asked how they travel to work: walk 18, bus 26, car 20, cycle 8.

Angle for one person = 360° ÷ 72 = 5°
Walk 18 × 5 = 90°,  Bus 26 × 5 = 130°,  Car 20 × 5 = 100°,  Cycle 8 × 5 = 40°
Check: 90 + 130 + 100 + 40 = 360°

A pie chart shows proportions, not totals. Two pie charts with the same angle can stand for different numbers of people if the totals are different.

Time series. Plot time along the horizontal axis and join the points with straight lines. Describe the trend (the general rise or fall) and any repeating pattern, such as sales that peak every summer.

A line graph of example quarterly ice-cream sales over three years, 12 points joined by straight lines. The general trend rises steadily, shown by a dashed line. The points also rise and fall in a repeating pattern, with a peak every summer quarter (quarters 3, 7 and 11).
In a time series, describe the trend (the general rise or fall) and any repeating pattern, such as a summer peak.
Bars for 1 to 5 people have heights 6, 13, 9, 8 and 4 flats, totalling 40. The bars for 3, 4 and 5 people are highlighted: 9 + 8 + 4 = 21 of 40 flats. The bars are separate and the axis starts at 0.
Each bar height is the frequency, and adding the frequencies gives the number of flats.
A pie chart of 72 people has 5 degrees per person. Beach 130, City 90, Mountains 60, Cruise 20 and Other 60 degrees, adding to 360.
Angle = frequency x (360 / total), and the angles must add to 360 degrees.

Averages and spread (S4)

Measure How to find it
Mean Total of values ÷ number of values
Median Middle value when in order; position (n + 1)/2
Mode Most common value (or modal class for grouped data)
Range Largest − smallest; a measure of spread

Worked example 3 (frequency table). Goals scored in 20 matches:

Goals 0 1 2 3 4
Frequency 4 7 5 3 1
Total goals = 0×4 + 1×7 + 2×5 + 3×3 + 4×1 = 30
Mean = 30 ÷ 20 = 1.5 goals
Median: position (20 + 1)/2 = 10.5, so between the 10th and 11th values.
Running totals 4, 11, ... so the 10th and 11th are both 1. Median = 1
Mode = 1 (highest frequency).  Range = 4 − 0 = 4

Worked example 4 (grouped data). Time, t minutes, 40 people waited:

Time (minutes) 0 < t ≤ 10 10 < t ≤ 20 20 < t ≤ 30 30 < t ≤ 40
Frequency 4 15 12 9

Use the midpoint of each class as a stand-in for every value in it.

Midpoints: 5, 15, 25, 35
Σ(midpoint × frequency) = 20 + 225 + 300 + 315 = 860
Estimated mean = 860 ÷ 40 = 21.5 minutes
Modal class: 10 < t ≤ 20
Running totals 4, 19, 31, 40: the 20th and 21st values are in 20 < t ≤ 30

The mean is only an estimate because you do not know the exact values in each class.

Outliers

An outlier is a value far from the rest. For 12, 14, 15, 15, 16, 17, 18, 41 the mean is 18.5, but without 41 it is about 15.3. The median barely moves (15.5 to 15). The range drops from 29 to 6. When there is an outlier, the median is usually the better average, and the range is not a reliable measure of spread.

Comparing distributions

Always make two comparisons: one of an average and one of a spread, each in the context of the question.

Example: Class A has median 62 and range 30. Class B has median 55 and range 48. “On average Class A scored higher, because its median is higher. Class A’s scores were more consistent, because its range is smaller.”

A dot plot of eight delivery times: 9, 9, 9, 11, 12, 14, 22, 35. Mode 9, median 11.5, mean 121/8 = 15.1 and range 26. The outlier 35 pulls the mean above six of the eight values.
The mean, median, mode and range of one data set, with an outlier pulling the mean up.
A frequency table of pets owned by 20 students. Multiplying pets by students gives 0, 8, 8, 6 and 4, a total of 26, so the mean is 26 / 20 = 1.3. The mode is 1 and the median is 1.
For a frequency table, multiply each value by its frequency, add, then divide by the total frequency.
Heights of 40 students in four 10 cm classes. Each class midpoint (155, 165, 175, 185) times its frequency totals 6740, giving an estimated mean of 168.5 cm. The modal class is 160 to under 170.
Use the midpoint of each class, not a boundary, so the mean is only an estimate.
Two dot plots of minutes late for seven buses. Both routes have mean and median 4, but Route 1 has range 4 and Route 2 has range 8, so Route 1 is more consistent.
Compare data sets with one comment on an average and one on spread.

Histograms and cumulative frequency (S3) – Higher tier only

Histograms

In a histogram the area of each bar stands for the frequency. The height is the frequency density:

frequency density = frequency ÷ class width

Worked example 5. Heights of 76 plants:

Draws the source plant-height histogram for unequal intervals and shows frequencies represented by bar areas, not heights.
Draws the source plant-height histogram for unequal intervals and shows frequencies represented by bar areas, not heights. Select the image to enlarge it.
Height h (cm) 0 < h ≤ 10 10 < h ≤ 15 15 < h ≤ 20 20 < h ≤ 30 30 < h ≤ 50
Frequency 12 18 22 16 8
Frequency density 1.2 3.6 4.4 1.6 0.4

Estimate how many plants are between 12 cm and 25 cm:

10 < h ≤ 15: the part from 12 to 15 is 3/5 of the class → 3/5 × 18 = 10.8
15 < h ≤ 20: all 22
20 < h ≤ 30: the part from 20 to 25 is 5/10 of the class → 5/10 × 16 = 8
Total ≈ 10.8 + 22 + 8 = 40.8, about 41 plants

This assumes values are spread evenly within each class.

A histogram of plant heights with frequency density 1.2, 3.6, 4.4, 1.6 and 0.4 for classes 0-10, 10-15, 15-20, 20-30 and 30-50. The area from 12 to 25 is shaded: 10.8 + 22 + 8 gives about 41 plants.
Area is frequency, so shade the part you want and add the areas.

Cumulative frequency graphs

Add the frequencies as you go, then plot each running total at the upper class boundary. Join with a smooth curve or straight lines, starting from the lower boundary of the first class at a cumulative frequency of 0.

For 80 values in classes 0–10, 10–20, 20–30, 30–40, 40–50 with frequencies 8, 14, 26, 20, 12, plot (10, 8), (20, 22), (30, 48), (40, 68), (50, 80), starting at (0, 0).

  • Median: read across at 80 ÷ 2 = 40. It is about 27.
  • Lower quartile: read across at 80 ÷ 4 = 20. It is about 18.6.
  • Upper quartile: read across at 3 × 80 ÷ 4 = 60. It is about 36.
  • Interquartile range = upper quartile − lower quartile, about 17.4.

The values here come from straight lines between the points. A curve gives slightly different readings, which is why exam answers from graphs accept a range of values.

A cumulative frequency graph through the points (0, 0), (10, 8), (20, 22), (30, 48), (40, 68) and (50, 80). Reading across at 20, 40 and 60 gives the lower quartile about 18.6, the median about 27 and the upper quartile 36. The interquartile range is about 17.4.
Read across at n ÷ 4, n ÷ 2 and 3n ÷ 4, then down to the axis; these straight-line readings are approximate.
A histogram of ages of 120 people with unequal class widths. Frequency density = frequency / class width gives 2, 2.5, 3 and 1. The 10 to 30 class has most people (50) but the 30 to 40 bar is the tallest.
With unequal class widths the bar height is frequency density, and the bar area equals the frequency.
Six time classes with frequencies 6, 18, 36, 32, 20, 8 are shown as bars beside running-total bars that grow to 6, 24, 60, 92, 112, 120. The last total equals 120, the number of students.
Each running total adds the new class to the total so far, and the last must equal the total frequency.
Crosses at (0,0), (10,6), (20,24), (30,60), (40,92), (50,112), (60,120) are joined by one smooth S-shaped curve. The start at zero and the upper class boundary are highlighted.
Plot each running total at the upper class boundary, start at (0, 0) and join with a smooth curve.
On the curve for 120 students, reading across at 30, 60 and 90 and down gives LQ about 22, median 30 and UQ about 39 minutes. The IQR is 39 minus 22 = 17 minutes.
Read across at n/4, n/2 and 3n/4, then down; the IQR is UQ minus LQ.

Box plots and interquartile range (S4) – Higher tier only

A box plot shows five values: minimum, lower quartile (LQ), median, upper quartile (UQ) and maximum. The interquartile range (IQR = UQ − LQ) is the spread of the middle half of the data. It is not affected by extreme values.

Worked example 6. Data in order: 3, 5, 6, 8, 9, 10, 12, 13, 15, 18, 22 (n = 11).

Median: (11 + 1)/2 = 6th value = 10
LQ: (11 + 1)/4 = 3rd value = 6
UQ: 3 × (11 + 1)/4 = 9th value = 15
IQR = 15 − 6 = 9;  range = 22 − 3 = 19

Draw the box from 6 to 15 with a line at 10, and whiskers out to 3 and 22, above a scale.

To compare two box plots, compare the medians (average) and the IQRs (consistency), in context.

Eleven ordered values from 3 to 22 sit on a number line. The 3rd value 6 is the lower quartile, the 6th value 10 is the median and the 9th value 15 is the upper quartile. The box runs from 6 to 15 with a line at 10, and whiskers reach 3 and 22, so the IQR is 9 and the range is 19.
Pick the minimum, three quartiles and maximum from the ordered list and draw them on one scale.
Two curves on one grid: teachers have median 26 and IQR 11, students median 30 and IQR about 17. Bars show the middle half of each group. Teachers were faster and more consistent.
Compare a lower median (faster) and a smaller IQR (more consistent), in context.

Scatter graphs (S6)

A scatter graph plots pairs of values (bivariate data) to see if they are related.

  • Positive correlation: as one goes up, the other tends to go up.
  • Negative correlation: as one goes up, the other tends to go down.
  • No correlation: no clear pattern.
  • Strong correlation means the points lie close to a straight line; weak means they are more spread out.

A line of best fit is a straight line drawn by eye through the middle of the points, with roughly as many points above as below. Do not force it through the origin.

Worked example 7. Hours revised (x) and test score out of 100 (y) for 8 students: (2, 35), (3, 41), (4, 44), (5, 52), (6, 55), (7, 63), (8, 66), (9, 72). This shows strong positive correlation. The mean point is (5.5, 53.5), which is a good point for your line to pass through.

  • Using the line to estimate within the range of the data (e.g. 6.5 hours) is interpolation and is fairly reliable.
  • Using it outside the data (e.g. 20 hours) is extrapolation. The trend may not continue: a line through the mean point that fits these data predicts about 130 marks at 20 hours, which is impossible.

Correlation does not show causation. Ice-cream sales and sunburn cases both rise in summer. They are correlated because of a third factor, hot weather. Neither causes the other.

An outlier on a scatter graph is a point far from the pattern. Ignore it when drawing the line.

Hot weather is at the top. An arrow from hot weather leads to rising ice-cream sales and another arrow leads to rising sunburn cases. A dashed line between ice-cream sales and sunburn is crossed out: the two are correlated, but neither causes the other. The third factor, hot weather, causes both.
Ice-cream sales and sunburn rise together because of a third factor, hot weather, not because one causes the other.
A scatter graph of the 8 students: hours revised from 2 to 9 against test score, rising steadily (strong positive correlation). The mean point (5.5, 53.5) lies on the line of best fit. Reading at 6.5 hours gives about 59 marks, which is interpolation. Extending the line to 20 hours gives about 130 marks, which is extrapolation and impossible because the test is out of 100.
Inside the data the line is a fair guide; outside it the trend may not continue.
Three scatter diagrams: positive correlation (y rises with x), negative correlation (y falls as x rises) and zero correlation (no pattern, so no line of best fit).
Judge the correlation by eye before drawing any line.
A scatter diagram of revision hours and test score with a line of best fit through the mean point. Reading up from 6 hours gives a predicted score of about 67. The line is dashed beyond 8 hours where there is no data.
A line of best fit passes through the mean point and is used to predict inside the range of the data.
The line of best fit gives about 10.3 thousand dollars at age 6.5. A triangle shows a drop of 2.85 over 2 years, gradient -1.425, so the value falls about 1 400 dollars a year. Beyond age 9 the line is dotted in a shaded region marked outside the data; age 12 gives about 2.4, which is unreliable.
Readings inside the data are reliable; extending the line beyond age 9 is not.

Common errors

  • Drawing bar charts for continuous data with gaps, or histograms with frequency (not frequency density) on the vertical axis for unequal classes.
  • Dividing Σ(midpoint × frequency) by the number of classes instead of the total frequency.
  • Using the class width or upper boundary instead of the midpoint for an estimated mean.
  • Plotting cumulative frequency at midpoints instead of upper class boundaries.
  • Comparing only averages, or giving comparisons with no context.
  • Saying one variable “causes” the other because the correlation is strong.

Frequently asked questions

What is the difference between discrete and continuous data? Discrete data can only take separate values, usually counts such as the number of siblings. Continuous data is measured and can take any value in a range, such as height, time or mass.

What is the difference between primary and secondary data? Primary data is collected by you, for example through your own survey. Secondary data was collected by someone else, such as figures from a website or published table.

Does correlation mean one thing causes another? No. Correlation shows that two variables are related, but a third factor may link them. For example, ice-cream sales and sunburn cases both rise in summer.

What topics are in AQA GCSE Maths Topic 6 Statistics? Topic 6 Statistics (S1 to S6) in AQA GCSE Mathematics 8300 covers sampling, charts and diagrams, histograms and cumulative frequency, averages and spread, using samples to describe populations, and scatter graphs.

Next steps

Condense this into the revision notes, then test yourself with the practice questions. The Probability study guide covers the other half of this area of the course.

Official syllabus

AQA GCSE Mathematics (8300) specification, for teaching from September 2015, exams from May/June 2017, version 1.0, published by AQA – section 3.6 Statistics (S1 to S6).

Get free revision emails (optional)

Occasional emails with practice questions, worked explanations and links to free resources for the qualification and subjects you choose. No spam, and you can unsubscribe from any email. The free tools on this site never need an email.

Subjects (optional, up to 6)

Choose a qualification to see its subjects.

Related resources

Related articles

Studying this with a teacher

Working through Mathematics GCSE?

This page is free and stays free. If you would rather be taught it, Marlbridge runs Mathematics classes one-to-one and in small groups of up to 15, online in your own time zone. The first trial class is free. WhatsApp replies within an hour (8am–11pm Pakistan time, every day); email the same day.