Cumulative frequency diagrams
In a minute
A is a running total of the frequencies. Its graph is a , and from the curve you estimate the , the , the and any .
You'll be able to
- Build a cumulative frequency table and use it to count values.
- Draw a cumulative frequency diagram.
- Use the curve to estimate how many values are below or above a given value.
- Estimate the median, quartiles and interquartile range, and say what they show.
- Estimate a percentile, and the percentage of values below a given value.
Before you start2 quick questions on what this builds on
Revise first: 9.3 Averages and measures of spread
Cumulative frequency tables
A counts the values in each class. Often you want something else: how many values are up to a certain size.
Each running total belongs to the top of its class, the upper boundary. For , it counts every value up to .
Try it first, then reveal one step at a time.
- First class: runners took minutes or less.
- Add the next frequency: took minutes or less.
- Keep adding: , then , then .
- Check: the last total equals the number of runners, .
To count the values above a boundary, subtract its running total from the total frequency.
Try it first, then reveal one step at a time.
- parcels are kg or less.
- Subtract from the total: .
Drawing the diagram
A draws the running totals as a curve.
1. Plot each cumulative frequency at the upper boundary of its class.
2. Plot the start too: the lowest boundary, with cumulative frequency . No runner finished in minutes or less, so the race curve starts at , not .
3. Mark each point with a small cross. Join them with one smooth curve that never goes down.
Why the upper boundary? The total means runners had finished by minutes, the end of the class. It is not true at the middle, .
Try it first, then reveal one step at a time.
- The upper boundary of is . The running total there is .
- So plot . All the points are , , , , and .
To count from the curve, go up from a value to the curve, then across. The reading counts the values at or below it.
Try it first, then reveal one step at a time.
- Go up from minutes to the curve, then across. It reads about .
- So about runners took minutes or less.
- The rest took longer: .
Median, quartiles and interquartile range
For values, the is the middle one. On the curve, go across from , halfway up the total.
The cut the data into quarters. Go across from for the lower quartile (LQ) and for the upper quartile (UQ).
For a list you used . A curve gives only an estimate, so is fine.
The is UQ LQ, read on the horizontal axis. It is the spread of the middle half of the data.
Try it first, then reveal one step at a time.
- Half of the total: .
- Draw a line from across to the curve, then down. With g per small square, it lands at about g.
Try it first, then reveal one step at a time.
- Lower quartile: . Across from and down gives about g.
- Upper quartile: . Across from and down gives about g.
- IQR g.
- The middle half of the apples cover about g.
To compare two data sets, use an average and a spread. A higher median means larger values on average. A smaller IQR means more consistent values.
Percentiles
Percentiles cut the data into hundredths, so you can ask about any percentage.
The median is the th percentile. The lower and upper quartiles are the th and th percentiles.
Try it first, then reveal one step at a time.
- of is .
- Go across from to the curve, then down: about g.
- So about of the apples weigh g or less.
You can also go the other way: read the count at a value, then write it as a percentage of the total.
Try it first, then reveal one step at a time.
- Go up from g to the curve, then across: about apples.
- As a percentage: .
Common mistakes
- Plotting at the midpoint or lower boundary of a class. Plot at the upper boundary.
- Starting at when the first class starts higher. Start at the lowest boundary, with cumulative frequency .
- Going across from , and whatever the total. Use , and of the total frequency.
- Answering "how many are more than" with the reading. Subtract the reading from the total.
- Giving the IQR as two values. It is one value: UQ minus LQ, read on the horizontal axis.
Exam tips
- For **draw a cumulative frequency diagram**, the marks are for accurate points, including the start, and one smooth curve through them.
- For **use your graph to find an estimate**, draw your reading lines on the graph. A reading close to the true one earns the mark, so read carefully against the small squares.
- For **how many are more than** a value, show the reading and the subtraction from the total.
Key takeaways
Practise this
43 past-paper questions on 9.6 in the 0580 Extended bank, each with a mark scheme and a worked explanation.